Methodology_NN1 The_AT areal_JJ interpolation_NN1 problem_NN1 :_: estimating_VVG population_NN1 using_VVG remote_JJ sensing_NN1 in_II a_AT1 GIS_NN2 framework_NN1 Mitchel_NP1 Langford_NP1 ,_, David_NP1 J._NP1 Maguire_NP1 and_CC David_NP1 J._NP1 Unwin_NP1 Introduction_NN1 Data_NN integration_NN1 is_VBZ one_MC1 of_IO the_AT fundamental_JJ GIS_NN2 operations_NN2 (_( Burrough_NP1 1986_MC )_) ._. 
It_PPH1 involves_VVZ transformation_NN1 of_IO data_NN so_CS21 that_CS22 they_PPHS2 are_VBR reported_VVN at_II a_AT1 comparable_JJ geographical_JJ scale_NN1 ,_, projection_NN1 and_CC set_NN1 of_IO geographical_JJ units_NN2 ._. 
The_AT need_NN1 for_IF data_NN integration_NN1 arises_VVZ because_CS many_DA2 of_IO the_AT questions_NN2 which_DDQ scientists_NN2 and_CC social_JJ scientists_NN2 investigate_VV0 require_VV0 data_NN from_II a_AT1 wide_JJ range_NN1 of_IO sources_NN2 which_DDQ are_VBR only_RR reported_VVN on_II disparate_JJ spatial_JJ bases_NN2 ._. 
Data_NN integration_NN1 is_VBZ especially_RR a_AT1 problem_NN1 for_IF geographers_NN2 because_CS information_NN1 synthesis_NN1 is_VBZ at_II the_AT very_JJ heart_NN1 of_IO the_AT discipline_NN1 ._. 
The_AT difficulties_NN2 of_IO integrating_VVG data_NN are_VBR compounded_VVN by_II the_AT fact_NN1 that_CST there_EX are_VBR few_DA2 standards_NN2 governing_VVG the_AT way_NN1 in_II which_DDQ geographical_JJ data_NN are_VBR collected_VVN and_CC reported_VVN ._. 
Hearnshaw_VV0 et_RA21 al_RA22 ._. 
(_( 1989_MC )_) ,_, for_REX21 example_REX22 ,_, highlight_VV0 the_AT problem_NN1 in_II the_AT context_NN1 of_IO Leicestershire_NP1 ,_, a_AT1 county_NN1 in_II the_AT Midlands_NP1 of_IO England_NP1 ,_, and_CC show_VV0 the_AT difficulties_NN2 of_IO linking_VVG enumeration_NN1 district_NN1 (_( ED_NP1 )_) ,_, ward_NN1 ,_, parish_NN1 and_CC postcode_VV0 data_NN ._. 
Similarly_RR ,_, Openshaw_NP1 et_RA21 al_RA22 ._. 
(_( 1986_MC )_) report_VV0 that_CST the_AT data_NN in_II the_AT Domesday_NP1 System_NN1 are_VBR available_JJ for_IF 25_MC different_JJ and_CC incompatible_JJ types_NN2 of_IO areal_JJ unit_NN1 ._. 
Strategies_NN2 for_IF integrating_VVG data_NN reported_VVN at_II different_JJ geographical_JJ scales_NN2 and_CC for_IF different_JJ map_NN1 projections_NN2 have_VH0 received_VVN considerable_JJ attention_NN1 over_II the_AT past_JJ few_DA2 decades_NNT2 (_( Robinson_NP1 et_RA21 al_RA22 ._. 
1984_MC ;_; Burrough_NP1 1986_MC )_) ._. 
Although_CS there_EX are_VBR still_RR a_AT1 number_NN1 of_IO problems_NN2 to_TO be_VBI solved_VVN ,_, in_II contrast_NN1 to_II the_AT difficulties_NN2 of_IO comparing_VVG data_NN reported_VVN for_IF different_JJ spatial_JJ units_NN2 ,_, these_DD2 aspects_NN2 of_IO data_NN integration_NN1 are_VBR fairly_RR well_RR developed_VVN and_CC understood_VVN ._. 
In_II this_DD1 chapter_NN1 we_PPIS2 concentrate_VV0 on_II the_AT problem_NN1 of_IO integrating_VVG geographical_JJ data_NN reported_VVN for_IF different_JJ areal_JJ spatial_JJ units_NN2 ,_, one_MC1 of_IO the_AT most_RGT intractable_JJ of_IO all_DB data_NN integration_NN1 problems_NN2 ._. 
A_AT1 fuller_JJR description_NN1 of_IO the_AT process_NN1 of_IO data_NN integration_NN1 is_VBZ presented_VVN by_II Flowerdew_NP1 and_CC Green_JJ in_II Chapter_NN1 4_MC ._. 
Areal_JJ interpolation_NN1 There_EX have_VH0 been_VBN a_AT1 number_NN1 of_IO published_JJ attempts_NN2 to_TO provide_VVI solutions_NN2 to_II the_AT cross-area_JJ aggregation_NN1 ,_, or_CC '_GE areal_JJ interpolation_NN1 '_GE problem_NN1 (_( Goodchild_NP1 and_CC Lam_NP1 1980_MC ;_; Lam_NP1 1983_MC )_) ._. 
These_DD2 have_VH0 been_VBN summarized_VVN and_CC classified_VVN by_II Flowerdew_NP1 and_CC Openshaw_NP1 (_( 1987_MC )_) and_CC several_DA2 examples_NN2 are_VBR given_VVN in_II Hearnshaw_NP1 et_RA21 al_RA22 ._. 
(_( 1989_MC )_) ._. 
The_AT areal_JJ interpolation_NN1 problem_NN1 can_VM be_VBI defined_VVN as_II the_AT transfer_NN1 of_IO data_NN from_II one_MC1 set_NN1 (_( source_NN1 units_NN2 or_CC zones_NN2 )_) to_II a_AT1 second_MD set_NN1 (_( target_NN1 units_NN2 )_) of_IO overlapping_JJ ,_, non-hierarchical_JJ areal_JJ units_NN2 ._. 
Where_CS the_AT source_NN1 zones_NN2 nest_VV0 hierarchically_RR into_II the_AT target_NN1 zones_NN2 ,_, for_REX21 example_REX22 UK_NP1 administrative_JJ EDs_NN2 nest_VV0 exactly_RR in_II wards_NN2 ,_, transfer_NN1 of_IO data_NN from_II the_AT source_NN1 units_NN2 to_II the_AT target_NN1 units_NN2 is_VBZ one_MC1 of_IO simple_JJ aggregation_NN1 ._. 
However_RR ,_, where_CS the_AT units_NN2 are_VBR overlapping_VVG and_CC non-hierarchical_JJ the_AT problem_NN1 is_VBZ more_RGR complex_JJ and_CC it_PPH1 is_VBZ this_DD1 that_CST we_PPIS2 investigate_VV0 here_RL ._. 
The_AT nature_NN1 of_IO the_AT problem_NN1 can_VM be_VBI illustrated_VVN with_IW an_AT1 example_NN1 ._. 
A_AT1 common_JJ problem_NN1 in_II geographical_JJ information_NN1 systems_NN2 (_( GIS_NN2 )_) ,_, and_CC one_PN1 which_DDQ has_VHZ been_VBN known_VVN about_RP for_IF many_DA2 years_NNT2 in_II the_AT context_NN1 of_IO choropleth_VVZ mapping_NN1 ,_, is_VBZ that_DD1 of_IO producing_VVG maps_NN2 from_II population_NN1 data_NN aggregated_VVN over_II selected_JJ arbitrary_JJ areal_JJ units_NN2 ._. 
In_II the_AT UK_NP1 ,_, the_AT decennial_JJ Census_NN1 of_IO Population_NN1 records_VVZ the_AT number_NN1 of_IO individual_JJ people_NN and_CC households_NN2 ,_, but_CCB almost_RR all_DB population_NN1 data_NN are_VBR reported_VVN as_CSA various_JJ aggregations_NN2 which_DDQ ensure_VV0 that_CST data_NN about_II individuals_NN2 can_VM not_XX be_VBI recovered_VVN ._. 
Unfortunately_RR ,_, the_AT aggregations_NN2 are_VBR generally_RR chosen_VVN for_IF non-statistical_JJ reasons_NN2 and_CC so_RR differ_VV0 from_II data_NN set_VVN to_II data_NN set_VV0 and_CC are_VBR often_RR unstable_JJ over_II time_NNT1 ._. 
This_DD1 makes_VVZ any_DD comparative_JJ analysis_NN1 of_IO data_NN recorded_VVN at_II different_JJ times_NNT2 or_CC using_VVG different_JJ aggregation_NN1 units_NN2 problematical_JJ ._. 
The_AT difficulty_NN1 is_VBZ particularly_RR important_JJ in_II studies_NN2 where_CS it_PPH1 is_VBZ necessary_JJ to_TO find_VVI the_AT denominator_NN1 ,_, in_BCL21 order_BCL22 to_TO estimate_VVI the_AT incidence_NN1 of_IO a_AT1 property_NN1 as_II a_AT1 proportion_NN1 of_IO some_DD population_NN1 total_NN1 ._. 
For_REX21 example_REX22 ,_, the_AT numerator_NN1 might_VM be_VBI a_AT1 count_NN1 of_IO people_NN in_II a_AT1 postcode_NN1 sector_NN1 who_PNQS have_VH0 ,_, say_VV0 ,_, a_AT1 cancer_NN1 ,_, yet_RR ,_, because_CS the_AT census_NN1 and_CC health_NN1 data_NN are_VBR reported_VVN for_IF different_JJ areal_JJ units_NN2 ,_, it_PPH1 is_VBZ not_XX possible_JJ to_TO find_VVI the_AT appropriate_JJ at-risk_JJ population_NN1 for_IF the_AT denominator_NN1 and_CC so_RR compute_VV0 a_AT1 reliable_JJ incidence_NN1 ratio_NN1 ._. 
Although_CS this_DD1 example_NN1 relates_VVZ to_II what_DDQ have_VH0 been_VBN termed_VVN '_GE imposed_JJ '_GE areal_JJ units_NN2 ,_, that_CST have_VH0 no_AT landscape_NN1 reality_NN1 (_( Unwin_NP1 1981_MC )_) ,_, it_PPH1 is_VBZ increasingly_RR necessary_JJ to_TO create_VVI aggregations_NN2 of_IO population_NN1 over_NN1 '_GE natural_JJ '_GE areal_JJ units_NN2 such_II21 as_II22 a_AT1 soil_NN1 association_NN1 or_CC outcrop_NN1 of_IO a_AT1 particular_JJ rock_NN1 type_NN1 ._. 
The_AT solutions_NN2 to_II this_DD1 type_NN1 of_IO problem_NN1 utilize_VV0 either_RR point_VV0 interpolation_NN1 or_CC areal_JJ interpolation_NN1 methods_NN2 (_( Lam_NP1 1983_MC )_) and_CC these_DD2 will_VM be_VBI examined_VVN in_II turn_NN1 ._. 
The_AT point_NN1 interpolation_NN1 methods_NN2 essentially_RR use_VV0 a_AT1 point_NN1 ,_, usually_RR the_AT centroid_NN1 ,_, as_CSA a_AT1 surrogate_NN1 for_IF the_AT areal_JJ units_NN2 and_CC then_RT apply_VV0 conventional_JJ point_NN1 interpolation_NN1 methods_NN2 ._. 
The_AT crudest_JJT of_IO these_DD2 involves_VVZ matching_VVG the_AT centroids_NN2 of_IO source_NN1 and_CC target_NN1 units_NN2 by_II minimum_JJ distance_NN1 nearest_II neighbour_NN1 techniques_NN2 ._. 
A_AT1 slightly_RR less_RGR crude_JJ method_NN1 is_VBZ to_TO use_VVI the_AT centroids_NN2 to_TO generate_VVI a_AT1 continuous_JJ surface_NN1 by_II interpolation_NN1 ._. 
The_AT target_NN1 zones_NN2 are_VBR then_RT overlain_VVN and_CC the_AT interpolated_JJ value_NN1 is_VBZ transferred_VVN into_II the_AT zones_NN2 ._. 
However_RR ,_, this_DD1 technique_NN1 is_VBZ particularly_RR unsuitable_JJ for_IF population_NN1 estimation_NN1 because_CS as_CSA Tobler_NP1 (_( 1979_MC )_) points_VVZ out_RP ,_, there_EX is_VBZ a_AT1 danger_NN1 that_CST people_NN can_VM be_VBI created_VVN or_CC removed_VVD ._. 
Martin_NP1 (_( 1988_MC ,_, 1989_MC )_) describes_VVZ a_AT1 simple_JJ algorithm_NN1 that_CST uses_VVZ the_AT ED_NP1 centroids_NN2 ,_, with_IW a_AT1 spreading_JJ function_NN1 to_TO allocate_VVI people_NN to_II neighbouring_JJ grid_NN1 squares_NN2 ,_, which_DDQ incorporates_VVZ Tobler_NP1 's_GE idea_NN1 ._. 
An_AT1 intermediate_JJ method_NN1 between_II the_AT point_NN1 and_CC areal_JJ methods_NN2 uses_VVZ point-in-polygon_JJ techniques_NN2 (_( Burrough_NP1 1986_MC )_) to_TO locate_VVI the_AT centroids_NN2 of_IO the_AT source_NN1 units_NN2 in_II the_AT boundaries_NN2 of_IO the_AT target_NN1 units_NN2 ._. 
Although_CS relatively_RR unsophisticated_JJ ,_, this_DD1 method_NN1 is_VBZ reasonably_RR fast_JJ and_CC it_PPH1 seems_VVZ to_TO work_VVI satisfactorily_RR where_CS the_AT target_NN1 units_NN2 are_VBR much_RR larger_JJR than_CSN the_AT source_NN1 units_NN2 Also_RR in_II situations_NN2 where_RRQ the_AT target_NN1 units_NN2 are_VBR much_RR larger_JJR than_CSN the_AT source_NN1 units_NN2 ,_, polygon-in-polygon_JJ areal_JJ interpolation_NN1 techniques_NN2 can_VM be_VBI used_VVN to_TO obtain_VVI reasonable_JJ estimates_NN2 by_II locating_VVG the_AT source_NN1 units_NN2 within_II the_AT target_NN1 units_NN2 ._. 
A_AT1 simple_JJ decision_NN1 rule_NN1 can_VM be_VBI used_VVN to_TO decide_VVI what_DDQ to_TO do_VDI in_II the_AT case_NN1 of_IO source_NN1 units_NN2 which_DDQ cross_VV0 the_AT target_NN1 unit_NN1 boundaries_NN2 ._. 
The_AT most_RGT reliable_JJ and_CC accurate_JJ estimates_NN2 can_VM normally_RR be_VBI obtained_VVN using_VVG one_MC1 of_IO the_AT true_JJ cross-areal_JJ interpolation_NN1 methods_NN2 ._. 
Lam_NP1 (_( 1983_MC )_) suggests_VVZ that_CST there_EX are_VBR two_MC types_NN2 of_IO true_JJ areal_JJ interpolation_NN1 methods_NN2 :_: polygon_NN1 overlay_NN1 and_CC pycnophylactic_JJ interpolation_NN1 (_( Tobler_NP1 1979_MC )_) ._. 
These_DD2 are_VBR well_RR known_VVN and_CC have_VH0 been_VBN described_VVN elsewhere_RL (_( Lam_NP1 1983_MC ;_; Wagner_NP1 1989_MC )_) ._. 
The_AT above_JJ areal_JJ interpolation_NN1 methods_NN2 are_VBR alike_RR in_II that_DD1 ,_, other_II21 than_II22 the_AT purely_RR spatial_JJ information_NN1 given_VVN by_II the_AT boundaries_NN2 ,_, areas_NN2 or_CC centroids_NN2 of_IO the_AT areal_JJ units_NN2 involved_VVD ,_, no_AT extra_JJ information_NN1 is_VBZ used_VVN in_II interpolation_NN1 ._. 
In_II a_AT1 truly_RR integrated_JJ GIS_NN2 framework_NN1 (_( Jackson_NP1 and_CC Mason_NP1 1986_MC )_) it_PPH1 is_VBZ almost_RR certain_JJ to_TO be_VBI the_AT case_NN1 that_CST other_JJ potentially_RR useful_JJ information_NN1 is_VBZ available_JJ ._. 
This_DD1 additional_JJ information_NN1 can_VM be_VBI used_VVN for_IF '_GE intelligent_JJ '_GE interpolation_NN1 ._. 
Flowerdew_NP1 (_( 1988_MC )_) developed_VVD a_AT1 theoretically_RR sound_JJ and_CC very_RG general_JJ statistical_JJ approach_NN1 using_VVG Poisson_NP1 regression_NN1 which_DDQ incorporated_VVD an_AT1 additional_JJ binary_JJ variable_NN1 ._. 
He_PPHS1 predicted_VVD the_AT population_NN1 characteristics_NN2 of_IO parliamentary_JJ constituencies_NN2 in_II Lancashire_NP1 (_( the_AT target_NN1 units_NN2 )_) from_II district_NN1 level_NN1 data_NN (_( the_AT source_NN1 units_NN2 )_) using_VVG the_AT party_NN1 affiliation_NN1 of_IO the_AT constituency_NN1 's_GE Member_NN1 of_IO Parliament_NN1 as_II a_AT1 binary_JJ variable_NN1 ._. 
In_II this_DD1 chapter_NN1 ,_, we_PPIS2 develop_VV0 a_AT1 method_NN1 similar_JJ to_II Flowerdew_NP1 's_VBZ in_II which_DDQ we_PPIS2 use_VV0 GIS_NN2 techniques_NN2 to_TO enable_VVI areal_JJ interpolation_NN1 to_TO be_VBI informed_VVN by_II the_AT distribution_NN1 of_IO land-cover_JJ types_NN2 ,_, as_CSA inferred_VVN from_II a_AT1 classified_JJ Landsat_NP1 Thematic_JJ Mapper_NP1 (_( TM_NP1 )_) image_NN1 ,_, in_II both_DB2 the_AT source_NN1 (_( 1981_MC Census_NN1 wards_NN2 )_) and_CC target_NN1 (_( National_JJ Grid_NN1 kilometre_NNU1 squares_NN2 )_) units_NN2 ._. 
The_AT basic_JJ approach_NN1 The_AT basic_JJ idea_NN1 is_VBZ simple_JJ ._. 
A_AT1 suitable_JJ satellite_NN1 image_NN1 is_VBZ classified_VVN using_VVG image-processing_JJ techniques_NN2 to_TO identify_VVI the_AT various_JJ types_NN2 of_IO land_NN1 cover_NN1 which_DDQ exist_VV0 over_RP the_AT entire_JJ study_NN1 area_NN1 at_II very_RG fine_JJ spatial_JJ resolution_NN1 ._. 
Thus_RR ,_, for_IF each_DD1 census_NN1 ward_NN1 for_IF which_DDQ a_AT1 1981_MC population_NN1 total_NN1 exists_VVZ ,_, it_PPH1 is_VBZ possible_JJ to_TO determine_VVI the_AT number_NN1 of_IO pixels_NN2 classified_VVN as_II a_AT1 certain_JJ land-cover_JJ type_NN1 ._. 
It_PPH1 is_VBZ then_RT feasible_JJ to_TO express_VVI the_AT known_JJ ward_NN1 populations_NN2 as_II a_AT1 function_NN1 of_IO these_DD2 pixel_NN1 counts_NN2 ._. 
Provided_CS the_AT model_NN1 relating_VVG population_NN1 to_II these_DD2 land-cover_JJ data_NN is_VBZ reasonably_RR good_JJ ,_, the_AT resulting_JJ relationship_NN1 can_VM then_RT be_VBI used_VVN to_TO estimate_VVI populations_NN2 for_IF any_DD set_NN1 of_IO imposed_JJ or_CC natural_JJ areal_JJ units_NN2 that_CST can_VM also_RR be_VBI located_VVN on_II the_AT image_NN1 ._. 
The_AT use_NN1 of_IO remotely_RR sensed_VVN data_NN to_TO inform_VVI population_NN1 mapping_NN1 is_VBZ not_XX new_JJ ._. 
Previous_JJ attempts_NN2 to_TO estimate_VVI entire_JJ populations_NN2 ,_, especially_RR of_IO Third_MD World_NN1 cities_NN2 ,_, include_VV0 studies_NN2 by_II Ogrosky_NP1 (_( 1975_MC )_) ,_, Lo_UH and_CC Welch_NP1 (_( 1977_MC )_) and_CC Han_NP1 (_( 1985_MC )_) ._. 
Similarly_RR ,_, a_AT1 number_NN1 of_IO authors_NN2 have_VH0 used_VVN vertical_JJ aerial_NN1 photography_NN1 to_TO estimate_VVI residential_JJ densities_NN2 across_II individual_JJ cities_NN2 (_( see_VV0 ,_, for_REX21 example_REX22 ,_, Collins_NP1 and_CC El-Beik_NP1 1971_MC ;_; Hsu_NP1 1971_MC ;_; Clayton_NP1 and_CC Estes_NP1 1980_MC )_) while_CS others_NN2 have_VH0 experimented_VVN with_IW the_AT use_NN1 of_IO satellite_NN1 imagery_NN1 (_( Iisaka_NP1 and_CC Hegedus_NP1 1982_MC )_) ._. 
Convenient_JJ summaries_NN2 of_IO this_DD1 previous_JJ work_NN1 are_VBR to_TO be_VBI found_VVN in_II Lo_FW (_( 1986_MC )_) and_CC Griffiths_NP1 (_( 1988_MC )_) ._. 
The_AT work_NN1 described_VVN here_RL involves_VVZ four_MC distinct_JJ steps_NN2 ._. 
First_MD ,_, there_EX is_VBZ the_AT choice_NN1 of_IO satellite_NN1 image_NN1 and_CC the_AT production_NN1 of_IO a_AT1 land-cover_JJ map_NN1 ._. 
Second_MD ,_, there_EX is_VBZ the_AT overlay_NN1 on_II21 to_II22 this_DD1 map_NN1 of_IO the_AT 1981_MC Census_NN1 ward_NN1 boundaries_NN2 and_CC summation_NN1 of_IO pixels_NN2 falling_VVG into_II each_DD1 of_IO the_AT recognized_JJ land-cover_JJ types_NN2 ._. 
Third_MD ,_, there_EX is_VBZ the_AT statistical_JJ modelling_NN1 of_IO the_AT relationship_NN1 between_II population_NN1 and_CC land_NN1 cover_NN1 and_CC the_AT validation_NN1 of_IO the_AT resulting_JJ model_NN1 ._. 
Finally_RR ,_, at_II the_AT end_NN1 of_IO the_AT process_NN1 lies_VVZ the_AT estimation_NN1 of_IO the_AT population_NN1 over_II some_DD target_NN1 set_NN1 of_IO areal_JJ units_NN2 ,_, in_II our_APPGE case_NN1 the_AT kilometre_NNU1 squares_NN2 of_IO the_AT UK_NP1 National_JJ Grid_NN1 ._. 
Obtaining_VVG a_AT1 land-cover_JJ classification_NN1 A_ZZ1 Landsat_NP1 TM_NP1 image_NN1 of_IO Leicestershire_NP1 recorded_VVD on_II a_AT1 cloud-free_JJ day_NNT1 in_II July_NPM1 1984_MC constitutes_VVZ the_AT basic_JJ data_NN source_NN1 ._. 
This_DD1 area_NN1 was_VBDZ selected_VVN for_IF study_NN1 because_II21 of_II22 the_AT authors_NN2 '_GE familiarity_NN1 with_IW it_PPH1 ,_, the_AT adequate_JJ rural-urban_JJ contrast_NN1 and_CC the_AT availability_NN1 of_IO a_AT1 suitable_JJ image_NN1 ._. 
The_AT ground_NN1 resolution_NN1 of_IO a_AT1 TM_NP1 image_NN1 is_VBZ such_CS21 that_CS22 a_AT1 pixel_NN1 has_VHZ about_II a_AT1 30_MC m_NNO side_NN1 ,_, which_DDQ seems_VVZ appropriate_JJ for_IF the_AT scale_NN1 of_IO analysis_NN1 used_VVD ._. 
Indeed_RR ,_, it_PPH1 may_VM be_VBI that_CST an_AT1 image_NN1 of_IO higher_JJR spatial_JJ resolution_NN1 ,_, such_II21 as_II22 one_PN1 obtained_VVN by_II a_AT1 SPOT_NN1 satellite_NN1 (_( 20/10_MF m_NNU )_) ,_, would_VM prove_VVI counterproductive_JJ ,_, giving_VVG too_RG much_DA1 local_JJ variation_NN1 in_II pixel_NN1 reflectance_NN1 ._. 
The_AT full_JJ seven_MC bands_NN2 of_IO image_NN1 data_NN were_VBDR loaded_VVN into_II an_AT1 ERDAS_NN2 (_( Earth_NN1 Resources_NN2 Data_NN Analysis_NN1 System_NN1 )_) software_NN1 system_NN1 and_CC ground_NN1 control_NN1 point_NN1 information_NN1 entered_VVD to_TO rectify_VVI the_AT image_NN1 to_II National_JJ Grid_NN1 coordinates_NN2 ._. 
This_DD1 rectified_VVD image_NN1 was_VBDZ then_RT subsetted_VVN to_TO give_VVI an_AT1 area_NN1 covering_VVG the_AT 49_MC wards_NN2 that_CST make_VV0 up_RP the_AT districts_NN2 of_IO Oadby_NP1 and_CC Wigston_NP1 ,_, Leicester_NP1 ,_, and_CC Charnwood_NP1 in_II northern_JJ Leicestershire_NP1 (_( Fig._NN1 5.1_MC )_) ._. 
Since_CS a_AT1 faithful_JJ classification_NN1 of_IO land_NN1 cover_NN1 is_VBZ an_AT1 essential_JJ component_NN1 in_II the_AT proposed_JJ scheme_NN1 ,_, several_DA2 attempts_NN2 were_VBDR made_VVN in_II an_AT1 effort_NN1 to_TO obtain_VVI a_AT1 consistent_JJ and_CC suitably_RR detailed_JJ product_NN1 ._. 
An_AT1 unsupervised_JJ classification_NN1 provided_VVD an_AT1 acceptable_JJ separation_NN1 between_II rural_JJ and_CC urban_JJ cover_NN1 ,_, but_CCB finer_JJR detail_NN1 was_VBDZ concentrated_VVN in_II agricultural_JJ subdivisions_NN2 at_II the_AT expense_NN1 of_IO intra-urban_JJ differentiation_NN1 ._. 
Information_NN1 on_II agricultural_JJ land_NN1 use_NN1 is_VBZ unimportant_JJ to_II this_DD1 application_NN1 ,_, whereas_CS urban_JJ detail_NN1 illustrating_VVG housing_NN1 densities_NN2 or_CC alternative_JJ urban_JJ land_NN1 is_VBZ highly_RR desirable_JJ ._. 
A_AT1 second_MD classification_NN1 was_VBDZ made_VVN using_VVG a_AT1 supervised_JJ approach_NN1 so_CS21 that_CS22 we_PPIS2 could_VM capitalize_VVI on_II our_APPGE local_JJ knowledge_NN1 of_IO the_AT study_NN1 area_NN1 ._. 
The_AT supervised_JJ approach_NN1 enabled_VVD urban_JJ subdivision_NN1 of_IO dense_JJ and_CC less_RGR dense_JJ housing_NN1 ._. 
The_AT result_NN1 was_VBDZ an_AT1 improvement_NN1 over_II the_AT unsupervised_JJ product_NN1 ,_, and_CC yet_RR we_PPIS2 still_RR felt_VVD that_CST further_JJR urban_JJ differentiation_NN1 might_VM be_VBI possible_JJ ._. 
This_DD1 led_VVD us_PPIO2 to_TO undertake_VVI a_AT1 principal_JJ components_NN2 transformation_NN1 of_IO the_AT seven_MC bands_NN2 of_IO image_NN1 data_NN ,_, which_DDQ gave_VVD encouraging_JJ results_NN2 similar_JJ to_II those_DD2 obtained_VVN by_II Forster_NP1 (_( 1985a_FO )_) ._. 
As_II a_AT1 data_NN compression_NN1 exercise_NN1 ,_, the_AT first_MD four_MC components_NN2 of_IO the_AT transformation_NN1 reduced_VVD the_AT storage_NN1 requirement_NN1 by_II 42_MC per_NNU21 cent_NNU22 for_IF less_DAR than_CSN 1_MC1 per_NNU21 cent_NNU22 loss_NN1 in_II information_NN1 content_NN1 ._. 
However_RR ,_, of_IO great_JJ significance_NN1 was_VBDZ the_AT information_NN1 displayed_VVN in_II the_AT transformed_JJ images_NN2 ._. 
The_AT first_MD principal_JJ component_NN1 ,_, carrying_VVG 50_MC per_NNU21 cent_NNU22 of_IO the_AT original_JJ variance_NN1 ,_, was_VBDZ dominated_VVN by_II land-cover_JJ differences_NN2 in_II the_AT rural_JJ area_NN1 ._. 
The_AT second_MD component_NN1 (_( 40_MC per_NNU21 cent_NNU22 of_IO variance_NN1 )_) revealed_VVD the_AT urban_JJ distribution_NN1 very_RG clearly_RR with_IW a_AT1 strong_JJ rural/urban_JJ differentiation_NN1 ._. 
The_AT third_MD component_NN1 ,_, despite_II carrying_VVG only_RR 6_MC per_NNU21 cent_NNU22 of_IO the_AT original_JJ variance_NN1 ,_, revealed_VVD intra-urban_JJ differentiation_NN1 with_IW a_AT1 clarity_NN1 unseen_JJ in_II any_DD RGB_NP1 (_( red-green-blue_JJ (_( colour_NN1 )_) )_) ,_, composite_JJ taken_VVN from_II the_AT original_JJ bands_NN2 ._. 
The_AT fourth_MD component_NN1 (_( 2_MC per_NNU21 cent_NNU22 of_IO variance_NN1 )_) was_VBDZ dominated_VVN by_II the_AT thermal_JJ infra-red_JJ band_NN1 ,_, recording_VVG thermal_JJ characteristics_NN2 of_IO the_AT surface_NN1 at_II poor_JJ spatial_JJ resolution_NN1 ._. 
Finally_RR ,_, the_AT remaining_JJ three_MC components_NN2 seemed_VVD to_TO consist_VVI of_IO uninterpretable_JJ noise_NN1 ._. 
As_CSA previously_RR stated_VVN ,_, rural_JJ land_NN1 cover_NN1 was_VBDZ not_XX an_AT1 important_JJ consideration_NN1 ,_, and_CC so_RR the_AT first_MD component_NN1 image_NN1 was_VBDZ discarded_VVN ,_, together_RL with_IW the_AT last_MD three_MC ,_, leaving_VVG components_NN2 2_MC ,_, 3_MC and_CC 4_MC ._. 
When_CS these_DD2 components_NN2 were_VBDR displayed_VVN as_II an_AT1 RGB_NP1 composite_JJ they_PPHS2 continued_VVD to_TO show_VVI good_JJ urban_JJ differentiation_NN1 ,_, closely_RR matching_VVG both_DB2 our_APPGE local_JJ knowledge_NN1 of_IO the_AT area_NN1 and_CC the_AT ground_NN1 truth_NN1 data_NN that_CST were_VBDR collected_VVN ._. 
The_AT final_JJ attempt_NN1 at_II classification_NN1 involved_VVD a_AT1 supervised_JJ classification_NN1 of_IO principal_JJ components_NN2 ,_, 2_MC ,_, 3_MC and_CC 4_MC ._. 
This_DD1 produced_VVD the_AT land-cover_JJ classification_NN1 shown_VVN for_IF a_AT1 part_NN1 of_IO the_AT area_NN1 in_II Fig._NN1 5.2_MC ._. 
Other_JJ ,_, similar_JJ attempts_NN2 to_TO produce_VVI urban_JJ cover_NN1 classification_NN1 using_VVG satellite_NN1 data_NN are_VBR discussed_VVN in_II Jackson_NP1 et_RA21 al_RA22 ._. 
(_( 1980_MC )_) ,_, Forster_NP1 (_( 1983_MC ,_, 1985b_FO )_) ,_, Wheeler_NP1 (_( 1985_MC )_) ,_, Lam_NP1 et_RA21 al_RA22 ._. 
(_( 1987_MC )_) and_CC Griffiths_NP1 (_( 1988_MC )_) ._. 
In_II an_AT1 interesting_JJ variation_NN1 of_IO the_AT work_NN1 described_VVN here_RL ,_, Quarmby_NP1 et_RA21 al_RA22 ._. 
(_( 1988a_FO )_) outline_VV0 the_AT use_NN1 of_IO census_NN1 data_NN to_TO derive_VVI training_NN1 sets_VVZ for_IF classifying_VVG urban_JJ areas_NN2 into_II housing_VVG types_NN2 ._. 
Quarmby_NP1 et_RA21 al_RA22 ._. 
(_( 1988b_FO )_) also_RR discuss_VV0 the_AT use_NN1 of_IO SPOT_NN1 data_NN for_IF monitoring_VVG urban_JJ land-use_NN1 change_NN1 in_II south-east_ND1 England_NP1 ._. 
For_IF analytical_JJ purposes_NN2 ,_, the_AT 12_MC categories_NN2 of_IO land_NN1 cover_NN1 recognized_VVN during_II the_AT classification_NN1 stage_NN1 were_VBDR compressed_VVN into_II five_MC representing_VVG :_: 1_MC1 ._. 
Industry_NN1 and_CC commerce_NN1 (_( presumably_RR detected_VVN by_II large_JJ single_JJ buildings_NN2 and_CC different_JJ building_NN1 materials_NN2 )_) ;_; 2_MC ._. 
Dense_JJ residential_JJ areas_NN2 ;_; 3_MC ._. 
Ordinary_JJ residential_JJ areas_NN2 ;_; 4_MC ._. 
Areas_NN2 that_CST logically_RR have_VH0 no_AT population_NN1 (_( quarries_NN2 ,_, woods_NN2 ,_, water_NN1 bodies_NN2 )_) ;_; 5_MC ._. 
All_DB agricultural_JJ land_NN1 uses_NN2 ._. 
Overlaying_VVG census_NN1 ward_NN1 boundaries_NN2 In_II the_AT second_MD phase_NN1 of_IO the_AT work_NN1 ,_, digitized_VVD 1981_MC Census_NN1 ward_NN1 boundaries_NN2 and_CC population_NN1 totals_NN2 were_VBDR compiled_VVN using_VVG GIMMS_NP2 and_CC SASPAC_NP1 (_( at_II Manchester_NP1 Computer_NN1 Centre_NN1 )_) ._. 
These_DD2 data_NN were_VBDR transferred_VVN into_II ERDAS_NN2 and_CC rasterized_VVN ._. 
ERDAS_NP2 GIS_NN2 functions_NN2 were_VBDR used_VVN to_TO overlay_VVI the_AT ward_NN1 boundaries_NN2 on_II21 to_II22 the_AT classified_JJ image_NN1 and_CC then_RT count_VV0 the_AT number_NN1 of_IO pixels_NN2 of_IO each_DD1 recognized_VVD land-cover_JJ type_NN1 within_II each_DD1 ward_NN1 ._. 
The_AT result_NN1 was_VBDZ a_AT1 data_NN matrix_NN1 giving_VVG pixel_NN1 counts_VVZ for_IF five_MC land-cover_JJ types_NN2 together_RL with_IW the_AT recorded_JJ population_NN1 for_IF the_AT 49_MC wards_NN2 ._. 
These_DD2 data_NN are_VBR summarized_VVN in_II Table_NN1 5.1_MC and_CC Fig._NN1 5.3_MC is_VBZ a_AT1 map_NN1 of_IO the_AT population_NN1 density_NN1 ._. 
Towards_II a_AT1 statistical_JJ model_NN1 Table_NN1 5.2_MC shows_VVZ the_AT matrix_NN1 of_IO intercorrelations_NN2 between_II these_DD2 six_MC variables_NN2 ._. 
The_AT ward_NN1 population_NN1 ,_, the_AT response_NN1 variable_NN1 ,_, has_VHZ a_AT1 relatively_RR high_JJ positive_JJ correlation_NN1 with_IW the_AT '_GE industry_NN1 and_CC commerce_NN1 '_GE ,_, '_GE dense_JJ residential_JJ '_GE and_CC '_GE ordinary_JJ residential_JJ '_GE variables_NN2 and_CC low_JJ negative_JJ correlations_NN2 with_IW '_" nobody_PN1 '_GE and_CC '_GE agriculture_NN1 '_GE ._. 
There_EX are_VBR equally_RR logical_JJ intercorrelations_NN2 between_II the_AT various_JJ land-cover_JJ pixel_NN1 counts_NN2 which_DDQ are_VBR used_VVN as_II the_AT carrier_NN1 variables_NN2 ._. 
Our_APPGE intention_NN1 is_VBZ to_TO express_VVI the_AT ward_NN1 populations_NN2 as_CSA some_DD function_NN1 of_IO the_AT pixel_NN1 counts_VVZ for_IF each_DD1 land-cover_JJ type_NN1 ._. 
It_PPH1 should_VM be_VBI apparent_JJ that_CST there_EX is_VBZ a_AT1 very_RG large_JJ number_NN1 of_IO possible_JJ models_NN2 that_CST might_VM be_VBI used_VVN and_CC ,_, indeed_RR ,_, a_AT1 number_NN1 of_IO considerations_NN2 involved_JJ in_II making_VVG this_DD1 choice_NN1 (_( see_VV0 Dunn_NP1 1989_MC )_) ._. 
In_RR21 particular_RR22 ,_, three_MC such_DA considerations_NN2 influenced_VVN our_APPGE analysis_NN1 and_CC selection_NN1 of_IO carrier_NN1 variables_NN2 ._. 
First_MD ,_, there_EX is_VBZ a_AT1 conflict_NN1 between_II a_AT1 desire_NN1 to_TO maximize_VVI the_AT fit_NN1 obtained_VVN ,_, by_II inclusion_NN1 of_IO as_RG many_DA2 statistically_RR significant_JJ terms_NN2 as_CSA possible_JJ ,_, and_CC simple_JJ logic_NN1 which_DDQ suggests_VVZ that_CST the_AT correct_JJ form_NN1 of_IO any_DD fitted_JJ model_NN1 should_VM be_VBI a_AT1 linear_JJ weighted_JJ sum_NN1 of_IO only_RR those_DD2 land-cover_JJ types_NN2 that_CST actually_RR contain_VV0 mostly_RR housing_VVG (_( '_GE dense_JJ '_GE and_CC '_GE resid_NN1 '_GE )_) without_IW any_DD intercept_VV0 term_NN1 ._. 
In_II such_DA a_AT1 model_NN1 the_AT weights_NN2 themselves_PPX2 will_VM have_VHI a_AT1 direct_JJ interpretation_NN1 as_II the_AT average_JJ population_NN1 density_NN1 per_II Landsat_NP1 pixel_NN1 ._. 
Second_MD ,_, although_CS there_EX is_VBZ no_AT reason_NN1 to_TO expect_VVI any_DD model_NN1 to_TO be_VBI other_II21 than_II22 linear_JJ and_CC additive_JJ ,_, there_EX is_VBZ an_AT1 argument_NN1 ,_, presented_VVN by_II Flowerdew_NP1 (_( 1988_MC )_) ,_, which_DDQ says_VVZ that_CST since_CS the_AT response_NN1 variable_NN1 compromises_NN2 count_VV0 data_NN ,_, the_AT appropriate_JJ regression_NN1 model_NN1 should_VM combine_VVI a_AT1 Poisson_NP1 error_NN1 distribution_NN1 with_IW an_AT1 identity_NN1 link_NN1 function_NN1 (_( see_VV0 Aitkin_NP1 et_RA21 al_RA22 ._. 
1989_MC ;_; 217_MC )_) ._. 
Third_MD ,_, there_EX are_VBR a_AT1 number_NN1 of_IO criteria_NN2 for_IF goodness_NN1 of_IO fit_JJ ._. 
Although_CS the_AT usual_JJ coefficient_NN1 of_IO determination_NN1 or_CC scaled_JJ deviance_NN1 can_VM be_VBI used_VVN to_TO indicate_VVI the_AT global_JJ fit_NN1 of_IO any_DD specified_JJ model_NN1 ,_, it_PPH1 is_VBZ also_RR important_JJ to_TO examine_VVI model_NN1 performance_NN1 when_CS estimating_VVG populations_NN2 over_II areal_JJ units_NN2 other_II21 than_II22 the_AT wards_NN2 from_II which_DDQ the_AT models_NN2 were_VBDR derived_VVN ._. 
In_II this_DD1 study_NN1 ,_, this_DD1 was_VBDZ achieved_VVN by_II examining_VVG the_AT estimated_JJ kilometre_NNU1 square_JJ populations_NN2 together_RL with_IW those_DD2 which_DDQ were_VBDR reported_VVN for_IF the_AT same_DA areal_JJ units_NN2 in_II the_AT 1971_MC Census_NN1 of_IO Population_NN1 ._. 
Three_MC methods_NN2 were_VBDR examined_VVN in_II detail_NN1 using_VVG the_AT MINITAB_NN1 and_CC GLIM_VV0 packages_NN2 ._. 
A_ZZ1 '_GE Shotgun_NN1 '_GE model_NN1 The_AT overall_JJ regression_NN1 equation_NN1 ,_, calibrated_VVN by_II ordinary_JJ least_DAT squares_NN2 ,_, relating_VVG population_NN1 to_II the_AT carrier_NN1 variables_NN2 is_VBZ Table_NN1 5.3_MC gives_VVZ further_JJR details_NN2 ._. 
It_PPH1 will_VM be_VBI seen_VVN that_CST although_CS the_AT overall_JJ fit_JJ ,_, at_II 84_MC per_NNU21 cent_NNU22 is_VBZ good_JJ ,_, only_JJ industry_NN1 ,_, dense_JJ residential_JJ and_CC ordinary_JJ residential_JJ have_VH0 significant_JJ regression_NN1 constants_NN2 ._. 
The_AT same_DA set_NN1 of_IO variables_NN2 used_VVN in_II a_AT1 model_NN1 with_IW a_AT1 Poisson_NP1 error_NN1 term_NN1 gives_VVZ the_AT model_NN1 which_DDQ is_VBZ remarkably_RR similar_JJ ._. 
Figure_NN1 5.4_MC shows_VVZ the_AT standardized_JJ residuals_NN2 for_IF this_DD1 model_NN1 ._. 
Four_MC wards_NN2 are_VBR particularly_RR badly_RR fitted_VVN ._. 
In_II the_AT western_JJ central_JJ area_NN1 ,_, two_MC rural_JJ wards_NN2 (_( Charnwood_NP1 16_MC and_CC 17_MC )_) actually_RR have_VH0 almost_RR doubled_VVN the_AT fitted_JJ value_NN1 ._. 
In_II the_AT west_ND1 of_IO Leicester_NP1 's_GE urban_JJ area_NN1 ,_, Newton_NP1 ward_NN1 ,_, with_IW population_NN1 20_MC 440_MC ,_, has_VHZ a_AT1 fitted_JJ value_NN1 of_IO only_RR 13_MC 495_MC and_CC is_VBZ badly_RR fitted_VVN by_II all_DB the_AT models_NN2 examined_VVD ._. 
This_DD1 is_VBZ probably_RR due_II21 to_II22 the_AT large_JJ preponderance_NN1 of_IO low-rise_JJ multi-storey_JJ council_NN1 housing_NN1 ._. 
Finally_RR ,_, there_EX is_VBZ an_AT1 over-prediction_NN1 of_IO the_AT population_NN1 of_IO central_JJ Leicester_NP1 itself_PPX1 (_( The_AT Castle_NN1 ward_NN1 )_) ,_, which_DDQ is_VBZ another_DD1 feature_NN1 shown_VVN by_II all_DB the_AT models_NN2 ._. 
Although_CS this_DD1 Shotgun_NN1 model_NN1 gives_VVZ the_AT best_RRT fit_VV0 to_II the_AT observed_JJ ward_NN1 population_NN1 data_NN ,_, it_PPH1 is_VBZ logically_RR flawed_JJ in_II at_RR21 least_RR22 two_MC respects_NN2 ._. 
First_MD ,_, it_PPH1 is_VBZ evident_JJ that_CST the_AT correct_JJ form_NN1 of_IO any_DD model_NN1 linking_VVG population_NN1 to_TO land_VVI cover_NN1 should_VM not_XX have_VHI an_AT1 intercept_VV0 constant_JJ ;_; if_CS there_EX is_VBZ no_AT residential_JJ cover_NN1 ,_, then_RT there_EX should_VM be_VBI no_AT population_NN1 ._. 
Second_MD ,_, three_MC of_IO the_AT variables_NN2 incorporated_VVN into_II it_PPH1 have_VH0 negative_JJ signs_NN2 ,_, indicating_VVG that_CST as_II the_AT number_NN1 of_IO pixels_NN2 of_IO these_DD2 types_NN2 increases_VVZ ,_, so_RR population_NN1 decreases_VVZ ._. 
At_II first_MD sight_NN1 ,_, this_DD1 seems_VVZ reasonable_JJ ,_, but_CCB when_RRQ the_AT model_NN1 is_VBZ used_VVN to_TO estimate_VVI populations_NN2 of_IO any_DD other_JJ arbitrary_JJ areas_NN2 there_EX is_VBZ a_AT1 possibility_NN1 that_CST ,_, if_CS there_EX is_VBZ very_RG little_DA1 housing_NN1 ,_, it_PPH1 will_VM predict_VVI negative_JJ populations_NN2 ._. 
A_AT1 further_JJR point_NN1 is_VBZ that_CST the_AT Shotgun_NN1 model_NN1 includes_VVZ a_AT1 large_JJ number_NN1 of_IO carrier_NN1 variables_NN2 and_CC that_CST it_PPH1 is_VBZ worth_II further_JJR investigation_NN1 to_TO see_VVI if_CSW a_AT1 more_RGR parsimonious_JJ model_NN1 provides_VVZ a_AT1 satisfactory_JJ relationship_NN1 ._. 
A_ZZ1 '_GE Focused_JJ '_GE model_NN1 It_PPH1 is_VBZ thus_RR argued_VVN that_CST any_DD statistical_JJ model_NN1 linking_VVG pixel_NN1 counts_NN2 of_IO land_NN1 cover_NN1 to_II population_NN1 should_VM be_VBI simple_JJ ,_, linear_JJ ,_, additive_JJ and_CC without_IW any_DD intercept_VV0 constant_JJ ._. 
In_II such_DA a_AT1 model_NN1 ,_, the_AT individual_JJ coefficients_NN2 have_VH0 a_AT1 direct_JJ interpretation_NN1 as_II the_AT average_JJ density_NN1 of_IO people_NN in_II each_DD1 30_MC m_NNO square_JJ pixel_NN1 of_IO the_AT specified_JJ type_NN1 ._. 
The_AT ordinary_JJ least_DAT squares_NN2 focused_VVD model_NN1 is_VBZ pop_NN1 81_MC =_FO 9.89_MC dense_JJ +_FO 5.03_MC resid_JJ whereas_CS specification_NN1 of_IO Poisson_NP1 errors_NN2 gives_VVZ pop_NN1 81_MC =_FO 10.93_MC dense_JJ +_FO 4.82_MC resid_JJ Note_NN1 that_CST the_AT estimated_JJ population_NN1 densities_NN2 for_IF a_AT1 30_MC m_NNO square_JJ pixel_NN1 for_IF the_AT two_MC types_NN2 of_IO residential_JJ housing_NN1 identified_VVN using_VVG the_AT Landsat_NP1 data_NN seem_VV0 reasonable_JJ and_CC that_CST this_DD1 model_NN1 can_VM not_XX produce_VVI negative_JJ population_NN1 estimates_NN2 ._. 
In_II the_AT intercept_VV0 constant_JJ form_NN1 ,_, this_DD1 greatly_RR simplified_JJ model_NN1 has_VHZ an_AT1 R2_FO of_IO 81.8_MC per_NNU21 cent_NNU22 ._. 
The_AT reduction_NN1 to_II just_RR the_AT two_MC land-cover_JJ categories_NN2 ,_, concerned_JJ with_IW housing_NN1 ,_, loses_VVZ only_RR 3_MC per_NNU21 cent_NNU22 explanation_NN1 when_CS compared_VVN with_IW the_AT Shotgun_NN1 model._NNU for_IF these_DD2 re_II ;_; sons_NN2 ,_, the_AT Focused_JJ model_NN1 is_VBZ considered_VVN to_TO be_VBI superior_JJ to_II the_AT Shotgun_NN1 model_NN1 ._. 
Figure_NN1 5.5_MC shows_VVZ the_AT standardized_JJ residuals_NN2 for_IF the_AT least_DAT squares_NN2 Focused_VVD model_NN1 ._. 
The_AT fit_NN1 is_VBZ reasonable_JJ over_II most_DAT of_IO the_AT area_NN1 and_CC ,_, as_CSA might_VM be_VBI expected_VVN ,_, is_VBZ broadly_RR similar_JJ to_II that_DD1 given_VVN by_II the_AT Shotgun_NN1 model_NN1 ._. 
Again_RT it_PPH1 tends_VVZ to_TO under-predict_VVI values_NN2 for_IF wards_NN2 surrounding_VVG the_AT centre_NN1 of_IO Leicester_NP1 (_( Humberstone_NP1 ,_, Aylestone_NP1 and_CC Newton_NP1 )_) and_CC massively_RR over-predicts_VVZ for_IF The_AT Castle_NN1 ward_NN1 which_DDQ includes_VVZ the_AT city_NN1 centre_NN1 ._. 
Here_RL an_AT1 actual_JJ ward_NN1 population_NN1 of_IO 10_MC 149_MC is_VBZ predicted_VVN to_TO be_VBI 18_MC 373_MC ._. 
A_ZZ1 '_GE Simple_JJ '_GE model_NN1 The_AT final_JJ model_NN1 investigated_VVN is_VBZ the_AT simplest_JJT of_IO all_DB ._. 
In_II this_DD1 the_AT '_GE dense_JJ '_GE and_CC '_GE residential_JJ '_GE variables_NN2 (_( equivalent_JJ to_TO '_GE houses_NN2 '_GE )_) were_VBDR summed_VVN and_CC regressed_VVN against_II population_NN1 in_II a_AT1 simple_JJ linear_JJ model_NN1 ._. 
Figure_NN1 5.6_MC shows_VVZ that_CST the_AT plot_NN1 of_IO '_GE houses_NN2 '_GE against_II '_GE population_NN1 '_GE is_VBZ remarkably_RR linear_JJ ,_, with_IW a_AT1 correlation_NN1 coefficient_NN1 of_IO 0.869_MC ._. 
The_AT ordinary_JJ least_DAT squares_NN2 model_NN1 ,_, with_IW forced_JJ zero_NN1 intercept_VV0 term_NN1 ,_, is_VBZ pop_NN1 81_MC =_FO 5.41_MC houses_NN2 or_CC ,_, with_IW Poisson_NP1 errors_NN2 :_: pop_VV0 81_MC =_FO 5.40_MC houses_NN2 This_DD1 result_NN1 implies_VVZ that_CST each_DD1 pixel_NN1 classified_VVN as_CSA having_VHG people_NN living_VVG in_II it_PPH1 ,_, and_CC there_EX are_VBR 86_MC 178_MC of_IO these_DD2 ,_, will_NN1 on_II average_VVI contain_VV0 5.4_MC people_NN ._. 
Examined_VVN in_II detail_NN1 (_( Fig._NN1 5.7_MC )_) this_DD1 simple_JJ model_NN1 repeats_VVZ the_AT problems_NN2 of_IO fitting_VVG the_AT city_NN1 wards_NN2 identified_VVN by_II both_DB2 the_AT Shotgun_NN1 and_CC the_AT Focused_JJ models_NN2 ,_, but_CCB adds_VVZ to_II this_DD1 a_AT1 tendency_NN1 towards_II relative_JJ over-prediction_NN1 in_II the_AT rural_JJ wards_NN2 to_II the_AT north_ND1 and_CC east_ND1 of_IO the_AT area_NN1 ._. 
An_AT1 areal_JJ interpolation_NN1 application_NN1 The_AT final_JJ step_NN1 is_VBZ to_TO use_VVI the_AT results_NN2 of_IO these_DD2 models_NN2 to_TO predict_VVI the_AT distribution_NN1 of_IO population_NN1 over_II other_JJ arbitrary_JJ areal_JJ units_NN2 for_IF which_DDQ pixel_NN1 land-cover_NN1 counts_NN2 can_VM be_VBI obtained_VVN ._. 
A_AT1 particularly_RR severe_JJ test_NN1 is_VBZ to_TO use_VVI the_AT results_NN2 to_TO estimate_VVI the_AT population_NN1 of_IO the_AT area_NN1 over_II the_AT kilometre_NNU1 square_JJ grid_NN1 given_VVN by_II the_AT UK_NP1 Ordnance_NN1 Survey_NN1 ._. 
In_II this_DD1 ,_, the_AT estimation_NN1 not_XX only_RR involves_VVZ a_AT1 cross-area_JJ element_NN1 ,_, it_PPH1 also_RR embraces_VVZ a_AT1 considerable_JJ areal_JJ disaggregation_NN1 ,_, from_II an_AT1 average_JJ ward_NN1 size_NN1 of_IO 7.7_MC km&amp;sup2_FO ;_; (_( range_NN1 0.7438.7_MC km&amp;sup2_FO ;_; )_) down_RP to_II the_AT individual_JJ kilometre_NNU1 square_NN1 ._. 
Figures_NN2 5.85.10_MC have_VH0 been_VBN produced_VVN using_VVG UNIRAS_NN2 to_TO show_VVI the_AT predicted_JJ 1984_MC population_NN1 for_IF the_AT Shotgun_NN1 ,_, Focused_VVN and_CC Simple_JJ models_NN2 outlined_VVN above_RL ._. 
All_DB three_MC pick_VV0 out_RP the_AT major_JJ features_NN2 of_IO the_AT population_NN1 density_NN1 variation_NN1 across_II northern_JJ Leicestershire_NP1 but_CCB ,_, of_RR21 course_RR22 ,_, as_CSA they_PPHS2 stand_VV0 they_PPHS2 are_VBR impossible_JJ to_TO validate_VVI in_II any_DD direct_JJ way_NN1 ._. 
First_MD ,_, the_AT 1981_MC Census_NN1 of_IO Population_NN1 was_VBDZ not_XX made_VVN available_JJ for_IF grid_NN1 squares_NN2 and_CC ,_, second_NNT1 ,_, our_APPGE estimates_NN2 were_VBDR produced_VVN using_VVG a_AT1 1984_MC satellite_NN1 image_NN1 and_CC 1981_MC Census_NN1 ward_NN1 data_NN ._. 
There_EX are_VBR two_MC indirect_JJ ways_NN2 of_IO validating_VVG these_DD2 estimates_NN2 ._. 
First_MD ,_, they_PPHS2 can_VM be_VBI compared_VVN with_IW the_AT actual_JJ distribution_NN1 of_IO population_NN1 as_CSA reported_VVN on_II a_AT1 grid_NN1 square_NN1 basis_NN1 by_II the_AT 1971_MC Census_NN1 of_IO Population_NN1 (_( Fig._NN1 5.11_MC )_) ._. 
Figures_NN2 5.125.14_MC show_VV0 the_AT differences_NN2 between_II the_AT actual_JJ population_NN1 for_IF 1971_MC and_CC the_AT predicted_JJ population_NN1 for_IF 1984_MC ._. 
Table_NN1 5.4_MC provides_VVZ a_AT1 summary_NN1 of_IO the_AT overall_JJ fit_NN1 in_II the_AT form_NN1 of_IO the_AT root_NN1 mean_JJ squares_NN2 (_( RMS_NN1 )_) of_IO the_AT differences_NN2 for_IF each_DD1 map_NN1 ._. 
All_DB point_NN1 to_II precisely_RR the_AT same_DA problems_NN2 as_CSA were_VBDR identified_VVN when_CS evaluating_VVG the_AT original_JJ model_NN1 fits_NN2 ._. 
The_AT average_JJ ward_NN1 population_NN1 is_VBZ 9488_MC (_( Table_NN1 5.1_MC )_) and_CC so_RR the_AT RMS_NN1 errors_NN2 are_VBR comparatively_RR small_JJ (_( the_AT range_NN1 of_IO RMS_NN1 errors_NN2 is_VBZ from_II 815.13_MC for_IF the_AT Shotgun_NN1 ordinary_NN1 least_DAT squares_NN2 (_( OLS_NP2 )_) to_II 1035.29_MC for_IF the_AT Simple_JJ Poisson_NP1 )_) ._. 
There_EX is_VBZ little_DA1 difference_NN1 in_II the_AT values_NN2 obtained_VVD using_VVG OLS_NN2 and_CC Poisson_NP1 regression_NN1 and_CC the_AT Simple_JJ models_NN2 produce_VV0 only_RR slightly_RR higher_JJR errors_NN2 ._. 
Figure_NN1 5.12_MC shows_VVZ the_AT differences_NN2 given_VVN using_VVG the_AT Shotgun_NN1 model_NN1 which_DDQ has_VHZ the_AT lowest_JJT RMS_NN1 value_NN1 ._. 
Throughout_II most_DAT of_IO the_AT area_NN1 ,_, the_AT differences_NN2 are_VBR less_DAR than_CSN 1000_MC ._. 
In_II fact_NN1 ,_, only_RR 30_MC (_( or_CC 5.3_MC per_NNU21 cent_NNU22 )_) grid_NN1 squares_NN2 have_VH0 absolute_JJ differences_NN2 greater_JJR than_CSN this_DD1 ._. 
Unfortunately_RR ,_, some_DD urban_JJ squares_NN2 in_II Leicester_NP1 and_CC Loughborough_NP1 have_VH0 large_JJ individual_JJ differences_NN2 which_DDQ have_VH0 a_AT1 total_JJ range_NN1 from_II -4466_MC to_II +6976_MC ._. 
Thirteen_MC years_NNT2 ,_, between_II the_AT 1971_MC Census_NN1 and_CC the_AT 1984_MC image_NN1 used_VVD ,_, is_VBZ a_AT1 long_JJ time_NNT1 and_CC has_VHZ encompassed_VVN considerable_JJ population_NN1 change_NN1 in_II both_RR magnitude_NN1 and_CC distribution_NN1 ._. 
The_AT general_JJ tendency_NN1 of_IO this_DD1 model_NN1 to_TO overestimate_VVI central_JJ city_NN1 population_NN1 and_CC underestimate_VVI around_RP the_AT urban_JJ fringe_NN1 is_VBZ consistent_JJ with_IW the_AT known_JJ pattern_NN1 of_IO change_NN1 ._. 
Figure_NN1 5.13_MC shows_VVZ the_AT same_DA data_NN for_IF the_AT Focused_JJ model_NN1 which_DDQ ,_, although_CS it_PPH1 has_VHZ much_RR the_AT same_DA proportion_NN1 of_IO grid_NN1 squares_NN2 with_IW differences_NN2 in_II31 excess_II32 of_II33 1000_MC people_NN ,_, now_RT gives_VVZ a_AT1 total_JJ range_NN1 from_II -4799_MC to_II +8339_MC ._. 
Even_RR Fig._NN1 14_MC ,_, for_IF the_AT Simple_JJ model_NN1 ,_, has_VHZ only_RR 6.5_MC per_NNU21 cent_NNU22 of_IO squares_NN2 with_IW absolute_JJ differences_NN2 greater_JJR than_CSN 1000_MC but_CCB ,_, as_CSA might_VM be_VBI expected_VVN ,_, the_AT range_NN1 is_VBZ now_RT from_II -3506_MC to_II +11_MC 204_MC ._. 
A_AT1 second_MD indirect_JJ way_NN1 of_IO validating_VVG the_AT regressions_NN2 is_VBZ to_TO produce_VVI models_NN2 relating_VVG population_NN1 to_II 1981_MC land_NN1 cover_NN1 calibrated_VVN using_VVG the_AT 1971_MC populations_NN2 ._. 
If_CS the_AT general_JJ approach_NN1 is_VBZ fairly_RR robust_JJ ,_, it_PPH1 ought_VMK to_TO be_VBI that_CST the_AT relationship_NN1 between_II population_NN1 and_CC land_NN1 cover_NN1 is_VBZ fairly_RR stable_JJ ,_, so_CS21 that_CS22 we_PPIS2 would_VM expect_VVI the_AT model_NN1 coefficients_NN2 to_TO be_VBI similar_JJ to_II those_DD2 obtained_VVN from_II the_AT ward_NN1 data_NN ._. 
Table_NN1 5.5_MC presents_VVZ these_DD2 coefficients_NN2 ,_, estimated_VVD using_VVG ordinary_JJ least_DAT squares_NN2 ,_, for_IF exactly_RR equivalent_JJ models_NN2 to_II those_DD2 reported_VVN in_II section_NN1 5.6_MC ._. 
It_PPH1 can_VM be_VBI seen_VVN that_CST in_II every_AT1 case_NN1 the_AT two_MC sets_NN2 of_IO coefficients_NN2 have_VH0 the_AT same_DA sign_NN1 and_CC general_JJ magnitude_NN1 ._. 
However_RR ,_, even_RR allowing_VVG for_IF the_AT fact_NN1 that_CST the_AT population_NN1 and_CC satellite_NN1 data_NN were_VBDR collected_VVN 13_MC years_NNT2 apart_RL ,_, the_AT differences_NN2 are_VBR greater_JJR than_CSN we_PPIS2 would_VM have_VHI liked_VVN ._. 
Conclusions_NN2 and_CC further_RRR work_VV0 We_PPIS2 regard_VV0 these_DD2 results_NN2 as_CSA encouraging_JJ ._. 
They_PPHS2 suggest_VV0 that_CST it_PPH1 is_VBZ possible_JJ to_TO use_VVI satellite_NN1 remotely_RR sensed_VVN data_NN to_TO help_VVI inform_VVI population_NN1 mapping_VVG by_II allowing_VVG cross-area_JJ estimation_NN1 and_CC some_DD element_NN1 of_IO areal_JJ disaggregation_NN1 ._. 
Given_CS21 that_CS22 the_AT remotely_RR sensed_VVN data_NN add_VV0 a_AT1 great_JJ deal_NN1 of_IO information_NN1 to_II these_DD2 processes_NN2 this_DD1 is_VBZ hardly_RR surprising_JJ ,_, but_CCB this_DD1 general_JJ approach_NN1 is_VBZ relatively_RR easy_JJ to_TO carry_VVI out_RP in_II a_AT1 GIS_NN2 environment_NN1 ._. 
We_PPIS2 are_VBR currently_RR extending_VVG this_DD1 work_NN1 in_II three_MC ways_NN2 ._. 
We_PPIS2 are_VBR applying_VVG this_DD1 technique_NN1 in_II another_DD1 part_NN1 of_IO the_AT East_NP1 Midlands_NP2 ,_, as_CSA a_AT1 further_JJR test_NN1 of_IO the_AT utility_NN1 of_IO this_DD1 approach_NN1 ._. 
We_PPIS2 are_VBR working_VVG on_II other_JJ methods_NN2 of_IO classifying_VVG images_NN2 ,_, including_II textural_JJ classifiers_NN2 ,_, in_BCL21 order_BCL22 to_TO improve_VVI the_AT statistical_JJ modelling_NN1 ._. 
Finally_RR ,_, we_PPIS2 are_VBR trying_VVG to_TO obtain_VVI independent_JJ 1_MC1 km_NNU grid_NN1 square_NN1 estimates_VVZ for_IF the_AT 1981_MC population_NN1 ._. 
Error_NN1 propagation_NN1 :_: a_AT1 Monte_NP1 Carlo_NP1 simulation_NN1 Stan_NP1 Openshaw_NP1 ,_, Martin_NP1 Charlton_NP1 and_CC Steve_NP1 Carver_NP1 Introduction_NN1 The_AT absence_NN1 of_IO facilities_NN2 within_II GIS_NN2 software_NN1 for_IF handling_VVG the_AT effects_NN2 of_IO input_NN1 data_NN uncertainty_NN1 and_CC possible_JJ error_NN1 propagation_NN1 by_II GIS_NN2 operations_NN2 creates_VVZ a_AT1 question_NN1 mark_NN1 over_II the_AT safe_JJ utilization_NN1 of_IO many_DA2 aspects_NN2 of_IO the_AT technology_NN1 ._. 
The_AT problem_NN1 arises_VVZ because_CS it_PPH1 is_VBZ thought_VVN that_CST the_AT positional_JJ errors_NN2 and_CC attribute_VV0 uncertainties_NN2 which_DDQ are_VBR characteristic_JJ of_IO all_DB spatial_JJ databases_NN2 ,_, may_VM be_VBI propagated_VVN and_CC amplified_VVN by_II GIS_NN2 operations_NN2 and_CC thus_RR adversely_RR affect_VV0 some_DD or_CC all_DB applications_NN2 ._. 
These_DD2 input_VV0 data_NN uncertainties_NN2 are_VBR attributable_JJ to_II a_AT1 number_NN1 of_IO sources_NN2 ranging_VVG from_II errors_NN2 in_II the_AT original_JJ cartographic_JJ map_NN1 documents_VVZ through_RP to_II the_AT effects_NN2 of_IO the_AT GIS_NN2 operations_NN2 themselves_PPX2 ._. 
Specific_JJ types_NN2 of_IO error_NN1 in_II digital_JJ map_NN1 data_NN include_VV0 :_: locational_JJ uncertainty_NN1 ,_, measurement_NN1 errors_NN2 in_II digitization_NN1 ,_, map-scale-dependent_JJ accuracy_NN1 and_CC resolution_NN1 ,_, and_CC classification_NN1 and_CC generalization_NN1 error_NN1 in_II both_DB2 categorical_JJ and_CC remotely_RR sensed_VVN map_NN1 data_NN ._. 
For_IF thematic_JJ maps_NN2 the_AT principle_NN1 cause_NN1 of_IO error_NN1 is_VBZ the_AT original_JJ map_NN1 document_NN1 and_CC its_APPGE conversion_NN1 to_II digital_JJ form_NN1 ._. 
For_IF remotely_RR sensed_VVN data_NN error_NN1 depends_VVZ on_II the_AT accuracy_NN1 of_IO the_AT sensing_NN1 device_NN1 and_CC the_AT pixel_NN1 classification_NN1 technology_NN1 used_VVN in_II image_NN1 processing_NN1 ._. 
There_EX is_VBZ nothing_PN1 new_JJ about_II the_AT existence_NN1 of_IO map-related_JJ errors_NN2 ._. 
Error_NN1 and_CC uncertainty_NN1 have_VH0 always_RR been_VBN a_AT1 feature_NN1 of_IO cartographic_JJ information_NN1 ._. 
Previously_RR ,_, however_RR ,_, the_AT worst_JJT effects_NN2 were_VBDR largely_RR avoided_VVN by_II a_AT1 combination_NN1 of_IO the_AT expertise_NN1 of_IO the_AT cartographer_NN1 ,_, who_PNQS knew_VVD about_II the_AT inherent_JJ generalization_NN1 implicit_JJ in_II analogue_JJ maps_NN2 ,_, while_CS the_AT difficulty_NN1 of_IO manipulating_VVG maps_NN2 by_II manual_JJ means_NN precluded_VVD most_DAT forms_NN2 of_IO analysis_NN1 likely_JJ to_TO be_VBI sensitive_JJ to_II the_AT effects_NN2 of_IO error_NN1 ._. 
The_AT advent_NN1 of_IO GIS_NN2 has_VHZ significantly_RR changed_VVN all_DB this_DD1 ._. 
The_AT ease_NN1 of_IO use_NN1 and_CC flexibility_NN1 of_IO GIS_NN2 allow_VV0 the_AT user_NN1 to_TO perform_VVI operations_NN2 on_II map_NN1 data_NN that_CST were_VBDR previously_RR impossible_JJ on_II a_AT1 large_JJ scale_NN1 ._. 
The_AT typical_JJ end-user_NN1 of_IO GIS_NN2 output_NN1 will_VM probably_RR care_VVI or_CC know_VVI little_RR about_II the_AT cartographic_JJ and_CC uncertainty_NN1 characteristics_NN2 of_IO the_AT map_NN1 data_NN being_VBG used_VVN ,_, while_CS the_AT GIS_NN2 itself_PPX1 has_VHZ no_AT procedures_NN2 for_IF handling_VVG the_AT varying_JJ accuracy_NN1 and_CC reliability_NN1 of_IO the_AT digital_JJ map_NN1 data_NN being_VBG processed_VVN ._. 
A_ZZ1 GIS_NN2 gives_VVZ the_AT user_NN1 complete_JJ freedom_NN1 to_TO combine_VVI ,_, overlay_VV0 and_CC analyse_VVI data_NN from_II many_DA2 different_JJ sources_NN2 ,_, regardless_RR of_IO scale_NN1 ,_, accuracy_NN1 ,_, resolution_NN1 and_CC quality_NN1 of_IO the_AT original_JJ map_NN1 documents_NN2 and_CC without_IW any_DD regard_NN1 for_IF the_AT accuracy_NN1 characteristics_NN2 of_IO the_AT data_NN themselves_PPX2 ._. 
The_AT mixing_NN1 of_IO geographical_JJ information_NN1 from_II different_JJ map_NN1 scales_NN2 and_CC sources_NN2 is_VBZ a_AT1 key_JJ aspect_NN1 of_IO GIS_NN2 functionality_NN1 ,_, but_CCB it_PPH1 does_VDZ raise_VVI the_AT question_NN1 as_II21 to_II22 what_DDQ effects_VVZ the_AT combination_NN1 of_IO different_JJ levels_NN2 of_IO data_NN uncertainty_NN1 has_VHZ on_II both_DB2 the_AT output_NN1 maps_NN2 and_CC on_II the_AT data_NN derived_VVN from_II spatial_JJ query_NN1 and_CC analysis_NN1 ._. 
It_PPH1 must_VM be_VBI recognized_VVN that_CST there_EX are_VBR many_DA2 good_JJ reasons_NN2 for_IF wishing_VVG to_TO combine_VVI data_NN in_II these_DD2 ways_NN2 ,_, but_CCB a_AT1 major_JJ problem_NN1 arises_VVZ because_CS GIS_NN2 packages_NN2 fail_VV0 to_TO offer_VVI any_DD means_NN of_IO keeping_VVG track_NN1 of_IO the_AT effects_NN2 of_IO error_NN1 propagation_NN1 and_CC how_RRQ it_PPH1 affects_VVZ the_AT results_NN2 ._. 
Despite_II research_NN1 into_II some_DD aspects_NN2 of_IO the_AT error_NN1 propagation_NN1 issue_NN1 in_II spatial_JJ data_NN processing_NN1 (_( e.g._REX Blakemore_NP1 1984_MC ;_; Chrisman_NP1 1984_MC ;_; Drummond_NP1 1987_MC ;_; Goodchild_NP1 and_CC Dubuc_NP1 1987_MC ;_; Walsh_NP1 et_RA21 al_RA22 ._. 
1987_MC )_) ,_, Burrough_NP1 (_( 1986:103_MC )_) correctly_RR points_VVZ out_RP that_DD1 '_VBZ It_NN1 is_VBZ remarkable_JJ that_CST there_EX have_VH0 been_VBN so_RG few_DA2 studies_NN2 on_II the_AT whole_JJ problem_NN1 of_IO residual_JJ variation_NN1 and_CC how_RRQ errors_NN2 arise_VV0 ,_, or_CC are_VBR created_VVN and_CC propagated_VVN in_II geographical_JJ information_NN1 processing_NN1 ,_, and_CC what_DDQ the_AT effects_NN2 of_IO these_DD2 errors_NN2 might_VM be_VBI on_II the_AT results_NN2 of_IO studies_NN2 made_VVD ._. 
'_" This_DD1 neglect_NN1 is_VBZ widely_RR perceived_VVN to_TO be_VBI a_AT1 major_JJ unresolved_JJ problem_NN1 ._. 
Such_DA is_VBZ its_APPGE importance_NN1 that_CST the_AT National_JJ Center_NN1 for_IF Geographic_JJ Information_NN1 and_CC Analysis_NN1 (_( NCGIA_NP1 )_) in_II the_AT USA_NP1 ,_, has_VHZ placed_VVN this_DD1 issue_NN1 first_MD in_II its_APPGE list_NN1 of_IO research_NN1 priorities_NN2 (_( NCGIA_NP1 1989_MC )_) ._. 
This_DD1 chapter_NN1 is_VBZ concerned_JJ with_IW developing_JJ methods_NN2 able_JK to_TO provide_VVI estimates_NN2 of_IO the_AT confidence_NN1 regions_NN2 around_II GIS_NN2 map-based_JJ outputs_NN2 by_II taking_VVG into_II account_NN1 certain_JJ selected_JJ sources_NN2 of_IO uncertainty_NN1 affecting_VVG spatial_JJ databases_NN2 ._. 
A_ZZ1 Monte_NP1 Carlo_NP1 simulation-based_JJ approach_NN1 is_VBZ used_VVN as_II a_AT1 general_JJ means_NN of_IO estimating_VVG the_AT effects_NN2 of_IO input_NN1 data_NN uncertainty_NN1 on_II the_AT map_NN1 outputs_VVZ after_II an_AT1 arbitrary_JJ sequence_NN1 of_IO GIS_NN2 operations_NN2 ._. 
The_AT objective_NN1 is_VBZ to_TO identify_VVI and_CC handle_VVI the_AT effects_NN2 of_IO data_NN uncertainty_NN1 in_II a_AT1 GIS_NN2 by_II defining_VVG uncertainty_NN1 envelopes_NN2 to_TO create_VVI '_GE credibility_NN1 regions_NN2 '_GE around_II the_AT results_NN2 ._. 
This_DD1 is_VBZ considered_VVN to_TO be_VBI the_AT minimum_NN1 needed_VVN to_TO allow_VVI a_AT1 GIS_NN2 to_TO function_VVI in_II a_AT1 mixed_JJ data_NN environment_NN1 ._. 
Sources_NN2 of_IO error_NN1 in_II GIS_NN2 Error_NN1 and_CC uncertainty_NN1 are_VBR common_JJ features_NN2 of_IO cartographic_JJ information_NN1 ,_, so_CS it_PPH1 is_VBZ hardly_RR surprising_JJ that_CST these_DD2 aspects_NN2 are_VBR also_RR present_JJ in_II digital_JJ versions_NN2 of_IO analogue_JJ maps_NN2 ._. 
It_PPH1 follows_VVZ ,_, therefore_RR ,_, that_CST no_AT map-related_JJ spatial_JJ data_NN exist_VV0 which_DDQ are_VBR wholly_RR error-free_JJ ._. 
There_EX are_VBR many_DA2 different_JJ causes_NN2 of_IO uncertainty_NN1 and_CC those_DD2 which_DDQ are_VBR explicitly_RR due_II21 to_II22 GIS-based_JJ manipulations_NN2 of_IO geographic_JJ information_NN1 are_VBR merely_RR a_AT1 more_RGR recent_JJ problem_NN1 ._. 
However_RR ,_, it_PPH1 is_VBZ also_RR obvious_JJ that_CST the_AT power_NN1 of_IO GIS_NN2 has_VHZ the_AT potential_NN1 dramatically_RR to_TO increase_VVI both_DB2 the_AT magnitude_NN1 and_CC importance_NN1 of_IO errors_NN2 in_II spatial_JJ databases_NN2 ._. 
Burrough_NP1 (_( 1986_MC )_) identifies_VVZ three_MC main_JJ groups_NN2 of_IO factors_NN2 that_CST govern_VV0 the_AT errors_NN2 which_DDQ may_VM be_VBI associated_VVN with_IW spatial_JJ data_NN processing_NN1 ._. 
These_DD2 are_VBR :_: 1_MC1 ._. 
Obvious_JJ sources_NN2 of_IO error_NN1 ;_; ,_, 2_MC ._. 
Errors_NN2 resulting_VVG from_II natural_JJ variations_NN2 or_CC from_II ordinal_JJ measurements_NN2 ;_; and_CC 3_MC ._. 
Errors_NN2 arising_VVG through_II processing_NN1 ._. 
Group_NN1 1_MC1 errors_NN2 include_VV0 such_DA sources_NN2 of_IO age_NN1 as_CSA the_AT data_NN ,_, areal_JJ coverage_NN1 ,_, map_NN1 scale_NN1 and_CC density_NN1 of_IO observations_NN2 ._. 
Group_NN1 2_MC errors_NN2 include_VV0 positional_JJ accuracy_NN1 ,_, attribute_VV0 uncertainty_NN1 ,_, and_CC generalization_NN1 arising_VVG from_II data_NN classification_NN1 and_CC spatial_JJ variations_NN2 in_II map_NN1 quality_NN1 ._. 
Group_NN1 3_MC errors_NN2 include_VV0 those_DD2 arising_VVG through_II the_AT processing_NN1 of_IO geographical_JJ data_NN ,_, such_II21 as_II22 numerical_JJ computing_NN1 errors_NN2 ,_, faulty_JJ topological_JJ analyses_NN2 and_CC errors_NN2 in_II interpolation_NN1 ._. 
The_AT three_MC groups_NN2 represent_VV0 errors_NN2 of_IO increasing_JJ complexity_NN1 and_CC difficulty_NN1 of_IO handling_NN1 ,_, such_CS21 that_CS22 those_DD2 in_II groups_NN2 2_MC and_CC 3_MC require_VV0 intimate_JJ knowledge_NN1 of_IO the_AT data_NN ,_, their_APPGE structure_NN1 and_CC the_AT algorithms_NN2 used_VVN and_CC some_DD means_NN of_IO converting_VVG this_DD1 knowledge_NN1 into_II quantitative_JJ measures_NN2 of_IO impact_NN1 ._. 
It_PPH1 is_VBZ useful_JJ to_TO start_VVI by_II examining_VVG some_DD of_IO these_DD2 sources_NN2 of_IO error_NN1 ._. 
Digitizing_VVG error_NN1 Despite_II the_AT availability_NN1 of_IO hardware_NN1 for_IF the_AT automated_JJ conversion_NN1 of_IO geographic_JJ data_NN from_II paper_NN1 maps_NN2 to_II digital_JJ form_NN1 (_( e.g._REX optical_JJ scanners_NN2 )_) much_RR data_NN input_NN1 to_II GIS_NN2 is_VBZ still_RR done_VDN by_II hand_NN1 using_VVG a_AT1 digitizing_JJ table_NN1 ._. 
As_II a_AT1 result_NN1 of_IO human_NN1 and_CC other_JJ complicating_JJ factors_NN2 involved_VVD ,_, a_AT1 high_JJ level_NN1 of_IO error_NN1 is_VBZ often_RR present_JJ in_II digital_JJ map_NN1 data_NN ._. 
Manual_JJ digitizing_NN1 is_VBZ consequently_RR recognized_VVN as_II a_AT1 significant_JJ source_NN1 of_IO map_NN1 error_NN1 in_II GIS_NN2 (_( Otawa_NP1 1987_MC ;_; Keefer_NP1 et_RA21 al_RA22 ._. 
1988_MC )_) ._. 
However_RR ,_, error_NN1 introduced_VVN into_II digital_JJ map_NN1 databases_NN2 through_II the_AT digitizing_JJ process_NN1 is_VBZ often_RR ignored_VVN because_CS the_AT characteristics_NN2 of_IO digitizing_VVG error_NN1 have_VH0 not_XX been_VBN fully_RR defined_VVN and_CC because_CS no_AT practical_JJ means_NN of_IO handling_VVG input_NN1 data_NN uncertainty_NN1 exist_VV0 within_II proprietary_JJ GIS_NN2 software_NN1 ._. 
Sources_NN2 of_IO error_NN1 in_II the_AT digitizing_JJ process_NN1 can_VM be_VBI broken_VVN down_RP into_II two_MC main_JJ streams_NN2 :_: source_NN1 map_NN1 error_NN1 and_CC operational_JJ error_NN1 ._. 
Source_NN1 map_NN1 error_NN1 includes_VVZ the_AT accumulated_JJ error_NN1 of_IO the_AT map_NN1 being_VBG digitized_VVN ,_, while_CS operational_JJ error_NN1 includes_VVZ those_DD2 errors_NN2 propagated_VVN during_II the_AT digitizing_JJ process_NN1 itself_PPX1 ._. 
In_II looking_VVG at_II the_AT first_MD of_IO these_DD2 it_PPH1 must_VM be_VBI recognized_VVN that_CST no_AT map_NN1 ,_, however_RGQV detailed_JJ or_CC carefully_RR compiled_VVN ,_, can_VM perfectly_RR represent_VVI the_AT '_GE ground_NN1 truth_NN1 '_GE ,_, while_CS in_II the_AT second_MD instance_NN1 ,_, the_AT digitizing_JJ process_NN1 merely_RR serves_VVZ to_TO compound_VVI the_AT errors_NN2 present_VV0 in_II the_AT original_JJ map_NN1 (_( Poiker_NP1 1982_MC ;_; Blakemore_NP1 1984_MC )_) ._. 
Going_VVG back_RP to_II the_AT source_NN1 map_NN1 ,_, the_AT level_NN1 of_IO detail_NN1 and_CC accuracy_NN1 depends_VVZ very_RG much_DA1 on_II the_AT map_NN1 scale_NN1 ._. 
Large-scale_JJ maps_NN2 can_VM be_VBI topologically_RR very_RG detailed_JJ and_CC have_VH0 complex_JJ legends_NN2 while_CS smaller-scale_JJ maps_NN2 are_VBR more_RGR generalized_JJ ._. 
This_DD1 scale-related_JJ generalization_NN1 gives_VVZ rise_NN1 to_II both_DB2 locational_JJ errors_NN2 and_CC attribute_VV0 uncertainty_NN1 ._. 
Consider_VV0 line_NN1 thickness_NN1 on_II paper_NN1 maps_NN2 ._. 
All_DB lines_NN2 on_II maps_NN2 ,_, depicting_VVG features_NN2 such_II21 as_II22 roads_NN2 ,_, contours_NN2 and_CC boundaries_NN2 ,_, are_VBR drawn_VVN so_BCL21 as_BCL22 to_TO be_VBI easily_RR visible_JJ to_II the_AT user_NN1 ._. 
The_AT thickness_NN1 of_IO some_DD lines_NN2 are_VBR even_RR emphasized_VVN to_TO accentuate_VVI their_APPGE importance_NN1 (_( e.g._REX major_JJ roads_NN2 )_) ._. 
This_DD1 can_VM be_VBI misleading_JJ (_( e.g._REX a_AT1 road_NN1 shown_VVN as_II a_AT1 line_NN1 1_MC1 mm_NNU thick_JJ on_II a_AT1 1_MC1 :_: 50000_MC scale_NN1 map_NN1 implies_VVZ a_AT1 width_NN1 of_IO 50_MC m_NNO on_II the_AT ground_NN1 an_AT1 obvious_JJ exaggeration_NN1 )_) and_CC as_II a_AT1 result_NN1 leads_VVZ to_II difficulties_NN2 in_II the_AT digitizing_JJ process_NN1 ._. 
Common_JJ sense_NN1 suggests_VVZ that_CST the_AT true_JJ course_NN1 of_IO a_AT1 line_NN1 on_II a_AT1 map_NN1 is_VBZ along_II its_APPGE mid-point_NN1 ._. 
However_RR ,_, as_CSA the_AT operator_NN1 traces_VVZ a_AT1 line_NN1 extra_JJ errors_NN2 arise_VV0 because_CS it_PPH1 is_VBZ impossible_JJ to_TO follow_VVI the_AT centre_NN1 of_IO the_AT line_NN1 exactly_RR and_CC some_DD displacement_NN1 of_IO the_AT cursor_NN1 on_II either_DD1 side_NN1 of_IO the_AT line_NN1 is_VBZ inevitable_JJ ._. 
Burrough_NP1 (_( 1986_MC )_) suggests_VVZ that_CST the_AT area_NN1 of_IO the_AT map_NN1 covered_VVN by_II lines_NN2 can_VM be_VBI assumed_VVN to_TO be_VBI an_AT1 area_NN1 of_IO uncertainty_NN1 ._. 
In_II one_MC1 example_NN1 he_PPHS1 states_VVZ that_CST a_AT1 1_MC1 :_: 25_MC 000_MC scale_NN1 soil_NN1 map_NN1 measuring_VVG 400_MC by_II 600_MC mm_NNU may_VM have_VHI as_RG much_DA1 as_CSA 24_MC 000_MC mm_NNU of_IO drawn_VVN lines_NN2 ,_, covering_VVG 24_MC 000_MC mm&amp;sup2_FO ;_; or_CC 10_MC per_NNU21 cent_NNU22 of_IO the_AT total_JJ map_NN1 area_NN1 ._. 
Blakemore_NP1 (_( 1984_MC )_) attempts_VVZ to_TO estimate_VVI the_AT uncertainty_NN1 caused_VVN by_II cartographic_JJ line_NN1 thickness_NN1 and_CC associated_JJ digitizer_NN1 error_NN1 by_II adapting_VVG an_AT1 idea_NN1 forwarded_VVD originally_RR by_II Perkal_NP1 (_( 1966_MC )_) ._. 
Perkal_NP1 defined_VVD a_AT1 distance_NN1 (_( '_GE epsilon_NN1 '_GE )_) about_II a_AT1 cartographic_JJ line_NN1 as_II a_AT1 means_NN of_IO objective_JJ generalization_NN1 ._. 
Blakemore_NP1 inverted_VVD the_AT concept_NN1 ,_, suggesting_VVG that_CST it_PPH1 can_VM be_VBI used_VVN to_TO indicate_VVI an_AT1 error_NN1 band_NN1 about_II a_AT1 digitized_JJ line_NN1 ._. 
This_DD1 can_VM be_VBI applied_VVN to_II polygon_NN1 overlay_VV0 operations_NN2 to_TO ascribe_VVI descriptive_JJ levels_NN2 of_IO certainty_NN1 to_II the_AT resulting_JJ map_NN1 ._. 
A_AT1 point-in-polygon_JJ problem_NN1 was_VBDZ used_VVN as_II an_AT1 example_NN1 which_DDQ results_VVZ in_II five_MC classes_NN2 of_IO answer_NN1 depending_II21 on_II22 the_AT position_NN1 of_IO the_AT point_NN1 relative_II21 to_II22 the_AT digitized_JJ boundary_NN1 and_CC the_AT '_GE epsilon_NN1 '_GE distance_NN1 (_( Fig._NN1 6.1_MC )_) ._. 
Empirical_JJ investigation_NN1 using_VVG UK_NP1 Department_NN1 of_IO Industry_NN1 1_MC1 km_NNU square_JJ data_NN overlaid_VVN on_II polygon_NN1 boundaries_NN2 of_IO north-west_ND1 England_NP1 employment_NN1 office_NN1 areas_NN2 revealed_VVD that_CST only_RR 60_MC per_NNU21 cent_NNU22 of_IO the_AT points_NN2 in_II the_AT industry_NN1 database_NN1 could_VM be_VBI positively_RR assigned_VVN to_II an_AT1 employment_NN1 office_NN1 area_NN1 by_II being_VBG within_II a_AT1 polygon_NN1 and_CC away_II21 from_II22 the_AT boundary_NN1 by_II a_AT1 distance_NN1 greater_JJR than_CSN epsilon_NN1 ._. 
The_AT accuracy_NN1 of_IO a_AT1 digital_JJ representation_NN1 of_IO a_AT1 line_NN1 depends_VVZ not_XX only_RR on_II the_AT ability_NN1 of_IO the_AT person_NN1 using_VVG the_AT digitizer_NN1 to_TO follow_VVI the_AT centre_NN1 of_IO the_AT line_NN1 on_II the_AT map_NN1 exactly_RR but_CCB also_RR on_II the_AT number_NN1 of_IO points_NN2 they_PPHS2 input_VV0 to_TO describe_VVI the_AT shape_NN1 of_IO the_AT line_NN1 (_( Aldred_JJ 1972_MC )_) ._. 
Converting_VVG a_AT1 line_NN1 on_II a_AT1 map_NN1 into_II a_AT1 series_NN of_IO x_ZZ1 ,_, y_ZZ1 coordinates_NN2 involves_VVZ a_AT1 sampling_NN1 process_NN1 ._. 
The_AT number_NN1 of_IO sample_NN1 points_NN2 required_VVN accurately_RR to_TO copy_VVI a_AT1 straight_JJ line_NN1 is_VBZ much_DA1 less_DAR (_( 2_MC )_) than_CSN that_DD1 required_VVN for_IF a_AT1 curve_NN1 or_CC complex_JJ line_NN1 feature_NN1 (_( &gt;&gt;2_FO )_) ._. 
The_AT relative_JJ error_NN1 associated_VVN with_IW digitizing_VVG straight_JJ lines_NN2 (_( e.g._REX power_NN1 lines_NN2 )_) is_VBZ considerably_RR less_DAR than_CSN that_DD1 associated_VVN with_IW digitizing_VVG line_NN1 features_NN2 made_VVD up_RP of_IO complex_JJ curves_NN2 (_( e.g._REX coastlines_NN2 )_) ._. 
Although_CS a_AT1 number_NN1 of_IO methods_NN2 are_VBR available_JJ for_IF sampling_VVG only_RR those_DD2 vertices_NN2 which_DDQ describe_VV0 the_AT basic_JJ shape_NN1 of_IO the_AT line_NN1 (_( e.g._REX Douglas_NP1 and_CC Peucker_NP1 1973_MC )_) ,_, the_AT difference_NN1 between_II digitizing_VVG straight_JJ lines_NN2 and_CC complex_JJ curves_NN2 suggests_VVZ that_CST Blakemore_NP1 's_GE use_NN1 of_IO distance_NN1 '_GE epsilon_NN1 '_GE is_VBZ an_AT1 over-simplification_NN1 of_IO reality_NN1 ._. 
This_DD1 could_VM be_VBI improved_VVN by_II more_RGR detailed_JJ consideration_NN1 of_IO the_AT processes_NN2 of_IO error_NN1 propagation_NN1 inherent_JJ in_II digitizing_VVG procedures_NN2 ._. 
Several_DA2 empirical_JJ studies_NN2 have_VH0 examined_VVN digitizing_VVG error_NN1 by_II comparing_VVG digital_JJ data_NN to_II their_APPGE source_NN1 maps_NN2 (_( Traylor_NP1 1979_MC ;_; Otawa_NP1 1987_MC ;_; Keefer_NP1 et_RA21 al_RA22 ._. 
1988_MC )_) ._. 
From_II such_DA studies_NN2 appropriate_JJ models_NN2 of_IO digitizer_NN1 error_NN1 could_VM be_VBI formulated_VVN ._. 
Errors_NN2 in_II digital_JJ overlay_NN1 analysis_NN1 Much_DA1 of_IO the_AT functionality_NN1 of_IO GIS_NN2 lies_VVZ with_IW their_APPGE ability_NN1 to_TO overlay_VVI one_MC1 or_CC more_DAR digital_JJ maps_NN2 for_IF the_AT purposes_NN2 of_IO Boolean_NN1 or_CC network_NN1 analyses_NN2 ._. 
This_DD1 kind_NN1 of_IO map_NN1 analysis_NN1 used_VMK to_TO be_VBI done_VDN manually_RR (_( before_CS the_AT advent_NN1 of_IO practical_JJ GIS_NN2 )_) by_II overlaying_VVG transparent_JJ map_NN1 sheets_NN2 ,_, establishing_VVG the_AT required_JJ spatial_JJ relationships_NN2 and_CC drawing_VVG the_AT new_JJ map_NN1 on_II a_AT1 clean_JJ top_JJ sheet_NN1 with_IW felt_NN1 pens_NN2 (_( McHarg_NP1 1969_MC )_) ._. 
Digital_JJ cartography_NN1 promised_VVD a_AT1 more_RGR efficient_JJ and_CC flexible_JJ way_NN1 of_IO doing_VDG this_DD1 kind_NN1 of_IO work_NN1 ._. 
However_RR ,_, making_VVG the_AT step_NN1 between_II simple_JJ paper_NN1 overlays_VVZ and_CC digital_JJ overlay_NN1 proved_VVD difficult_JJ ._. 
Much_DA1 research_NN1 and_CC development_NN1 was_VBDZ carried_VVN out_RP into_II this_DD1 problem_NN1 during_II the_AT 1970s_MC2 (_( e.g._REX McAlpine_NP1 and_CC Cook_NP1 1971_MC ;_; Goodchild_NP1 1978_MC )_) ._. 
Results_NN2 from_II this_DD1 work_NN1 have_VH0 ,_, however_RR ,_, created_VVN more_DAR questions_NN2 about_II data_NN quality_NN1 and_CC error_NN1 propagation_NN1 ._. 
These_DD2 questions_NN2 need_VV0 to_TO be_VBI answered_VVN ,_, at_RR21 least_RR22 in_RR21 part_RR22 ,_, before_CS GIS_NN2 can_VM realize_VVI their_APPGE full_JJ potential_NN1 ._. 
McAlpine_NP1 and_CC Cook_NP1 (_( 1971_MC )_) experimented_VVD with_IW the_AT polygon_NN1 overlay_VV0 problem_NN1 by_II overlaying_VVG different_JJ sized_JJ hexagons_NN2 over_II a_AT1 hexagonal_JJ grid_NN1 with_IW random_JJ orientation_NN1 and_CC displacement_NN1 ._. 
Their_APPGE results_NN2 showed_VVD a_AT1 surprisingly_RR large_JJ proportion_NN1 of_IO small_JJ ,_, '_GE sliver_NN1 '_GE polygons_NN2 on_II the_AT output_NN1 maps_NN2 ._. 
When_CS applied_VVN to_II a_AT1 case_NN1 study_NN1 ,_, involving_VVG the_AT overlay_NN1 of_IO three_MC maps_NN2 of_IO Papua_NP1 New_NP1 Guinea_NP1 containing_VVG 7_MC ,_, 42_MC and_CC 101_MC polygons_NN2 respectively_RR ,_, the_AT output_NN1 map_NN1 was_VBDZ comprised_VVN of_IO 304_MC polygons_NN2 ,_, of_IO which_DDQ those_DD2 less_DAR than_CSN 3.8_MC km&amp;sup2_FO ;_; ;_; amounted_VVD to_II 38_MC per_NNU21 cent_NNU22 of_IO the_AT total_JJ area_NN1 ._. 
McAlpine_NP1 and_CC Cook_NP1 evaluated_VVD these_DD2 results_NN2 by_II classifying_VVG the_AT derived_JJ polygons_NN2 by_II size_NN1 and_CC boundary_NN1 complexity_NN1 ._. 
A_ZZ1 10_MC per_NNU21 cent_NNU22 random_JJ sample_NN1 of_IO derived_JJ polygons_NN2 was_VBDZ evaluated_VVN to_TO determine_VVI the_AT measure_NN1 of_IO agreement_NN1 between_II the_AT initial_NN1 and_CC derived_VVD descriptions_NN2 ._. 
It_PPH1 was_VBDZ found_VVN that_CST approximately_RR 30_MC per_NNU21 cent_NNU22 of_IO the_AT map_NN1 area_NN1 was_VBDZ made_VVN up_RP of_IO polygons_NN2 which_DDQ did_VDD not_XX agree_VVI with_IW the_AT initial_JJ map_NN1 descriptions_NN2 ._. 
Goodchild_NP1 (_( 1978_MC )_) suggests_VVZ that_CST the_AT number_NN1 of_IO derived_JJ polygons_NN2 is_VBZ more_RRR a_AT1 function_NN1 of_IO boundary_NN1 complexity_NN1 than_CSN the_AT actual_JJ number_NN1 of_IO polygons_NN2 ._. 
When_CS the_AT boundaries_NN2 of_IO overlaid_VVN polygon_NN1 networks_NN2 are_VBR highly_RR correlated_VVN (_( e.g._REX as_II21 for_II22 certain_JJ types_NN2 of_IO administrative_JJ areas_NN2 )_) serious_JJ problems_NN2 are_VBR created_VVN through_II the_AT introduction_NN1 of_IO large_JJ numbers_NN2 of_IO '_" spurious_58 '_GE sliver_NN1 polygons_NN2 ._. 
In_II the_AT case_NN1 of_IO administrative_JJ boundaries_NN2 ,_, different_JJ units_NN2 present_VV0 in_II different_JJ maps_NN2 often_RR share_VV0 a_AT1 common_JJ boundary_NN1 (_( e.g._REX a_AT1 river_NN1 or_CC stretch_NN1 of_IO coastline_NN1 )_) ._. 
These_DD2 common_JJ boundaries_NN2 will_VM have_VHI been_VBN digitized_VVN separated_JJ ,_, and_CC so_RR will_VM not_XX ,_, therefore_RR ,_, coincide_VV0 exactly_RR ._. 
Furthermore_RR ,_, the_AT more_RGR accurately_RR common_JJ boundaries_NN2 are_VBR digitized_VVN in_II each_DD1 map_NN1 and_CC the_AT more_DAR coordinates_NN2 are_VBR used_VVN ,_, then_RT the_AT larger_JJR the_AT number_NN1 of_IO spurious_JJ sliver_NN1 polygons_NN2 produced_VVN (_( Goodchild_NP1 1978_MC )_) ._. 
Elimination_NN1 procedures_NN2 are_VBR available_JJ in_II GIS_NN2 software_NN1 to_TO remove_VVI sliver_NN1 polygons_NN2 on_II the_AT basis_NN1 of_IO minimum_JJ area_NN1 (_( e.g._REX ARC/INFO_NN1 ,_, ESRI_NP1 1987_MC )_) ._. 
It_PPH1 is_VBZ likely_JJ ,_, however_RR ,_, that_CST the_AT resulting_JJ common_JJ boundary_NN1 will_VM be_VBI moved_VVN from_II its_APPGE true_JJ position_NN1 by_II the_AT elimination_NN1 procedure_NN1 ,_, thereby_RR introducing_VVG further_JJR uncertainty_NN1 into_II the_AT output_NN1 map_NN1 ._. 
An_AT1 attempt_NN1 to_TO estimate_VVI the_AT cumulative_JJ effect_NN1 of_IO thematic_JJ map_NN1 errors_NN2 in_II digital_JJ overlay_NN1 analysis_NN1 has_VHZ been_VBN made_VVN by_II Newcomer_NN1 and_CC Szajgin_NP1 (_( 1984_MC )_) ._. 
Using_VVG conditional_JJ probability_NN1 theory_NN1 they_PPHS2 suggest_VV0 that_CST the_AT accuracy_NN1 of_IO maps_NN2 resulting_VVG from_II overlay_NN1 analysis_NN1 is_VBZ determined_VVN by_II the_AT number_NN1 of_IO map_NN1 layers_NN2 ,_, their_APPGE accuracy_NN1 and_CC the_AT coincidence_NN1 of_IO errors_NN2 at_II the_AT same_DA position_NN1 in_II several_DA2 map_NN1 layers_NN2 ._. 
As_II a_AT1 result_NN1 the_AT accuracy_NN1 of_IO a_AT1 composite_JJ map_NN1 from_II overlay_NN1 analysis_NN1 is_VBZ generally_RR less_DAR than_CSN the_AT accuracy_NN1 of_IO the_AT least_RGT accurate_JJ map_NN1 layer_NN1 used_VVN (_( Newcomer_NN1 and_CC Szajgin_NP1 1984_MC )_) ._. 
Newcomer_NN1 and_CC Szajgin_NP1 go_VV0 on_RP to_TO define_VVI the_AT lower_JJR and_CC upper_JJ bounds_NN2 of_IO accuracy_NN1 in_II the_AT output_NN1 map_NN1 ._. 
The_AT upper_JJ bound_NN1 is_VBZ at_RR21 best_RR22 equal_JJ to_II the_AT accuracy_NN1 of_IO the_AT least_RGT accurate_JJ map_NN1 layer_NN1 (_( when_CS all_DB errors_NN2 in_II other_JJ layers_NN2 are_VBR coincident_JJ in_II their_APPGE location_NN1 )_) ,_, while_CS the_AT lower_JJR bound_NN1 results_NN2 when_RRQ the_AT errors_NN2 in_II each_DD1 map_NN1 layer_NN1 occur_VV0 at_II unique_JJ locations_NN2 ._. 
Other_JJ researchers_NN2 have_VH0 compared_VVN spatial_JJ databases_NN2 with_IW ground_NN1 truth_NN1 observations._NNU for_REX21 example_REX22 ,_, Walsh_NP1 et_RA21 al_RA22 ._. 
(_( 1987_MC )_) made_VVD an_AT1 assessment_NN1 of_IO the_AT error_NN1 inherent_JJ in_II Landsat_NP1 data_NN ,_, digital_JJ terrain_NN1 models_NN2 (_( DTMs_NP1 )_) and_CC digital_JJ thematic_JJ soil_NN1 maps_NN2 by_II comparing_VVG the_AT data_NN from_II these_DD2 sources_NN2 with_IW information_NN1 collected_VVN in_II the_AT field_NN1 ._. 
They_PPHS2 found_VVD that_CST the_AT error_NN1 ranged_VVN from_II 43_MC per_NNU21 cent_NNU22 for_IF the_AT soil_NN1 maps_NN2 to_II 83_MC per_NNU21 cent_NNU22 for_IF the_AT DTMs_NP1 ._. 
Using_VVG the_AT techniques_NN2 described_VVN by_II Newcomer_NN1 and_CC Szajgin_NP1 (_( 1984_MC )_) Walsh_NP1 et_RA21 al_RA22 ._. 
(_( 1987_MC )_) found_VVD that_CST the_AT combination_NN1 of_IO inherent_JJ and_CC operational_JJ error_NN1 ranged_VVN from_II 71_MC per_NNU21 cent_NNU22 (_( upper_JJ bound_NN1 )_) to_II 83_MC per_NNU21 cent_NNU22 (_( lower_RRR bound_VVN )_) ._. 
The_AT problem_NN1 is_VBZ knowing_VVG what_DDQ these_DD2 results_NN2 mean_VV0 in_II the_AT context_NN1 of_IO a_AT1 particular_JJ application_NN1 ._. 
Errors_NN2 associated_VVN with_IW vector_NN1 to_II raster_NN1 conversion_NN1 There_EX are_VBR generally_RR recognized_VVN to_TO be_VBI two_MC major_JJ sources_NN2 of_IO error_NN1 when_CS converting_VVG vector_NN1 maps_NN2 (_( either_RR paper_NN1 or_CC digital_JJ )_) into_II gridded_JJ or_CC raster_NN1 format_NN1 ._. 
These_DD2 are_VBR coding_VVG errors_NN2 and_CC topological_JJ mismatch_NN1 errors_NN2 ._. 
The_AT first_MD ,_, and_CC probably_RR most_RGT obvious_JJ ,_, source_NN1 of_IO error_NN1 is_VBZ associated_VVN with_IW the_AT problem_NN1 of_IO coding_VVG those_DD2 grid_NN1 cells_NN2 which_DDQ contain_VV0 parts_NN2 of_IO several_DA2 different_JJ vector_NN1 polygons_NN2 each_DD1 in_II similar_JJ proportions_NN2 ._. 
How_RRQ should_VM the_AT grid_NN1 cell_NN1 be_VBI coded_VVN ?_? 
This_DD1 problem_NN1 arises_VVZ because_CS each_DD1 grid_NN1 cell_NN1 can_VM only_RR have_VHI a_AT1 single_JJ attribute_NN1 value_NN1 and_CC because_CS the_AT chosen_JJ grid_NN1 dimensions_NN2 are_VBR too_RG large_JJ to_TO resolve_VVI the_AT spatial_JJ detail_NN1 required_VVN ._. 
One_MC1 answer_NN1 is_VBZ to_TO use_VVI a_AT1 finer_JJR grid_NN1 to_TO achieve_VVI finer_JJR detail_NN1 (_( Walsh_NP1 et_RA21 al_RA22 ._. 
1987_MC )_) ,_, but_CCB this_DD1 causes_VVZ a_AT1 corresponding_JJ increase_NN1 in_II the_AT size_NN1 of_IO the_AT raster_NN1 database_NN1 ._. 
The_AT second_MD source_NN1 of_IO error_NN1 is_VBZ the_AT problem_NN1 of_IO topological_JJ mismatch_NN1 when_CS a_AT1 polygon_NN1 map_NN1 is_VBZ represented_VVN by_II a_AT1 grid_NN1 ._. 
Again_RT this_DD1 can_VM be_VBI reduced_VVN by_II using_VVG a_AT1 finer_JJR grid_NN1 ,_, but_CCB problems_NN2 with_IW the_AT increased_JJ amount_NN1 of_IO data_NN storage_NN1 required_VVN still_RR occur_VV0 ._. 
Methodological_JJ outline_NN1 In_II conclusion_NN1 to_II their_APPGE paper_NN1 Walsh_NP1 et_RA21 al_RA22 ._. 
(_( 1987_MC )_) state_VV0 that_DD1 '_VBZ additional_JJ research_NN1 in_II inherent_JJ and_CC operational_JJ error_NN1 assessment_NN1 is_VBZ warranted_VVN ._. 
Research_NN1 in_II methods_NN2 for_IF cartographic_JJ display_NN1 of_IO errors_NN2 ,_, and_CC the_AT setting_NN1 of_IO statistical_JJ confidence_NN1 limits_NN2 of_IO databases_NN2 ,_, are_VBR particularly_RR required_JJ '_GE (_( Walsh_NP1 et_RA21 al_RA22 ._. 
1987_MC ;_; 1429_MC )_) ._. 
Expanding_VVG on_II this_DD1 it_PPH1 is_VBZ possible_JJ to_TO identify_VVI five_MC major_JJ tasks_NN2 in_II the_AT study_NN1 of_IO error_NN1 propagation_NN1 within_II GIS_NN2 ._. 
These_DD2 are_VBR :_: 1_MC1 ._. 
The_AT development_NN1 of_IO mathematical_JJ models_NN2 to_TO represent_VVI the_AT uncertainty_NN1 characteristics_NN2 of_IO digital_JJ map_NN1 databases_NN2 ;_; 2_MC ._. 
The_AT development_NN1 of_IO procedures_NN2 for_IF estimating_VVG the_AT effects_NN2 of_IO input_NN1 data_NN uncertainties_NN2 and_CC their_APPGE propagation_NN1 through_II GIS_NN2 ;_; 3_MC ._. 
The_AT application_NN1 of_IO these_DD2 models_NN2 and_CC techniques_NN2 to_II a_AT1 representative_JJ range_NN1 of_IO case_NN1 studies_NN2 to_TO derive_VVI empirical_JJ estimations_NN2 of_IO likely_JJ error_NN1 levels_NN2 in_II GIS_NN2 output_VV0 ;_; 4_MC ._. 
The_AT development_NN1 of_IO techniques_NN2 to_TO utilize_VVI output_NN1 data_NN uncertainty_NN1 estimates_NN2 ;_; and_CC 5_MC ._. 
The_AT incorporation_NN1 of_IO the_AT technology_NN1 as_CSA standard_JJ GIS_NN2 tools_NN2 ._. 
Task_NN1 1_MC1 is_VBZ of_IO critical_JJ importance_NN1 ._. 
A_AT1 means_NN of_IO modelling_VVG the_AT distribution_NN1 of_IO errors_NN2 is_VBZ needed_VVN ._. 
One_MC1 way_NN1 of_IO meeting_VVG this_DD1 need_NN1 is_VBZ to_TO take_VVI some_DD commonly_RR used_JJ map_NN1 data_NN sources_NN2 (_( both_DB2 analogue_JJ and_CC digital_JJ )_) and_CC perform_VV0 numerical_JJ experiments_NN2 on_II them_PPHO2 to_TO investigate_VVI the_AT effects_NN2 of_IO scale_NN1 and_CC map_NN1 resolution_NN1 on_II data_NN accuracy_NN1 ._. 
In_II doing_VDG so_RR there_EX are_VBR a_AT1 number_NN1 of_IO important_JJ areas_NN2 to_TO consider_VVI :_: mathematical_JJ modelling_NN1 of_IO the_AT distribution_NN1 of_IO positional_JJ errors_NN2 in_II digital_JJ maps_NN2 derived_VVN from_II source_NN1 maps_NN2 of_IO different_JJ scales_NN2 ;_; the_AT level_NN1 and_CC distribution_NN1 of_IO error_NN1 which_DDQ characterize_VV0 digital_JJ map_NN1 data_NN ;_; and_CC the_AT types_NN2 of_IO probability_NN1 models_NN2 which_DDQ best_RRT represent_VV0 these_DD2 errors_NN2 ._. 
The_AT long-term_JJ aim_NN1 of_IO this_DD1 research_NN1 should_VM be_VBI to_TO obtain_VVI a_AT1 reasonable_JJ representation_NN1 of_IO the_AT various_JJ errors_NN2 that_CST exist_VV0 ,_, although_CS practical_JJ considerations_NN2 would_VM limit_VVI the_AT scope_NN1 and_CC complexity_NN1 of_IO the_AT representations_NN2 used_VVD ._. 
It_PPH1 should_VM be_VBI possible_JJ to_TO obtain_VVI reasonable_JJ error_NN1 models_NN2 relating_VVG to_II positional_JJ accuracy_NN1 ,_, classification_NN1 errors_NN2 ,_, attribute_VV0 uncertainty_NN1 and_CC normal_JJ sampling_NN1 aspects_NN2 of_IO spatial_JJ data_NN ._. 
Such_DA an_AT1 approach_NN1 conforms_VVZ to_II Chrisman_NP1 's_GE concept_NN1 of_IO a_AT1 total_JJ error_NN1 model_NN1 as_II a_AT1 decomposable_JJ set_NN1 of_IO stochastic_JJ processes_NN2 which_DDQ operate_VV0 simultaneously_RR ._. 
As_CSA more_DAR components_NN2 of_IO the_AT total_JJ error_NN1 model_NN1 become_VV0 available_JJ as_II a_AT1 result_NN1 of_IO research_NN1 then_RT they_PPHS2 could_VM be_VBI adopted_VVN ._. 
As_II a_AT1 result_NN1 ,_, any_DD error-handling_NN1 and_CC estimation_NN1 procedures_NN2 should_VM be_VBI sufficiently_RR general_JJ as_CSA to_TO be_VBI able_JK to_TO incorporate_VVI any_DD new_JJ error_NN1 model_NN1 that_CST may_VM emerge_VVI at_II a_AT1 later_JJR date_NN1 ._. 
Task_NN1 2_MC may_VM be_VBI solved_VVN by_II using_VVG a_AT1 Monte_NP1 Carlo_NP1 approach_NN1 ._. 
The_AT generality_NN1 and_CC flexibility_NN1 of_IO such_DA a_AT1 procedure_NN1 are_VBR a_AT1 very_RG attractive_JJ feature_NN1 of_IO the_AT technology_NN1 ._. 
Here_RL ,_, a_AT1 Monte_NP1 Carlo_NP1 based_VVD simulation_NN1 procedure_NN1 for_IF estimating_VVG the_AT impact_NN1 of_IO error_NN1 in_II GIS_NN2 is_VBZ proposed_VVN and_CC developed_VVN ._. 
The_AT concept_NN1 is_VBZ simple_JJ and_CC universal_JJ in_II its_APPGE applicability_NN1 ._. 
It_PPH1 does_VDZ ,_, however_RR ,_, present_VVI a_AT1 number_NN1 of_IO operational_JJ difficulties_NN2 that_CST require_VV0 basic_JJ research_NN1 before_CS it_PPH1 can_VM be_VBI widely_RR used_VVN ._. 
The_AT procedure_NN1 is_VBZ to_TO simulate_VVI the_AT effects_NN2 of_IO input_NN1 data_NN uncertainty_NN1 by_II a_AT1 Monte_NP1 Carlo_NP1 approach_NN1 that_CST is_VBZ often_RR used_VVN in_II statistics_NN to_TO perform_VVI exact_JJ significance_NN1 tests_NN2 ._. 
Each_DD1 input_NN1 data_NN source_NN1 is_VBZ assumed_VVN to_TO be_VBI characterized_VVN by_II an_AT1 error_NN1 model_NN1 that_CST represents_VVZ reasonable_JJ estimates_NN2 of_IO the_AT levels_NN2 and_CC nature_NN1 of_IO the_AT data_NN uncertainty_NN1 thought_VVD to_TO be_VBI present_JJ ._. 
This_DD1 allows_VVZ the_AT map_NN1 data_NN to_TO be_VBI replaced_VVN by_II probability_NN1 distributions_NN2 of_IO known_JJ form_NN1 and_CC parameters_NN2 ._. 
A_AT1 single_JJ simulation_NN1 is_VBZ made_VVN by_II generating_VVG random_JJ numbers_NN2 from_II these_DD2 probability_NN1 distributions_NN2 and_CC adding_VVG them_PPHO2 to_II the_AT observed_JJ geographic_JJ coordinates_NN2 (_( defining_VVG point_NN1 ,_, line_NN1 or_CC area_NN1 features_NN2 )_) with_IW these_DD2 random_JJ values_NN2 ._. 
This_DD1 process_NN1 is_VBZ repeated_VVN for_IF each_DD1 source_NN1 of_IO data_NN input_NN1 ._. 
The_AT randomized_JJ input_NN1 map_NN1 data_NN are_VBR then_RT subject_II21 to_II22 an_AT1 arbitrary_JJ sequence_NN1 of_IO GIS_NN2 operations_NN2 ._. 
The_AT final_JJ output_NN1 results_NN2 are_VBR saved_VVN ._. 
The_AT entire_JJ process_NN1 is_VBZ repeated_VVN M_MC times_NNT2 (_( where_CS M_ZZ1 may_VM be_VBI up_RG21 to_RG22 100_MC )_) ._. 
Research_NN1 is_VBZ required_VVN to_TO determine_VVI the_AT appropriate_JJ value_NN1 of_IO M._NN1 If_CS the_AT GIS_NN2 output_NN1 is_VBZ merely_RR numeric_JJ ,_, then_RT the_AT distribution_NN1 of_IO M_NN1 results_NN2 gives_VVZ some_DD indication_NN1 of_IO the_AT effects_NN2 of_IO input_NN1 data_NN uncertainty_NN1 ._. 
If_CS the_AT result_NN1 is_VBZ a_AT1 map_NN1 ,_, then_RT the_AT total_JJ set_NN1 of_IO output_NN1 maps_NN2 can_VM be_VBI used_VVN to_TO draw_VVI confidence_NN1 intervals_NN2 or_CC '_GE credibility_NN1 regions_NN2 '_GE which_DDQ can_VM be_VBI overlaid_VVN on_II the_AT deterministic_JJ results_NN2 as_II a_AT1 visual_JJ indication_NN1 of_IO the_AT effects_NN2 of_IO data_NN uncertainty_NN1 ._. 
For_REX21 example_REX22 ,_, suppose_VV0 a_AT1 series_NN of_IO map_NN1 overlays_VVZ is_VBZ being_VBG used_VVN to_TO define_VVI suitable_JJ locations_NN2 for_IF a_AT1 facility_NN1 ._. 
The_AT certainty_NN1 of_IO the_AT output_NN1 map_NN1 being_VBG correct_JJ can_VM be_VBI estimated_VVN using_VVG Monte_NP1 Carlo_NP1 simulation_NN1 ._. 
In_II a_AT1 vector_NN1 GIS_NN2 the_AT set_NN1 of_IO M_ZZ1 different_JJ output_NN1 maps_NN2 would_VM be_VBI rasterized_VVN and_CC a_AT1 count_NN1 made_VVN of_IO the_AT frequency_NN1 that_CST each_DD1 cell_NN1 appears_VVZ in_II the_AT final_JJ map_NN1 ._. 
Those_DD2 cells_NN2 in_II locations_NN2 which_DDQ are_VBR little_RR affected_VVN by_II the_AT sensitivity_NN1 analyses_NN2 would_VM accumulate_VVI high_JJ counts_NN2 ._. 
They_PPHS2 could_VM then_RT be_VBI converted_VVN back_RP into_II vector_NN1 form_NN1 as_CSA polygon_NN1 data_NN and_CC superimposed_VVN on_II the_AT deterministic_JJ results_NN2 ._. 
There_EX is_VBZ a_AT1 useful_JJ analogy_NN1 here_RL with_IW Monte_NP1 Carlo_NP1 significance_NN1 testing_NN1 which_DDQ could_VM be_VBI used_VVN to_TO add_VVI a_AT1 further_JJR degree_NN1 of_IO inferential_JJ screening_NN1 to_II assessing_VVG the_AT descriptive_JJ significance_NN1 of_IO the_AT uncertainty_NN1 bands_VVZ in_II the_AT context_NN1 of_IO a_AT1 particular_JJ application_NN1 ._. 
This_DD1 simulation_NN1 procedure_NN1 is_VBZ totally_RR independent_JJ of_IO the_AT error_NN1 models_NN2 used_VVN and_CC the_AT nature_NN1 and_CC sequence_NN1 of_IO the_AT GIS_NN2 operations_NN2 employed_VVD ._. 
The_AT GIS_NN2 component_NN1 can_VM include_VVI all_DB manner_NN1 of_IO map_NN1 manipulation_NN1 ,_, evaluation_NN1 and_CC statistical_JJ procedures_NN2 ._. 
However_RR ,_, there_EX are_VBR a_AT1 number_NN1 of_IO basic_JJ research_NN1 tasks_NN2 that_CST need_VV0 to_TO be_VBI resolved_VVN before_II this_DD1 technique_NN1 can_VM be_VBI widely_RR applied_VVN ._. 
These_DD2 include_VV0 :_: defining_VVG suitable_JJ values_NN2 for_IF the_AT number_NN1 of_IO simulations_NN2 required_VVN in_II a_AT1 GIS_NN2 environment_NN1 rather_II21 than_II22 a_AT1 statistical_JJ hypothesis-testing_JJ one_PN1 ;_; assessing_VVG the_AT utility_NN1 of_IO kernel_NN1 estimators_NN2 as_II an_AT1 approximation_NN1 that_CST may_VM allow_VVI smaller_JJR numbers_NN2 of_IO simulations_NN2 to_TO be_VBI used_VVN ;_; investigating_VVG the_AT possibility_NN1 of_IO predicting_VVG the_AT final_JJ output_NN1 regions_NN2 without_IW performing_VVG large_JJ numbers_NN2 of_IO simulations_NN2 ;_; and_CC investigating_VVG possible_JJ hardware_NN1 solutions_NN2 to_TO speed_VVI up_RP the_AT simulation_NN1 process_NN1 ._. 
The_AT Monte_NP1 Carlo_NP1 simulation_NN1 technique_NN1 offers_VVZ an_AT1 effective_JJ means_NN of_IO identifying_VVG and_CC demonstrating_VVG the_AT effects_NN2 of_IO data_NN uncertainty_NN1 in_II a_AT1 number_NN1 of_IO case_NN1 studies_NN2 drawn_VVN from_II the_AT natural_JJ and_CC social_JJ sciences_NN2 ._. 
This_DD1 is_VBZ important_JJ as_II a_AT1 means_NN of_IO drawing_VVG end-user_NN1 attention_NN1 to_II the_AT problems_NN2 of_IO error_NN1 propagation_NN1 ._. 
The_AT resulting_JJ '_GE error_NN1 audits_NN2 '_GE would_VM also_RR provide_VVI a_AT1 platform_NN1 for_IF illustrating_VVG to_II vendors_NN2 the_AT importance_NN1 of_IO installing_VVG error_NN1 estimators_NN2 in_II GIS_NN2 software_NN1 ._. 
A_AT1 practical_JJ means_NN of_IO identifying_VVG approximate_JJ levels_NN2 of_IO output_NN1 uncertainty_NN1 also_RR requires_VVZ that_CST some_DD basic_JJ recommendations_NN2 are_VBR made_VVN about_II how_RRQ this_DD1 variability_NN1 can_VM be_VBI retained_VVN ,_, used_JJ and_CC passed_VVN on_RP to_II subsequent_JJ operations_NN2 and_CC applications_NN2 using_VVG the_AT data_NN ._. 
Some_DD US_NP1 research_NN1 groups_NN2 appear_VV0 to_TO be_VBI tackling_VVG part_NN1 of_IO this_DD1 problem_NN1 by_II tagging_VVG databases_NN2 with_IW error_NN1 information_NN1 ._. 
Attention_NN1 might_VM be_VBI better_RRR focused_VVN on_II more_RGR technical_JJ questions_NN2 such_II21 as_II22 how_RRQ this_DD1 error_NN1 information_NN1 may_VM be_VBI used_VVN in_II GIS_NN2 and_CC spatial_JJ analysis_NN1 ._. 
For_REX21 example_REX22 ,_, development_NN1 of_IO spatial_JJ retrieval_NN1 techniques_NN2 and_CC nearest_JJT neighbour_NN1 analyses_NN2 which_DDQ can_VM operate_VVI with_IW fuzzy_JJ data_NN ._. 
Other_JJ issues_NN2 relate_VV0 to_II investigating_VVG how_RRQ this_DD1 uncertainty_NN1 information_NN1 can_VM best_RRT be_VBI presented_VVN to_II the_AT user_NN1 ._. 
Finally_RR ,_, the_AT simulation_NN1 approach_NN1 and_CC associated_JJ error_NN1 models_NN2 should_VM be_VBI capable_JJ of_IO being_VBG incorporated_VVN into_II standard_JJ GIS_NN2 software_NN1 ._. 
The_AT principal_JJ problem_NN1 area_NN1 concerns_VVZ the_AT additional_JJ amount_NN1 of_IO computation_NN1 that_CST may_VM be_VBI necessary_JJ successfully_RR to_TO implement_VVI Monte_NP1 Carlo_NP1 based_VVD techniques_NN2 ._. 
It_PPH1 is_VBZ possible_JJ ,_, however_RR ,_, that_DD1 quick_JJ approximations_NN2 could_VM be_VBI devised_VVN that_CST would_VM reduce_VVI the_AT work-load_NN1 by_II a_AT1 factor_NN1 of_IO 5_MC ,_, while_CS improvements_NN2 in_II hardware_NN1 over_II the_AT next_MD few_DA2 years_NNT2 may_VM well_RR absorb_VVI the_AT rest_NN1 ._. 
Monte_NP1 Carlo_NP1 simulation_NN1 methodology_NN1 Some_DD preliminaries_NN2 Although_CS the_AT methodology_NN1 is_VBZ of_IO general_JJ applicability_NN1 ,_, it_PPH1 has_VHZ been_VBN developed_VVN here_RL using_VVG ARC/INFO_NN1 running_VVG under_II the_AT VMS_NN1 operating_NN1 system_NN1 on_II a_AT1 microVAX_NN1 2_MC ._. 
Some_DD of_IO the_AT terminology_NN1 ,_, therefore_RR ,_, relates_VVZ to_II the_AT ARC/INFO_NN1 GIS_NN2 software_NN1 (_( see_VV0 Table_NN1 6.1_MC )_) ._. 
The_AT basic_JJ sequence_NN1 of_IO operations_NN2 The_AT nature_NN1 of_IO the_AT simulation_NN1 methodology_NN1 is_VBZ outlined_VVN in_II Fig._NN1 6.2_MC ._. 
The_AT simulation_NN1 of_IO input_NN1 data_NN uncertainty_NN1 involves_VVZ replacing_VVG the_AT deterministic_JJ input_NN1 data_NN values_NN2 by_II those_DD2 from_II a_AT1 probability_NN1 distribution_NN1 that_CST reflects_VVZ an_AT1 appropriate_JJ error_NN1 model_NN1 ._. 
The_AT GIS_NN2 operations_NN2 are_VBR then_RT performed_VVN and_CC the_AT results_NN2 saved_VVN for_IF evaluation_NN1 ._. 
The_AT GIS_NN2 software_NN1 used_VVN in_II this_DD1 case_NN1 study_NN1 is_VBZ the_AT widely_RR used_JJ ARC/INFO_NN1 package_NN1 (_( ESRI_NP1 1987_MC )_) ._. 
The_AT methodology_NN1 is_VBZ operationalized_VVN in_II a_AT1 macro_NN1 which_DDQ calls_VVZ a_AT1 separate_JJ program_NN1 to_TO perturb_VVI the_AT input_NN1 data_NN without_IW changing_VVG the_AT topology_NN1 ._. 
This_DD1 is_VBZ achieved_VVN by_II accessing_VVG the_AT binary_JJ arc_NN1 files_NN2 directly_RR from_II a_AT1 FORTRAN_NP1 program_NN1 and_CC then_RT replacing_VVG the_AT deterministic_JJ values_NN2 by_II probabilistic_JJ ones_NN2 ,_, checking_VVG in_II the_AT process_NN1 that_CST the_AT implicit_JJ topology_NN1 is_VBZ retained_VVN ._. 
There_EX are_VBR a_AT1 number_NN1 of_IO approaches_NN2 to_II implementing_VVG the_AT methodology_NN1 in_II ARC/INFO_NN1 ._. 
The_AT sequence_NN1 of_IO events_NN2 is_VBZ :_: 1_MC1 ._. 
Perturb_VV0 the_AT arc_NN1 data_NN for_IF the_AT coverages_NN2 ,_, preserving_JJ polygon_NN1 topology_NN1 ;_; 2_MC ._. 
Carry_VV0 out_RP the_AT GIS_NN2 operations_NN2 on_II the_AT perturbed_JJ coverage_NN1 ;_; 3_MC ._. 
Rasterize_VV0 the_AT results_NN2 ;_; Steps_NN2 13_MC are_VBR repeated_VVN 100_MC times._NNU 4_MC ._. 
Calculate_VV0 frequencies_NN2 for_IF each_DD1 raster_NN1 ;_; and_CC 5_MC ._. 
Map_VV0 the_AT results_NN2 for_IF a_AT1 given_JJ probability_NN1 level_NN1 ._. 
Steps_NN2 13_MC are_VBR contained_VVN within_II the_AT macro_NN1 ,_, written_VVN in_II the_AT ARC/INFO_NN1 macro-language_NN1 ,_, AML_NP1 ._. 
Step_NN1 1_MC1 is_VBZ implemented_VVN in_II a_AT1 separate_JJ macro_NN1 ._. 
Step_NN1 4_MC is_VBZ implemented_VVN in_II a_AT1 VAX_NN1 DCL_MC command_NN1 procedure_NN1 ._. 
The_AT results_NN2 are_VBR then_RT mapped_VVN on_II a_AT1 micro_NN1 ,_, using_VVG a_AT1 Postscript_NP1 driver_NN1 (_( Adobe_NN1 Systems_NN2 1985_MC )_) of_IO the_AT authors_NN2 '_GE own_DA devising_VVG (_( step_NN1 5_MC )_) ._. 
The_AT first_MD major_JJ problem_NN1 is_VBZ to_TO access_VVI the_AT arc_NN1 data_NN ._. 
In_II ARC/INFO_NN1 ,_, the_AT binary_JJ arc/node_NN1 topology_NN1 information_NN1 is_VBZ kept_VVN in_II one_MC1 subdirectory_NN1 ,_, while_CS the_AT tables_NN2 which_DDQ access_VV0 these_DD2 attributes_NN2 are_VBR stored_VVN in_II another_DD1 subdirectory_NN1 ._. 
Initial_JJ attempts_NN2 used_VVD the_AT ARC_NP1 UNGENERATE_NP1 command_VV0 to_TO extract_VVI the_AT binary_JJ data_NN in_II ASCII_NN1 character_NN1 form_NN1 for_IF input_NN1 to_II the_AT FORTRAN_NP1 perturbation_NN1 program_NN1 ,_, followed_VVN by_II a_AT1 GENERATE_VV0 and_CC BUILD_VV0 to_TO re-create_VVI the_AT coverage_NN1 ._. 
There_EX are_VBR two_MC disadvantages_NN2 to_II this_DD1 approach_NN1 ;_; first_MD ,_, it_PPH1 is_VBZ enormously_RR time_NNT1 consuming_VVG and_CC second_NNT1 ,_, the_AT polygon_NN1 topology_NN1 of_IO the_AT input_NN1 coverage_NN1 is_VBZ not_XX preserved_VVN ,_, so_CS one_PN1 can_VM not_XX work_VVI back_RP to_II the_AT initial_JJ coverage_NN1 attributes_NN2 ._. 
The_AT easiest_JJT solution_NN1 is_VBZ to_TO access_VVI the_AT binary_JJ files_NN2 directly_RR ._. 
All_DB the_AT information_NN1 is_VBZ then_RT contained_VVN in_RP only_RR one_MC1 file_NN1 ._. 
The_AT ARC/INFO_NN1 programmer_NN1 's_GE guide_NN1 (_( Aronson_NP1 1985_MC )_) give_VV0 some_DD clues_NN2 as_II21 to_II22 what_DDQ to_TO expect_VVI ,_, but_CCB the_AT actual_JJ file_NN1 structure_NN1 can_VM only_RR be_VBI determined_VVN by_II inspection_NN1 ._. 
Fortunately_RR ,_, the_AT VAX_NN1 command_VV0 DIR/FULL_FU gives_VVZ enough_DD information_NN1 to_TO allow_VVI a_AT1 FORTRAN_NP1 program_NN1 to_TO be_VBI written_VVN which_DDQ can_VM read_VVI the_AT data_NN block_NN1 by_II block_NN1 ,_, and_CC then_RT replace_VV0 the_AT arc_NN1 coordinates_NN2 by_II randomized_JJ values_NN2 drawn_VVN from_II some_DD appropriate_JJ probability_NN1 distribution_NN1 ._. 
The_AT complexity_NN1 of_IO the_AT record_NN1 structure_NN1 provides_VVZ extra_JJ problems_NN2 for_IF coding_VVG the_AT perturbation_NN1 program_NN1 ._. 
Integer_NN1 and_CC real_JJ data_NN representations_NN2 are_VBR mixed_VVN together_RL in_II the_AT logical_JJ records_NN2 in_II the_AT data_NN ._. 
The_AT trick_NN1 is_VBZ to_TO recognize_VVI which_DDQ are_VBR which_DDQ ,_, and_CC deal_VVI with_IW them_PPHO2 accordingly_RR ._. 
The_AT header/trailer_NN1 and_CC topology_NN1 information_NN1 must_VM also_RR be_VBI preserved_VVN exactly_RR ,_, and_CC only_RR the_AT coordinate_NN1 data_NN perturbed._NNU finally_RR ,_, the_AT node_NN1 information_NN1 can_VM be_VBI perturbed_VVN only_RR once_RR ,_, which_DDQ involves_VVZ creating_VVG a_AT1 node_NN1 table_NN1 into_II which_DDQ the_AT perturbed_JJ nodes_NN2 are_VBR stored_VVN as_CSA they_PPHS2 are_VBR read_VVN in_RP ,_, and_CC from_II which_DDQ they_PPHS2 can_VM be_VBI retrieved_VVN in_II later_JJR references_NN2 to_II the_AT same_DA node_NN1 ._. 
The_AT GIS_NN2 overlay_VV0 operations_NN2 follow_VV0 next_MD ._. 
In_II this_DD1 case_NN1 ,_, the_AT operations_NN2 are_VBR a_AT1 series_NN of_IO overlays_NN2 ,_, designed_VVN to_TO remove_VVI infeasible_JJ areas_NN2 from_II the_AT base_NN1 coverage_NN1 ._. 
Lastly_RR the_AT polygon_NN1 data_NN in_II the_AT output_NN1 coverage_NN1 are_VBR rasterized_VVN and_CC the_AT rasterized_JJ coverage_NN1 stored_VVD ._. 
Rasterization_NN1 is_VBZ used_VVN to_TO provide_VVI an_AT1 easy_JJ way_NN1 of_IO visualizing_VVG the_AT uncertainty_NN1 in_II the_AT output_NN1 results_NN2 ._. 
The_AT object_NN1 of_IO the_AT final_JJ analysis_NN1 is_VBZ to_TO create_VVI a_AT1 data_NN set_VV0 which_DDQ contains_VVZ ,_, for_IF each_DD1 raster_NN1 ,_, the_AT number_NN1 of_IO times_NNT2 it_PPH1 is_VBZ simulated_VVN to_TO be_VBI inside_II the_AT areas_NN2 resulting_VVG from_II the_AT sequence_NN1 of_IO GIS_NN2 operations_NN2 ._. 
There_EX are_VBR several_DA2 ways_NN2 of_IO achieving_VVG this_DD1 ._. 
The_AT methodology_NN1 used_VVD here_RL is_VBZ a_AT1 three-stage_JJ operation_NN1 ;_; first_MD ,_, the_AT binary_JJ raster_NN1 files_NN2 or_CC single_JJ variable_NN1 files_NN2 (_( SVFs_NP1 )_) are_VBR expanded_VVN into_II character_NN1 form_NN1 ,_, compressed_VVN (_( Held_VVN 1983_MC )_) ,_, and_CC a_AT1 row_NN1 number_NN1 added_VVD ;_; second_MD ,_, the_AT files_NN2 area_NN1 is_VBZ sorted_VVN by_II the_AT row_NN1 number_NN1 so_CS21 that_CS22 all_DB records_NN2 with_IW the_AT same_DA row_NN1 number_NN1 are_VBR contiguous_JJ in_II the_AT file_NN1 ;_; third_MD ,_, a_AT1 count_NN1 is_VBZ made_VVN of_IO all_DB the_AT rasters_NN2 in_II each_DD1 row_NN1 which_DDQ are_VBR simulated_VVN to_TO be_VBI inside_II the_AT required_JJ areas_NN2 ._. 
These_DD2 frequencies_NN2 can_VM then_RT be_VBI mapped_VVN to_TO yield_VVI credibility_NN1 regions_NN2 corresponding_VVG to_II particular_JJ significance_NN1 levels_NN2 ._. 
A_AT1 case_NN1 study_NN1 Searching_VVG for_IF radwaste_NN1 sites_NN2 Openshaw_NP1 et_RA21 al_RA22 ._. 
(_( 1989_MC )_) discuss_VV0 the_AT use_NN1 of_IO ARC/INFO_NN1 to_TO identify_VVI feasible_JJ sites_NN2 for_IF radwaste_NN1 dumps_NN2 ._. 
The_AT GIS_NN2 operations_NN2 constitute_VV0 little_DA1 more_DAR than_CSN a_AT1 sequence_NN1 of_IO map_NN1 overlays_VVZ in_II the_AT form_NN1 of_IO a_AT1 Boolean_NN1 search_NN1 ._. 
This_DD1 process_NN1 is_VBZ re-examined_VVN here_RL with_IW particular_JJ attention_NN1 to_II northern_JJ England_NP1 to_TO reduce_VVI the_AT computational_JJ resources_NN2 required_VVN ._. 
The_AT object_NN1 of_IO the_AT overlay_NN1 exercise_NN1 is_VBZ to_TO determine_VVI potentially_RR feasible_JJ sites_NN2 for_IF the_AT dumping_NN1 of_IO waste_NN1 material_NN1 from_II the_AT nuclear_JJ industry_NN1 ,_, in_RR21 particular_RR22 ,_, low-_JJ and_CC intermediate-level_JJ radioactive_JJ waste_NN1 ._. 
This_DD1 chapter_NN1 is_VBZ not_XX itself_PPX1 concerned_JJ with_IW the_AT mechanics_NN2 or_CC politics_NN1 of_IO the_AT matter_NN1 ,_, the_AT example_NN1 chosen_VVN is_VBZ merely_RR a_AT1 convenient_JJ one_PN1 for_IF purposes_NN2 of_IO illustrating_VVG a_AT1 very_RG common_JJ GIS_NN2 procedure_NN1 ._. 
For_IF the_AT dumping_NN1 of_IO such_DA radioactive_JJ matter_NN1 four_MC criteria_NN2 need_VV0 to_TO be_VBI considered_VVN :_: 1_MC1 ._. 
The_AT waste_NN1 dump_NN1 must_VM be_VBI on_II a_AT1 site_NN1 with_IW a_AT1 suitable_JJ underlying_JJ geology_NN1 ;_; 2_MC ._. 
The_AT site_NN1 must_VM be_VBI accessible_JJ from_II the_AT UK_NP1 rail_NN1 network_NN1 ;_; 3_MC ._. 
The_AT site_NN1 must_VM not_XX be_VBI within_II a_AT1 conservation_JJ area_NN1 (_( that_REX21 is_REX22 ,_, a_AT1 national_JJ park_NN1 ,_, area_NN1 of_IO outstanding_JJ natural_JJ beauty_NN1 ,_, etc_RA ...._... )_) ;_; and_CC 4_MC ._. 
The_AT site_NN1 must_VM be_VBI away_II21 from_II22 areas_NN2 of_IO population_NN1 concentration_NN1 ._. 
Four_MC coverages_NN2 are_VBR used_VVN in_II the_AT case_NN1 study_NN1 ._. 
The_AT underlying_JJ suitable_JJ geologies_NN2 ,_, defined_VVN by_II the_AT British_JJ Geological_JJ Survey_NN1 (_( Chapman_NP1 et_RA21 al_RA22 ._. 
1986_MC )_) ,_, were_VBDR digitized_VVN from_II 1_MC1 :_: 625_MC 000_MC scale_NN1 maps_NN2 ._. 
The_AT rail_NN1 network_NN1 was_VBDZ digitized_VVN from_II 1_MC1 :_: 250_MC 000_MC maps_NN2 ,_, and_CC buffered_VVD at_II 3_MC km_NNU ._. 
The_AT national_JJ park_NN1 data_NN were_VBDR digitized_VVN from_II 1_MC1 :_: 250_MC 000_MC maps_NN2 ._. 
The_AT population_NN1 counts_VVZ those_DD2 grid_NN1 squares_NN2 with_IW a_AT1 population_NN1 greater_JJR than_CSN 490_MC residents_NN2 ,_, representing_VVG a_AT1 population_NN1 density_NN1 of_IO greater_JJR than_CSN 4.9_MC persons/ha_FU ._. 
The_AT coverages_NN2 are_VBR shown_VVN in_II Fig._NN1 6.3_MC and_CC the_AT error_NN1 estimates_VVZ for_IF them_PPHO2 are_VBR contained_VVN in_II Table_NN1 6.2_MC ._. 
The_AT first_MD step_NN1 involves_VVZ processing_VVG a_AT1 suitable_JJ geology_NN1 coverage_NN1 (_( GEOLOGY_NN1 )_) ._. 
The_AT ARC/INFO_NN1 ERASE_VV0 command_NN1 (_( the_AT Boolean_NN1 equivalent_NN1 of_IO NOT_XX )_) is_VBZ used_VVN to_TO remove_VVI the_AT constrained_JJ areas_NN2 from_II the_AT coverage_NN1 to_TO build_VVI up_RP a_AT1 composite_JJ coverage_NN1 ._. 
The_AT elimination_NN1 sequence_NN1 is_VBZ presented_VVN in_II Fig._NN1 6.4_MC ._. 
Those_DD2 areas_NN2 which_DDQ are_VBR outside_II the_AT rail_NN1 buffer_NN1 ,_, RAILBUF_NP1 ,_, are_VBR removed_VVN (_( to_TO give_VVI ERASRG_NP1 )_) ,_, and_CC next_MD the_AT conservation_JJ areas_NN2 ,_, CONSV_NP1 (_( to_TO give_VVI ERASRGC_NP1 )_) ._. 
finally_RR ,_, the_AT densely_RR populated_VVN areas_NN2 ,_, POP490_FO ,_, are_VBR deleted_VVN to_TO give_VVI a_AT1 set_NN1 of_IO feasible_JJ areas_NN2 for_IF site_NN1 investigation_NN1 (_( ERASRGCP_NP1 )_) ._. 
The_AT characteristics_NN2 of_IO each_DD1 of_IO the_AT coverages_NN2 are_VBR shown_VVN in_II Tables_NN2 6.36.6_MC ,_, and_CC the_AT ARC_NN1 command_NN1 sequence_NN1 to_TO obtain_VVI the_AT final_JJ overlay_NN1 is_VBZ in_II Table_NN1 6.7_MC ._. 
It_PPH1 is_VBZ difficult_JJ to_TO attempt_VVI to_TO determine_VVI what_DDQ are_VBR suitable_JJ estimates_NN2 of_IO the_AT error_NN1 in_II the_AT map_NN1 ._. 
ARC/INFO_NN1 works_VVZ to_II a_AT1 minimum_JJ tolerance_NN1 the_AT smallest_JJT distance_NN1 between_II any_DD pair_NN of_IO coordinates_NN2 known_VVN as_II the_AT fuzzy_JJ tolerance_NN1 ._. 
As_II a_AT1 default_NN1 ,_, the_AT fuzzy_JJ tolerance_NN1 is_VBZ 1/50_MF 000_MC of_IO the_AT width_NN1 of_IO the_AT base_NN1 map_NN1 ,_, in_II this_DD1 case_NN1 ,_, 0.48_MC mapunit_NN1 ._. 
As_II the_AT mapunits_NN2 are_VBR 10_MC m_NNO ,_, the_AT fuzzy_JJ tolerance_NN1 represents_VVZ 4.8_MC m_NNO on_II the_AT ground_NN1 ._. 
The_AT population_NN1 data_NN arc_NN1 rather_RG more_RGR problematic_JJ ,_, since_CS they_PPHS2 are_VBR generated_VVN from_II the_AT '_GE centroids_NN2 '_GE of_IO census_NN1 enumeration_NN1 districts_NN2 (_( EDs_NN2 )_) ;_; the_AT boundaries_NN2 of_IO the_AT census_NN1 EDs_NN2 do_VD0 not_XX conveniently_RR follow_VVI a_AT1 grid_NN1 ,_, it_PPH1 is_VBZ assumed_VVN they_PPHS2 have_VH0 a_AT1 similar_JJ fuzzy_JJ tolerance_NN1 to_II the_AT other_JJ coverages_NN2 ._. 
Clearly_RR ,_, the_AT quantization_NN1 noise_NN1 which_DDQ is_VBZ introduced_VVN to_II the_AT data_NN requires_VVZ a_AT1 rather_RG more_RGR sophisticated_JJ error_NN1 model_NN1 ._. 
The_AT case_NN1 study_NN1 uses_VVZ the_AT error_NN1 estimates_VVZ in_II Table_NN1 6.2_MC as_II a_AT1 guide_NN1 to_II likely_JJ levels_NN2 of_IO uncertainty_NN1 until_CS further_JJR research_NN1 yields_VVZ more_RGR realistic_JJ models_NN2 ._. 
The_AT simulation_NN1 is_VBZ run_VVN 100_MC times_NNT2 ._. 
In_II each_DD1 simulation_NN1 :_: 1_MC1 ._. 
Leftovers_NN2 from_II any_DD previous_JJ run_NN1 are_VBR removed_VVN ;_; 2_MC ._. 
A_AT1 copy_NN1 is_VBZ taken_VVN of_IO each_DD1 of_IO the_AT unperturbed_JJ coverages_NN2 ;_; 3_MC ._. 
All_DB nodes_NN2 and_CC vertices_NN2 of_IO each_DD1 coverage_NN1 are_VBR perturbed_VVN ;_; 4_MC ._. 
The_AT file_NN1 structures_NN2 are_VBR reorganized_VVN into_II a_AT1 form_NN1 the_AT ARC_NN1 i/o_NN1 routines_NN2 can_VM deal_VVI with_IW ._. 
This_DD1 step_NN1 is_VBZ highly_RR VAX/VMS_NP1 specific_JJ and_CC could_VM probably_RR be_VBI replaced_VVN by_II some_DD more_DAR universal_JJ coding_NN1 in_II the_AT perturbation_NN1 software._NNU 5_MC ._. 
The_AT overlay_NN1 sequence_NN1 takes_VVZ place_NN1 ;_; remnants_NN2 of_IO any_DD previous_JJ failures_NN2 are_VBR removed_VVN ;_; and_CC 6_MC ._. 
The_AT output_NN1 vector_NN1 coverage_NN1 is_VBZ rastered_VVN ,_, prior_II21 to_II22 plotting_VVG ._. 
Step_NN1 6_MC introduces_VVZ a_AT1 series_NN of_IO problems_NN2 ._. 
The_AT rasterization_NN1 process_NN1 introduces_VVZ noise_NN1 ._. 
For_IF the_AT experiments_NN2 here_RL ,_, a_AT1 raster_NN1 size_NN1 of_IO 1_MC1 km_NNU (_( 100_MC mapunits_NN2 )_) is_VBZ used_VVN ._. 
Given_CS21 that_CS22 the_AT fuzzy_JJ tolerance_NN1 is_VBZ 25_MC mapunits_NN2 ,_, then_RT the_AT raster_NN1 size_NN1 should_VM be_VBI that_DD1 or_CC less_RRR ,_, to_TO remove_VVI the_AT quantization_NN1 noise_NN1 ._. 
However_RR ,_, reducing_VVG the_AT raster_NN1 size_NN1 from_II 100_MC to_II 25_MC mapunits_NN2 increases_VVZ the_AT disc_NN1 space_NN1 required_VVN to_TO store_VVI the_AT raster_NN1 files_NN2 by_II a_AT1 factor_NN1 of_IO 16_MC as_II31 well_II32 as_II33 increasing_JJ processing_NN1 times_NNT2 ._. 
ARC/INFO_NN1 is_VBZ not_XX well_RR tuned_VVN to_TO handle_VVI this_DD1 sort_NN1 of_IO problem_NN1 ._. 
Simulation_NN1 results_VVZ The_AT error_NN1 model_NN1 used_VVD here_RL is_VBZ an_AT1 obvious_JJ over-simplification_NN1 of_IO reality_NN1 ._. 
Each_DD1 vertex_NN1 in_II the_AT coverage_NN1 was_VBDZ perturbed_VVN by_II selecting_VVG random_JJ numbers_NN2 from_II a_AT1 uniform_JJ distribution_NN1 ,_, with_IW ranges_NN2 defined_VVN in_II Table_NN1 6.2_MC ._. 
Steps_NN2 were_VBDR taken_VVN to_TO ensure_VVI that_CST the_AT nodes_NN2 were_VBDR not_XX perturbed_VVN more_DAR than_CSN once_RR ,_, but_CCB no_AT checks_NN2 were_VBDR made_VVN to_TO determine_VVI whether_CSW the_AT perturbation_NN1 process_NN1 caused_VVD sliver_NN1 polygons_NN2 to_TO be_VBI created_VVN ._. 
The_AT fuzzy_JJ tolerance_NN1 of_IO 25_MC mapunits_NN2 was_VBDZ chosen_VVN to_TO try_VVI to_TO eliminate_VVI these_DD2 from_II the_AT final_JJ coverage_NN1 ._. 
The_AT process_NN1 required_VVN 16_MC 140_MC s_ZZ1 of_IO CPU_NN1 time_NNT1 and_CC 20_MC h_ZZ1 of_IO elapsed_JJ time_NNT1 on_II a_AT1 lightly_RR loaded_JJ microVAX_NN1 2_MC ,_, under_II VMS_NP1 ._. 
The_AT resulting_JJ maps_NN2 are_VBR shown_VVN in_II Fig._NN1 6.5_MC ._. 
There_EX are_VBR seven_MC maps_NN2 ,_, showing_VVG 90_MC ,_, 95_MC ,_, 96_MC ,_, 97_MC ,_, 98_MC ,_, 99_MC and_CC 100_MC per_NNU21 cent_NNU22 ,_, '_GE credibility_NN1 regions_NN2 '_GE ._. 
The_AT credible_JJ region_NN1 is_VBZ that_DD1 which_DDQ contains_VVZ the_AT feasible_JJ region_NN1 n_ZZ1 or_CC more_DAR times_NNT2 out_II21 of_II22 M_MC ;_; in_II this_DD1 case_NN1 M_ZZ1 is_VBZ 100_MC ._. 
As_II the_AT '_GE credibility_NN1 level_NN1 '_GE increases_NN2 ,_, so_RR fewer_DAR and_CC fewer_DAR areas_NN2 are_VBR included_VVN in_II the_AT map_NN1 ,_, reflecting_VVG the_AT increasing_JJ confidence_NN1 in_II the_AT map_NN1 ._. 
More_DAR areas_NN2 are_VBR included_VVN in_II the_AT map_NN1 where_CS the_AT user_NN1 is_VBZ willing_JJ to_TO be_VBI wrong_JJ 10_MC times_NNT2 out_II21 of_II22 100_MC ,_, but_CCB far_RR fewer_DAR in_II that_DD1 which_DDQ they_PPHS2 are_VBR willing_JJ to_TO run_VVI the_AT risk_NN1 of_IO being_VBG wrong_JJ once_RR in_II a_AT1 hundred_NNO times_NNT2 ;_; a_AT1 comparison_NN1 of_IO the_AT 90_MC and_CC 99_MC per_NNU21 cent_NNU22 maps_NN2 will_VM make_VVI this_DD1 clear_JJ (_( figs_NN2 6.5a_FO and_CC f_ZZ1 )_) ._. 
There_EX is_VBZ little_DA1 difference_NN1 between_II the_AT 90_MC ,_, 95_MC ,_, 96_MC and_CC 97_MC per_NNU21 cent_NNU22 maps_NN2 (_( Figs_NN2 6.5a-d_FO )_) ,_, but_CCB between_II the_AT 97_MC and-98_MC per_NNU21 cent_NNU22 maps_NN2 a_AT1 considerable_JJ region_NN1 in_II the_AT north-east_ND1 corner_NN1 of_IO the_AT map_NN1 is_VBZ lost_VVN (_( Figs_NN2 6.5d-e_NNU )_) ._. 
Between_II 99_MC and_CC l00_FO per_NNU21 cent_NNU22 ,_, virtually_RR all_DB the_AT map_NN1 (_( figs_NN2 6.5f-g_NNU )_) is_VBZ lost_VVN with_IW the_AT exception_NN1 of_IO the_AT region_NN1 in_II the_AT south-east_ND1 corner_NN1 of_IO the_AT map_NN1 ._. 
The_AT overlay_NN1 process_NN1 has_VHZ been_VBN most_RGT reliable_JJ here_RL ._. 
Comparing_VVG this_DD1 map_NN1 with_IW the_AT ERASRGCP_NN1 map_NN1 from_II Fig._NN1 6.4_MC will_VM reveal_VVI the_AT effect_NN1 of_IO trying_VVG to_TO take_VVI uncertainty_NN1 in_II the_AT input_NN1 data_NN into_II account_NN1 explicitly_RR on_II the_AT resulting_JJ map_NN1 pattern_NN1 ._. 
There_EX are_VBR far_RR fewer_DAR areas_NN2 in_II the_AT simulated_JJ map_NN1 about_II which_DDQ one_PN1 would_VM risk_VVI a_AT1 definitive_JJ statement_NN1 than_CSN in_II the_AT map_NN1 from_II the_AT unsimulated_JJ process_NN1 ._. 
Discussion_NN1 The_AT main_JJ problem_NN1 with_IW using_VVG a_AT1 Monte_NP1 Carlo_NP1 approach_NN1 is_VBZ its_APPGE slowness_NN1 ._. 
A_AT1 lightly_RR loaded_JJ microVAX_NN1 2_MC can_VM expect_VVI to_TO take_VVI over_RG 24_MC h_ZZ1 of_IO elapsed_JJ time_NNT1 to_TO carry_VVI out_RP 100_MC simulations_NN2 on_II a_AT1 small_JJ problem_NN1 ._. 
Fortunately_RR the_AT process_NN1 is_VBZ explicitly_RR parallel_JJ and_CC were_VBDR it_PPH1 to_TO be_VBI run_VVN on_II multiprocessor_NN1 hardware_NN1 then_RT there_EX could_VM be_VBI substantial_JJ speed-up_NN1 ._. 
The_AT simulation_NN1 process_NN1 is_VBZ parallel_RR ,_, since_CS one_MC1 simulation_NN1 does_VDZ not_XX depend_VVI on_II the_AT outcome_NN1 of_IO a_AT1 previous_JJ one_PN1 ._. 
The_AT selection_NN1 of_IO a_AT1 suitable_JJ error_NN1 model_NN1 is_VBZ problematic_JJ ._. 
One_MC1 approach_NN1 is_VBZ to_TO carry_VVI out_RP experiments_NN2 with_IW a_AT1 digitizing_JJ table_NN1 in_BCL21 order_BCL22 to_TO determine_VVI empirically_RR an_AT1 appropriate_JJ distribution_NN1 for_IF digitizing_VVG errors_NN2 ._. 
Synthetic_JJ data_NN which_DDQ are_VBR themselves_PPX2 a_AT1 GIS-derived_JJ product_NN1 may_VM be_VBI the_AT most_RGT difficult_JJ to_TO deal_VVI with_IW ._. 
The_AT grid_NN1 square_NN1 population_NN1 data_NN used_VVD here_RL are_VBR a_AT1 case_NN1 in_II point_NN1 ._. 
These_DD2 data_NN had_VHD to_TO be_VBI estimated_VVN from_II census_NN1 ED_NP1 information_NN1 ._. 
The_AT aggregation_NN1 process_NN1 merely_RR allocates_VVZ to_II a_AT1 grid_NN1 square_VV0 the_AT populations_NN2 of_IO the_AT ED_NP1 whose_DDQGE '_GE centroid_NN1 '_GE happens_VVZ to_TO fall_VVI in_II that_DD1 grid_NN1 square_NN1 ._. 
In_II the_AT case_NN1 where_CS two_MC or_CC more_DAR EDs_NN2 fall_VV0 in_II the_AT same_DA grid_NN1 square_NN1 ,_, then_RT the_AT grid_NN1 square_NN1 's_GE population_NN1 is_VBZ the_AT sum_NN1 of_IO those_DD2 of_IO the_AT individual_JJ EDs_NN2 ._. 
It_PPH1 can_VM also_RR be_VBI argued_VVN that_CST the_AT error_NN1 present_NN1 in_II this_DD1 data_NN ,_, given_VVN its_APPGE crudity_NN1 ,_, would_VM swamp_VVI any_DD of_IO the_AT other_JJ data_NN ._. 
However_RR ,_, it_PPH1 is_VBZ only_RR fair_JJ to_TO point_VVI out_RP that_CST the_AT population_NN1 distribution_NN1 is_VBZ uneven_JJ ,_, and_CC any_DD swamping_NN1 will_VM take_VVI place_NN1 in_II urban_JJ areas_NN2 ._. 
In_II the_AT context_NN1 of_IO radwaste_NN1 disposal_NN1 ,_, this_DD1 is_VBZ not_XX critical_JJ ,_, but_CCB for_IF some_DD other_JJ operations_NN2 ,_, it_PPH1 obviously_RR is_VBZ ._. 
There_EX might_VM be_VBI alternatives_NN2 to_II using_VVG a_AT1 grid_NN1 ._. 
Very_RG few_DA2 '_GE natural_JJ '_GE data_NN come_VV0 in_II grid_NN1 form_NN1 ._. 
One_PN1 might_VM buffer_VVI the_AT ED_NP1 centroids_NN2 ,_, assign_VV0 populations_NN2 to_II the_AT resulting_JJ zones_NN2 ,_, then_RT determine_VV0 the_AT population_NN1 densities_NN2 ._. 
The_AT choice_NN1 of_IO buffer_NN1 distance_NN1 would_VM need_VVI to_TO be_VBI the_AT subject_NN1 of_IO some_DD experiment_NN1 ._. 
The_AT use_NN1 of_IO Theissen_NP1 polygons_NN2 has_VHZ been_VBN suggested_VVN ,_, although_CS the_AT statistical_JJ properties_NN2 of_IO processes_NN2 giving_VVG rise_NN1 to_II such_DA areas_NN2 are_VBR poor_JJ surrogates_NN2 for_IF digitized_JJ ED_NN1 boundaries_NN2 ._. 
The_AT varying_JJ size_NN1 of_IO EDs_NN2 in_II urban_JJ and_CC rural_JJ areas_NN2 ,_, and_CC the_AT discontinuous_JJ nature_NN1 of_IO the_AT population_NN1 distribution_NN1 inside_II the_AT larger_JJR rural_JJ EDs_NN2 is_VBZ also_RR difficult_JJ to_TO parametrize_VVI ._. 
The_AT selection_NN1 of_IO a_AT1 raster_NN1 size_NN1 presents_VVZ some_DD other_JJ problems_NN2 ._. 
The_AT authors_NN2 used_VVD an_AT1 IBM_NP1 4216010_MC PagePrinter_VV0 as_II an_AT1 output_NN1 device_NN1 ._. 
This_DD1 is_VBZ a_AT1 device_NN1 which_DDQ supports_VVZ PostScript_NP1 ,_, and_CC which_DDQ has_VHZ a_AT1 specified_JJ resolution_NN1 of_IO 300_MC pixels/inch_FU ,_, both_DB2 the_AT vertical_JJ and_CC horizontal_JJ direction_NN1 ._. 
In_II the_AT case_NN1 study_NN1 maps_NN2 ,_, 1_MC1 pixel_NN1 represents_VVZ a_AT1 square_NN1 of_IO size_NN1 118_MC m_NNO ._. 
With_IW a_AT1 l200_FO dpi_NN2 printer_NN1 (_( such_II21 as_II22 a_AT1 phototypesetter_NN1 )_) ,_, each_DD1 pixel_NN1 would_VM represent_VVI a_AT1 square_NN1 of_IO size_NN1 29.5_MC m_NNO ._. 
These_DD2 set_VV0 minimum_JJ sizes_NN2 for_IF output_NN1 ._. 
It_PPH1 should_VM be_VBI noted_VVN that_CST all_DB the_AT experiments_NN2 thus_RR far_RR have_VH0 used_VVN vector_NN1 representations_NN2 ._. 
It_PPH1 may_VM be_VBI easier_JJR to_TO implement_VVI it_PPH1 in_II a_AT1 raster_NN1 environment_NN1 ,_, but_CCB much_RR would_VM depend_VVI on_II the_AT size_NN1 of_IO the_AT rasters_NN2 chosen_VVN to_TO represent_VVI the_AT areas_NN2 ._. 
Clearly_RR this_DD1 is_VBZ yet_RR another_DD1 area_NN1 for_IF further_JJR research_NN1 ._. 
Conclusions_NN2 This_DD1 chapter_NN1 has_VHZ introduced_VVN some_DD of_IO the_AT issues_NN2 surrounding_VVG the_AT propagation_NN1 of_IO error_NN1 in_II GIS_NN2 and_CC described_VVD the_AT preliminary_JJ application_NN1 of_IO a_AT1 Monte_NP1 Carlo_NP1 approach_NN1 to_II assessing_VVG their_APPGE effects_NN2 ._. 
The_AT empirical_JJ research_NN1 carried_VVD out_RP here_RL is_VBZ intended_VVN to_TO test_VVI the_AT feasibility_NN1 of_IO this_DD1 approach_NN1 ,_, as_II31 well_II32 as_II33 providing_VVG some_DD general_JJ indications_NN2 of_IO what_DDQ the_AT effects_NN2 might_VM be_VBI and_CC of_IO one_MC1 way_NN1 by_II which_DDQ they_PPHS2 can_VM be_VBI represented_VVN on_II the_AT final_JJ output_NN1 map_NN1 ._. 
In_II so_RR doing_VDG it_PPH1 is_VBZ recognized_VVN that_CST many_DA2 of_IO the_AT questions_NN2 raised_VVN are_VBR left_VVN open_JJ for_IF subsequent_JJ investigation_NN1 and_CC that_CST the_AT results_NN2 are_VBR initial_JJ and_CC tentative_JJ ._. 
Although_CS conceptually_RR clear_JJ ,_, the_AT method_NN1 is_VBZ not_XX easy_JJ to_TO implement_VVI with_IW current_JJ hardware_NN1 ._. 
There_EX is_VBZ likely_JJ to_TO be_VBI substantial_JJ increases_NN2 in_II CPU_NN1 times_NNT2 ,_, although_CS it_PPH1 might_VM be_VBI possible_JJ to_TO reduce_VVI this_DD1 by_II a_AT1 factor_NN1 of_IO 5_MC ._. 
However_RR ,_, with_IW faster_JJR hardware_NN1 likely_JJ to_TO be_VBI on_II the_AT market_NN1 in_II the_AT near_JJ future_NN1 and_CC the_AT possibility_NN1 of_IO the_AT emergence_NN1 of_IO parallel_JJ GIS_NN2 machines_NN2 ,_, there_EX is_VBZ some_DD justification_NN1 for_IF believing_VVG that_DD1 extra_JJ effort_NN1 is_VBZ both_RR worth_II while_NNT1 and_CC acceptable_JJ ._. 
The_AT challenge_NN1 is_VBZ to_TO resolve_VVI the_AT outstanding_JJ questions_NN2 and_CC perfect_VVI the_AT technology_NN1 as_RG soon_RR as_CSA possible_JJ ._. 
Perusal_NN1 of_IO Fig._NN1 6.5_MC indicates_VVZ that_CST it_PPH1 does_VDZ appear_VVI to_TO work_VVI ,_, and_CC that_CST it_PPH1 offers_VVZ a_AT1 pragmatic_JJ solution_NN1 that_CST could_VM probably_RR be_VBI developed_VVN further_RRR into_II a_AT1 general-purpose_JJ GIS_NN2 '_GE error_NN1 button_NN1 '_GE ._. 
User_NN1 interfaces_VVZ Jonathan_NP1 F._NP1 Raper_NP1 Introduction_NN1 Geographical_JJ information_NN1 systems_NN2 (_( GIS_NN2 )_) make_VV0 considerable_JJ demands_NN2 on_II the_AT user_NN1 :_: the_AT wide_JJ variety_NN1 of_IO data_NN types_NN2 recorded_VVN in_II digital_JJ maps_NN2 ,_, the_AT complex_JJ data_NN structures_NN2 used_VMK to_TO organize_VVI them_PPHO2 and_CC the_AT range_NN1 of_IO operations_NN2 available_JJ ,_, amount_NN1 to_II a_AT1 formidable_JJ obstacle_NN1 for_IF most_DAT users_NN2 with_IW standard_JJ requirements_NN2 ._. 
As_II such_DA ,_, the_AT quality_NN1 of_IO interfaces_NN2 to_II GIS_NN2 has_VHZ taken_VVN on_RP a_AT1 considerable_JJ importance_NN1 in_II31 terms_II32 of_II33 awareness_NN1 ,_, training_NN1 and_CC usage_NN1 ,_, both_RR to_II the_AT providers_NN2 of_IO GIS_NN2 software_NN1 and_CC users_NN2 of_IO GIS_NN2 alike_RR (_( Rhind_NP1 ,_, et_RA21 al_RA22 ._. 
1989_MC )_) ._. 
However_RR ,_, there_EX are_VBR many_DA2 aspects_NN2 to_II the_AT definition_NN1 of_IO an_AT1 interface_NN1 for_IF systems_NN2 as_RG complex_JJ as_CSA GIS_NN2 ,_, and_CC the_AT solutions_NN2 to_II this_DD1 problem_NN1 are_VBR developing_VVG extremely_RR rapidly_RR at_II the_AT time_NNT1 of_IO writing_NN1 ._. 
Accordingly_RR ,_, this_DD1 chapter_NN1 aims_VVZ to_TO set_VVI out_RP the_AT requirements_NN2 for_IF a_AT1 fully_RR configured_VVN GIS_NN2 interface_VV0 ,_, and_CC profiles_VVZ the_AT development_NN1 of_IO a_AT1 new_JJ GIS_NN2 user_NN1 interface_NN1 system_NN1 called_VVN UGIX_NN1 ._. 
This_DD1 model_NN1 is_VBZ also_RR used_VVN to_TO define_VVI a_AT1 research_NN1 agenda_NN1 for_IF the_AT next_MD 5_MC years_NNT2 ;_; the_AT reader_NN1 may_VM judge_VVI the_AT accuracy_NN1 of_IO this_DD1 analysis_NN1 by_II the_AT commercial_JJ reality_NN1 of_IO available_JJ systems_NN2 during_II the_AT early_JJ and_CC mid-1990s_MC2 ._. 
There_EX is_VBZ a_AT1 wide_JJ recognition_NN1 that_CST the_AT problems_NN2 of_IO poor_JJ interfaces_NN2 are_VBR of_IO considerable_JJ importance_NN1 to_II the_AT development_NN1 of_IO GIS_NN2 ._. 
The_AT UK_NP1 Government_NN1 Committee_NN1 of_IO Enquiry_NN1 chaired_VVN by_II Lord_NNB Chorley_NP1 (_( DoE_NP1 1987_MC )_) on_II the_AT Handling_NN1 of_IO Geographic_JJ Information_NN1 suggested_VVN in_II recommendation_NN1 59_MC that_CST GIS_NN2 technology_NN1 projects_NN2 be_VBI promoted_VVN since_CS the_AT report_NN1 noted_VVD that_CST the_AT existing_JJ interfaces_NN2 to_II GIS_NN2 systems_NN2 were_VBDR poor_JJ ._. 
The_AT shortcomings_NN2 of_IO GIS_NN2 user_NN1 environments_NN2 can_VM be_VBI divided_VVN into_II two_MC groups_NN2 :_: 1_MC1 ._. 
Task_NN1 complexity_NN1 the_AT system_NN1 functions_NN2 implementing_VVG spatial_JJ operations_NN2 are_VBR often_RR complex_JJ and_CC obscure_VV0 ;_; 2_MC ._. 
Poor_JJ database_NN1 view_NN1 tools_NN2 few_DA2 systems_NN2 make_VV0 available_JJ database_NN1 views_VVZ meaningful_JJ to_II the_AT user_NN1 ._. 
Accordingly_RR ,_, it_PPH1 is_VBZ clearly_RR desirable_JJ that_CST improved_JJ interfaces_NN2 be_VBI developed_VVN so_CS21 that_CS22 the_AT range_NN1 of_IO spatial_JJ operations_NN2 in_II a_AT1 system_NN1 can_VM be_VBI organized_VVN into_II standardized_JJ functions_NN2 or_CC tasks_NN2 ,_, and_CC so_CS21 that_CS22 the_AT spatial_JJ database_NN1 can_VM be_VBI visualized_VVN ._. 
The_AT quality_NN1 of_IO GIS_NN2 user_NN1 interfaces_NN2 is_VBZ also_RR an_AT1 important_JJ factor_NN1 in_II the_AT acceptance_NN1 ,_, uptake_NN1 and_CC efficiency_NN1 of_IO the_AT integrated_JJ GIS_NN2 which_DDQ are_VBR currently_RR on_II the_AT market_NN1 ._. 
In_II a_AT1 recent_JJ study_NN1 by_II Willis_NP1 and_CC Nutter_NP1 (_( 1990_MC )_) of_IO 136_MC publicly_RR funded_JJ utilities_NN2 and_CC municipalities_NN2 in_II the_AT UK_NP1 57_MC per_NNU21 cent_NNU22 stated_VVD that_CST they_PPHS2 were_VBDR inhibited_VVN in_II their_APPGE GIS_NN2 developments_NN2 by_II '_GE a_AT1 lack_NN1 of_IO staff_NN with_IW the_AT right_JJ expertise_NN1 '_GE ._. 
At_II a_AT1 time_NNT1 when_RRQ there_EX is_VBZ a_AT1 national_JJ and_CC international_JJ shortage_NN1 of_IO staff_NN skilled_JJ in_II the_AT computer_NN1 handling_NN1 of_IO geographical_JJ data_NN (_( Rhind_NP1 and_CC Mounsey_NP1 1989_MC )_) ,_, '_GE ease_NN1 of_IO use_NN1 '_GE is_VBZ a_AT1 vital_JJ criterion_NN1 for_IF the_AT selection_NN1 of_IO an_AT1 appropriate_JJ GIS_NN2 ._. 
It_PPH1 is_VBZ generally_RR accepted_VVN that_CST a_AT1 system_NN1 which_DDQ is_VBZ easy_JJ to_TO use_VVI can_VM help_VVI cut_JJ recruitment_NN1 and_CC training_NN1 costs_NN2 ,_, and_CC help_VV0 retain_VVI staff_NN ._. 
Organizations_NN2 which_DDQ are_VBR in_II the_AT process_NN1 of_IO implementing_VVG GIS_NN2 strategies_NN2 also_RR appreciate_VV0 that_CST the_AT '_GE ease_NN1 of_IO use_NN1 '_GE factor_NN1 is_VBZ a_AT1 key_JJ control_NN1 over_II how_RGQ quickly_RR GIS_NN2 programmes_NN2 can_VM be_VBI implemented_VVN ,_, and_CC therefore_RR the_AT speed_NN1 with_IW which_DDQ financial_JJ targets_NN2 for_IF paying_VVG off_RP capital_NN1 costs_NN2 can_VM be_VBI met_VVN ._. 
It_PPH1 may_VM also_RR be_VBI true_JJ that_DD1 '_VBZ ease_NN1 of_IO use_NN1 '_GE can_NN1 influence_VVI the_AT quality_NN1 of_IO work_NN1 done_VDN and_CC the_AT effectiveness_NN1 of_IO a_AT1 GIS_NN2 as_II a_AT1 decision_NN1 support_NN1 system_NN1 ._. 
To_TO illustrate_VVI this_DD1 in_II the_AT negative_JJ ,_, Beard_NP1 (_( 1989_MC )_) for_REX21 example_REX22 ,_, showed_VVD how_RRQ '_GE use_NN1 error_NN1 '_GE in_II GIS_NN2 was_VBDZ an_AT1 important_JJ but_CCB neglected_JJ aspect_NN1 of_IO quality_NN1 control_NN1 in_II GIS_NN2 ._. 
In_II summary_NN1 ,_, the_AT user_NN1 interface_NN1 is_VBZ a_AT1 vital_JJ element_NN1 of_IO any_DD GIS_NN2 ._. 
Long_RR ignored_VVN as_II an_AT1 esoteric_JJ aspect_NN1 of_IO GIS_NN2 design_VV0 while_CS GIS_NN2 development_NN1 was_VBDZ driven_VVN by_II the_AT need_NN1 to_TO extend_VVI functionality_NN1 ,_, the_AT user_NN1 interface_NN1 is_VBZ now_RT beginning_VVG to_TO attract_VVI its_APPGE due_JJ attention_NN1 ._. 
However_RR ,_, the_AT implementation_NN1 of_IO a_AT1 GIS_NN2 user_NN1 interface_NN1 involves_VVZ considerably_RR more_DAR than_CSN the_AT improvement_NN1 of_IO the_AT human-computer_JJ interaction_NN1 (_( HCI_NP1 )_) process_VV0 ._. 
Since_CS GIS_NN2 are_VBR conceptually_RR complex_JJ and_CC involve_VV0 diverse_JJ operations_NN2 ranging_VVG from_II data_NN modelling_VVG to_II geometric_JJ transformations_NN2 ,_, improving_VVG the_AT HCI_NN1 can_VM not_XX be_VBI a_AT1 complete_JJ solution_NN1 to_II the_AT improvements_NN2 of_IO GIS_NN2 use_VV0 ._. 
Consideration_NN1 also_RR needs_VVZ to_TO be_VBI given_VVN to_II the_AT embedding_NN1 of_IO knowledge_NN1 ,_, task_NN1 definitions_NN2 and_CC database_NN1 view_NN1 manipulations_NN2 into_II such_DA interfaces_NN2 ._. 
A_AT1 key_JJ assumption_NN1 ,_, therefore_RR ,_, which_DDQ remains_VVZ to_TO be_VBI tested_VVN is_VBZ whether_CSW a_AT1 GIS_NN2 user_NN1 interface_NN1 should_VM condition_NN1 GIS_NN2 use_VV0 ._. 
While_CS there_EX are_VBR many_DA2 who_PNQS would_VM argue_VVI that_CST a_AT1 measure_NN1 of_IO technical_JJ knowledge_NN1 is_VBZ desirable_JJ in_II those_DD2 who_PNQS use_VV0 a_AT1 GIS_NN2 and_CC a_AT1 protection_NN1 against_II the_AT misuse_NN1 of_IO a_AT1 powerful_JJ tool_NN1 ,_, it_PPH1 must_VM now_RT be_VBI established_VVN that_DD1 maximum_NN1 '_GE achievable_JJ '_GE use_NN1 of_IO a_AT1 software_NN1 system_NN1 owes_VVZ much_RR to_II the_AT creation_NN1 of_IO a_AT1 structured_JJ use_NN1 environment_NN1 ,_, with_IW logic_NN1 controls_NN2 built_VVN into_II the_AT interface_NN1 ._. 
The_AT user_NN1 is_VBZ then_RT free_VV0 to_TO choose_VVI the_AT environment_NN1 which_DDQ best_RRT matches_VVZ their_APPGE use_NN1 characteristics_NN2 or_CC which_DDQ improves_VVZ their_APPGE aggregate_JJ efficiency_NN1 measured_VVN in_II time_NNT1 ,_, error_NN1 or_CC quality_NN1 terms_NN2 ._. 
Characteristics_NN2 of_IO a_AT1 user_NN1 interface_VV0 A_AT1 user_NN1 interface_NN1 ,_, at_II its_APPGE most_RGT basic_JJ ,_, consists_VVZ simply_RR of_IO a_AT1 system_NN1 for_IF communication_NN1 with_IW the_AT computer_NN1 ._. 
Since_CS the_AT demise_NN1 of_IO the_AT punch-card_JJ system_NN1 for_IF entering_VVG commands_NN2 this_DD1 has_VHZ normally_RR been_VBN undertaken_VVN using_VVG a_AT1 keyboard_NN1 and_CC a_AT1 screen_NN1 driven_VVN by_II a_AT1 cathode_NN1 ray_NN1 tube_NN1 (_( CRT_NP1 )_) or_CC liquid_JJ crystal_NN1 display_NN1 (_( LCD_NN1 )_) ,_, although_CS voice_NN1 entry_NN1 of_IO data_NN and_CC touch_NN1 screens_NN2 have_VH0 also_RR been_VBN used_VVN ._. 
Typical_JJ computer_NN1 operating_NN1 systems_NN2 use_VV0 (_( and_CC continue_VV0 to_TO use_VVI )_) '_GE command_NN1 line_NN1 interfaces_NN2 '_GE where_CS a_AT1 user_NN1 enters_VVZ a_AT1 syntactically_RR correct_JJ command_NN1 at_II a_AT1 prompt_NN1 in_BCL21 order_BCL22 to_TO give_VVI an_AT1 instruction_NN1 ._. 
Many_DA2 computer_NN1 applications_NN2 use_VV0 this_DD1 kind_NN1 of_IO interface_NN1 which_DDQ is_VBZ common_JJ on_II all_DB machine_VV0 ranges_NN2 and_CC sizes_NN2 ,_, e.g._REX MS-DOS_NN1 for_IF IBM-compatible_JJ personal_JJ computers_NN2 ._. 
However_RR ,_, in_II the_AT last_MD 5_MC years_NNT2 the_AT command_NN1 line_NN1 interface_NN1 has_VHZ begun_VVN to_TO be_VBI replaced_VVN by_II the_AT graphical_JJ user_NN1 interface_NN1 (_( GUI_NP1 )_) ._. 
The_AT GUI_NN1 is_VBZ an_AT1 audio-visual_JJ display_NN1 on_II the_AT computer_NN1 screen_NN1 which_DDQ presents_VVZ a_AT1 screen_NN1 metaphor_NN1 for_IF the_AT actions_NN2 which_DDQ the_AT computer_NN1 or_CC program_NN1 can_VM carry_VVI out_RP ._. 
Typically_RR the_AT means_NN of_IO issuing_JJ commands_NN2 using_VVG this_DD1 system_NN1 has_VHZ been_VBN via_II a_AT1 window-icon-mouse-pop-up_NN1 menu_NN1 (_( WIMP_NP1 )_) display_VV0 ,_, where_CS the_AT mouse_NN1 is_VBZ used_VVN to_TO point_VVI at_II icons_NN2 or_CC menus_NN2 in_II windows_NN2 (_( which_DDQ are_VBR subsets_NN2 of_IO the_AT screen_NN1 working_VVG area_NN1 )_) ._. 
The_AT options_NN2 available_JJ are_VBR indicated_VVN to_II the_AT user_NN1 by_II the_AT range_NN1 of_IO identifiable_JJ screen_NN1 objects_NN2 (_( Helander_NP1 1988_MC )_) ._. 
Early_RR work_VV0 on_II the_AT concept_NN1 of_IO WIMP_NP1 user_NN1 interfaces_NN2 developed_VVN from_II research_NN1 on_II the_AT Smalltalk-80_MC project_NN1 at_II the_AT XeroxPARC_NP1 in_II California_NP1 in_II the_AT early_JJ 1980s_MC2 (_( Goldberg_NP1 and_CC Robson_NP1 1983_MC )_) ._. 
Work_VV0 by_II Smith_NP1 et_RA21 al_RA22 ._. 
(_( 1983_MC )_) and_CC Sneiderman_NP1 (_( 1983_MC )_) developed_VVD the_AT concepts_NN2 behind_II moving_VVG and_CC selecting_VVG screen_NN1 representations_NN2 ,_, which_DDQ has_VHZ become_VVN important_JJ to_II all_DB graphics-oriented_JJ applications_NN2 (_( such_II21 as_II22 GIS_NN2 )_) ._. 
The_AT GUI_NN1 for_IF the_AT Apple_NN1 Macintosh_NP1 (_( first_MD released_VVN in_II 1984_MC )_) was_VBDZ the_AT first_MD to_TO become_VVI widely_RR used_VVN and_CC its_APPGE popularity_NN1 helped_VVN ensure_VVI that_CST other_JJ GUIs_NN1 were_VBDR developed_VVN for_IF PC_NN1 compatibles_NN2 and_CC UNIX_NN1 platforms_NN2 ._. 
This_DD1 development_NN1 has_VHZ defined_VVN a_AT1 new_JJ and_CC higher_JJR standard_NN1 for_IF interfaces_NN2 which_DDQ has_VHZ become_VVN common_JJ in_II all_DB areas_NN2 of_IO data_NN processing_NN1 ._. 
Apple_NN1 's_GE publication_NN1 Human_NN1 Interface_NN1 Guidelines_NN2 (_( Apple_NN1 Computer_NN1 1987_MC )_) set_VVD out_RP the_AT 10_MC chief_JJ characteristics_NN2 of_IO its_APPGE own_DA GUI_NN2 as_CSA follows_VVZ :_: 1_MC1 ._. 
Metaphors_NN2 from_II the_AT real_JJ world_NN1 ;_; 2_MC ._. 
Direct_JJ manipulation_NN1 by_II the_AT user_NN1 ;_; ,_, 3_MC ._. 
See_VV0 and_CC point_NN1 (_( instead_CS21 of_CS22 remember_VV0 and_CC type_NN1 )_) ;_; 4_MC ._. 
Consistency_NN1 ;_; 5_MC ._. 
WYSIWYG_JJ (_( what_DDQ you_PPY see_VV0 is_VBZ what_DDQ you_PPY get_VV0 )_) ;_; 6_MC ._. 
User_NN1 control_NN1 ;_; 7_MC ._. 
Feedback_NN1 and_CC dialogue_NN1 ;_; 8_MC ._. 
Forgiveness_NN1 ;_; 9_MC ._. 
Perceived_JJ stability_NN1 ;_; 10_MC ._. 
Aesthetic_JJ integrity_NN1 ._. 
These_DD2 principles_NN2 have_VH0 now_RT been_VBN generally_RR adopted_VVN across_II all_DB platforms_NN2 ,_, although_CS since_CS the_AT Macintosh_NP1 GUI_NP1 is_VBZ embedded_VVN in_II the_AT system_NN1 architecture_NN1 most_DAT of_IO these_DD2 characteristics_NN2 are_VBR enforced_VVN in_II Macintosh_NP1 software_NN1 design_NN1 and_CC engineering_NN1 ._. 
However_RR ,_, this_DD1 is_VBZ not_XX generally_RR the_AT case_NN1 ,_, and_CC a_AT1 whole_JJ variety_NN1 of_IO GUIs_NP1 have_VH0 developed_VVN ,_, each_DD1 adopting_VVG a_AT1 slightly_RR different_JJ subset_NN1 of_IO these_DD2 principles_NN2 as_CSA guidelines_NN2 ._. 
The_AT range_NN1 of_IO GUIs_NN1 which_DDQ have_VH0 now_RT developed_VVN are_VBR surveyed_VVN in_II Hayes_NP1 and_CC Baran_NP1 (_( 1989_MC )_) ._. 
On_II the_AT basis_NN1 of_IO the_AT analysis_NN1 of_IO 12_MC GUIs_NN1 they_PPHS2 suggest_VV0 that_CST the_AT GUI_NN1 is_VBZ composed_VVN of_IO three_MC main_JJ components_NN2 ._. 
Firstly_RR ,_, the_AT windowing_JJ system_NN1 is_VBZ a_AT1 set_NN1 of_IO tools_NN2 for_IF creating_VVG windows_NN2 and_CC their_APPGE characteristics_NN2 an_AT1 example_NN1 of_IO this_DD1 is_VBZ X_ZZ1 Windows_NN2 ._. 
The_AT second_MD main_JJ component_NN1 of_IO the_AT GUI_NN1 is_VBZ the_AT imaging_NN1 model_NN1 which_DDQ controls_VVZ the_AT drawing_NN1 of_IO the_AT screen_NN1 representations_NN2 such_II21 as_II22 fonts_NN2 and_CC icons_NN2 ;_; an_AT1 example_NN1 of_IO this_DD1 is_VBZ Display_NN1 Postscript_NP1 ._. 
The_AT third_MD component_NN1 is_VBZ the_AT application_NN1 program_NN1 interface_NN1 which_DDQ acts_VVZ as_II an_AT1 interface_NN1 to_II the_AT program_NN1 operations_NN2 and_CC controls_VVZ feedback_NN1 from_II the_AT screen_NN1 representations_NN2 ._. 
In_II31 response_II32 to_II33 the_AT buoyant_JJ demand_NN1 for_IF GUIs_NP1 almost_RR all_DB vendors_NN2 of_IO computer_NN1 systems_NN2 now_RT offer_VV0 such_DA an_AT1 interface_NN1 product_NN1 ._. 
Some_DD GUI_JJ components_NN2 are_VBR becoming_VVG standard_NN1 across_II a_AT1 variety_NN1 of_IO vendor_NN1 platforms_NN2 ;_; thus_RR ,_, the_AT windowing_JJ system_NN1 X_ZZ1 Windows_NP1 is_VBZ shared_VVN by_II a_AT1 number_NN1 of_IO systems_NN2 running_VVG UNIX_NN1 ._. 
Several_DA2 GIS_NN2 have_VH0 already_RR begun_VVN to_TO use_VVI GUIs_NN1 to_TO make_VVI their_APPGE systems_NN2 more_RRR user_NN1 friendly_NN1 using_VVG the_AT standard_JJ platform_NN1 interface_NN1 tools_NN2 as_CSA described_VVN above_RL ._. 
One_MC1 example_NN1 is_VBZ AlperRecords_NP2 from_II Alper_NP1 Systems_NN2 of_IO Cambridge_NP1 ,_, UK_NP1 which_DDQ uses_VVZ SUN_NN1 interface_NN1 tools_NN2 ,_, and_CC is_VBZ based_VVN on_II pop-up_JJ hierarchical_JJ menus_NN2 ._. 
However_RR ,_, a_AT1 standard_JJ GUI_NN2 can_VM only_RR translate_VVI the_AT command_NN1 structure_NN1 into_II graphical_JJ cues_NN2 and_CC will_VM still_RR be_VBI a_AT1 '_GE function-oriented_NN1 '_GE rather_II21 than_II22 a_AT1 '_GE task-oriented_NN1 '_GE system_NN1 ._. 
(_( This_DD1 means_VVZ that_CST the_AT user_NN1 sees_58 '_GE primitive_NN1 '_GE spatial_JJ operations_NN2 and_CC not_XX a_AT1 whole_JJ spatial_JJ problem_NN1 which_DDQ may_VM involve_VVI a_AT1 sequence_NN1 of_IO primitive_JJ operations_NN2 )_) ._. 
Hence_RR ,_, the_AT use_NN1 of_IO a_AT1 GIS_NN2 with_IW a_AT1 GUI_NN1 can_VM only_RR improve_VVI user_NN1 productivity_NN1 in_II '_GE use_NN1 '_GE factors_NN2 ,_, for_REX21 example_REX22 by_II increasing_VVG the_AT speed_NN1 of_IO use_NN1 and_CC reducing_VVG errors_NN2 ,_, and_CC may_VM only_RR help_VVI the_AT user_NN1 with_IW a_AT1 previously_RR substantial_JJ knowledge_NN1 of_IO GIS_NN2 ._. 
Thus_RR ,_, due_II21 to_II22 the_AT sheer_JJ complexity_NN1 of_IO spatial_JJ data_NN and_CC the_AT operations_NN2 available_JJ ,_, this_DD1 can_VM only_RR be_VBI a_AT1 partial_JJ solution_NN1 to_II the_AT general_JJ problem_NN1 of_IO user_NN1 interaction_NN1 with_IW a_AT1 GIS_NN2 (_( Gould_NP1 1989_MC )_) ._. 
Recent_JJ research_NN1 into_II GIS_NN2 user_NN1 interfaces_NN2 Research_VV0 into_II the_AT form_NN1 and_CC content_NN1 of_IO user_NN1 interfaces_NN2 for_IF GIS_NN2 has_VHZ accelerated_VVN rapidly_RR in_II the_AT last_MD couple_NN1 of_IO years_NNT2 ._. 
Initially_RR this_DD1 work_NN1 has_VHZ followed_VVN from_II the_AT general_JJ acceptance_NN1 of_IO the_AT WIMP_NN1 system_NN1 as_II an_AT1 effective_JJ form_NN1 of_IO human-computer_JJ interface_NN1 ._. 
Within_II GIS_NN2 the_AT unique_JJ appeal_NN1 of_IO the_AT GUI_NN1 has_VHZ been_VBN the_AT desire_NN1 to_TO improve_VVI the_AT ease_NN1 of_IO use_NN1 for_IF systems_NN2 involving_VVG complex_JJ graphical_JJ display_NN1 and_CC spatial_JJ manipulation_NN1 ._. 
Hence_RR the_AT primary_JJ interest_NN1 has_VHZ been_VBN in_II :_: 1_MC1 ._. 
The_AT development_NN1 of_IO new_JJ task-oriented_JJ interfaces_NN2 based_VVN on_II standard_JJ windowing_JJ systems_NN2 ;_; and_CC 2_MC ._. 
The_AT creation_NN1 of_IO standard_JJ query_NN1 formats_NN2 which_DDQ are_VBR independent_JJ of_IO the_AT database_NN1 structure_NN1 ._. 
Initial_JJ work_NN1 in_II the_AT GIS_NN2 field_VV0 on_II user_NN1 interfaces_NN2 has_VHZ concentrated_VVN on_II the_AT screen_NN1 designs_NN2 appropriate_VV0 for_IF spatial_JJ data_NN handling_VVG ._. 
Thus_RR ,_, Egenhofer_NP1 and_CC Frank_NP1 (_( 1988_MC )_) suggested_VVD a_AT1 way_NN1 in_II which_DDQ the_AT WYSIWYG_JJ principle_NN1 of_IO '_" what_DDQ you_PPY see_VV0 is_VBZ what_DDQ you_PPY get_VV0 '_" could_VM be_VBI extended_VVN to_II GIS_NN2 and_CC outlined_VVN the_AT components_NN2 of_IO their_APPGE own_DA system_NN1 which_DDQ included_VVD a_AT1 study_NN1 of_IO the_AT selection_NN1 of_IO objects_NN2 and_CC areas_NN2 ,_, legends_NN2 and_CC query_VV0 specification_NN1 ._. 
Subsequent_JJ work_NN1 by_II Cassel_NP1 and_CC Parker_NP1 (_( 1989_MC )_) carried_VVD out_RP tests_NN2 on_II the_AT effectiveness_NN1 of_IO different_JJ designs_NN2 of_IO GUI_NP1 ,_, focusing_VVG especially_RR on_II the_AT use_NN1 of_IO icons_NN2 to_TO represent_VVI classes_NN2 of_IO spatial_JJ data_NN ._. 
This_DD1 work_NN1 demonstrated_VVD some_DD of_IO the_AT special_JJ problems_NN2 which_DDQ have_VH0 to_TO be_VBI solved_VVN to_TO design_VVI good_JJ GUIs_NN1 for_IF GIS_NN2 :_: thus_RR ,_, an_AT1 icon_NN1 of_IO a_AT1 single_JJ pine_NN1 tree_NN1 caused_VVD some_DD confusion_NN1 when_CS presented_VVN to_II a_AT1 forester_NN1 since_CS he_PPHS1 could_VM not_XX identify_VVI its_APPGE species_NN !_! 
This_DD1 indicates_VVZ how_RRQ legends_NN2 and_CC graphical_JJ cues_NN2 must_VM generalize_VVI a_AT1 complex_JJ reality_NN1 to_TO provide_VVI appropriate_JJ '_GE cues_NN2 for_IF action_NN1 '_GE (_( Mark_NP1 1989_MC )_) ._. 
Gould_NP1 (_( 1989_MC )_) ,_, in_II a_AT1 review_NN1 of_IO the_AT HCI_JJ literature_NN1 ,_, also_RR noted_VVD the_AT potential_JJ to_TO move_VVI beyond_II simple_JJ GUIs_NN1 in_II interface_NN1 design_NN1 and_CC identified_VVN three_MC areas_NN2 of_IO potential_JJ improvement_NN1 ._. 
First_MD ,_, he_PPHS1 considered_VVD that_CST graphic_JJ interaction_NN1 was_VBDZ presently_RR a_AT1 limited_JJ form_NN1 of_IO process-response_JJ system_NN1 ,_, and_CC recommended_VVD that_CST designers_NN2 of_IO GIS_NN2 interfaces_NN2 look_VV0 for_IF ways_NN2 to_TO exploit_VVI the_AT sophisticated_JJ forms_NN2 of_IO human_JJ visual_JJ thinking_NN1 (_( Verplank_NP1 1988_MC )_) ._. 
Second_MD ,_, he_PPHS1 suggested_VVD that_CST the_AT interface_NN1 itself_PPX1 exploit_VV0 multi-media_NN techniques_NN2 of_IO sound_NN1 and_CC video_NN1 to_TO issue_VVI and_CC support_VVI commands._NNU finally_RR ,_, he_PPHS1 suggested_VVD that_CST interface_VV0 design_NN1 ought_VMK to_TO concentrate_VVI on_II handling_NN1 '_GE fuzzy_JJ '_GE queries_NN2 and_CC parsing_NN1 '_GE natural_JJ language_NN1 '_GE concepts_NN2 ._. 
One_MC1 recent_JJ study_NN1 of_IO user_NN1 interfaces_NN2 by_II Campari_NP1 et_RA21 al_RA22 ._. 
(_( 1990_MC )_) has_VHZ employed_VVN graphical_JJ concepts_NN2 in_II the_AT design_NN1 of_IO GIS_NN2 operations_NN2 ._. 
This_DD1 work_NN1 ,_, described_VVN as_II a_AT1 visual_JJ language_NN1 approach_NN1 to_II GIS_NN2 ,_, used_VVD a_AT1 visual_JJ programming_NN1 language_NN1 to_TO define_VVI map_NN1 operations_NN2 ._. 
This_DD1 consisted_VVD of_IO a_AT1 graphic_JJ workspace_NN1 and_CC primitive_JJ GIS_NN2 operations_NN2 shown_VVN as_CSA icons_NN2 which_DDQ the_AT user_NN1 selected_VVN from_II a_AT1 palette_NN1 and_CC assembled_VVN in_II the_AT workspace_NN1 ._. 
The_AT whole_JJ set_NN1 of_IO commands_NN2 encapsulated_VVN in_II the_AT visual_JJ procedure_NN1 was_VBDZ then_RT executed_VVN sequentially_RR ._. 
Hence_RR ,_, another_DD1 key_JJ aspect_NN1 of_IO spatial_JJ user_NN1 interface_NN1 design_NN1 is_VBZ that_CST it_PPH1 must_VM seek_VVI both_RR to_II structure_NN1 and_CC filter_NN1 human_NN1 perception_NN1 of_IO space_NN1 and_CC spatial_JJ relationships_NN2 when_CS expressing_VVG an_AT1 operation_NN1 and_CC passing_VVG it_PPH1 to_II a_AT1 GIS_NN2 ._. 
Research_NN1 into_II general_JJ aspects_NN2 of_IO the_AT perception_NN1 of_IO spatial_JJ relationships_NN2 began_VVD long_RR before_II the_AT special_JJ problems_NN2 presented_VVN by_II GIS_NN2 existed_VVD ,_, and_CC is_VBZ exemplified_VVN by_II the_AT classic_JJ work_NN1 of_IO Gould_NP1 and_CC White_NP1 (_( 1974_MC )_) on_II '_GE mental_JJ maps_NN2 '_GE ._. 
In_II the_AT GIS_NN2 context_NN1 such_DA issues_NN2 are_VBR becoming_VVG increasingly_RR important_JJ as_CSA GIS_NN2 use_NN1 grows_VVZ rapidly_RR and_CC reaches_VVZ a_AT1 wider_JJR audience_NN1 :_: this_DD1 concern_NN1 has_VHZ created_VVN a_AT1 new_JJ sub-discipline_NN1 of_IO '_GE spatial_JJ language_NN1 '_GE research_NN1 ._. 
Early_RR work_VV0 on_II spatial_JJ languages_NN2 was_VBDZ based_VVN upon_II the_AT principles_NN2 of_IO spatial_JJ inference_NN1 and_CC retrieval_NN1 ,_, and_CC was_VBDZ not_XX specifically_RR digitally_RR based_VVN (_( McDermott_NP1 1980_MC )_) ._. 
Thus_RR ,_, Palmer_NP1 and_CC Frank_NP1 (_( 1988_MC )_) distinguish_VV0 between_II the_AT interpretation_NN1 of_IO a_AT1 man-made_JJ image_NN1 (_( whose_DDQGE meaning_NN1 is_VBZ to_TO be_VBI reconstructed_VVN )_) ,_, and_CC a_AT1 '_GE natural_JJ '_GE image_NN1 (_( whose_DDQGE meaning_NN1 has_VHZ to_TO be_VBI extracted_VVN and_CC standardized_VVN with_IW belief_NN1 systems_NN2 )_) ._. 
Since_CS maps_NN2 are_VBR man-made_JJ the_AT chief_JJ objective_NN1 of_IO a_AT1 spatial_JJ language_NN1 is_VBZ to_TO guide_VVI in_II the_AT elucidation_NN1 of_IO the_AT meaning(s)_NN2 attached_VVN ._. 
Palmer_NP1 and_CC Frank_NP1 (_( 1988_MC )_) also_RR identify_VV0 the_AT key_JJ components_NN2 of_IO a_AT1 spatial_JJ language_NN1 as_II a_AT1 family_NN1 of_IO spatial_JJ objects_NN2 ,_, the_AT geometry_NN1 of_IO space_NN1 (_( as_CSA mapped_VVN and_CC perceived_VVN )_) and_CC the_AT mappings_NN2 of_IO the_AT objects_NN2 into_II the_AT space_NN1 ._. 
The_AT importance_NN1 of_IO the_AT cognitive_JJ structuring_NN1 of_IO space_NN1 is_VBZ expressed_VVN by_II Mark_NP1 and_CC Frank_NP1 (_( 1989_MC )_) ,_, who_PNQS set_VVD out_RP the_AT research_NN1 agenda_NN1 for_IF the_AT '_GE Spatial_JJ Languages_NN2 '_GE Research_NN1 Initiative_NN1 of_IO the_AT US_NP1 National_JJ Center_NN1 for_IF Geographic_JJ Information_NN1 and_CC Analysis_NN1 (_( NCGIA_NP1 )_) ._. 
Cognitive_JJ aspects_NN2 of_IO spatial_JJ perception_NN1 place_NN1 constraints_NN2 on_II the_AT definition_NN1 of_IO a_AT1 spatial_JJ language_NN1 intended_VVD to_TO be_VBI used_VVN as_II part_NN1 of_IO a_AT1 user_NN1 dialogue_NN1 with_IW the_AT computer_NN1 ._. 
Thus_RR ,_, Mark_NP1 (_( 1989_MC )_) suggests_VVZ that_CST the_AT '_GE image-schema_NN1 '_GE can_VM be_VBI used_JJ to_II structure_NN1 perceptions_NN2 of_IO space_NN1 and_CC provides_VVZ examples_NN2 drawn_VVN from_II Johnson_NP1 (_( 1987_MC )_) such_II21 as_II22 CONTAINER_NN1 (_( associated_VVN with_IW the_AT preposition_NN1 '_GE IN_II '_GE )_) and_CC PLATFORM_NN1 (_( associated_VVN with_IW the_AT preposition_NN1 '_GE ON_II '_GE )_) ._. 
This_DD1 consideration_NN1 can_VM affect_VVI the_AT process_NN1 of_IO data_NN modelling_VVG since_II the_AT cognitive_JJ models_NN2 available_JJ define_VV0 the_AT '_GE entities_NN2 thought_VVD to_TO be_VBI relevant_JJ to_II the_AT task_NN1 in_II hand_NN1 '_GE ._. 
Mark_NP1 (_( 1989_MC )_) applies_VVZ this_DD1 consideration_NN1 to_II the_AT digital_JJ line_NN1 graph_NN1 ,_, extended_VVN (_( DLG-E_NP1 )_) model_NN1 of_IO cartographic_JJ representation_NN1 used_VVN by_II the_AT US_NP1 Geological_JJ Survey_NN1 mapping_NN1 programme_NN1 and_CC comments_NN2 that_CST the_AT rules_NN2 and_CC feature_NN1 classes_NN2 adopted_VVD do_VD0 not_XX relate_VVI well_RR to_II an_AT1 image-schema_NN1 ,_, and_CC hence_RR will_VM be_VBI difficult_JJ to_TO generalize_VVI ._. 
Similarly_RR ,_, different_JJ linguistic_JJ traditions_NN2 structure_NN1 space_NN1 in_II distinct_JJ ways_NN2 (_( Mark_NP1 1988_MC )_) ._. 
Examples_NN2 of_IO linguistic_JJ structuring_NN1 (_( with_IW examples_NN2 in_II English_NN1 )_) include_VV0 :_: 1_MC1 ._. 
Preposition_NN1 ,_, e.g._REX near_JJ ,_, between_II ,_, among_II ;_; 2_MC ._. 
Motion_NN1 ,_, e.g._REX across_II ;_; 3_MC ._. 
Demonstratives_NN2 ,_, e.g._REX this_DD1 ,_, that_DD1 ;_; 4_MC ._. 
Viewpoint_NN1 ,_, e.g._REX further_JJR down_II the_AT road_NN1 ;_; 5_MC ._. 
Distribution_NN1 of_IO attention_NN1 ,_, e.g._REX reference_NN1 frame_NN1 such_II21 as_II22 a_AT1 field_NN1 ._. 
Awareness_NN1 of_IO these_DD2 characteristics_NN2 of_IO language_NN1 can_VM help_VVI in_II the_AT design_NN1 of_IO a_AT1 generic_JJ spatial_JJ language_NN1 with_IW which_DDQ a_AT1 wide_JJ variety_NN1 of_IO users_NN2 can_VM interact_VVI with_IW a_AT1 spatial_JJ database_NN1 ._. 
A_AT1 clear_JJ objective_NN1 in_II this_DD1 field_NN1 must_VM surely_RR be_VBI the_AT evaluation_NN1 of_IO the_AT language_NN1 of_IO groups_NN2 of_IO users_NN2 ._. 
However_RR ,_, attempts_NN2 to_TO translate_VVI the_AT Geographical_JJ Information_NN1 Systems_NN2 Tutor_NN1 (_( GISTutor_NP1 )_) (_( Raper_NP1 and_CC Green_JJ 1989_MC )_) into_II a_AT1 number_NN1 of_IO European_JJ languages_NN2 have_VH0 encountered_VVN two_MC main_JJ difficulties_NN2 :_: first_MD ,_, the_AT local_JJ adoption_NN1 of_IO English_NN1 for_IF spatial_JJ terms_NN2 (_( and_CC therefore_RR concepts_NN2 ?_? )_) due_II21 to_II22 the_AT English_JJ lexicon_NN1 of_IO most_RGT commercially_RR available_JJ GlS_NP1 ,_, and_CC second_NNT1 the_AT difficulty_NN1 of_IO providing_VVG for_IF translation_NN1 of_IO terms_NN2 with_IW no_AT direct_JJ counterpart_NN1 ._. 
The_AT spatial_JJ query_NN1 language_NN1 tools_NN2 which_DDQ a_AT1 GIS_NN2 uses_VVZ for_IF general_JJ interaction_NN1 with_IW spatial_JJ data_NN and_CC for_IF the_AT elucidation_NN1 of_IO spatial_JJ relationships_NN2 have_VH0 also_RR been_VBN the_AT subject_NN1 of_IO recent_JJ research_NN1 (_( Egenhofer_NP1 1989_MC )_) ._. 
However_RR ,_, spatial_JJ query_NN1 tools_NN2 are_VBR intrinsically_RR linked_VVN to_II the_AT data_NN model_NN1 employed_VVD :_: hence_RR the_AT use_NN1 of_IO multiple_JJ spatial_JJ data_NN structures_NN2 has_VHZ led_VVN to_II the_AT proliferation_NN1 of_IO query_NN1 processors_NN2 ._. 
The_AT result_NN1 has_VHZ been_VBN that_CST in_II many_DA2 GIS_NN2 spatial_JJ queries_NN2 initiated_VVN by_II the_AT user_NN1 must_VM be_VBI formulated_VVN in_II31 terms_II32 of_II33 a_AT1 software-specific_JJ command_NN1 language_NN1 ._. 
However_RR ,_, the_AT fact_NN1 that_CST many_DA2 new_JJ systems_NN2 use_VV0 relational_JJ database_NN1 storage_NN1 for_IF spatial_JJ and_CC non-spatial_JJ data_NN has_VHZ led_VVN to_II a_AT1 wide_JJ interest_NN1 in_II the_AT adaptation_NN1 of_IO the_AT structured_JJ query_NN1 language_NN1 (_( SQL_NN1 )_) for_IF spatially-related_JJ query_NN1 operations_NN2 in_II GIS_NN2 ._. 
Thus_RR ,_, SQL_NN1 is_VBZ now_RT seen_VVN as_CSA one_MC1 possible_JJ means_NN of_IO establishing_VVG spatial_JJ queries_NN2 across_II many_DA2 different_JJ spatial_JJ data_NN models_NN2 and_CC data_NN structures_NN2 ._. 
Hence_RR ,_, Herring_NN et_RA21 al_RA22 ._. 
(_( 1988_MC )_) described_VVD the_AT design_NN1 of_IO SQL_NN1 extensions_NN2 to_II a_AT1 new_JJ object-oriented_JJ GIS_NN2 ;_; Abel_NP1 (_( 1988_MC )_) has_VHZ reported_VVN work_NN1 on_II the_AT creation_NN1 of_IO SQL-based_JJ spatial_JJ extensions_NN2 to_II a_AT1 relational_JJ database_NN1 in_II the_AT '_GE Spatial_JJ Information_NN1 in_II a_AT1 Relational_JJ Open-architecture_JJ database_NN1 management_NN1 system_NN1 '_GE (_( SIRO-DBMS_NP1 )_) project_NN1 ;_; and_CC Ingram_NP1 and_CC Phillips_NP1 (_( 1988_MC )_) have_VH0 designed_VVN spatial_JJ extensions_NN2 to_II a_AT1 hybrid_JJ GIS_NN2 data_NN model_NN1 ._. 
Egenhofer_NP1 (_( 1989_MC )_) also_RR reported_VVD the_AT test_NN1 development_NN1 of_IO a_AT1 prototype_NN1 extended_VVD SQL_NN1 (_( XSQL_NP1 )_) ,_, adding_VVG concepts_NN2 of_IO spatial_JJ relations_NN2 ,_, spatial_JJ operations_NN2 ,_, graph_NN1 representation_NN1 and_CC context_NN1 ._. 
However_RR ,_, when_CS he_PPHS1 tested_VVD the_AT system_NN1 he_PPHS1 found_VVD that_CST the_AT queries_NN2 became_VVD far_RR too_RG long_JJ and_CC complex_JJ to_TO parse_VVI and_CC execute_VVI for_IF user-initiated_JJ queries_NN2 ._. 
This_DD1 may_VM indicate_VVI that_DD1 user_NN1 interfaces_NN2 to_TO query_VVI structures_NN2 can_VM not_XX yet_RR replicate_VVI human_JJ processes_NN2 of_IO spatial_JJ reasoning_NN1 in_II a_AT1 reasonable_JJ time_NNT1 ,_, or_CC that_CST in_II software_NN1 terms_NN2 it_PPH1 is_VBZ inefficient_JJ to_TO extend_VVI SQL_NN1 ,_, and_CC that_CST an_AT1 alternative_NN1 must_VM be_VBI found_VVN ._. 
Hence_RR the_AT link_NN1 between_II the_AT user_NN1 interface_NN1 and_CC the_AT spatial_JJ query_NN1 processor_NN1 is_VBZ a_AT1 problem_NN1 of_IO key_JJ importance_NN1 ,_, and_CC the_AT subject_NN1 of_IO considerable_JJ debate_NN1 ._. 
Egenhofer_NP1 (_( 1989_MC )_) ,_, in_II work_NN1 on_II the_AT testing_NN1 of_IO query_NN1 processors_NN2 ,_, suggests_VVZ that_CST it_PPH1 is_VBZ not_XX possible_JJ to_TO hide_VVI the_AT implementation_NN1 structure_NN1 of_IO the_AT query_NN1 at_II the_AT interface_NN1 ._. 
However_RR ,_, note_VV0 that_CST Rhind_NP1 et_RA21 al_RA22 ._. 
(_( 1989_MC )_) suggest_VV0 that_CST a_AT1 Universal_JJ Geographic_JJ Information_NN1 eXecutive_NN1 (_( UGIX_NP1 )_) can_VM be_VBI created_VVN to_TO interface_VVI between_II the_AT user_NN1 and_CC a_AT1 wide_JJ variety_NN1 of_IO different_JJ systems_NN2 ._. 
However_RR ,_, the_AT UGIX_NN1 concept_NN1 is_VBZ based_VVN on_II the_AT key_JJ assumption_NN1 that_CST current_JJ GIS_NN2 users_NN2 are_VBR motivated_VVN and_CC capable_JJ of_IO exercising_VVG spatial_JJ reasoning_NN1 ._. 
These_DD2 users_NN2 want_VV0 access_NN1 to_TO tools_NN2 to_TO extend_VVI the_AT spatial_JJ database_NN1 as_RG well_RR as_CSA to_TO make_VVI spatial_JJ analyses_NN2 ._. 
This_DD1 may_VM indicate_VVI that_CST a_AT1 different_JJ approach_NN1 is_VBZ required_VVN for_IF '_GE public_NN1 '_GE spatial_JJ information_NN1 systems_NN2 used_VVN by_II '_GE spatial_JJ professionals_NN2 '_GE ._. 
These_DD2 systems_NN2 may_VM have_VHI to_TO be_VBI restricted_VVN to_II passive_JJ use_NN1 based_VVN on_II the_AT extraction_NN1 of_IO information_NN1 in_II well-defined_JJ forms_NN2 ._. 
An_AT1 example_NN1 of_IO this_DD1 would_VM be_VBI the_AT systems_NN2 recently_RR installed_VVN in_II Shell_NN1 petrol_NN1 stations_NN2 in_II the_AT UK_NP1 developed_VVN by_II Action_NN1 Information_NN1 Management_NN1 which_DDQ use_VV0 touch_NN1 screen_NN1 technology_NN1 to_TO view_VVI '_GE road_NN1 atlas_NN1 '_GE type_NN1 maps_NN2 marked_VVN with_IW symbols_NN2 indicating_VVG travel_NN1 hazards_NN2 on_II major_JJ routes_NN2 ._. 
The_AT information_NN1 content_NN1 of_IO maps_NN2 is_VBZ also_RR an_AT1 issue_NN1 for_IF user_NN1 interfaces_NN2 ._. 
Thus_RR ,_, Bertin_NP1 (_( 1983_MC )_) defined_VVD the_AT concept_NN1 of_IO a_AT1 '_GE map-to-see_VV0 '_" as_58 '_GE a_AT1 clear_JJ graphic_JJ representation_NN1 which_DDQ can_VM be_VBI comprehended_VVN in_II a_AT1 short_JJ moment_NN1 '_GE ._. 
Such_DA an_AT1 efficient_JJ map_NN1 design_NN1 should_VM provide_VVI the_AT map_NN1 user_NN1 with_IW answers_NN2 to_II all_DB possible_JJ questions_NN2 related_VVN to_II the_AT map_NN1 ,_, whatever_DDQV their_APPGE type_NN1 or_CC complexity_NN1 ,_, in_II a_AT1 single_JJ moment_NN1 of_IO perception_NN1 ._. 
Every_AT1 other_JJ kind_NN1 of_IO map_NN1 must_VM be_VBI a_AT1 '_GE map-to-read_JJ '_GE and_CC requires_VVZ some_DD study_NN1 ;_; it_PPH1 is_VBZ implicit_JJ in_II experiments_NN2 carried_VVN out_RP by_II Kraak_NP1 (_( 1989_MC )_) that_CST such_DA maps_NN2 are_VBR not_XX suitable_JJ for_IF temporary_JJ display_NN1 on_II a_AT1 computer_NN1 screen_NN1 ._. 
Kraak_VV0 also_RR showed_VVD that_CST maps_NN2 illustrating_VVG the_AT values_NN2 of_IO a_AT1 variable_NN1 in_II 2D_NNU shaded_JJ polygons_NN2 best_RRT formed_VVN a_AT1 '_GE map-to-see_VV0 '_" when_RRQ they_PPHS2 were_VBDR displayed_VVN with_IW the_AT variable_NN1 visualized_VVD in_II 3D_NNU ._. 
The_AT design_NN1 of_IO help_NN1 systems_NN2 and_CC tutors_NN2 for_IF GIS_NN2 also_RR presents_VVZ some_DD significant_JJ challenges_NN2 to_II the_AT developers_NN2 of_IO GIS_NN2 user_NN1 interfaces_NN2 ._. 
However_RR ,_, the_AT availability_NN1 of_IO such_DA systems_NN2 may_VM be_VBI a_AT1 crucial_JJ component_NN1 of_IO the_AT user_NN1 interface_NN1 (_( Rhind_NP1 et_RA21 al_RA22 ._. 
1989_MC )_) ._. 
Difficulties_NN2 include_VV0 the_AT need_NN1 to_TO show_VVI complex_JJ graphic_JJ operations_NN2 ,_, the_AT explanation_NN1 of_IO algorithms_NN2 and_CC file_NN1 structures_NN2 ,_, and_CC the_AT application_NN1 of_IO GIS_NN2 techniques_NN2 to_II decision-making_NN1 in_II a_AT1 variety_NN1 of_IO application_NN1 areas_NN2 ._. 
One_MC1 example_NN1 of_IO a_AT1 solution_NN1 to_II this_DD1 problem_NN1 is_VBZ the_AT creation_NN1 of_IO GISTutor_NN1 which_DDQ is_VBZ described_VVN by_II Raper_NP1 and_CC Green_NP1 (_( 1989_MC )_) ._. 
Hence_RR at_RR21 present_RR22 three_MC main_JJ kinds_NN2 of_IO problems_NN2 are_VBR encountered_VVN in_II the_AT use_NN1 of_IO existing_JJ systems_NN2 :_: first_MD ,_, the_AT generally_RR inappropriate_JJ design_NN1 of_IO GUIs_NN1 for_IF GIS_NN2 purposes_NN2 ;_; second_MD ,_, the_AT specification_NN1 of_IO the_AT language_NN1 of_IO interaction_NN1 for_IF spatial_JJ operations_NN2 and_CC queries_NN2 ;_; and_CC ,_, third_MD ,_, the_AT limited_JJ content_NN1 and_CC rigid_JJ structure_NN1 of_IO standard_JJ help_NN1 systems_NN2 ._. 
It_PPH1 is_VBZ therefore_RR suggested_VVN that_CST the_AT design_NN1 of_IO a_AT1 user_NN1 interface_NN1 must_VM be_VBI rooted_VVN in_II the_AT creation_NN1 of_IO a_AT1 user_NN1 environment_NN1 which_DDQ integrates_VVZ the_AT data-modelling_JJ process_NN1 ,_, the_AT availability_NN1 of_IO spatial_JJ operations_NN2 and_CC queries_NN2 ,_, and_CC the_AT system_NN1 of_IO assistance_NN1 and_CC concept_NN1 support_NN1 within_II the_AT interface_NN1 ._. 
User_NN1 interface_NN1 development_NN1 :_: a_AT1 case_NN1 study_NN1 of_IO UGIX_NN1 In_II an_AT1 attempt_NN1 to_TO define_VVI the_AT basis_NN1 on_II which_DDQ a_AT1 comprehensive_JJ GIS_NN2 user_NN1 environment_NN1 could_VM be_VBI built_VVN Rhind_NP1 et_RA21 al_RA22 ._. 
(_( 1989_MC )_) proposed_VVD the_AT Universal_JJ Geographic_JJ Information_NN1 eXecutive_NN1 (_( UGIX_NP1 )_) as_II a_AT1 system_NN1 design_NN1 ._. 
It_PPH1 contains_VVZ three_MC main_JJ modules_NN2 viz._REX (_( A_ZZ1 )_) containing_VVG the_AT screen_NN1 interfaces_NN2 ,_, dialogues_NN2 and_CC spatial_JJ command_NN1 processor_NN1 ;_; (_( B_ZZ1 )_) containing_VVG a_AT1 help_NN1 and_CC information_NN1 system_NN1 for_IF a_AT1 GIS_NN2 ;_; and_CC (_( C_ZZ1 )_) an_AT1 expert_NN1 system_NN1 shell_NN1 or_CC high-level_JJ system_NN1 access_NN1 module_NN1 ._. 
The_AT structure_NN1 of_IO UGIX_NN1 is_VBZ illustrated_VVN in_II Fig._NN1 7.1_MC ._. 
This_DD1 section_NN1 describes_VVZ the_AT approach_NN1 taken_VVN in_II the_AT UGIX_NN1 project_NN1 ,_, through_II first_MD and_CC second_MD generation_NN1 implementations_NN2 ._. 
The_AT first_MD generation_NN1 approach_NN1 to_TO interface_VVI design_NN1 within_II the_AT UGIX_NN1 project_NN1 has_VHZ been_VBN to_II prototype_NN1 using_VVG Hypercard_NP1 for_IF the_AT Apple_NN1 Macintosh_NP1 ,_, where_CS the_AT Hypercard_NP1 application_NN1 (_( complete_JJ with_IW in-built_JJ communications_NN2 software_NN1 )_) acts_VVZ as_II a_AT1 client_NN1 to_II a_AT1 host_NN1 processor_NN1 '_GE running_VVG the_AT GIS_NN2 application_NN1 software_NN1 ._. 
This_DD1 approach_NN1 is_VBZ similar_JJ to_II the_AT one_PN1 used_VVD by_II Cowan_NP1 and_CC Love_NN1 (_( 1988_MC )_) to_TO create_VVI an_AT1 interface_NN1 to_II the_AT S_NP1 Carolina_NP1 Historic_JJ Preservation_NN1 Office_NN1 GIS_NN2 database_NN1 ._. 
Hypercard_NP1 with_IW its_APPGE standard_JJ set_NN1 of_IO buttons_NN2 ,_, scrolling_VVG boxes_NN2 and_CC cards_NN2 makes_VVZ use_NN1 of_IO the_AT GlS_NP1 less_RGR daunting_JJ for_IF the_AT less_RGR technically_RR aware_JJ user_NN1 ._. 
In_RR21 addition_RR22 ,_, with_IW the_AT rich_JJ graphics_NN environment_NN1 available_JJ in_II Hypercard_NP1 it_PPH1 is_VBZ possible_JJ to_TO show_VVI a_AT1 graphic_JJ to_TO illustrate_VVI the_AT effect_NN1 of_IO various_JJ options_NN2 available_JJ at_II any_DD one_MC1 point_NN1 ._. 
It_PPH1 is_VBZ also_RR desirable_JJ to_TO display_VVI all_DB the_AT commands_NN2 available_JJ to_II the_AT user_NN1 at_II one_MC1 time_NNT1 ,_, with_IW a_AT1 pop-up_NN1 explanation_NN1 for_IF each_DD1 option_NN1 ._. 
Screen_NN1 design_NN1 has_VHZ involved_VVN the_AT standardization_NN1 of_IO button_NN1 and_CC text_NN1 field_NN1 formats_NN2 as_II31 well_II32 as_II33 card_NN1 and_CC background_NN1 layout_NN1 for_IF different_JJ areas_NN2 of_IO activity_NN1 such_II21 as_II22 :_: ._. 
1_MC1 ._. 
Introduction_NN1 and_CC explanation_NN1 (_( using_VVG a_AT1 map_NN1 guide_NN1 )_) ;_; 2_MC ._. 
Map_NN1 and_CC file_NN1 selection_NN1 (_( using_VVG standard_JJ Macintosh_NP1 file_NN1 selection_NN1 dialogue_NN1 )_) ;_; 3_MC ._. 
Session_NNT1 screens_NN2 for_IF command_NN1 processing_NN1 ;_; 4_MC ._. 
Help_VV0 environment_NN1 (_( UGIX_NP1 (_( B_ZZ1 )_) based_VVN on_II GISTutor_NP1 II_MC )_) ;_; ,_, 5_MC ._. 
Library_NN1 for_IF maps_NN2 and_CC images_NN2 generated_VVN in_II the_AT GIS_NN2 (_( along_II21 with_II22 a_AT1 button_NN1 to_TO redraw_VVI them_PPHO2 )_) ._. 
Screen_NN1 metaphors_NN2 have_VH0 been_VBN developed_VVN for_IF each_DD1 of_IO these_DD2 areas_NN2 to_TO make_VVI location_NN1 in_II the_AT system_NN1 a_AT1 graphical_JJ attribute_NN1 ._. 
Currently_RR this_DD1 system_NN1 interface_NN1 '_GE shell_NN1 '_GE is_VBZ being_VBG implemented_VVN for_IF the_AT GIS_NN2 ARC/INFO_NN1 ._. 
During_II the_AT second_MD phase_NN1 of_IO the_AT UGIX_NN1 project_NN1 a_AT1 new_JJ interface_NN1 '_GE shell_NN1 '_GE is_VBZ being_VBG developed_VVN as_II part_NN1 of_IO a_AT1 generic_JJ GIS-independent_JJ approach_NN1 to_II GIS_NN2 user_NN1 interfaces_NN2 (_( Raper_NP1 and_CC Bundock_VV0 1991_MC )_) ._. 
The_AT first-generation_JJ system_NN1 known_VVN as_98 '_GE HyperArc_NP1 '_GE is_VBZ currently_RR under_II test_NN1 with_IW users_NN2 at_II '_GE novice_NN1 '_GE and_CC '_GE competent_JJ '_GE levels_NN2 of_IO expertise_NN1 :_: in_II31 addition_II32 to_II33 feedback_NN1 on_II the_AT use_NN1 of_IO the_AT system_NN1 ,_, the_AT aim_NN1 of_IO the_AT evaluation_NN1 phase_NN1 is_VBZ to_TO define_VVI a_AT1 core_NN1 area_NN1 of_IO functionality_NN1 in_RR21 common_RR22 use_VV0 to_TO help_VVI optimize_VVI the_AT system_NN1 structure_NN1 ._. 
A_AT1 key_JJ consideration_NN1 for_IF the_AT development_NN1 of_IO a_AT1 spatial_JJ language_NN1 interface_NN1 to_II GlS_NP1 is_VBZ that_CST many_DA2 different_JJ groups_NN2 of_IO users_NN2 exist_VV0 within_II the_AT spatial_JJ data-handling_JJ community_NN1 ._. 
It_PPH1 is_VBZ recognized_VVN within_II the_AT UGIX_NN1 design_NN1 that_CST to_II some_DD high-level_JJ users_NN2 skilled_JJ in_II the_AT manipulation_NN1 of_IO spatial_JJ data_NN a_AT1 command_NN1 line_NN1 interface_NN1 to_II a_AT1 GIS_NN2 offers_VVZ the_AT maximum_JJ flexibility_NN1 and_CC power_NN1 available_JJ ._. 
However_RR ,_, it_PPH1 is_VBZ considered_VVN that_CST during_II the_AT first_MD stage_NN1 of_IO the_AT expansion_NN1 of_IO GIS_NN2 use_VV0 the_AT largest_JJT group_NN1 of_IO new_JJ users_NN2 are_VBR non-computer-literate_JJ professionals_NN2 who_PNQS frequently_RR have_VH0 a_AT1 highly_RR developed_JJ insight_NN1 into_II spatial_JJ data_NN ._. 
It_PPH1 should_VM therefore_RR be_VBI emphasized_VVN that_CST the_AT first_MD objective_NN1 within_II the_AT UGIX_NN1 project_NN1 is_VBZ to_TO improve_VVI access_NN1 to_II existing_JJ GlS_NP1 by_II converting_VVG the_AT current_JJ function-orientation_NN1 of_IO the_AT native_JJ system_NN1 interface_NN1 to_II a_AT1 task-oriented_JJ interface_NN1 usable_JJ by_II a_AT1 spatially_RR aware_JJ user_NN1 (_( see_VV0 Fig._NN1 7.2_MC for_IF a_AT1 taxonomy_NN1 of_IO interfaces_NN2 for_IF ARC/INFO_NN1 at_II the_AT time_NNT1 of_IO writing_VVG in_II 1990_MC )_) ._. 
The_AT second_MD generation_NN1 of_IO the_AT UGIX_NN1 project_NN1 aims_VVZ to_TO build_VVI on_II the_AT experience_NN1 of_IO constructing_VVG such_DA task-oriented_JJ interfaces_NN2 to_TO create_VVI generic_JJ interfaces_NN2 capable_JJ of_IO communication_NN1 with_IW any_DD GIS_NN2 ._. 
The_AT basic_JJ principle_NN1 of_IO the_AT UGIX_NN1 design_NN1 is_VBZ that_CST in_BCL21 order_BCL22 to_TO make_VVI a_AT1 GIS_NN2 easy_JJ to_TO use_VVI the_AT process_NN1 of_IO making_VVG a_AT1 database_NN1 enquiry_NN1 ,_, plotting_VVG a_AT1 map_NN1 or_CC carrying_VVG out_RP spatial_JJ analysis_NN1 must_VM be_VBI broken_VVN down_RP into_II manageable_JJ parts_NN2 ,_, linked_VVN by_II a_AT1 pathway_NN1 for_IF the_AT user_NN1 to_TO follow_VVI ._. 
Following_VVG such_DA a_AT1 path_NN1 and_CC gaining_VVG experience_NN1 with_IW the_AT alternative_JJ options_NN2 is_VBZ an_AT1 excellent_JJ way_NN1 to_TO improve_VVI a_AT1 user_NN1 's_GE end-to-end_JJ understanding_NN1 of_IO the_AT components_NN2 of_IO spatial_JJ data_NN processing_NN1 ._. 
In_BCL21 order_BCL22 to_TO create_VVI this_DD1 form_NN1 of_IO interface_NN1 environment_NN1 this_DD1 project_NN1 is_VBZ engaged_VVN in_II the_AT definition_NN1 of_IO a_AT1 language_NN1 of_IO spatial_JJ interaction_NN1 to_TO handle_VVI a_AT1 dialogue_NN1 with_IW a_AT1 GIS_NN2 through_II a_AT1 system_NN1 shell_NN1 ._. 
In_II the_AT development_NN1 of_IO this_DD1 shell_NN1 it_PPH1 is_VBZ considered_VVN that_CST the_AT selection_NN1 of_IO options_NN2 itself_PPX1 should_VM drive_VVI the_AT commands_NN2 presented_VVN to_II the_AT user_NN1 for_IF action_NN1 ._. 
Appropriate_JJ information_NN1 needed_VVN for_IF a_AT1 user_NN1 to_TO make_VVI a_AT1 decision_NN1 is_VBZ also_RR retrieved_VVN before_II presenting_VVG the_AT command_NN1 options_NN2 ,_, for_REX21 example_REX22 ,_, maps_NN2 with_IW the_AT correct_JJ specifications_NN2 (_( e.g._REX with_IW topological_JJ relations_NN2 already_RR created_VVN ,_, when_CS this_DD1 is_VBZ necessary_JJ for_IF the_AT operation_NN1 )_) ._. 
Thus_RR ,_, the_AT structuring_NN1 of_IO a_AT1 wide_JJ range_NN1 of_IO task-related_JJ paths_NN2 through_II a_AT1 GIS_NN2 has_VHZ formed_VVN one_MC1 of_IO the_AT main_JJ development_NN1 objectives_NN2 in_II this_DD1 project_NN1 ._. 
Another_DD1 key_JJ objective_NN1 in_II developing_VVG a_AT1 GIS_NN2 user_NN1 interface_NN1 is_VBZ to_TO improve_VVI the_AT data-handling_JJ procedures_NN2 to_TO support_VVI the_AT user_NN1 's_GE concept_NN1 of_IO maps_NN2 as_CSA views_NN2 of_IO spatial_JJ data_NN ._. 
Thus_RR ,_, utilities_NN2 to_TO manage_VVI map_NN1 files_NN2 are_VBR being_VBG implemented_VVN within_II HyperArc_NP1 which_DDQ use_VV0 icons_NN2 to_TO access_VVI ARC/INFO_NN1 data_NN ._. 
Search_NN1 routines_NN2 to_TO find_VVI maps_NN2 with_IW particular_JJ names_NN2 ,_, to_TO sort_VVI maps_NN2 by_II type_NN1 (_( e.g._REX point_NN1 or_CC polygon_NN1 based_VVN )_) ,_, to_TO access_VVI the_AT map_NN1 tiling_NN1 system_NN1 and_CC to_TO select_VVI the_AT part_NN1 of_IO the_AT sheet_NN1 to_TO view_VVI have_VH0 also_RR been_VBN created_VVN ._. 
In_II the_AT first_MD generation_NN1 of_IO this_DD1 project_NN1 the_AT user_NN1 specifies_VVZ spatial_JJ queries_NN2 using_VVG system_NN1 implementation_NN1 concepts_NN2 which_DDQ are_VBR made_VVN comprehensible_JJ (_( and_CC refined_JJ through_II testing_VVG )_) ._. 
However_RR ,_, a_AT1 new_JJ query_NN1 language_NN1 interface_NN1 based_VVN on_II SQL_NN1 is_VBZ under_II development_NN1 for_IF the_AT second_MD generation_NN1 of_IO UGIX_NP1 (_( Raper_NP1 and_CC Bundock_VV0 1991_MC )_) ._. 
Conclusions_NN2 This_DD1 chapter_NN1 has_VHZ attempted_VVN to_TO summarize_VVI the_AT recent_JJ developments_NN2 in_II user_NN1 interfaces_NN2 for_IF GIS_NN2 ._. 
It_PPH1 can_VM be_VBI said_VVN in_II conclusion_NN1 that_CST these_DD2 developments_NN2 have_VH0 become_VVN central_JJ to_II the_AT future_NN1 of_IO the_AT '_GE GIS_NN2 revolution_NN1 '_GE of_IO the_AT late_JJ 1980s_MC2 ,_, since_CS it_PPH1 is_VBZ now_RT clear_VV0 that_CST the_AT massive_JJ (_( and_CC desirable_JJ )_) growth_NN1 of_IO the_AT use_NN1 of_IO GIS_NN2 can_VM not_XX be_VBI supported_VVN without_IW improvements_NN2 to_II the_AT use_NN1 environment_NN1 experienced_VVN by_II the_AT user_NN1 ._. 
It_PPH1 is_VBZ predicted_VVN that_CST such_DA developments_NN2 will_VM gain_VVI common_JJ ground_NN1 in_II the_AT elaboration_NN1 of_IO the_AT GUI_JJ concept_NN1 as_CSA it_PPH1 has_VHZ in_II the_AT computer-aided_JJ design_NN1 world_NN1 ,_, but_CCB will_VM quickly_RR diversify_VVI in_II31 terms_II32 of_II33 the_AT actual_JJ products_NN2 which_DDQ emerge_VV0 ._. 
Already_RR a_AT1 variety_NN1 of_IO vendors_NN2 are_VBR attempting_VVG to_TO create_VVI user_NN1 interfaces_NN2 which_DDQ go_VV0 beyond_II GUIs_NP1 ,_, but_CCB not_XX have_VH0 appeared_VVN at_II the_AT time_NNT1 of_IO writing_VVG in_II 1990_MC ._. 
Finally_RR ,_, it_PPH1 is_VBZ important_JJ to_TO note_VVI that_CST the_AT user_NN1 interface_NN1 '_GE debate_NN1 '_GE can_NN1 also_RR be_VBI interpreted_VVN as_CSA another_DD1 facet_NN1 of_IO the_AT data-modelling_JJ debate_NN1 in_II GIS_NN2 ._. 
User_NN1 interfaces_NN2 are_VBR not_XX just_RR pretty_JJ screen_NN1 representations_NN2 :_: as_CSA their_APPGE use_NN1 is_VBZ extended_VVN they_PPHS2 will_VM come_VVI to_TO express_VVI the_AT whole_JJ nature_NN1 of_IO the_AT system_NN1 data_NN model_NN1 ,_, and_CC will_VM probably_RR become_VVI highly_RR specialized_VVN as_II the_AT interfaces_NN2 move_VV0 from_II function-oriented_NN1 to_II task-oriented_NN1 forms_NN2 ._. 
Hence_RR in_II the_AT 1990s_MC2 the_AT user_NN1 may_VM not_XX gain_VVI freedom_NN1 of_IO interaction_NN1 with_IW spatial_JJ data_NN but_CCB become_VV0 a_AT1 captive_NN1 of_IO a_AT1 particular_JJ design_NN1 and_CC control_NN1 purpose_NN1 ._. 
Hence_RR ,_, the_AT challenge_NN1 for_IF GIS_NN2 user_NN1 interface_NN1 design_NN1 is_VBZ to_TO gain_VVI credibility_NN1 for_IF generic_JJ forms_NN2 of_IO interface_NN1 :_: it_PPH1 is_VBZ this_DD1 form_NN1 of_IO GIS_NN2 use_VV0 that_CST will_VM ensure_VVI wider_JJR and_CC more_RGR informed_JJ use_NN1 of_IO GIS_NN2 in_II the_AT years_NNT2 ahead_RL ._. 
Acknowledgements_NN2 The_AT authors_NN2 of_IO this_DD1 chapter_NN1 acknowledges_VVZ the_AT help_NN1 of_IO ESRI_NN2 in_II the_AT funding_NN1 of_IO work_NN1 leading_VVG to_II the_AT prototyping_NN1 of_IO the_AT first_MD generation_NN1 of_IO UGIX_NN1 ._. 
Apple_NN1 Computer_NN1 (_( UK_NP1 )_) have_VH0 also_RR assisted_VVN in_II this_DD1 work_NN1 by_II providing_VVG laboratory_NN1 Macintosh_NP1 computers_NN2 which_DDQ now_RT form_VV0 the_AT Apple_NN1 Mapping_NN1 Centre_NN1 at_II Birkbeck_NP1 College_NN1 ._. 
Potential_JJ Applications_NN2 Overview_NN1 Ian_NP1 Masser_NP1 Given_VVN the_AT extent_NN1 to_II which_DDQ geographic_JJ information_NN1 management_NN1 is_VBZ applications_NN2 driven_VVN ,_, it_PPH1 is_VBZ important_JJ to_TO consider_VVI in_II some_DD depth_NN1 the_AT characteristics_NN2 of_IO some_DD potential_JJ application_NN1 fields_NN2 ._. 
In_II the_AT process_NN1 it_PPH1 will_VM be_VBI necessary_JJ to_TO examine_VVI not_XX only_RR the_AT extent_NN1 to_II which_DDQ geographic_JJ issues_NN2 feature_VV0 on_II the_AT policy_NN1 issues_NN2 agenda_NN1 in_II these_DD2 fields_NN2 but_CCB also_RR the_AT institutional_JJ context_NN1 which_DDQ governs_VVZ decision-making_JJ and_CC the_AT extent_NN1 to_II which_DDQ data_NN are_VBR available_JJ for_IF research_NN1 and_CC policy_NN1 analysis_NN1 purposes_NN2 ._. 
Questions_NN2 on_II data_NN availability_NN1 are_VBR likely_JJ to_TO be_VBI particularly_RR important_JJ where_CS users_NN2 are_VBR heavily_RR dependent_JJ on_II secondary_JJ sources_NN2 ._. 
In_II situations_NN2 of_IO this_DD1 kind_NN1 provisions_NN2 governing_VVG access_NN1 and_CC format_NN1 may_VM exert_VVI powerful_JJ constraints_NN2 on_II the_AT use_NN1 that_CST can_VM be_VBI made_VVN of_IO these_DD2 data_NN ._. 
The_AT first_MD four_MC chapters_NN2 in_II this_DD1 part_NN1 examine_VV0 issues_NN2 relating_VVG to_II geographic_JJ information_NN1 handling_VVG in_II31 terms_II32 of_II33 four_MC potential_JJ applications_NN2 fields_NN2 ._. 
The_AT authors_NN2 were_VBDR asked_VVN to_TO review_VVI the_AT current_JJ state_NN1 of_IO the_AT art_NN1 in_II their_APPGE fields_NN2 with_IW particular_JJ reference_NN1 to_II the_AT likely_JJ requirements_NN2 of_IO the_AT future_JJ policy_NN1 issues_NN2 agenda_NN1 in_II31 terms_II32 of_II33 the_AT demands_NN2 that_CST they_PPHS2 make_VV0 on_II geographic_JJ information_NN1 management_NN1 ._. 
Authors_NN2 were_VBDR also_RR asked_VVN to_TO evaluate_VVI critically_RR the_AT state_NN1 of_IO geographic_JJ information_NN1 provision_NN1 in_II31 terms_II32 of_II33 the_AT needs_NN2 of_IO users_NN2 and_CC to_TO consider_VVI the_AT impact_NN1 of_IO user_NN1 requirements_NN2 on_II the_AT development_NN1 of_IO geographic_JJ information_NN1 handling_NN1 methodology_NN1 ._. 
In_II this_DD1 way_NN1 the_AT chapters_NN2 in_II this_DD1 part_NN1 present_NN1 a_AT1 number_NN1 of_IO perspectives_NN2 which_DDQ essentially_RR complement_VV0 and_CC amplify_VV0 the_AT discussion_NN1 in_II the_AT previous_JJ part_NN1 ._. 
In_II this_DD1 case_NN1 ,_, however_RR ,_, methodological_JJ issues_NN2 are_VBR tackled_VVN from_II a_AT1 largely_RR user_NN1 standpoint_NN1 ,_, whereas_CS in_II the_AT previous_JJ part_NN1 they_PPHS2 were_VBDR viewed_VVN as_CSA substantive_JJ problems_NN2 in_II their_APPGE own_DA right_NN1 ._. 
In_II effect_NN1 the_AT product_NN1 of_IO the_AT discussion_NN1 in_II both_DB2 parts_NN2 is_VBZ a_AT1 matrix_NN1 where_CS the_AT methodological_JJ issues_NN2 are_VBR listed_VVN along_II one_MC1 axis_NN1 and_CC particular_JJ applications_NN2 fields_NN2 on_II the_AT other_JJ ._. 
Figure_NN1 8.1_MC summarizes_VVZ the_AT main_JJ features_NN2 of_IO this_DD1 matrix_NN1 ._. 
There_EX are_VBR important_JJ differences_NN2 between_II the_AT chapters_NN2 in_II this_DD1 part_NN1 both_RR in_II the_AT nature_NN1 of_IO the_AT demands_NN2 that_CST they_PPHS2 make_VV0 on_II geographic_JJ information_NN1 and_CC also_RR in_II the_AT relative_JJ importance_NN1 that_CST is_VBZ attached_VVN to_II geographic_JJ criteria_NN2 as_II21 against_II22 other_JJ issues_NN2 by_II planners_NN2 and_CC decision-makers_NN2 in_II these_DD2 fields_NN2 ._. 
The_AT first_MD three_MC chapters_NN2 dealing_VVG with_IW environmental_JJ monitoring_NN1 ,_, natural_JJ and_CC technological_JJ hazards_NN2 and_CC settlements_NN2 and_CC infrastructure_NN1 applications_NN2 make_VV0 use_NN1 of_IO a_AT1 very_RG wide_JJ range_NN1 of_IO geographic_JJ information_NN1 ._. 
A_AT1 major_JJ feature_NN1 of_IO these_DD2 applications_NN2 is_VBZ the_AT extent_NN1 to_II which_DDQ they_PPHS2 involve_VV0 overlays_NN2 of_IO a_AT1 variety_NN1 of_IO information_NN1 drawn_VVN from_II different_JJ sources_NN2 on_II a_AT1 topographic_JJ map_NN1 base._NNU for_IF this_DD1 reason_NN1 they_PPHS2 make_VV0 particularly_RR heavy_JJ demands_NN2 on_II current_JJ GIS_NN2 technology_NN1 both_RR because_II21 of_II22 the_AT size_NN1 of_IO the_AT data_NN sets_NN2 and_CC the_AT diversity_NN1 of_IO the_AT data_NN structures_NN2 that_CST are_VBR involved_VVN ._. 
Brown_NP1 's_GE chapter_NN1 on_II geodemographics_NN1 discusses_VVZ an_AT1 application_NN1 field_NN1 where_CS primary_JJ data_NN collected_VVN for_IF operational_JJ purposes_NN2 are_VBR typically_RR evaluated_VVN in_II31 relation_II32 to_II33 secondary_JJ data_NN derived_VVN from_II sources_NN2 such_II21 as_II22 the_AT Census_NN1 of_IO Population_NN1 ._. 
The_AT range_NN1 of_IO data_NN sets_NN2 involved_VVN and_CC the_AT demands_NN2 that_CST are_VBR made_VVN on_II GIS_NN2 technology_NN1 in_II applications_NN2 of_IO this_DD1 kind_NN1 are_VBR often_RR more_RGR limited_JJ than_CSN those_DD2 of_IO previous_JJ applications_NN2 fields_NN2 ._. 
However_RR ,_, this_DD1 is_VBZ offset_VVN by_II the_AT demands_NN2 that_CST are_VBR made_VVN on_II spatial_JJ analysis_NN1 and_CC modelling_VVG methodology_NN1 ._. 
Rhind_NP1 's_GE chapter_NN1 on_II environmental_JJ monitoring_NN1 and_CC prediction_NN1 discusses_VVZ the_AT role_NN1 of_IO geographic_JJ information_NN1 management_NN1 in_II promoting_VVG sustainable_JJ development_NN1 on_II a_AT1 global_JJ scale_NN1 in_II the_AT context_NN1 of_IO the_AT issues_NN2 identified_VVN in_II the_AT Brundtland_NP1 report_NN1 ._. 
It_PPH1 points_VVZ to_II the_AT impact_NN1 that_CST the_AT findings_NN2 of_IO global_JJ environmental_JJ research_NN1 are_VBR likely_JJ to_TO have_VHI on_II most_DAT sectors_NN2 of_IO the_AT UK_NP1 economy_NN1 ._. 
This_DD1 contribution_NN1 also_RR indicates_VVZ the_AT wide_JJ variety_NN1 of_IO groups_NN2 that_CST are_VBR involved_JJ in_II monitoring_VVG environmental_JJ change_NN1 and_CC draws_VVZ attention_NN1 to_II the_AT role_NN1 that_CST the_AT European_JJ Community_NN1 is_VBZ increasingly_RR playing_VVG in_II co-ordinating_VVG these_DD2 efforts_NN2 ._. 
One_MC1 of_IO the_AT examples_NN2 of_IO work_NN1 in_II progress_NN1 discusses_VVZ research_NN1 that_CST has_VHZ been_VBN carried_VVN out_RP for_IF the_AT European_JJ Commission_NN1 as_II part_NN1 of_IO the_AT CORINE_JJ Project_NN1 ._. 
This_DD1 highlights_VVZ the_AT challenges_NN2 for_IF research_NN1 in_II a_AT1 field_NN1 categorized_VVD by_II a_AT1 lack_NN1 of_IO data_NN for_IF some_DD countries_NN2 ,_, varying_VVG definitions_NN2 and_CC standards_NN2 in_II use_NN1 ,_, and_CC unreasonably_RR high_JJ user_NN1 expectations_NN2 ._. 
The_AT chapter_NN1 as_II a_AT1 whole_NN1 draws_VVZ attention_NN1 to_II a_AT1 number_NN1 of_IO key_JJ methodological_JJ issues_NN2 that_CST are_VBR involved_JJ in_II environmental_JJ monitoring_NN1 and_CC prediction_NN1 ._. 
These_DD2 include_VV0 the_AT highly_RR specialized_JJ nature_NN1 of_IO much_DA1 of_IO the_AT technical_JJ analysis_NN1 ,_, the_AT limited_JJ capacity_NN1 for_IF checking_VVG the_AT accuracy_NN1 of_IO many_DA2 of_IO the_AT findings_NN2 of_IO this_DD1 kind_NN1 of_IO research_NN1 and_CC the_AT consequent_JJ need_NN1 for_IF the_AT development_NN1 of_IO knowledge-based_JJ inference_NN1 machines_NN2 for_IF extracting_VVG useful_JJ information_NN1 from_II secondary_NN1 and_CC often_RR proxy_NN1 data_NN ._. 
In_II Chapter_NN1 10_MC Gatrell_NP1 and_CC Vincent_NP1 consider_VV0 a_AT1 wide_JJ range_NN1 of_IO natural_JJ and_CC technological_JJ hazards_NN2 including_II industrial_JJ hazards_NN2 ,_, risk_NN1 assessment_NN1 ,_, hazardous_JJ waste_NN1 disposal_NN1 ,_, emergency_NN1 planning_NN1 ,_, natural_JJ hazards_NN2 and_CC environmental_JJ epidemiology_NN1 ._. 
The_AT substantive_JJ content_NN1 of_IO much_DA1 of_IO this_DD1 research_NN1 straddles_VVZ the_AT social_JJ ,_, environmental_JJ and_CC mathematical_JJ sciences_NN2 ._. 
To_TO model_VVI the_AT dispersal_NN1 of_IO a_AT1 plume_NN1 of_IO toxic_JJ gas_NN1 generated_VVN by_II a_AT1 chemical_JJ explosion_NN1 and_CC its_APPGE impact_NN1 on_II the_AT surrounding_JJ area_NN1 ,_, for_REX21 example_REX22 ,_, an_AT1 understanding_NN1 of_IO atmospheric_JJ dispersion_NN1 models_NN2 ,_, epidemiology_NN1 and_CC population_NN1 distribution_NN1 would_VM be_VBI required_VVN ._. 
In_II applications_NN2 of_IO this_DD1 kind_NN1 the_AT authors_NN2 argue_VV0 that_CST geographic_JJ information_NN1 management_NN1 in_RR21 general_RR22 and_CC GIS_NN2 in_RR21 particular_RR22 have_VH0 an_AT1 important_JJ role_NN1 to_TO play_VVI and_CC they_PPHS2 give_VV0 a_AT1 number_NN1 of_IO examples_NN2 from_II research_NN1 in_II progress_NN1 to_TO illustrate_VVI the_AT current_JJ state_NN1 of_IO the_AT art_NN1 in_II this_DD1 field_NN1 ._. 
The_AT discussion_NN1 of_IO operational_JJ problems_NN2 and_CC prospects_NN2 further_RRR amplifies_VVZ many_DA2 of_IO the_AT issues_NN2 raised_VVN in_II preceding_JJ chapters_NN2 ,_, particularly_RR with_II31 respect_II32 to_II33 the_AT problems_NN2 caused_VVN by_II the_AT dependency_NN1 on_II secondary_JJ data_NN sources_NN2 and_CC the_AT need_NN1 for_IF much_DA1 more_DAR detailed_JJ geographic_JJ information_NN1 to_TO overcome_VVI some_DD of_IO the_AT problems_NN2 associated_VVN with_IW linking_VVG data_NN sets_NN2 ._. 
Like_II most_DAT of_IO the_AT other_JJ contributors_NN2 to_II this_DD1 section_NN1 ,_, Gatrell_NP1 and_CC Vincent_NP1 see_VV0 decision_NN1 support_NN1 systems_NN2 as_II a_AT1 major_JJ area_NN1 for_IF future_JJ development_NN1 in_II the_AT geographic_JJ information_NN1 management_NN1 field_NN1 ._. 
Given_CS21 that_CS22 '_" most_DAT GIS_NN2 are_VBR rather_RG dumb_JJ systems_NN2 ,_, requiring_VVG intelligent_JJ ,_, very_JJ computer_NN1 literate_JJ users_NN2 '_GE ,_, managing_VVG natural_JJ and_CC technological_JJ hazards_NN2 means_VVZ that_CST this_DD1 knowledge_NN1 base_NN1 must_VM be_VBI built_VVN into_II the_AT system_NN1 so_CS21 that_CS22 it_PPH1 can_VM be_VBI utilized_VVN quickly_RR by_II untrained_JJ users_NN2 after_II a_AT1 disaster_NN1 has_VHZ taken_VVN place_NN1 or_CC in_II an_AT1 emergency_NN1 ._. 
The_AT settlement_NN1 and_CC infrastructure_NN1 field_NN1 is_VBZ also_RR characterized_VVN by_II a_AT1 wide_JJ range_NN1 of_IO users_NN2 with_IW a_AT1 great_JJ diversity_NN1 of_IO interests_NN2 ._. 
These_DD2 include_VV0 not_XX only_RR central_JJ and_CC local_JJ government_NN1 but_CCB also_RR major_JJ utility_NN1 companies_NN2 and_CC transport_NN1 suppliers_NN2 ._. 
According_II21 to_II22 Shepherd_NP1 these_DD2 users_NN2 share_VV0 a_AT1 common_JJ need_NN1 for_IF up-to-date_JJ information_NN1 on_II the_AT amount_NN1 ,_, capacity_NN1 and_CC condition_NN1 of_IO urban_JJ land_NN1 ,_, transport_NN1 networks_NN2 ,_, utilities_NN2 ,_, hospitals_NN2 ,_, schools_NN2 and_CC other_JJ major_JJ communal_JJ facilities_NN2 ._. 
In_II31 addition_II32 to_II33 these_DD2 data_NN requirements_NN2 ,_, such_DA users_NN2 are_VBR also_RR interested_JJ in_II developing_VVG their_APPGE capabilities_NN2 for_IF modelling_VVG the_AT changing_JJ relationships_NN2 between_II the_AT supply_NN1 and_CC demand_VV0 for_IF infrastructure_NN1 in_II the_AT context_NN1 of_IO both_RR public_JJ and_CC private_JJ sector_NN1 decision-making_NN1 ._. 
Shepherd_NN1 ,_, more_RRR than_CSN any_DD other_JJ contributor_NN1 to_II this_DD1 section_NN1 ,_, draws_VVZ attention_NN1 to_II the_AT extent_NN1 to_II which_DDQ current_JJ efforts_NN2 in_II the_AT geographic_JJ information_NN1 management_NN1 field_NN1 are_VBR constrained_VVN by_II the_AT highly_RR fragmented_JJ nature_NN1 of_IO both_RR data_NN suppliers_NN2 and_CC users_NN2 both_RR in_II sectoral_NN1 and_CC spatial_JJ terms_NN2 ,_, the_AT great_JJ diversity_NN1 of_IO data_NN sets_VVZ in_II varying_JJ formats_NN2 and_CC the_AT difficulties_NN2 presented_VVN by_II institutional_JJ factors_NN2 such_II21 as_II22 copyright_NN1 provision_NN1 ,_, data_NN ownership_NN1 and_CC the_AT control_NN1 of_IO access_NN1 to_II data_NN ._. 
Digital_JJ data_NN occupy_VV0 a_AT1 central_JJ position_NN1 in_II this_DD1 applications_VVZ field_NN1 and_CC a_AT1 number_NN1 of_IO examples_NN2 are_VBR given_VVN to_TO illustrate_VVI the_AT extent_NN1 to_II which_DDQ their_APPGE development_NN1 has_VHZ been_VBN hampered_VVN or_CC facilitated_VVN in_II various_JJ countries_NN2 by_II the_AT policies_NN2 of_IO the_AT organizations_NN2 with_IW responsibility_NN1 for_IF digital_JJ data_NN provision_NN1 ._. 
In_II Chapter_NN1 12_MC Brown_NP1 reviews_VVZ the_AT development_NN1 of_IO geodemographics_NN2 from_II its_APPGE origins_NN2 in_II the_AT computer-based_JJ social_JJ area_NN1 analysis_NN1 of_IO US_NP1 census_NN1 tract_NN1 data_NN from_II 1950_MC onwards_RL ._. 
He_PPHS1 describes_VVZ the_AT features_NN2 of_IO the_AT four_MC main_JJ systems_NN2 that_CST are_VBR currently_RR in_II use_NN1 in_II the_AT UK_NP1 and_CC illustrates_VVZ their_APPGE application_NN1 with_II31 reference_II32 to_II33 examples_NN2 drawn_VVN from_II retail_JJ analysis_NN1 and_CC store_NN1 location_NN1 ,_, market_VV0 research_NN1 ,_, credit_NN1 rating_NN1 and_CC target_NN1 marketing_NN1 ._. 
Brown_NP1 demonstrates_VVZ the_AT extent_NN1 to_II which_DDQ the_AT growth_NN1 of_IO research_NN1 activities_NN2 in_II this_DD1 field_NN1 has_VHZ led_VVN to_II the_AT emergence_NN1 of_IO improved_JJ cluster_NN1 analytic_JJ methods_NN2 for_IF the_AT multidimensional_JJ classification_NN1 of_IO people_NN according_II21 to_II22 the_AT type_NN1 of_IO area_NN1 in_II which_DDQ they_PPHS2 live_VV0 ._. 
In_II the_AT last_MD part_NN1 of_IO this_DD1 contribution_NN1 ,_, Brown_NP1 identifies_VVZ a_AT1 number_NN1 of_IO priorities_NN2 for_IF future_JJ research_NN1 ._. 
Like_II Gatrell_NP1 and_CC Vincent_NP1 and_CC Shepherd_NN1 before_II him_PPHO1 ,_, he_PPHS1 is_VBZ concerned_JJ with_IW the_AT lack_NN1 of_IO detailed_JJ geographical_JJ information_NN1 particularly_RR in_II this_DD1 case_NN1 with_II31 reference_II32 to_II33 the_AT years_NNT2 between_II censuses_NN2 ._. 
Spatial_JJ linkages_NN2 also_RR feature_VV0 prominently_RR in_II Brown_NP1 's_GE discussion_NN1 which_DDQ amplifies_VVZ many_DA2 of_IO the_AT points_NN2 made_VVN in_II the_AT earlier_JJR contribution_NN1 from_II Openshaw_NP1 in_II Chapter_NN1 3_MC ._. 
The_AT analysis_NN1 points_VVZ to_II the_AT need_NN1 for_IF the_AT development_NN1 of_IO a_AT1 fuzzy_JJ geodemographics_NN1 methodology_NN1 which_DDQ takes_VVZ account_NN1 of_IO the_AT effects_NN2 of_IO different_JJ types_NN2 of_IO spatial_JJ errors_NN2 ._. 
Other_JJ major_JJ research_NN1 priorities_NN2 for_IF the_AT future_NN1 in_II this_DD1 applications_VVZ field_NN1 are_VBR the_AT development_NN1 of_IO response-modelling_JJ techniques_NN2 to_TO explore_VVI the_AT factors_NN2 affecting_VVG consumer_NN1 choice_NN1 ._. 
This_DD1 is_VBZ particularly_RR important_JJ in_II a_AT1 field_NN1 where_RRQ many_DA2 of_IO the_AT demands_NN2 made_VVN by_II users_NN2 are_VBR highly_RR specialized_VVN in_II nature_NN1 ._. 
In_II summary_NN1 ,_, then_RT ,_, as_CSA fig._NN1 8.1_MC shows_NN2 ,_, many_DA2 of_IO the_AT issues_NN2 raised_VVN by_II this_DD1 review_NN1 of_IO geographic_JJ information_NN1 handling_VVG in_II the_AT four_MC potential_JJ application_NN1 fields_NN2 are_VBR essentially_RR complementary_JJ to_II those_DD2 discussed_VVN in_II the_AT part_NN1 on_II methodological_JJ developments_NN2 ._. 
Spatial_JJ analysis_NN1 features_NN2 particularly_RR prominent_JJ on_II the_AT research_NN1 agenda_NN1 relating_VVG to_II natural_JJ and_CC technological_JJ hazards_NN2 and_CC geodemographics_NN2 ._. 
Questions_NN2 relating_VVG to_II data_NN integration_NN1 emerge_VV0 as_CSA being_VBG particularly_RR critical_JJ in_II31 terms_II32 of_II33 environmental_JJ monitoring_NN1 and_CC settlement_NN1 and_CC infrastructure_NN1 applications_NN2 ._. 
Spatial_JJ linkages_NN2 and_CC questions_NN2 relating_VVG to_II spatial_JJ representation_NN1 feature_NN1 prominently_RR in_II the_AT settlements_NN2 and_CC infrastructure_NN1 and_CC geodemographics_NN1 fields_NN2 ._. 
The_AT evaluation_NN1 of_IO error_NN1 propagation_NN1 is_VBZ a_AT1 matter_NN1 of_IO considerable_JJ significance_NN1 in_II the_AT environmental_JJ monitoring_NN1 and_CC settlements_NN2 and_CC infrastructure_NN1 fields_NN2 ._. 
Finally_RR ,_, in_II all_DB applications_NN2 fields_NN2 ,_, the_AT findings_NN2 of_IO the_AT reviews_NN2 above_RL suggest_VV0 that_CST there_EX is_VBZ a_AT1 need_NN1 for_IF improvements_NN2 in_II user_NN1 interfaces_NN2 of_IO all_DB kinds_NN2 ._. 
This_DD1 is_VBZ particularly_RR evident_JJ in_II the_AT demand_NN1 for_IF the_AT development_NN1 of_IO knowledge-based_JJ decision_NN1 support_NN1 systems_NN2 and_CC modelling_VVG capabilities_NN2 ._. 
Such_DA demands_NN2 highlight_VV0 the_AT extent_NN1 to_II which_DDQ geographic_JJ information_NN1 management_NN1 must_VM be_VBI regarded_VVN as_II a_AT1 means_NN to_II an_AT1 end_NN1 rather_II21 than_II22 an_AT1 end_NN1 in_II itself_PPX1 when_CS viewed_VVN from_II the_AT standpoints_NN2 of_IO the_AT needs_NN2 of_IO planners_NN2 and_CC decision-makers_NN2 ._. 
The_AT final_JJ chapter_NN1 in_II this_DD1 section_NN1 rounds_VVZ off_II the_AT discussion_NN1 of_IO applications-related_JJ research_NN1 with_IW an_AT1 examination_NN1 of_IO organizational_JJ issues_NN2 that_CST need_VV0 to_TO be_VBI taken_VVN account_NN1 of_IO in_II handling_VVG geographic_JJ information_NN1 ._. 
Campbell_NP1 presents_VVZ a_AT1 conceptual_JJ framework_NN1 for_IF the_AT analysis_NN1 of_IO the_AT organizational_JJ issues_NN2 involved_JJ in_II the_AT development_NN1 and_CC implementation_NN1 of_IO GIS_NN2 ._. 
This_DD1 contribution_NN1 cuts_VVZ across_II many_DA2 of_IO the_AT questions_NN2 relating_VVG to_II particular_JJ applications_NN2 that_CST were_VBDR raised_VVN in_II preceding_JJ chapters_NN2 and_CC draws_VVZ attention_NN1 to_II the_AT dangers_NN2 of_IO neglecting_VVG institutional_JJ and_CC personal_JJ considerations_NN2 in_II practice_NN1 ._. 
All_RR too_RG often_RR ,_, as_CSA the_AT experience_NN1 of_IO many_DA2 mature_JJ applications_NN2 indicates_VVZ ,_, technical_JJ and_CC data-related_JJ matters_NN2 dominate_VV0 the_AT system_NN1 development_NN1 phase_NN1 and_CC technically_RR elegant_JJ solutions_NN2 are_VBR devised_VVN which_DDQ fail_VV0 to_TO take_VVI enough_DD account_NN1 of_IO user_NN1 needs_NN2 and_CC capabilities_NN2 ._. 
Campbell_NP1 discusses_VVZ the_AT role_NN1 of_IO geographic_JJ information_NN1 in_II organizations_NN2 and_CC the_AT process_NN1 of_IO GIS_NN2 diffusion_NN1 drawing_VVG upon_II research_NN1 from_II a_AT1 variety_NN1 of_IO social_JJ science_NN1 disciplines_NN2 ._. 
The_AT findings_NN2 of_IO the_AT analysis_NN1 are_VBR summarized_VVN in_II two_MC sets_NN2 of_IO research_NN1 questions_NN2 for_IF all_DB applications_NN2 fields_NN2 ._. 
In_II this_DD1 way_NN1 she_PPHS1 sets_VVZ out_RP a_AT1 research_NN1 agenda_NN1 for_IF applications-related_JJ research_NN1 which_DDQ serves_VVZ a_AT1 similar_JJ function_NN1 to_II the_AT spatial_JJ analysis_NN1 research_NN1 agenda_NN1 set_VVN out_RP by_II Openshaw_NP1 at_II the_AT beginning_NN1 of_IO Part_NN1 Two_MC ._. 
Environmental_JJ monitoring_NN1 and_CC prediction_NN1 David_NP1 Rhind_NP1 Introduction_NN1 Monitoring_NN1 of_IO the_AT state_NN1 of_IO and_CC the_AT changes_NN2 in_II the_AT environment_NN1 is_VBZ now_RT universally_RR recognized_VVN as_CSA being_VBG of_IO profound_JJ importance_NN1 to_II mankind_NN1 (_( Tickell_NP1 1986_MC ,_, 1989_MC )_) ._. 
This_DD1 awareness_NN1 is_VBZ most_RGT obviously_RR manifested_VVN by_II the_AT frequent_JJ appearance_NN1 of_IO press_NN1 reports_VVZ on_II ozone_NN1 depletion_NN1 ,_, on_II the_AT greenhouse_NN1 effect_NN1 and_CC possible_JJ global_JJ warming_NN1 ,_, on_II nitrate_NN1 pollution_NN1 ,_, on_II deforestation_NN1 in_II the_AT Tropics_NN ,_, on_II decreasing_VVG biological_JJ diversity_NN1 and_CC on_II many_DA2 other_JJ environmentally_RR related_JJ topics_NN2 ._. 
Usually_RR these_DD2 reports_NN2 concentrate_VV0 on_II prophecies_NN2 of_IO a_AT1 forthcoming_JJ Armageddon_NN1 but_CCB many_DA2 also_RR describe_VV0 a_AT1 dire_JJ contemporary_JJ situation_NN1 ._. 
Thus_RR the_AT algal_JJ bloom_NN1 off_II the_AT Italian_JJ Adriatic_NP1 shore_NN1 in_II 1989_MC is_VBZ reckoned_VVN to_TO have_VHI caused_VVN the_AT holiday_NN1 trade_NN1 and_CC local_JJ fisherman_NN1 losses_NN2 running_VVG into_II tens_MC2 if_CS not_XX hundreds_NNO2 of_IO millions_NNO2 of_IO pounds_NN2 ._. 
More_RGR globally_RR ,_, the_AT Brundtland_NP1 Report_NN1 (_( WCED_NP1 1987_MC )_) noted_VVD that_CST ,_, in_II the_AT 900_MC days_NNT2 during_II which_DDQ the_AT World_NN1 Commission_NN1 was_VBDZ at_II work_NN1 ,_, 1_MC1 ._. 
The_AT African_JJ drought_NN1 put_VVD the_AT lives_NN2 of_IO 35_MC million_NNO people_NN at_II risk_NN1 and_CC may_VM have_VHI killed_VVN 1_MC1 million_NNO of_IO them_PPHO2 ;_; 2_MC ._. 
The_AT Bhopal_NP1 accident_NN1 killed_VVD 2000_MC and_CC injured_VVD 200_MC 000_MC more_DAR ;_; 3_MC ._. 
The_AT Chernobyl_NP1 disaster_NN1 caused_VVD environmental_JJ damage_NN1 across_II Europe_NP1 ;_; 4_MC ._. 
A_AT1 chemical_JJ fire_NN1 in_II Switzerland_NP1 caused_VVD toxic_JJ materials_NN2 to_TO be_VBI transported_VVN by_II the_AT Rhine_NP1 at_RR21 least_RR22 as_CS31 far_CS32 as_CS33 the_AT Netherlands_NP1 ;_; 5_MC ._. 
At_RR21 least_RR22 60_MC million_NNO people_NN died_VVD of_IO diarrhoeal_JJ diseases_NN2 caused_VVN by_II malnutrition_NN1 and_CC dirty_JJ water_NN1 ._. 
As_II a_AT1 consequence_NN1 of_IO these_DD2 and_CC other_JJ disasters_NN2 ,_, of_IO rapid_JJ and_CC world-wide_JJ broadcasting_NN1 of_IO them_PPHO2 and_CC of_IO the_AT concerned_JJ action_NN1 of_IO many_DA2 individuals_NN2 ,_, the_AT environment_NN1 is_VBZ now_RT '_GE moving_VVG fast_RR up_RP the_AT political_JJ agenda_NN1 '_GE (_( Patten_NP1 1989_MC )_) ._. 
The_AT then_JJ Prime_JJ Minister_NN1 also_RR emphasized_VVD the_AT political_JJ awakening_NN1 to_II the_AT environment_NN1 in_II statements_NN2 such_II21 as_II22 :_: '_" it_PPH1 is_VBZ possible_JJ that_CST with_IW all_DB these_DD2 enormous_JJ changes_NN2 (_( population_NN1 ,_, agricultural_JJ ,_, use_NN1 of_IO fossil_NN1 fuels_NN2 )_) concentrated_VVD into_II such_DA a_AT1 short_JJ period_NN1 of_IO time_NNT1 ,_, we_PPIS2 have_VH0 unwittingly_RR begun_VVN a_AT1 massive_JJ experiment_NN1 with_IW the_AT system_NN1 of_IO this_DD1 planet_NN1 itself_PPX1 '_GE (_( Thatcher_NP1 1988_MC )_) ._. 
The_AT political_JJ concern_NN1 is_VBZ not_XX altruistic_JJ ._. 
Mrs_NNB Thatcher_NP1 ,_, a_AT1 recent_JJ convert_NN1 to_II environmental_JJ conservation_NN1 ,_, clearly_RR held_VVD the_AT opinion_NN1 that_CST environmental_JJ problems_NN2 are_VBR not_XX only_RR soluble_JJ but_CCB that_DD1 solving_VVG them_PPHO2 can_VM be_VBI shown_VVN to_TO be_VBI cost-effective_JJ :_: In_II the_AT past_NN1 when_CS we_PPIS2 identified_VVD forms_NN2 of_IO pollution_NN1 ,_, we_PPIS2 have_VH0 shown_VVN our_APPGE capacity_NN1 to_TO act_VVI effectively_RR ._. 
The_AT great_JJ London_NP1 smogs_NN2 are_VBR now_RT only_RR a_AT1 nightmare_NN1 of_IO the_AT past_NN1 ._. 
We_PPIS2 have_VH0 cut_VVN airborne_JJ lead_NN1 by_II 50_MC per_NNU21 cent_NNU22 ._. 
We_PPIS2 are_VBR spending_VVG 4_MC billion_NNO on_II cleansing_VVG the_AT Mersey_NP1 Basin_NN1 alone_RR ,_, and_CC the_AT Thames_NP1 now_RT has_VHZ the_AT cleanest_JJT metropolitan_JJ estuary_NN1 in_II the_AT world_NN1 ._. 
Even_CS21 though_CS22 this_DD1 kind_NN1 of_IO action_NN1 may_VM cost_VVI a_AT1 lot_NN1 ,_, I_PPIS1 believe_VV0 it_PPH1 to_TO be_VBI money_NN1 well_RR and_CC necessarily_RR spent_VVN because_CS the_AT health_NN1 of_IO our_APPGE economy_NN1 and_CC the_AT health_NN1 of_IO our_APPGE environment_NN1 are_VBR totally_RR dependent_JJ upon_II each_PPX221 other_PPX222 ...._... 
The_AT Government_NN1 espouses_VVZ the_AT concept_NN1 of_IO sustainable_JJ economic_JJ development_NN1 ._. 
Stable_JJ prosperity_NN1 can_VM be_VBI achieved_VVN throughout_II the_AT world_NN1 provided_CS the_AT environment_NN1 is_VBZ nurtured_VVN and_CC safeguarded_VVN (_( Thatcher_NP1 1988_MC )_) ._. 
The_AT primary_JJ official_JJ manifestation_NN1 of_IO this_DD1 UK_NP1 commitment_NN1 to_II the_AT environment_NN1 has_VHZ been_VBN the_AT government_NN1 's_GE White_JJ Paper_NN1 (_( HMSO_NP1 1990_MC )_) ._. 
While_CS criticized_VVN by_II many_DA2 on_II the_AT grounds_NN2 that_CST it_PPH1 had_VHD few_DA2 '_GE teeth_NN2 '_GE and_CC represented_VVD a_AT1 compromise_NN1 between_II the_AT DoE_NP1 and_CC all_DB other_JJ departments_NN2 (_( notably_RR those_DD2 of_IO Agriculture_NN1 and_CC of_IO Transport_NN1 )_) ,_, it_PPH1 was_VBDZ none_RR31 the_RR32 less_RR33 a_AT1 major_JJ achievement_NN1 ._. 
Its_APPGE main_JJ sections_NN2 dealt_VVN with_IW the_AT government_NN1 's_GE approach_NN1 to_II the_AT environment_NN1 (_( nationally_RR ,_, at_II the_AT European_JJ level_NN1 and_CC globally_RR )_) ,_, the_AT nature_NN1 and_CC effects_NN2 of_IO the_AT greenhouse_NN1 effect_NN1 ,_, environmental_JJ problems_NN2 in_II both_DB2 rural_JJ and_CC urban_JJ areas_NN2 (_( especially_RR in_II31 relation_II32 to_II33 land-use_NN1 matters_NN2 )_) ,_, pollution_NN1 control_NN1 and_CC enhancing_VVG awareness_NN1 of_IO environmental_JJ matters_NN2 ._. 
Central_JJ to_II this_DD1 approach_NN1 is_VBZ said_VVN to_TO be_VBI an_AT1 emphasis_NN1 upon_II guidance_NN1 from_II the_AT best_JJT available_JJ scientific_JJ evidence_NN1 ._. 
There_EX is_VBZ also_RR a_AT1 clear_JJ recognition_NN1 of_IO the_AT need_NN1 for_IF monitoring_VVG and_CC modelling_NN1 of_IO the_AT state_NN1 of_IO the_AT natural_JJ and_CC man-influenced_JJ environment_NN1 ._. 
Though_CS there_EX has_VHZ been_VBN a_AT1 change_NN1 of_IO Prime_JJ Minister_NN1 and_CC of_IO the_AT Secretary_NN1 of_IO State_NN1 for_IF the_AT Environment_NN1 since_CS the_AT publication_NN1 of_IO the_AT White_JJ Paper_NN1 ,_, the_AT environmental_JJ cause_NN1 seems_VVZ as_RG strong_JJ as_CSA ever_RR ._. 
Thus_RR ,_, The_AT Times_NNT2 of_IO 17_MC December_NPM1 1990_MC reported_VVD that_CST a_AT1 group_NN1 of_IO 40_MC of_IO the_AT most_RGT senior_JJ Civil_JJ Servants_NN2 had_VHD met_VVN the_AT previous_JJ weekend_NNT1 to_II discuss_58 '_GE ways_NN2 in_II which_DDQ the_AT government_NN1 machine_NN1 could_VM respond_VVI better_RRR to_II national_JJ and_CC international_JJ problems_NN2 such_II21 as_II22 the_AT threat_NN1 of_IO global_JJ climate_NN1 change_NN1 '_GE ._. 
Pearce_NP1 et_RA21 al_RA22 ._. 
(_( 1989_MC )_) have_VH0 argued_VVN that_CST the_AT effects_NN2 of_IO government_NN1 ,_, commercial_JJ and_CC other_JJ actions_NN2 on_II the_AT environment_NN1 can_VM and_CC should_VM be_VBI assessed_VVN on_II a_AT1 cost/benefit_NN1 basis_NN1 and_CC this_DD1 seems_VVZ likely_JJ to_TO be_VBI the_AT basis_NN1 on_II which_DDQ at_RR21 least_RR22 the_AT UK_NP1 government_NN1 will_VM proceed_VVI ._. 
Given_VVN such_DA a_AT1 lead_NN1 ,_, it_PPH1 is_VBZ scarcely_RR surprising_JJ that_CST British_JJ scientists_NN2 have_VH0 sought_VVN to_TO demonstrate_VVI the_AT practical_JJ and_CC financial_JJ consequences_NN2 of_IO environmental_JJ events_NN2 and_CC processes_NN2 ;_; Table_NN1 9.1_MC ,_, for_REX21 example_REX22 illustrates_VVZ their_APPGE perception_NN1 of_IO the_AT importance_NN1 of_IO global_JJ environmental_JJ research_NN1 to_II the_AT UK_NP1 domestic_JJ economy_NN1 ._. 
If_CS it_PPH1 now_RT seems_VVZ self-evident_JJ that_CST monitoring_NN1 of_IO the_AT global_JJ environment_NN1 is_VBZ necessary_JJ ,_, indeed_RR is_VBZ even_RR vital_JJ ,_, the_AT prediction_NN1 of_IO what_DDQ is_VBZ likely_JJ to_TO happen_VVI is_VBZ almost_RR as_CSA important_JJ :_: the_AT 170_MC 000_MC people_NN in_II the_AT Maldives_NP1 are_VBR understandably_RR worried_VVN about_II the_AT prospect_NN1 of_IO global_JJ sea-level_JJ change_NN1 since_CS no_AT part_NN1 of_IO the_AT islands_NN2 is_VBZ more_DAR than_CSN 2_MC m_NNO above_II present_JJ sea-level_NN1 !_! 
Yet_RR ,_, as_CSA we_PPIS2 shall_VM see_VVI below_RL ,_, the_AT prediction_NN1 of_IO the_AT activities_NN2 of_IO natural_JJ and_CC man-influenced_JJ environmental_JJ systems_NN2 requires_VVZ an_AT1 understanding_NN1 of_IO the_AT interaction_NN1 of_IO processes_NN2 that_CST ,_, in_II many_DA2 cases_NN2 ,_, we_PPIS2 simply_RR do_VD0 not_XX yet_RR have_VHI ._. 
Even_RR monitoring_VVG of_IO the_AT present_NN1 is_VBZ very_RG imperfect_JJ ,_, as_CSA witness_NN1 the_AT order_NN1 of_IO magnitude_NN1 variations_NN2 in_II estimates_NN2 now_RT current_JJ for_IF deforestation_NN1 rates_NN2 ._. 
This_DD1 chapter_NN1 ,_, then_RT ,_, attempts_NN2 to_TO review_VVI some_DD of_IO the_AT problems_NN2 in_II environmental_JJ monitoring_NN1 and_CC prediction_NN1 ,_, describes_VVZ some_DD of_IO the_AT relevant_JJ scientific_JJ and_CC organizational_JJ activities_NN2 and_CC looks_VVZ ahead_RL to_II what_DDQ is_VBZ likely_JJ to_TO happen_VVI next_MD ._. 
The_AT nature_NN1 of_IO environmental_JJ processes_NN2 Notwithstanding_II Mrs_NNB Thatcher_NP1 's_GE confidence_NN1 that_CST we_PPIS2 can_VM '_GE fix_NN1 '_GE environmental_JJ problems_NN2 ,_, the_AT difficulties_NN2 associated_VVN with_IW a_AT1 task_NN1 even_RR as_CSA apparently_RR simple_JJ as_CSA monitoring_VVG a_AT1 state_NN1 variable_NN1 are_VBR considerable_JJ ._. 
Some_DD of_IO the_AT reasons_NN2 for_IF this_DD1 the_AT difficulty_NN1 of_IO deciding_VVG on_II appropriate_JJ proxy_NN1 variables_NN2 and_CC on_II economical_JJ yet_RR non-biasing_JJ temporal_JJ and_CC spatial_JJ sampling_NN1 frameworks_NN2 ,_, taking_VVG account_NN1 of_IO the_AT relative_JJ importance_NN1 of_IO aperiodic_JJ and_CC rare_JJ events_NN2 as_CSA compared_VVN to_II near-continuous_JJ processes_NN2 ,_, processing_VVG the_AT vast_JJ volumes_NN2 of_IO data_NN usually_RR involved_VVN and_CC organizing_VVG the_AT multidisciplinary_NN1 and_CC (_( often_RR )_) multinational_JJ researchers_NN2 involved_VVN are_VBR discussed_VVN briefly_RR later_RRR ._. 
Even_RR at_II this_DD1 stage_NN1 ,_, however_RR ,_, it_PPH1 is_VBZ important_JJ to_TO stress_VVI that_CST we_PPIS2 have_VH0 a_AT1 very_RG incomplete_JJ understanding_NN1 of_IO many_DA2 processes_NN2 in_II the_AT natural_JJ environment_NN1 ._. 
The_AT next_MD section_NN1 therefore_RR attempts_VVZ to_TO summarize_VVI what_DDQ we_PPIS2 do_VD0 know_VVI ;_; it_PPH1 is_VBZ derived_VVN chiefly_RR from_II the_AT Earth_NN1 System_NN1 Sciences_NN2 Committee_NN1 (_( ESSC_NP1 )_) (_( 1988_MC )_) ._. 
Temporal_JJ and_CC spatial_JJ scales_NN2 of_IO operation_NN1 Even_RR ignoring_VVG the_AT effects_NN2 of_IO man_NN1 ,_, earth_NN1 system_NN1 processes_NN2 operate_VV0 at_II a_AT1 variety_NN1 of_IO spatial_JJ and_CC temporal_JJ scales_NN2 (_( see_VV0 Fig._NN1 9.1_MC ;_; see_VV0 also_RR Rosswall_NP1 et_RA21 al_RA22 ._. 
1988_MC )_) ._. 
Thus_RR microclimate_JJ effects_NN2 can_VM have_VHI major_JJ short-term_JJ implications_NN2 for_IF local_JJ populations_NN2 (_( such_II21 as_II22 those_DD2 brought_VVN about_RP by_II the_AT photochemical_JJ smogs_NN2 of_IO Los_NP1 Angeles_NP1 )_) ,_, are_VBR often_RR applicable_JJ only_RR over_II areas_NN2 of_IO a_AT1 few_DA2 hectares_NNU2 or_CC less_RRR and_CC may_VM operate_VVI on_II a_AT1 diurnal_JJ cycle_NN1 ._. 
In_II contrast_NN1 ,_, soil_NN1 creation_NN1 (_( cf._VV0 soil_NN1 degradation_NN1 )_) may_VM require_VVI hundreds_NNO2 or_CC thousands_NNO2 of_IO years_NNT2 ._. 
In_RL21 between_RL22 are_VBR variations_NN2 in_II phenomena_NN2 such_II21 as_II22 El_NP1 Nino_NP1 ._. 
The_AT sampling_NN1 theorem_NN1 indicates_VVZ that_CST we_PPIS2 must_VM sample_VVI these_DD2 phenomena_NN2 at_II a_AT1 frequency_NN1 of_IO no_AT less_DAR than_CSN half_DB their_APPGE wavelength_NN1 if_CS we_PPIS2 are_VBR to_TO avoid_VVI bias_NN1 in_II measurement_NN1 ._. 
In_II this_DD1 chapter_NN1 ,_, the_AT primary_JJ concern_NN1 is_VBZ with_IW changes_NN2 occurring_VVG over_II time_NNT1 spans_VVZ from_II minutes_NNT2 to_II centuries_NNT2 and_CC from_II spatial_JJ wavelengths_NN2 from_II about_II 1_MC1 m_NNO to_II about_RG 100_MC km_NNU ._. 
It_PPH1 is_VBZ ,_, however_RR ,_, impossible_JJ to_TO separate_VVI out_RP processes_NN2 acting_VVG at_II different_JJ scales_NN2 and_CC over_II different_JJ periods_NN2 :_: a_AT1 single_JJ volcanic_JJ eruption_NN1 lasting_JJ hours_NNT2 or_CC days_NNT2 sometimes_RT leads_VVZ to_TO dust_VVI and_CC gas_NN1 being_VBG ejected_VVN into_II the_AT atmosphere_NN1 over_II a_AT1 period_NN1 of_IO months_NNT2 and_CC this_DD1 in_II turn_NN1 may_VM lead_VVI to_II global_JJ climatic_JJ fluctuations_NN2 over_II years_NNT2 or_CC decades_NNT2 ._. 
Conversely_RR ,_, a_AT1 large-scale_JJ and_CC long-period_JJ phenomenon_NN1 such_II21 as_II22 the_AT movement_NN1 of_IO tectonic_JJ plates_NN2 can_VM lead_VVI to_II local_JJ stress_NN1 accumulation_NN1 and_CC to_II its_APPGE release_NN1 by_II an_AT1 earthquake_NN1 in_II a_AT1 few_DA2 seconds_NNT2 and_CC with_IW highly_RR local_JJ effects_NN2 ._. 
This_DD1 therefore_RR has_VHZ the_AT following_JJ consequences_NN2 :_: 1_MC1 ._. 
We_PPIS2 should_VM know_VVI something_PN1 about_II the_AT periodicity_NN1 of_IO the_AT variables_NN2 we_PPIS2 are_VBR monitoring_VVG if_CS we_PPIS2 are_VBR not_XX to_TO risk_VVI either_RR unnecessarily_RR detailed_JJ data_NN collection_NN1 or_CC aliasing_NN1 of_IO trends_NN2 ;_; 2_MC ._. 
Since_CS we_PPIS2 know_VV0 that_CST the_AT periodicity_NN1 of_IO some_DD variables_NN2 is_VBZ longer_JJR than_CSN the_AT monitoring_NN1 already_RR carried_VVN out_RP by_II human_JJ beings_NN2 ,_, we_PPIS2 must_VM make_VVI use_NN1 of_IO surrogate_NN1 or_CC proxy_NN1 variables_NN2 to_TO extend_VVI our_APPGE range_NN1 of_IO knowledge_NN1 (_( such_II21 as_II22 via_II tree_NN1 rings_NN2 ,_, palynology_NN1 ,_, oxygen_NN1 isotope_NN1 ratios_NN2 from_II ice_NN1 cores_NN2 and_CC geomorphological_JJ evidence_NN1 of_IO environmental_JJ change_NN1 )_) ;_; 3_MC ._. 
Since_CS we_PPIS2 are_VBR not_XX yet_RR able_JK to_TO predict_VVI all_DB the_AT interactions_NN2 which_DDQ may_VM be_VBI important_JJ in_II a_AT1 study_NN1 of_IO environmental_JJ problems_NN2 ,_, collecting_VVG potentially_RR useful_JJ data_NN as_II31 well_II32 as_II33 those_DD2 believed_VVN central_JJ to_II the_AT problem_NN1 under_II study_NN1 is_VBZ often_RR wise_JJ ;_; 4_MC ._. 
We_PPIS2 must_VM accept_VVI that_CST most_DAT data_NN collection_NN1 exercises_NN2 will_VM inevitably_RR be_VBI a_AT1 compromise_NN1 on_II the_AT grounds_NN2 of_IO cost_NN1 and_CC hence_RR less_DAR than_CSN perfect_JJ data_NN sets_NN2 will_VM be_VBI produced_VVN ;_; 5_MC ._. 
Irrespective_II21 of_II22 the_AT use_NN1 of_IO surrogates_NN2 ,_, monitoring_VVG environmental_JJ processes_NN2 only_RR makes_VVZ sense_NN1 over_II time-scales_NN2 which_DDQ are_VBR extended_VVN by_II human_JJ standards_NN2 (_( hence_RR the_AT IGBP_NP1 (_( see_VV0 below_RL )_) is_VBZ designed_VVN to_TO run_VVI for_IF two_MC or_CC three_MC decades_NNT2 )_) ;_; 6_MC ._. 
Changes_NN2 in_II political_JJ priorities_NN2 may_VM simply_RR result_VVI in_II the_AT need_NN1 for_IF informed_JJ guesswork_NN1 if_CS suitable_JJ data_NN are_VBR not_XX available_JJ ._. 
Much_DA1 attention_NN1 will_VM be_VBI given_VVN throughout_II this_DD1 chapter_NN1 to_II monitoring_VVG the_AT environment_NN1 on_II a_AT1 global_JJ basis_NN1 ._. 
This_DD1 is_VBZ not_XX merely_RR because_II21 of_II22 its_APPGE topical_JJ nature_NN1 ._. 
The_AT main_JJ reason_NN1 is_VBZ that_CST though_CS some_DD problems_NN2 are_VBR local_JJ and_CC are_VBR best_RRT studied_VVN that_DD1 way_NN1 many_DA2 others_NN2 (_( even_RR involving_VVG a_AT1 limited_JJ geographical_JJ extent_NN1 )_) are_VBR best_RRT understood_VVN within_II a_AT1 global_JJ framework_NN1 ._. 
Thus_RR no_AT meteorologist_NN1 would_VM attempt_VVI to_TO forecast_VVI the_AT weather_NN1 2_MC or_CC 3_MC days_NNT2 ahead_RL without_IW using_VVG a_AT1 mathematical_JJ model_NN1 based_VVN on_II observations_NN2 from_II all_DB around_II the_AT world_NN1 ._. 
Moreover_RR ,_, recent_JJ work_NN1 has_VHZ shown_VVN that_RG much_DA1 of_IO the_AT heat_NN1 energy_NN1 in_II the_AT seas_NN2 of_IO the_AT North_ND1 Atlantic_NP1 (_( every_AT1 square_JJ kilometre_NNU1 of_IO which_DDQ gives_VVZ off_RP as_RG much_DA1 energy_NN1 as_CSA a_AT1 nuclear_JJ power_NN1 station_NN1 and_CC hence_RR influences_VVZ our_APPGE weather_NN1 )_) is_VBZ absorbed_VVN from_II sunlight_NN1 in_II the_AT tropical_JJ Pacific_NP1 and_CC is_VBZ carried_VVN by_II ocean_NN1 currents_NN2 through_II the_AT Drake_NN1 Passage_NN1 and_CC up_RP into_II the_AT Atlantic_NP1 ._. 
Thus_RR ,_, to_TO understand_VVI some_DD local_JJ and_CC many_DA2 regional_JJ ,_, national_JJ and_CC continental_JJ and_CC all_DB global_JJ environmental_JJ events_NN2 ,_, we_PPIS2 need_VV0 to_TO have_VHI a_AT1 global_JJ perspective_NN1 and_CC hold_VV0 global_JJ databases_NN2 (_( Rhind_NP1 and_CC Mounsey_NP1 1989_MC ;_; Mounsey_NP1 and_CC Tomlinson_NP1 1988_MC )_) ._. 
In_II no_AT way_NN1 ,_, however_RR ,_, does_VDZ this_DD1 invalidate_VVI the_AT need_NN1 for_IF microscale_NN1 studies_NN2 for_REX21 instance_REX22 ,_, while_CS global_JJ data_NN sets_NN2 may_VM indicate_VVI the_AT extent_NN1 of_IO algal_JJ blooms_NN2 and_CC may_VM suggest_VVI causes_NN2 of_IO them_PPHO2 ,_, only_RR detailed_JJ study_NN1 of_IO the_AT algae_NN2 themselves_PPX2 can_VM provide_VVI understanding_NN1 of_IO why_RRQ the_AT bloom_NN1 is_VBZ occurring_VVG ._. 
The_AT processes_NN2 themselves_PPX2 A_AT1 major_JJ complication_NN1 ,_, however_RR ,_, is_VBZ that_CST the_AT environment_NN1 can_VM rarely_RR be_VBI treated_VVN as_CSA in_II a_AT1 laboratory_NN1 experiment_NN1 ._. 
Given_VVN this_DD1 ,_, determining_VVG the_AT nature_NN1 of_IO the_AT interactions_NN2 between_II the_AT variables_NN2 becomes_VVZ a_AT1 matter_NN1 of_IO major_JJ difficulty_NN1 ._. 
Figure_NN1 9.2_MC is_VBZ taken_VVN from_II ESSC_NP1 (_( 1988_MC )_) and_CC is_VBZ an_AT1 attempt_NN1 to_TO show_VVI the_AT interactions_NN2 between_II environmental_JJ processes_NN2 ,_, as_II31 well_II32 as_II33 the_AT indicators_NN2 and_CC implications_NN2 of_IO change_NN1 in_II the_AT state_NN1 of_IO the_AT environment_NN1 ._. 
In_II essence_NN1 ,_, the_AT figure_NN1 can_VM be_VBI summarized_VVN as_CSA showing_VVG processes_NN2 within_II the_AT physical_JJ climate_NN1 system_NN1 and_CC within_II the_AT biogeochemical_JJ one_PN1 ,_, these_DD2 systems_NN2 being_VBG interwoven_VVN by_II the_AT global_JJ role_NN1 of_IO water_NN1 and_CC increasingly_RR affected_VVN by_II human_JJ activities_NN2 ._. 
The_AT main_JJ components_NN2 of_IO the_AT first_MD system_NN1 are_VBR well_RR known_VVN :_: the_AT atmosphere_NN1 is_VBZ its_APPGE engine_NN1 and_CC atmospheric_JJ circulation_NN1 is_VBZ driven_VVN by_II differences_NN2 in_II solar_JJ heating_NN1 between_II the_AT equatorial_JJ and_CC polar_JJ regions_NN2 ._. 
We_PPIS2 are_VBR ,_, however_RR ,_, becoming_VVG ever_RR more_RGR aware_JJ of_IO the_AT massive_JJ influence_NN1 of_IO oceans_NN2 on_II our_APPGE climate_NN1 ;_; research_NN1 into_II ocean/atmosphere_NN1 interactions_NN2 is_VBZ now_RT a_AT1 priority_NN1 ._. 
In_II ocean_NN1 dynamics_NN studies_NN2 ,_, important_JJ topics_NN2 are_VBR heat_NN1 storage_NN1 ,_, circulation_NN1 and_CC the_AT role_NN1 of_IO sea_NN1 ice_NN1 ._. 
Atmospheric_JJ circulation_NN1 (_( and_CC ,_, in_II turn_NN1 ,_, precipitation_NN1 and_CC temperature_NN1 )_) is_VBZ also_RR greatly_RR affected_VVN by_II the_AT balance_NN1 of_IO thermal_JJ energy_NN1 at_II the_AT land_NN1 surface_NN1 ,_, itself_PPX1 affected_VVN by_II the_AT vegetation_NN1 and_CC terrain_NN1 characteristics_NN2 ._. 
The_AT biogeochemical_JJ system_NN1 embraces_VVZ the_AT totality_NN1 of_IO the_AT movements_NN2 of_IO the_AT key_JJ elements_NN2 essential_JJ to_II life_NN1 carbon_NN1 ,_, nitrogen_NN1 ,_, oxygen_NN1 ,_, phosphorus_NN1 and_CC others_NN2 through_II what_DDQ has_VHZ been_VBN described_VVN as_II the_AT total_JJ earth_NN1 system_NN1 ._. 
The_AT oceans_NN2 and_CC their_APPGE biota_NN2 play_VV0 a_AT1 central_JJ role_NN1 in_II the_AT carbon_NN1 and_CC nutrient_NN1 cycles_NN2 ;_; global_JJ ocean_NN1 circulation_NN1 is_VBZ also_RR critical_JJ to_II the_AT global_JJ carbon_NN1 budget_NN1 ._. 
Trace_VV0 gases_NN2 in_II the_AT atmosphere_NN1 are_VBR key_JJ components_NN2 of_IO the_AT cycles_NN2 of_IO such_DA elements_NN2 as_CSA carbon_NN1 ,_, nitrogen_NN1 ,_, oxygen_NN1 ,_, sulphur_NN1 and_CC the_AT halogens_NN2 ;_; as_CSA is_VBZ now_RT well_RR known_VVN ,_, their_APPGE concentrations_NN2 are_VBR much_RR influenced_VVN by_II biogenic_JJ and_CC anthropogenic_JJ activities_NN2 and_CC the_AT concentrations_NN2 of_IO such_DA gases_NN2 as_CSA carbon_NN1 dioxide_NN1 ,_, ozone_NN1 ,_, methane_NN1 ,_, nitrous_JJ oxide_NN1 and_CC the_AT chlorofluoromethanes_NN2 (_( CFMs_NP1 )_) strongly_RR affect_VV0 radiative_JJ transfer_NN1 and_CC provide_VV0 a_AT1 link_NN1 with_IW the_AT physical_JJ climate_NN1 system_NN1 ._. 
A_AT1 particular_JJ cause_NN1 for_IF concern_NN1 is_VBZ that_CST the_AT concentrations_NN2 of_IO the_AT two_MC main_JJ CFMs_NP1 is_VBZ increasing_VVG by_RP about_RG 5_MC per_NNU21 cent_NNU22 per_RA21 annum_RA22 while_CS the_AT equivalent_JJ figures_NN2 for_IF methane_NN1 ,_, nitrous_JJ oxide_NN1 and_CC carbon_NN1 dioxide_NN1 are_VBR 1_MC1 ,_, 0.4_MC and_CC 0.3_MC per_NNU21 cent_NNU22 ._. 
Finally_RR ,_, the_AT chemistry_NN1 of_IO the_AT stratosphere_NN1 is_VBZ dominated_VVN by_II the_AT photochemical_JJ production_NN1 and_CC catalytic_JJ destruction_NN1 of_IO ozone_NN1 ._. 
There_EX are_VBR several_DA2 important_JJ ways_NN2 in_II which_DDQ biogeochemical_JJ processes_NN2 can_VM influence_VVI climate_NN1 and_CC hence_RR habitability_NN1 (_( except_CS where_CS the_AT latter_DA is_VBZ maintained_VVN at_II enormous_JJ cost_NN1 )_) ._. 
If_CS ,_, for_REX21 instance_REX22 ,_, the_AT concentrations_NN2 of_IO certain_JJ trace_NN1 gases_NN2 (_( see_VV0 above_RL )_) continue_VV0 to_TO increase_VVI ,_, it_PPH1 has_VHZ been_VBN predicted_VVN that_CST the_AT earth_NN1 's_GE surface_NN1 temperature_NN1 will_VM increase_VVI by_II an_AT1 amount_NN1 comparable_JJ to_II that_DD1 since_CS the_AT last_MD major_JJ phase_NN1 of_IO the_AT most_RGT recent_JJ glaciation_NN1 (_( about_RG 18_MC 000_MC years_NNT2 BP_NP1 )_) though_CS there_EX is_VBZ some_DD dispute_NN1 about_II the_AT magnitude_NN1 of_IO the_AT likely_JJ change_NN1 ._. 
The_AT converse_JJ case_NN1 the_AT effect_NN1 of_IO physical_JJ climate_NN1 changes_NN2 on_II biogeochemical_JJ processes_NN2 is_VBZ also_RR of_IO obvious_JJ importance_NN1 :_: changes_NN2 in_II annual_JJ averages_NN2 and_CC cycles_NN2 of_IO temperature_NN1 and_CC precipitation_NN1 ,_, together_RL with_IW the_AT probabilities_NN2 of_IO extreme_JJ events_NN2 such_II21 as_II22 prolonged_JJ droughts_NN2 or_CC frosts_NN2 ,_, are_VBR major_JJ determinants_NN2 of_IO terrestrial_JJ ecosystem_NN1 type_NN1 ,_, at_RR21 least_RR22 in_II areas_NN2 largely_RR unaffected_JJ by_II human_JJ activity_NN1 ._. 
Human-induced_JJ changes_NN2 to_II the_AT environment_NN1 have_VH0 increased_VVN rapidly_RR as_II a_AT1 consequence_NN1 of_IO growing_JJ population_NN1 numbers_NN2 and_CC of_IO numerous_JJ developments_NN2 in_II technology._NNU fossil_NN1 fuel_NN1 combustion_NN1 ,_, for_REX21 instance_REX22 ,_, has_VHZ led_VVN to_II increased_JJ concentrations_NN2 of_IO sulphate_NN1 in_II precipitation_NN1 (_( '_GE acid_NN1 rain'_VVG )_) and_CC many_DA2 rivers_NN2 ,_, lakes_NN2 and_CC estuaries_NN2 have_VH0 been_VBN greatly_RR affected_VVN by_II phosphates_NN2 from_II agricultural_JJ ,_, urban_JJ or_CC industrial_JJ sources_NN2 ._. 
It_PPH1 should_VM be_VBI stressed_VVN ,_, however_RR ,_, that_CST while_CS the_AT outlines_NN2 of_IO human_JJ activities_NN2 are_VBR clear_JJ there_EX are_VBR still_RR many_DA2 uncertainties_NN2 :_: for_REX21 instance_REX22 ,_, we_PPIS2 know_VV0 that_CST the_AT atmospheric_JJ concentration_NN1 of_IO carbon_NN1 dioxide_NN1 has_VHZ risen_VVN about_RG 9_MC per_NNU21 cent_NNU22 since_II 1959_MC ._. 
But_CCB calculations_NN2 of_IO the_AT total_JJ carbon_NN1 released_VVN by_II fossil_NN1 fuel_NN1 combustion_NN1 imply_VV0 that_CST ,_, if_CS all_DB such_DA carbon_NN1 remained_VVD in_II the_AT atmosphere_NN1 ,_, the_AT concentration_NN1 would_VM have_VHI increased_VVN by_II twice_RR as_RG much_DA1 ._. 
The_AT absolute_JJ difference_NN1 between_II the_AT two_MC figures_NN2 seems_VVZ to_TO be_VBI due_II21 to_II22 two_MC different_JJ factors_NN2 :_: absorption_NN1 by_II the_AT oceans_NN2 and_CC other_JJ reductions_NN2 in_II carbon_NN1 stocks_NN2 in_II living_JJ biota_NN2 ,_, the_AT latter_DA being_NN1 strongly_RR affected_VVN by_II human_JJ activities_NN2 since_CS the_AT Industrial_JJ Revolution_NN1 ._. 
Does_VDZ God_NP1 play_VVI dice_NN ?_? 
Much_RR environmental_JJ prediction_NN1 is_VBZ predicated_VVN upon_II a_AT1 logical_JJ positivist_NN1 or_CC Newtonian_JJ deterministic_JJ basis_NN1 ._. 
The_AT environmental_JJ systems_NN2 are_VBR effectively_RR regarded_VVN as_II machines_NN2 whose_DDQGE workings_NN2 can_VM all_DB be_VBI discovered_VVN ,_, described_VVN in_II differential_JJ equation_NN1 form_NN1 and_CC whose_DDQGE future_JJ activities_NN2 can_VM thus_RR be_VBI predicted_VVN in_II detail_NN1 if_CS their_APPGE initial_JJ states_NN2 are_VBR known_VVN ._. 
An_AT1 argument_NN1 by_II many_DA2 atmospheric_JJ physicists_NN2 ,_, for_REX21 example_REX22 ,_, is_VBZ that_DD1 shortcomings_NN2 in_II the_AT accuracy_NN1 of_IO weather_NN1 prediction_NN1 over_II periods_NN2 of_IO more_DAR than_CSN a_AT1 few_DA2 days_NNT2 largely_RR results_VVZ from_II the_AT unsophisticated_JJ nature_NN1 of_IO existing_JJ models_NN2 ,_, the_AT lack_NN1 of_IO suitable_JJ data_NN and_CC inadequate_JJ computer_NN1 power_NN1 (_( see_VV0 Fig._NN1 9.3_MC ,_, derived_VVN from_II Tyler_NP1 1989_MC )_) ._. 
In_II contrast_NN1 ,_, the_AT advocates_NN2 of_IO chaos_NN1 theory_NN1 (_( e.g._REX Stewart_NP1 1989_MC )_) argue_VV0 that_CST some_DD natural_JJ systems_NN2 are_VBR so_RG sensitive_JJ to_II tiny_JJ stimuli_NN2 that_DD1 even_NNT1 '_GE the_AT faint_JJ beat_NN1 of_IO a_AT1 butterfly_NN1 's_GE wing_NN1 could_VM be_VBI the_AT ultimate_JJ trigger_NN1 for_IF a_AT1 hurricane_NN1 '_GE ._. 
In_II essence_NN1 ,_, they_PPHS2 argue_VV0 that_CST tiny_JJ differences_NN2 in_II the_AT initial_JJ conditions_NN2 of_IO many_DA2 systems_NN2 can_VM lead_VVI to_II widely_RR different_JJ outcomes_NN2 since_CS the_AT systems_NN2 exhibit_VV0 stochastic_JJ behaviour_NN1 within_II a_AT1 deterministic_JJ framework_NN1 ._. 
If_CS so_RR ,_, there_EX is_VBZ little_DA1 chance_NN1 of_IO making_VVG accurate_JJ long-term_JJ forecasts_NN2 of_IO these_DD2 systems_NN2 ._. 
This_DD1 has_VHZ been_VBN recognized_VVN by_II ,_, among_II others_NN2 ,_, ESSC_NP1 (_( 1988_MC )_) ._. 
The_AT authors_NN2 of_IO that_DD1 report_NN1 accepted_VVD that_CST there_EX is_VBZ a_AT1 theoretical_JJ limit_NN1 of_IO 23_MC weeks_NNT2 in_II explicit_JJ prediction_NN1 of_IO day-to-day_JJ fluctuations_NN2 in_II weather_NN1 because_II21 of_II22 the_AT influence_NN1 of_IO small-scale_JJ turbulence_NN1 within_II the_AT large-scale_JJ dynamics_NN ;_; such_DA limits_NN2 imply_VV0 that_DD1 medium_NN1 and_CC long-term_JJ prediction_NN1 must_VM inevitably_RR be_VBI stochastic_JJ ._. 
To_TO complicate_VVI matters_NN2 still_RR further_RRR ,_, however_RR ,_, it_PPH1 seems_VVZ at_RR21 least_RR22 possible_JJ that_CST the_AT components_NN2 of_IO the_AT earth_NN1 system_NN1 may_VM best_RRT be_VBI predicted_VVN deterministically_RR at_II some_DD spatial_JJ and_CC temporal_JJ scales_NN2 but_CCB stochastically_RR at_II others_NN2 and_CC that_CST the_AT basis_NN1 of_IO prediction_NN1 will_VM differ_VVI for_IF different_JJ components_NN2 ._. 
As_II an_AT1 example_NN1 ,_, atmospheric_JJ temperature_NN1 shows_VVZ marked_JJ day-to-day_JJ variability_NN1 ,_, a_AT1 strong_JJ seasonal_JJ cycle_NN1 with_IW some_DD year-to-year_JJ variations_NN2 ,_, a_AT1 weak_JJ minimum_JJ variability_NN1 for_IF 1020_MC year_NNT1 averages_NN2 and_CC then_RT increased_JJ variation_NN1 again_RT at_II longer_JJR time-scales_NN2 (_( ESSC_NP1 1988_MC )_) ._. 
It_PPH1 need_VM hardly_RR be_VBI pointed_VVN out_RP that_CST coping_VVG with_IW such_DA multiple_JJ databases_NN2 on_II which_DDQ modelling_NN1 is_VBZ to_TO be_VBI carried_VVN out_RP is_VBZ not_XX a_AT1 trivial_JJ exercise_NN1 ._. 
Finally_RR ,_, it_PPH1 is_VBZ perhaps_RR worth_II concluding_VVG this_DD1 section_NN1 by_II emphasizing_VVG that_CST there_EX is_VBZ still_RR much_DA1 controversy_NN1 among_II the_AT scientific_JJ community_NN1 about_II the_AT magnitude_NN1 of_IO many_DA2 of_IO the_AT effects_NN2 being_VBG discussed_VVN ._. 
In_RR21 general_RR22 ,_, the_AT research_NN1 work_NN1 of_IO the_AT international_JJ groups_NN2 of_IO scientists_NN2 ,_, including_II those_DD2 involved_JJ in_II the_AT Inter-Governmental_JJ Committee_NN1 and_CC in_II the_AT UK_NP1 's_GE Inter-Agency_NP1 Committee_NN1 on_II Global_JJ Environmental_JJ Change_NN1 ,_, has_VHZ progressively_RR reduced_VVN the_AT magnitude_NN1 of_IO the_AT predicted_JJ effects_NN2 of_IO ,_, for_REX21 example_REX22 ,_, global_JJ warming_NN1 ._. 
In_II giving_VVG the_AT 1990_MC Robens_NP2 Coal_NN1 Science_NN1 Lecture_NN1 at_II the_AT Royal_JJ Institution_NN1 ,_, Sir_NNB John_NP1 Mason_NP1 (_( a_AT1 former_DA Director_NN1 of_IO the_AT Meteorological_JJ Office_NN1 )_) urged_VVD caution_NN1 in_II interpreting_VVG the_AT results_NN2 of_IO the_AT massive_JJ atmospheric_JJ and_CC oceanic_JJ modelling_NN1 exercises_NN2 ;_; as_II the_AT equilibrium_NN1 models_NN2 are_VBR refined_VVN to_TO take_VVI account_NN1 of_IO the_AT thermal_JJ lag_NN1 in_II oceans_NN2 ,_, etc._RA ,_, lower_JJR values_NN2 of_IO predicted_JJ changes_NN2 are_VBR likely_JJ ._. 
Thus_RR ,_, while_CS almost_RR all_DB scientists_NN2 are_VBR at_II the_AT time_NNT1 of_IO writing_VVG convinced_JJ from_II the_AT evidence_NN1 that_CST change_VV0 in_II climate_NN1 is_VBZ under_RR21 way_RR22 and_CC that_DD1 regional_JJ effects_NN2 may_VM be_VBI much_RR greater_JJR than_CSN the_AT global_JJ average_NN1 ,_, the_AT magnitude_NN1 of_IO the_AT latter_DA is_VBZ now_RT thought_VVN likely_JJ to_TO be_VBI about_RG half_DB of_IO what_DDQ was_VBDZ predicted_VVN only_RR 5_MC years_NNT2 ago_RA ._. 
Who_PNQS is_VBZ involved_VVN and_CC what_DDQ is_VBZ going_VVG on_RP ?_? 
If_CS we_PPIS2 consider_VV0 the_AT environment_NN1 in_II the_AT broadest_JJT sense_NN1 ,_, much_DA1 monitoring_NN1 of_IO it_PPH1 has_VHZ been_VBN in_II progress_NN1 for_IF many_DA2 years_NNT2 and_CC by_II an_AT1 enormous_JJ diversity_NN1 of_IO organizations_NN2 in_II the_AT UK_NP1 ;_; these_DD2 include_VV0 the_AT DoE_NP1 ,_, the_AT Forestry_NN1 Commission_NN1 ,_, the_AT Nature_NN1 Conservancy_NN1 ,_, the_AT Health_NN1 and_CC Safety_NN1 Executive_NN1 ,_, local_JJ authorities_NN2 ,_, the_AT component_NN1 members_NN2 of_IO the_AT Natural_JJ Environment_NN1 Research_NN1 Council_NN1 (_( such_II21 as_II22 the_AT Institute_NN1 of_IO Terrestrial_JJ Ecology_NN1 and_CC the_AT British_JJ Antarctic_JJ Survey_NN1 )_) ,_, individuals_NN2 or_CC research_NN1 groups_NN2 in_II academia_NN1 ,_, the_AT CEGB_NP1 ,_, the_AT Friends_NN2 of_IO the_AT Earth_NN1 ,_, and_CC UK_NP1 Centre_NN1 for_IF Economic_JJ and_CC Environmental_JJ Development_NN1 (_( CEED_NP1 )_) (_( 1989_MC )_) ._. 
The_AT diversity_NN1 of_IO monitoring_VVG organizations_NN2 seems_VVZ just_RR as_RG great_JJ in_II other_JJ countries_NN2 ;_; numerous_JJ multinational_JJ organizations_NN2 ,_, both_RR quasi-governmental_JJ (_( e.g._REX the_AT UN_NP1 Environment_NN1 Programme_NN1 (_( UNEP_NP1 )_) )_) and_CC non-governmental_JJ (_( e.g._REX the_AT International_JJ Union_NN1 for_IF the_AT Conservation_NN1 of_IO Nature_NN1 (_( IUCN_NP1 )_) and_CC the_AT International_JJ Institute_NN1 for_IF Environment_NN1 and_CC Development_NN1 (_( IIED_NP1 )_) )_) are_VBR also_RR active_JJ in_II the_AT field_NN1 ._. 
Moreover_RR ,_, commercial_JJ monitoring_NN1 of_IO the_AT environment_NN1 is_VBZ now_RT also_RR routine_NN1 ,_, notably_RR to_TO predict_VVI crop_NN1 yields_VVZ in_II the_AT main_JJ grain-growing_JJ areas_NN2 and_CC hence_RR facilitate_VV0 the_AT buying_NN1 of_IO futures_NN2 ._. 
The_AT international_JJ scientific_JJ involvement_NN1 Scientists_NN2 have_VH0 both_DB2 promoted_VVN and_CC responded_VVD to_II the_AT increasing_JJ interest_NN1 in_II the_AT environment_NN1 ._. 
Thus_RR ,_, at_II the_AT global_JJ scale_NN1 ,_, the_AT International_JJ Council_NN1 of_IO Scientific_JJ Unions_NN2 (_( ICSU_NP1 )_) has_VHZ endorsed_VVN the_AT setting_NN1 up_RP of_IO the_AT International_JJ Geosphere_NP1 Biosphere_NP1 Project_NP1 (_( IGBP_NP1 )_) to_TO '_" describe_VVI and_CC understand_VVI the_AT interactive_JJ physical_JJ ,_, chemical_JJ and_CC biological_JJ processes_NN2 that_CST regulate_VV0 the_AT total_JJ Earth_NN1 system_NN1 ,_, the_AT unique_JJ environment_NN1 it_PPH1 provides_VVZ for_IF life_NN1 ,_, the_AT changes_NN2 that_CST are_VBR occurring_VVG in_II that_DD1 system_NN1 and_CC the_AT manner_NN1 by_II which_DDQ these_DD2 changes_NN2 are_VBR influenced_VVN by_II human_JJ actions_NN2 '_GE (_( Report_NN1 of_IO ad_JJ21 hoc_JJ22 Planning_NN1 Group_NN1 for_IF IGBP_NP1 1986_MC ,_, quoted_VVN in_II IGBP_NP1 1988_MC )_) ._. 
Priority_NN1 in_II the_AT IGBP_NN1 is_VBZ intended_VVN to_TO be_VBI given_VVN to_II those_DD2 areas_NN2 which_DDQ deal_VV0 with_IW key_JJ interactions_NN2 and_CC significant_JJ changes_NN2 on_II the_AT time-scales_NN2 of_IO decades_NNT2 to_II centuries_NNT2 ,_, that_CST most_DAT affect_VV0 the_AT biosphere_NN1 ,_, that_CST are_VBR most_RGT susceptible_JJ to_II human_JJ perturbations_NN2 and_CC those_DD2 that_CST will_VM most_RGT likely_RR lead_VVI to_II a_AT1 practical_JJ ,_, predictive_JJ capability_NN1 for_IF global_JJ change_NN1 ._. 
This_DD1 is_VBZ to_TO be_VBI put_VVN into_II operation_NN1 by_II concentrating_VVG on_II four_MC themes_NN2 :_: 1_MC1 ._. 
Documenting_VVG and_CC predicting_VVG global_JJ change_NN1 ;_; 2_MC ._. 
Observing_VVG and_CC improving_VVG our_APPGE understanding_NN1 of_IO dominant_JJ forcing_NN1 functions_NN2 ;_; 3_MC ._. 
Improving_VVG our_APPGE understanding_NN1 of_IO interactive_JJ phenomena_NN2 in_II the_AT total_JJ Earth_NN1 system_NN1 ;_; 4_MC ._. 
Assessing_VVG those_DD2 effects_NN2 of_IO global_JJ change_NN1 which_DDQ will_VM be_VBI large_JJ scale_NN1 and_CC cause_VV0 major_JJ modifications_NN2 to_II both_DB2 renewable_JJ and_CC non-renewable_JJ resources_NN2 ._. 
By_II 1990_MC ,_, IGBP_NP1 had_VHD progressed_VVN as_CS31 far_CS32 as_CS33 defining_VVG a_AT1 set_NN1 of_IO seven_MC core_NN1 projects_NN2 (_( IGBP_NP1 1990_MC )_) addressing_VVG these_DD2 four_MC themes_NN2 ._. 
In_RR21 addition_RR22 ,_, plans_NN2 for_IF regional_JJ research_NN1 centres_NN2 and_CC for_IF the_AT data_NN and_CC information_NN1 systems_NN2 for_IF IGBP_NP1 (_( known_VVN as_II IGBP-DIS_NN2 )_) had_VHD been_VBN announced_VVN ._. 
The_AT latter_DA are_VBR particularly_RR relevant_JJ to_II this_DD1 chapter_NN1 ._. 
Based_VVN in_II Paris_NP1 and_CC working_VVG under_II Professor_NNB I._NP1 Rasool_NP1 ,_, it_PPH1 is_VBZ intended_VVN to_TO concentrate_VVI on_II managerial_JJ and_CC policy_NN1 aspects_NN2 of_IO the_AT task_NN1 in_II the_AT first_MD 2_MC years_NNT2 (_( 199092_MC )_) but_CCB then_RT to_TO expand_VVI considerably_RR thereafter_RT ._. 
Tasks_NN2 agreed_VVN for_IF the_AT IGBP-DIS_NN2 include_VV0 making_VVG available_JJ a_AT1 '_GE directory_NN1 of_IO data_NN directories_NN2 '_GE modelled_VVN on_II NASA_NP1 's_GE Master_NN1 Directory_NN1 ,_, providing_VVG data_NN sets_NN2 for_IF education_NN1 and_CC training_NN1 and_CC undertaking_VVG a_AT1 land_NN1 cover_NN1 pilot_NN1 study_NN1 ._. 
Until_CS 1990_MC ,_, remarkably_RR little_JJ interaction_NN1 had_VHD taken_VVN place_NN1 between_II the_AT '_GE hard_JJ science_NN1 '_GE and_CC the_AT social_JJ science_NN1 communities_NN2 on_II global_JJ monitoring_NN1 and_CC prediction_NN1 ._. 
It_PPH1 is_VBZ noteworthy_JJ ,_, however_RR ,_, that_CST the_AT Royal_JJ Society_NN1 made_VVD a_AT1 plea_NN1 for_IF just_RR this_DD1 type_NN1 of_IO multidisciplinary_JJ work_NN1 in_II its_APPGE submission_NN1 on_II greenhouse_NN1 gases_NN2 to_II the_AT House_NN1 of_IO Lords_NP '_GE Select_JJ Committee_NN1 on_II Science_NN1 and_CC Technology_NN1 (_( Royal_JJ Society_NN1 1989_MC :_: 40_MC )_) ._. 
Fortunately_RR ,_, in_II31 addition_II32 to_II33 the_AT actions_NN2 of_IO the_AT world_NN1 '_GE hard_JJ science_NN1 '_GE community_NN1 ,_, a_AT1 parallel_JJ effort_NN1 is_VBZ now_RT being_VBG initiated_VVN by_II social_JJ scientists_NN2 (_( Fuchs_NP1 1989_MC ;_; Jacobson_NP1 and_CC Price_NP1 1990_MC )_) ._. 
Currently_RR this_DD1 is_VBZ organized_VVN and_CC sponsored_VVN by_II the_AT International_JJ Social_JJ Science_NN1 Council_NN1 (_( ISSC_NP1 )_) ,_, the_AT UN_NP1 University_NN1 and_CC the_AT International_JJ Federation_NN1 of_IO Institutes_NN2 of_IO Advanced_JJ Studies_NN2 ._. 
Its_APPGE objectives_NN2 are_VBR to_II :_: 1_MC1 ._. 
Improve_VV0 scientific_JJ understanding_NN1 and_CC increase_NN1 awareness_NN1 of_IO the_AT complex_JJ dynamics_NN governing_VVG human_JJ interaction_NN1 with_IW the_AT total_JJ Earth_NN1 system_NN1 ;_; 2_MC ._. 
Study_NN1 ,_, explore_VV0 and_CC anticipate_VV0 social_JJ change_NN1 affecting_VVG the_AT global_JJ environment_NN1 ;_; 3_MC ._. 
Identify_VV0 broad_JJ social_JJ strategies_NN2 to_TO prevent_VVI or_CC mitigate_VVI undesirable_JJ impacts_NN2 of_IO global_JJ change_NN1 or_CC to_TO adapt_VVI to_II changes_NN2 which_DDQ are_VBR unavoidable_JJ ;_; 4_MC ._. 
Analyse_VV0 policy_NN1 options_NN2 for_IF dealing_VVG with_IW global_JJ environmental_JJ change_NN1 and_CC promoting_VVG the_AT goal_NN1 of_IO sustainable_JJ development_NN1 ._. 
Jacobson_NN1 and_CC Price_NP1 (_( 1990_MC )_) provide_VV0 a_AT1 useful_JJ summary_NN1 of_IO the_AT ways_NN2 in_II which_DDQ human_JJ factors_NN2 contribute_VV0 to_TO ,_, and_CC are_VBR affected_VVN by_II ,_, likely_JJ global_JJ changes_NN2 ;_; in_RR21 particular_RR22 ,_, they_PPHS2 stress_VV0 the_AT importance_NN1 of_IO the_AT data-handling_JJ issues_NN2 and_CC the_AT role_NN1 of_IO GIS_NN2 in_II coping_VVG with_IW some_DD of_IO the_AT problems_NN2 of_IO data_NN integration_NN1 and_CC manipulation_NN1 ._. 
The_AT ISSC_NP1 proposed_VVD its_APPGE own_DA work_NN1 programme_NN1 in_II the_AT Human_JJ Dimensions_NN2 of_IO Global_JJ Change_NN1 in_II late_JJ 1990_MC ._. 
While_CS these_DD2 organizations_NN2 intrinsically_RR have_VH0 a_AT1 world-wide_JJ membership_NN1 and_CC are_VBR concerned_JJ with_IW research_NN1 into_II global_JJ environmental_JJ change_NN1 ,_, others_NN2 share_VV0 the_AT same_DA interest_NN1 but_CCB from_II a_AT1 more_RGR restricted_JJ geographical_JJ base_NN1 ._. 
A_AT1 leading_JJ example_NN1 of_IO this_DD1 is_VBZ the_AT European_JJ Science_NN1 Foundation_NN1 which_DDQ has_VHZ sponsored_VVN research_NN1 programmes_NN2 in_II both_DB2 the_AT physical_JJ and_CC social_JJ science_NN1 aspects_NN2 of_IO environmental_JJ change_NN1 (_( ESF_NP1 1990_MC )_) ._. 
Governmental_JJ and_CC political_JJ involvement_NN1 The_AT spread_NN1 of_IO environmental_JJ awareness_NN1 is_VBZ readily_RR demonstrated_VVN :_: in_II 1972_MC ,_, 26_MC countries_NN2 had_VHD environmental_JJ and_CC natural_JJ resource_NN1 management_NN1 agencies_NN2 whereas_CS ,_, a_AT1 decade_NNT1 later_RRR the_AT number_NN1 was_VBDZ 144_MC (_( WEC_NP1 1985_MC )_) ._. 
In_II most_DAT countries_NN2 ,_, the_AT governmental_JJ and_CC political_JJ concern_NN1 with_IW the_AT environment_NN1 has_VHZ come_VVN somewhat_RR later_RRR than_CSN that_DD1 of_IO scientists_NN2 or_CC concerned_JJ lay_JJ people_NN :_: the_AT early_JJ development_NN1 of_IO the_AT Green_JJ Party_NN1 in_II the_AT Federal_NP1 Republic_NP1 of_IO Germany_NP1 is_VBZ something_PN1 of_IO an_AT1 exception_NN1 ._. 
Such_DA a_AT1 commitment_NN1 to_II the_AT environment_NN1 is_VBZ ,_, none_RR31 the_RR32 less_RR33 ,_, both_RR welcome_JJ and_CC necessary_JJ if_CS resources_NN2 are_VBR to_TO be_VBI made_VVN available_JJ to_TO tackle_VVI existing_JJ ,_, let_II21 alone_II22 forthcoming_JJ ,_, environmental_JJ problems_NN2 ._. 
The_AT '_GE conversion_NN1 '_GE of_IO the_AT UK_NP1 government_NN1 has_VHZ been_VBN briefly_RR described_VVN earlier_RRR ;_; it_PPH1 is_VBZ manifested_VVN in_II their_APPGE July_NPM1 1989_MC commitment_NN1 to_TO spend_VVI 10m._NNU on_II climate_NN1 change_NN1 research_NN1 in_II 1989/90_MC and_CC the_AT confident_JJ request_NN1 to_II them_PPHO2 from_II the_AT Advisory_JJ Board_NN1 for_IF the_AT Research_NN1 Councils_NN2 (_( ABRC_NP1 )_) for_IF an_AT1 extra_JJ 11m._NNU in_II 1990_MC and_CC 13m._NNU in_II the_AT two_MC succeeding_JJ years_NNT2 for_IF additional_JJ environmental_JJ research_NN1 ._. 
Perhaps_RR the_AT most_RGT significant_JJ development_NN1 of_IO all_DB ,_, however_RR ,_, is_VBZ the_AT increasing_JJ role_NN1 of_IO the_AT European_JJ Commission_NN1 ._. 
Their_APPGE role_NN1 in_II bringing_VVG environmental_JJ impact_NN1 assessment_NN1 (_( EIA_NP1 )_) on_II21 to_II22 the_AT statute_NN1 book_NN1 is_VBZ well_RR known_VVN ,_, but_CCB the_AT Commission_NN1 has_VHZ now_RT made_VVN a_AT1 proposal_NN1 (_( CEC_NP1 1989_MC )_) for_IF the_AT '_GE establishment_NN1 of_IO the_AT European_JJ Environment_NN1 Agency_NN1 and_CC the_AT European_JJ Environment_NN1 Monitoring_NN1 and_CC Information_NN1 Network_NN1 '_GE ._. 
The_AT Commission_NN1 argues_VVZ that_CST there_EX is_VBZ at_RR21 present_RR22 no_AT monitoring_NN1 of_IO environmental_JJ quality_NN1 and_CC trends_NN2 on_II a_AT1 European_JJ scale_NN1 ,_, nor_CC any_DD guarantee_VV0 that_CST the_AT results_NN2 of_IO environmental_JJ monitoring_NN1 will_VM be_VBI comparable_JJ on_II a_AT1 Community-wide_JJ basis_NN1 (_( a_AT1 realization_NN1 brought_VVN about_RP through_II the_AT CORINE_JJ programme_NN1 described_VVN later_RRR )_) ._. 
The_AT objectives_NN2 of_IO the_AT new_JJ agency_NN1 would_VM be_VBI to_TO assist_VVI the_AT Community_NN1 and_CC the_AT Member_NN1 States_VVZ to_TO achieve_VVI the_AT goals_NN2 set_VVN out_RP in_II the_AT Treaty_NN1 of_IO Rome_NP1 and_CC in_II successive_JJ environmental_JJ action_NN1 programmes_NN2 ;_; it_PPH1 must_VM also_RR ,_, however_RR ,_, be_VBI seen_VVN in_II the_AT context_NN1 of_IO the_AT European_JJ Council_NN1 's_GE adoption_NN1 on_II 2_MC December_NPM1 1988_MC of_IO the_AT Rhodes_NP1 Declaration_NN1 on_II the_AT Environment_NN1 and_CC the_AT environment_NN1 chapter_NN1 in_II the_AT Single_JJ European_JJ Act_NN1 which_DDQ comes_VVZ into_II effect_NN1 in_II 1992_MC ._. 
The_AT detailed_JJ specific_JJ roles_NN2 of_IO '_GE the_AT System_NN1 '_GE (_( comprising_VVG both_RR agency_NN1 and_CC network_NN1 )_) are_VBR envisaged_VVN as_II :_: 1_MC1 ._. 
To_TO provide_VVI the_AT Community_NN1 ,_, the_AT Member_NN1 States_NP1 and_CC participating_VVG third_MD countries_NN2 with_IW the_AT objective_NN1 (_( sic_RR )_) information_NN1 requested_VVN for_IF the_AT formulation_NN1 and_CC implementation_NN1 of_IO sound_NN1 and_CC effective_JJ environmental_JJ policies_NN2 ;_; 2_MC ._. 
In_RR21 particular_RR22 ,_, to_TO provide_VVI technical_JJ ,_, scientific_JJ and_CC economic_JJ information_NN1 requested_VVN by_II the_AT Commission_NN1 in_II its_APPGE tasks_NN2 of_IO identification_NN1 ,_, preparation_NN1 and_CC assessment_NN1 of_IO the_AT implementation_NN1 and_CC results_NN2 of_IO environmental_JJ action_NN1 and_CC legislation_NN1 ;_; 3_MC ._. 
To_TO stimulate_VVI the_AT development_NN1 and_CC the_AT application_NN1 within_II the_AT agency_NN1 of_IO techniques_NN2 of_IO environmental_JJ modelling_NN1 and_CC forecasting_VVG in_BCL21 order_BCL22 that_CST adequate_JJ preventative_JJ action_NN1 can_VM be_VBI taken_VVN at_II the_AT appropriate_JJ time_NNT1 ;_; 4_MC ._. 
To_TO help_VVI ensure_VVI the_AT harmonization_NN1 and_CC comparability_NN1 of_IO environmental_JJ data_NN in_II the_AT Community_NN1 as_II31 well_II32 as_II33 the_AT integration_NN1 of_IO European_JJ environmental_JJ data_NN into_II international_JJ environmental_JJ monitoring_NN1 programmes_NN2 ,_, such_II21 as_II22 those_DD2 established_VVN within_II the_AT framework_NN1 of_IO the_AT United_JJ Nations_NN2 and_CC its_APPGE system_NN1 of_IO agencies_NN2 ._. 
Considerable_JJ stress_NN1 is_VBZ laid_VVN by_II the_AT Commission_NN1 on_II collaboration_NN1 with_IW national_JJ data-gathering_JJ agencies_NN2 and_CC with_IW states_NN2 adjacent_II21 to_II22 the_AT Community_NN1 's_GE borders_NN2 ._. 
It_PPH1 is_VBZ envisaged_VVN that_CST the_AT totality_NN1 of_IO activities_NN2 will_VM extend_VVI the_AT existing_JJ State_NN1 of_IO the_AT European_JJ Environment_NN1 report_NN1 (_( CEC_NP1 1988_MC )_) through_II addressing_VVG the_AT issues_NN2 of_IO the_AT quality_NN1 and_CC sensitivity_NN1 of_IO the_AT environment_NN1 and_CC also_RR the_AT pressures_NN2 upon_II it_PPH1 ._. 
Priorities_NN2 for_IF the_AT new_JJ agency_NN1 include_VV0 study_NN1 of_IO atmospheric_JJ emissions_NN2 and_CC quality_NN1 ,_, water_NN1 resources_NN2 and_CC quality_NN1 (_( including_II the_AT marine_JJ environment_NN1 )_) ,_, soil_NN1 erosion_NN1 and_CC pollution_NN1 and_CC important_JJ land_NN1 resources_NN2 ,_, and_CC biotopes_NN2 and_CC nature_NN1 conservation_NN1 ._. 
At_II the_AT time_NNT1 of_IO writing_NN1 ,_, the_AT proposal_NN1 has_VHZ been_VBN approved_VVN and_CC discussions_NN2 are_VBR in_II progress_NN1 on_II the_AT location_NN1 of_IO the_AT agency_NN1 's_GE headquarters_NN ._. 
It_PPH1 seems_VVZ certain_JJ to_TO have_VHI major_JJ and_CC growing_JJ effects_NN2 upon_II environmental_JJ monitoring_NN1 and_CC prediction_NN1 in_II Europe_NP1 and_CC on_II the_AT policies_NN2 of_IO component_NN1 governments_NN2 ._. 
Yet_RR this_DD1 is_VBZ by_RR31 no_RR32 means_RR33 the_AT only_JJ plan_NN1 which_DDQ the_AT Community_NN1 has_VHZ in_II31 relation_II32 to_II33 the_AT environment_NN1 :_: Directorate-General_NN1 XII_MC (_( that_DD1 for_IF Science_NN1 ,_, Research_NN1 and_CC Development_NN1 )_) has_VHZ published_VVN details_NN2 of_IO an_AT1 extensive_JJ R&amp;D_NN1 programme_NN1 (_( CEC_NP1 1990a_FO )_) and_CC the_AT Commission_NN1 as_II a_AT1 whole_NN1 (_( CEC_NP1 1990b_FO )_) has_VHZ made_VVN public_RR its_APPGE plans_NN2 to_TO develop_VVI regular_JJ official_JJ statistics_NN of_IO the_AT environment_NN1 ._. 
The_AT private_JJ sector_NN1 Certain_JJ aspects_NN2 of_IO environmental_JJ monitoring_NN1 have_VH0 long_RR been_VBN carried_VVN out_RP by_II the_AT private_JJ sector_NN1 ,_, notably_RR that_DD1 of_IO crop_NN1 states_VVZ on_II a_AT1 world-wide_JJ basis_NN1 ._. 
Derived_VVN from_II satellite_NN1 imagery_NN1 at_II comparatively_RR low_JJ resolution_NN1 ,_, predicted_VVD yields_NN2 for_IF different_JJ crops_NN2 in_II different_JJ nation_NN1 states_NN2 become_VV0 of_IO commercial_JJ value_NN1 ._. 
However_RR ,_, significant_JJ resources_NN2 are_VBR also_RR being_VBG made_VVN available_JJ by_II the_AT private_JJ sector_NN1 where_CS there_EX is_VBZ no_AT immediate_JJ commercial_JJ gain_NN1 :_: thus_RR IBM_NP1 UK_NP1 Ltd_JJ has_VHZ donated_VVN 3.5m._NNU of_IO computing_NN1 equipment_NN1 to_TO UNEP_VVI and_CC IBM_NP1 in_RR21 general_RR22 is_VBZ actively_RR supporting_VVG other_JJ R&amp;D_NN1 which_DDQ can_VM be_VBI expected_VVN to_TO facilitate_VVI '_GE sustainable_JJ development_NN1 '_GE :_: IBM_NP1 Europe_NP1 ,_, for_REX21 instance_REX22 ,_, has_VHZ invested_VVN $16m._NNU in_II its_APPGE Bergen_NP1 scientific_JJ centre_NN1 to_TO make_VVI it_PPH1 the_AT focal_JJ point_NN1 for_IF the_AT company_NN1 's_GE environmental_JJ modelling_NN1 and_CC a_AT1 centre_NN1 for_IF information_NN1 on_II the_AT environment_NN1 and_CC sustainable_JJ development_NN1 ._. 
On_II a_AT1 more_RGR modest_JJ basis_NN1 ,_, IBM_NP1 UK_NP1 have_VH0 agreed_VVN to_TO fund_VVI a_AT1 pilot_NN1 global_JJ population_NN1 database_NN1 being_VBG built_VVN in_II Birkbeck_NP1 College_NN1 to_TO permit_VVI assessment_NN1 of_IO the_AT impact_NN1 of_IO environmental_JJ change_NN1 on_II the_AT world_NN1 s_ZZ1 population_NN1 ._. 
Many_DA2 other_JJ companies_NN2 ,_, of_RR21 course_RR22 ,_, are_VBR supporting_VVG environmental_JJ work_NN1 but_CCB IBM_NP1 seems_VVZ almost_RR to_TO be_VBI unique_JJ in_II its_APPGE corporate_JJ commitment_NN1 to_TO use_VVI of_IO high_JJ technology_NN1 for_IF facilitating_NN1 '_GE sustainable_JJ development_NN1 '_GE ._. 
Some_DD examples_NN2 of_IO environmental_JJ monitoring_NN1 projects_VVZ The_AT local_JJ scale_NN1 At_II the_AT local_JJ scale_NN1 ,_, it_PPH1 is_VBZ usually_RR impossible_JJ to_TO neglect_VVI the_AT effects_NN2 of_IO multiple_JJ processes_NN2 operating_VVG ._. 
An_AT1 example_NN1 of_IO environmental_JJ modelling_NN1 and_CC prediction_NN1 where_CS the_AT interactions_NN2 between_II different_JJ processes_NN2 and_CC dynamic_JJ feedback_NN1 is_VBZ taken_VVN into_II account_NN1 is_VBZ given_VVN by_II Haber_NP1 and_CC Schaller_NP1 (_( 1988_MC )_) ._. 
Their_APPGE work_NN1 formed_VVD part_NN1 of_IO one_MC1 UNESCO_NP1 Man_NN1 and_CC Biosphere_NP1 (_( MAB_NP1 )_) project_NN1 ,_, carried_VVN out_RP in_II the_AT Berchtesgaden_NN1 National_JJ Park_NN1 in_II Bavaria_NP1 ._. 
Following_VVG Holling_NP1 (_( 1978_MC )_) and_CC Grossman_NP1 (_( 1983_MC )_) ,_, they_PPHS2 conceptualized_VVD the_AT man/environment_NN1 relationship_NN1 at_II a_AT1 series_NN of_IO hierarchical_JJ levels_NN2 :_: 1_MC1 ._. 
The_AT lowest_JJT ,_, or_CC process_NN1 ,_, level_NN1 ,_, is_VBZ directly_RR and_CC (_( usually_RR )_) obviously_RR connected_VVN to_II perceived_JJ reality_NN1 ._. 
Thus_RR processes_NN2 and_CC interrelationships_NN2 are_VBR mostly_RR obvious_JJ ,_, simple_JJ and_CC linear_JJ ;_; data_NN are_VBR usually_RR readily_RR available_JJ and_CC may_VM be_VBI voluminous_JJ ._. 
This_DD1 level_NN1 is_VBZ readily_RR handled_VVN in_II commercial_JJ GIS_NN2 ;_; ,_, 2_MC ._. 
The_AT intermediate_JJ ,_, or_CC dynamic_JJ ,_, level_NN1 is_VBZ concerned_JJ with_IW less_RGR frequent_JJ and_CC often_RR irregular_JJ events_NN2 ,_, e.g._REX a_AT1 temporary_JJ lack_NN1 of_IO water_NN1 because_II21 of_II22 fire_NN1 ,_, frost_NN1 or_CC excessive_JJ demand_NN1 from_II an_AT1 increasing_JJ populace_NN1 ._. 
Characteristically_RR ,_, the_AT data_NN used_VVN may_VM be_VBI more_RGR difficult_JJ to_TO obtain_VVI and_CC are_VBR usually_RR more_RGR spatially_RR aggregated_VVN ;_; modelling_VVG software_NN1 which_DDQ permits_VVZ feedback_NN1 loops_NN2 is_VBZ essential_JJ ;_; 3_MC ._. 
The_AT top_NN1 ,_, or_CC strategic_JJ ,_, level_NN1 is_VBZ the_AT most_RGT difficult_JJ to_TO explore_VVI ;_; the_AT external_JJ influences_NN2 are_VBR often_RR unpredictable_JJ by_II formal_JJ means_NN ._. 
Relevant_JJ data_NN are_VBR therefore_RR difficult_JJ to_TO identify_VVI and_CC scenario_NN1 building_NN1 is_VBZ one_MC1 of_IO the_AT few_DA2 approaches_NN2 available_JJ ._. 
The_AT Haber_NN1 and_CC Schaller_NP1 approach_NN1 is_VBZ to_TO '_GE soft-couple_NN1 '_GE these_DD2 different_JJ levels_NN2 and_CC their_APPGE accompanying_JJ approaches_NN2 using_VVG an_AT1 ecological_JJ balance_NN1 model_NN1 ._. 
Results_NN2 of_IO running_VVG this_DD1 with_IW different_JJ scenarios_NN2 include_VV0 '_GE time_NNT1 slice_NN1 '_GE maps_NN2 of_IO possible_JJ outcomes_NN2 and_CC of_IO the_AT exchange_NN1 of_IO material_NN1 and_CC energy_NN1 within_II the_AT study_NN1 area_NN1 ._. 
While_CS ultimately_RR qualitative_JJ ,_, it_PPH1 seems_VVZ to_TO offer_VVI a_AT1 method_NN1 for_IF coping_VVG with_IW different_JJ types_NN2 of_IO process_NN1 and_CC incorporating_VVG stochastic_JJ and_CC truly_RR random_JJ events_NN2 ._. 
The_AT national_JJ and_CC continental_JJ scale_NN1 :_: the_AT CORINE_JJ project_NN1 The_AT CORINE_JJ programme_NN1 represents_VVZ the_AT product_NN1 of_IO much_DA1 activity_NN1 by_II the_AT European_JJ Commission_NN1 's_GE Directorate-General_NN1 (_( XI_NN1 )_) for_IF the_AT Environment_NN1 ._. 
It_PPH1 originates_VVZ from_II an_AT1 Italian_JJ request_NN1 to_II the_AT Council_NN1 of_IO Ministers_NN2 in_II 1973_MC to_TO identify_VVI environmentally_RR '_GE balanced_JJ '_GE and_CC '_GE unbalanced_JJ '_GE areas_NN2 in_II the_AT Community_NN1 ;_; the_AT first_MD attempts_NN2 to_TO do_VDI this_DD1 were_VBDR unsuccessful_JJ and_CC ,_, though_CS by_II 1981_MC it_PPH1 was_VBDZ clear_JJ that_CST a_AT1 new_JJ approach_NN1 based_VVN on_II an_AT1 environmental_JJ information_NN1 system_NN1 was_VBDZ the_AT most_RGT promising_JJ one_PN1 ,_, funding_VVG for_IF this_DD1 was_VBDZ not_XX secured_VVN until_II 1985_MC ._. 
A_AT1 4-year_JJ experimental_JJ programme_NN1 to_TO '_" collect_VVI ,_, co-ordinate_VV0 and_CC ensure_VVI the_AT consistency_NN1 of_IO information_NN1 on_II the_AT state_NN1 of_IO the_AT environment_NN1 and_CC natural_JJ resources_NN2 in_II the_AT European_JJ Communities_NN2 '_GE was_VBDZ set_VVN up_RP and_CC labelled_JJ CORINE_NN1 ._. 
The_AT objectives_NN2 of_IO CORINE_NP1 were_VBDR to_TO be_VBI achieved_VVN through_II bringing_VVG together_RL existing_JJ data_NN holdings_NN2 in_II the_AT member_NN1 states_NN2 ,_, developing_VVG methods_NN2 for_IF holding_VVG ,_, analysing_VVG and_CC presenting_VVG the_AT data_NN ,_, and_CC encouraging_VVG the_AT exchange_NN1 of_IO data_NN ._. 
A_AT1 number_NN1 of_IO priority_NN1 areas_NN2 were_VBDR identified_VVN ,_, including_II biotopes_NN2 of_IO importance_NN1 for_IF nature_NN1 conservation_NN1 ,_, acid_NN1 deposition_NN1 and_CC its_APPGE effects_NN2 on_II soils_NN2 and_CC biotopes_NN2 and_CC the_AT protection_NN1 of_IO the_AT environment_NN1 of_IO the_AT Mediterranean_JJ region_NN1 (_( Briggs_NP1 and_CC Martin_NP1 1988_MC )_) ._. 
Many_DA2 achievements_NN2 can_VM be_VBI claimed_VVN by_II the_AT CORINE_JJ project_NN1 team_NN1 ._. 
As_II a_AT1 result_NN1 of_IO some_DD 40_MC sub-projects_NN2 involving_VVG all_DB the_AT Member_NN1 States_NP1 (_( and_CC often_RR several_DA2 groups_NN2 within_II each_DD1 one_PN1 )_) ,_, databases_NN2 of_IO topography_NN1 ,_, soils_NN2 ,_, water_NN1 resources_NN2 and_CC quality_NN1 ,_, biotopes_NN2 ,_, atmospheric_JJ emissions_NN2 ,_, climate_NN1 ,_, soil_NN1 erosion_NN1 and_CC administrative_JJ boundaries_NN2 have_VH0 been_VBN built_VVN up_RP for_IF the_AT whole_JJ Community_NN1 and_CC can_VM be_VBI interrelated_VVN ;_; other_JJ data_NN sets_NN2 (_( such_II21 as_II22 land_NN1 cover_NN1 ,_, derived_VVN from_II satellite_NN1 imagery_NN1 )_) have_VH0 already_RR been_VBN compiled_VVN for_IF some_DD of_IO the_AT Member_NN1 States_NP1 ._. 
Various_JJ aspects_NN2 of_IO the_AT CORINE_JJ project_NN1 have_VH0 been_VBN described_VVN in_II papers_NN2 by_II Rhind_NP1 et_RA21 al_RA22 ._. 
(_( 1986_MC )_) ,_, Wiggins_NP1 et_RA21 al_RA22 ._. 
(_( 1987_MC )_) ,_, Briggs_NP1 and_CC Martin_NP1 (_( 1988_MC )_) ,_, Briggs_NP1 and_CC Mounsey_NP1 (_( 1989_MC )_) and_CC CEC_NP1 (_( 1990c_FO )_) ._. 
For_IF the_AT present_JJ purposes_NN2 ,_, however_RR ,_, the_AT key_JJ findings_NN2 are_VBR :_: 1_MC1 ._. 
Many_DA2 necessary_JJ data_NN sets_NN2 were_VBDR unavailable_JJ for_IF reasons_NN2 of_IO administrative_JJ inadequacies_NN2 ,_, confidentiality_NN1 constraints_NN2 ,_, cost_NN1 or_CC non-collection_NN1 in_II certain_JJ countries_NN2 ;_; 2_MC ._. 
The_AT unavoidable_JJ reliance_NN1 upon_II existing_JJ data_NN sets_NN2 led_VVN to_II many_DA2 problems_NN2 due_II21 to_II22 non-harmonization_NN1 even_RR of_IO commonly_RR used_JJ variables_NN2 ,_, e.g._REX eight_MC different_JJ definitions_NN2 of_IO potential_JJ evapo-transpiration_NN1 were_VBDR in_II use_NN1 in_II the_AT Member_NN1 States_NP1 and_CC '_GE maximum_JJ temperature_NN1 '_GE at_II each_DD1 weather_NN1 station_NN1 was_VBDZ defined_VVN in_II at_RR21 least_RR22 four_MC different_JJ ways_NN2 ;_; 3_MC ._. 
The_AT compilation_NN1 of_IO a_AT1 spatially_RR coherent_JJ database_NN1 from_II mapped_JJ information_NN1 compiled_VVN by_II different_JJ organizations_NN2 at_II different_JJ scales_NN2 and_CC on_II different_JJ topographic_JJ bases_NN2 (_( see_VV0 Rhind_NP1 and_CC Clark_NP1 1988_MC )_) is_VBZ often_RR difficult_JJ ;_; 4_MC ._. 
User_NN1 expectations_NN2 are_VBR often_RR unreasonably_RR high_JJ ;_; use_NN1 of_IO administrative_JJ computing_NN1 systems_NN2 encourages_VVZ bureaucrats_NN2 to_TO believe_VVI that_DD1 environmental_JJ data_NN can_VM just_RR as_RG readily_RR be_VBI analysed_VVN and_CC used_VVN to_TO support_VVI decisions_NN2 ._. 
The_AT '_GE fuzzy_JJ '_GE nature_NN1 of_IO much_DA1 geographically_RR distributed_VVN data_NN ,_, the_AT inherent_JJ errors_NN2 in_II it_PPH1 ,_, the_AT effects_NN2 of_IO the_AT processing_NN1 algorithms_NN2 used_VVN and_CC the_AT assumption_NN1 of_IO known_JJ interactions_NN2 between_II variables_NN2 ensures_VVZ that_CST completely_RR routine_JJ use_NN1 by_II unskilled_JJ users_NN2 is_VBZ unlikely_JJ to_TO be_VBI possible_JJ in_II the_AT foreseeable_JJ future_NN1 ;_; 5_MC ._. 
More_RGR detailed_JJ data_NN are_VBR required_VVN for_IF particular_JJ tasks_NN2 as_II the_AT role_NN1 of_IO CORINE_NP1 grows_VVZ ._. 
The_AT European_JJ Commission_NN1 's_GE proposal_NN1 for_IF an_AT1 environment_NN1 agency_NN1 (_( see_VV0 above_RL )_) will_VM subsume_VVI the_AT CORINE_JJ team_NN1 and_CC databases_NN2 ._. 
The_AT global_JJ scale_NN1 :_: NASA_NP1 and_CC GEMS/GRID_FU Without_IW question_NN1 ,_, the_AT largest_JJT '_GE player_NN1 '_GE in_II the_AT global_JJ environment_NN1 arena_NN1 is_VBZ NASA_NP1 ._. 
The_AT organization_NN1 has_VHZ initiated_VVN many_DA2 of_IO the_AT schemes_NN2 which_DDQ now_RT are_VBR under_II discussion_NN1 the_AT IGBP_NN1 ,_, for_REX21 instance_REX22 ,_, has_VHZ been_VBN strongly_RR influenced_VVN by_II NASA_NP1 proposals_NN2 ._. 
A_AT1 summary_NN1 of_IO current_JJ activities_NN2 and_CC future_JJ plans_NN2 is_VBZ given_VVN in_II NASA_NP1 (_( 1988b_FO )_) ._. 
By_II31 way_II32 of_II33 example_NN1 ,_, current_JJ programmes_NN2 are_VBR collecting_VVG global_JJ data_NN on_II stratospheric_JJ ozone_NN1 ,_, on_II sea-surface_NN1 and_CC sea-ice_JJ variables_NN2 and_CC on_II the_AT earth_NN1 's_GE radiation_NN1 budget_NN1 ._. 
Approved_JJ future_JJ programmes_NN2 include_VV0 the_AT Upper_JJ Atmosphere_NN1 Research_NN1 Satellite_NN1 ,_, the_AT NASA_NP1 Scatterometer_NP1 and_CC the_AT joint_JJ NASA/CNES_FU Ocean_NN1 Topography_NN1 Experiment_NN1 ._. 
In_RR21 addition_RR22 ,_, however_RR ,_, funding_NN1 has_VHZ recently_RR been_VBN approved_VVN for_IF a_AT1 much_RR more_RGR ambitious_JJ scheme_NN1 the_AT Earth_NN1 Observing_VVG System_NN1 (_( EOS_NP2 )_) ._. 
This_DD1 is_VBZ conceived_VVN not_XX as_II a_AT1 set_NN1 of_IO hardware_NN1 but_CCB as_II a_AT1 comprehensive_JJ information_NN1 system_NN1 focused_VVN on_II the_AT needs_NN2 identified_VVN by_II the_AT ESSC_NP1 (_( see_VV0 above_RL )_) and_CC anticipating_VVG somewhat_RR those_DD2 of_IO IGBP_NN1 ._. 
It_PPH1 will_VM collate_VVI data_NN from_II the_AT two_MC proposed_JJ space_NN1 station_NN1 polar_JJ platforms_NN2 ,_, one_MC1 European_JJ Space_NN1 Agency_NN1 platform_NN1 ,_, a_AT1 Japanese_JJ one_PN1 and_CC also_RR from_II the_AT manned_JJ space_NN1 station_NN1 ._. 
On_II this_DD1 basis_NN1 ,_, a_AT1 data_NN flow_NN1 of_IO no_AT less_DAR than_CSN 1_MC1 terabyte_NN1 (_( i.e._REX 10&amp;sup12_FO ;_; bytes_NNU2 )_) /day_FU is_VBZ anticipated_VVN by_II the_AT mid-1990s_MC2 ._. 
A_AT1 major_JJ user_NN1 of_IO NASA_NP1 data_NN and_CC skills_NN2 has_VHZ been_VBN the_AT UN_NP1 Environment_NN1 Programme._NP1 formed_VVD in_II the_AT wake_NN1 of_IO the_AT 1972_MC Stockholm_NP1 conference_NN1 ,_, UNEP_VV0 soon_RR spawned_VVD the_AT Global_JJ Environment_NN1 Monitoring_NN1 System_NN1 (_( GEMS_NN2 )_) as_II31 well_II32 as_II33 an_AT1 international_JJ register_NN1 of_IO potentially_RR toxic_JJ chemicals_NN2 and_CC a_AT1 global_JJ network_NN1 of_IO sources_NN2 to_TO locate_VVI and_CC provide_VVI technical_JJ ,_, scientific_JJ and_CC management_NN1 information_NN1 on_II the_AT environment_NN1 ._. 
The_AT need_NN1 for_IF a_AT1 mechanism_NN1 to_TO handle_VVI the_AT global_JJ environmental_JJ data_NN was_VBDZ defined_VVN in_II the_AT early_JJ 1980s_MC2 and_CC the_AT Global_JJ Resources_NN2 Information_NN1 Database_NN1 (_( GRID_NN1 )_) resulted_VVD ._. 
Mooneyhan_NP1 (_( 1988_MC )_) has_VHZ described_VVN the_AT progress_NN1 of_IO the_AT pilot_NN1 stage_NN1 of_IO GRID_NN1 which_DDQ culminated_VVD in_II approval_NN1 for_IF a_AT1 full_JJ operational_JJ phase_NN1 ;_; GRID_NN1 is_VBZ now_RT involved_JJ in_II integrating_VVG ,_, storing_VVG and_CC exploiting_VVG a_AT1 variety_NN1 of_IO global_JJ environmental_JJ databases_NN2 ,_, mostly_RR acquired_VVN from_II NASA_NP1 ._. 
In_II31 addition_II32 to_II33 the_AT work_NN1 being_VBG carried_VVN out_RP in_II the_AT UNEP_NN1 HQ_NN1 at_II Nairobi_NP1 and_CC in_II the_AT GEMS/GRID_FU site_NN1 in_II Geneva_NP1 ,_, a_AT1 series_NN of_IO GRID_NN1 regional_JJ nodes_NN2 are_VBR now_RT being_VBG set_VVN up_RP world-wide_JJ ,_, each_DD1 equipped_VVN with_IW the_AT same_DA hardware_NN1 and_CC software_NN1 and_CC local_JJ subsets_NN2 of_IO the_AT data_NN ._. 
The_AT first_MD of_IO these_DD2 is_VBZ in_II Bangkok_NP1 and_CC another_DD1 is_VBZ to_TO be_VBI opened_VVN in_II Latin_JJ America_NP1 ._. 
Moreover_RR ,_, national_JJ nodes_NN2 are_VBR also_RR being_VBG set_VVN up_RP ;_; the_AT recent_JJ IBM_NP1 gift_NN1 to_II GRID_NN1 (_( see_VV0 above_RL )_) has_VHZ ensured_VVN that_CST powerful_JJ microcomputers_NN2 have_VH0 been_VBN installed_VVN in_II many_DA2 African_JJ countries_NN2 ,_, together_RL with_IW national_JJ and_CC continental_JJ data_NN sets_NN2 ._. 
Training_NN1 of_IO unskilled_JJ staff_NN to_TO operate_VVI this_DD1 equipment_NN1 and_CC to_TO exploit_VVI the_AT scientific_JJ data_NN is_VBZ now_RT a_AT1 major_JJ role_NN1 for_IF GRID_NN1 staff_NN ._. 
Technical_JJ aspects_NN2 and_CC problems_NN2 It_PPH1 follows_VVZ from_II much_DA1 of_IO the_AT above_JJ that_CST there_EX are_VBR major_JJ challenges_NN2 in_II use_NN1 of_IO environmental_JJ data_NN ._. 
These_DD2 are_VBR best_RRT considered_VVN by_II itemizing_VVG characteristics_NN2 of_IO contemporary_JJ environmental_JJ monitoring_NN1 and_CC prediction_NN1 :_: 1_MC1 ._. 
The_AT use_NN1 of_IO secondary_JJ data_NN ,_, rather_II21 than_II22 primary_JJ data_NN collected_VVN by_II the_AT data_NN analyst_NN1 ;_; 2_MC ._. 
The_AT need_NN1 ,_, at_RR21 least_RR22 with_IW data_NN on_II the_AT terrestrial_JJ environment_NN1 ,_, for_IF an_AT1 inference_NN1 process_NN1 to_TO extract_VVI useful_JJ information_NN1 from_II the_AT secondary_NN1 (_( and_CC often_RR proxy_NN1 )_) data_NN ;_; 3_MC ._. 
The_AT data_NN volumes_NN2 collected_VVN are_VBR often_RR voluminous_JJ even_RR by_II the_AT standards_NN2 of_IO contemporary_JJ computing_NN1 facilities_NN2 ;_; 4_MC ._. 
The_AT analyses_NN2 carried_VVD out_RP are_VBR often_RR arcane_JJ in_II detail_NN1 to_II all_RR21 but_RR22 a_AT1 small_JJ group_NN1 of_IO '_GE high_JJ priests_NN2 '_GE ;_; 5_MC ._. 
We_PPIS2 have_VH0 only_RR a_AT1 limited_JJ capacity_NN1 for_IF checking_VVG the_AT accuracy_NN1 of_IO many_DA2 environmental_JJ monitoring_NN1 results_NN2 ,_, let_VVD alone_RR of_IO predictions_NN2 ._. 
Some_DD of_IO these_DD2 characteristics_NN2 are_VBR now_RT addressed_VVN in_RP slightly_RR more_DAR detail_NN1 ;_; it_PPH1 will_VM be_VBI obvious_JJ that_CST such_DA problems_NN2 merit_VV0 a_AT1 paper_NN1 in_II themselves_PPX2 ._. 
Remote_JJ sensing_NN1 as_II a_AT1 data_NN collection_NN1 process_NN1 As_II the_AT need_NN1 for_IF more_RGR extensive_JJ ,_, high-resolution_NN1 yet_RR consistent_JJ and_CC up-to-date_JJ data_NN becomes_VVZ more_RGR pressing_JJ ,_, increased_JJ use_NN1 of_IO remote_JJ sensing_NN1 seems_VVZ inevitable_JJ ._. 
No_AT other_JJ data_NN collection_NN1 methodology_NN1 obviates_VVZ the_AT short_JJ distance_NN1 variation_NN1 in_II data_NN sets_NN2 induced_VVN by_II variations_NN2 in_II data_NN collection_NN1 and_CC aggregation_NN1 methodology_NN1 typically_RR encountered_VVN between_II adjacent_JJ nation_NN1 states_NN2 ._. 
The_AT problem_NN1 is_VBZ now_RT recognized_VVN ,_, especially_RR in_II Europe_NP1 ,_, and_CC strenuous_JJ efforts_NN2 are_VBR being_VBG made_VVN to_TO harmonise_VVI the_AT data_NN sets_NN2 (_( e.g._REX through_II the_AT use_NN1 of_IO standard_JJ classifications_NN2 ,_, such_II21 as_II22 that_DD1 for_IF land_NN1 use_NN1 produced_VVN by_II the_AT Conference_NN1 of_IO European_JJ Statisticians_NN2 in_II 1989_MC )_) ._. 
At_RR21 present_RR22 ,_, however_RR ,_, ground-collected_JJ environmental_JJ data_NN are_VBR far_RG less_DAR than_CSN ideal_JJ ._. 
Recognizing_VVG this_DD1 ,_, NASA_NP1 (_( 1988a_FO )_) concluded_VVD that_CST forthcoming_JJ global_JJ science_NN1 projects_NN2 would_VM need_VVI land_NN1 surface_NN1 altitude_NN1 data_NN at_II three_MC resolutions_NN2 :_: 1_MC1 km_NNU resolution_NN1 in_II XY_FO and_CC 10100_MC m_NNO resolution_NN1 in_II Z_ZZ1 on_II a_AT1 global_JJ basis_NN1 ,_, 100_MC m_ZZ1 XY_FO resolution_NN1 and_CC 110_MC m_NNO resolution_NN1 in_II Z_ZZ1 for_IF regional_JJ databases_NN2 and_CC 10_MC m_ZZ1 XY_FO resolution_NN1 and_CC 0.11_MC m_NNO in_II Z_ZZ1 for_IF local_JJ studies_NN2 ._. 
They_PPHS2 argued_VVD that_CST existing_JJ maps_NN2 and_CC digitized_VVD files_NN2 from_II them_PPHO2 are_VBR unable_JK to_TO meet_VVI these_DD2 needs_NN2 at_II global_JJ or_CC regional_JJ scale_NN1 and_CC only_RR remote_JJ sensing_NN1 could_VM help_VVI in_II the_AT short_JJ term_NN1 :_: the_AT availability_NN1 of_IO stereometric_JJ data_NN from_II the_AT French_JJ SPOT_NN1 satellite_NN1 has_VHZ already_RR led_VVN to_II proposals_NN2 for_IF automated_JJ creation_NN1 of_IO global_JJ digital_JJ elevation_NN1 models_NN2 with_IW a_AT1 spatial_JJ (_( XY_FO )_) resolution_NN1 of_IO about_RG 30_MC m_NNU (_( Muller_NP1 1989_MC )_) ._. 
For_IF many_DA2 terrestrial_JJ purposes_NN2 ,_, however_RR ,_, a_AT1 substantial_JJ inference_NN1 process_NN1 must_VM take_VVI place_NN1 to_TO convert_VVI the_AT radiometric_JJ measurements_NN2 into_II useful_JJ information_NN1 (_( in_II contrast_NN1 ,_, more_RGR direct_JJ measurements_NN2 of_IO sea_NN1 state_NN1 ,_, etc._RA are_VBR possible_JJ )_) ._. 
The_AT value_NN1 of_IO this_DD1 information_NN1 is_VBZ inherently_RR dependent_JJ upon_II the_AT quality_NN1 of_IO the_AT inference_NN1 process_NN1 and_CC that_CST in_II turn_NN1 depends_VVZ upon_II the_AT complexity_NN1 of_IO the_AT scene_NN1 ,_, the_AT spatial_JJ ,_, temporal_JJ and_CC spectral_JJ resolution_NN1 of_IO the_AT sensors_NN2 and_CC the_AT particular_JJ algorithms_NN2 used_VVD ._. 
Thus_RR far_RR ,_, for_REX21 instance_REX22 ,_, accuracies_NN2 of_IO land_NN1 cover_NN1 (_( let_II21 alone_II22 what_DDQ is_VBZ often_RR needed_VVN ,_, i.e._REX land_NN1 use_NN1 )_) determined_VVN from_II Landsat_NP1 and_CC SPOT_VV0 imagery_NN1 for_IF the_AT UK_NP1 have_VH0 rarely_RR been_VBN higher_JJR than_CSN 70_MC per_NNU21 cent_NNU22 unless_CS trivial_JJ classifications_NN2 (_( e.g._REX built/unbuilt_FU land_NN1 )_) have_VH0 been_VBN used_VVN ._. 
One_MC1 solution_NN1 to_II this_DD1 is_VBZ to_TO '_" densify_VVI '_GE the_AT ground_NN1 control_NN1 and_CC another_DD1 is_VBZ to_TO use_VVI contextually_RR based_VVN classifiers_NN2 rather_II21 than_II22 the_AT traditional_JJ ,_, spectrally_RR based_VVN ones_NN2 ._. 
But_CCB the_AT ground_NN1 conditions_NN2 in_II some_DD parts_NN2 of_IO Europe_NP1 make_VV0 good_JJ results_NN2 extremely_RR difficult_JJ to_TO obtain_VVI :_: some_DD areas_NN2 of_IO Portugal_NP1 ,_, for_REX21 instance_REX22 ,_, have_VH0 field_NN1 sizes_NN2 of_IO only_RR a_AT1 few_DA2 metres_NNU2 and_CC exhibit_VV0 multi-level_JJ and_CC multi-seasonal_JJ cropping_NN1 ._. 
A_AT1 further_JJR difficulty_NN1 is_VBZ the_AT cost_NN1 of_IO the_AT data_NN in_II31 relation_II32 to_II33 its_APPGE anticipated_JJ benefits_NN2 ._. 
Relatively_RR few_DA2 cost/benefit_VV0 studies_NN2 have_VH0 yet_RR been_VBN carried_VVN out_RP except_CS in_II such_DA areas_NN2 as_CSA crop_NN1 prediction_NN1 and_CC ocean_NN1 navigation_NN1 (_( where_CS they_PPHS2 are_VBR generally_RR shown_VVN to_TO be_VBI highly_RR advantageous_JJ in_II saving_VVG someone_PN1 else_RR 's_GE money_NN1 )_) ._. 
It_PPH1 may_VM be_VBI that_CST a_AT1 study_NN1 commissioned_VVN from_II contractors_NN2 by_II the_AT British_JJ National_JJ Space_NN1 Centre_NN1 will_VM provide_VVI useful_JJ evidence_NN1 on_II the_AT cash_NN1 value_NN1 of_IO remote_JJ sensing_NN1 ;_; it_PPH1 is_VBZ particularly_RR timely_JJ in_II31 view_II32 of_II33 the_AT appointment_NN1 of_IO Professor_NNB Pearce_NP1 (_( see_VV0 above_RL )_) as_II the_AT adviser_NN1 to_II the_AT UK_NP1 Secretary_NN1 of_IO State_NN1 for_IF the_AT Environment_NN1 ._. 
Other_JJ work_NN1 on_II assessment_NN1 of_IO costs_NN2 and_CC benefits_NN2 of_IO such_DA data_NN have_VH0 been_VBN carried_VVN out_RP for_IF the_AT French_JJ government_NN1 by_II Professor_NNB Didier_NP1 (_( 1990_MC )_) ._. 
Finally_RR ,_, since_CS data_NN are_VBR often_RR costly_JJ to_TO acquire_VVI ,_, it_PPH1 follows_VVZ that_CST summaries_NN2 of_IO them_PPHO2 e.g._REX The_AT State_NN1 of_IO the_AT European_JJ Environment_NN1 report_NN1 (_( CEC_NP1 1988_MC )_) and_CC the_AT Environmental_JJ Data_NN Report_NN1 of_IO the_AT UN_NP1 EP_NP1 (_( GEMS_NN2 MARC_NN1 1989_MC )_) are_VBR of_IO considerable_JJ value_NN1 ._. 
Equally_RR ,_, the_AT '_GE signposting_NN1 '_GE role_NN1 now_RT being_VBG adopted_VVN by_II UNEP_NP1 's_GE GRID_NN1 project_NN1 which_DDQ is_VBZ expected_VVN to_TO offer_VVI on-line_JJ access_NN1 to_II a_AT1 catalogue_NN1 of_IO environmental_JJ data_NN sets_NN2 and_CC details_NN2 of_IO the_AT responsible_JJ agencies_NN2 in_II 1991_MC is_VBZ to_TO be_VBI welcomed_VVN ._. 
The_AT data_NN problem_NN1 and_CC partial_JJ solutions_NN2 The_AT lack_NN1 of_IO harmonization_NN1 of_IO definitions_NN2 and_CC methods_NN2 of_IO collection_NN1 of_IO environmental_JJ data_NN have_VH0 already_RR been_VBN outlined_VVN above_RL ._. 
In_II the_AT European_JJ Community_NN1 of_IO 12_MC nation_NN1 states_NN2 ,_, for_REX21 instance_REX22 ,_, no_AT less_DAR than_CSN eight_MC procedures_NN2 for_IF calculating_VVG potential_JJ evapo-transpiration_NN1 have_VH0 been_VBN in_II use_NN1 !_! 
The_AT need_NN1 for_IF harmonization_NN1 is_VBZ obvious_JJ and_CC ,_, in_II this_DD1 respect_NN1 ,_, must_VM follow_VVI from_II the_AT successful_JJ pioneering_JJ achievements_NN2 of_IO EUROSTAT_NN1 in_II harmonizing_VVG the_AT definitions_NN2 in_II trade_NN1 ,_, demographic_JJ and_CC other_NN1 '_GE social_JJ science_NN1 '_GE statistics_NN (_( see_VV0 CEC_NP1 1990b_FO )_) ._. 
An_AT1 additional_JJ and_CC severe_JJ problem_NN1 is_VBZ that_CST the_AT volumes_NN2 of_IO data_NN which_DDQ are_VBR already_RR collected_VVN are_VBR huge_JJ by_II the_AT standards_NN2 of_IO only_RR a_AT1 decade_NNT1 ago_RA ._. 
By_II 1995/96_MC ,_, they_PPHS2 will_VM be_VBI very_RG much_DA1 greater_JJR ,_, notably_RR from_II the_AT EOS_NP2 programme_NN1 ._. 
NASA_NP1 has_VHZ announced_VVN plans_NN2 to_TO collect_VVI up_RG21 to_RG22 1_MC1 terabyte_NN1 (_( 10&amp;sup12_FO ;_; bytes_NNU2 )_) /day_FU ._. 
Since_CS remote_JJ sensing_NN1 data_NN sets_NN2 (_( especially_RR those_DD2 pertaining_II21 to_II22 terrestrial_JJ areas_NN2 of_IO the_AT globe_NN1 )_) typically_RR need_VV0 much_RR pre-processing_JJ to_TO calibrate_VVI ,_, transform_VV0 and_CC then_RT perform_VV0 an_AT1 inference_NN1 process_NN1 and_CC hence_RR convert_VV0 measurements_NN2 of_IO the_AT radiation_NN1 reflected_VVN or_CC emitted_VVD by_II small_JJ areas_NN2 of_IO ground_NN1 into_II useful_JJ (_( e.g._REX land_NN1 cover_NN1 )_) data_NN ,_, the_AT processing_NN1 power_NN1 required_VVN is_VBZ going_VVGK to_TO be_VBI formidable_JJ ._. 
Other_JJ ,_, non-remote-sensing_JJ data_NN sets_NN2 are_VBR smaller_JJR in_II volume_NN1 but_CCB more_RGR complex_JJ in_II structure_NN1 and_CC in_II their_APPGE characteristics_NN2 ._. 
To_TO hold_VVI all_DB of_IO the_AT 1/50_MF 000_MC scale_NN1 topographic_JJ map_NN1 coverage_NN1 alone_RR for_IF the_AT European_JJ Community_NN1 would_VM probably_RR require_VVI about_RG 3_MC terabytes_NN2 of_IO storage_NN1 ._. 
Parenthetically_RR ,_, the_AT form_NN1 in_II which_DDQ the_AT data_NN may_VM be_VBI held_VVN can_VM affect_VVI dramatically_RR the_AT scale_NN1 of_IO the_AT problem_NN1 faced_VVD ._. 
It_PPH1 would_VM be_VBI technically_RR possible_JJ ,_, for_REX21 instance_REX22 ,_, to_TO hold_VVI in_II less_DAR than_CSN 1_MC1 terabyte_NN1 detailed_JJ information_NN1 for_IF every_AT1 individual_NN1 in_II the_AT world_NN1 ,_, akin_JJ to_II that_DD1 collected_VVD about_II everyone_PN1 in_II the_AT UK_NP1 through_II the_AT Census_NN1 of_IO Population_NN1 ._. 
In_II practice_NN1 ,_, legislation_NN1 generally_RR forbids_VVZ such_DA data_NN being_VBG held_VVN thus_RR and_CC distributed_VVN in_II anything_PN1 other_II21 than_II22 area_NN1 aggregate_NN1 form_NN1 ;_; the_AT cross-tabulations_NN2 commonly_RR employed_VVN (_( e.g._REX to_TO give_VVI tables_NN2 of_IO population_NN1 numbers_NN2 broken_VVN down_RP by_II age_NN1 and_CC sex_NN1 )_) often_RR result_VV0 in_II a_AT1 great_JJ multiplication_NN1 of_IO the_AT data_NN volumes_NN2 ._. 
This_DD1 all_DB presents_NN2 major_JJ problems_NN2 of_IO data_NN storage_NN1 ,_, handling_VVG (_( especially_RR for_IF global_JJ data_NN sets_VVZ )_) ,_, display_NN1 and_CC dissemination_NN1 ._. 
Fortunately_RR ,_, technical_JJ developments_NN2 have_VH0 ensured_VVN that_CST the_AT increase_NN1 in_II computing_NN1 power_NN1 per_II unit_NN1 cost_NN1 has_VHZ been_VBN growing_VVG at_II about_II an_AT1 order_NN1 of_IO magnitude_NN1 every_AT1 6_MC years_NNT2 over_II the_AT last_MD three_MC decades_NNT2 ._. 
Indeed_RR ,_, this_DD1 may_VM grossly_RR under-represent_VVI the_AT present_JJ situation_NN1 :_: the_AT rapid_JJ spread_NN1 of_IO UNIX-based_JJ systems_NN2 seems_VVZ to_TO facilitate_VVI competition_NN1 and_CC ,_, at_II the_AT time_NNT1 of_IO writing_NN1 ,_, performance_NN1 of_IO workstations_NN2 per_II unit_NN1 cost_NN1 seems_VVZ to_TO be_VBI increasing_VVG at_II about_RG 50_MC per_NNU21 cent_NNU22 per_RA21 annum_RA22 ._. 
Moreover_RR ,_, the_AT advent_NN1 of_IO low-cost_JJ ,_, high-density-storage_JJ devices_NN2 like_II CD-ROMs_NP2 may_VM well_RR make_VVI substantial_JJ data_NN sets_VVZ available_JJ even_RR to_II those_DD2 with_IW modest_JJ computing_NN1 power_NN1 :_: a_AT1 CD-ROM_NN1 ,_, for_REX21 instance_REX22 ,_, can_VM hold_VVI about_RG 600_MC megabytes_NNU2 (_( i.e._REX 10&amp;sup6_FO ;_; bytes_NNU2 )_) ,_, can_VM be_VBI reproduced_VVN currently_RR for_IF about_II 1_MC1 and_CC read_VVN on_II a_AT1 device_NN1 costing_VVG about_RG 400_MC ._. 
Yet_RR ,_, despite_II such_DA revolutionary_JJ changes_NN2 in_II technology_NN1 ,_, novel_JJ solutions_NN2 will_VM still_RR be_VBI required_VVN to_TO make_VVI effective_JJ use_NN1 of_IO the_AT new_JJ data_NN sets_NN2 ._. 
One_MC1 such_DA solution_NN1 is_VBZ Goodchild_NP1 and_CC Yang_NP1 's_GE (_( 1989_MC )_) scheme_NN1 for_IF a_AT1 hierarchical_JJ spatial_JJ data_NN structure_NN1 to_TO handle_VVI data_NN for_IF the_AT spherical_JJ earth_NN1 ._. 
A_AT1 particular_JJ problem_NN1 is_VBZ that_RG little_DA1 or_CC no_AT quantitative_JJ regard_NN1 has_VHZ been_VBN paid_VVN until_II recently_RR to_II the_AT effects_NN2 of_IO quality_NN1 variation_NN1 in_II environmental_JJ data_NN ._. 
The_AT literature_NN1 of_IO the_AT last_MD few_DA2 years_NNT2 is_VBZ thronged_VVN with_IW papers_NN2 proposing_VVG that_CST these_DD2 must_VM be_VBI taken_VVN into_II account_NN1 but_CCB with_IW few_DA2 realistic_JJ proposals_NN2 on_II how_RRQ this_DD1 is_VBZ to_TO be_VBI achieved_VVN ._. 
Goodchild_NP1 (_( 1988_MC )_) has_VHZ made_VVN perhaps_RR the_AT best_JJT summary_NN1 of_IO the_AT problem_NN1 and_CC possible_JJ solutions_NN2 to_II date_NN1 ._. 
Finally_RR ,_, relatively_RR little_JJ emphasis_NN1 has_VHZ been_VBN given_VVN to_II the_AT management_NN1 of_IO data_NN as_II a_AT1 corporate_JJ resource_NN1 ,_, available_JJ to_II a_AT1 wide_JJ community_NN1 :_: many_DA2 environmental_JJ data_NN are_VBR collected_VVN on_II a_AT1 project-by-project_JJ basis_NN1 and_CC '_GE lost_JJ '_GE once_RR the_AT urgency_NN1 of_IO the_AT initial_JJ research_NN1 is_VBZ over_RP ._. 
However_RR ,_, the_AT 1990_MC report_NN1 of_IO the_AT Data_NN and_CC Facilities_NN2 Working_JJ Group_NN1 of_IO the_AT UK_NP1 Inter-Agency_NP1 Committee_NN1 on_II Global_JJ Environmental_JJ Change_NN1 argued_VVD strongly_RR that_DD1 planning_NN1 for_IF the_AT dissemination_NN1 and_CC maintenance_NN1 of_IO such_DA data_NN was_VBDZ a_AT1 vital_JJ role_NN1 ._. 
It_PPH1 also_RR noted_VVD that_DD1 ownership_NN1 of_IO and_CC access_NN1 to_II environmental_JJ data_NN was_VBDZ a_AT1 policy_NN1 matter_NN1 of_IO the_AT greatest_JJT importance_NN1 ;_; this_DD1 may_VM be_VBI more_RGR difficult_JJ in_II '_GE within_II country_NN1 '_GE data_NN than_CSN with_IW global_JJ data_NN since_CS the_AT latter_DA are_VBR often_RR already_RR the_AT subject_NN1 of_IO international_JJ exchange_NN1 agreements_NN2 at_II zero_MC cost_NN1 ._. 
Global_JJ environmental_JJ research_NN1 and_CC the_AT social_JJ sciences_NN2 Thus_RR far_RR ,_, the_AT emphasis_NN1 has_VHZ been_VBN very_RG much_DA1 on_II the_AT natural_JJ science_NN1 aspects_NN2 of_IO environmental_JJ monitoring_NN1 and_CC prediction_NN1 ._. 
That_DD1 is_VBZ unremarkable_JJ since_CS the_AT anticipation_NN1 of_IO such_DA problems_NN2 and_CC assessments_NN2 of_IO some_DD of_IO their_APPGE causes_NN2 and_CC magnitudes_NN2 can_VM only_RR be_VBI made_VVN by_II environmental_JJ scientists_NN2 ._. 
Yet_RR the_AT consequences_NN2 of_IO any_DD significant_JJ changes_NN2 are_VBR a_AT1 matter_NN1 of_IO the_AT utmost_JJ concern_NN1 to_II planners_NN2 and_CC to_II politicians_NN2 at_II all_DB levels_NN2 ,_, the_AT quantification_NN1 of_IO the_AT costs_NN2 and_CC benefits_NN2 of_IO alternative_JJ policy_NN1 strategies_NN2 is_VBZ very_RG much_RR a_AT1 matter_NN1 for_IF economists_NN2 ,_, and_CC the_AT adaptation_NN1 of_IO societies_NN2 to_II massive_JJ change_NN1 is_VBZ a_AT1 particular_JJ interest_NN1 to_II sociologists_NN2 and_CC others_NN2 ._. 
For_IF geographers_NN2 who_PNQS span_VV0 the_AT conventional_JJ (_( and_CC now_RT archaic_JJ )_) dichotomy_NN1 between_II the_AT natural_JJ and_CC the_AT environmental_JJ sciences_NN2 all_DB of_IO these_DD2 aspects_NN2 need_VV0 to_TO be_VBI woven_VVN together_RL to_TO anticipate_VVI the_AT likely_JJ spatial_JJ patterns_NN2 of_IO the_AT effects_NN2 of_IO massive_JJ change_NN1 ,_, the_AT redistributions_NN2 in_II trade_NN1 ,_, health_NN1 and_CC wealth_NN1 which_DDQ they_PPHS2 will_VM bring_VVI about_RP and_CC the_AT '_GE knock_NN1 on_II '_GE effects_NN2 these_DD2 consequences_NN2 themselves_PPX2 will_VM have_VHI on_II society_NN1 and_CC the_AT environment_NN1 ._. 
For_IF social_JJ scientists_NN2 ,_, then_RT ,_, environmental_JJ monitoring_NN1 and_CC prediction_NN1 are_VBR a_AT1 matter_NN1 of_IO considerable_JJ significance_NN1 ,_, especially_RR as_58 '_GE globalization_NN1 '_GE of_IO markets_NN2 and_CC economies_NN2 proceeds_VVZ ._. 
The_AT Economic_JJ and_CC Social_JJ Research_NN1 Council_NN1 (_( ESRC_NP1 )_) commissioned_VVD two_MC reports_NN2 on_II global_JJ environmental_JJ change_NN1 to_TO guide_VVI or_CC at_RR21 least_RR22 inform_VVI the_AT allocation_NN1 of_IO research_NN1 funding_NN1 ._. 
These_DD2 reports_NN2 (_( Pearce_NP1 1990_MC ;_; Turner_NP1 1990_MC )_) may_VM be_VBI taken_VVN as_CSA informed_VVN views_NN2 from_II at_RR21 least_RR22 the_AT economists_NN2 in_II the_AT UK_NP1 social_JJ science_NN1 community_NN1 on_II the_AT social_JJ science_NN1 aspects_NN2 of_IO global_JJ environmental_JJ change_NN1 ._. 
Pearce_NP1 (_( 1990_MC )_) ,_, for_REX21 instance_REX22 ,_, presented_VVD an_AT1 economist_NN1 's_GE views_NN2 of_IO the_AT high-priority_JJ topics_NN2 for_IF research_NN1 as_CSA follows_VVZ :_: 1_MC1 ._. 
The_AT theory_NN1 and_CC practice_NN1 of_IO international_JJ agreements_NN2 ,_, viewed_VVN in_II a_AT1 game_NN1 theory_NN1 context_NN1 ;_; 2_MC ._. 
The_AT theory_NN1 of_IO optimal_JJ behaviour_NN1 under_II scientific_JJ uncertainty_NN1 ;_; ,_, 3_MC ._. 
Cost-benefit_JJ frameworks_NN2 ,_, especially_RR for_IF dealing_VVG with_IW man-induced_JJ climatic_JJ effects_NN2 ;_; 4_MC ._. 
Definition_NN1 of_IO the_AT impacts_NN2 of_IO climate_NN1 change_NN1 on_II eastern_JJ Europe_NP1 and_CC the_AT developing_JJ countries_NN2 ;_; 5_MC ._. 
Appropriate_JJ policies_NN2 on_II non-carbon_JJ dioxide_NN1 greenhouse_NN1 gases_NN2 ;_; 6_MC ._. 
Assessment_NN1 of_IO the_AT ecological_JJ values_NN2 of_IO tropical_JJ forest_NN1 ;_; 7_MC ._. 
Economic_JJ instruments_NN2 for_IF reducing_VVG greenhouse_NN1 gases_NN2 ;_; 8_MC ._. 
The_AT international_JJ politics_NN1 of_IO climate_NN1 change_NN1 ;_; 9_MC ._. 
National_JJ and_CC global_JJ energy_NN1 options_NN2 given_VVN the_AT greenhouse_NN1 effect_NN1 ;_; 10_MC ._. 
Trade_NN1 and_CC global_JJ change_NN1 ._. 
Turner_NP1 (_( 1990_MC )_) argued_VVD that_CST the_AT priorities_NN2 are_VBR as_CSA follows_VVZ :_: 1_MC1 ._. 
The_AT socio-economic_JJ consequences_NN2 of_IO sea-level_JJ rise_NN1 ;_; 2_MC ._. 
Techniques_NN2 for_IF the_AT valuation_NN1 of_IO environmental_JJ resources_NN2 and_CC effects_NN2 ;_; 3_MC ._. 
An_AT1 assessment_NN1 of_IO management_NN1 tools_NN2 ,_, institutional_JJ adaptation_NN1 and_CC policy_NN1 integration_NN1 in_II31 regard_II32 to_II33 global_JJ environmental_JJ change_NN1 ;_; 4_MC ._. 
Studies_NN2 of_IO agriculture_NN1 in_II31 relation_II32 to_II33 climate_NN1 change_NN1 ;_; 5_MC ._. 
Understanding_NN1 of_IO how_RGQ environmental_JJ issues_NN2 and_CC technical_JJ change_NN1 interact_VV0 ;_; 6_MC ._. 
The_AT interrelationship_NN1 between_II environmental_JJ ethics_NN and_CC economics_NN1 ._. 
There_EX is_VBZ ,_, then_RT ,_, no_AT shortage_NN1 of_IO social_JJ science_NN1 research_NN1 which_DDQ is_VBZ claimed_VVN to_TO be_VBI necessary_JJ if_CS we_PPIS2 are_VBR to_TO cope_VVI with_IW environmental_JJ change_NN1 ._. 
Some_DD of_IO it_PPH1 is_VBZ being_VBG carried_VVN out_RP :_: in_II March_NPM1 1991_MC ESRC_NP1 announced_VVD a_AT1 6m_NNU programme_NN1 of_IO research_NN1 on_II Global_JJ Environmental_JJ Change_NN1 ._. 
However_RR ,_, perhaps_RR most_RGT critical_JJ of_IO all_DB in_II the_AT early_JJ stages_NN2 is_VBZ social_JJ science_NN1 practice_NN1 rather_CS21 than_CS22 research_VVI ._. 
In_II the_AT USA_NP1 at_RR21 least_RR22 ,_, there_EX needs_VVZ to_TO be_VBI an_AT1 improvement_NN1 in_II scientists_NN2 '_GE understandings_NN2 of_IO how_RRQ to_TO influence_VVI the_AT allocation_NN1 of_IO funding_NN1 for_IF remedial_JJ measures_NN2 or_CC for_IF further_JJR research_NN1 (_( see_VV0 Kitsos_NN2 and_CC Ashe_VV0 1989_MC )_) ._. 
All_RR too_RG often_RR ,_, it_PPH1 seems_VVZ that_CST the_AT coupling_NN1 of_IO an_AT1 as_RR21 yet_RR22 imperfect_JJ scientific_JJ understanding_NN1 to_II policy-making_NN1 is_VBZ tenuous_JJ (_( see_VV0 Table_NN1 9.2_MC )_) ._. 
Fortunately_RR ,_, considerable_JJ high-quality_JJ work_NN1 has_VHZ already_RR begun_VVN on_II the_AT legal_JJ ,_, political_JJ and_CC institutional_JJ aspects_NN2 of_IO global_JJ environmental_JJ change_NN1 (_( see_VV0 ,_, for_REX21 instance_REX22 ,_, Nitze_NP1 1990_MC )_) ._. 
What_DDQ next_MD ?_? 
The_AT contents_NN2 of_IO this_DD1 chapter_NN1 have_VH0 scarcely_RR touched_VVN upon_II such_DA important_JJ environmental_JJ research_NN1 as_CSA the_AT World_NP1 Ocean_NNL1 Climate_NN1 Experiment_NN1 or_CC the_AT World_NN1 Climate_NN1 Research_NN1 Programme_NN1 ._. 
Nor_CC has_VHZ there_EX been_VBN any_DD discussion_NN1 of_IO the_AT growing_JJ requirement_NN1 in_II the_AT UK_NP1 (_( as_CSA elsewhere_RL )_) to_TO carry_VVI out_RP an_AT1 environmental_JJ impact_NN1 assessment_NN1 for_IF all_DB major_JJ developments_NN2 ._. 
The_AT consequences_NN2 for_IF some_DD nations_NN2 or_CC subgroups_VVZ in_II a_AT1 population_NN1 of_IO being_VBG without_IW access_NN1 to_II environmental_JJ data_NN or_CC the_AT tools_NN2 for_IF analysing_VVG them_PPHO2 while_CS others_NN2 have_VH0 such_DA access_NN1 are_VBR potentially_RR profound_JJ as_CSA the_AT European_JJ Commission_NN1 has_VHZ recognized_VVN in_II issuing_VVG a_AT1 directive_NN1 on_II public_NN1 access_NN1 to_II environmental_JJ data_NN (_( CEC_NP1 1990d_NNU )_) yet_RR space_VV0 limitations_NN2 have_VH0 also_RR precluded_VVN their_APPGE discussion_NN1 ._. 
Despite_II these_DD2 shortcomings_NN2 dictated_VVN by_II the_AT space_NN1 available_JJ ,_, what_DDQ has_VHZ been_VBN said_VVN should_VM be_VBI sufficient_JJ to_TO illustrate_VVI the_AT enormous_JJ and_CC diverse_JJ scope_NN1 of_IO environmental_JJ monitoring_NN1 and_CC prediction_NN1 ._. 
Such_DA is_VBZ its_APPGE diversity_NN1 that_CST measuring_VVG the_AT total_JJ expenditure_NN1 on_II it_PPH1 seems_VVZ quite_RG impossible_JJ and_CC the_AT number_NN1 of_IO '_GE actors_NN2 '_GE already_RR involved_VVN is_VBZ vast_JJ ._. 
What_DDQ is_VBZ clear_JJ ,_, however_RR ,_, is_VBZ that_DD1 monitoring_NN1 is_VBZ rarely_RR a_AT1 simple_JJ exercise_NN1 ,_, that_CST the_AT time-scales_NN2 of_IO the_AT sponsors_NN2 of_IO monitoring_NN1 and_CC environmental_JJ research_NN1 are_VBR rarely_RR long_JJ term_NN1 though_CS most_DAT science_NN1 is_VBZ inevitably_RR of_IO this_DD1 nature_NN1 and_CC that_CST the_AT increasingly_RR multinational_JJ nature_NN1 of_IO such_DA work_NN1 necessitates_VVZ a_AT1 degree_NN1 of_IO management_NN1 not_XX always_RR present_JJ in_II local_JJ or_CC nationally_RR based_VVN schemes_NN2 ._. 
In_II these_DD2 circumstances_NN2 ,_, it_PPH1 is_VBZ scarcely_RR surprising_JJ that_CST there_EX is_VBZ evidence_NN1 of_IO international_JJ and_CC interdisciplinary_JJ dispute_NN1 ,_, as_CSA Terney_NP1 (_( 1989_MC )_) has_VHZ chronicled_VVN in_II31 regard_II32 to_II33 IGBP_NP1 ._. 
Moreover_RR ,_, as_CSA Newby_NP1 (_( 1990_MC )_) has_VHZ pointed_VVN out_RP ,_, the_AT complexities_NN2 of_IO environmental_JJ research_NN1 have_VH0 created_VVN a_AT1 new_JJ kind_NN1 of_IO relationship_NN1 between_II research_NN1 and_CC policy_NN1 :_: '_GE in_II the_AT past_NN1 ,_, ..._... the_AT relationship_NN1 was_VBDZ predicated_VVN on_II the_AT belief_NN1 that_CST science_NN1 provided_CS decision-makers_NN2 with_IW objective_NN1 '_GE hard_JJ '_GE facts_NN2 on_II which_DDQ to_TO base_VVI their_APPGE soft_JJ ,_, value-ridden_JJ policies_NN2 ...._... 
But_CCB now_RT we_PPIS2 find_VV0 scientists_NN2 delivering_VVG only_RR '_GE soft_JJ '_GE ,_, uncertain_JJ '_GE facts_NN2 '_GE to_II decision-makers_NN2 facing_NN1 '_GE hard_JJ '_GE decisions_NN2 ._. 
'_GE Yet_RR ,_, even_CS21 if_CS22 all_DB of_IO this_DD1 causes_VVZ great_JJ uncertainty_NN1 among_II scientists_NN2 ,_, it_PPH1 is_VBZ still_RR a_AT1 reasonable_JJ prediction_NN1 that_CST a_AT1 decade_NNT1 ahead_RL will_VM see_VVI both_RR environmental_JJ monitoring_NN1 and_CC prediction_NN1 treated_VVN as_II an_AT1 everyday_JJ need_NN1 and_CC activity_NN1 in_II major_JJ organizations_NN2 ._. 
We_PPIS2 shall_VM also_RR see_VVI the_AT existence_NN1 of_IO binding_JJ requirements_NN2 upon_II national_JJ governments_NN2 certainly_RR in_II the_AT European_JJ Community_NN1 to_TO collect_VVI at_RR21 least_RR22 basic_JJ environmental_JJ indicators_NN2 in_II a_AT1 harmonized_JJ form_NN1 (_( CEC_NP1 1990b_FO )_) ._. 
More_DAR than_CSN that_DD1 ,_, it_PPH1 seems_VVZ highly_RR likely_JJ that_CST the_AT social_JJ science_NN1 aspects_NN2 of_IO environmental_JJ problems_NN2 will_VM underpin_VVI the_AT actions_NN2 of_IO governments_NN2 and_CC international_JJ agencies_NN2 ._. 
Perhaps_RR the_AT single_JJ most_RGT important_JJ conclusion_NN1 to_TO be_VBI drawn_VVN from_II this_DD1 review_NN1 is_VBZ that_CST the_AT monitoring_NN1 ,_, modelling_VVG and_CC management_NN1 of_IO the_AT global_JJ environment_NN1 are_VBR tasks_NN2 which_DDQ uniquely_RR require_VV0 the_AT integration_NN1 of_IO skills_NN2 and_CC techniques_NN2 from_II many_DA2 disciplines_NN2 ._. 
Building_VVG such_DA research_NN1 teams_NN2 is_VBZ essential_JJ ._. 
Underlying_VVG it_PPH1 all_DB ,_, however_RR ,_, is_VBZ one_MC1 obvious_JJ fact_NN1 :_: that_CST the_AT gathering_NN1 and_CC analysis_NN1 of_IO geographically_RR distributed_VVN environmental_JJ data_NN form_VV0 the_AT necessary_JJ starting-point_NN1 for_IF ensuring_VVG the_AT success_NN1 of_IO our_APPGE future_NN1 on_II earth_NN1 ._. 
Acknowledgements_NN2 Thanks_NN2 are_VBR due_II21 to_II22 Tina_NP1 Buckle_NN1 and_CC Jonathan_NP1 Raper_NP1 for_IF remastering_VVG the_AT diagrams_NN2 ._. 
The_AT UK_NP1 ESRC_NP1 supported_VVD the_AT work_NN1 of_IO the_AT South_ND1 East_ND1 Regional_JJ Research_NN1 Laboratory_NN1 in_II Birkbeck_NP1 College_NN1 ,_, University_NN1 of_IO London_NP1 ,_, from_II which_DDQ this_DD1 review_NN1 has_VHZ grown_VVN ._. 
Managing_JJ natural_JJ and_CC technological_JJ hazards_NN2 Anthony_NP1 C._NP1 Gatrell_NP1 and_CC Peter_NP1 Vincent_NP1 Introduction_NN1 Few_DA2 areas_NN2 of_IO the_AT application_NN1 of_IO GIS_NN2 technology_NN1 can_VM be_VBI as_RG socially_RR significant_JJ ,_, or_CC environmentally_RR relevant_JJ ,_, as_CSA the_AT management_NN1 of_IO emergencies_NN2 and_CC disasters_NN2 due_II21 to_II22 natural_JJ and_CC technological_JJ hazards_NN2 ._. 
Whether_CSW attempting_VVG to_TO construct_VVI a_AT1 database_NN1 of_IO resources_NN2 for_IF use_NN1 in_II planning_VVG responses_NN2 to_II nuclear_JJ emergencies_NN2 ,_, developing_VVG optimal_JJ routes_NN2 for_IF scheduling_VVG the_AT safe_JJ transport_NN1 of_IO hazardous_JJ substances_NN2 ,_, or_CC monitoring_VVG the_AT health_NN1 implications_NN2 of_IO a_AT1 disaster_NN1 ,_, GIS_NN2 can_VM assist_VVI in_II identifying_VVG possibilities_NN2 and_CC formulating_VVG solutions_NN2 ._. 
The_AT 1990s_MC2 have_VH0 been_VBN declared_VVN the_AT International_JJ Decade_NNT1 for_IF Natural_JJ Disaster_NN1 Reduction_NN1 ._. 
Let_VV0 us_PPIO2 hope_VVI that_CST G1S_FO can_VM ,_, in_II some_DD small_JJ way_NN1 ,_, help_VV0 mitigate_VVI the_AT suffering_NN1 and_CC hardship_NN1 felt_VVD by_II all_DB those_DD2 afflicted_VVN by_II the_AT effects_NN2 of_IO hazards_NN2 which_DDQ ,_, as_CSA we_PPIS2 have_VH0 learnt_VVN with_IW bitter_JJ recent_JJ experience_NN1 in_II the_AT UK_NP1 ,_, can_VM occur_VVI in_II an_AT1 untimely_JJ fashion_NN1 in_II the_AT most_RGT improbable_JJ of_IO places_NN2 ._. 
Research_NN1 on_II hazards_NN2 is_VBZ multidisciplinary_JJ and_CC straddles_VVZ the_AT social_JJ ,_, environmental_JJ and_CC mathematical_JJ sciences_NN2 ._. 
If_CS ,_, for_REX21 instance_REX22 ,_, we_PPIS2 wish_VV0 to_TO model_VVI the_AT dispersal_NN1 of_IO a_AT1 plume_NN1 of_IO toxic_JJ gas_NN1 resulting_VVG from_II a_AT1 chemical_JJ explosion_NN1 ,_, assess_VV0 its_APPGE possible_JJ impact_NN1 on_II human_JJ health_NN1 and_CC evaluate_VV0 likely_JJ evacuation_NN1 scenarios_NN2 of_IO the_AT population_NN1 at_II risk_NN1 we_PPIS2 would_VM require_VVI ,_, for_REX21 example_REX22 ,_, a_AT1 knowledge_NN1 of_IO atmospheric_JJ dispersion_NN1 models_NN2 ,_, epidemiology_NN1 and_CC population_NN1 distribution_NN1 ._. 
Some_DD researchers_NN2 suggest_VV0 that_CST hazards_NN2 (_( and_CC emergencies_NN2 that_CST might_VM result_VVI )_) can_VM be_VBI regarded_VVN as_II either_RR natural_JJ or_CC technological_JJ ._. 
We_PPIS2 shall_VM elaborate_VVI on_II this_DD1 simple_JJ division_NN1 below_RL but_CCB suggest_VV0 here_RL that_CST the_AT wealth_NN1 of_IO literature_NN1 on_II the_AT former_DA (_( see_VV0 ,_, for_REX21 instance_REX22 ,_, Burton_NP1 et_RA21 al_RA22 ._. 
1978_MC and_CC Perry_NP1 1981_MC )_) has_VHZ yet_RR to_TO be_VBI matched_VVN by_II a_AT1 similar_JJ volume_NN1 of_IO work_NN1 on_II the_AT latter_DA ._. 
There_EX are_VBR signs_NN2 that_CST this_DD1 imbalance_NN1 is_VBZ being_VBG rectified_VVN (_( Zeigler_NP1 et_RA21 al_RA22 ._. 
1983_MC )_) ._. 
Hazard_NN1 research_NN1 ,_, like_II few_DA2 other_JJ GIS_NN2 application_NN1 areas_NN2 ,_, is_VBZ not_XX only_RR stretching_VVG the_AT present_JJ technology_NN1 to_II its_APPGE limits_NN2 but_CCB is_VBZ a_AT1 quite_RG remarkable_JJ focus_NN1 of_IO international_JJ effort_NN1 into_II several_DA2 important_JJ research-related_JJ areas_NN2 such_II21 as_II22 expert_NN1 systems_NN2 and_CC simulation_NN1 studies_NN2 ._. 
In_II this_DD1 sense_NN1 the_AT scene_NN1 is_VBZ clearly_RR one_MC1 of_IO hazard_NN1 problems_NN2 looking_VVG for_IF improved_JJ GIS_NN2 rather_II21 than_II22 for_IF GIS_NN2 looking_VVG for_IF good_JJ problems_NN2 ,_, as_CSA is_VBZ all_RR too_RG often_RR the_AT case_NN1 ._. 
Hazard_NN1 research_NN1 What_DDQ are_VBR hazards_NN2 ?_? 
Formally_RR ,_, a_AT1 hazard_NN1 can_VM be_VBI defined_VVN as_II :_: '_GE a_AT1 physical_JJ situation_NN1 with_IW a_AT1 potential_NN1 for_IF human_JJ injury_NN1 ,_, damage_VV0 to_II property_NN1 ,_, damage_VV0 to_II the_AT environment_NN1 ,_, or_CC some_DD combination_NN1 of_IO these_DD2 '_GE (_( Health_NN1 and_CC Safety_NN1 Executive_NN1 1989_MC :_: 30_MC )_) ._. 
A_AT1 hazard_NN1 is_VBZ a_AT1 threat_NN1 which_DDQ ,_, given_VVN a_AT1 set_NN1 of_IO circumstances_NN2 ,_, may_VM become_VVI translated_VVN into_II a_AT1 realized_JJ event_NN1 ._. 
What_DDQ we_PPIS2 choose_VV0 to_TO call_VVI this_DD1 event_NN1 depends_VVZ upon_II factors_NN2 such_II21 as_II22 its_APPGE magnitude_NN1 :_: we_PPIS2 speak_VV0 for_REX21 instance_REX22 of_IO accidents_NN2 ,_, emergencies_NN2 ,_, disasters_NN2 and_CC catastrophes_NN2 ,_, each_DD1 of_IO which_DDQ carries_VVZ connotations_NN2 concerning_II the_AT scale_NN1 of_IO the_AT event_NN1 and_CC each_DD1 of_IO which_DDQ will_VM have_VHI a_AT1 set_NN1 of_IO human_NN1 and/or_CC environmental_JJ consequences_NN2 ._. 
On_II 30_MC March_NPM1 1956_MC one_MC1 of_IO the_AT most_RGT powerful_JJ volcanic_JJ disturbances_NN2 this_DD1 century_NNT1 ,_, the_AT Bezymianny_JJ eruption_NN1 ,_, occurred_VVD ._. 
Few_DA2 people_NN noticed_VVD the_AT event_NN1 as_CSA it_PPH1 took_VVD place_NN1 in_II an_AT1 uninhabited_JJ part_NN1 of_IO Kamchatcka_NP1 and_CC caused_VVD no_AT known_JJ casualties_NN2 ._. 
In_II contrast_NN1 ,_, the_AT extrusion_NN1 of_IO a_AT1 small_JJ volume_NN1 of_IO lava_NN1 from_II a_AT1 secondary_JJ cone_NN1 on_II the_AT slopes_NN2 of_IO Tristan_NP1 da_NP1 Cunha_NP1 became_VVD the_AT focus_NN1 of_IO global_JJ interest_NN1 when_RRQ ,_, during_II October_NPM1 1961_MC ,_, the_AT island_NN1 's_GE lobster_NN1 processing_NN1 plant_NN1 was_VBDZ smothered_VVN and_CC the_AT 300_MC people_NN of_IO the_AT island_NN1 's_GE only_JJ settlement_NN1 were_VBDR evacuated_VVN to_II the_AT UK_NP1 ._. 
These_DD2 two_MC contrasting_JJ events_NN2 emphasize_VV0 the_AT central_JJ fact_NN1 that_CST even_RR natural_JJ hazards_NN2 are_VBR not_XX defined_VVN solely_RR by_II the_AT characteristics_NN2 of_IO the_AT event_NN1 but_CCB by_II the_AT interaction_NN1 of_IO those_DD2 events_NN2 with_IW the_AT human_JJ occupation_NN1 of_IO the_AT threatened_JJ area_NN1 ._. 
As_CSA noted_VVN above_RL ,_, the_AT simple_JJ division_NN1 of_IO hazards_NN2 is_VBZ between_II those_DD2 which_DDQ are_VBR natural_JJ and_CC those_DD2 which_DDQ are_VBR technological_JJ or_CC human-induced_NN1 ._. 
Each_DD1 of_IO these_DD2 may_VM be_VBI further_RRR subdivided_VVN and_CC the_AT typology_NN1 shown_VVN in_II Fig._NN1 10.1_MC draws_NN2 upon_II the_AT classic_JJ work_NN1 of_IO Burton_NP1 and_CC Kates_NP2 (_( 1964_MC )_) in_II separating_VVG geophysical_JJ from_II biological_JJ hazards_NN2 and_CC Zeigler_NP1 et_RA21 al_RA22 ._. 
(_( 1983_MC )_) in_II distinguishing_JJ public_NN1 (_( or_CC what_DDQ we_PPIS2 might_VM call_VVI societal_JJ )_) hazards_NN2 from_II private_JJ (_( individual_NN1 )_) ones_NN2 ._. 
The_AT further_JJR subdivision_NN1 of_IO natural_JJ hazards_NN2 is_VBZ self-explanatory_JJ as_CSA is_VBZ that_DD1 of_IO individual_JJ technological_JJ hazards_NN2 ._. 
Zeigler_NP1 divides_VVZ the_AT public_JJ hazards_NN2 into_II those_DD2 which_DDQ result_VV0 from_II the_AT production_NN1 of_IO raw_JJ materials_NN2 and_CC manufactured_JJ goods_NN2 ,_, those_DD2 which_DDQ originate_VV0 in_II transport_NN1 and_CC transmission_NN1 and_CC those_DD2 which_DDQ threaten_VV0 the_AT public_NN1 in_II their_APPGE role_NN1 as_CSA consumers_NN2 ._. 
The_AT United_JJ Nations_NN2 Disaster_NN1 Relief_NN1 Organization_NN1 (_( UNDRO_NP1 )_) has_VHZ produced_VVN a_AT1 classification_NN1 of_IO disasters_NN2 ,_, derived_VVN on_II the_AT basis_NN1 of_IO case_NN1 data_NN (_( Table_NN1 10.1_MC )_) ._. 
UNDRO_NP1 recognizes_VVZ three_MC basic_JJ types_NN2 of_IO disaster_NN1 :_: accidents_NN2 ,_, natural_JJ events_NN2 ,_, and_CC other_JJ disasters_NN2 ._. 
The_AT differentiation_NN1 of_IO accidents_NN2 from_II other_JJ types_NN2 is_VBZ useful_JJ in_II the_AT context_NN1 of_IO emergency_NN1 planning_NN1 and_CC is_VBZ not_XX so_RG clearly_RR conveyed_VVN by_II other_JJ hazard_NN1 typologies_NN2 ._. 
However_RR ,_, the_AT temptation_NN1 in_II all_DB classification_NN1 exercises_NN2 is_VBZ to_TO produce_VVI partitional_JJ schemes_NN2 and_CC this_DD1 should_VM really_RR be_VBI resisted_VVN here_RL ._. 
It_PPH1 is_VBZ more_RGR useful_JJ (_( Johnson_NP1 1983_MC )_) to_TO adopt_VVI a_AT1 non-partitional_JJ scheme_NN1 (_( fig._NN1 10.1_MC )_) that_CST recognizes_VVZ ,_, for_REX21 instance_REX22 ,_, food_NN1 poisoning_VVG as_RG both_RR a_AT1 personal_JJ and_CC consumer_NN1 hazard_NN1 ,_, or_CC lead_VV0 pollution_NN1 (_( e.g._REX from_II car_NN1 exhausts_VVZ )_) as_CSA both_RR a_AT1 meteorological_JJ event_NN1 and_CC technological_JJ hazard_NN1 (_( both_RR public_JJ and_CC private_JJ )_) ._. 
An_AT1 air_NN1 crash_NN1 (_( as_CSA at_II Lockerbie_NP1 )_) affects_VVZ an_AT1 entire_JJ community_NN1 directly_RR ,_, as_II31 well_II32 as_II33 being_VBG an_AT1 occupational_JJ hazard_NN1 and_CC risk_NN1 to_II the_AT individual_JJ traveller_NN1 ._. 
Hillslope_VV0 failure_NN1 (_( as_CSA at_II Aberfan_NP1 in_II South_ND1 Wales_NP1 in_II 1966_MC )_) is_VBZ a_AT1 geomorphic_JJ hazard_NN1 but_CCB may_VM arise_VVI because_II21 of_II22 raw_JJ material_NN1 extraction_NN1 ._. 
Dutch_JJ elm_NN1 disease_NN1 is_VBZ a_AT1 floral_JJ hazard_NN1 but_CCB is_VBZ exacerbated_VVN by_II the_AT transport_NN1 of_IO infected_JJ logs_NN2 ._. 
Hohensemer_NP1 et_RA21 al_RA22 ._. 
(_( 1983_MC )_) make_VV0 the_AT important_JJ point_NN1 that_CST hazards_NN2 are_VBR multidimensional_JJ in_II nature_NN1 and_CC suggest_VV0 12_MC dimensions_NN2 along_II which_DDQ any_DD hazard_NN1 may_VM be_VBI scored_VVN (_( see_VV0 Table_NN1 10.2_MC )_) ._. 
These_DD2 factors_NN2 will_VM ,_, of_RR21 course_RR22 ,_, be_VBI important_JJ in_II deciding_VVG on_II the_AT role_NN1 that_CST GIS_NN2 can_VM play_VVI in_II hazard_NN1 monitoring_NN1 and_CC emergency_NN1 planning_NN1 ._. 
For_REX21 instance_REX22 ,_, some_DD events_NN2 will_VM persist_VVI briefly_RR ,_, have_VH0 immediate_JJ consequences_NN2 ,_, minor_JJ transgenerational_JJ effects_NN2 and_CC have_VH0 little_DA1 or_CC no_AT potential_NN1 for_IF non-human_JJ mortality_NN1 (_( an_AT1 air_NN1 crash_NN1 is_VBZ an_AT1 example_NN1 )_) ._. 
Others_NN2 ,_, such_II21 as_II22 a_AT1 chemical_JJ explosion_NN1 involving_VVG the_AT release_NN1 of_IO a_AT1 toxic_JJ gas_NN1 ,_, may_VM persist_VVI for_IF a_AT1 week_NNT1 or_CC more_RRR ,_, have_VH0 delayed_VVN consequences_NN2 ,_, may_VM affect_VVI a_AT1 future_JJ generation_NN1 if_CS the_AT toxic_JJ chemical_NN1 is_VBZ mutagenic_JJ and_CC may_VM have_VHI a_AT1 significant_JJ impact_NN1 on_II plant_NN1 and_CC animal_NN1 populations_NN2 ._. 
There_EX may_VM be_VBI more_DAR scope_NN1 for_IF a_AT1 GIS_NN2 approach_VV0 in_II the_AT latter_DA than_CSN in_II the_AT former_DA situations_NN2 ._. 
The_AT Hohensemer_NP1 scheme_NN1 is_VBZ not_XX ,_, however_RR ,_, beyond_II criticism._NNU for_REX21 example_REX22 ,_, can_VV0 hazards_NN2 really_RR be_VBI thought_VVN of_IO as_CSA having_VHG any_DD intentionality_NN1 ?_? 
Industrial_JJ hazards_NN2 and_CC risk_NN1 assessment_NN1 Once_RR hazards_NN2 have_VH0 been_VBN identified_VVN there_RL may_VM follow_VVI efforts_NN2 to_TO assess_VVI the_AT hazard_NN1 risk_NN1 ._. 
By_II '_GE risk_NN1 '_GE we_PPIS2 understand_VV0 '_GE the_AT likelihood_NN1 of_IO a_AT1 specified_JJ undesired_JJ event_NN1 occurring_VVG within_II a_AT1 specified_JJ period_NN1 or_CC in_II specified_JJ circumstances_NN2 '_GE (_( Health_NN1 and_CC Safety_NN1 Executive_NN1 1989_MC :_: 30_MC )_) ._. 
It_PPH1 may_VM ,_, of_RR21 course_RR22 ,_, be_VBI easier_JJR to_TO quantify_VVI risk_NN1 for_IF natural_JJ hazards_NN2 because_II21 of_II22 historical_JJ records_NN2 and_CC statistical_JJ estimation_NN1 of_IO recurrence_NN1 intervals_NN2 ._. 
Technological_JJ hazards_NN2 are_VBR more_RGR unpredictable_JJ and_CC may_VM be_VBI extremely_RR rare_JJ ,_, so_CS21 that_CS22 it_PPH1 becomes_VVZ difficult_JJ to_TO use_VVI probability_NN1 concepts_NN2 to_TO quantify_VVI risk_NN1 ,_, especially_RR those_DD2 which_DDQ rely_VV0 on_II notions_NN2 of_IO relative_JJ frequency_NN1 ._. 
The_AT identification_NN1 of_IO chemical_JJ hazards_NN2 requires_VVZ that_CST the_AT locations_NN2 of_IO sites_NN2 be_VBI known_VVN ,_, together_RL with_IW the_AT dangerous_JJ substances_NN2 stored_VVN or_CC processed_VVD there_RL ._. 
In_II the_AT UK_NP1 (_( see_VV0 Petts_NN2 1988_MC for_IF a_AT1 splendid_JJ overview_NN1 )_) the_AT Health_NN1 and_CC Safety_NN1 Executive_NN1 must_VM be_VBI notified_VVN of_IO sites_NN2 where_RRQ hazardous_JJ substances_NN2 in_II31 excess_II32 of_II33 threshold_NN1 limits_NN2 are_VBR handled_VVN ._. 
Such_DA substances_NN2 and_CC thresholds_NN2 include_VV0 ,_, for_REX21 instance_REX22 ,_, 25_MC tonnes_NNU2 of_IO flammable_JJ liquid_JJ petroleum_NN1 gas_NN1 and_CC the_AT explosive_JJ sodium_NN1 chlorate_NN1 ,_, and_CC 10_MC tonnes_NNU2 of_IO toxic_JJ chlorine_NN1 gas_NN1 ._. 
Some_DD 1600_MC installations_NN2 are_VBR notifiable_JJ ,_, along_II21 with_II22 thousands_NNO2 of_IO kilometres_NNU2 of_IO pipelines_NN2 (_( mostly_RR high-pressure_JJ natural_JJ gas_NN1 transmission_NN1 lines_NN2 )_) ._. 
Following_VVG a_AT1 number_NN1 of_IO major_JJ accidents_NN2 involving_VVG industrial_JJ activities_NN2 (_( including_II those_DD2 at_II Flixborough_NP1 ,_, Humberside_NP1 in_II 1974_MC ;_; Beek_NP1 ,_, the_AT Netherlands_NP1 in_II 1975_MC ;_; and_CC Seveso_NP1 and_CC Manfredonia_NP1 in_II Italy_NP1 in_II 1976_MC )_) ,_, the_AT EC_NP1 issued_VVD a_AT1 Directive_NN1 on_II Major_JJ Accident_NN1 Hazards_NN2 ._. 
The_AT Directive_NN1 ,_, which_DDQ is_VBZ commonly_RR referred_VVN to_II as_II the_AT '_GE Seveso_NP1 '_GE Directive_NN1 ,_, was_VBDZ issued_VVN in_II June_NPM1 1982_MC and_CC in_II the_AT UK_NP1 its_APPGE requirements_NN2 were_VBDR embodied_VVN in_II the_AT Control_NN1 of_IO Major_JJ Accident_NN1 Hazard_NN1 (_( CIMAH_NP1 )_) Regulations_NN2 (_( 1984_MC )_) ._. 
Among_II other_JJ things_NN2 ,_, these_DD2 regulations_NN2 require_VV0 operators_NN2 of_IO hazardous_JJ installations_NN2 to_TO draw_VVI up_RP on-site_JJ and_CC off-site_JJ emergency_NN1 plans_NN2 and_CC to_TO give_VVI information_NN1 to_II members_NN2 of_IO the_AT public_NN1 who_PNQS live_VV0 or_CC work_VV0 near_II such_DA installations_NN2 (_( Houston_NP1 1987_MC )_) ._. 
There_EX are_VBR presently_RR about_RG 300_MC CIMAH_NN1 sites_NN2 ._. 
However_RR ,_, many_DA2 potentially_RR hazardous_JJ chemical_JJ installations_NN2 are_VBR not_XX covered_VVN by_II such_DA regulations_NN2 ._. 
Clearly_RR ,_, a_AT1 plant_NN1 which_DDQ stores_NN2 1_MC1 tonne_NNU1 less_DAR than_CSN is_VBZ required_VVN for_IF notification_NN1 slips_VVZ through_II the_AT net_NN1 ,_, while_CS large_JJ numbers_NN2 of_IO handling_VVG sites_NN2 (_( warehouses_NN2 ,_, for_REX21 instance_REX22 )_) are_VBR not_XX covered_VVN ._. 
The_AT regulations_NN2 have_VH0 obvious_JJ implications_NN2 for_IF G1S_FO work_NN1 ,_, as_CSA illustrated_VVN below_RL ._. 
In_II Halton_NP1 Borough_NN1 ,_, Cheshire_NP1 ,_, for_REX21 instance_REX22 ,_, there_EX are_VBR 13_MC notifiable_JJ installations_NN2 ,_, with_IW over_RG 20_MC 000_MC people_NN living_VVG within_II the_AT defined_JJ '_GE consultation_NN1 zones_NN2 '_GE (_( Petts_VVZ 1988_MC )_) ._. 
It_PPH1 is_VBZ also_RR relevant_JJ to_TO point_VVI out_RP here_RL the_AT work_NN1 of_IO the_AT Health_NN1 and_CC Safety_NN1 Executive_NN1 in_II developing_JJ techniques_NN2 for_IF quantitative_JJ risk_NN1 assessment_NN1 (_( Health_NN1 and_CC Safety_NN1 Executive_NN1 1989_MC )_) ._. 
Such_DA techniques_NN2 include_VV0 the_AT use_NN1 of_IO a_AT1 '_GE risk_NN1 assessment_NN1 tool_NN1 '_GE (_( RISKAT_NP1 )_) ,_, inputs_NN2 to_II which_DDQ include_VV0 data_NN on_II wind_NN1 direction_NN1 and_CC speed_NN1 ;_; the_AT output_NN1 of_IO which_DDQ is_VBZ a_AT1 map_NN1 of_IO calculated_JJ risk_NN1 contours_NN2 ._. 
In_II societal_JJ risk_NN1 assessment_NN1 '_GE it_PPH1 is_VBZ also_RR necessary_JJ to_TO input_VVI data_NN on_II the_AT distribution_NN1 of_IO population_NN1 around_II the_AT installation_NN1 ._. 
This_DD1 can_VM be_VBI a_AT1 difficult_JJ and_CC time-consuming_JJ process_NN1 ,_, and_CC inevitably_RR some_DD simplifying_JJ assumptions_NN2 are_VBR needed_JJ '_GE (_( Health_NN1 and_CC Safety_NN1 Executive_NN1 1989_MC :_: 17_MC )_) ._. 
Again_RT ,_, the_AT G1S_FO implications_NN2 are_VBR clear_JJ and_CC are_VBR discussed_VVN below_RL ._. 
Hazardous_JJ waste_NN1 disposal_NN1 The_AT disposal_NN1 of_IO both_DB2 nuclear_JJ and_CC non-nuclear_JJ wastes_NN2 has_VHZ proved_VVN to_TO be_VBI of_IO major_JJ public_JJ concern_NN1 in_II recent_JJ years_NNT2 ._. 
The_AT rather_RG clumsy_JJ handling_NN1 by_II NIREX_NN1 of_IO the_AT search_NN1 for_IF sites_NN2 to_TO store_VVI low-_JJ and_CC intermediate-level_JJ wastes_NN2 ,_, together_RL with_IW very_RG recent_JJ worries_NN2 about_II the_AT import_NN1 of_IO toxic_JJ wastes_NN2 such_II21 as_II22 polychlorinated_JJ biphenyls_NN2 (_( PCBs_NN2 )_) ,_, are_VBR but_CCB two_MC examples_NN2 of_IO such_DA concern_NN1 in_II the_AT UK_NP1 ._. 
However_RR ,_, while_CS the_AT former_DA has_VHZ been_VBN given_VVN considerable_JJ attention_NN1 by_II geographers_NN2 (_( summarized_VVN by_II Openshaw_NP1 et_RA21 al_RA22 ._. 
1989_MC )_) there_EX has_VHZ been_VBN relatively_RR little_RR such_DA work_NN1 in_II the_AT UK_NP1 on_II the_AT disposal_NN1 of_IO non-nuclear_JJ wastes_NN2 ._. 
This_DD1 contrasts_VVZ with_IW the_AT situation_NN1 in_II the_AT USA_NP1 (_( Greenberg_NP1 and_CC Anderson_NP1 1984_MC )_) ,_, where_CS the_AT clean-up_NN1 of_IO abandoned_JJ sites_NN2 is_VBZ a_AT1 priority_NN1 ._. 
Certainly_RR ,_, in_II31 terms_II32 of_II33 the_AT volume_NN1 of_IO waste_NN1 generated_VVN in_II the_AT UK_NP1 non-nuclear_JJ hazardous_JJ industrial_JJ wastes_NN2 should_VM be_VBI high_JJ on_II the_AT environmental_JJ agenda_NN1 ;_; while_CS less_DAR than_CSN 1_MC1 million_NNO tonnes_NNU2 of_IO all_DB nuclear_JJ wastes_NN2 were_VBDR generated_VVN in_II 198586_MC ,_, 10_MC times_NNT2 that_DD1 amount_NN1 were_VBDR generated_VVN of_IO hazardous_JJ wastes_NN2 (_( Openshaw_NP1 et_RA21 al_RA22 ._. 
1989_MC :_: 13_MC )_) ._. 
Attention_NN1 should_VM be_VBI drawn_VVN to_II the_AT three_MC Annual_JJ Reports_NN2 of_IO the_AT Hazardous_JJ Waste_NN1 Inspectorate_NN1 (_( now_RT merged_VVN with_IW other_JJ branches_NN2 of_IO the_AT DoE_NP1 to_TO form_VVI Her_APPGE Majesty_NN1 's_GE Inspectorate_NN1 of_IO Pollution_NN1 )_) ._. 
These_DD2 contain_VV0 some_DD valuable_JJ material_NN1 on_II disposal_NN1 practices_NN2 ,_, and_CC volumes_NN2 of_IO waste_NN1 generated_VVN and_CC disposed_JJ of_IO by_II each_DD1 waste_NN1 regulation_NN1 authority_NN1 (_( WRA_NP1 )_) counties_NN2 in_II England_NP1 ,_, districts_NN2 in_II Wales_NP1 ,_, Scotland_NP1 and_CC Northern_NP1 Ireland_NP1 ._. 
Each_DD1 WRA_NN1 is_VBZ required_VVN to_TO produce_VVI a_AT1 waste_NN1 disposal_NN1 plan_NN1 and_CC these_DD2 are_VBR starting_VVG to_TO appear_VVI ._. 
They_PPHS2 too_RR contain_VV0 useful_JJ material_NN1 on_RP ,_, for_REX21 instance_REX22 ,_, the_AT locations_NN2 of_IO sites_NN2 ,_, though_CS individual_JJ site_NN1 licences_NN2 must_VM be_VBI consulted_VVN for_IF details_NN2 of_IO what_DDQ each_DD1 is_VBZ permitted_VVN to_TO take_VVI in_II the_AT way_NN1 of_IO hazardous_JJ wastes_NN2 ._. 
What_DDQ is_VBZ not_XX known_VVN is_VBZ information_NN1 on_II the_AT locations_NN2 of_IO former_DA sites_NN2 ,_, at_II which_DDQ quite_RG toxic_JJ materials_NN2 may_VM have_VHI been_VBN dumped_VVN ,_, though_CS recent_JJ work_NN1 by_II Egger_NP1 (_( 1989_MC )_) in_II Austria_NP1 has_VHZ mapped_VVN a_AT1 vast_JJ number_NN1 of_IO such_DA sites_NN2 ._. 
It_PPH1 is_VBZ these_DD2 ,_, rather_II21 than_II22 existing_JJ ,_, licensed_JJ sites_NN2 which_DDQ should_VM be_VBI a_AT1 focus_NN1 of_IO research_NN1 activity_NN1 ,_, particularly_RR when_CS examining_VVG possible_JJ links_NN2 to_II ill_JJ health_NN1 ._. 
We_PPIS2 simply_RR do_VD0 not_XX have_VHI enough_DD evidence_NN1 yet_RR to_TO back_VVI up_RP the_AT claim_NN1 that_DD1 '_VBZ people_NN are_VBR almost_RR certainly_RR ill_JJ ,_, dead_JJ ,_, or_CC dying_VVG because_II21 of_II22 these_DD2 sloppy_JJ waste_NN1 disposal_NN1 activities_NN2 '_GE (_( Openshaw_NP1 et_RA21 al_RA22 ._. 
1989_MC :_: 12_MC )_) ._. 
Research_NN1 is_VBZ in_II progress_NN1 ,_, however_RR ,_, to_TO assess_VVI possible_JJ health_NN1 effects_NN2 of_IO poor_JJ incineration_NN1 (_( Diggle_NP1 et_RA21 al_RA22 ._. 
1990_MC )_) ._. 
Particular_JJ attention_NN1 should_VM be_VBI focused_VVN on_II hospital_NN1 incinerators_NN2 ,_, since_CS these_DD2 have_VH0 been_VBN subject_II21 to_II22 Crown_NN1 immunity_NN1 for_IF many_DA2 years_NNT2 and_CC most_DAT are_VBR technically_RR ill-equipped_JJ to_TO burn_VVI at_II sufficiently_RR high_JJ temperatures_NN2 for_IF adequate_JJ disposal_NN1 of_IO wastes_NN2 (_( Gatrell_NP1 and_CC Lovett_NP1 1991_MC )_) ._. 
We_PPIS2 need_VV0 to_TO produce_VVI an_AT1 inventory_NN1 of_IO all_DB such_DA sites_NN2 ,_, together_RL with_IW a_AT1 publicly_RR available_JJ ,_, comprehensive_JJ database_NN1 on_II all_DB sites_NN2 accepting_VVG hazardous_JJ industrial_JJ wastes_NN2 ._. 
Emergency_NN1 planning_VVG It_PPH1 is_VBZ convenient_JJ to_TO retain_VVI the_AT nuclear/non-nuclear_JJ distinction_NN1 when_CS discussing_VVG the_AT background_NN1 to_II emergency_NN1 planning_NN1 ._. 
Nuclear_JJ emergency_NN1 planning_VVG in_II the_AT UK_NP1 has_VHZ been_VBN reviewed_VVN by_II Matthews_NP1 and_CC Pepper_NN1 (_( 1981_MC )_) ._. 
Nuclear_JJ power_NN1 stations_NN2 must_VM prepare_VVI a_AT1 plan_NN1 and_CC submit_VVI this_DD1 to_II the_AT Nuclear_JJ Installations_NN2 Inspectorate_NN1 and_CC Health_NN1 and_CC Safety_NN1 Executive_NN1 for_IF approval_NN1 ._. 
Evacuation_NN1 plans_NN2 are_VBR required_VVN to_TO cover_VVI an_AT1 area_NN1 within_II a_AT1 distance_NN1 of_IO 2.4_MC km_NNU ;_; these_DD2 detail_VV0 the_AT roles_NN2 to_TO be_VBI played_VVN by_II the_AT emergency_NN1 services_NN2 ._. 
The_AT very_RG narrow_JJ emergency_NN1 planning_NN1 zones_NN2 contrast_VV0 with_IW those_DD2 in_II the_AT USA_NP1 (_( Collins_NP1 1981_MC )_) ,_, where_CS planning_NN1 is_VBZ required_VVN up_RG21 to_RG22 16_MC km_NNU from_II the_AT site_NN1 to_TO cover_VVI exposure_NN1 to_II radiation_NN1 plumes_NN2 ._. 
Local_JJ authorities_NN2 in_II the_AT UK_NP1 that_CST have_VH0 declared_VVN themselves_58 '_GE nuclear-free_JJ '_GE have_VH0 been_VBN highly_RR critical_JJ of_IO British_JJ planning_NN1 ,_, calling_VVG for_IF an_AT1 extension_NN1 of_IO emergency_NN1 zones_NN2 to_II at_RR21 least_RR22 24_MC km_NNU ._. 
Zeigler_NP1 et_RA21 al_RA22 ._. 
(_( 1983_MC )_) should_VM be_VBI consulted_VVN for_IF material_NN1 relating_VVG to_II evacuation_NN1 behaviour_NN1 in_II the_AT wake_NN1 of_IO the_AT incident_NN1 at_II Three_NP1 Mile_NP1 Island_NNL1 ._. 
In_II the_AT UK_NP1 little_JJ work_NN1 of_IO any_DD description_NN1 seems_VVZ to_TO have_VHI been_VBN done_VDN on_II human_JJ behaviour_NN1 in_II the_AT aftermath_NN1 of_IO releases_NN2 of_IO hazardous_JJ substances_NN2 ,_, nor_CC is_VBZ much_RR available_JJ on_II public_JJ attitudes_NN2 to_II emergency_NN1 planning_NN1 ._. 
Indeed_RR ,_, no_AT major_JJ studies_NN2 of_IO evacuation_NN1 due_II21 to_II22 hazards_NN2 have_VH0 been_VBN undertaken_VVN in_II recent_JJ times_NNT2 and_CC this_DD1 is_VBZ a_AT1 widely_RR recognized_VVN deficiency_NN1 which_DDQ no_AT research_NN1 funding_NN1 body_NN1 seems_VVZ willing_JJ to_TO rectify_VVI ._. 
In_II the_AT USA_NP1 on_II the_AT other_JJ hand_NN1 ,_, a_AT1 good_JJ deal_NN1 of_IO useful_JJ research_NN1 in_II these_DD2 areas_NN2 has_VHZ been_VBN undertaken_VVN ._. 
This_DD1 research_NN1 is_VBZ brought_VVN together_RL in_II very_RG useful_JJ reports_NN2 produced_VVN by_II the_AT Oak_NN1 Ridge_NN1 National_JJ Laboratory_NN1 ,_, Tennessee_NP1 (_( Sorenson_NP1 et_RA21 al_RA22 ._. 
1987_MC ;_; Vogt_NP1 and_CC Sorenson_NP1 1987_MC )_) and_CC the_AT Disaster_NN1 Research_NN1 Centre_NN1 ,_, University_NN1 of_IO Delaware_NP1 ._. 
Perry_NP1 (_( 1985_MC )_) has_VHZ written_VVN an_AT1 interesting_JJ book_NN1 on_II emergency_NN1 planning_NN1 and_CC evacuation_NN1 ,_, covering_VVG both_RR natural_JJ and_CC technological_JJ hazards_NN2 ._. 
For_IF him_PPHO1 ,_, comprehensive_JJ emergency_NN1 planning_NN1 and_CC management_NN1 involve_VV0 :_: prevention_NN1 ,_, protection_NN1 ,_, response_NN1 and_CC restoration_NN1 ._. 
By_II prevention_NN1 we_PPIS2 understand_VV0 the_AT elimination_NN1 or_CC reduction_NN1 in_II risk_NN1 ,_, perhaps_RR by_II adequate_JJ zoning_NN1 or_CC safe_JJ routeing_NN1 of_IO shipments_NN2 ,_, for_REX21 instance_REX22 ._. 
Included_VVN in_II protection_NN1 are_VBR methods_NN2 for_IF detecting_VVG likely_JJ hazard_NN1 events_NN2 and_CC in_II warning_VVG those_DD2 likely_JJ to_TO be_VBI affected_VVN ._. 
Response_NN1 includes_VVZ search_NN1 and_CC rescue_NN1 (_( including_II the_AT need_NN1 to_TO cope_VVI with_IW secondary_JJ threats_NN2 such_II21 as_II22 contamination_NN1 of_IO water_NN1 supplies_VVZ after_II an_AT1 explosion_NN1 ,_, or_CC fire_NN1 following_VVG an_AT1 earthquake_NN1 )_) ._. 
As_CSA Perry_NP1 and_CC others_NN2 have_VH0 noted_VVN ,_, relatively_RR few_DA2 resources_NN2 are_VBR devoted_JJ to_II the_AT first_MD two_MC aspects_NN2 ,_, the_AT bulk_NN1 going_VVG to_II response_NN1 and_CC restoration_NN1 (_( rebuilding_NN1 )_) ._. 
It_PPH1 should_VM be_VBI clear_JJ from_II this_DD1 brief_JJ description_NN1 that_CST these_DD2 issues_NN2 are_VBR inherently_RR geographical_JJ ._. 
In_II one_MC1 of_IO the_AT few_DA2 studies_NN2 of_IO evacuation_NN1 behaviour_NN1 after_II a_AT1 non-nuclear_JJ hazard_NN1 event_NN1 Liverman_NN1 and_CC Wilson_NP1 (_( 1981_MC )_) give_VV0 weight_NN1 to_II this_DD1 argument_NN1 ._. 
In_II 1979_MC a_AT1 train_NN1 carrying_VVG chlorine_NN1 ,_, liquid_JJ petroleum_NN1 ,_, toluene_NN1 and_CC propane_NN1 was_VBDZ derailed_VVN in_II Mississauga_NP1 ,_, Ontario_NP1 ,_, causing_VVG a_AT1 series_NN of_IO explosions_NN2 ._. 
This_DD1 led_VVD to_II the_AT evacuation_NN1 of_IO 250_MC 000_MC people_NN which_DDQ ,_, although_CS staggered_VVN over_II several_DA2 days_NNT2 ,_, led_VVN to_II great_JJ pressure_NN1 on_II transport_NN1 arteries_NN2 ._. 
An_AT1 important_JJ finding_NN1 was_VBDZ that_CST many_DA2 relied_VVN on_II their_APPGE own_DA transport_NN1 and_CC went_VVN to_TO stay_VVI with_IW friends_NN2 and_CC relatives_NN2 some_DD distance_NN1 away_RL ,_, rather_II21 than_II22 relying_VVG on_II the_AT 19_MC official_JJ evacuation_NN1 reception_NN1 centres_NN2 ._. 
Despite_II this_DD1 ,_, the_AT evacuation_NN1 was_VBDZ successful_JJ ,_, partly_RR because_CS the_AT accident_NN1 occurred_VVD at_II a_AT1 weekend_NNT1 and_CC the_AT immediately_RR affected_VVN area_NN1 was_VBDZ sparsely_RR populated_VVN ._. 
In_RR21 addition_RR22 ,_, the_AT police_NN2 were_VBDR well_RR prepared_VVN ,_, stimulated_VVD largely_RR by_II implementing_VVG a_AT1 plan_NN1 designed_VVN more_RRR to_TO cope_VVI with_IW an_AT1 emergency_NN1 at_II the_AT nearby_JJ Toronto_NP1 International_JJ Airport_NN1 than_CSN for_IF the_AT event_NN1 in_II question_NN1 ._. 
It_PPH1 is_VBZ not_XX without_IW interest_NN1 to_TO note_VVI that_CST in_II the_AT USA_NP1 there_EX is_VBZ a_AT1 Federal_JJ Emergency_NN1 Management_NN1 Agency_NN1 (_( FEMA_NP1 )_) ,_, established_VVN in_II 1978_MC ,_, charged_VVN with_IW dealing_VVG with_IW all_DB types_NN2 of_IO emergency_NN1 planning_NN1 and_CC management_NN1 ._. 
In_II the_AT UK_NP1 there_EX is_VBZ no_AT equivalent_JJ ,_, though_CS the_AT spate_NN1 of_IO disasters_NN2 in_II the_AT past_JJ 4_MC years_NNT2 (_( involving_VVG over_RG 1000_MC lives_NN2 lost_VVN )_) has_VHZ led_VVN to_II several_DA2 calls_NN2 for_IF such_DA a_AT1 national_JJ body_NN1 to_TO be_VBI established_VVN ._. 
Natural_JJ hazards_NN2 In_II this_DD1 short_JJ space_NN1 it_PPH1 is_VBZ not_XX possible_JJ to_TO go_VVI into_II any_DD depth_NN1 on_II the_AT nature_NN1 of_IO particular_JJ natural_JJ hazards_NN2 but_CCB only_RR to_II highlight_NN1 those_DD2 aspects_NN2 relevant_JJ to_II GIS_NN2 ._. 
Needless_JJ to_TO say_VVI ,_, the_AT study_NN1 of_IO natural_JJ hazards_NN2 is_VBZ important_JJ ,_, a_AT1 statement_NN1 brought_VVN into_II sharp_JJ focus_NN1 by_II the_AT fact_NN1 that_CST natural_JJ hazards_NN2 account_VV0 for_IF up_RG21 to_RG22 4_MC per_NNU21 cent_NNU22 of_IO total_JJ deaths_NN2 in_II the_AT world_NN1 each_DD1 year_NNT1 (_( Mitchell_NP1 ,_, 1974_MC )_) ._. 
In_II 1970_MC ,_, for_REX21 example_REX22 ,_, more_DAR than_CSN 200_MC 000_MC people_NN died_VVD in_II the_AT cyclone_NN1 and_CC flooding_NN1 of_IO Bangladesh_NP1 and_CC in_II 1979_MC the_AT hurricanes_NN2 David_NP1 and_CC Frederick_NP1 caused_VVD more_DAR than_CSN $3bn_NNU damage_NN1 in_II the_AT USA_NP1 ._. 
Shah_NP1 (_( 1983_MC )_) made_VVD a_AT1 comprehensive_JJ investigation_NN1 of_IO natural_JJ disaster_NN1 reports_NN2 for_IF the_AT period_NN1 197480_MC and_CC concluded_VVD that_CST both_DB2 disaster_NN1 frequency_NN1 and_CC magnitude_NN1 (_( measured_VVN in_II31 terms_II32 of_II33 lives_NN2 lost_VVN )_) are_VBR increasing_VVG ._. 
High_JJ death_NN1 tolls_NN2 are_VBR still_RR a_AT1 characteristic_NN1 of_IO less_RGR developed_JJ ,_, low-income_JJ countries_NN2 ._. 
Surprisingly_RR ,_, fog_NN1 is_VBZ the_AT third_MD most_RGT important_JJ hazard_NN1 in_II Europe_NP1 after_II floods_NN2 and_CC volcanoes_NN2 ._. 
Research_NN1 on_II natural_JJ hazards_NN2 has_VHZ a_AT1 long_JJ tradition_NN1 in_II geography_NN1 going_VVG back_RP more_DAR than_CSN half_DB a_AT1 century_NNT1 ._. 
Much_DA1 of_IO the_AT impetus_NN1 for_IF such_DA research_NN1 arose_VVD in_II the_AT USA_NP1 from_II observations_NN2 on_II river_NN1 basin_NN1 management_NN1 and_CC the_AT need_NN1 to_TO reduce_VVI flood_NN1 damage_NN1 ._. 
The_AT vast_JJ majority_NN1 of_IO natural_JJ hazards_NN2 studies_NN2 have_VH0 ignored_VVN the_AT physical_JJ nature_NN1 of_IO the_AT events_NN2 per_RR21 se_RR22 and_CC have_VH0 concentrated_VVN on_II behavioural_JJ issues_NN2 such_II21 as_II22 the_AT perception_NN1 and_CC estimation_NN1 of_IO hazards_NN2 ._. 
This_DD1 emphasis_NN1 is_VBZ partly_RR due_II21 to_II22 the_AT background_NN1 of_IO the_AT researchers_NN2 involved_VVN (_( many_DA2 are_VBR human_JJ geographers_NN2 and_CC sociologists_NN2 )_) and_CC partly_RR because_CS it_PPH1 is_VBZ only_RR really_RR since_CS the_AT Second_MD World_NN1 War_NN1 that_CST major_JJ hazards_NN2 have_VH0 been_VBN monitored_VVN systematically_RR ,_, culminating_VVG ,_, of_RR21 course_RR22 ,_, in_II the_AT last_MD decade_NNT1 or_CC so_RR in_II the_AT use_NN1 of_IO satellite_NN1 technology_NN1 ._. 
Brief_JJ reference_NN1 to_II two_MC examples_NN2 of_IO natural_JJ hazards_NN2 may_VM illustrate_VVI the_AT potential_JJ relevance_NN1 of_IO a_AT1 GIS_NN2 approach_NN1 ,_, applications_NN2 of_IO which_DDQ are_VBR considered_VVN below_RL ._. 
Jones_NP1 et_RA21 al_RA22 ._. 
(_( 1989_MC )_) have_VH0 examined_VVN the_AT distribution_NN1 of_IO some_DD 8500_MC landslides_NN2 in_II the_AT UK_NP1 ._. 
Their_APPGE maps_NN2 show_VV0 clearly_RR an_AT1 increased_JJ propensity_NN1 for_IF landsliding_VVG in_II the_AT north_ND1 and_CC west_ND1 of_IO the_AT UK_NP1 ._. 
More_RRR to_II the_AT point_NN1 ,_, however_RR ,_, is_VBZ the_AT fact_NN1 that_CST this_DD1 distribution_NN1 can_VM be_VBI modelled_VVN in_II31 terms_II32 of_II33 slopes_NN2 ,_, rainfall_NN1 ,_, rock_NN1 types_NN2 and_RR31 so_RR32 on_RR33 ._. 
In_II other_JJ words_NN2 the_AT actual_JJ hazard_NN1 distribution_NN1 can_VM be_VBI understood_VVN and_CC areas_NN2 at_II risk_NN1 identified_VVN ._. 
The_AT potential_JJ benefits_NN2 to_II insurance_NN1 companies_NN2 ,_, engineers_NN2 and_CC planners_NN2 are_VBR clear_JJ ._. 
Second_MD ,_, a_AT1 well-known_JJ winter_NNT1 hazard_NN1 in_II the_AT UK_NP1 ,_, with_IW marked_JJ spatial_JJ localization_NN1 ,_, is_VBZ fog_NN1 ._. 
Predictive_JJ models_NN2 of_IO its_APPGE distribution_NN1 have_VH0 clear_JJ implications_NN2 for_IF road_NN1 and_CC air_NN1 safety_NN1 ._. 
Perry_NP1 (_( 1981_MC )_) describes_VVZ a_AT1 fog_NN1 potential_NN1 index_NN1 that_CST is_VBZ a_AT1 function_NN1 of_IO local_JJ topography_NN1 ,_, environmental_JJ features_NN2 affecting_VVG fog_NN1 formation_NN1 and_CC distance_NN1 from_II standing_VVG water_NN1 bodies_NN2 ._. 
Reliance_NN1 is_VBZ currently_RR placed_VVN on_II a_AT1 network_NN1 of_IO monitoring_VVG stations_NN2 ,_, often_RR sparsely_RR located_VVN and_CC frequently_RR outside_JJ fog-prone_JJ areas_NN2 ;_; GIS_NN2 could_VM implement_VVI a_AT1 predictive_JJ model_NN1 based_VVN on_II digitized_JJ map_NN1 features_NN2 and_CC indeed_RR the_AT model_NN1 could_VM also_RR be_VBI used_VVN in_II site_NN1 allocation_NN1 studies_NN2 for_IF airports_NN2 and_CC the_AT like_JJ ._. 
The_AT majority_NN1 of_IO natural_JJ hazard_NN1 research_NN1 has_VHZ focused_VVN on_II single_JJ hazards_NN2 but_CCB of_RR21 course_RR22 any_DD area_NN1 may_VM be_VBI prone_JJ to_II more_DAR than_CSN one_MC1 hazard_NN1 type_NN1 ._. 
This_DD1 idea_NN1 led_VVD Hewitt_NP1 and_CC Burton_NP1 (_( 1971_MC )_) to_TO develop_VVI the_AT concept_NN1 of_IO '_GE all-hazards-at-a-place_NN1 '_GE ._. 
To_TO achieve_VVI this_DD1 they_PPHS2 concentrated_VVD on_II the_AT whole_JJ spectrum_NN1 of_IO damaging_JJ events_NN2 in_II an_AT1 area_NN1 and_CC explored_VVD their_APPGE aggregate_JJ impact_NN1 ._. 
The_AT implications_NN2 for_IF GIS_NN2 in_II the_AT development_NN1 of_IO '_GE hazardousness_NN1 '_GE indices_NN2 are_VBR obvious_JJ and_CC if_CS time-series_JJ data_NN are_VBR available_JJ there_EX is_VBZ the_AT potential_JJ to_TO produce_VVI a_AT1 variety_NN1 of_IO probability_NN1 maps_NN2 ._. 
The_AT collection_NN1 of_IO historical_JJ data_NN on_II natural_JJ hazards_NN2 is_VBZ important_JJ since_CS it_PPH1 is_VBZ clear_JJ that_CST their_APPGE spatial_JJ pattern_NN1 varies_VVZ through_II time_NNT1 ._. 
Many_DA2 areas_NN2 that_CST appear_VV0 to_TO be_VBI hazard-free_JJ on_II current_JJ maps_NN2 may_VM merely_RR by_II passing_VVG through_II a_AT1 temporary_JJ period_NN1 of_IO quiescence_NN1 ._. 
This_DD1 is_VBZ perhaps_RR best_RRT illustrated_VVN with_II31 reference_II32 to_II33 earthquakes_NN2 ._. 
For_REX21 example_REX22 ,_, in_II 1692_MC ,_, the_AT whole_NN1 of_IO southern_JJ England_NP1 was_VBDZ affected_VVN by_II an_AT1 earthquake_NN1 whose_DDQGE epicentre_NN1 was_VBDZ located_VVN in_II what_DDQ is_VBZ now_RT Belgium_NP1 ;_; a_AT1 good_JJ deal_NN1 of_IO minor_JJ structural_JJ damage_NN1 was_VBDZ inflicted_VVN (_( Morse_NP1 1983_MC )_) ._. 
Presumably_RR the_AT builders_NN2 of_IO the_AT Channel_NN1 Tunnel_NN1 are_VBR aware_JJ of_IO this_DD1 ?_? 
Environmental_JJ epidemiology_NN1 This_DD1 is_VBZ ,_, of_RR21 course_RR22 ,_, a_AT1 vast_JJ subject_NN1 area_NN1 ,_, embracing_VVG the_AT impact_NN1 of_IO the_AT natural_JJ and_CC technological_JJ environments_NN2 on_II human_JJ health_NN1 ._. 
It_PPH1 covers_VVZ ,_, for_REX21 instance_REX22 ,_, the_AT associations_NN2 between_II water_NN1 and_CC air_NN1 pollution_NN1 and_CC ill_JJ health_NN1 ,_, links_VVZ between_II geological_JJ and_CC pedological_JJ environments_NN2 and_CC ill_JJ health_NN1 ,_, as_II31 well_II32 as_II33 associations_NN2 between_II specific_JJ technological_JJ hazards_NN2 (_( e.g._REX high-voltage_JJ power_NN1 lines_NN2 )_) and_CC possible_JJ health_NN1 effects_NN2 ._. 
We_PPIS2 can_VM only_RR consider_VVI a_AT1 small_JJ fraction_NN1 of_IO this_DD1 work_NN1 here_RL ;_; little_DA1 consideration_NN1 is_VBZ given_VVN to_II occupational_JJ environments_NN2 ,_, for_REX21 instance_REX22 ._. 
In_II the_AT UK_NP1 such_DA research_NN1 is_VBZ widespread_JJ though_CS there_EX are_VBR specialist_JJ environmental_JJ epidemiology_NN1 research_NN1 institutes_NN2 at_II Southampton_NP1 and_CC Cardiff_NP1 (_( funded_VVN by_II the_AT Medical_JJ Research_NN1 Council_NN1 )_) ,_, and_CC others_NN2 at_II the_AT University_NN1 of_IO Surrey_NP1 (_( Robens_NP2 Institute_NN1 )_) and_CC at_II the_AT London_NP1 School_NN1 for_IF Hygiene_NN1 and_CC Tropical_JJ Medicine_NN1 ._. 
The_AT latter_DA houses_NN2 the_AT Small_JJ Area_NN1 Health_NN1 Statistics_NN Research_NN1 Unit_NN1 (_( SAHSRU_NP1 )_) ,_, which_DDQ has_VHZ a_AT1 specific_JJ remit_NN1 to_TO conduct_VVI research_NN1 into_II morbidity_NN1 and_CC mortality_NN1 at_II local_JJ scales_NN2 (_( Elliot_NP1 1988_MC )_) ._. 
The_AT SAHSRU_NN1 liaises_VVZ closely_RR with_IW OPCS_NP1 ._. 
The_AT kinds_NN2 of_IO databases_NN2 used_VVN by_II these_DD2 groups_NN2 are_VBR discussed_VVN below_RL ._. 
Greenberg_NP1 (_( 1983_MC )_) and_CC a_AT1 collection_NN1 of_IO papers_NN2 in_II Greenberg_NP1 (_( 1987_MC )_) review_VV0 some_DD of_IO the_AT literature_NN1 on_II environmental_JJ epidemiology_NN1 ._. 
It_PPH1 seems_VVZ worth_II stating_VVG at_II the_AT outset_NN1 that_CST there_EX are_VBR two_MC ways_NN2 geographical_JJ research_NN1 in_II this_DD1 general_JJ area_NN1 can_VM proceed_VVI ._. 
One_PN1 is_VBZ to_TO collect_VVI morbidity_NN1 and_CC mortality_NN1 data_NN (_( usually_RR the_AT latter_DA is_VBZ more_RGR widely_RR available_JJ ,_, though_CS the_AT former_DA more_RGR useful_JJ )_) and_CC to_TO examine_VVI spatial_JJ distributions_NN2 for_IF clustering_NN1 ._. 
This_DD1 has_VHZ itself_PPX1 generated_VVN a_AT1 substantial_JJ literature_NN1 (_( see_VV0 the_AT papers_NN2 in_II Elliot_NP1 1988_MC for_IF an_AT1 overview_NN1 )_) and_CC is_VBZ nowhere_RL better_RRR exemplified_VVN than_CSN in_II Openshaw_NP1 's_GE geographical_JJ analysis_NN1 machine_NN1 (_( see_VV0 Ch._NN1 2_MC )_) ._. 
We_PPIS2 then_RT search_VV0 retrospectively_RR for_IF evidence_NN1 of_IO possible_JJ environmental_JJ associations_NN2 ._. 
An_AT1 alternative_JJ approach_NN1 is_VBZ to_TO begin_VVI by_II hypothesizing_VVG an_AT1 environmental_JJ determinant_NN1 of_IO ill_JJ health_NN1 and_CC to_TO collect_VVI data_NN to_TO test_VVI an_AT1 explicit_JJ hypothesis_NN1 ._. 
If_CS an_AT1 association_NN1 is_VBZ suggested_VVN then_RT we_PPIS2 would_VM want_VVI to_TO collect_VVI data_NN in_II other_JJ areas_NN2 to_TO assess_VVI the_AT hypothesis_NN1 further_RRR ._. 
Of_RR21 course_RR22 ,_, these_DD2 two_MC approaches_NN2 are_VBR not_XX clear-cut_JJ in_II practice_NN1 ;_; in_II many_DA2 cases_NN2 the_AT hypotheses_NN2 of_IO interest_NN1 will_VM have_VHI themselves_PPX2 been_VBN generated_VVN by_II inductive_JJ mapping_NN1 ._. 
One_MC1 example_NN1 of_IO an_AT1 hypothesis_NN1 which_DDQ has_VHZ been_VBN given_VVN serious_JJ attention_NN1 in_II recent_JJ years_NNT2 is_VBZ the_AT possible_JJ link_NN1 between_II electromagnetic_JJ fields_NN2 generated_VVN by_II high-voltage_JJ power_NN1 lines_NN2 or_CC supply_VV0 cables_NN2 and_CC ill_JJ health_NN1 ._. 
Some_DD work_NN1 (_( Perry_NP1 et_RA21 al_RA22 ._. 
1981_MC )_) suggested_VVD associations_NN2 with_IW mental_JJ illness_NN1 and_CC suicide_NN1 ,_, while_CS others_NN2 (_( e.g._REX Wertheimer_NP1 and_CC Leeper_NP1 1982_MC )_) have_VH0 examined_VVN links_NN2 to_II various_JJ cancers_NN2 ,_, including_II leukaemias_NN2 ._. 
A_AT1 good_JJ bibliography_NN1 is_VBZ provided_VVN in_II Perry_NP1 and_CC Pearl_NP1 (_( 1988_MC )_) ._. 
The_AT Central_JJ Electricity_NN1 Generating_JJ Board_NN1 (_( now_RT privatized_VVN and_CC split_VVN into_II two_MC companies_NN2 )_) is_VBZ funding_VVG a_AT1 large-scale_JJ ,_, 2-year_JJ project_NN1 to_TO assess_VVI links_NN2 to_II childhood_NN1 cancer_NN1 ,_, while_CS another_DD1 recently_RR announced_VVN study_NN1 aims_VVZ to_TO look_VVI at_II links_NN2 to_II sudden_JJ infant_NN1 death_NN1 ._. 
One_MC1 obvious_JJ problem_NN1 ,_, with_IW implications_NN2 for_IF any_DD GIS_NN2 input_VV0 ,_, is_VBZ that_DD1 the_AT effects_NN2 are_VBR highly_RR localized_VVN ,_, with_IW little_DA1 measurable_JJ field_NN1 effect_NN1 more_RRR than_CSN a_AT1 few_DA2 dozen_NNO metres_NNU2 away_II21 from_II22 the_AT route_NN1 centre-lines_NN2 ._. 
Accuracy_NN1 of_IO digital_JJ data_NN collection_NN1 (_( either_RR of_IO the_AT power_NN1 lines_NN2 or_CC residential_JJ locations_NN2 )_) is_VBZ paramount_JJ ._. 
There_EX seems_VVZ to_TO be_VBI a_AT1 growing_JJ interest_NN1 in_II examining_VVG associations_NN2 between_II ill_JJ health_NN1 and_CC the_AT geological_JJ environment_NN1 ._. 
GIS_NN2 work_VV0 in_II this_DD1 field_NN1 (_( e.g._REX Matthews_NP1 1989_MC )_) is_VBZ reviewed_VVN later_RRR ,_, but_CCB we_PPIS2 note_VV0 here_RL the_AT work_NN1 of_IO the_AT applied_JJ geochemistry_NN1 group_NN1 at_II Imperial_JJ College_NN1 (_( responsible_JJ for_IF a_AT1 series_NN of_IO atlases_NN2 on_II regional_JJ geochemistry_NN1 )_) and_CC the_AT existence_NN1 of_IO a_AT1 Society_NN1 for_IF Environmental_JJ Geochemistry_NN1 and_CC Health_NN1 ._. 
One_MC1 area_NN1 of_IO recent_JJ debate_NN1 concerns_VVZ radon_NN1 gas_NN1 and_CC its_APPGE links_NN2 to_II lung_NN1 cancer_NN1 ._. 
This_DD1 is_VBZ taken_VVN very_RG seriously_RR in_II the_AT USA_NP1 ,_, where_CS the_AT Environmental_JJ Protection_NN1 Agency_NN1 estimates_VVZ that_CST 10_MC 00020_MC 000_MC deaths_NN2 from_II lung_NN1 cancer_NN1 each_DD1 year_NNT1 are_VBR due_II21 to_II22 radon_NN1 gas_NN1 exposure_NN1 ._. 
Estimates_NN2 from_II Scandinavia_NP1 suggest_VV0 that_CST 1030_MC per_NNU21 cent_NNU22 of_IO such_DA deaths_NN2 are_VBR caused_VVN by_II exposure_NN1 to_II radon_NN1 (_( Richardson_NP1 1988_MC :_: 27081_MC )_) ._. 
The_AT gas_NN1 is_VBZ trapped_VVN in_II well-insulated_JJ houses_NN2 and_CC is_VBZ produced_VVN by_II the_AT decay_NN1 of_IO uranium_NN1 in_II rocks_NN2 and_CC soil_NN1 ;_; thus_RR it_PPH1 has_VHZ been_VBN measured_VVN at_II high_JJ concentrations_NN2 in_II parts_NN2 of_IO Cornwall_NP1 and_CC Devon_NP1 ._. 
There_EX are_VBR great_JJ difficulties_NN2 in_II assessing_VVG health_NN1 effects_NN2 ,_, however_RR ,_, since_CS exposure_NN1 varies_VVZ greatly_RR from_II house_NN1 to_II house_NN1 within_II small_JJ neighbourhoods_NN2 ._. 
Carefully_RR designed_VVN case-control_JJ studies_NN2 are_VBR therefore_RR required_VVN to_II measure_NN1 the_AT separate_JJ effect_NN1 of_IO radon_NN1 exposure_NN1 after_II controlling_VVG for_IF smoking_JJ behaviour_NN1 and_CC other_JJ risk_NN1 factors_NN2 ._. 
Other_JJ health_NN1 concerns_NN2 which_DDQ might_VM merit_VVI attention_NN1 by_II those_DD2 involved_JJ in_II the_AT Regional_JJ Research_NN1 Laboratory_NN1 (_( RRL_NP1 )_) initiative_NN1 include_VV0 air_NN1 and_CC water_NN1 pollution_NN1 ._. 
The_AT debate_NN1 about_II lead_NN1 pollution_NN1 from_II vehicle_NN1 exhaust_NN1 emissions_NN2 is_VBZ well_RR known_VVN ;_; less_DAR publicity_NN1 has_VHZ been_VBN given_VVN to_II PAHs_NN2 (_( polychlorinated_JJ aromatic_JJ hydrocarbons_NN2 )_) ,_, a_AT1 ubiquitous_JJ product_NN1 of_IO combustion_NN1 processes_NN2 and_CC known_VVN to_TO be_VBI carcinogenic_JJ (_( Lioy_NP1 and_CC Daisey_NP1 1987_MC )_) ._. 
Research_NN1 at_II the_AT North_ND1 West_ND1 Regional_JJ Research_NN1 Laboratory_NN1 (_( NWRRL_NP1 )_) aims_VVZ to_TO create_VVI buffers_NN2 around_II busy_JJ roads_NN2 and_CC junctions_NN2 and_CC to_TO use_VVI data_NN from_II the_AT Cancer_NN1 Registry_NN1 to_TO examine_VVI possible_JJ links_NN2 between_II proximity_NN1 to_II such_DA sources_NN2 of_IO pollution_NN1 and_CC prevalence_NN1 of_IO lung_NN1 cancer_NN1 ._. 
Those_DD2 with_IW an_AT1 interest_NN1 in_II health_NN1 and_CC air_NN1 pollution_NN1 should_VM note_VVI the_AT availability_NN1 of_IO a_AT1 very_RG detailed_JJ database_NN1 from_II the_AT Warren_NP1 Springs_NN2 Laboratory_NN1 ,_, containing_VVG details_NN2 of_IO smoke_NN1 and_CC sulphur_NN1 dioxide_NN1 emissions_NN2 at_II a_AT1 large_JJ number_NN1 of_IO sites_NN2 in_II the_AT UK_NP1 ._. 
The_AT possible_JJ link_NN1 between_II nitrates_NN2 in_II water_NN1 and_CC stomach_NN1 cancer_NN1 is_VBZ the_AT subject_NN1 of_IO much_DA1 debate_NN1 (_( Beresford_NP1 1985_MC )_) ,_, as_CSA is_VBZ the_AT association_NN1 between_II aluminium_NN1 content_NN1 and_CC Alzheimer_NP1 's_GE disease_NN1 ._. 
The_AT problems_NN2 of_IO assessing_VVG such_DA links_NN2 are_VBR ,_, of_RR21 course_RR22 ,_, huge_JJ ,_, but_CCB there_EX is_VBZ scope_NN1 for_IF a_AT1 contribution_NN1 ._. 
The_AT role_NN1 of_IO GIS_NN2 in_II hazard_NN1 management_NN1 scenarios_NN2 examples_NN2 As_CSA we_PPIS2 have_VH0 already_RR implied_VVN ,_, GIS_NN2 has_VHZ a_AT1 role_NN1 to_TO play_VVI in_II all_DB aspects_NN2 of_IO hazard_NN1 research_NN1 ,_, from_II hazard_NN1 monitoring_NN1 ,_, risk_NN1 assessment_NN1 and_CC emergency_NN1 planning_VVG to_II coping_VVG with_IW an_AT1 event_NN1 and_CC evaluating_VVG its_APPGE consequences_NN2 ._. 
The_AT examples_NN2 described_VVN below_RL illustrate_VV0 all_DB these_DD2 ,_, though_CS as_RR21 yet_RR22 there_RL do_VD0 not_XX appear_VVI to_TO be_VBI many_DA2 instances_NN2 of_IO disasters_NN2 in_II the_AT management_NN1 of_IO which_DDQ GIS_NN2 has_VHZ played_VVN a_AT1 significant_JJ part_NN1 ._. 
Technological_JJ hazards_NN2 ,_, emergency_NN1 planning_NN1 and_CC GIS_NN2 A_ZZ1 considerable_JJ amount_NN1 of_IO research_NN1 has_VHZ been_VBN undertaken_VVN to_TO apply_VVI GIS_NN2 methods_NN2 to_II evacuation_NN1 scenarios_NN2 around_II hazardous_JJ sites_NN2 ._. 
Much_DA1 of_IO this_DD1 research_NN1 has_VHZ concentrated_VVN on_II network_NN1 analysis_NN1 of_IO the_AT road_NN1 system_NN1 ._. 
Reference_NN1 was_VBDZ made_VVN above_RL to_II so-called_JJ CIMAH_NN1 sites_NN2 ,_, major_JJ hazardous_JJ installations_NN2 for_IF which_DDQ on-site_JJ and_CC off-site_JJ emergency_NN1 plans_NN2 must_VM be_VBI prepared_VVN ._. 
Early_JJ research_NN1 at_II the_AT NWRRL_NP1 focused_VVD on_II developing_VVG a_AT1 GIS_NN2 for_IF use_NN1 by_II the_AT police_NN2 and_CC county_NN1 emergency_NN1 planning_NN1 officers_NN2 in_II Cumbria_NP1 that_CST would_VM aid_VVI in_II the_AT management_NN1 of_IO a_AT1 possible_JJ explosion_NN1 at_II a_AT1 chemical_JJ factory_NN1 in_II Whitehaven_NP1 ,_, west_ND1 Cumbria_NP1 (_( Vincent_NP1 et_RA21 al_RA22 ._. 
1988_MC ;_; Dunn_NP1 1989_MC )_) ._. 
The_AT police_NN2 required_VVN a_AT1 portable_JJ system_NN1 that_CST could_VM be_VBI implemented_VVN on_II an_AT1 IBM-compatible_JJ microcomputer_NN1 ._. 
This_DD1 was_VBDZ written_VVN at_II the_AT NWRRL_NP1 using_VVG FORTRAN_NP1 77_MC and_CC the_AT Graphical_JJ Kernel_NN1 System_NN1 ._. 
The_AT GIS_NN2 is_VBZ menu-based_JJ and_CC is_VBZ exceptionally_RR user-friendly_JJ (_( Pl._NP1 10.1_MC )_) ._. 
It_PPH1 opens_VVZ with_IW an_AT1 outline_NN1 map_NN1 and_CC the_AT superimposition_NN1 of_IO the_AT rather_RG rigid_JJ zones_NN2 used_VVN by_II the_AT emergency_NN1 planners_NN2 ._. 
Full_JJ zoom_NN1 facilities_NN2 are_VBR available_JJ at_II any_DD stage_NN1 ._. 
In_II any_DD emergency_NN1 it_PPH1 is_VBZ of_RR21 course_RR22 vital_JJ to_TO have_VHI high-quality_JJ data_NN on_II the_AT distribution_NN1 of_IO population_NN1 and_CC resources_NN2 ._. 
The_AT system_NN1 incorporates_VVZ digitized_JJ data_NN on_II the_AT distribution_NN1 of_IO all_DB domestic_JJ and_CC non-domestic_JJ properties_NN2 in_II the_AT town_NN1 together_RL with_IW digitized_JJ road_NN1 centre-line_NN1 data_NN ._. 
Both_DB2 sets_NN2 of_IO data_NN were_VBDR provided_VVN by_II Pinpoint_JJ Analysis_NN1 Ltd_JJ ,_, who_PNQS are_VBR undertaking_VVG this_DD1 digitizing_JJ task_NN1 for_IF the_AT entire_JJ country_NN1 ._. 
Other_JJ spatially_RR referenced_VVD data_NN incorporated_VVN into_II the_AT system_NN1 include_VV0 those_DD2 on_II the_AT distribution_NN1 of_IO schools_NN2 ,_, medical_JJ and_CC transport_NN1 services_NN2 which_DDQ can_VM be_VBI displayed_VVN and_CC supplementary_JJ information_NN1 about_II them_PPHO2 obtained_VVD ._. 
The_AT GIS_NN2 permits_VVZ population_NN1 estimates_VVZ to_TO be_VBI made_VVN for_IF any_DD arbitrary_JJ rectangular_JJ region_NN1 on_II the_AT map_NN1 ._. 
Most_RGT significantly_RR ,_, it_PPH1 incorporates_VVZ a_AT1 shortest_JJT path_NN1 algorithm_NN1 which_DDQ uses_VVZ the_AT road_NN1 network_NN1 to_TO find_VVI optimal_JJ routes_NN2 between_II user-defined_JJ start_NN1 and_CC end_VV0 nodes_NN2 ._. 
Although_CS predating_VVG the_AT RRL_NP1 initiative_NN1 by_II several_DA2 years_NNT2 and_CC ,_, indeed_RR the_AT emergence_NN1 of_IO GIS_NN2 as_II a_AT1 major_JJ research_NN1 area_NN1 we_PPIS2 should_VM remind_VVI ourselves_PPX2 of_IO Openshaw_NP1 's_GE work_NN1 on_II appraising_VVG nuclear_JJ reactor_NN1 sites_NN2 ._. 
Using_VVG what_DDQ would_VM now_RT be_VBI called_VVN GIS_NN2 skills_NN2 ,_, Openshaw_NP1 (_( 1980_MC )_) examined_VVD over_RG 13_MC 000_MC 1_MC1 km_NNU grid_NN1 squares_NN2 in_II the_AT UK_NP1 which_DDQ intersect_VV0 the_AT coastline_NN1 and_CC related_VVD these_DD2 to_II data_NN from_II the_AT 1971_MC Census_NN1 (_( which_DDQ were_VBDR made_VVN available_JJ for_IF such_DA grid_NN1 squares_NN2 )_) ._. 
He_PPHS1 demonstrated_VVD that_CST there_EX is_VBZ an_AT1 abundance_NN1 of_IO '_GE remote_JJ '_GE sites_NN2 ,_, in_II contrast_NN1 to_II the_AT official_JJ view_NN1 that_CST such_DA sites_NN2 were_VBDR increasingly_RR hard_JJ to_TO find_VVI ._. 
He_PPHS1 suggested_VVD that_CST in_II early_JJ spatial_JJ searches_NN2 for_IF sites_NN2 '_GE only_RR a_AT1 small_JJ number_NN1 were_VBDR ever_RR identified_VVN in_II the_AT first_MD place_NN1 because_CS rigorous_JJ searches_NN2 could_VM not_XX be_VBI performed_VVN by_II manual_JJ means_NN with_IW poor_JJ quality_NN1 data_NN '_GE (_( Openshaw_NP1 1980_MC :_: 289_MC )_) ._. 
Heywood_NP1 and_CC Cornelius_NP2 (_( 1989_MC )_) have_VH0 shown_VVN how_RRQ GIS_NN2 can_VM be_VBI used_VVN for_IF monitoring_VVG possible_JJ radiation_NN1 releases_NN2 and_CC thus_RR the_AT relevance_NN1 of_IO GIS_NN2 in_II nuclear_JJ emergency_NN1 planning_NN1 and_CC monitoring_NN1 ._. 
They_PPHS2 propose_VV0 a_AT1 national_JJ radiological_JJ spatial_JJ information_NN1 system_NN1 and_CC ,_, in_II a_AT1 pilot_NN1 project_NN1 in_II Cumbria_NP1 ,_, have_VH0 integrated_VVN several_DA2 layers_NN2 of_IO data_NN within_II ARC/INFO_NN1 to_TO show_VVI what_DDQ this_DD1 might_VM involve_VVI (_( Fig._NN1 10.2_MC )_) ._. 
Such_DA data_NN include_VV0 those_DD2 from_II the_AT Population_NN1 and_CC Agricultural_JJ Censuses_NN2 as_II31 well_II32 as_II33 point_NN1 data_NN from_II rain_NN1 gauge_NN1 sites_NN2 and_CC radiation_NN1 monitoring_NN1 stations_NN2 ._. 
As_II an_AT1 example_NN1 ,_, the_AT system_NN1 is_VBZ used_VVN to_TO determine_VVI restriction_NN1 zones_NN2 for_IF the_AT movement_NN1 of_IO sheep_NN after_II the_AT explosion_NN1 at_II Chernobyl_NP1 in_II 1986_MC ._. 
This_DD1 is_VBZ done_VDN by_II finding_VVG areas_NN2 of_IO high_JJ rainfall_NN1 and_CC parishes_NN2 with_IW high_JJ sheep_NN populations_NN2 ._. 
The_AT map_NN1 overlay_NN1 of_IO point_NN1 and_CC area_NN1 information_NN1 shows_VVZ generally_RR good_JJ agreement_NN1 with_IW the_AT restriction_NN1 zones_NN2 imposed_VVN by_II the_AT Ministry_NN1 of_IO Agriculture_NN1 ,_, Fisheries_NN2 and_CC Food_NN1 ,_, with_IW the_AT exception_NN1 of_IO a_AT1 small_JJ area_NN1 to_II the_AT south-east_ND1 of_IO Cumbria_NP1 ,_, bordering_VVG on_II Lancashire_NP1 ._. 
The_AT British_JJ government_NN1 carried_VVD out_RP a_AT1 review_NN1 of_IO its_APPGE arrangements_NN2 for_IF handling_VVG nuclear_JJ accidents_NN2 overseas_RL following_VVG the_AT Chernobyl_NP1 accident_NN1 ._. 
The_AT DoE_NP1 ,_, through_II Her_APPGE Majesty_NN1 's_GE Inspectorate_NN1 of_IO Pollution_NN1 (_( HMIP_NP1 )_) ,_, was_VBDZ given_VVN the_AT task_NN1 of_IO co-ordinating_VVG a_AT1 national_JJ response_NN1 plan_NN1 ,_, installing_VVG the_AT radiation_NN1 monitoring_NN1 network_NN1 and_CC acting_VVG as_II lead_NN1 government_NN1 department_NN1 in_II an_AT1 emergency_NN1 (_( HMSO_NP1 1988_MC )_) ._. 
The_AT monitoring_NN1 network_NN1 has_VHZ become_VVN known_VVN as_II the_AT radioactive_JJ incident_NN1 monitoring_NN1 network_NN1 (_( RIMNET_NP1 )_) and_CC has_VHZ ,_, as_CSA its_APPGE prime_JJ responsibility_NN1 ,_, the_AT detection_NN1 of_IO abnormal_JJ increases_NN2 in_II radiation_NN1 levels_NN2 within_II the_AT UK_NP1 of_IO the_AT kind_NN1 that_CST might_VM arise_VVI from_II an_AT1 overseas_JJ nuclear_JJ accident_NN1 ._. 
The_AT full_JJ implementation_NN1 of_IO the_AT RIMNET_NN1 system_NN1 will_VM take_VVI some_DD time_NNT1 ._. 
By_II the_AT end_NN1 of_IO 1988_MC gamma_NN1 radiation_NN1 dose_NN1 monitors_NN2 had_VHD been_VBN installed_VVN at_II 46_MC meteorological_JJ stations_NN2 as_II part_NN1 of_IO phase_NN1 1_MC1 of_IO the_AT development_NN1 programme_NN1 ._. 
Radiation_NN1 readings_NN2 are_VBR taken_VVN every_AT1 hour_NNT1 at_II each_DD1 station_NN1 and_CC are_VBR transmitted_VVN to_II a_AT1 central_JJ computer_NN1 which_DDQ analyses_VVZ the_AT data_NN ._. 
The_AT current_JJ criterion_NN1 for_IF alert_NN1 is_VBZ two_MC readings_NN2 of_IO more_DAR than_CSN three_MC times_NNT2 normal_JJ background_NN1 ,_, either_RR successively_RR at_II the_AT same_DA site_NN1 or_CC simultaneously_RR at_II two_MC adjacent_JJ sites_NN2 ._. 
The_AT dissemination_NN1 of_IO information_NN1 in_II the_AT event_NN1 of_IO an_AT1 incident_NN1 is_VBZ dealt_VVN with_IW at_II three_MC levels_NN2 :_: the_AT first_MD is_VBZ the_AT normal_JJ system_NN1 of_IO press_NN1 notices_NN2 ;_; the_AT second_NNT1 is_VBZ the_AT use_NN1 of_IO teletext_NN1 and_CC viewdata_NN1 (_( Ceefax_NP1 and_CC Prestel_NP1 )_) to_TO get_VVI direct_JJ to_II members_NN2 of_IO the_AT public_NN1 at_II their_APPGE homes_NN2 or_CC places_NN2 of_IO work_NN1 ;_; the_AT third_MD is_VBZ the_AT use_NN1 of_IO Telecom_NP1 Gold_NN1 electronic_JJ mail_NN1 to_TO get_VVI more_RGR technical_JJ details_NN2 to_II various_JJ official_JJ bodies_NN2 ,_, such_II21 as_II22 water_NN1 authorities_NN2 (_( Jackson_NP1 1989_MC )_) ._. 
Phase_NN1 2_MC of_IO the_AT RIMNET_NN1 scheme_NN1 will_VM be_VBI completed_VVN by_II the_AT end_NN1 of_IO 1991_MC when_RRQ between_II 80_MC and_CC 90_MC monitoring_NN1 sites_NN2 will_VM have_VHI been_VBN established_VVN ._. 
Given_VVN very_RG wide_JJ regional_JJ and_CC local_JJ variations_NN2 in_II deposition_NN1 ,_, questions_NN2 must_VM be_VBI raised_VVN about_II the_AT effectiveness_NN1 of_IO such_DA a_AT1 small_JJ network_NN1 ;_; local_JJ authorities_NN2 will_VM need_VVI to_TO supplement_VVI it_PPH1 with_IW their_APPGE own_DA networks_NN2 ._. 
There_EX are_VBR important_JJ analytical_JJ problems_NN2 to_TO be_VBI addressed_VVN concerning_II the_AT siting_NN1 of_IO such_DA monitoring_NN1 equipment_NN1 and_CC these_DD2 are_VBR elaborated_VVN on_II later_JJR ._. 
Abroad_RL ,_, Kaspar_NP1 (_( 1981_MC )_) has_VHZ discussed_VVN a_AT1 computer-based_JJ exercise_NN1 ,_, with_IW GIS_NN2 overtones_VVZ ,_, involving_VVG the_AT simulated_JJ evacuation_NN1 of_IO the_AT population_NN1 living_VVG in_II the_AT vicinity_NN1 of_IO the_AT Neckarwestheim_NP1 nuclear_JJ power_NN1 plant_NN1 in_II Baden_NP1 Wrttemberg_NP1 ._. 
In_II the_AT scenario_NN1 he_PPHS1 outlines_NN2 ,_, an_AT1 area_NN1 affected_VVN by_II radioactive_JJ releases_NN2 is_VBZ determined_VVN ,_, the_AT population_NN1 within_II that_DD1 area_NN1 is_VBZ allocated_VVN to_II reception_NN1 areas_NN2 (_( with_IW capacity_NN1 constraints_NN2 )_) ,_, evacuation_NN1 routes_NN2 are_VBR identified_VVN and_CC traffic_NN1 restrictions_NN2 implemented_VVN ._. 
Little_DA1 detail_NN1 is_VBZ provided_VVN ,_, however_RR ,_, and_CC it_PPH1 is_VBZ thus_RR difficult_JJ to_TO evaluate_VVI his_APPGE work_NN1 fully_RR ._. 
Dangermond_NP1 (_( 1985_MC )_) describes_VVZ the_AT use_NN1 of_IO the_AT ARC/INFO_NN1 NETWORK_NN1 module_NN1 for_IF use_NN1 in_II emergency_NN1 planning_NN1 situations_NN2 ._. 
He_PPHS1 describes_VVZ how_RRQ a_AT1 road_NN1 network_NN1 is_VBZ digitized_VVN and_CC subsequently_RR analysed_VVN using_VVG allocation_NN1 ,_, districting_VVG and_CC routeing_VVG algorithms_NN2 ._. 
Allocation_NN1 determines_VVZ which_DDQ arcs_NN2 in_II the_AT network_NN1 will_VM be_VBI allocated_VVN to_II a_AT1 particular_JJ node_NN1 or_CC centre_NN1 ;_; districting_VVG makes_VVZ it_PPH1 possible_JJ to_TO outline_VVI rapidly_RR sets_VVZ of_IO polygons_NN2 in_BCL21 order_BCL22 to_TO define_VVI specific_JJ areas_NN2 of_IO interest_NN1 (_( districts_NN2 )_) and_CC to_TO summarize_VVI their_APPGE characteristics_NN2 ;_; routeing_VVG provides_VVZ a_AT1 minimum_JJ path_NN1 algorithm_NN1 through_II the_AT network_NN1 ,_, the_AT arcs_NN2 of_IO which_DDQ can_VM be_VBI assigned_VVN weights_NN2 according_II21 to_II22 road_NN1 conditions_NN2 ,_, road_NN1 capacity_NN1 and_RR31 so_RR32 on_RR33 ._. 
Dangermond_NN1 gives_VVZ several_DA2 examples_NN2 in_II his_APPGE paper_NN1 including_II the_AT use_NN1 of_IO NETWORK_NN1 for_IF the_AT allocation_NN1 of_IO emergency_NN1 vehicles_NN2 ,_, optimum_JJ routeing_NN1 of_IO fire_NN1 engines_NN2 from_II garages_NN2 to_II the_AT accident_NN1 scene_NN1 and_CC the_AT movement_NN1 of_IO spills_NN2 through_II sewers_NN2 and_CC river_NN1 networks_NN2 ._. 
When_CS combined_VVN with_IW the_AT other_JJ facilities_NN2 in_II ARC/INFO_NN1 ,_, quite_RG complex_JJ disaster_NN1 management_NN1 scenarios_NN2 can_VM be_VBI handled_VVN ._. 
For_REX21 example_REX22 ,_, earthquake_NN1 fault_NN1 zones_NN2 can_VM be_VBI combined_VVN with_IW data_NN on_II housing_VVG density_NN1 and_CC structural_JJ details_NN2 to_TO predict_VVI earthquake_NN1 damage_NN1 levels_NN2 and_CC these_DD2 data_NN then_RT related_VVN to_II the_AT network_NN1 model_NN1 ._. 
Vincent_NP1 (_( 1989_MC )_) illustrates_VVZ the_AT use_NN1 of_IO ARC/INFO_NN1 and_CC its_APPGE NETWORK_NN1 module_NN1 for_IF finding_VVG safe_JJ routes_NN2 for_IF chlorine_NN1 transhipment_NN1 on_II the_AT road_NN1 and_CC rail_NN1 systems_NN2 of_IO north-west_ND1 England_NP1 (_( Pl._NP1 10.2_MC )_) ._. 
In_II this_DD1 case_NN1 ,_, population_NN1 density_NN1 data_NN for_IF a_AT1 1_MC1 km_NNU buffer_NN1 on_II either_DD1 side_NN1 of_IO network_NN1 arcs_NN2 were_VBDR overlaid_VVN on_II the_AT network_NN1 itself_PPX1 so_BCL21 as_BCL22 to_TO act_VVI as_CSA weights_NN2 for_IF the_AT minimum_JJ path_NN1 algorithm_NN1 ._. 
Chosen_JJ routes_NN2 were_VBDR those_DD2 which_DDQ minimized_VVD population_NN1 rather_II21 than_II22 distance_NN1 ._. 
Hazardous_JJ waste_NN1 disposal_NN1 There_EX can_VM be_VBI little_RR doubt_VV0 that_CST a_AT1 GIS_NN2 approach_NN1 has_VHZ wide_JJ applicability_NN1 in_II all_DB sorts_NN2 of_IO location_NN1 problems_NN2 where_RRQ the_AT goal_NN1 is_VBZ to_TO minimize_VVI (_( rather_CS21 than_CS22 ,_, as_CSA conventionally_RR ,_, maximize_VV0 )_) accessibility_NN1 to_II a_AT1 population_NN1 ._. 
A_AT1 particularly_RR good_JJ application_NN1 is_VBZ to_II the_AT search_NN1 for_IF potential_JJ sites_NN2 for_IF the_AT disposal_NN1 of_IO nuclear_JJ and_CC non-nuclear_JJ (_( but_CCB none_RR31 the_RR32 less_RR33 hazardous_JJ )_) wastes_VVZ ._. 
We_PPIS2 discuss_VV0 here_RL some_DD relevant_JJ work_NN1 on_II the_AT siting_NN1 of_IO facilities_NN2 for_IF the_AT disposal_NN1 of_IO nuclear_JJ wastes_NN2 before_II examining_VVG those_DD2 for_IF non-nuclear_JJ wastes_NN2 ._. 
The_AT major_JJ contribution_NN1 to_II informed_JJ debate_NN1 about_II the_AT search_NN1 for_IF sites_NN2 for_IF the_AT disposal_NN1 of_IO nuclear_JJ wastes_NN2 has_VHZ come_VVN from_II a_AT1 group_NN1 of_IO geographers_NN2 (_( Openshaw_NP1 et_RA21 al_RA22 ._. 
1989_MC )_) and_CC a_AT1 GIS_NN2 approach_VV0 to_II this_DD1 problem_NN1 can_VM pay_VVI high_JJ dividends_NN2 ._. 
The_AT work_NN1 of_IO Carver_NP1 (_( see_VV0 Openshaw_NP1 et_RA21 al_RA22 ._. 
1989_MC :_: Ch._NN1 7_MC )_) is_VBZ an_AT1 interesting_JJ example_NN1 ._. 
One_MC1 advantage_NN1 is_VBZ the_AT ability_NN1 to_TO model_VVI a_AT1 range_NN1 of_IO alternatives_NN2 by_II allowing_VVG particular_JJ criteria_NN2 to_TO enter_VVI or_CC be_VBI omitted_VVN from_II the_AT polygon_NN1 overlay_VV0 operation_NN1 ._. 
Conceptually_RR ,_, what_DDQ is_VBZ offered_VVN is_VBZ little_DA1 more_DAR than_CSN an_AT1 automated_JJ '_GE sieve_NN1 mapping_NN1 '_GE that_DD1 land-use_NN1 planners_NN2 have_VH0 used_VVN for_IF many_DA2 years_NNT2 ,_, but_CCB the_AT whole_JJ procedure_NN1 is_VBZ speeded_VVN up_RP by_II many_DA2 orders_NN2 of_IO magnitude_NN1 using_VVG GIS_NN2 software_NN1 ._. 
The_AT authors_NN2 examine_VV0 both_RR near-surface_JJ sites_NN2 and_CC deep_JJ repositories_NN2 (_( which_DDQ use_VV0 more_RGR restricted_JJ population_NN1 criteria_NN2 and_CC allow_VV0 offshore_JJ locations_NN2 to_TO enter_VVI )_) ._. 
Data_NN layers_NN2 include_VV0 geological_JJ maps_NN2 ,_, conservation_JJ areas_NN2 ,_, transport_NN1 routes_NN2 ,_, petrochemical_JJ facilities_NN2 and_CC demographic_JJ data_NN ._. 
As_CSA different_JJ layers_NN2 are_VBR added_VVN sequentially_RR we_PPIS2 see_VV0 how_RRQ the_AT proportion_NN1 of_IO available_JJ land_NN1 shrinks_NN2 (_( Pl._NP1 10.3_MC )_) ;_; for_REX21 instance_REX22 ,_, when_CS geological_JJ criteria_NN2 are_VBR set_VVN in_II the_AT search_NN1 for_IF a_AT1 deep_JJ repository_NN1 ,_, 25_MC per_NNU21 cent_NNU22 of_IO land_NN1 area_NN1 remains_VVZ and_CC the_AT addition_NN1 of_IO the_AT population_NN1 density_NN1 layer_NN1 reduces_VVZ this_DD1 to_II 24_MC per_NNU21 cent_NNU22 ._. 
However_RR ,_, adding_VVG a_AT1 proximity_NN1 to_TO transport_VVI constraint_NN1 lowers_VVZ this_DD1 to_II 12_MC per_NNU21 cent_NNU22 ;_; and_RR31 so_RR32 on_RR33 ._. 
We_PPIS2 can_VM not_XX pretend_VVI GIS_NN2 is_VBZ a_AT1 solution_NN1 to_II the_AT political_JJ problems_NN2 of_IO negotiation_NN1 and_CC justification_NN1 but_CCB at_II the_AT very_RG least_RRT the_AT openness_NN1 of_IO the_AT procedures_NN2 goes_VVZ some_DD way_NN1 towards_II satisfying_VVG a_AT1 worried_JJ public_NN1 ._. 
GIS_NN2 research_VV0 into_II site_NN1 selection_NN1 for_IF non-nuclear_JJ hazardous_JJ waste_NN1 has_VHZ been_VBN almost_RR exclusively_RR conducted_VVN in_II North_ND1 America_NP1 and_CC has_VHZ yet_RR to_TO be_VBI matched_VVN in_II the_AT UK_NP1 ._. 
Jensen_NP1 and_CC Christiansen_NP1 (_( 1986_MC )_) ,_, working_VVG in_II the_AT south-eastern_JJ USA_NP1 ,_, use_VV0 a_AT1 raster_NN1 approach_NN1 (_( with_IW a_AT1 pixel_NN1 resolution_NN1 of_IO 20_MC m&amp;sup2_FO ;_; )_) and_CC Boolean_NN1 overlay_VV0 to_TO '_" weed_VVI out_RP '_GE particular_JJ sites_NN2 ,_, such_II21 as_II22 those_DD2 which_DDQ are_VBR poorly_RR drained_VVN ,_, environmentally_RR sensitive_JJ or_CC inaccessible_JJ (_( see_VV0 also_RR Stewart_NP1 1987_MC )_) ._. 
Environmental_JJ monitoring_NN1 of_IO sites_NN2 for_IF the_AT disposal_NN1 of_IO toxic_JJ waste_NN1 is_VBZ of_IO particular_JJ concern_NN1 in_II the_AT USA_NP1 ._. 
The_AT Comprehensive_JJ Environmental_JJ Response_NN1 ,_, Compensation_NN1 and_CC Liability_NN1 Act_NN1 (_( CERCLA_NP1 )_) legislation_NN1 of_IO 1980_MC (_( '_GE Superfund_NP1 '_GE )_) requires_VVZ the_AT cleaning_NN1 up_RP of_IO the_AT worst_JJT landfill_NN1 sites_NN2 ,_, of_IO which_DDQ some_DD 800_MC had_VHD been_VBN identified_VVN by_II the_AT Environmental_JJ Protection_NN1 Agency_NN1 by_II 1986_MC (_( Foresman_NP1 1986_MC )_) ._. 
A_AT1 major_JJ worry_NN1 has_VHZ been_VBN the_AT threat_NN1 posed_VVN by_II groundwater_NN1 pollution_NN1 and_CC some_DD research_NN1 has_VHZ been_VBN conducted_VVN on_II using_VVG GIS_NN2 to_TO monitor_VVI this_DD1 (_( Merchant_NN1 et_RA21 al_RA22 ._. 
1987_MC ;_; Barringer_NP1 et_RA21 al_RA22 ._. 
1987_MC ;_; von_NP1 Braun_NP1 1988_MC )_) ._. 
This_DD1 is_VBZ an_AT1 important_JJ topic_NN1 since_II 95_MC per_NNU21 cent_NNU22 of_IO rural_JJ households_NN2 in_II the_AT USA_NP1 consume_VV0 only_JJ groundwater_NN1 ,_, while_CS half_DB the_AT US_NP1 population_NN1 consumes_VVZ at_RR21 least_RR22 some_DD groundwater_NN1 (_( Merchant_NN1 et_RA21 al_RA22 ._. 
1987_MC )_) ._. 
Clearly_RR ,_, the_AT research_NN1 demands_NN2 access_VV0 to_TO exceptionally_RR good_JJ environmental_JJ data_NN ,_, usually_RR in_II three_MC dimensions_NN2 ._. 
As_CSA Foresman_NP1 (_( 1986_MC )_) suggests_VVZ ,_, we_PPIS2 might_VM expect_VVI remote_JJ sensing_NN1 technology_NN1 to_TO play_VVI a_AT1 major_JJ role_NN1 ._. 
Merchant_NN1 et_RA21 al_RA22 ._. 
(_( 1987_MC )_) for_REX21 example_REX22 ,_, use_VV0 the_AT image_NN1 processing_NN1 system_NN1 ERDAS_NN2 to_TO construct_VVI an_AT1 index_NN1 of_IO vulnerability_NN1 to_II groundwater_NN1 pollution_NN1 ;_; this_DD1 uses_VVZ variables_NN2 such_II21 as_II22 depth_NN1 to_II water_NN1 table_NN1 ,_, soil_NN1 type_NN1 ,_, slope_VV0 and_RR31 so_RR32 on_RR33 (_( see_VV0 Estes_NP1 et_RA21 al_RA22 ._. 
1987_MC for_IF a_AT1 similar_JJ approach_NN1 )_) ._. 
Barringer_NP1 et_RA21 al_RA22 ._. 
(_( 1987_MC )_) overlay_VV0 data_NN on_II soil_NN1 type_NN1 and_CC geology_NN1 with_IW that_DD1 of_IO groundwater_NN1 quality_NN1 (_( hardness_NN1 and_CC corrosivity_NN1 measures_NN2 )_) and_CC build_VV0 a_AT1 linear_JJ model_NN1 which_DDQ shows_VVZ that_CST the_AT background_NN1 environmental_JJ variables_NN2 are_VBR good_JJ predictors_NN2 of_IO the_AT water_NN1 quality_NN1 indices_NN2 ._. 
Again_RT ,_, a_AT1 raster-based_JJ approach_NN1 is_VBZ used_VVN ._. 
A_AT1 further_JJR example_NN1 is_VBZ work_NN1 by_II von_NP1 Braun_NP1 (_( 1988_MC )_) who_PNQS uses_VVZ pMAP_NN1 to_TO look_VVI at_II exposure_NN1 of_IO organic_JJ and_CC metallic_JJ compounds_NN2 in_II groundwater_NN1 in_II the_AT vicinity_NN1 of_IO Tucson_NP1 airport_NN1 ,_, Arizona_NP1 ._. 
Using_VVG plume_NN1 models_NN2 that_CST predict_VV0 the_AT movement_NN1 of_IO contaminants_NN2 she_PPHS1 is_VBZ able_JK to_TO intersect_VVI the_AT results_NN2 with_IW data_NN on_II current_JJ well_JJ locations_NN2 and_CC to_TO assess_VVI which_DDQ water_NN1 supply_NN1 areas_NN2 are_VBR worst_RRT contaminated_VVN ._. 
Overlaying_VVG the_AT results_NN2 with_IW residential_JJ location_NN1 data_NN provides_58 '_GE maps_NN2 delineating_VVG location-specific_JJ exposures_NN2 for_IF each_DD1 residence_NN1 in_II each_DD1 time-period_NN1 '_GE (_( von_NP1 Braun_NP1 1988_MC ;_; 1160_MC )_) ._. 
In_II principle_NN1 ,_, it_PPH1 is_VBZ then_RT possible_JJ to_TO compare_VVI such_DA '_GE doses_NN2 '_GE with_IW epidemiological_JJ data_NN ,_, once_RR21 again_RR22 indicating_VVG a_AT1 link_NN1 from_II a_AT1 hazard_NN1 event_NN1 to_II assessment_NN1 of_IO health_NN1 outcomes_NN2 ._. 
Attention_NN1 should_VM be_VBI drawn_VVN to_II Foresman_NP1 's_GE comment_NN1 that_CST the_AT US_NP1 Geological_JJ Survey_NN1 (_( USGS_NP2 )_) (_( at_II the_AT EROS_NP1 Data_NN Centre_NN1 )_) have_VH0 successfully_RR linked_VVN ARC/INFO_NN1 to_II a_AT1 model_NN1 of_IO groundwater_NN1 flow_NN1 (_( Foresman_NP1 1986_MC :_: 260_MC )_) ._. 
This_DD1 kind_NN1 of_IO link_NN1 between_II GIS_NN2 and_CC other_JJ modelling_NN1 programs_NN2 is_VBZ to_TO be_VBI warmly_RR welcomed_VVN and_CC parallels_NN2 work_VV0 at_II the_AT NWRRL_NP1 in_II linking_VVG GIS_NN2 and_CC air_NN1 pollution_NN1 plume_NN1 models_NN2 ._. 
Natural_JJ hazards_NN2 and_CC disaster_NN1 management_NN1 Several_DA2 studies_NN2 in_II recent_JJ papers_NN2 have_VH0 shown_VVN the_AT effectiveness_NN1 of_IO GIS_NN2 in_II mitigating_VVG the_AT effects_NN2 of_IO major_JJ natural_JJ hazards_NN2 ._. 
Berke_VV0 and_CC Ruch_NP1 (_( 1985_MC )_) ,_, for_REX21 example_REX22 ,_, describe_VV0 a_AT1 GIS_NN2 for_IF assessing_VVG the_AT impacts_NN2 of_IO hurricanes_NN2 on_II the_AT Texas_NP1 Gulf_NN1 coast_NN1 ._. 
Hurricane_NN1 damage_NN1 through_II high_JJ winds_NN2 and_CC tidal_JJ surges_NN2 causes_VVZ an_AT1 immense_JJ amount_NN1 of_IO destruction_NN1 and_CC poses_VVZ a_AT1 major_JJ threat_NN1 to_II many_DA2 coastal_JJ communities_NN2 in_II the_AT USA_NP1 ._. 
It_PPH1 has_VHZ been_VBN estimated_VVN ,_, for_REX21 example_REX22 ,_, that_CST during_II the_AT mid_JJ 1970s_MC2 some_DD 50_MC million_NNO US_NP1 citizens_NN2 were_VBDR subject_II21 to_II22 hurricane_NN1 winds_NN2 of_IO over_RG 160_MC km/h_FU and_CC 6_MC million_NNO were_VBDR subject_II21 to_II22 hurricane_NN1 surge_NN1 (_( Brinkman_NP1 1975_MC )_) ._. 
The_AT system_NN1 described_VVN by_II Berke_NP1 and_CC Ruch_NP1 (_( 1985_MC )_) is_VBZ raster-based_JJ and_CC has_VHZ an_AT1 interesting_JJ modelling_NN1 component_NN1 called_VVN SLOSH_NP1 (_( Sea_NN1 ,_, Lake_NN1 and_CC Overland_NN1 Surges_VVZ from_II Hurricanes_NN2 )_) ._. 
SLOSH_NP1 is_VBZ able_JK to_TO estimate_VVI still_JJ water_NN1 surge_NN1 by_II grid_NN1 cell_NN1 for_IF various_JJ hurricane_NN1 intensities_NN2 ,_, angles_NN2 of_IO approach_NN1 to_II the_AT shoreline_NN1 ,_, forward_JJ movement_NN1 speeds_NN2 and_CC landfall_NN1 locations_NN2 ._. 
Additional_JJ algorithms_NN2 are_VBR used_VVN to_TO model_VVI total_JJ building_NN1 damage_NN1 costs_VVZ for_IF each_DD1 grid_NN1 cell_NN1 based_VVN on_II water-depth_NN1 and_CC speed-depth_JJ rating_NN1 curves_NN2 ._. 
Hobeika_NN1 and_CC Jamei_NP1 (_( 1985_MC )_) describe_VV0 the_AT MASSeVACuation_NN1 (_( MASSVAC_NP1 )_) simulation_NN1 software_NN1 designed_VVN to_TO analyse_VVI and_CC evaluate_VVI traffic_NN1 evacuation_NN1 plans_NN2 given_VVN a_AT1 natural_JJ disaster_NN1 in_II an_AT1 urban_JJ area_NN1 ._. 
The_AT system_NN1 comprises_VVZ three_MC interrelated_JJ modules_NN2 :_: community_NN1 and_CC disaster_NN1 type_NN1 characteristics_NN2 module_NN1 ;_; population_NN1 distribution_NN1 module_NN1 ;_; network_NN1 evacuation_NN1 module_NN1 ._. 
The_AT type_NN1 of_IO natural_JJ disaster_NN1 determines_VVZ the_AT time_NNT1 period_NN1 within_II which_DDQ the_AT road_NN1 network_NN1 has_VHZ to_TO be_VBI evacuated_VVN ._. 
The_AT control_NN1 and_CC management_NN1 strategies_NN2 are_VBR directly_RR correlated_VVN to_II each_DD1 specific_JJ disaster_NN1 scenario_NN1 ._. 
MASSVAC_NP1 is_VBZ quite_RG flexible_JJ and_CC various_JJ user-defined_JJ road_NN1 network_NN1 options_NN2 are_VBR available_JJ such_II21 as_II22 :_: traffic_NN1 signal_NN1 timings_NN2 on_II the_AT road_NN1 network_NN1 ;_; one-way_JJ traffic_NN1 ;_; reserved_JJ lanes_NN2 for_IF special_JJ vehicles_NN2 such_II21 as_II22 those_DD2 of_IO the_AT emergency_NN1 services_NN2 ._. 
Hobeika_NN1 and_CC Jamei_NP1 have_VH0 tested_VVN MASSVAC_NN1 on_II various_JJ evacuation_NN1 scenarios_NN2 for_IF Virginia_NP1 Beach_NP1 City_NN1 in_BCL21 order_BCL22 to_TO determine_VVI evacuation_NN1 times_NNT2 for_IF flood_NN1 and_CC hurricane_NN1 conditions_NN2 ._. 
In_RR21 general_RR22 ,_, they_PPHS2 suggest_VV0 that_CST the_AT model_NN1 provided_CS reasonable_JJ and_CC reliable_JJ evacuation_NN1 times_NNT2 ._. 
Southworth_NP1 et_RA21 al_RA22 ._. 
(_( 1989_MC )_) have_VH0 enhanced_VVN the_AT MASSVAC_JJ model_NN1 with_IW the_AT addition_NN1 of_IO a_AT1 colour_NN1 graphics_NN module_NN1 which_DDQ they_PPHS2 call_VV0 MVOPL_NN1 ._. 
This_DD1 module_NN1 has_VHZ been_VBN developed_VVN using_VVG IBM_NP1 's_GE Graphics_NN Kernel_NP1 System_NN1 and_CC provides_VVZ a_AT1 number_NN1 of_IO display_NN1 screens_NN2 ._. 
The_AT most_RGT interesting_JJ are_VBR those_DD2 which_DDQ portray_VV0 population_NN1 flows_VVZ as_RG scaled_JJ ribbons_NN2 on_II the_AT network_NN1 diagram_NN1 ,_, and_CC evacuation_NN1 shelter_NN1 capacities_NN2 as_CSA bar_NN1 diagrams_NN2 ._. 
The_AT diagrams_NN2 are_VBR updated_VVN periodically_RR during_II the_AT course_NN1 of_IO the_AT simulation_NN1 ._. 
In_II dry_JJ parts_NN2 of_IO the_AT world_NN1 ,_, such_II21 as_II22 Australia_NP1 ,_, forest_NN1 and_CC bush_NN1 fires_NN2 are_VBR an_AT1 important_JJ natural_JJ hazard_NN1 and_CC some_DD attention_NN1 has_VHZ been_VBN given_VVN to_II the_AT ways_NN2 in_II which_DDQ GIS_NN2 might_VM help_VVI examine_VVI their_APPGE potential_JJ impacts_NN2 ._. 
Kessell_NP1 (_( 1988_MC )_) describes_VVZ a_AT1 PC-based_JJ GIS_NN2 called_VVN PREPLAN_NP1 (_( PRistine_JJ Environment_NN1 Planning_NN1 LANguage_NN1 )_) which_DDQ is_VBZ a_AT1 natural_JJ area_NN1 management_NN1 ,_, land-use_NN1 planning_NN1 and_CC fire_NN1 modelling_NN1 system_NN1 developed_VVN for_IF the_AT Australian_JJ National_JJ Parks_NN2 and_CC Wildlife_NN1 Service_NN1 ,_, PREPLAN_NP1 comprises_VVZ four_MC modules_NN2 :_: a_AT1 simple_JJ raster-based_JJ GIS_NN2 ;_; a_AT1 grid_NN1 cell_NN1 resource_NN1 database_NN1 ;_; a_AT1 wide_JJ range_NN1 of_IO vegetation_NN1 ,_, fuel_NN1 ,_, fire_NN1 behaviour_NN1 ,_, erosion_NN1 and_CC land-use_NN1 models_NN2 ;_; tabular_JJ ,_, statistical_JJ and_CC colour_NN1 graphics_NN output_NN1 system_NN1 ._. 
With_IW this_DD1 system_NN1 real-time_NN1 fire_NN1 growth_NN1 can_VM be_VBI modelled_VVN ,_, taking_VVG into_II account_NN1 changes_NN2 in_II terrain_NN1 ,_, fuel_NN1 and_CC temporal_JJ changes_NN2 in_II the_AT weather_NN1 ._. 
Kessell_NN1 points_VVZ out_RP that_DD1 uncontrollable_JJ fires_NN2 are_VBR recurrent_JJ phenomena_NN2 in_II many_DA2 parts_NN2 of_IO Australia_NP1 ,_, causing_VVG damage_NN1 in_II the_AT hundreds_NNO2 of_IO millions_NNO2 of_IO dollars_NNU2 and_CC significant_JJ loss_NN1 of_IO life_NN1 ._. 
Systems_NN2 such_II21 as_II22 PREPLAN_NP1 can_VM provide_VVI the_AT user_NN1 with_IW sound_NN1 models_NN2 of_IO the_AT behaviour_NN1 patterns_NN2 of_IO fire_NN1 and_CC are_VBR therefore_RR of_IO enormous_JJ value_NN1 in_II land-use_NN1 planning_NN1 ._. 
One_MC1 of_IO the_AT most_RGT widespread_JJ natural_JJ hazards_NN2 is_VBZ flooding_NN1 ,_, and_CC the_AT potential_NN1 for_IF loss_NN1 of_IO property_NN1 ,_, risk_VV0 to_II life_NN1 and_CC general_JJ social_JJ disruption_NN1 is_VBZ large_JJ ._. 
One_MC1 interesting_JJ example_NN1 of_IO a_AT1 GIS_NN2 approach_VV0 to_TO flood_VVI damage_NN1 estimation_NN1 is_VBZ the_AT ANUFLOOD_NN1 package_NN1 which_DDQ was_VBDZ developed_VVN in_II the_AT early_JJ 1980s_MC2 following_VVG detailed_JJ flood_NN1 damage_NN1 studies_NN2 for_IF flood-prone_JJ coastal_JJ towns_NN2 in_II northern_JJ New_NP1 South_NP1 Wales_NP1 ,_, Australia_NP1 (_( Smith_NP1 and_CC Greenaway_NP1 ,_, 1988_MC )_) ._. 
The_AT ANUFLOOD_NN1 system_NN1 requires_VVZ locational_JJ and_CC building_NN1 type_NN1 data_NN on_II individual_JJ properties_NN2 ,_, stage-damage_JJ curves_NN2 relating_VVG the_AT average_JJ damage_NN1 that_CST would_VM result_VVI from_II overfloor_NN1 flooding_NN1 to_II differing_JJ depths_NN2 for_IF each_DD1 property_NN1 type_NN1 ,_, and_CC lastly_RR ,_, flood_VV0 frequency_NN1 in_II31 terms_II32 of_II33 flood_NN1 height_NN1 expressed_VVN as_II a_AT1 probability_NN1 ._. 
Output_NN1 from_II the_AT system_NN1 is_VBZ both_RR tabular_JJ and_CC graphical_JJ ._. 
Of_IO particular_JJ interest_NN1 is_VBZ the_AT construction_NN1 of_IO maps_NN2 showing_VVG the_AT expected_JJ annual_JJ amounts_NN2 of_IO damage_NN1 for_IF gridded_JJ flood-prone_JJ areas_NN2 ._. 
ANUFLOOD_VV0 also_RR contains_VVZ modules_NN2 which_DDQ permit_VV0 the_AT investigation_NN1 of_IO various_JJ flood_NN1 damage_NN1 mitigation_NN1 options_NN2 ,_, such_II21 as_II22 property_NN1 height_NN1 raising_NN1 ,_, levee_NN1 construction_NN1 and_CC flood_VV0 proofing_NN1 ._. 
Environmental_JJ epidemiology_NN1 The_AT possibilities_NN2 sketched_VVD above_RL relate_VV0 to_II emergency_NN1 planning_NN1 ,_, the_AT search_NN1 for_IF hazardous_JJ sites_NN2 ,_, and_CC risk_NN1 assessment_NN1 ._. 
To_II what_DDQ extent_NN1 can_VM GIS_NN2 help_VVI if_CSW an_AT1 event_NN1 occurs_VVZ ?_? 
What_DDQ ,_, for_REX21 instance_REX22 ,_, of_IO the_AT health_NN1 consequences_NN2 ?_? 
There_EX is_VBZ a_AT1 rich_JJ tradition_NN1 of_IO epidemiological_JJ work_NN1 in_II geography_NN1 and_CC with_IW the_AT increasing_JJ availability_NN1 of_IO spatially_RR referenced_VVD medical_JJ data_NN the_AT scope_NN1 for_IF GIS_NN2 applications_NN2 is_VBZ wide_JJ (_( Gatrell_NP1 1987_MC )_) ._. 
For_REX21 instance_REX22 ,_, Diggle_NP1 et_RA21 al_RA22 ._. 
(_( 1990_MC )_) assess_VV0 the_AT hypothesis_NN1 that_CST cancer_NN1 of_IO the_AT larynx_NN1 is_VBZ associated_VVN with_IW proximity_NN1 to_II a_AT1 now-closed_JJ incinerator_NN1 and_CC develop_VV0 a_AT1 statistical_JJ model_NN1 from_II the_AT theory_NN1 of_IO spatial_JJ point_NN1 processes_NN2 to_TO test_VVI this_DD1 ._. 
This_DD1 would_VM have_VHI been_VBN impossible_JJ without_IW the_AT link_NN1 provided_VVN from_II postcoded_JJ data_NN (_( increasingly_RR available_JJ from_II sources_NN2 such_II21 as_II22 cancer_NN1 registries_NN2 )_) to_II Ordnance_NN1 Survey_NN1 (_( OS_NP2 )_) grid_NN1 references_NN2 ._. 
Such_DA a_AT1 link_NN1 is_VBZ provided_VVN by_II the_AT Central_JJ Postcode_NN1 Directory_NN1 ,_, which_DDQ matches_VVZ all_DB 1.5_MC million_NNO unit_NN1 postcodes_VVZ in_II the_AT UK_NP1 to_II grid_NN1 references_NN2 ._. 
When_CS the_AT cases_NN2 of_IO laryngeal_JJ cancer_NN1 are_VBR compared_VVN with_IW a_AT1 null_JJ distribution_NN1 (_( provided_VVN by_II the_AT distribution_NN1 of_IO much_DA1 more_DAR common_JJ lung_NN1 cancer_NN1 )_) the_AT hypothesis_NN1 is_VBZ given_VVN strong_JJ support_NN1 ._. 
Although_CS the_AT proposed_JJ method_NN1 is_VBZ novel_JJ ,_, it_PPH1 is_VBZ but_CCB one_MC1 of_IO an_AT1 increasingly_RR large_JJ number_NN1 of_IO methods_NN2 designed_VVN to_TO detect_VVI and_CC model_NN1 '_GE clusters_NN2 '_GE of_IO disease_NN1 (_( see_VV0 ,_, for_REX21 example_REX22 ,_, Cuzick_NP1 and_CC Edwards_NP1 1990_MC )_) ._. 
The_AT work_NN1 of_IO Diggle_NP1 et_RA21 al_RA22 ._. 
(_( 1990_MC )_) examined_VVD only_RR a_AT1 single_JJ point_NN1 source_NN1 ,_, though_CS it_PPH1 generalizes_VVZ to_II multiple_JJ sources_NN2 and_CC to_II linear_JJ hazards_NN2 as_RR21 well_RR22 ._. 
As_II a_AT1 second_MD example_NN1 of_IO health_NN1 work_NN1 linked_VVN to_TO hazard_VVI studies_NN2 we_PPIS2 may_VM cite_VVI the_AT research_NN1 being_VBG conducted_VVN by_II Cross_NP1 (_( 1989_MC )_) on_II childhood_NN1 leukaemia_NN1 ._. 
There_EX are_VBR numerous_JJ hypotheses_NN2 concerning_II the_AT aetiology_NN1 of_IO this_DD1 disease_NN1 (_( from_II ionizing_VVG radiation_NN1 to_II viral_JJ transmission_NN1 )_) and_CC Cross_NP1 's_GE approach_NN1 is_VBZ to_TO collect_VVI digital_JJ data_NN designed_VVN to_TO test_VVI some_DD of_IO these_DD2 hypotheses_NN2 and_CC to_TO use_VVI ARC/INFO_NN1 to_TO display_VVI and_CC analyse_VVI results_NN2 ._. 
For_REX21 instance_REX22 ,_, one_MC1 hypothesis_NN1 concerns_VVZ the_AT role_NN1 of_IO electromagnetic_JJ radiation_NN1 ,_, a_AT1 source_NN1 of_IO which_DDQ is_VBZ overhead_JJ high-voltage_JJ transmission_NN1 lines_NN2 as_CSA noted_VVN earlier_RRR ._. 
If_CS these_DD2 are_VBR digitized_VVN and_CC buffer_NN1 zones_NN2 constructed_VVN around_II them_PPHO2 (_( Fig._NN1 10.3_MC )_) we_PPIS2 can_VM quite_RG easily_RR test_VVI hypotheses_NN2 about_II prevalence_NN1 in_II such_DA buffer_NN1 zones_NN2 as_CSA compared_VVN with_IW prevalence_NN1 outside_RL ._. 
However_RR ,_, we_PPIS2 should_VM bear_VVI in_II mind_NN1 the_AT earlier_JJR point_NN1 about_II the_AT resolution_NN1 of_IO the_AT data_NN ,_, since_CS such_DA electromagnetic_JJ effects_NN2 have_VH0 a_AT1 very_RG weak_JJ effect_NN1 beyond_II perhaps_RR 50_MC m_NNU (_( Wertheimer_NP1 and_CC Leeper_NP1 1982_MC )_) ._. 
There_EX has_VHZ been_VBN a_AT1 suspicion_NN1 for_IF many_DA2 years_NNT2 that_CST high_JJ rates_NN2 of_IO stomach_NN1 cancer_NN1 in_II parts_NN2 of_IO Wales_NP1 may_VM be_VBI linked_VVN to_TO soil_VVI chemistry_NN1 ,_, specifically_RR such_DA factors_NN2 as_CSA zinc/copper_NN1 ratios_NN2 ,_, and_CC to_TO lead_VVI content_NN1 in_II water_NN1 supplies_NN2 ._. 
Matthews_NP1 (_( 1989_MC )_) has_VHZ used_VVN data_NN from_II the_AT Regional_JJ Cancer_NN1 Registry_NN1 (_( morbidity_NN1 rather_II21 than_II22 mortality_NN1 data_NN )_) and_CC data_NN on_II trace_NN1 elements_NN2 to_TO examine_VVI this_DD1 link_NN1 ._. 
Since_CS the_AT latter_DA are_VBR available_JJ for_IF a_AT1 5_MC km_NNU raster_NN1 the_AT registration_NN1 data_NN are_VBR assigned_VVN to_II these_DD2 grid_NN1 squares_NN2 using_VVG the_AT OS_NP2 grid_NN1 reference_NN1 ._. 
As_CSA with_IW any_DD such_DA study_NN1 there_EX are_VBR the_AT usual_JJ problems_NN2 of_IO ecological_JJ analysis_NN1 and_CC influences_NN2 of_IO migration_NN1 ,_, together_RL with_IW issues_NN2 of_IO separating_VVG out_RP the_AT risk_NN1 factors_NN2 under_II investigation_NN1 from_II possible_JJ confounding_JJ factors_NN2 ._. 
In_II the_AT USA_NP1 ,_, Morrill_NP1 's_GE report_NN1 on_II the_AT beginnings_NN2 of_IO a_AT1 $15m._NNU study_NN1 into_II the_AT possible_JJ radiation_NN1 doses_NN2 received_VVN as_CSA result_NN1 of_IO radiation_NN1 releases_VVZ by_II the_AT Hanford_NP1 (_( Washington_NP1 State_NN1 )_) nuclear_JJ complex_NN1 would_VM seem_VVI to_TO cry_VVI out_RP for_IF GIS_NN2 skills_NN2 to_TO be_VBI applied_VVN (_( Morrill_NP1 1989_MC )_) ._. 
The_AT research_NN1 project_NN1 will_VM employ_VVI plume_NN1 dispersion_NN1 models_NN2 in_BCL21 order_BCL22 to_TO model_VVI dispersal_NN1 by_II wind_NN1 and_CC will_VM need_VVI to_TO integrate_VVI this_DD1 and_CC other_JJ environmental_JJ data_NN with_IW detailed_JJ demographic_JJ records_NN2 ._. 
To_II some_DD extent_NN1 this_DD1 line_NN1 of_IO approach_NN1 was_VBDZ anticipated_VVN by_II Johnson_NP1 (_( 1981_MC )_) in_II his_APPGE study_NN1 of_IO exposure_NN1 to_II plutonium_NN1 fall-out_NN1 from_II the_AT Rocky_JJ Flats_NN2 nuclear_JJ weapons_NN2 plant_VV0 near_II Denver_NP1 (_( see_VV0 Zeigler_NP1 et_RA21 al_RA22 ._. 
1983_MC :_: 478_MC )_) ._. 
Using_VVG soil_NN1 samples_NN2 he_PPHS1 constructed_VVD a_AT1 contoured_JJ risk_NN1 surface_NN1 and_CC then_RT mapped_VVN excess_JJ cancers_NN2 in_II various_JJ distance_NN1 bands_NN2 ,_, detecting_VVG a_AT1 marked_JJ distance_NN1 decay_NN1 effect_NN1 ._. 
Research_NN1 problems_NN2 and_CC prospects_NN2 The_AT work_NN1 reviewed_VVD above_RL suggests_VVZ that_CST the_AT application_NN1 of_IO GIS_NN2 techniques_NN2 to_TO hazard_VVI assessment_NN1 ,_, emergency_NN1 planning_NN1 and_CC environmental_JJ monitoring_NN1 is_VBZ both_RR actually_RR and_CC potentially_RR an_AT1 important_JJ research_NN1 area_NN1 with_IW many_DA2 challenging_JJ problems_NN2 ._. 
Undoubtedly_RR ,_, the_AT GISs_NN1 of_IO the_AT 1990s_MC2 will_VM look_VVI very_RG different_JJ from_II their_APPGE rather_RG primitive_JJ ancestors_NN2 ._. 
Computer_NN1 power_NN1 is_VBZ becoming_VVG cheaper_JJR ,_, digital_JJ data_NN more_RGR readily_RR available_JJ ,_, and_CC GISs_NP1 are_VBR becoming_VVG hybrid_JJ systems_NN2 involving_VVG other_JJ technologies_NN2 ._. 
In_II this_DD1 last_MD section_NN1 we_PPIS2 highlight_VV0 some_DD important_JJ research_NN1 areas_NN2 as_CSA we_PPIS2 see_VV0 them_PPHO2 ._. 
Doubtless_RR there_EX are_VBR many_DA2 others_NN2 ,_, such_DA is_VBZ the_AT surge_NN1 of_IO interest_NN1 in_II GIS_NN2 technology_NN1 ._. 
Data_NN availability_NN1 We_PPIS2 should_VM first_MD consider_VVI the_AT availability_NN1 and_CC quality_NN1 of_IO data_NN concerning_II the_AT populations_NN2 who_PNQS may_VM be_VBI at_II risk_NN1 from_II the_AT storage_NN1 and_CC transport_NN1 of_IO hazardous_JJ substances_NN2 ,_, from_II toxic_JJ air_NN1 pollution_NN1 ,_, from_II flooding_NN1 and_RR31 so_RR32 on_RR33 ._. 
Such_DA data_NN are_VBR crucial_JJ ,_, either_RR for_IF assisting_VVG the_AT emergency_NN1 services_NN2 to_TO make_VVI assessments_NN2 about_II resources_NN2 they_PPHS2 need_VV0 to_TO deploy_VVI in_II such_DA events_NN2 ,_, or_CC for_IF high-quality_JJ estimates_NN2 to_TO be_VBI made_VVN of_IO risk_NN1 ._. 
Within_II the_AT public_JJ domain_NN1 in_II the_AT UK_NP1 we_PPIS2 must_VM rely_VVI in_RR21 general_RR22 on_II data_NN from_II the_AT most_RGT recent_JJ Population_NN1 Census_NN1 ,_, the_AT lowest_JJT level_NN1 being_VBG that_DD1 for_IF enumeration_NN1 districts_NN2 (_( EDs_NN2 )_) ._. 
As_CSA is_VBZ well_RR known_VVN (_( Rhind_NP1 1983_MC )_) these_DD2 contain_VV0 on_II average_NN1 perhaps_RR 150200_MC households_NN2 and_CC 400500_MC people_NN ._. 
The_AT boundaries_NN2 of_IO EDs_NN2 have_VH0 not_XX been_VBN widely_RR digitized_VVN ,_, unlike_II the_AT higher-level_JJ electoral_JJ wards_NN2 ._. 
However_RR ,_, the_AT centroids_NN2 of_IO EDs_NN2 are_VBR available_JJ as_CSA 100_MC m_NNO grid_NN1 references_NN2 and_CC artificial_JJ ED_NP1 '_GE polygons_NN2 '_GE can_VM be_VBI created_VVN if_CS necessary_JJ ._. 
In_II some_DD instances_NN2 such_DA data_NN will_VM suffice_VVI ;_; in_II many_DA2 cases_NN2 they_PPHS2 will_VM have_VHI to_TO suffice_VVI as_CSA nothing_PN1 better_RRR is_VBZ available_JJ ._. 
But_CCB in_II some_DD areas_NN2 EDs_NN2 are_VBR very_RG extensive_JJ physical_JJ units_NN2 and_CC the_AT shapes_NN2 will_VM be_VBI quite_RG distorted_JJ ._. 
Furthermore_RR ,_, the_AT scales_NN2 at_II which_DDQ population_NN1 estimates_NN2 are_VBR often_RR required_JJ means_NN that_CST even_RR EDs_NN2 are_VBR too_RG coarse_JJ for_IF risk_NN1 assessments_NN2 ._. 
For_REX21 instance_REX22 ,_, in_II performing_VVG its_APPGE quantitative_JJ risk_NN1 assessments_NN2 the_AT Health_NN1 and_CC Safety_NN1 Executive_NN1 will_VM rarely_RR deal_VVI with_IW distances_NN2 from_II hazardous_JJ sites_NN2 in_II31 excess_II32 of_II33 1_MC1 km_NNU (_( Petts_VVZ 1988_MC )_) and_CC census_NN1 data_NN will_VM not_XX offer_VVI sufficient_JJ resolution_NN1 for_IF accurate_JJ population_NN1 estimates_VVZ to_TO be_VBI made_VVN ._. 
It_PPH1 should_VM be_VBI pointed_VVN out_RP ,_, however_RR ,_, that_CST the_AT National_JJ Radiological_JJ Protection_NN1 Board_NN1 currently_RR uses_VVZ 1_MC1 km_NNU grid_NN1 square_NN1 resolution_NN1 population_NN1 data_NN from_II the_AT 1971_MC Census_NN1 in_II its_APPGE radiological_JJ protection_NN1 studies_NN2 (_( Hallam_NP1 et_RA21 al_RA22 ._. 
1981_MC )_) ._. 
It_PPH1 would_VM seem_VVI that_DD1 ,_, despite_II the_AT costs_NN2 involved_VVD ,_, they_PPHS2 should_VM consider_VVI investing_VVG in_II the_AT very_RG detailed_JJ digital_JJ data_NN (_( Pinpoint_VV0 Address_NN1 Code_NN1 (_( PAC_NP1 )_) )_) provided_VVD by_II Pinpoint_JJ Analysis_NN1 Ltd_JJ ._. 
This_DD1 area_NN1 of_IO application_NN1 would_VM appear_VVI to_TO be_VBI an_AT1 obvious_JJ market_NN1 for_IF such_DA a_AT1 product_NN1 ._. 
A_AT1 useful_JJ research_NN1 exercise_NN1 would_VM be_VBI to_TO compare_VVI the_AT use_NN1 of_IO census_NN1 and_CC PAC_NP1 data_NN in_II deriving_VVG quantitative_JJ risk_NN1 assessments_NN2 and_CC this_DD1 is_VBZ an_AT1 interesting_JJ research_NN1 problem_NN1 ._. 
Even_RR with_IW PAC_NP1 data_NN some_DD assumptions_NN2 would_VM need_VVI to_TO be_VBI made_VVN about_II numbers_NN2 of_IO persons_NN2 per_II household_NN1 if_CS good_JJ population_NN1 estimates_NN2 were_VBDR to_TO be_VBI made_VVN ._. 
More_RGR generally_RR ,_, we_PPIS2 need_VV0 to_TO consider_VVI the_AT availability_NN1 of_IO large_JJ databases_NN2 for_IF hazard_NN1 studies_NN2 and_CC emergency_NN1 planning_NN1 and_CC this_DD1 is_VBZ an_AT1 area_NN1 ripe_JJ for_IF major_JJ initiatives_NN2 ._. 
In_II the_AT UK_NP1 we_PPIS2 seem_VV0 to_TO be_VBI lagging_VVG well_RR behind_II the_AT USA_NP1 in_II the_AT construction_NN1 of_IO databases_NN2 on_II hazardous_JJ sites_NN2 and_CC especially_RR hazard_VV0 events_NN2 such_II21 as_II22 those_DD2 involving_VVG toxic_JJ releases_NN2 ._. 
For_REX21 instance_REX22 ,_, Lioy_NP1 and_CC Daisey_NP1 (_( 1987_MC )_) ,_, in_II their_APPGE work_NN1 on_II the_AT monitoring_NN1 of_IO toxic_JJ air_NN1 pollution_NN1 in_II New_NP1 Jersey_NP1 ,_, construct_VV0 what_DDQ they_PPHS2 term_VV0 a_AT1 '_GE microinventory_JJ '_GE for_IF each_DD1 of_IO three_MC towns_NN2 ,_, comprising_VVG the_AT locations_NN2 of_IO chemical_JJ plants_NN2 ,_, metal_NN1 processing_NN1 plants_NN2 ,_, paint_VV0 spray_NN1 manufacturers_NN2 and_RR31 so_RR32 on_RR33 ._. 
Some_DD attribute_NN1 information_NN1 is_VBZ available_JJ for_IF these_DD2 ._. 
More_RGR closely_RR linked_VVN to_II GIS_NN2 is_VBZ work_NN1 by_II McMaster_NP1 and_CC Johnson_NP1 (_( 1986_MC )_) who_PNQS perform_VV0 a_AT1 very_RG detailed_JJ inventory_NN1 of_IO hazardous_JJ materials_NN2 (_( both_DB2 stored_VVN and_CC transported_VVN )_) in_II Santa_NP1 Monica_NP1 ,_, California_NP1 ._. 
Although_CS the_AT raster_NN1 displays_NN2 are_VBR very_RG crude_JJ it_PPH1 is_VBZ simple_JJ to_TO answer_VVI spatial_JJ queries_NN2 concerning_II the_AT locations_NN2 of_IO people_NN (_( for_REX21 instance_REX22 ,_, particular_JJ age_NN1 groups_NN2 )_) within_II certain_JJ distances_NN2 of_IO toxic_JJ hazard_NN1 sites_NN2 ._. 
We_PPIS2 have_VH0 data_NN on_II some_DD hazardous_JJ sites_NN2 (_( e.g._REX those_DD2 regulated_VVN by_II CIMAH_NP1 )_) in_II the_AT UK_NP1 but_CCB a_AT1 much_RR more_RGR comprehensive_JJ approach_NN1 should_VM be_VBI adopted_VVN ,_, especially_RR if_CS we_PPIS2 are_VBR serious_JJ in_II the_AT search_NN1 for_IF environmental_JJ associations_NN2 with_IW chronic_JJ diseases_NN2 ._. 
The_AT same_DA is_VBZ true_JJ for_IF toxic_JJ releases_NN2 ._. 
In_II the_AT USA_NP1 the_AT Environmental_JJ Protection_NN1 Agency_NN1 produces_VVZ an_AT1 Acute_JJ Hazards_NN2 Events_NN2 Database_NN1 while_CS the_AT Department_NN1 of_IO Transportation_NN1 has_VHZ a_AT1 database_NN1 recording_NN1 details_NN2 of_IO 25_MC 000_MC releases_NN2 of_IO toxic_JJ materials_NN2 ;_; such_DA details_NN2 include_VV0 locational_JJ information_NN1 ,_, type_NN1 and_CC amount_NN1 of_IO material_NN1 released_VVN ,_, mode_NN1 of_IO release_NN1 and_CC resulting_JJ injuries_NN2 or_CC deaths_NN2 (_( Cutter_NN1 and_CC Solecki_NP1 1989_MC )_) ._. 
Can_VM we_PPIS2 not_XX use_VVI GIS_NN2 technology_NN1 to_TO construct_VVI '_GE risk_NN1 mosaics_NN2 '_GE (_( Zeigler_NP1 et_RA21 al_RA22 ._. 
1983_MC )_) ,_, at_II a_AT1 variety_NN1 of_IO spatial_JJ scales_NN2 ,_, akin_JJ to_II Hewitt_NP1 and_CC Burton_NP1 's_GE '_GE hazardousness_NN1 '_GE of_IO a_AT1 place_NN1 ?_? 
In_II the_AT UK_NP1 only_RR a_AT1 handful_NN1 of_IO county_NN1 emergency_NN1 planning_NN1 departments_NN2 (_( Cumbria_NP1 ,_, for_REX21 instance_REX22 )_) have_VH0 built_VVN up_RP resource_NN1 databases_NN2 in_II digital_JJ form_NN1 and_CC this_DD1 needs_VVZ to_TO be_VBI promoted_VVN ._. 
Further_RRR afield_RL ,_, we_PPIS2 need_VV0 to_TO recognize_VVI that_CST some_DD disasters_NN2 (_( most_RGT obviously_RR Chernobyl_NP1 )_) will_VM have_VHI implications_NN2 for_IF more_DAR than_CSN one_MC1 country_NN1 so_CS21 that_CS22 environmental_JJ monitoring_NN1 should_VM take_VVI place_NN1 with_IW some_DD kind_NN1 of_IO international_JJ co-ordination_NN1 ._. 
There_EX will_VM also_RR be_VBI movements_NN2 of_IO hazardous_JJ cargoes_NN2 across_II frontiers_NN2 and_CC there_EX are_VBR clearly_RR problems_NN2 of_IO risk_NN1 assessment_NN1 and_CC planning_VVG for_IF emergencies_NN2 at_II or_CC near_II international_JJ boundaries_NN2 ,_, or_CC at_II sea_NN1 and_CC airports_NN2 ._. 
As_CSA noted_VVN above_RL ,_, the_AT work_NN1 of_IO Fedra_NP1 (_( 1989_MC )_) is_VBZ a_AT1 major_JJ step_NN1 in_II this_DD1 direction_NN1 ,_, and_CC links_NN2 need_VV0 to_TO be_VBI forged_VVN between_II initiatives_NN2 in_II the_AT UK_NP1 and_CC the_AT rest_NN1 of_IO Europe_NP1 ._. 
In_II31 terms_II32 of_II33 databases_NN2 on_II hazardous_JJ wastes_NN2 ,_, attention_NN1 has_VHZ already_RR been_VBN drawn_VVN to_II the_AT waste_NN1 disposal_NN1 plans_NN2 prepared_VVN by_II WRAs_NN2 ._. 
These_DD2 give_VV0 details_NN2 of_IO present_JJ site_NN1 locations_NN2 and_CC some_DD indication_NN1 of_IO permitted_JJ wastes_NN2 ,_, though_CS little_JJ or_CC no_AT information_NN1 on_II closed_JJ and_CC abandoned_JJ sites_NN2 ._. 
However_RR ,_, for_IF details_NN2 of_IO disposal_NN1 the_AT individual_JJ site_NN1 licences_NN2 must_VM be_VBI consulted_VVN ._. 
The_AT former_DA Hazardous_JJ Waste_NN1 Inspectorate_NN1 operated_VVN a_AT1 large_JJ database_NN1 on_II all_DB sites_NN2 ,_, with_IW details_NN2 of_IO permitted_JJ wastes_NN2 ,_, including_II allowable_JJ quantities_NN2 ._. 
The_AT database_NN1 was_VBDZ created_VVN by_II a_AT1 commercial_JJ company_NN1 ,_, Aspinwalls_VVZ Ltd_JJ ._. 
There_EX are_VBR numerous_JJ potentially_RR usable_JJ databases_NN2 for_IF environmental_JJ epidemiological_JJ work_NN1 ._. 
OPCS_NP1 ,_, of_RR21 course_RR22 ,_, is_VBZ a_AT1 major_JJ data_NN repository_NN1 for_IF medical_JJ statistics_NN and_CC their_APPGE various_JJ Monitor_NN1 Series_NN are_VBR of_IO use_NN1 ,_, though_CS only_RR at_II the_AT scale_NN1 of_IO district_NN1 health_NN1 authority_NN1 and_CC above_RL (_( see_VV0 Gatrell_NP1 and_CC Lovett_NP1 1988_MC for_IF an_AT1 example_NN1 )_) ._. 
Regional_JJ health_NN1 authorities_NN2 will_VM from_RR41 time_RR42 to_RR43 time_RR44 produce_VVI mortality_NN1 data_NN at_II ward_NN1 level_NN1 ._. 
However_RR ,_, if_CS cancers_NN2 are_VBR of_IO interest_NN1 it_PPH1 is_VBZ more_RGR appropriate_JJ to_TO obtain_VVI data_NN from_II regional_JJ cancer_NN1 registries_NN2 ;_; these_DD2 data_NN are_VBR morbidity_NN1 rather_II21 than_II22 mortality_NN1 data_NN and_CC are_VBR therefore_RR more_RGR valuable_JJ ._. 
The_AT data_NN are_VBR postcoded_VVN and_CC may_VM have_VHI some_DD occupational_JJ details_NN2 ._. 
For_REX21 example_REX22 ,_, we_PPIS2 have_VH0 received_VVN data_NN from_II the_AT Regional_JJ Cancer_NN1 Registry_NN1 in_II Manchester_NP1 comprising_VVG all_DB cancers_NN2 of_IO the_AT larynx_NN1 and_CC lung_NN1 notified_VVN between_II 1974_MC and_CC 1984_MC ._. 
This_DD1 large_JJ data_NN set_VV0 ,_, to_II which_DDQ grid_NN1 references_NN2 have_VH0 now_RT been_VBN added_VVN ,_, comprises_VVZ information_NN1 about_II nearly_RR 40_MC 000_MC individuals_NN2 ._. 
It_PPH1 has_VHZ now_RT been_VBN copied_VVN into_II INGRES_NP1 ,_, a_AT1 relational_JJ database_NN1 management_NN1 system_NN1 ,_, permitting_VVG queries_NN2 to_TO be_VBI made_VVN about_II subsets_NN2 of_IO the_AT data_NN ;_; for_REX21 instance_REX22 ,_, all_DB males_NN2 between_II the_AT ages_NN2 of_IO 25_MC and_CC 64_MC suffering_JJ from_II laryngeal_JJ cancer_NN1 ._. 
Other_JJ data_NN sources_NN2 on_II cancers_NN2 are_VBR independent_JJ registries_NN2 of_IO malignancies_NN2 ._. 
An_AT1 example_NN1 is_VBZ that_CST on_II leukaemias_NN2 ,_, managed_VVN by_II the_AT Leukaemia_NN1 Research_NN1 Fund_NN1 Centre_NN1 in_II Leeds_NP1 (_( Alexander_NP1 et_RA21 al_RA22 ._. 
1989_MC )_) ._. 
It_PPH1 should_VM be_VBI noted_VVN that_CST there_EX are_VBR frequently_RR major_JJ discrepancies_NN2 between_II the_AT regional_JJ registries_NN2 and_CC the_AT specialist_NN1 registries_NN2 ,_, partly_RR because_CS the_AT latter_DA receive_VV0 notifications_NN2 directly_RR from_II consultants_NN2 ._. 
Indeed_RR ,_, working_VVG directly_RR with_IW consultants_NN2 proves_VVZ invaluable_JJ in_II gaining_VVG access_NN1 to_II high-quality_JJ morbidity_NN1 data_NN ._. 
Two_MC other_JJ potential_JJ sources_NN2 of_IO data_NN are_VBR worth_II mentioning_VVG ._. 
First_MD ,_, attention_NN1 should_VM be_VBI drawn_VVN to_II the_AT new_JJ directive_NN1 from_II the_AT Department_NN1 of_IO Health_NN1 to_II regional_JJ and_CC district_NN1 health_NN1 authorities_NN2 to_TO review_VVI the_AT health_NN1 of_IO the_AT population_NN1 ._. 
This_DD1 means_VVZ in_II practice_NN1 that_CST district_NN1 medical_JJ officers_NN2 are_VBR charged_VVN with_IW preparing_VVG an_AT1 annual_JJ report_NN1 ,_, presenting_VVG and_CC interpreting_VVG epidemiological_JJ data_NN ,_, identifying_VVG local_JJ health_NN1 problems_NN2 and_CC evaluating_VVG service_NN1 outcomes_NN2 ._. 
Lancaster_NP1 University_NN1 is_VBZ already_RR heavily_RR involved_JJ in_II the_AT work_NN1 of_IO Preston_NP1 District_NN1 Health_NN1 Authority_NN1 in_II this_DD1 area_NN1 ._. 
Second_MD ,_, family_NN1 health_NN1 service_NN1 authorities_NN2 (_( FHSAs_NP2 )_) ,_, are_VBR also_RR required_VVN to_TO take_VVI on_RP these_DD2 roles_NN2 ,_, but_CCB from_II the_AT viewpoint_NN1 of_IO primary_JJ health_NN1 care_NN1 services_NN2 ._. 
FHSAs_NP2 (_( which_DDQ correspond_VV0 to_II non-metropolitan_JJ counties_NN2 and_CC to_II metropolitan_JJ boroughs_NN2 )_) hold_VV0 large_JJ databases_NN2 on_II services_NN2 provided_VVN by_II general_JJ practitioners_NN2 ;_; for_REX21 instance_REX22 ,_, uptake_NN1 of_IO immunization_NN1 and_CC screening_VVG for_IF breast_NN1 and_CC cervical_JJ cancers_NN2 ._. 
Is_VBZ it_PPH1 ,_, we_PPIS2 wonder_VV0 ,_, possible_JJ to_TO contemplate_VVI bringing_VVG together_RL some_DD of_IO these_DD2 data_NN sources_NN2 to_TO form_VVI a_AT1 National_JJ Online_JJ Health_NN1 Information_NN1 System_NN1 (_( NOHIS_NP1 )_) ,_, to_TO parallel_VVI that_DD1 for_IF employment_NN1 and_CC unemployment_NN1 (_( Townsend_NP1 et_RA21 al_RA22 ._. 
1987_MC )_) ?_? 
This_DD1 might_VM include_VVI data_NN on_II morbidity_NN1 and_CC mortality_NN1 ,_, uptake_NN1 of_IO preventative_JJ medicine_NN1 ,_, service_NN1 use_NN1 and_CC availability_NN1 ,_, at_II both_RR primary_JJ and_CC hospital_NN1 level_NN1 ._. 
It_PPH1 would_VM be_VBI a_AT1 useful_JJ management_NN1 tool_NN1 ,_, particularly_RR in_II31 view_II32 of_II33 the_AT reorganization_NN1 taking_VVG place_NN1 in_II the_AT National_JJ Health_NN1 Service_NN1 ._. 
It_PPH1 should_VM not_XX be_VBI forgotten_VVN that_CST hazards_NN2 themselves_PPX2 have_VH0 geographical_JJ distributions_NN2 yet_RR there_EX has_VHZ been_VBN little_DA1 effort_NN1 in_II the_AT UK_NP1 to_TO develop_VVI public_JJ domain_NN1 databases_NN2 ._. 
Are_VBR there_RL hazard-prone_JJ areas_NN2 ?_? 
Is_VBZ there_EX a_AT1 regional_JJ differentiation_NN1 in_II hazard_NN1 type_NN1 ?_? 
These_DD2 are_VBR ,_, of_RR21 course_RR22 ,_, commercially_RR important_JJ questions_NN2 and_CC a_AT1 good_JJ deal_NN1 of_IO money_NN1 is_VBZ going_VVG into_II the_AT collection_NN1 of_IO such_DA data_NN by_II insurance_NN1 companies_NN2 ._. 
One_PN1 well_RR known_VVN database_NN1 is_VBZ Major_JJ Hazard_NN1 Incident_NN1 Data_NN Service_NN1 (_( MHIDAS_NP2 )_) global_JJ industrial_JJ hazards_NN2 database_NN1 which_DDQ is_VBZ currently_RR being_VBG developed_VVN by_II the_AT Safety_NN1 and_CC Reliability_NN1 Directorate_NN1 on_II31 behalf_II32 of_II33 the_AT Major_NNB Hazards_NP1 Assessment_NN1 Unit_NN1 of_IO the_AT UK_NP1 Health_NN1 and_CC Safety_NN1 Directorate_NN1 ._. 
This_DD1 system_NN1 currently_RR runs_VVZ in_II dBASE_NN1 II_MC and_CC contains_VVZ in_II31 excess_II32 of_II33 3000_MC items_NN2 of_IO professionally_RR assessed_VVN information_NN1 on_II major_JJ hazards_NN2 coded_VVN into_II some_DD 24_MC separate_JJ fields_NN2 ._. 
Unfortunately_RR ,_, the_AT spatial_JJ referencing_NN1 is_VBZ poor_JJ and_CC with_IW some_DD effort_NN1 this_DD1 could_VM easily_RR be_VBI remedied_VVN ._. 
It_PPH1 is_VBZ ,_, for_REX21 example_REX22 ,_, little_RR use_VV0 when_CS recording_VVG a_AT1 chemical_JJ spill_NN1 into_II an_AT1 Egyptian_JJ river_NN1 to_TO have_VHI entered_VVN into_II the_AT appropriate_JJ field_NN1 '_GE the_AT Nile_NP1 '_GE after_II all_DB ,_, this_DD1 river_NN1 is_VBZ some_DD 6480_MC km_NNU long_JJ !_! 
Here_RL ,_, in_II our_APPGE view_NN1 ,_, is_VBZ a_AT1 classic_JJ case_NN1 of_IO the_AT absence_NN1 of_IO a_AT1 geographical_JJ input_NN1 rendering_VVG the_AT data_NN of_IO limited_JJ value_NN1 ._. 
In_II some_DD cases_NN2 ,_, good_JJ hazard_NN1 data_NN exist_VV0 but_CCB have_VH0 simply_RR not_XX yet_RR found_VVD their_APPGE way_NN1 into_II a_AT1 GIS_NN2 ._. 
A_AT1 case_NN1 in_II point_NN1 is_VBZ inland_RL and_CC coastal_JJ flooding_NN1 ._. 
Most_DAT river_NN1 authorities_NN2 have_VH0 detailed_VVN maps_NN2 showing_VVG the_AT geographical_JJ limits_NN2 of_IO floods_NN2 in_II31 relation_II32 to_II33 their_APPGE return_NN1 periods_NN2 ._. 
Within_II the_AT next_MD decade_NNT1 ,_, as_CSA the_AT privatized_JJ water_NN1 companies_NN2 develop_VV0 their_APPGE GIS_NN2 ,_, we_PPIS2 shall_VM see_VVI a_AT1 massive_JJ amount_NN1 of_IO data_NN conversion_NN1 in_II this_DD1 area_NN1 ._. 
This_DD1 development_NN1 ,_, together_RL with_IW the_AT availability_NN1 of_IO digital_JJ terrain_NN1 models_NN2 ,_, will_VM allow_VVI quite_RG sophisticated_JJ flood_NN1 forecasting_VVG models_NN2 to_TO be_VBI developed_VVN within_II a_AT1 GIS_NN2 framework_NN1 ._. 
As_CSA sea-levels_NN2 rise_VV0 in_II the_AT next_MD 2030_MC years_NNT2 due_II21 to_II22 the_AT '_GE greenhouse_NN1 '_GE effect_NN1 ,_, this_DD1 very_RG practical_JJ use_NN1 of_IO GIS_NN2 technology_NN1 is_VBZ likely_JJ to_TO become_VVI commonplace_JJ and_CC of_IO considerable_JJ practical_JJ importance_NN1 ._. 
Spatial_JJ statistics_NN Another_DD1 way_NN1 in_II which_DDQ there_EX needs_VVZ to_TO be_VBI a_AT1 rapprochement_NN1 between_II statistical_JJ methods_NN2 and_CC GIS_NN2 is_VBZ in_II the_AT area_NN1 of_IO spatial_JJ sampling_NN1 and_CC estimation_NN1 from_II spatial_JJ samples_NN2 ._. 
This_DD1 is_VBZ clearly_RR an_AT1 important_JJ issue_NN1 in_II radiation_NN1 monitoring_VVG as_II the_AT discussion_NN1 of_IO RIMNET_NN1 above_RL implies_VVZ ._. 
As_II an_AT1 example_NN1 ,_, Peirson_NP1 (_( 1988_MC )_) has_VHZ reviewed_VVN the_AT evidence_NN1 on_II artificial_JJ radioactivity_NN1 in_II Cumbria_NP1 and_CC shows_VVZ maps_NN2 of_IO the_AT distribution_NN1 of_IO certain_JJ radionuclides_NN2 as_CSA contoured_JJ surfaces_NN2 ._. 
The_AT point_NN1 worth_II making_VVG here_RL is_VBZ that_CST there_EX are_VBR many_DA2 alternative_JJ schemes_NN2 for_IF interpolating_VVG from_II irregular_JJ point_NN1 data_NN (_( Lam_NP1 1983_MC )_) ._. 
One_MC1 method_NN1 in_RR21 particular_RR22 ,_, kriging_VVG (_( now_RT available_JJ as_II part_NN1 of_IO the_AT UNIRAS_NP2 software_NN1 )_) ,_, offers_NN2 not_XX only_RR a_AT1 contoured_JJ surface_NN1 but_CCB also_RR an_AT1 estimate_NN1 of_IO the_AT standard_JJ error_NN1 at_II any_DD point_NN1 on_II the_AT map_NN1 ._. 
Where_CS this_DD1 is_VBZ high_RR it_PPH1 suggests_VVZ that_CST an_AT1 additional_JJ sampling_NN1 station_NN1 might_VM be_VBI set_VVN up_RP ._. 
Such_DA techniques_NN2 are_VBR now_RT standard_JJ in_II environmental_JJ science_NN1 (_( see_VV0 Streit_NN1 1981_MC for_IF an_AT1 application_NN1 to_II rainfall_NN1 monitoring_NN1 )_) and_CC need_VV0 to_TO be_VBI promoted_VVN vigorously_RR in_II hazard_NN1 studies_NN2 (_( Estes_NP1 et_RA21 al_RA22 ._. 
1987_MC )_) ._. 
Related_VVN to_II this_DD1 problem_NN1 is_VBZ that_DD1 of_IO making_VVG population_NN1 estimates_VVZ for_IF new_JJ zones_NN2 that_CST result_VV0 from_II overlay_NN1 operations_NN2 ._. 
We_PPIS2 may_VM want_VVI to_TO count_VVI numbers_NN2 living_VVG within_II plumes_NN2 ,_, circles_NN2 of_IO fixed_JJ radius_NN1 ,_, buffer_NN1 zones_NN2 and_RR31 so_RR32 on_RR33 ._. 
Existing_JJ GIS_NN2 techniques_NN2 usually_RR base_VV0 the_AT estimates_NN2 on_II area-weighted_JJ shares_NN2 ,_, so_CS21 that_CS22 a_AT1 new_JJ polygon_NN1 created_VVN by_II '_GE slicing_JJ '_GE an_AT1 existing_JJ area_NN1 in_II half_DB is_VBZ assigned_VVN half_DB the_AT population_NN1 of_IO that_DD1 source_NN1 unit_NN1 ._. 
This_DD1 may_VM be_VBI quite_RG absurd_JJ in_II some_DD cases_NN2 ,_, notably_RR where_CS population_NN1 is_VBZ spatially_RR clustered_VVN within_II a_AT1 physically_RR large_JJ ED_NN1 ._. 
At_II the_AT NWRRL_NP1 work_NN1 is_VBZ under_RR21 way_RR22 (_( Flowerdew_NP1 and_CC Green_JJ 1989_MC )_) to_TO solve_VVI such_DA problems_NN2 using_VVG the_AT EM_FU algorithm_NN1 and_CC the_AT statistical_JJ modelling_NN1 package_NN1 GLIM_NN1 ,_, which_DDQ has_VHZ been_VBN interfaced_VVN with_IW ARC/INFO_NN1 ._. 
Operations_NN2 research_VV0 methods_NN2 and_CC simulation_NN1 While_CS there_EX is_VBZ some_DD research_NN1 activity_NN1 in_II linking_VVG statistical_JJ models_NN2 to_II GIS_NN2 there_EX would_VM seem_VVI to_TO be_VBI scope_NN1 for_IF further_JJR applications_NN2 of_IO operations_NN2 research_VV0 techniques_NN2 to_II GIS_NN2 problems_NN2 ._. 
Within_II some_DD sophisticated_JJ GIS_NN2 such_II21 as_II22 ARC/INFO_NN1 there_EX are_VBR some_DD useful_JJ algorithms_NN2 (_( such_II21 as_II22 shortest_JJT path_NN1 calculations_NN2 )_) ,_, but_CCB GIS_NN2 developers_NN2 have_VH0 in_RR21 general_RR22 failed_VVN to_TO realize_VVI the_AT natural_JJ links_NN2 between_II GIS_NN2 functions_NN2 and_CC many_DA2 operations_NN2 research_VV0 techniques_NN2 ._. 
Over_II the_AT last_MD 20_MC years_NNT2 or_CC so_RR a_AT1 large_JJ theoretical_JJ literature_NN1 has_VHZ been_VBN developed_VVN to_TO provide_VVI tools_NN2 for_IF the_AT solution_NN1 of_IO many_DA2 problems_NN2 associated_VVN with_IW emergency_NN1 planning_NN1 (_( Kolesar_NP1 1981_MC )_) ._. 
One_MC1 important_JJ class_NN1 of_IO problems_NN2 is_VBZ the_AT selection_NN1 of_IO locations_NN2 for_IF the_AT emergency_NN1 service_NN1 units_NN2 ._. 
The_AT choice_NN1 of_IO locations_NN2 need_VM not_XX be_VBI for_IF permanent_JJ bases_NN2 ,_, but_CCB can_VM also_RR include_VVI dynamic_JJ repositioning_NN1 of_IO units_NN2 as_CSA circumstances_NN2 dictate_VV0 (_( Kolesar_NP1 and_CC Walker_NP1 1974_MC )_) ._. 
In_II any_DD major_JJ emergency_NN1 a_AT1 typical_JJ problem_NN1 is_VBZ the_AT assignment_NN1 of_IO resources_NN2 to_II demands_NN2 ._. 
If_CS resources_NN2 are_VBR held_VVN at_II one_MC1 set_NN1 of_IO locations_NN2 and_CC the_AT demands_NN2 are_VBR at_II another_DD1 and_CC the_AT costs_NN2 of_IO transportation_NN1 are_VBR known_VVN ,_, how_RRQ do_VD0 we_PPIS2 allocate_VVI the_AT resources_NN2 most_RGT efficiently_RR ?_? 
In_II a_AT1 major_JJ emergency_NN1 ,_, for_REX21 example_REX22 ,_, there_EX might_VM be_VBI a_AT1 need_NN1 to_TO supply_VVI blood_NN1 products_NN2 held_VVN at_II regional_JJ blood_NN1 banks_NN2 to_II hospitals_NN2 ._. 
This_DD1 ,_, of_RR21 course_RR22 ,_, is_VBZ the_AT classic_JJ transportation_NN1 problem_NN1 and_CC its_APPGE solution_NN1 is_VBZ well_RR known_VVN ._. 
As_CS31 far_CS32 as_CS33 we_PPIS2 know_VV0 the_AT microcomputer-based_JJ emergency_NN1 response_NN1 system_NN1 devised_VVN by_II Belardo_NP1 et_RA21 al_RA22 ._. 
(_( 1983_MC )_) ,_, is_VBZ the_AT only_JJ GIS-like_JJ software_NN1 to_TO incorporate_VVI this_DD1 algorithm_NN1 ._. 
In_II this_DD1 system_NN1 ,_, the_AT solution_NN1 to_II the_AT transportation_NN1 problem_NN1 is_VBZ shown_VVN graphically_RR on_II the_AT road_NN1 network_NN1 ,_, which_DDQ is_VBZ displayed_VVN on_II the_AT computer_NN1 screen_NN1 ._. 
As_II a_AT1 last_MD example_NN1 of_IO the_AT potential_JJ role_NN1 of_IO operations_NN2 research_VV0 in_II GIS_NN2 and_CC disaster_NN1 management_NN1 we_PPIS2 are_VBR reminded_VVN of_IO the_AT logistic_JJ problems_NN2 that_CST faced_VVD the_AT Peel_NN1 Regional_JJ Police_NN2 Force_NN1 during_II the_AT Mississauga_NP1 evacuation_NN1 mentioned_VVD previously_RR (_( Scanlon_NP1 and_CC Padgham_NP1 1980_MC )_) ._. 
During_II the_AT early_JJ states_NN2 of_IO the_AT phased_JJ evacuation_NN1 the_AT logistical_JJ problem_NN1 facing_VVG the_AT police_NN2 was_VBDZ the_AT street-by-street_JJ warning_NN1 of_IO the_AT population_NN1 to_TO make_VVI ready_JJ for_IF evacuation_NN1 ._. 
Speed_NN1 was_VBDZ of_IO the_AT essence_NN1 and_CC manpower_NN1 was_VBDZ limited_VVN ._. 
Here_RL ,_, then_RT ,_, is_VBZ a_AT1 familiar_JJ operations_NN2 research_VV0 network_NN1 problem_NN1 :_: all_DB roads_NN2 have_VH0 to_TO be_VBI traversed_VVN such_CS21 that_CS22 the_AT journey_NN1 length/time_FU is_VBZ a_AT1 minimum_NN1 ._. 
This_DD1 is_VBZ the_AT so-called_JJ Chinese_JJ postman_NN1 problem_NN1 ,_, which_DDQ provides_VVZ a_AT1 Hamiltonian_JJ circuit_NN1 through_II the_AT arcs_NN2 (_( as_CSA compared_VVN with_IW the_AT travelling_NN1 salesman_NN1 problem_NN1 which_DDQ is_VBZ a_AT1 minimum_JJ route_NN1 through_II the_AT network_NN1 nodes_NN2 )_) ._. 
In_II fact_NN1 ,_, the_AT police_NN2 force_NN1 lacked_VVD access_NN1 to_II any_DD digital_JJ information_NN1 and_CC although_CS the_AT evacuation_NN1 went_VVD off_RP without_IW a_AT1 hitch_NN1 there_EX is_VBZ little_RR doubt_VV0 that_CST a_AT1 GIS_NN2 containing_VVG the_AT Chinese_JJ postman_NN1 algorithm_NN1 could_VM have_VHI provided_VVN efficient_JJ routes_NN2 and_CC also_RR have_VH0 guaranteed_VVN that_CST no_AT road_NN1 was_VBDZ missed_VVN from_II the_AT warning_NN1 sweep_NN1 ._. 
Simulating_VVG the_AT likely_JJ consequences_NN2 of_IO real_JJ events_NN2 is_VBZ an_AT1 important_JJ ,_, but_CCB almost_RR untouched_JJ ,_, GIS_NN2 research_VV0 area_NN1 ._. 
As_II an_AT1 educational_JJ tool_NN1 ,_, simulation_NN1 is_VBZ not_XX only_RR an_AT1 invaluable_JJ training_NN1 aid_NN1 for_IF personnel_NN2 who_PNQS may_VM one_MC1 day_NNT1 be_VBI confronted_VVN with_IW a_AT1 real_JJ situation_NN1 but_CCB is_VBZ sometimes_RT the_AT only_JJ method_NN1 for_IF dealing_VVG with_IW improbable_JJ but_CCB possible_JJ events_NN2 ._. 
In_II the_AT UK_NP1 the_AT Institution_NN1 of_IO Chemical_NN1 Engineers_NN2 has_VHZ produced_VVN a_AT1 computer_NN1 simulation_NN1 program_NN1 specifically_RR for_IF training_NN1 personnel_NN2 at_II large_JJ chemical_JJ plants_NN2 who_PNQS may_VM one_MC1 day_NNT1 have_VHI to_TO manage_VVI major_JJ fires_NN2 and_CC toxic_JJ releases_NN2 (_( Institution_NN1 of_IO Chemical_NN1 Engineers_NN2 ,_, no_AT date_NN1 )_) ._. 
Although_CS not_XX a_AT1 thoroughbred_JJ GIS_NN2 ,_, this_DD1 package_NN1 ,_, based_VVN on_II an_AT1 IBM_NP1 PC_NN1 ,_, does_VDZ make_VVI use_NN1 of_IO map_NN1 data_NN and_CC site_NN1 plans_NN2 can_VM be_VBI displayed_VVN ._. 
Decision_NN1 support_NN1 systems_NN2 and_CC GIS_NN2 A_ZZ1 decision_NN1 support_NN1 system_NN1 (_( DSS_NP1 )_) can_VM be_VBI considered_VVN as_II an_AT1 integration_NN1 of_IO computer_NN1 hardware_NN1 and_CC software_NN1 specifically_RR designed_VVN to_TO complement_VVI the_AT human_JJ thought_NN1 process_NN1 in_II problem-solving_NN1 ,_, decision-making_JJ and_CC information_NN1 processing_NN1 (_( Benbasat_NP1 1977_MC )_) ._. 
According_II21 to_II22 Berke_NP1 and_CC Stubbs_NP1 (_( 1989_MC )_) a_AT1 DSS_NP1 can_VM often_RR be_VBI conceptualized_VVN as_II a_AT1 tool_NN1 to_TO be_VBI used_VVN as_II part_NN1 of_IO an_AT1 interactive_JJ learning_NN1 process_NN1 allowing_VVG the_AT user_NN1 to_TO undertake_VVI '_" what_DDQ if_CS '_GE analyses_NN2 and_CC view_VV0 the_AT consequences_NN2 of_IO such_DA alternatives_NN2 ._. 
The_AT basic_JJ components_NN2 of_IO a_AT1 DSS_NP1 comprise_VV0 :_: data_NN storage_NN1 files_NN2 ;_; data_NN analysis_NN1 models_NN2 ;_; display_VV0 and_CC interactive_JJ use_NN1 technology_NN1 ._. 
These_DD2 three_MC components_NN2 are_VBR managed_VVN by_II :_: a_AT1 data_NN management_NN1 subsystem_NN1 ;_; a_AT1 model_NN1 management_NN1 subsystem_NN1 ;_; and_CC an_AT1 interactive_JJ dialogue_NN1 subsystem_NN1 ._. 
Berke_VV0 and_CC Stubbs_NP1 indicate_VV0 that_CST the_AT interactive_JJ dialogue_NN1 subsystem_NN1 and_CC the_AT display_NN1 and_CC interactive_JJ use_NN1 component_NN1 are_VBR particularly_RR critical_JJ for_IF the_AT effective_JJ use_NN1 of_IO a_AT1 DSS_NP1 because_CS they_PPHS2 provide_VV0 the_AT interaction_NN1 between_II the_AT user_NN1 and_CC the_AT machine_NN1 ._. 
These_DD2 features_NN2 isolate_VV0 the_AT user_NN1 from_II the_AT technicalities_NN2 of_IO the_AT computer_NN1 and_CC encourage_VV0 a_AT1 dialogue_NN1 based_VVN on_II the_AT user_NN1 's_GE judgements_NN2 rather_II21 than_II22 imposing_VVG the_AT hardware_NN1 engineer_NN1 's_GE or_CC computer_NN1 programmer_NN1 's_GE discipline_NN1 on_II the_AT user_NN1 ._. 
In_II this_DD1 sense_NN1 ARC/INFO_NN1 macros_NN2 ,_, do_VD0 ,_, to_II a_AT1 certain_JJ extent_NN1 ,_, shield_VV0 the_AT novice_NN1 from_II a_AT1 bewildering_JJ number_NN1 of_IO options_NN2 ,_, though_CS such_DA a_AT1 system_NN1 is_VBZ most_RGT certainly_RR not_XX conversational_JJ and_CC in_II several_DA2 respects_NN2 is_VBZ very_RG simple_JJ as_CSA compared_VVN with_IW a_AT1 comprehensive_JJ DSS_NP1 ._. 
The_AT integration_NN1 of_IO GIS_NN2 and_CC DSS_NP1 for_IF hazard_NN1 mitigation_NN1 is_VBZ an_AT1 active_JJ research_NN1 area_NN1 in_II the_AT USA_NP1 ,_, Japan_NP1 and_CC Europe_NP1 ._. 
Dong_NP1 et_RA21 al_RA22 ._. 
(_( 1988_MC )_) ,_, for_REX21 example_REX22 ,_, have_VH0 developed_VVN microcomputer-based_JJ DSS_NP1 for_IF earthquake_NN1 hazard_NN1 mitigation_NN1 ._. 
This_DD1 system_NN1 provides_VVZ a_AT1 map_NN1 display_NN1 of_IO the_AT threatened_JJ region_NN1 and_CC access_NN1 to_II spatially_RR referenced_VVD data_NN to_TO facilitate_VVI hazard_NN1 zoning_NN1 for_IF a_AT1 given_JJ earthquake_NN1 magnitude_NN1 ._. 
In_RR21 addition_RR22 ,_, the_AT software_NN1 calculates_VVZ damage_NN1 estimates_NN2 and_CC building_NN1 repair_NN1 costs_VVZ for_IF any_DD geographical_JJ area_NN1 ._. 
Fedra_NN1 and_CC Reitsma_NP1 (_( 1989_MC )_) describe_VV0 a_AT1 DSS-GIS_NN2 developed_VVN by_II the_AT Advanced_JJ Computer_NN1 Applications_NN2 group_NN1 at_II the_AT International_JJ Institute_NN1 for_IF Applied_JJ Systems_NN2 Analysis_NN1 (_( IIASA_NP1 )_) ,_, Austria_NP1 ,_, which_DDQ provides_VVZ an_AT1 interactive_JJ ,_, graphics-oriented_JJ framework_NN1 and_CC post-processor_NN1 for_IF the_AT risk_NN1 assessment_NN1 package_NN1 SAFETI_NN2 (_( Technica_NP1 1984_MC )_) ._. 
Raw_JJ data_NN such_II21 as_II22 plant_NN1 locations_NN2 ,_, weather_NN1 data_NN and_CC population_NN1 distribution_NN1 can_VM be_VBI displayed_VVN as_CSA overlays_VVZ on_II a_AT1 basic_JJ land-use_NN1 map_NN1 ._. 
The_AT graphical_JJ interface_NN1 provides_VVZ a_AT1 link_NN1 to_II SAFETI_NP1 's_GE databases_NN2 and_CC consequence_NN1 modelling_NN1 and_CC risk_NN1 estimates_NN2 and_CC risk_NN1 contours_NN2 can_VM be_VBI produced_VVN ._. 
Expert_NN1 systems_NN2 are_VBR likely_JJ to_TO have_VHI a_AT1 significant_JJ impact_NN1 on_II DSS_NP1 and_CC GIS_NN2 in_II hazard_NN1 management_NN1 (_( Fedra_NP1 and_CC Reitsma_NP1 1989_MC )_) ._. 
Important_JJ characteristics_NN2 of_IO expert_NN1 systems_NN2 are_VBR that_CST they_PPHS2 provide_VV0 advice_NN1 in_II problem-solving_NN1 based_VVN on_II the_AT knowledge_NN1 of_IO experts_NN2 ,_, facilitate_VV0 learning_VVG through_II experience_NN1 and_CC allow_VV0 the_AT use_NN1 of_IO natural_JJ language_NN1 processing_NN1 ._. 
One_MC1 of_IO the_AT most_RGT advanced_JJ applications_NN2 of_IO expert_NN1 systems_NN2 to_II DSS_NP1 and_CC GIS_NN2 is_VBZ a_AT1 hazardous_JJ substances_NN2 and_CC industrial_JJ risk_NN1 management_NN1 system_NN1 (_( IRIMS_VVZ Ispra_NP1 Risk_NN1 Management_NN1 Support_NN1 System_NN1 )_) developed_VVD at_II IIASA_NP1 in_II31 conjunction_II32 with_II33 the_AT Joint_JJ Research_NN1 Centre_NN1 of_IO the_AT Commission_NN1 of_IO the_AT European_JJ Communities_NN2 at_II Ispra_NP1 ,_, Italy_NP1 (_( Fedra_NP1 et_RA21 al_RA22 ._. 
1987_MC ;_; Fedra_NP1 1986_MC )_) ._. 
This_DD1 large_JJ system_NN1 ,_, designed_VVN for_IF a_AT1 SUN_NN1 workstation_NN1 has_VHZ the_AT following_JJ main_JJ elements_NN2 :_: an_AT1 intelligent_JJ user_NN1 interface_NN1 ;_; an_AT1 information_NN1 system_NN1 including_II knowledge_NN1 bases_NN2 ,_, databases_NN2 ,_, inference_NN1 machine_NN1 and_CC database_NN1 management_NN1 system_NN1 ;_; a_AT1 simulation_NN1 system_NN1 ;_; a_AT1 DSS_NP1 ._. 
The_AT large_JJ geographical_JJ database_NN1 held_VVN in_II the_AT system_NN1 covers_VVZ Europe_NP1 as_RG far_RR east_ND1 as_CSA the_AT Urals_NP2 ._. 
Roads_NN2 ,_, railways_NN2 ,_, lakes_NN2 ,_, rivers_NN2 ,_, major_JJ settlements_NN2 ,_, chemical_JJ storage_NN1 facilities_NN2 and_CC political_JJ boundaries_NN2 can_VM be_VBI displayed_VVN and_CC used_JJ as_CSA overlays_VVZ for_IF the_AT various_JJ modules_NN2 that_CST examine_VV0 the_AT impacts_NN2 of_IO chemical_NN1 spills_VVZ into_II the_AT atmosphere_NN1 and_CC river_NN1 systems_NN2 ._. 
IRIMS_NN2 uses_VVZ a_AT1 number_NN1 of_IO expert_NN1 systems_NN2 techniques_NN2 including_II :_: a_AT1 language_NN1 input_NN1 parser_NN1 based_VVN on_II sideways_RL chaining_VVG and_CC rule_NN1 values_NN2 which_DDQ allow_VV0 Bayesian_JJ probabilistic_JJ reasoning_NN1 to_TO identify_VVI possible_JJ user_NN1 intentions/hypotheses_NN2 ;_; fuzzy_JJ set_NN1 methods_NN2 to_TO translate_VVI uncertainty_NN1 and_CC ambiguity_NN1 in_II the_AT databases_NN2 or_CC user_NN1 specifications_NN2 into_II linguistic_JJ or_CC graphical_JJ descriptions_NN2 ;_; and_CC various_JJ rule-based_JJ pre-_JJ and_CC post-processors_NN2 to_TO define_VVI appropriate_JJ context-dependent_JJ default_NN1 input_NN1 values_NN2 ._. 
Even_RR in_II this_DD1 brief_JJ description_NN1 of_IO DSS_NP1 it_PPH1 can_VM be_VBI see_VV0 that_CST ,_, as_CSA presently_RR configured_VVN ,_, most_DAT GIS_NN2 are_VBR rather_RG dumb_JJ systems_NN2 requiring_VVG intelligent_JJ ,_, very_RG computer-literate_JJ users_NN2 ._. 
But_CCB in_II an_AT1 emergency_NN1 or_CC disaster_NN1 situation_NN1 ,_, the_AT knowledge_NN1 base_NN1 must_VM be_VBI in_II the_AT computer_NN1 system_NN1 and_CC there_EX should_VM be_VBI no_AT skill_NN1 barrier_NN1 to_TO access_VVI the_AT system_NN1 ._. 
In_II this_DD1 respect_NN1 we_PPIS2 believe_VV0 that_CST the_AT next_MD generation_NN1 of_IO GIS_NN2 will_VM be_VBI more_RGR akin_JJ to_II DSS_NP1 than_CSN to_II sophisticated_JJ mapping_NN1 packages_NN2 ,_, as_CSA is_VBZ presently_RR the_AT case_NN1 ._. 
The_AT development_NN1 of_IO such_DA GIS_NN2 will_VM surely_RR be_VBI at_II the_AT forefront_NN1 of_IO any_DD research_NN1 agenda_NN1 ._. 
Real-time_JJ GIS_NN2 The_AT ultimate_JJ test_NN1 of_IO the_AT role_NN1 of_IO GIS_NN2 in_II emergency/hazard_NN1 situations_NN2 is_VBZ their_APPGE response_NN1 time_NNT1 as_CSA external_JJ information_NN1 is_VBZ fed_VVN into_II the_AT system_NN1 ._. 
Two_MC approaches_NN2 can_VM be_VBI adopted_VVN to_TO minimize_VVI the_AT response_NN1 time_NNT1 :_: a_AT1 simple_JJ approach_NN1 is_VBZ to_TO devise_VVI portable_JJ GIS_NN2 which_DDQ can_VM be_VBI taken_VVN to_II the_AT hazard_NN1 ;_; a_AT1 more_RGR complex_JJ solution_NN1 is_VBZ to_TO couple_VVI the_AT GIS_NN2 to_II real-time_JJ monitoring_NN1 systems_NN2 ._. 
Probably_RR the_AT outstanding_JJ example_NN1 of_IO the_AT portability_NN1 approach_NN1 is_VBZ the_AT USA_NP1 's_GE National_JJ Oceanic_JJ and_CC Atmospheric_JJ Administration_NN1 's_GE (_( NOAA_NP1 )_) award-winning_JJ emergency_NN1 response_NN1 package_NN1 called_VVN CAMEO_NP1 (_( NOAA_NP1 1988_MC )_) ._. 
The_AT best_JJT implementation_NN1 of_IO CAMEO_NP1 runs_VVZ on_II an_AT1 Apple_NN1 Macintosh_NP1 and_CC makes_VVZ full_JJ use_NN1 of_IO the_AT Hypercard_NP1 environment_NN1 ._. 
The_AT system_NN1 comes_VVZ with_IW three_MC components_NN2 :_: CODEBREAKER_NN1 which_DDQ uses_VVZ intelligent_JJ software_NN1 to_TO identify_VVI hazardous_JJ chemical_JJ labels_NN2 ,_, summarize_VV0 the_AT chemical_JJ properties_NN2 and_CC indicate_VV0 health_NN1 risks_NN2 ;_; MAP_NN1 which_DDQ can_VM display_VVI digital_JJ and_CC scanned_VVD maps_NN2 of_IO the_AT incident_NN1 site_NN1 ;_; AIR_NN1 MODEL_NN1 which_DDQ is_VBZ an_AT1 atmospheric_JJ modelling_NN1 program_NN1 linked_VVN to_II CODEBREAKER_NN1 and_CC MAP_NN1 ._. 
AIR_NN1 MODEL_NN1 also_RR has_VHZ the_AT facility_NN1 to_TO link_VVI on-line_JJ to_TO weather_VVI data_NN sources_NN2 and_CC can_VM estimate_VVI the_AT scope_NN1 of_IO potential_JJ downwind_JJ hazard_NN1 zones_NN2 ._. 
CAMEO_NP1 is_VBZ designed_VVN to_TO be_VBI carried_VVN on_RP emergency_NN1 vehicles_NN2 and_CC has_VHZ undergone_VVN successful_JJ on-scene_JJ field_NN1 trials_NN2 with_IW the_AT Seattle_NP1 Fire_NN1 Department_NN1 ._. 
An_AT1 alternative_JJ approach_NN1 to_II real-time_JJ monitoring_NN1 is_VBZ to_TO develop_VVI software_NN1 capable_JJ of_IO receiving_VVG data_NN from_II external_JJ monitoring_NN1 systems_NN2 ._. 
Southworth_NP1 et_RA21 al_RA22 ._. 
(_( 1989_MC )_) describe_VV0 an_AT1 operational_JJ ,_, prototype_NN1 Real-time_NN1 Traffic_NN1 Monitoring_NN1 and_CC Analysis_NN1 System_NN1 (_( RTMAS_NP2 )_) whose_DDQGE purpose_NN1 is_VBZ to_TO record_VVI and_CC warn_VVI of_IO the_AT build-up_NN1 of_IO significant_JJ urban_JJ population_NN1 evacuation_NN1 in_II times_NNT2 of_IO threat_NN1 ._. 
RTMAS_NN2 integrates_VVZ three_MC software_NN1 components_NN2 :_: GURU_NN1 ,_, an_AT1 expert_NN1 system_NN1 ;_; AUTOBOX_NP1 ,_, a_AT1 time-series_JJ module_NN1 ;_; CROSSTALK_VV0 proprietary_JJ software_NN1 for_IF transferring_VVG traffic_NN1 count_NN1 data_NN from_II roadside_NN1 counters_NN2 to_II the_AT IBM_NP1 PS2_FO ._. 
In_II tracking_VVG an_AT1 evacuation_NN1 build-up_NN1 ,_, RTMAS_NN2 collects_VVZ and_CC cumulates_VVZ net_JJ traffic_NN1 movements_NN2 ,_, building_VVG up_RP a_AT1 quantitative_JJ picture_NN1 of_IO the_AT timing_NN1 and_CC directions_NN2 of_IO evacuation_NN1 ._. 
RTMAS_NP2 has_VHZ been_VBN developed_VVN for_IF an_AT1 IBM_NP1 PS2/Model_NN1 80_MC connected_VVN via_II modem_NN1 to_II software_NN1 initiated_VVN by_II direct_JJ dial_NN1 up_II21 to_II22 a_AT1 sequence_NN1 of_IO traffic_NN1 counters_NN2 ._. 
Currently_RR ,_, the_AT system_NN1 monitors_VVZ traffic_NN1 movements_NN2 in_II Florida_NP1 and_CC Georgia_NP1 but_CCB ultimately_RR national_JJ coverage_NN1 could_VM be_VBI achieved_VVN ._. 
Systems_NN2 such_II21 as_II22 RTMAS_NN2 have_VH0 great_JJ practical_JJ potential_NN1 in_II those_DD2 emergency_NN1 situations_NN2 where_RRQ there_EX is_VBZ likely_JJ to_TO be_VBI self-evacuation_JJ by_II the_AT public_NN1 ._. 
For_REX21 example_REX22 ,_, there_EX were_VBDR major_JJ traffic_NN1 congestion_NN1 problems_NN2 ,_, as_II a_AT1 result_NN1 of_IO self-evacuation_NN1 ,_, during_II Hurricane_NN1 Hugo_NP1 which_DDQ threatened_VVD large_JJ parts_NN2 of_IO the_AT south-west_ND1 of_IO the_AT USA_NP1 in_II October_NPM1 1989_MC ._. 
In_II a_AT1 similar_JJ vein_NN1 ,_, Martin_NP1 (_( 1989_MC )_) describes_VVZ work_NN1 by_II Marconi_NP1 Defence_NN1 Systems_NN2 which_DDQ integrates_VVZ an_AT1 automatic_JJ vehicle_NN1 location_NN1 system_NN1 with_IW Marconi_NP1 's_GE GIS_NN2 (_( called_VVN Tactic_NN1 Plus_NN1 )_) ._. 
Emergency_NN1 vehicles_NN2 can_VM be_VBI supplied_VVN with_IW computers_NN2 that_CST track_VV0 their_APPGE spatial_JJ location_NN1 ._. 
This_DD1 information_NN1 can_VM be_VBI continuously_RR monitored_VVN by_II a_AT1 receiving_JJ station_NN1 and_CC then_RT delivered_VVN to_II the_AT map_NN1 component_NN1 of_IO the_AT GIS_NN2 ._. 
What_DDQ seems_VVZ to_TO be_VBI a_AT1 very_RG advanced_JJ GIS_NN2 with_IW real-time_JJ monitoring_NN1 capacity_NN1 for_IF use_NN1 in_II emergency_NN1 planning_NN1 has_VHZ been_VBN recently_RR developed_VVN by_II Plessey_NP1 Defence_NN1 Systems_NN2 ._. 
The_AT software_NN1 ,_, called_VVN GENERICS_NP1 ,_, is_VBZ from_II the_AT same_DA stable_JJ as_CSA Plessey_NP1 's_GE defence_NN1 and_CC military_JJ research_NN1 ._. 
Quite_RG obviously_RR ,_, there_EX is_VBZ a_AT1 good_JJ deal_NN1 of_IO ,_, as_RR21 yet_RR22 ,_, classified_JJ technology_NN1 which_DDQ might_VM one_MC1 day_NNT1 be_VBI available_JJ for_IF all_DB sorts_NN2 of_IO hazard_NN1 mitigation_NN1 purposes_NN2 ._. 
Conclusions_NN2 It_PPH1 is_VBZ useful_JJ to_TO return_VVI to_II the_AT earlier_JJR typology_NN1 of_IO hazards_NN2 (_( Fig._NN1 10.1_MC )_) and_CC to_TO stress_VVI the_AT links_NN2 that_CST might_VM be_VBI made_VVN between_II those_DD2 working_VVG on_II natural_JJ hazards_NN2 and_CC those_DD2 working_VVG in_II the_AT socio-economic_JJ domains_NN2 ._. 
There_EX are_VBR several_DA2 examples_NN2 of_IO natural_JJ hazards_NN2 whose_DDQGE impact_NN1 might_VM benefit_VVI from_II a_AT1 GIS_NN2 perspective_NN1 that_CST cuts_VVZ across_II the_AT science-social_JJ science_NN1 divide_NN1 ._. 
These_DD2 include_VV0 snow_NN1 and_CC ice_NN1 hazards_NN2 ,_, where_CS impacts_NN2 on_II transport_NN1 and_CC the_AT need_NN1 to_TO plan_VVI road_NN1 salting_VVG or_CC gritting_VVG strategies_NN2 are_VBR obvious_JJ (_( see_VV0 Perry_NP1 et_RA21 al_RA22 ._. 
1986_MC for_IF the_AT beginnings_NN2 of_IO a_AT1 GIS_NN2 approach_NN1 )_) ._. 
It_PPH1 would_VM be_VBI foolish_JJ to_TO imagine_VVI that_CST GIS_NN2 can_VM assist_VVI in_II all_DB hazard_NN1 studies_NN2 ,_, emergencies_NN2 and_CC disasters_NN2 ._. 
It_PPH1 is_VBZ hard_JJ to_TO imagine_VVI what_DDQ role_NN1 it_PPH1 could_VM have_VHI played_VVN in_II the_AT prevention_NN1 or_CC aftermath_NN1 of_IO events_NN2 such_II21 as_II22 the_AT Piper_NP1 Alpha_NN1 oil_NN1 platform_NN1 explosion_NN1 in_II the_AT North_NP1 Sea_NNL1 ,_, the_AT Zeebrugge_NP1 ferry_NN1 disaster_NN1 or_CC the_AT sinking_NN1 of_IO the_AT Marchioness_NN1 pleasure_NN1 boat_NN1 on_II the_AT Thames_NP1 ,_, for_REX21 instance_REX22 ._. 
But_CCB it_PPH1 might_VM have_VHI assisted_VVN in_II suggesting_VVG search_NN1 areas_NN2 for_IF victims_NN2 of_IO the_AT Lockerbie_NP1 air_NN1 crash_NN1 and_CC ,_, at_II a_AT1 very_RG different_JJ scale_NN1 ,_, in_II helping_VVG firemen_NN2 navigate_VVI their_APPGE way_NN1 around_II King_NP1 's_GE Cross_NN1 underground_JJ rail_NN1 station_NN1 ._. 
Why_RRQ ,_, for_REX21 instance_REX22 ,_, could_VM there_EX not_XX be_VBI a_AT1 library_NN1 of_IO scanned_JJ plans_NN2 of_IO all_DB underground_JJ stations_NN2 ,_, showing_VVG the_AT detailed_JJ topography_NN1 and_CC assisting_VVG fire_NN1 personnel_NN2 unfamiliar_JJ with_IW the_AT site_NN1 to_TO direct_VVI their_APPGE efforts_NN2 in_II the_AT search_NN1 for_IF casualties_NN2 ?_? 
In_RR21 general_RR22 ,_, however_RR ,_, GIS_NN2 surely_RR has_VHZ a_AT1 function_NN1 to_TO perform_VVI in_II simulation_NN1 and_CC role-playing_NN1 exercises_NN2 ._. 
The_AT Cumbria_NP1 emergency_NN1 planning_NN1 system_NN1 outlined_VVN above_RL can_VM ,_, for_REX21 instance_REX22 ,_, be_VBI used_VVN as_II an_AT1 experimental_JJ tool_NN1 ,_, permitting_VVG police_NN2 and_CC emergency_NN1 planners_NN2 to_TO simulate_VVI a_AT1 hazard_NN1 event_NN1 and_CC to_TO call_VVI up_RP information_NN1 on_II resources_NN2 ,_, to_TO identify_VVI possible_JJ areas_NN2 for_IF evacuation_NN1 ,_, to_TO plan_VVI routes_NN2 through_II the_AT transport_NN1 network_NN1 ,_, and_RR31 so_RR32 on_RR33 ._. 
There_EX must_VM be_VBI many_DA2 opportunities_NN2 for_IF user-friendly_JJ GIS_NN2 to_TO be_VBI employed_VVN in_II these_DD2 kinds_NN2 of_IO scenarios_NN2 ,_, either_RR as_CSA special-purpose_JJ systems_NN2 or_CC by_II employing_VVG proprietary_JJ GIS_NN2 packages_NN2 with_IW easy-to-use_JJ macros_NN2 ._. 
Anyone_PN1 working_VVG in_II the_AT GIS_NN2 research_VV0 field_NN1 is_VBZ conscious_JJ of_IO the_AT links_NN2 that_CST need_VV0 to_TO be_VBI made_VVN to_II groups_NN2 of_IO researchers_NN2 in_II other_JJ disciplines_NN2 ._. 
Nowhere_RL ,_, it_PPH1 seems_VVZ ,_, is_VBZ this_RG more_RGR apparent_JJ than_CSN in_II the_AT area_NN1 under_II consideration_NN1 here_RL ._. 
We_PPIS2 need_VV0 the_AT skills_NN2 of_IO environmental_JJ scientists_NN2 ,_, statisticians_NN2 ,_, epidemiologists_NN2 and_CC chemical_JJ engineers_NN2 ,_, not_XX to_TO mention_VVI those_DD2 of_IO the_AT geographer_NN1 and_CC planner_NN1 ._. 
Given_VVN the_AT spate_NN1 of_IO recent_JJ disasters_NN2 ,_, notably_RR in_II the_AT UK_NP1 the_AT time_NNT1 is_VBZ ripe_JJ for_IF a_AT1 major_JJ initiative_NN1 in_II the_AT field_NN1 of_IO emergency_NN1 planning_NN1 and_CC hazard_NN1 studies_NN2 ._. 
Acknowledgements_NN2 The_AT Economic_JJ and_CC Social_JJ Research_NN1 Council_NN1 is_VBZ thanked_VVN for_IF its_APPGE promotion_NN1 of_IO GIS_NN2 research_NN1 and_CC for_IF funding_VVG the_AT NWRRL_NP1 at_II Lancaster_NP1 University_NN1 ._. 
