new developments in geographical information …gisteac/gis_book_abridged/...we are delighted that...

1 INTRODUCTION

We are delighted that the publisher of GeographicalInformation Systems: Principles, Techniques,Management and Applications (the ‘Big Book’ of GIS)has decided to publish this abridged edition of theoriginal two-volume set. Since its publication in 1999,we believe that the title has become established as thepremier reference compendium in its field, continuinga tradition begun with Geographical InformationSystems: Principles and Applications in 1991. Bothtitles, we hope, are part of the history of GIS: asencyclopaedic reference works they have providedreliable reference points during the evolution of thesubject; and as thoroughly cross-referenced essays,commissioned from leading authors, we think theystill bear comparison with the best of all theencyclopaedias and handbooks that are nowappearing in an increasingly crowded market place.

Our motivations for producing this abridgedversion are straightforward and simple. Followingpublication of this major work in 1999, we producedwhat we believe was an equally successful textbook,designed as an advanced introductory guide to GIS –the marketplace seemed to think this too, for thattitle, Geographic Information Systems and Science,sold over 25,000 copies in the three years followingits publication. In the meantime, the ‘Big Book’experienced a sustained level of sales to libraries andresearch organisations, and in 2004 a successfulChinese language version was published. We thoughtof the textbook (Longley, Goodchild, Maguire andRhind 2001) as providing a primer to the fast-developing field of GIS, while the ‘Big Book’ would

remain a key reference point that would reinforce itskey concepts, techniques and management practices.Fundamental to our view of GIS is that principlesand best practices are enduring, while the moreephemeral issues that pervade more techno-centricguides are not. Thus when the success of thetextbook led the publisher to ask us for an expandedSecond Edition (Longley, Goodchild, Maguire andRhind 2005), we were keen to maintain a link withthe research compendium that has been formative toour own understanding of GIS.

We were also keen that the book reach the widestpossible audience – for example, individuals whowould use the material to provide backgroundinformation while undertaking courses in GISeducation; those working in institutions that are newto GIS and might not have access to a copy of theoriginal boxed set in their libraries; and smallorganisations for whom the pricing of the originalwork was prohibitive. We are very pleased that thepublisher, John Wiley and Sons, Inc., was amenableto the view that their substantial initial investment inthe second edition of ‘Big Book’ had been largelyamortised, and that a low-cost abridged versioncould be published. We should add that we aregrateful that similar magnanimity underlay theiragreement to post those chapters of GeographicalInformation Systems: Principles and Applications (thefirst edition of the work: Maguire, Goodchild andRhind 1991) on the Website of the textbook(www.wiley.com/go/longley), some years before freedownloads became widespread. The economics ofthis project have nevertheless dictated that we reducethe number of chapters that are available in printform – although all of the contents of the originalwork are retained on the CD that is provided withthe book.

ix

Preface to the Abridged Edition

New Developments in Geographical InformationSystems: Principles, Techniques, Management

and ApplicationsP A LONGLEY, M F GOODCHILD, D J MAGUIRE, D W RHIND1

1 Incorporating written suggestions of many of thecontributors to the second ‘Big Book’ of GIS.

03_735450_fpref.qxd 2/24/05 2:17 PM Page ix

GIS is, without doubt, a fast-moving field, albeitone that is centred upon principles, techniques andmanagement practices that are less transient than itssoftware or its technology. We were guided by threecriteria in selecting the chapters that would be retainedin print form. Our first criterion was the relevance ofthe material that each retains. Much has happenedsince this work was first published in 1999, and itwould be foolish to suggest that the practices ofactually doing GIS have also not developed andchanged in recent years. Because the use of GIS is nowso pervasive, it is also the case that GIS is rarelythought of as a novel solution. For this reason, wehave retained the original Applications part of thebook in the CD-ROM only. Interested readers areinvited to consult the Applications boxes in ourtextbook (Longley, Goodchild, Maguire and Rhind2005) for recent case studies that we think exemplifyparticularly new or novel characteristics of real-worldapplications. Our second criterion concerned thebalance of this abridged edition: we wished to retainmaterial that was representative of all of the parts andsections in all of the remaining three parts (Principles,Techniques and Management) of the original. Thethird criterion was the degree to which the materialdirectly reinforces that discussed in the second editionof Geographic Information Systems and Science – inorder to reinforce our own view that the two workscan be viewed as complementary. Thus the reader willfind that nearly all of the chapters included in thisabridged edition are cited in the Guides to FurtherReading that are at the chapter ends in the Longley,Goodchild, Maguire and Rhind (2005) textbook.

In most all instances, we believe that the textbookenables the full benefits of this reference work to beunlocked. In the remainder of this extended update,however, we provide some additional observationsthat help to update the chapters in the original.These have been provided with the very significantassistance of the authors of many of the originalchapters. We are very grateful, therefore, to thefollowing for their inputs: Prue Adler, Luc Anselin,Richard Aspinall, Kate Beard, Yvan Bédard, TorBernhardsen, Antonio Câmara, Heather Campbell,David Coleman, Michael Curry, Peter Dale, BillDavenhall, Ian Dowman, Sue Elshaw Thrall, RobertFincham, Manfred Fischer, Peter Fisher, Pip Forer,Greg Forsyth, Art Getis, Steve Guptill, GerardHeuvelink, Ron Johnston, Russ Johnson, Menno-Jan Kraak, Werner Kuhn, Art Lange, MaryLarsgaard, Robin McLaren, Dave MacDevette, Jeff

Meyers, Helena Mitasova, Nancy Obermeyer, Petervan Oosterom, John Pickles, Mark Sondheim, DavidSwann, Grant Thrall, David Unwin, Nigel Walters,Rob Weibel and Mike Worboys.

2 PRINCIPLES (PART ONE)

(a) Space and time in GIS

A significant proportion of the opening part of thework focuses upon the disciplinary setting of GIS. Itis interesting that most of the recent handbooks andencyclopaedias on GIS pay rather less attention toissues of spatial representation and the apparentcentral relevance of geography to them. Indeed,whilst some (e.g., Bossler et al 2002) areundoubtedly about GIS as an area of activity, theyprefer to use the term ‘geospatial’ (implying a subsetof the adjective ‘spatial’ applied specifically to theEarth’s surface and near-surface) rather than‘geographical’ – particularly when referring totechnologies or generic techniques.

On balance, we think this is a pity – the term‘GIS’ is distinctive, widely recognised in governmentand business, and suggests the clear anddemonstrable link to geographical informationscience (GIScience) that underlies our own view ofthe development of the field. We suggest that thenewer term, by contrast, which fails to confer anyunequivocal advantages, is likely to confuse theuninitiated and seemingly presents applications asdevoid of real-world scientific context. For us, theterm ‘geospatial’ seems to imply a preoccupationwith technology and low-order data concepts, ratherthan the higher levels of the chain of understanding(evidence, knowledge and wisdom) that we havediscussed elsewhere (Longley, Goodchild, Maguireand Rhind 2005: Section 1.2). A new term maynevertheless be attractive to some, such as thosedeveloping instruments and technologies (e.g. GPSand remote sensing) associated with academicdisciplines other than geography, and new businessentrants to the field that may not have track recordsin GIS.

The rise of the term ‘geospatial’ also seems toconvey to some a greater sense of ‘hard science’ andhas sometimes gone un-noticed in the discipline ofGeography. In some parts of the world (but notablynot the United States), it seems to us that thediscipline of Geography has developed a verycomplacent attitude to investment in its flagship

x

P A Longley, M F Goodchild, D J Maguire, and D W Rhind

03_735450_fpref.qxd 2/24/05 2:17 PM Page x

specialism: GIS is widely recognised in high schools,is central to transferable geographical skills, is key tooutreach to potential new recruits to the disciplineand is crucial to justification of education funding asa part laboratory-based discipline. We observe that adecreasing real share of new student recruitment, aneed to recruit skilled researchers from otherdisciplines, and a downturn in real share of researchawards are seeming consequences of such neglect.

These are concerns with GIS as an appliedproblem-solving technology, and are thus not thecentral preoccupations of Helen Couclelis (Chapter2: Space, time, geography) or Ron Johnston (Chapter3: Geography and GIS). But these concernsnevertheless have ramifications for the practice ofGeography and the vitality of the discipline. Today,the Geography taught in many quite respectableuniversities largely fails systematically to address thecore organizing principles that are inherent to anywidely recognized definition of the subject (see alsothe contribution of Pip Forer and David Unwin,Chapter 54: Enabling progress in GIS and education).Ron Johnston observes the periodic disciplinaryrumblings about the inevitable ethical inconsistencyof anything that can be measured, allied withassertions that quantitative measurement is eitherpassé, no longer practiced, or relevant only to ‘thin’preliminary description of geographical problems(e.g., Cloke et al 2004). Some traditional quantitativegeographers have suggested that numbers arerelevant to the vitality of the discipline, in ways thatscarcely begin to capitalize on the potential andachievements of GIScience (Johnston et al 2003),while some external observers remain bemused atthe failure of the discipline to agree on issues ofcontent, coherence and relevance. Meanwhile, GISpractitioners continue to respond to interest withinother disciplines and remain part of the GIS successstory.

Six years on, it seems odd to recall the passionsthat were sparked by the early debates betweensocial theorists and GIScientists, and that were soaccurately recorded in the chapter by John Pickles(Chapter 4: Arguments, debates and dialogues: theGIS–social theory debate and the concern foralternatives). The dialog that began with the FridayHarbor meeting in 1993, and continued throughResearch Initiative 19 of the National Center forGeographic Information and Analysis (NCGIA) hascontinued, and today it would be inconceivable thata text in GIS written by a geographer would not

address the social context and social impacts of thetechnology. As the author writes, referring to thewell-known arguments presented by C.P. Snow inthe 1950s about deep divisions in society between thetechnological orientation of the sciences on the onehand and the more reflective orientation of thehumanities on the other (Snow 1961), ‘the best ofour students and young faculty members working inGIS and social theory are no longer...wrapped up inthe ‘Two Cultures’ arguments and ways of thinking.’

Recently several significant new directions haveemerged in the dialog. Most conspicuous of these,perhaps, is the growing interest in actor-networktheory and the work of Bruno Latour, led by suchGIScientists as Chrisman, Harvey, and Schuurman.It asks, in essence, how advances in GIScience can beunderstood in the context of the individuals whomade them, their networks of interaction, and otheraspects of their social setting. What, for example, canwe learn from the particular circumstancessurrounding Roger Tomlinson’s proposal for aCanada Geographic Information System in the 1960s(Foresman 1998), or Ian McHarg’s (1995) interest inmodeling planning decisions through the overlay oftransparent representations of different variables?Who were they talking with in that period, and whatwas the nature of their relationships with others?

Pickles also notes a third recent trend: ‘...theconvergence of discussions about information andcommunication technologies, public-participationGIS (PPGIS), and broader writings aboutrepresentational technologies. One branch of PPGIShas to do with issues of cost, access, open source,technical assistance, user-interface design, the Web,etc. These have moved along quickly, as [we] know.Another part of this, however, is the question ofagency and voice – crudely put, who acts, whodecides, who designs, and who inputs?’ This secondarea seems to Pickles to be moving much moreslowly, but the questions it asks seem moreimportant, and he anticipates increasing interest inthe years to come. Issues of public participation inGIS are addressed in detail by Craig et al (2002).

Several real breakthroughs have been achieved inthe past six years in the area of map generalisation.Notable among them are the work of Bader andBarrault on the application of energy-minimisationtechniques from engineering physics, and the carefulanalysis of feature shape employed in techniquesdeveloped for road generalisation at IGN. Progresshas also been made in algorithms to control the


xi

03_735450_fpref.qxd 2/24/05 2:17 PM Page xi

generalisation process, by exploiting optimalselections and combinations from among the host ofspecialized techniques that have been developed byresearchers over the past two decades. Ware et al(2003) used combinatorial optimisation techniquesincluding simulated annealing and geneticalgorithms, integrating several algorithms into acomprehensive process. In the AGENT project(http://agent.ign.fr), a multi-agent system was builtthat is even more flexible and basically allowsintegrating any generalisation algorithm into acomprehensive process using a constraint-basedapproach (Barrault et al 2001).

Author Robert Weibel writes: ‘there is anincreasing demand by map and data producers, inparticular public mapping agencies, to be able toautomate the production process of maps anddatabases more thoroughly, owing to new productrequirements such as data or maps for navigationand for location-based services (LBS). In response tothese requirements, the leading softwaremanufacturers have improved their productsconsiderably in recent years, both in terms offunctionality and quality. For example, Laser-Scan,which has a tradition of concentrating on themapping market, was a partner in the AGENTproject. The agent-based prototype system ofAGENT has since been turned into a commercialproduct called Clarity. Intergraph, another softwarevendor with a strong presence in the mappingmarket, offers its own generalisation software underthe name DynaGen. Even ESRI, where mapping isbut one of many markets, has added moregeneralisation-related functions into its system inrecent years. Additionally, there are smaller vendorsthat offer their own solutions, such as Axes Systemswith the Axpand software.’

Generalisation research is no longer restricted toautomation of the production of paper maps, but ismoving in new directions in response to newproducts, technologies, and demands. The extremelyconstrained screen area and resolution of hand-helddevices, such as cellphones, is one such area, andothers include automated generalisation of three-dimensional objects such as buildings, and thecreation of the multi-resolution databases that areneeded in vehicle guidance systems.

By 1999 it was abundantly clear that the Internetand Web were having a profound impact on manyaspects of human activity. Over the past six yearssince the original publication of this book, several

new directions have emerged in cartographicresearch and practice, driven in large part by thepotential of these new communication technologies.As we also comment elsewhere in this review andupdate, dramatic reductions in the cost of software,and simultaneous increases in power, have enabledlarge numbers of non-experts to engage in map-making, and to make their products freely availableto others. A specific design approach has emerged,motivated by the limited bandwidth of the Interneton the one hand, and by easy access to techniques oftransparency, blinking and shading on the other.The Scalable Vector Graphics (SVG) format hasbeen adopted widely in the cartographic community,leading to increasing facility of interaction andinteroperability.

‘In the GIScience community’, writes authorMenno-Jan Kraak (Chapter 11: Visualising spatialdistributions), ‘the abundance of geospatial data hascreated a new role for maps in exploratoryenvironments where they are used to stimulate (visual)thinking about geospatial patterns, relationships andtrends. The context where maps like this operate is theworld of geovisualisation (Dykes et al 2004) which canbe described as a loosely bounded domain thataddresses the visual exploration, analysis, synthesisand presentation of geospatial data by integratingapproaches from disciplines including cartographywith those from scientific visualisation, imageanalysis, information visualisation, exploratory dataanalysis and GIScience.’

(b) Data quality

Interest in representation and ontologies is alsorelevant to the issues of data quality and uncertaintythat are inherent in them. Shi et al (2002) provideperhaps the starkest statement of this problem whenthey claim that ‘for many types of geographical data,there is no clear concept of truth, so that models thataddress the differences between measurement andtruth are clearly inappropriate’ (cited by Heuvelink2003, 817). This invites consideration of semantics(termed discord in Peter Fisher’s Chapter 13: Modelsof uncertainty in spatial data) and how discord arisesthrough the social construction of information (alsoalluded to by John Pickles in his comments, above).Peter Fisher’s contribution remains a very usefulframework for thinking about uncertainty in GIS.However, the Shi et al (2002) volume now providesan additional valuable and wide-ranging update on


xii

03_735450_fpref.qxd 2/24/05 2:17 PM Page xii

spatial data quality issues of theory, method andapplication, and provides evidence that we now havea much fuller and more rounded view of the sources,operation and consequences of uncertainty in GIS.This volume also demonstrates the broadening ofinterest that has occurred in the topic over time,and illustrates how we have moved well beyondpreoccupations with uncertainty propagated inmap overlay operations and statistical models forrepresenting positional uncertainty (discussed byHoward Veregin in Chapter 12: Data qualityparameters). Today, there is a much more establishedinterest in visualisation and communication ofuncertainty, the pitfalls of decision-making underuncertainty, and the quest to develop error-sensitiveGIS.

Within the statistical perspective, GerardHeuvelink’s treatment of uncertainty (Chapter 14:Propagation of error in spatial modelling with GIS)remains a holistic overview of the sources andoperation of quantitative attribute errors that is ofenduring importance. He identifies three respects inwhich the methodological discussion of Chapter 14might be updated. First, he observes that MonteCarlo simulation is very much taking over from theTaylor series approximation method and other moreanalytically based methods for uncertaintypropagation analyses. Algorithms for stochasticsimulation of spatial phenomena, he argues, havebecome much richer over recent years – particularlywith regard to problems involving categorical spatialdata and the stochastic simulation of objects. Second,and related to this, he notes progress in problems ofuncertainty involving categorical spatial attributes(e.g. Kyriakidis and Duncan 2001). He observes thatproblems of uncertainty with categorical attributesare more difficult than their continuous-attributecounterparts. Third, he sees progress in handling errorpropagation arising out of positional uncertainty, asdiscussed by Shi et al (2002).

More generally, the recent literature providessome evidence that spatial uncertainty analysis isenjoying greater real-world professional application.However, it is the case today that uncertaintyanalysis remains the preserve of the GIS specialist,and that it remains a labour-intensive task. Time willtell whether uncertainty analysis and risk analysiswill become more standard elements in spatialanalysis.

Kate Beard and Barbara Buttenfield’scontribution (Chapter 15: Detecting and evaluating

errors by graphical methods) remains an importantearly graphical evaluation of uncertainty (or morespecifically, errors). In 1999 it seemed that furtheradvances in the detection and evaluation of errorsby graphical methods depended on furtherrefinement of error models. Since then, new methodsand models have been developed, and a generaltrend is evident away from summary-levelmeasurements, such as root-mean-square error(RMSE), toward local statistics that reflect non-stationarity. They concur with Gerard Heuvelinkthat greater quality assessment is now supportedthrough use of Monte Carlo simulations to generatemultiple realisations of error processes. Author KateBeard writes ‘These improvements have raised newchallenges for graphic display. Finer detail in spatialvariation can be more difficult to communicate andthe side-by-side and sequenced graphic displaysoriginally proposed for quality depiction are lessable to support user perception and association offine spatial variations in quality with the datadistribution. With simulations the visualisationchallenge is to convey effectively the uncertaintyrepresented by large sets of possible realisations.’

Progress has been made in the development oftechniques linked to specific error types andassessment contexts, and such work has becomemore central to the visual analysis of spatialdistributions discussed in Section 1(a) above. Forexample, Menno-Jan Kraak (Chapter 11: Visualisingspatial distributions) has recently described avisualisation tool for fuzzy attribute classificationthat uses a collection of multiple and dynamicallylinked visual displays including images, parallelcoordinate plots, and a 3D feature space plot thatusers can interact with to explore the classificationprocess. Rather than simply viewing static displaysof error, this approach creates an environment forinteractive exploration that potentially leads to abetter understanding of uncertainty.

Gerard Heuvelink, Kate Beard and BarbaraButtenfield share a frustration that little of thisresearch has found its way into the commercial GISpackages and into mainstream use. The need forquality assessment has if anything grown in the pastfew years with the explosion of Web applicationsand the growing availability of spatial data. Whilethe majority of the spatial data distributed over theWeb now have metadata associated with them,quality descriptions are not as complete or as usefulas they might be for prospective users. Work by


xiii

03_735450_fpref.qxd 2/24/05 2:17 PM Page xiii

Devillers et al (2002) and Bédard et al (2004) makessome progress on this front. Their work targets data-quality exploration in an interactive environmentassociated with a SOLAP (spatial on-line analyticprocessing) architecture that allows users to isolateand explore individual data-quality variables or toview data-quality metadata at several differentaggregation levels from an individual data value, toobject classes, to an entire data set.

(c) Spatial analysis

Spatial analysis is very much the engine that drivesresearch applications of GIS. In technological terms,the chapters in this section assimilate many of theimplications of the ongoing improvements incomputer memory and processing speed, and thesedevelopments have continued to impact profoundlyupon the field of GIS. The chapter by StanOpenshaw and Seraphim Alvanides (Chapter 18:Applying geocomputation to the analysis of spatialdistributions) exemplifies an approach to spatialanalysis that continues to develop using theincreased power of computing: although widelyapplicable analytical solutions to the modifiableareal unit problem remain as elusive as ever, thegeocomputational approach to zone design providesan example of the use of research techniques inspatial analysis in applied problem solving – forexample, the 2001 UK Census zones were designedaround some of the principles set out in Chapter 18.

The contributions to this section of the book alsovariously flag a number of longstanding issues aboutthe ways in which we carry out our scientificinvestigations and seek to improve the kind offindings that we are able to generate. The linkage ofthese issues to those of spatial, temporal andcognitive representation, and to visualisation (alldiscussed in Section 1(a) of this book) hasstrengthened in recent years. This is largely becauseimprovements in computation have facilitatedgreater disaggregation and a greater focus uponindividuals and micro-scale events and occurrences,alongside improved representation of temporaldynamics and, in socio-economic applications, morerealistic simulation of spatial behaviour.

Developments in digital data infrastructures havealso fuelled interest in spatial analysis techniques thatperform well using large numbers of georeferencedobservations. Data-mining techniques (Miller andHan 2001) in the geocomputational paradigm

provide the most obvious examples. Tremendousprogress has been made in the past six years in thedevelopment and dissemination of readily accessibletools for advanced forms of spatial analysis. LucAnselin’s (Chapter 17: Interactive techniques andexploratory spatial data analysis) GeoDa consolidatesand extends the available set of tools for exploratoryspatial data analysis, local statistics and spatialregression. It has been developed through the Centerfor Spatially Integrated Social Science(www.csiss.org) and has been downloaded over 4,000times. At the same time, ESRI’s ArcGIS 9.0 hasgreatly expanded the set of tools available either asextensions or as parts of the core of this popularGIS; GeoVISTA Studio from Pennsylvania StateUniversity has become a powerful collection ofopen-source tools with an emphasis on visualisation(www.geovistastudio.psu.edu); and Serge Rey’sSTARS (Space-Time Analysis of Regional Systems;stars-py.sourceforge.net) adds another open-sourcepackage focused on statistical techniques to thisrapidly growing collection.

Historically, GIS has provided a medium withinwhich spatial analysis techniques, that often predatethe innovation of GIS by decades, could berehabilitated for real-world problem solving in data-rich environments. The pace of development isevident in the increasing recognition of theimportance of space in academic disciplines outsideGeography. As this rehabilitation nears itscompletion, the techniques that are most used arethose that deliver the most in real-worldapplications. Thus while some of the techniquesreviewed by Art Getis (Chapter 16: Spatial statistics)and Manfred Fischer (Chapter 19: Spatial analysis:retrospect and prospect) have withered in usage,others – notably Bayesian and non-parametricstatistics, have become much more widely used. It isalso important to note that the softwareenvironment of GIS is also leading to thedevelopment of new spatial analysis methods –geographically weighted regression (Fotheringham etal 2002) being perhaps the best example. This is anexample of a technique that allows the uniquecharacteristics of localities to be examined withinwhat is ultimately a global statistical generalisation –this kind of statistical sensitivity to context is alsoillustrated by locally sensitive measures of spatialautocorrelation, such as the Ord and Getis Ostatistic (Ord and Getis 2001).

Finally, it is also clear that practical problems of


xiv

03_735450_fpref.qxd 2/24/05 2:17 PM Page xiv

extending theory in applied contexts andaccommodating ethical concerns are becomingincreasingly important: problems of ownership ofdata, copyright, and the limits posed by changingconceptions of ownership are changing the way weare able to share and communicate spatialinformation, and require ever more of spatial analystsin relation to privacy issues (see Michael Curry,Chapter 55: Rethinking privacy in a geocoded world).

3 TECHNIQUES (PART TWO)

(a) GIS architecture issues

As suggested above, the original edition of this bookpostdated the innovation of the Internet, but far-reaching changes in peer-to-peer networking (DavidColeman, Chapter 22: GIS in networkedenvironments) have occurred in recent years. Thenotion of using peer-to-peer technology andstandards to access distributed spatial data holdingsbegan attracting the attention of the GIScommunity in late 2000. In suggesting the scale ofchange that has occurred since this book was firstpublished, David Coleman cites OECD statisticsthat suggest the number of Internet subscribersworldwide has more than tripled to over 300 millionsince 1999, and the number of those subscriberswith broadband connections has increased to over100 million over the same period. He points out thatthe sharing of digital data (whether as downloadedmusic, movies, images, games, software orgeographical data) through peer-to-peer networkscontinues to increase at an unprecedented rate: thenumber of people logged on simultaneously topopular file-sharing networks approached 10 millionin April 2004, a rise of 30% from the same period ayear earlier.

GIS interoperability (Mark Sondheim, KennGardels and Kurt Buehler, Chapter 24: GISinteroperability) is key to driving such data exchange.Using Open Geospatial Consortium (OGC;www.opengeospatial.org), U.S. Federal GeographicData Committee (FGDC; www.fdgc.gov) and ISOspecifications along with IT industry standards (e.g.SOAP/XML, .Net and Java), emerging Webmapping services based upon common standards areenabling users not only to access geospatial datafrom different servers around the world, but also todetermine their fitness for a particular application,

conflate (not just register) them to a common base,and then communicate the results to others.Geoportals (Maguire and Longley 2005) such asGeospatial One Stop (www.geodata.govv: see alsoSection 4(a) below) have become important ways offacilitating access to data (see also David Rhind,Chapter 56: National and international geospatialdata policies). Much of this progress has been drivenby developments in the wider informationtechnology arena, and their adaptation to the specialneeds of GIS. For example, the success of XML(eXtensible Markup Language) has led to thedevelopment of GML (Geography MarkupLanguage) as a GIS data transfer format; andseveral of the leading database management systemshave added spatial capabilities, in the form of IBMDB2 Spatial Extender and Oracle Spatial, forexample.

Author Mark Sondheim writes ‘The OpenGeospatial Consortium and related ISO(International Standards Organisation) activitieshave been increasingly successful in developingspecifications highly relevant to GIS interoperability.The most popular of these is certainly the Web MapService; however, Web Feature Service, WebCoverage Service, Location Based Services, WebCoordinate Transformation Service, Web RegistryService (as a profile of the OGC Catalogue Service)and others are beginning to be implemented as well.Of course GML is a key development of the OGCand integral to some of these services. It is of notethat both commercial and open-source versions ofsome of these services are now available. This bodeswell for their broader acceptance.’

An important area of GIS that we did notanticipate in 1999 is the convergence of networkedGIS with communications and positioningtechnologies in location-based services. Today, GISdata and software are increasingly accessed remotely,allowing the user to move away from the desktopand hence to apply GIS anywhere – as ‘distributedGIS’. Limited GIS services are already available incommon mobile devices such as cell phones, and areincreasingly being installed in vehicles. In the futureit is clear that GIS will become both more mobileand ubiquitous, and will be based around distributeddata, distributed users and distributed software(Longley, Goodchild, Maguire and Rhind 2005:241-59). Author David Coleman notes that thenumber of cell phone users worldwide has almostquadrupled since this book was published in 1999;


xv

03_735450_fpref.qxd 2/24/05 2:17 PM Page xv

and that many of these cell phones contain tinyGlobal Positioning System (GPS) receivers thatallow service providers to know the location of thecaller in the event of an emergency or (with userconsent) to monitor his or her activity patterns. In asimilar way, sensor webs are being created,composed of intelligent digital devices that exchangemany kinds of information with people and othermachines around them. These are creating new typesof networks which will ultimately provide importantsources of input to GIS and Web-mapping systems.

Distributed GIS offers enormous advantages, inreducing duplication of effort, allowing users to takeadvantage of remotely located data and servicesthrough simple devices, and providing ways ofcombining information gathered through the senseswith information provided from digital sources.However, progress on a number of fronts is difficultbecause of complications resulting from thedifficulties of interacting with devices in field settings,limitations placed on communication bandwidth andreliability, and limitations inherent in batterytechnology. Thus for the foreseeable future, we arelikely to continue to associate GIS with the desktop,where rapid developments in software (Elshaw Thralland Thrall, Chapter 23: Desktop GIS software)remain central to the development of the field.

Sue Elshaw Thrall and Grant Thrall reflect thatthe compounded effect of developments in desktopsoftware has led to a bifurcation in GIS. On the onehand, specialised GIS software continues to addspatial analysis applications while, on the other,many core functions are no longer recognised as GISper se. With respect to the latter, they point out thatend mass-market GIS users are only rarely awarethat they are using high-technology geography. Thusinteractive street displays and route finders arestandard fixtures in up-market automobiles;camping stores sell GPS-enabled wristwatch-likedevices that use GPS and digital terrain maps foroff-road route finding; joggers use similar devices ashigh-tech alternatives to the pedometer; and Internetmapping has become ubiquitous, through servicessuch as mapquest.com and yell.com. Perhaps mostimpressive of all, Microsoft’s MapPoint finds routes,calculates drive times, and interfaces with both GPSdevices and standard Excel software, therebybecoming much more than geographically enabledspreadsheets. MapPoint is not yet presented as aserious contender to challenge mainstream GISfunctionality: but if and when the general public

becomes familiar with and expects GISfunctionality, MapPoint will be there to seamlesslydeliver to and interface with Microsoft Officeapplications.

An update of core GIS software offerings and theways in which they may be customised (DavidMaguire, Chapter 25: GIS customisation) is providedby Longley, Goodchild, Maguire and Rhind (2005:157-75). Proprietary geographically enabledprogramming languages have been replaced withintegration of GIS functionality within standardprogramming languages such as Microsoft’s VisualBasic, Visual C++, C# and Java. These developmenttools have assisted deployment of GIS via wirelessinterfaces to a range of hand-held devices, and assuch have contributed to the development oflocation-based services. A major new developmentin this arena is the adoption of distributedarchitectures (such the .Net and Java frameworks)for implementing enterprise systems. Web servicesand service-oriented architectures (SOA) are theunderlying technologies that sew together legacy andnew systems.

A final point is that GIS software and dataproducts are sensitive to the end uses to which theyare put. Recent years have seen increasing usage ofGIS in the areas of business and business geography(Thrall and Campins 2004). It is sometimes the casethat such users use different spatial terminology (e.g.‘trade areas’) and are reluctant to becomeacquainted with the established terminology of GIS.Thus GIS market development has entailedrepackaging of standard offerings into business-specific application packages, such as ESRI’sBusiness Analyst, that use novel user interfaces toguide the user through solving a business problem,rather than requiring mastery of a list of GISfunctional logic.

(b) Spatial databases

Relational database technology (Mike Worboys,Chapter 26: Relational databases and beyond)remains and looks set to continue to be at the heartof geographical data management and analysis forthe foreseeable future. Author Mike Worboys reflectsthat object-oriented database management systems(DBMS) have not turned out to be replacements forrelational systems. Instead, he sees that the mostpopular model, and that implemented by the largedatabase players, has been to incorporate object-


xvi

03_735450_fpref.qxd 2/24/05 2:17 PM Page xvi

oriented concepts in extended relational databasesystems. The challenges set out in Mike Worboys’chapter still remain: constraint databases remain anelusive route towards incorporating more expressivepower; and the dynamic nature of the world is stillnot addressed by the current round of databasetechnology (e.g. see Worboys 2005). Elsewhere moreprogress has been in evidence: geosensor datamanagement suggests clear and important goals forthe future of database science in the context of real-time data management; the linkage betweenagent-based computational models and databasetechnology provides interesting possibilities for themanagement and analysis of dynamic spatial data;and event-oriented models are becoming importantfor the development of our ability to represent andreason about dynamic geographical phenomena.

Most of the current generation of DBMSs incommercial systems (e.g. Oracle, Informix, DB2,etc.) and open source software (e.g. PostgreSQL,MySQL) support geographical data to some degree.This represents a significant improvement on thesituation when Peter van Oosterom (Chapter 27:Spatial access methods) prepared his contribution,although the spatial indexing methods that aresupported remain limited to variants of simple grids,R-trees and quadtrees. He notes that there remainchallenges to using these methods to manage agreater range of topological structures, such aslinear networks and TINs, which have beenresponded to in part by Oracle’s, LaserScan’s andESRI’s commercial systems. He sees a growing areasof interest in creating methods to support 3D datawithin GIS and geo-DBMSs: this entailsimplementation of appropriate data types (e.g.polyhedra), operators (e.g. for computing volumeand intersection), 3D spatial access methods andcomplex 3D structures (polyhedral partitions of 3Dspace, TINs, etc.). He also notes that progresstowards the development of multi-scale spatialstorage and access methods has been very limited.

The past six years have seen rapid developmentand adoption of the database design environmentsdescribed by Yvan Bédard (Chapter 29: Principles ofspatial database analysis and design). Modellingtechniques are now generally accepted in thedevelopment of geographic databases for GIS, LBSand other position-aware applications, and severalGIS products now include modelling tools. Over thepast six years, UML has clearly established itself asa standard and is nowadays used on a regular basis

by the GIS industry, academia and standardisationbodies like the ISO/TC-211 and OGC. A goodexample of this evolution is the thousands ofdownloads of the spatially extended UML CASEtool ‘Perceptory’ and its use in over 40 countries – amajor increase in usage rates since the late 1990s. Inthe meantime, very flexible approaches to systemdesign have emerged in reaction to the highlydisciplined approaches typically represented by RUP(Rational Unified Process). These light but robustapproaches, stemming from the philosophy ofExtreme Programming, are described by theumbrella terms ‘Agile Development’ and ‘AgileModelling’. Yvan Bédard observes that numerousbooks describing these new approaches haveappeared in the past six years, including booksspecifically focused upon agile database techniques.Books comparing ‘heavy methods’ with ‘lightmethods’ have also appeared in order to appraiseusers of their relative merits.

New sources and types of data may generate theneed for new access methods. In particular, Peter vanOosterom identifies the need to support a basic‘point cloud’ data structure, in order to manage themassive amounts of point cloud data that are nowcollected by laser scanning (or multi-beam sonar),and for which traditional (raster and vector) datastructure access methods are not very effective.

Human interaction with GIS (Max Egenhoferand Werner Kuhn, Chapter 28: Interacting with GIS)is also changing in detail, if not in its fundamentalnature. Author Werner Kuhn identifies a number ofdevelopments in current thinking about userinterface issues. First, he observes that the userinterface of many applications is now a Webbrowser, or at least is being accessed through one.This has generally improved the possibility oftransferring knowledge from one application toanother (and thereby the usability of both). Second,Web interfaces have also led to an increasedemphasis on the use of remote services in preferenceto locally installed functions. This has had the effectof vastly simplifying the functionality of GIS. Third,the range of devices and their user interfaces hasbroadened: some of the most challenging user-interface problems are now those of small, mobiledevices such as cellphones (see the discussion oflocation-based services in Section 2(a) above).Fourth (and also echoing David Coleman’scomments in Section 2(a) above), user interaction isincreasingly becoming implicit, as more and more


xvii

03_735450_fpref.qxd 2/24/05 2:17 PM Page xvii

information about users becomes available throughsensors, including the determination of userlocation. Fifth, ‘clickable’ maps and other forms ofdynamic user interaction are now commonplace.And sixth, interactive visualisation of 3D and 4Ddata remains generally awkward and confined toresearch prototypes.

(c) Technical aspects of GIS data collection

There have been significant developments in theimage data available for use in GIS, subsequent tothe publication of Ian Dowman’s chapter (Chapter31: Encoding and validating data from maps andimages). He observes that not only hasphotogrammetry moved into the digital age, but anumber of new and important satellite sensors havecome into service (e.g. see Donnay et al 2001).Digital photogrammetric workstations wereavailable at the time that this book was originallypublished, but these have now become standard forphotogrammetric map production in mostdeveloped countries. In very recent years digitalairborne cameras have reached the market and arealready making an impact by cutting out filmprocessing and extending the range of conditionsunder which photography can be obtained. IanDowman notes that airborne LiDAR andinterferometric synthetic aperture radar (SAR) arenow used for directly generating digital elevationmodels (DEMs) and have allowed dense networks ofaccurate elevation points to be obtained forapplications such as 3D city models and floodprediction and management. High-resolution opticalsensors on satellites, with pixel sizes of 0.6m, arenow available, thus extending the accuracy of imagesavailable from space (cf. Mike Barnsley, Chapter 32:Digital remotely-sensed data and their characteristicsand Chapter 48: Jack Estes and Tom Loveland:Characteristics, sources, and management ofremotely-sensed data). More DEMs are nowavailable using data from space sensors, the mostnotable being a near-global DEM at 90m spacingand 10m vertical accuracy from the Shuttle RadarTopography Mission (SRTM); these data are freelyavailable over the Internet.

While there have been a number of technicalimprovements in Global Positioning Systemequipment, the basic technology fundamentallyremains the same (Art Lange and Chuck Gilbert,Chapter 33: Using GPS for GIS data capture). Art

Lange observes that there are now more and bettersources of differential correction such as the FederalAviation Administration’s Wide Area AugmentationSystem (WAAS) and its counterpart in Europe andJapan. Virtual Reference Station (VRS) networks areanother new source of differential correctionbeing implemented across Europe and morerecently in select areas within the United States. Theappeal of VRS technology is easy to understand: amobile GIS user equipped with a standard mapping-grade GPS unit connected to a cell phone with datatransfer capability can routinely collect location datawith sub-meter accuracy – without assistance from adedicated receiver used as a base station. Moreimportantly, differential correction is constantwithin the VRS network area. There is nodegradation of correction accuracy as the usermoves away from the differential signal source.

Art Lange anticipates that within the next fewyears, a new generation of GPS satellites will belaunched with improved access to the L2 Civilian(L2C) signal (the ‘second frequency’). This will leadto improved measurement accuracy by removingsome of the errors associated with atmosphericGPS signal distortion. The general trend of GPSreceivers for GIS applications has been to providegreater accuracy and more user-friendly features ata generally lower cost. The widespread availabilityof consumer-grade GPS receivers with built-inlarge map data bases has changed the waymany non-GIS professionals depend upon, andare affected by, the accuracy of the coordinates inGIS databases. GIS software for processing field-collected data has continued to evolve – forexample, the GPS extensions to ArcGIS andArcPad, which make the process of transferringfield-collected GPS/GIS data to the user’s databasea seamless operation.

While GPS-enabled devices have opened the doorfor location-based technologies, the emergence ofWi-Fi-based systems promises to bring the samelocation-tracking services indoors to hospitals,warehouses and large industrial complexes. Creatingwhat has been sometimes called ‘the Internet ofThings’ or ‘indoor GIS’, these systems are beingused to track moveable assets, such as machines,merchandise, animals, and even people carrying orwearing a new generation of tiny RFID (RadioFrequency Identification) tags which will storeimportant attribute information about the object orwearer of the tag.


xviii

03_735450_fpref.qxd 2/24/05 2:17 PM Page xviii

Here once again, the potential of these systemsalso highlights the important growing debate overwhat constitutes ‘appropriate’ access to personaldata and a person’s current location (see Chapter 55,Michael Curry: Rethinking privacy in a geocodedworld).

(d) Data transformation and linkage

Spatial interpolation continues to be a vitallyimportant part of GIS functionality, and the set oftools has matured significantly over the past sixyears. Increasingly, spatial interpolation services arebeing offered remotely over the Web, as the field ofscientific computing moves to a more integratedview of how standard operations are packaged andmade available.

The most dramatic recent development has beenthe increase in the size of the datasets that requirespatial interpolation or approximation. Newmapping technologies, such as LiDAR, real-timekinematic GPS, and automated sensors have madedata acquisition orders of magnitude more efficient.However, data processing and analysis often lagsbehind data acquisition – for example, millions ofgeoreferenced points can be measured by LiDARwithin a single hour but it takes much longer toproduce a bare-Earth digital elevation model. Theproperties of data produced by automatedtechnologies also require smoothing of noise and, asa result, spatial approximation rather than exactinterpolation has become increasingly important.

While the principles of the most commonly usedmethods (Kriging, splines, inverse distanceweighting, nearest neighbour, triangulated irregularnetwork) remain the same, their implementation hasbecome more robust and better adapted to dataheterogeneity, noise and large datasets. Researchinto model-based interpolation continues and isbeing used for specialised applications (meteorology,topography, groundwater). A fully automatedmethodology that would select the most appropriatemethod and optimise its parameters for a givendataset and application is still not available.However, the tools that help to select a suitablemethod (geostatistical analysis, visualisation, etc.)have improved significantly. Nevertheless, writesauthor Helena Mitasova, ‘spatial interpolation andapproximation remain a challenging task for manyGIS users that requires a solid understanding ofavailable methods and modelled phenomena.’

Author Antonio Câmara sees four main trendsunderlying the integration of GIS and virtualenvironments over the past six years. First, thepromise of immersive systems largely failed todeliver, and the field remains dominated by non-immersive desktop solutions. These benefit from thedevelopment of an extension to VRML (VirtualReality Markup Language) known as GeoVRML,which provides the capability to browse multi-resolution, tiled data that are streamed over the Web(Reddy et al 1999, 2001). Second, interoperabilityefforts, such as those promoted by the OpenGeoSpatial Consortium (www.opengeospatial.org),are also facilitating the integration of GIS withvirtual environments. SRI International is in theimplementation phase of an OGC Web Map Service(WMS) capable of generating GeoVRML output.

Third, improved representations of virtualterrains have become possible thanks to newtechniques of detail management and multi-resolution modeling of geographical data. The workof Losasso and Hoppe (2004) and Cignoni et al(2003) on interactive rendering of large-sizedtextured terrain surfaces is worth mentioning.Finally, mobility is a new trend in GIS due to theemergence of the wireless Internet and theavailability of communication-enabled mobiledevices. Romao et al (2003) discuss the promise oflocation-based augmented reality services, wherethree-dimensional synthetic images from databasesare superimposed on real images in mobile devices.

Many of the issues of emerging data infrastruc-tures, provision and access that are raised by MikeGoodchild and Paul Longley (Chapter 40: Thefuture of GIS and spatial analysis) are becoming stillmore significant with the gradual maturation ofadvanced information economies. They observe thatthis is particularly apparent in the socio-economicrealm, where small-area geodemographic measuresof social and economic conditions are becoming keyto GIS-based models of resource allocation. Currentresearch in geodemographics illustrates theinterdependences between classification method,variable selection and data source when devisingclassifications that work in the real world. Theyreiterate the point already made several times abovethat, in measurement terms, more data are collectedabout more aspects of our individual lifestyles thanat any point in the past, through routine interactionsbetween humans and machines. They see


xix

03_735450_fpref.qxd 2/24/05 2:17 PM Page xix

enlightened approaches to public and academic dataaccess (e.g. through geoportals: Maguire andLongley 2005) as key to making wide disseminationof socio-economic data a reality, and makingpossible an open debate about the remit andpotential of social measurement at neighbourhoodscales. Geodemographic systems based onframework socio-economic data can be successfully‘fused’ to census sources to provide richer depictionsof lifestyles – yet lifestyles data sources are usuallynot scientific in collection and requirereconsideration of the practices of science in theways identified in Chapter 40. If this can besuccessfully undertaken, the toolkit of spatialanalysis in GIS now makes it easier than ever beforeto match diverse data sources and accommodate theuncertainties created by scale and aggregationeffects.

4 MANAGEMENT (PART THREE)

As in most other things, the GIS world of 2005 is arather different place to that of 1999 when theoriginal version of this book was published. Beforewe discuss what has changed in the world of GIS, wefirst rehearse (very briefly) the external factors whichimpact on our subject matter and which differ fromthose of 1999.

(a) The changing context

One factor is common to the drivers for change notedin the Principles and Techniques section. Thecontinuing evolution of information technology hasbeen rapid. This has had some direct effects and manyindirect ones. The growth of storage capacity –typically around 20GB in 1999 and around 100GBnow – exemplifies the more/smaller/faster trend whichenables software functionality and applications thatwere hitherto impossible to become commonplace.The most obvious change is in regard to location-based services (LBS) which have penetrated bothconsumer cellphone and industrial and militarytracking markets. The dramatic growth in broadbanduptake and Internet bandwidth has also fuelled theinterest in such geographic Web services.

Aside from the technology, the most obviouschange of GIS drivers is that governments and theirleaders have changed. Nowhere has this been moreimportant than in the USA, the largest individual

‘engine’ of GIS. The replacement of Clinton andGore by Bush and Cheney resulted in a loss of overtsupport for GIS based on the former administration’sseemingly altruistic approach – this was theunderpinning of the 1994 Presidential ExecutiveOrder which triggered the US National Spatial DataInfrastructure. Yet federal government support inthe USA has not died: rather it has beentransformed into something driven by the twoimperatives of getting more efficient government andHomeland Security issues (see below). Like manyother governments, the US federal one has launchedan e-government initiative to enhance efficiency andservice to the citizens, with one of the initial batch ofprojects being a GIS metadata service: theGeospatial One Stop initiative (see Section 3(a)above and Box 20.6 in Longley, Goodchild, Maguireand Rhind 2005) builds upon much of the workdescribed by Guptill (Chapter 49: Metadata and datacatalogues). Underpinning the present situation isthat metadata standards have continued to evolve asthey have made their way through the variousnational and international standards bodies. Whilethis has helped in the syntax and coding ofmetadata, the underlying content has remainedpretty much intact. As more diverse user groupshave reviewed the metadata standards they haveadded elements of value to their specialty. As ageneral result, the number of possible data elementshas grown quite considerably. The major difficultywas, and continues to be, that very few parties arewilling to fully populate both the ‘mandatory’ and‘optional’ fields. Mandatory fields are usuallyreduced to the equivalent of ‘title, author, date’ andare usually populated. Since many other fields areoptional and left unfilled, any third party data useris left unaware as to the value of the data and itsfitness for particular uses. The only way out of thisproblem is if GIS software becomes a lot smarterand automatically encodes more fields of metadata.

The number of catalogues has continued to growbut there is as yet no good catalogue of catalogues.Thus much surfing is still required to discover whatis needed (or if it exists). The user interfaces for thecatalogues are also different, creating a learningcurve for novice users. Private sector sites haveemerged which point to their products and theproducts of their partners. Some of the products arerepackaged government data sets. This can causesome user confusion.

More generally, whilst the formal policies of


xx

03_735450_fpref.qxd 2/24/05 2:17 PM Page xx

national governments described in Rhind (Chapter56: National and international geospatial datapolicies) have not changed dramatically, someblurring has occurred at the edges, some potentiallyimportant international agreements have beenlaunched (see Longley, Goodchild, Maguire andRhind 2005, Chapter 20) and the advent ofcommercial suppliers has rendered the practicalimport of some policies much more limited:government matters less in GIS than it did. Perhapsbecause of this, the nature of the National Map isevolving rather differently in many countries (see, forinstance, Longley, Goodchild, Maguire and Rhind2005, Section 19.3) – most dramatically between theUSA and the UK.

The ‘new world order’ has transformed manyaspects of our world since 1999 and impacted onGIS. The growth of terrorism, notably manifested inthe events of 9/11 in the USA but replicated in manyways in other countries as far afield as Colombia,India and Russia, has triggered a renewed focus on‘homeland security’ and military campaigns such asthose in Afghanistan and Iraq. Immediately after9/11 there was a flurry of activity in removing GIfrom Websites so as to minimise aid to terrorists;after various studies it was concluded that this wasnot really necessary (see Baker et al 2004). But, moregenerally, GIS is clearly relevant to homelandsecurity issues. We can think of five stages in anymajor disaster, natural or man-made. They are:

● Risk assessment● Preparedness● Mitigation● Response● Recovery

It is now widely accepted that each of theseinherently involves use of geographic informationand GIS. Equally, some parts of all of these stagesinvolve human judgement, understanding of thecharacteristics of other organisations as well as ofdata and a clear understanding of what needs to bedone for the greater good – and a strong code ofethics. The consequences for those in GIS ofterrorism acts and of natural disasters alike (like theSouth East Asia tsunami disaster of December2004) has been a further impetus to the developmentof hardware and software, including sensingplatforms ranging from aircraft a few centimetresacross to commercial satellite imaging (supplied tothe military, the ‘anchor customers’) with resolutions

as great as 60cm (and getting still finer – 40cm ispredicted by 2007).

But the consequences of the changed geopoliticalsituation are not all those of Homeland Security.The addition of a further 10 countries to theEuropean Union, now an entity with 450 millionpeople, creates a formidably large trading bloc andone which is determined to become a major player inhigh technology and its uses; GIS is part of this, theEuropean Union proposing far-reaching plans for anew GI directive.

In one sense a ‘counterbalance’ to terrorism andhomeland security is the use of GIS to supporthumanitarian efforts arising both from natural andhuman causes. Both in-field hand-held mobile GISand office-based systems have been used to greateffect during such crises – the former to collect dataabout safety, infrastructure damage, disease, forexample, and the latter to plan missions such as fooddrops, and temporary hospital locations, as well asto brief executives and members of the press. Rapidaccess to current data can help save lives and reducehuman suffering.

It is not only military concerns and homelandsecurity issues that have become global. Globalscience has come of age. Concerns with ‘big issues’,such as climate change and global warming anddimming, have become widespread. Models of pastand projected changes to our habitat have typicallyused GIS: the geographical manifestations of anychanges have become highly charged political issues.And natural disasters – such as the South East Asiatsunami but also many others – have forced a needfor much-improved monitoring and early warningtools in which GIS has a key role to play.

(b) The results

Over the past few years we have seen a maturingrecognition of the potential of GIS in its originalheartland (North America, some parts of WesternEurope and Australasia) but also in the emergingeconomies of Asia, Europe and Latin America. Wehave estimated (Longley, Goodchild, Maguire andRhind 2005) that the likely number of active GISprofessional users must be well over two millionpeople world-wide. At least double that number ofindividuals will have had some direct experience ofGIS and perhaps an order of magnitude morepeople (i.e. well over 10 million) will have heardabout it (e.g. through such events as GIS Day). Yet


xxi

03_735450_fpref.qxd 2/24/05 2:17 PM Page xxi

more will – sometimes unknowingly – have usedelementary GIS capabilities in passive Web servicessuch as local mapping. We estimated that, in 2004,the total global expenditure caused by the use ofGIS could not be less that $20 to $25 billionannually; the sum is not precisely knowable and ispartly dependent on how we define GIS – but it islarge and continuing to grow every year. Oneindicator of this growth is the numbers attendingsoftware conferences: that hosted by ESRI has seennumbers grow from a start of 23 attendees in 1982 to12,000+ in 2004. Even if we take our highestestimate (10 million globally) of those having hadsome experience of GIS, this still means that onlyone person in 600 on Earth falls in this category.Clearly there is still a long way to go! Part of thisgrowth has come about by a spread of the same typeof applications worldwide and part from innovativeapplications. Section 5 describes how these haveexpanded and diversified.

The GIS industry has also become more mature.This is manifested in various ways. There is nowwidespread international agreement on the use ofstandards of various kinds, on mainstreamfunctionality and even the terminology used – withone exception (see below). The growth of thecommercial elements in GIS – software, systemintegration, consultancy, data provision, etc. – hasbecome global and has led to some consolidation.The US hegemony is being challenged: for example,the data and services company Tele-Atlas of theNetherlands purchased its equivalent, GDT of theUSA.

In becoming more mature, the industry hasbecome more commercial. We now know much moreabout the business of selling data and services basedupon sound economic principles (see for instance,Box 19.10 in Longley, Goodchild, Maguire andRhind 2005 or Shapiro and Varian 1999). We havealready identified the ease with which globalmonitoring by commercial satellite and data-servingorganisations is carried out. To an extent, GIS hasbecome democratised: mapping is now producedmore frequently by individuals via the Web than bynational mapping organisations and the underlyingservice is funded by advertising – a dramatic changeto the situation of a few years ago. The early workdescribed by Shiffer (Chapter 52: Managing publicdiscourse: towards the augmentation of GIS withmultimedia) has become much more commonplaceand public participation in GIS (PPGIS) has

blossomed with the fall of GIS software costs andthe spread of GIS skills (see Box 20.2 in Longley,Goodchild, Maguire and Rhind 2005).

Although business applications of GIS have notdeveloped at the rate anticipated when this book wasfirst published (see the comments of Sue ElshawThrall and Grant Thrall in Section 3 above), therehave nevertheless been three sustained developmentsin the area described by Mark Birkin, GrahamClarke and Martin Clarke in the book (Chapter 51:GIS for business and service planning). First,businesses have continued to create ever-increasingvolumes of data about their operations, customerbehaviour, and competitive environment (e.g.through loyalty cards, lifestyle data, stock controlsystems, etc). Many of these data are spatiallyreferenced with respect to the points of delivery andof consumption. However, Mark Birkin, GrahamClarke and Martin Clarke suggest that there ‘is asyet little evidence that retail and service businessesare able to determine key actions based on thisinformation, for example in the optimisation ofproduct mix or network configuration.’ In their view,off-the-shelf GIS packages offer appropriatecomponents for operations such as like spatialinteraction modelling, but still lack the flexibilityand sophistication required to support businessdecision-making, especially at the strategic level.Second, some of the increasing interest in spatialdecision support systems (e.g. Geertman andStillwell, 2002) has been directed at businessapplications, though this area of activity remainssmall relative to environment and physical planningapplications. Third, there has been interest in newmodelling techniques such as intelligent agents andcellular automata (see Longley and Batty 2003).Here again, however, the emphasis has not beenupon the tactical or strategic needs of business, orupon activity patterns and spatial behaviour, butrather has addressed more general problems ofurban form and structure. There has, however, beenan upsurge in interest in the use of geodemographics(a tool traditionally focused upon businessapplications) to issues of efficiency and effectivenessof public service delivery, and these are addressed inthe update on Tony Yeh’s contribution (Chapter 62:Urban planning and GIS) in Section 5(a) below.

In general, the capabilities of GIS to profile entirepopulations of a country (or beyond), groupingthose millions of individuals on the basis of theirinferred (e.g. purchasing) characteristics, has created


xxii

03_735450_fpref.qxd 2/24/05 2:17 PM Page xxii

commercially valuable knowledge. But this has adownside. The black forebodings about the capacityof the technology to destroy privacy expressed byMichael Curry (Chapter 55: Rethinking privacy in ageocoded world) have more than come to pass.Knowing where people are at any moment clearlyhas some benefits – e.g. if someone is being attackedand calls for help, the emergency can be targetedeffectively in real time. But combining this with adetailed or inferred knowledge of a variety of thevictim’s characteristics – and potentially producing acradle-to-grave narrative of an individual’s life posesreal challenges for the sort of society we wish to be.Curry has gone so far as to suggest that the impactof homeland security and the developments inlocational technology mean that we have returned insome senses to the 1960s and 1970s. He argues thatwhilst in the 1980s and 1990s it was widely believedthat business was the most significant threat toindividuals’ rights to privacy and autonomy, themore significant threat now appears to arise fromgovernment.

If the technology and practical experience ofdesigning, selecting and using GIS has come someway since 1999, the legal aspects remain complex.The law touches everything. During a career in GIS,we may have to deal with several manifestations ofthe law. These could include copyright and otherintellectual property rights (IPR), data protectionlaws, public access issues enabled (e.g.) throughFreedom of Information Acts (FOIA), and legalliability issues. But since laws of various sorts haveseveral roles – to regulate and incentivise thebehavior of citizens and to help resolve disputes andprotect the individual citizen – almost all aspects ofthe operations of organizations and individuals aresteered or constrained by them. One complication inareas such as GIS is that the law is always doomedto trail behind the development of new technology;laws only get enacted after (sometimes long after) atechnology appears. All those using GIS need to beaware that, whilst commerce is global, law – for themost part – is not. In essence, there is a geography ofthe law, the legal framework varying from country tocountry. The creation, maintenance anddissemination of ‘official’ (government-produced)geographic information are strongly influenced bynational laws and practice. The best recent overallsummary of this is Cho (2005).

The management of GIS has also come someway: chapters by Bernhardsen (Chapter 41: Choosing

a GIS), by Obermeyer (Chapter 42: Measuring thebenefits and costs of a GIS) and Sugarbaker(Chapter 43: Managing an operational GIS) stillcontain much of value. But since 1999 much focushas been put on ‘interoperability’ and the concept ofspatial data infrastructures (SDI). These two topicshave given a new dimension to the process ofchoosing a GIS. They are technically based on thedevelopment of standards created by ISO and OGC,such as the geography markup language (GML) andthe Web map service (WMS). These technicaldevelopments are discussed in Section 3(a) above.The driving force is the user’s requirements foraccess to data, and thus to increase the social benefitof data collected with only minor increase in thecosts. Any organisation producing georeferenceddata, and which is in the process of choosing a GIS,should now choose a system not only on the basis oftheir internal user requirements, but should seek tointegrate their own new system in the regional ornational spatial data infrastructure. Ourunderstanding of how to assess costs and projectedbenefits of systems has also improved considerably(see Thomas and Ospina 2004 and Tomlinson 2003).There are, for instance, far more examples of real-world benefit-cost analyses for GISimplementations than there were in 1999; goodexamples include the ‘best-GIS’ ESPRIT-ESSIProject n.21.580 at http://www.gisig.it/best-gis/Guides/chapter9/nine.htm or that written byDarlene Wilcox at http://www.geoplace.com/gw/2000/0200/0200wlcx.asp.

Given the much-enhanced commercial interest inGIS and GI, it is no great surprise to find thatuniversity departments outside the traditional onesof geography and surveying are now beginning toteach and research in our subject area. The highest-ranked business school world-wide on manyoccasions in the past decade has been the WhartonSchool of the University of Pennsylvania. It nowhas a senior member of staff working in GIS andthis trend is being followed elsewhere.

All this begs the question of what is needed andwhat is changing in GIS education. In truth, weknow relatively little about the backgrounds ofmany people who are active GIS practitioners sincesome at least seem to have become ‘accidentalgeographers’, drawn into GIS by the need to carryout spatial operations in their own, original, field ofendeavour. What they felt they needed is thereforeless than crystal-clear. Traditionally, however, much


xxiii

03_735450_fpref.qxd 2/24/05 2:17 PM Page xxiii

education in GIS has really been training, especiallyin how to use a particular set of software tools (seeForer and Unwin’s Chapter 54: Enabling progress inGIS and education). Forer has argued that we ‘may beproducing a range of people who have a simple toolfor simple problems, and aren’t impatient with thetool or aware of its multiple limitations’. Unwin, onthe other hand, has claimed that four developmentshave occurred since 1999 in GIS and education:

● The development of several Web-based distancelearning courses. This, he argues, says more aboutuniversities’ needs to expand their market ingeneral than it does about GIS;

● A near universal concentration on GIScience asopposed to GISystems, though training remains athriving industry;

● Greatly increased teaching of GIScience incontexts other than academic geography. Mainlythis seems to be in applications areas such asarchaeology and environmental science andmanagement rather than computer science.

● A divergence in trajectories between the US andUK. In the US, geography has done well out ofembracing GIScience, which has been thespearhead of a very substantial revival in thediscipline’s fortunes throughout education. This isnot so in the UK where, despite the best efforts ofmany people, it has not been greatly used in an‘embedded’ mode, to teach about other parts ofthe discipline (see also the comments on RonJohnston’s Chapter 3: Geography and GIS, inSection 2(a) above).

Given all this and since the world of applicationshas spread ever-wider, surely the old approaches areno longer appropriate and new core competences areneeded? To maximise the utility of GIS, we see theneed for education offerings in GIS and GI-relatedareas to extend beyond the ambit of commercialfirms and universities. We see the need for some GISeducation to be much more sophisticated in relationto critical theory and its application to understandthe downsides as well as the upsides of GIS and tohelp understand how it is changing some aspects ofsociety. We therefore conjecture the need for GISeducation now to include:

● Entrepreneurial skill development and leadership● The principles of geographical science● Understanding of and familiarity with GIS

technologies

● Understanding of organisations● Finance, investment criteria and risk management● Human resources policies and practice and ethics

relating to use of GIS● Legal constraints to local operations● Cultural differences between disciplines● Awareness of international differences in culture,

legal practice and policy priorities● Formal management training, including staff

development, team working using GIS, andpresentational and analytic skills

● Attempts to embed GIS and GIScience in themainstream of the academic discipline ofgeography as well as other disciplines

Clearly not all courses and learning needs toinclude all of this material and some of it is not bestlearned from courses. Some elements will beparticularly relevant to those engaged (as allprofessionals should be) in continuing professionaldevelopment. The introduction of local legal,cultural and application-related elements to GIScourses – as well as buttressing the global technical,business and management issues – will be to thebenefit of GIS practitioners, the discipline andbusiness (used in a wide sense) alike.

One characteristic which has become ever morecommonplace is the role of GIS partnerships. Thesenow operate at all levels, ranging from the very local(e.g. where lobby groups share resources and poolskills to produce maps to illustrate threats), to thecontinental (like the Permanent GIS Committeecomprising 55 countries in Asia) to the global. Theyalso range from the informal through to thosedefined via international treaty obligations, withcommercial partnerships being somewhere in themiddle. The most common manifestation of thesepartnerships is a national spatial data infrastructure:some 39 countries are now said to have one thoughwhat really is happening on the ground is somewhatvariable. How partnerships are best made to work inan era of Homeland Security, where safeguardingaccess to information is seen by some as crucial, isnot immediately clear.

5 APPLICATIONS (PART FOUR)

In the Introduction to the Applications section inthe book we classified GIS applications astraditional, developing and new. Traditional


xxiv

03_735450_fpref.qxd 2/24/05 2:17 PM Page xxiv

applications include military, government,education, utilities and also natural resources. In thelast six years these traditional areas have continuedto prosper and still constitute the lion’s share of GISapplication activity and revenue. The US NationalGeospatial-Intelligence Agency (NGA) continues tobe the largest single spender on GIS in the world.The developing applications of the mid-1990sinvolved general business uses such as banking andfinancial services, transportation logistics, real estateand market analysis. Consistent with Sue ElshawThrall and Grant Thrall’s comments in Section 3above, we observe that in the ensuing years theseapplications have not made the progress we hadexpected. In part this is a manifestation of thedownturn in the US economy in the early 2000s (itsstock market has yet to regain its all time high ofearly 2000), but other contributory factors includecommercial company inertia, an inability for GISvendors and consultants to simplify the technology,and a poor presentation of the business case.Nonetheless, there are notable showcase examples ofGIS in business, such as home delivery of electricalgoods at Sears in the USA, reinsurance riskmodelling by PartnerRe in Europe and, at a morelocal scale, direct marketing to banking customersby Arrowhead Credit Union, California (Thomasand Ospina 2004). Back in 1999 we hypothesizedthat small office/home office (SOHO) and personalor consumer applications would represent the nextwave of new GIS applications. While we wereincorrect in our SOHO prediction we were on targetwith our suggestion about personal/consumerapplications. There has been rapid growth in interestin location-based services. Some things that we didnot foresee in 1999 included growth in interest inhomeland security, in geoportals and in spatial datainfrastructures (SDIs: Maguire and Longley 2005).It was the events of September 11, 2001 that put theterm ‘homeland security’ into common parlance.The interest and funding for homeland securitygeographic information infrastructure projects hashelped propel the rather stagnant SDI communityinto the modern era (see Section 4).

Looking forward we envision a bright future forhand-held and mobile GIS applications, as well asdevelopment of a new field that may becharacterized as ‘indoor GIS’. The latter isconcerned with the location and movement ofresources within buildings. Using RFID (radiofrequency identification) tags (see Section 3(c)

above) and other technologies it is possible tomonitor the movement of people, and otherinanimate resources, around buildings. This hasmany implications for security and for facility andresource management.

(a) Operational applications

In the new millennium GIS continues to be used toeven greater effect in operational application areas.There can scarcely be a government in the world thatdoes not use GIS either directly or indirectly in oneway, shape or form for managing its assets, be theyland and property parcels and easements,road/street/highway networks, or information aboutcitizens (Tony Yeh, Chapter 62: Urban planning andGIS). GIS has played a central role in many of thebig digital/electronic government initiatives of thepast few years (Curtain et al 2004; Greene 2001;Song 2003). A similar picture can be painted forutilities (Jeff Meyers, Chapter 57: GIS in the utilities;Caroline Fry, Chapter 58: GIS in telecommunications)where GIS is extending from core networkmaintenance and mapping applications to a pluralityof applications such as tree trimming, customerrecruitment and retention, environmentalmanagement, pipeline routing, mobile workforcemanagement, and transportation logistics. There isalso a trend toward integrating GIS with otherexisting enterprise applications such as networkoptimization, SCADA (Supervisory Control andData Acquisition), CRM (Customer RelationshipManagement) and ERP (Enterprise ResourcePlanning) systems. After a period of rapidexpansion, the telecommunications industry is nowin a period of stability, and consolidation is theorder of the day in most geographies.

In Jeff Meyers’ opinion, two major and relatedchanges have driven the utility GIS industry in thepast five years. First, GIS has become mainstreamInformation Technology (IT). This change beganwith the ability to store the geometry (spatialcharacteristics) of features in garden-variety, openrelational database management systems. Partlyenabled by improved server and networkperformance, the open storage of GIS featuresbrought GIS into the fold of IT within utilities,initially encouraging and then demanding that ITprofessionals review platform requirements in thecontext of corporate standards. The standardtechnology trend continued with the development of


xxv

03_735450_fpref.qxd 2/24/05 2:17 PM Page xxv

core GIS tools written with standard programminglanguages. As programming standards emergedwithin the IT industry, GIS vendors (for the mostpart) adopted those standards, making GISimplementation and extension using standardprogramming languages possible.

The implications of standard technology aremany. Among the most important, open GIStechnology has reduced the barriers to ITacceptance within the utility organisation, andprovided a means for more standard IT support.Software engineers and computer scientists withstandard education can now be utilized as resourcesin GIS. Through data storage in non-proprietaryformats, management can confidently invest in GISto manage assets more effectively, without the fearthat data might become obsolete or be costly tomigrate from one proprietary format to another.Standard technology has led to integration of GISwithin the backbone IT of the enterprise, sharingdata and even function between core IT systemsthrough standard interface techniques. In thisrespect, open GIS technology has drasticallyreduced the life-cycle cost of spatially enabledsystems. And, perhaps most significantly, theavailability of GIS technology based on computingindustry standards has changed the debate aboutvendor selection from a preference-based choicebetween proprietary platforms to a focused decisionbased on business benefit and feature function.

The adoption of open, standard GIS technologyhas been a key driver in the second key change inutility GIS. As predicted and planned by industryexperts and observers, utilities have moved GIS froma departmental tool to an enterprise solution.Enabled by standard technology, and driven bycompetitive factors within the energy industry, manyof the best performing companies in the utilityindustry today use GIS as an everyday tool for assetmanagement, work design, and outage management.Through the integration of disparate data sourcesabout network assets, utilities use GIS today as aneveryday resource for managing assets andcommunicating change across enterprises that servelarge geographic areas. Additionally, standardcomputing technology has enabled the developmentof focused end-user GIS applications that allowbusiness people to interact with data aboutcustomers and assets in a spatial context withoutbecoming ‘GIS experts’. End user GIS tools thatsolve business problems lead to the demand for more

and more functionality. And since they can easily besupported through standard developmentenvironments, GIS core teams are happy to obligewith more development, which in turn leads todemands for more data and function, and theenterprise cycle sustains itself.

In terms of new GIS developments intransportation, Nigel Waters (Chapter 59:Transportation GIS: GIS-T) draws attention toSimon Lewis’ lists of GIS-T 10 accomplishmentsand 10 challenges for the future (http://www.gis-t.org/yr2003/gist2003sessions/gist2003session5.htm).The accomplishments are largely technological whilethe challenges for the future relate to concepts andframeworks, people and institutions. Others besidesLewis have their own lists (e.g. Fletcher 2003;http://www.gis-t.org/yr2003/gist2003sessions/gist2003session4.htm). Looking forward Watersbelieves that there is an ongoing need for GIS-TScience integration, temporal modelling oftransportation data, integration of web-basedmodelling, a new data model, markup language anddata standards, and greater public participation inGIS-T (for more discussion see: http://esri.com/industries/federal/gis-business/transportation2.html).Waters’ chapter on GIS-T discussed some of thepioneering attempts at integrating GIS-T andIntelligent Transportation Systems. This discussionconcentrated on work being carried out on theROMANSE project at Southampton, UK, andother parts of the European Community(http://www.romanse.org.uk). The field has movedforward since then but perhaps more slowly thanmight have been expected. This is especially true inCanada where a report by Transport Canada andIntergraph has detailed the problems and challengesof integrating ITS and GIS-T (http://www.its-sti.gc.ca/en/downloads/execsum/tp13224e.htm). Thethree main difficulties outlined in this report are thecost-recovery pricing policies of the CanadianGovernment and Statistics Canada, the lack ofpublic-private partnerships (one of the greatstrengths of the ROMANSE project) and the lack ofgovernment involvement (another strength ofROMANSE). The UK government, like itsCanadian counterpart, also implements punitivecost recovery policies, while the US remains one ofthe few countries where this is not the case. Thus it isin the US that GIS-T and ITS integration is likely totake off in the coming years. Key new works in thisfield include Lang’s review of applications (Lang


xxvi

03_735450_fpref.qxd 2/24/05 2:17 PM Page xxvi

1999) and Miller and Shaw (2001) that describesGIS-T principles and applications.

GIS provides the core organising framework foremergency management (Tom Cova, Chapter 60:GIS in emergency management). Today emergencymanagement activities are focused on three primaryobjectives: protecting life, property and theenvironment. All phases of emergency managementdepend on geographic information from a variety ofsources. The use of GIS ranges from displaying theeffects of events, to managing the actual incidentsthemselves at command posts or emergencyoperations centres. GIS is the only tool capable ofproviding a common operating picture foremergency management, planning, response andrecovery. State of the art emergency managementcentres allow managers to fuse real time GIS data(pertaining to the area affected or threatened by anevent) with fixed or static GIS infrastructure data.This provides users with dynamic ‘actionableinformation’ in near real time. This information canbe further enhanced within a GIS with eventmodelling tools and by importing real time weatherdata. With these tools, emergency managers canclose roads, order evacuations, and route publicsafety responders efficiently and accurately. Radke etal (2000) review the application challenges forGIScience and their implications for research,education, and policy for risk assessment, emergencypreparedness and response.

The awareness of the benefits of good landadministration has grown over the last five yearsaccording to Peter Dale and Robin McLaren (GIS inLand Administration, Chapter 61), in part because ofthe better understanding of the role of land inpoverty reduction. As a World Bank Policy ResearchReport states (World Bank 2003) ‘Land Policies areof fundamental importance to sustainable growth,good governance, and the well-being of andeconomic opportunities open to rural and urbandwellers – particularly poor people’. De Soto (2000)has argued that open, efficient and enforceableproperty rights could release trillions of dollars of‘dead capital’ in poor countries.

The recognition of the need for greater opennesswith regard to property rights has led to increasingopportunities to exploit land-related data. This hasbecome possible as a result of technologicaldevelopments both with regard to networking andalso through the use of GIS as a data analysis tool –in addition to its established roles in creating and

maintaining cadastral maps. Most countries areintroducing computerized systems for handling theirland administration data and some are now makingthese data available across networks in movestowards full electronic conveyancing, supported bye-signatures. In Europe, for example, all countrieshave computerized systems and the developmentemphasis has shifted from ‘design and build’ to‘sustain and maintain’. There are also movesthrough the European Union Land InformationService (EULIS) to exploit data across nationalboundaries so that all European countries mightparticipate in a pan-European land market.Conveyancing is increasingly being perceived as anintegral part of e-government initiatives and thisservice is being delivered by governments alongsideother ‘life events’, such as the registration of births,companies, marriages and deaths, using ‘citizenaccounts’.

The money to pay for the upgrade of hardwareand software systems has to come either fromgovernment or from charges for products andservices. To achieve the latter, the agencies deliveringland administration services must operate on abusiness basis and although they may not be run forprofit, they are increasingly being asked to recoverall their operating costs. While this is relatively easyfor those handling text data, such as computerizedversions of old deeds and paper documents, it stillremains a problem for many mapping agencieswhose costs are not so easy to recover.

What started as a question of how to fund thepurchase of new equipment has led to afundamental change in the culture of landadministration agencies, making them much morecustomer focused and conscious of the costs of theiroperations. In addition it is leading to a reappraisalof the role of the private sector in landadministration and the creation of various forms ofpublic-private partnerships. Ultimately, in manycountries the risk remains with the centralgovernment, but the delivery of various services nowinvolves a wider range of stakeholders and is basedon greater exploitation of the data that are nowavailable. Many countries now provide informationservices to a wide range of stakeholders, includingfinancial services, taxation agencies, economicdevelopment agencies, police, estate agents and thecitizen, within the constraints of their legislativeframeworks for access to information. Open accessto Land Administration information is sowing the


xxvii

03_735450_fpref.qxd 2/24/05 2:17 PM Page xxvii

seeds for the introduction of participatorydemocracy.

These various developments are relevant to theuse of GIS in planning (Tony Yeh, Chapter 62:Urban planning and GIS), where it has establishedroles in development control and generaladministration. The more strategic use of GIS inplan making has been explored by Benenson andTorrens (2004), particularly with regard to thedevelopment of applications of cellular automata.Yeh’s discussion makes little mention of the use ofGIS to plan expenditure on services that is incurredat the local level: recent years have seen increasinginterest in the use of geodemographics (Harris et al2005; Longley 2005) to understand and prescribelocal spatial patterns of resource expenditure inpolicing, health and education. The greatly increaseduse of GIS in policing is discussed in detail byChainey and Ratcliffe (2005).

David Swann observes that the last five years haveseen an acceptance that GIS provides criticalinfrastructure for defence and intelligence (Chapter63: Military applications of GIS). There has been avery swift deployment of commercial-off-the-shelf(COTS) GIS throughout the defence sector,representing a considerable advance from previousniche usage. The rapid uptake of GIS is part of abroader recognition that IT adoption creates andrequires a transformation of defence activities. Thisrevolution in military affairs manifests itself as‘Network Centric Operations’ (NCO) – the use ofthe network to connect decision making acrossmultiple defence domains. The horizontal nature ofNCO demands a move to a modern services-oriented architecture that provides transparentinteroperability between domains.

Modern COTS GIS platforms are capable ofproviding enterprise-class IT infrastructure formilitary applications. The resulting spatialinfrastructure couples the sensors that monitor thebattlespace to distributed geodatabases thatcontain knowledge of the battlespace (data, datamodels, process models, maps and metadata).These distributed geodatabases are coupled todistributed geoprocesses that provide powerfulanalysis and information filtering tools. Theresulting information can be served across thenetwork for fusion into embedded geovisualisationcomponents. At every stage, interoperability isrequired with other enterprise class technologiessuch as supply chain management (SCM) and

enterprise resource planning (ERP).As a tool providing defence-wide infrastructure,

GIS is simultaneously able to support war fightingmissions such as command and control, businessmissions such as installation management, and avariety of strategic intelligence missions. Given theenormous expense of spatial data collection andproduction, this re-use across a common spatialinformation infrastructure offers immediate costsavings in addition to important new capabilities.

The massive increase in electronic information isnot without its challenges. Prue Adler and MaryLarsgaard (Chapter 64: Applying GIS in libraries)highlight an explosive growth in the acquisition anduse of electronic resources. These include resourcesthat are available locally and via the Web. Researchresources may include digital information, resourcesthat are neither owned nor licensed by the library,resources digitized by the library or in partnershipwith others, and electronic publishing projects. As aresult, there has been a dramatic rise in the digitalservices provided by research libraries. At the sametime the continued evolution of informationtechnologies including network services has greatlyincreased the ability of the research user tomanipulate, analyse, and integrate data andinformation. There remain continuing challenges forlong-term preservation and access to GIS andrelated datasets such as curation, meeting the needsof data users, data authors, and data managers, andthe degree of centralisation of the support that isprovided.

Today many in the academic and researchcommunities are embracing open access – thesharing of information and data without restriction.In support of these efforts, many research librariesare establishing institutional repositories to capture,index and preserve the intellectual content of theresearchers, faculty and staff in digital format. MITLibrary’s DSpace is illustrative of this digital libraryinitiative and is available worldwide as open sourcesoftware. As a consequence, a large and growingbody of GIS information and datasets is availablewithout restriction. At the same time as we arewitnessing a move towards open access publishing,so there is also a rise in the use of licensing by manyGIS vendors that may restrict how a user may utilizedata, such as restrictions on how data may be sharedor used. Today, nearly all major map libraries have aGIS facility of some kind. As a general rule, thisdoes not include teaching classes, but may well


xxviii

03_735450_fpref.qxd 2/24/05 2:17 PM Page xxviii

include brief introductory sessions for individualsand groups.

Standard cataloging of Web resources is nowcommonplace. Libraries are now also deeplyinvolved in the creation of metadata records fordigital geospatial data. It is both understood andaccepted that metadata records may be loaded notinto the library’s online catalogue but rather intosome alternative database, with some form of Webinterface to facilitate query.

(b) Social and environmental applications

David MacDevette, Robert Fincham and GregForsyth’s account of the role of GIS in nation-building (Chapter 65: The rebuilding of a country:the role of GIS in South Africa) was written during atime of major upheaval. They offer someobservations that are pertinent to contextualisingGIS in today’s developing African sub-continentwhere the situation regarding GIS has continued toalter, in many respects. Most notably, it is evidentthat GIS must be irrevocably tied into a strategiccontext of poverty alleviation and development.These are now major policy concerns of governmentand they require pursuance of service delivery on ascale unprecedented in the past.

They also observe that there is a need for achange in mind set of those who are purveyors ofthe GIS story. The technology has it roots outside ofthe continent – a European- and American-centriccapacity that has been transferred to the developingworld. ‘Proudly South African’ is a new motto inbusiness that signals a new pride in indigenouscapacities. For those working in the GIS arena itmeans an African context with a real commitment tothe ideals and aspirations of the country as a whole.From an educational point of view there has to be astrong commitment to the notion of Africanscholarship. Such an ideal offers great prospects forcontributing to better standards of living for moreof the country’s people than was the case in the past.

MacDevette et al’s original contribution remindsus how GIS can contribute to major world eventsand can effect social and political change. In somerespects GIS’s role in South Africa is comparable tothe events surrounding 9/11 and the December 2004South East Asian tsunami referred to throughoutthis chapter. Since 1999 GIS in South Africa hasfollowed a course more similar to that in manydeveloped countries in America, Europe and Asia

Pacific. One recent exceptional GIS project in SouthAfrica has been Willem van Riet’s Peace ParksFoundation of South Africa. Under his leadershipthe Foundation has established transfrontierconservation areas (TFCAs), also known as peaceparks, which are large tracts of land that crossinternational boundaries. The purpose of theseparks is to employ conservation as a land use optionto benefit local people. Their pioneering spirit isforging international cooperation and is resulting inareas where wildlife can roam freely across borders.GIS has been used not only to help spark theadoption of the original peace parks concept, but tomaintain existing parks and promote the concept inother countries.

The use of GIS in Health and health careapplications (Tony Gatrell and Martin Senior,Chapter 66) has experienced significant recentgrowth, both in numbers of GIS practitioners and inorganizations using GIS. The outbreaks of SARS inAsia, West Nile Virus in the US and foot-and-mouthin the UK, and the threat of global bio-terrorism allillustrate why GIS activities are essential. Among theranks of the newest users are public health programmanagers responsible for health policy impactanalysis, public health preparedness and diseasesurveillance. For example, in England and Walesthere is now a network of regionally based PublicHealth Observatories in which GIS figures quiteprominently. Increasingly, academic health sciencecentres offer GIS as part of their health informaticsand healthcare training programs. For example, USNational Institutes of Health have recently funded abiomedical imaging research program that is GIS-centric, as well as a Geospatial Medicine program atDuke University Medical Center. Large public andprivate hospitals are establishing GIS departmentsto foster the wider use of GIS to meet the challengesof improving performance and lowering costs. On aglobal scale, the use of GIS to geographically enablenational public health surveillance systems is one ofthe fastest growing markets for GIS technology andservices. Perhaps one of the greatest unrealizedvalues of GIS to health organizations lies in creatingnew knowledge about the convergence ofenvironmental factors on human health status andhealthcare delivery outcomes. Cromley andMcLafferty (2002) provide a useful review ofprogress in the field.

Recent years have certainly seen very manyprofound political changes. Mark Horn’s discussion


xxix

03_735450_fpref.qxd 2/24/05 2:17 PM Page xxix

of GIS and the geography of politics (Chapter 67)described the principles of gerrymandering andmalapportionment and showed how they have beenviolated. While the effects of geographicaljurisdictions were writ large on the outcome of the2000 US Presidential Election (when the winningcandidate failed to win the popular vote), it isargued that understanding of geodemographic detailis crucial to the US out-turn in 2004 and that in theUK in 2005. It seems that the availability of newtechnology through GIS can serve the interests ofthose that seek to manipulate the geography of voterturn out for political gain. It will be interesting to seeif broadening of participation and engagementthrough PPGIS can counter such effects. On abrighter note GIS has recently been usedsuccessfully to communicate election results. CBSand other television stations used GIS in the 2004US Presidential Elections: novel visualizationtechniques such as 3D perspective views andcartograms were used in near real time, perhaps forthe first time on primetime television. In planningfor future elections GIS-based redistricting remainsas popular as ever.

The use of GIS in Monitoring land cover and land-use for urban and regional planning (Peter Bibby andJohn Shepherd, Chapter 68) dates back to the veryorigins of the GIS field. Recent advances in satellitegeo-imaging have allowed the extents of land coverto be estimated ever more accurately and cheaply.Precise ascription of land-use, however, remains adifficult task, not least because of new andchallenging policy requirements to differentiatebetween greater numbers of land use classes. Bibbyand Shepherd concentrated on how GIS has beenused to generate information for policy purposes,but, today, there remains a remarkable lack ofawareness and willingness for senior politicians andpolicy makers to embrace science and technology.

GIS and landscape conservation (RichardAspinall, Chapter 69) is a field closely related to landcover and land-use estimation, in that both involvewildlife and scenic resources, and have significantenvironmental policy implications. Richard Aspinallsuggests that in the past few years there has beenwider use of more flexible regression methods(generalized linear models and generalized additivemodels) for predictive modelling of speciesdistribution (e.g. Guisan et al 2002) anddevelopment of rendering of landscapes forlandscape assessment (e.g. Bishop et al 2001).

The business of agriculture is also an inherentlygeographic practice from the local to the globalscale. In Chapter 70, Local, national and globalapplications of GIS in agriculture, John Wilsonillustrated this with a number of examples.Improvements in global satellite-based sensors haveprovided many new opportunities for monitoringand modelling global changes. At a more local scalesome of the advances in geoprocessing and visualmodelling (for example, in Clark Labs IDRISI andESRI ArcGIS) have helped progress simulationanalysis and modelling efforts.

The final chapter in this section examined GIS inenvironmental monitoring and assessment (LarsLarsen, Chapter 71). A feature of this chapter wasthe interest in real-time monitoring, data quality andmathematical modelling. Increasing concern withthe impact of pollution on the environment and theexploitation of natural resources have been majorconcerns for the environmental lobby and have ledto interesting GIS applications in the last few years.The proceedings of the 4th International Conferenceon Integrating Geographic Information Systems andEnvironmental Modeling (Parks et al 2003) providea useful update on progress in the field.

6 CONCLUSIONS

So, is everything progressing well, with thecontinuing advent of new technology driving ever-greater benefits to mankind? That is anunsustainably simplistic view of the GIS world. Yes,many of the current GIS tools are being used forreal benefits to society and helping to generatewealth, employment, safety and improvements ofthe quality of life of some peoples at least. But, asever with technology, we are exposed to risksthrough over-enthusiastic use of databases or thepursuit of wealth or power and influence irrespectiveof wider considerations. The GIS communityexhibits all the usual characteristics of any group ofhuman beings: it is sometimes fractious (evendisputing the best name for the field — GIS orgeospatial systems/engineering, as discussed inSection 2(a) above) or finding it difficult to establishconsensus (e.g. on the best way forward for nationalspatial data infrastructures). It often lacksprofessionalism compared to some older professionsor disciplines (such as engineering) so greater focuson this, on ethics and on regulation will probably be


xxx

03_735450_fpref.qxd 2/24/05 2:17 PM Page xxx

appropriate. But what we have now is a globalmovement which recognises that geography and theprocesses and actions which are manifestedgeographically can influence the lives of all humans.The next decade or so will show us whether thisoptimistic scenario really is what happens.

References

Baker J, Lachman B, Frelinger D, O’Connell K, Hou A, TsengM, Orletsky D, Yost C 2004 Mapping the risks: assessing thehomeland security implications of publicly available geospatialinformation. Rand Corporation Monograph (seehttp://www.rand.org/publications/MG/MG142)

Barrault M, Regnauld N, Duchene C, Haire K, Baeijs C,Demazeau Y, Hardy P, Mackaness W, Ruas A, Weibel R2001 Integrating multi-agent, object-oriented, andalgorithmic techniques for improved automated mapgeneralization. Proceedings of the 20th InternationalCartographic Conference, Beijing China, 3: 2110-6

Bédard Y, Devillers R, Gervais M, Jeansoulin R 2004 Towardmultidimensional user manuals for geospatial datasets: legalissues and their considerations into the design of a technicalsolution Third International Symposium on Spatial DataQuality, Austria

Benenson I, Torrens P 2004 Geosimulation: Automata-basedmodeling of urban phenomena. Chichester, John Wiley &Sons

Bishop I D, Wherrett J R, Miller D R 2001 Landscape andUrban Planning 52: 227-39

Bossler J D, Jensen J R, McMaster R B, Rizos C 2001 Manualof geospatial science and technology. London, Taylor andFrancis

Chainey S, Ratcliffe J 2005 GIS and crime mapping.Chichester, John Wiley & Sons

Cho G 2005 Geographic information science: mastering thelegal issues. Chichester, John Wiley & Sons

Cignoni P, Ganovelli F, Gobbetti E, Marton F, Ponchio F,Scopigno R 2003 Planet-sized batched dynamic adaptivemeshes P-BDAM Proceedings IEEE Visualization IEEEComputer Society Press: 147-55

Cloke P J, Cook I, Crang P, Goodwin M, Painter J, Pluto C2004 Practising Human Geography. London, Sage

Craig W J, Harris T M, Weiner D (eds) 2002 Communityparticipation and geographic information systems. London,Taylor and Francis

Cromley E, McLafferty S 2002 GIS and public health. NewYork, Guilford Press

Curtain G G, Sommer M H, Vis-Sommer V (eds) 2004 Theworld of e-government Haworth, Haworth Press

De Soto H 2000 The mystery of capital: Why capitalismtriumphs in the West and fails everywhere else. New York,Basic Books

Devillers R, Gervais M, Bédard Y, Jeansoulin R 2002 Spatialdata quality: from metadata to quality indicators andcontextual end user manual OEEPE/ISPRS Joint Workshopon Spatial Data Quality Management, Istanbul

Donnay J-P, Barnsley M J, Longley P A 2001 Remote sensingand urban analysis. London, Taylor and Francis

Dykes J, MacEachren A M, Kraak M J editors 2004 Exploringgeovisualization Amsterdam, Elsevier

Foresman T W editor 1998 The history of geographicinformation systems: perspectives from the pioneers. UpperSaddle River, NJ, Prentice Hall

Fotheringham A S, Brunsdon C, Charlton M 2002 Chichester,John Wiley & Sons

Geertman S, Stillwell J C H (eds) 2002 Planning supportsystems in practice. Heidelberg, Springer

Greene R W 2002 Open access: GIS in e-government. RedlandsCA, ESRI Press

Guisan A, Edwards T J, Hastie T 2002 Generalized linear andgeneralized additive models in studies of species distributions:setting the scene. Ecological Modelling 157: 89-100

Harris R, Sleight P, Webber R 2005 Geodemographics, GIS andneighbourhood targeting. Chichester, John Wiley & Sons

Heuvelink G 2003 Review of ‘Spatial data quality’ (Eds Shi W,Fisher P F, Goodchild M F). International Journal ofGeographical Information Science 17: 816-8

Johnston R J, Hepple L W, Hoare A G, Jones K, Plummer P2003 Contemporary fiddling in human geography whileRome burns: has quantitative analysis been largelyabandoned – and should it? Geoforum 34: 157-61

Kyriakidis P C, Dungan J L 2001 A geostatistical approach formapping thematic classification accuracy and evaluating theimpact of inaccurate spatial data on ecological modelpredictions. Environmental and Ecological Statistics 8:311-30.

Lang L 1999 Transportation GIS. Redlands CA, ESRI Press

Longley P A 2005 A renaissance of geodemographics forpublic service delivery. Progress in Human Geography 29: inpress

Longley P A, Goodchild M F, Maguire D J, Rhind D W 2001Geographic information systems and science. Chichester,John Wiley & Sons

Longley P A, Goodchild M F, Maguire D J, Rhind D W 2005Geographic information systems and science (Second edition).Chichester, John Wiley & Sons

Losasso F, Hoppe H 2004 Geometry clipmaps: terrainrendering using nested regular grids. Proceedings, ACMSIGGRAPH 2004 ACM Press: 769-76

Maguire D J, Goodchild M F, Rhind D W 1991 Geographicalinformation systems: principles and applications. Harlow,Longman

Maguire D J, Longley P A 2005 The emergence of geoportalsand their role in spatial data infrastructures. Computers,Environment and Urban Systems 29: 3-14


xxxi

03_735450_fpref.qxd 2/24/05 2:17 PM Page xxxi

McHarg I L 1995 Design with nature. New York, John Wiley

Miller H J, Han J, 2001 Geographic data mining and knowledgediscovery. London, Taylor and Francis

Miller H J, Shaw S-L 2001 Geographic information systems fortransportation: principles and applications. New York, OxfordUniversity Press

Ord J K, Getis A 2001 Testing for local spatial autocorrelationin the presence of global autocorrelation. Journal ofRegional Science 41: 411-32

Parks B O, Clarke K M, Crane M P (eds) 2003 Proceedings ofthe 4th International Conference on Integrating GeographicInformation Systems and Environmental Modeling: ProblemsProspectus and Needs for Research Boulder, University ofColorado Press

Photogrammetric Engineering and Remote Sensing 2003Special issue on ‘the National Map’. Volume 69.

Radke J, Cova T, Sheridan M, Troy A, Lan M, Johnson R2000 Application challenges for GIScience: implications forresearch, education, and policy for risk assessment,emergency preparedness and response. URISA Journal12:15-30

Reddy M, Iverson L, Leclerc Y, Heller A 2001 GeoVRML:open Web-based 3D cartography. Proceedings of theInternational Cartographic Conference ICC2001, Beijing,6-10 August

Reddy M, Leclerc Y, Iverson L, Bletter N 1999 TerraVision II:visualizing massive terrain databases in VRML IEEEComputer Graphics and Applications Special Issue on VRML192: 30-8

Romao T, Dias A E, Danado J, Correia N, Trabuco A, SantosC, Nobre E, Câmara A 2002 Augmented reality with geo-referenced information for environmental management ACMGIS 2002 ACM Press: 175-180

Shapiro C, Varian H R 1999 Information rules. Cambridge,Mass., Harvard Business School Press

Shi W, Fisher P F, Goodchild M F, eds 2002 Spatial dataquality. London, Taylor and Francis

Snow C P 1961 The two cultures and the Scientific Revolution.New York, Cambridge University Press

Song H-J 2003 E-government: lessons learnt and challengesahead. In Proceedings, 8th International Seminar on GIS:envisioning cyber-geospace and spatially enabled e-government: 1-113 Seoul, Korea Research Institute forHuman Settlements

Thomas C, Ospina M 2004 Measuring up: the business case forGIS. Redlands, CA, ESRI Press

Thrall G I, Campins M 2004 Mapping the geospatialcommunity: parts 1-4. GeoSpatial Solutions (July, September,November) 14 (numbers 7, 9, 11); also 2005 15 (number 1)

Tomlinson R 2003 Thinking about GIS: geographic informationsystem planning for managers. Redlands, CA, ESRI Press

Ware M, Jones C, Thomas N 2003 Automated mapgeneralization with multiple operators: a simulated annealing

approach. International Journal of Geographical InformationScience 178: 743-69

Worboys M F 2005 Event-oriented approaches to geographicphenomena. International Journal of Geographic InformationScience, in press

World Bank 2003 Land Policies for Growth and PovertyReduction. Washington DC, World Bank

Addenda

Lubos Mitas and Helena Mitasova, Chapter 34: Spatialinterpolation

Equation 3 should read:

Barry Boots, Chapter 36: Spatial tessellations.Unfortunately some of the illustrations in this chapter were

corrupted in the production process, and it has not beenpossible to replace them in this abridged edition. ProfessorBoots has made correct versions available for interestedreaders to consult at the following URL:info.wlu.ca/~wwwgeog/facstaff/bootscorrect.htm

Additional Acknowledgment to the Abridged Edition

Paul Longley’s contribution to the Abridged Edition wascompleted during funding for ESRC AIM Fellowship RES-331-25-0001.


xxxii

( )N

z zr rVar z zh r h22

1 1rh

i j

ij

N 2h

. -= + -c ! ^ _^ ^

a

h ih h

k

9: CD$ .

03_735450_fpref.qxd 2/24/05 2:19 PM Page xxxii

new developments in geographical information …gisteac/gis_book_abridged/...we are delighted that...

Documents