CAMBRIDGE MANUALS IN ARCHAEOLOGY Saeed seeeeseeeseeeueaaeanpsteaeaeeaeeeeeeeeeaee GEOGRAPHICAL INFORMATION acne SYSTEMS IN ARCHAEOLOGY cera et ‘Grom Bch, James Conolly rchaesogy, Clive Oto, Pa Tyes and Alan Viaoe Taplonoms, Lee Lyman “archaclogy and Conservation, 2 eRe C, Der haroogy, & GBrowse Mark Lake Tania frac, Urivarity Coleg Leno CAMBRIDGE UNIVERSITY PRESSConnolly James and Mark Laks, 2008. Geogrephical Information Systems in Archaeology. Manuals in ‘Archaeology. Cambridge University Press, Ceméridge, UK. 2 rcthods. However, artifi- beyor ‘of mind and tackling. social construction of emotions (Cafiamero and de Velde 2000) idea that cognition is not somehow separate from engagement with the (Maris and te Bockhorst 1996). We suspect that these developments will ‘rough to GIS, perhaps initially in conjunction with the use of agent-based ition models (Lake 2004) First principles 2.1 Introduetion onelusion ‘the most powerful tec ce are elements of truth strengths of the use of “first principles’ of GIS: the software and hardware requirements, geodetic and car- tographic principles, and GIS data models. These provide the conceptual building ling what GIS is, how it works, and what igh some of these “first principles’ may be even more as a result of the integration of GIS wi lation and ongoing developments in artificial intellige ‘five groups: data acquisition, spatial data management, database managemer visualisation and spatial analysis. Some of the routine tasks performed under these headings are outlined in Fig. 2.1 and described in Bex 2.1. While each of these tasks.are important in themselves, above all GIS should cchnology that pro- terms of understanding spatial and even space-time rela- snd anthropozenic phenomena (Couclelis 1999). IndFig. 21 The ive man groupe of tn peered by GS ater Joost 1997, Fig.) soncered both with the more fundamental conceptual issues of spatial and ime selatiouships as wel a the impact geospatial technologies are having ‘and social sciences (Mazhle 1990; Curry 1998; Forer and 2.2.2 Geographic information ‘There is considerable overiap between 8 with each other. Geospatial information can also , such asthe patterning of erosion on the facades jowion of cut marks on bone (e.g. Marean ¢ 102). Most archaeclagicel data ~ whether artefacts, ecofacts, sapes— have spatial and aspatia!atibat . These attibutes include 2.2 The basics“18 Fastprincipes ‘causes profound distortion and spatial error tional and regional have inimisng eror within their own boundaries. At very term subregional or local, the surface of the Earth can there is comespondingly the last dstotion (ig. 2). Al projections wil distort 2.3 Cartographic principles 19 there are four projection groups ‘defined on the bass of how this distortion is managed. Conformal or ortkomorphic (in the correct angle om any line ‘other point on the map, family and then a projection ‘ype. Forexample, «conical projection plualged a fitting acone over ‘one of the polar regions as depicted in Fig. 24, which is then cut along a meridian asin Fig, ‘equal-arca conic project Azimuthl (or planar) projections tepresent the Barth's surface on a fat plane ‘sing a single point of contact rather than a line of tangency (Fig. 2.6), Azimuthal20 Firstprinciples Fig 25 Albers equal-ee conical projection mith ane aw of tngency lef and ein (shed ins). Te eng ap eo he igh, showing he ine fade ‘dvtaece easement slong te ipes of engi (orplaner) projections are usually used to map the poles alchou ‘occur anywhere on dhe Earth's surface. If polar, them the pro with concentric lines of latitude and radiating lines of longi the centr point ro any otber point an the map are accurate. ‘Cylindrical projections are conformal and so 90° angles ate maintained between | gency Boome neressngy toed ‘common cylindrical projection is the Mercator Projection, which set thecqvtor auf ne of engeney and sales the ydnenson ae) reduce the distortion at polar extremes. This projection gives a very misleading ‘view ofthe world as movement away fom the equator causes areas towards the top and bottom of the map to become disproportionately large in area (Snyder and 3 Mercator Projection (TM projection), rotates the cone 90° so that a meride Voxland 1989, p. 10). The invented by Johann Lambert (1728-1 jan becomes the line of tangency. axis but maintains north-south mes Projection, The TM Projection is one ofthe standard ways of mapping the globe, 2.3 Cartographic principles a g.22 fiaicu projection wth aoe of angecycomesponting tthe ese {i emaniin (Sato ne). The meslng ape he ph shone De Bae oF Inet paral nes Finally, the Universal Transverse Mercator Projection (UTM) is a twentieth- ‘system for spatial data collection when a local grid syste Tt addition tothe projection system used to make the 1 be aware of which mathematical approximation of the shape of the Fart was ‘used for the construction of a map. The Earth is not an exact ellipsoid, since the surface ig not smooth and the poles are not equidistant from the equator. Polar coordinates of latitude and longitude are therefore calculated using a machematical22 -Firstprinciples (GRS80), World Geodetic System (WGS84) and European Terrestrial Refers nt reealeulations of the ellipsoid used in ly use whichever elipsoid caloulations rapping in Great Britain i, although the Ordnance jr mapping desived from ‘Survey have adopted the ETRS89 elipsoi GPS receivers. ‘Coordinate transformation and reprojection ‘Maps that share a projection system (e.g, UTM) bat are based on different elipsoids (e.g, WGS84 versus NAD27) are not compat se different dary transformation or reprajection, To informationisrequired aboutthe existing and desired projection systems. Mi provide tools to transform maps {rom one geodetic datum to another, and dedicated software tools ate also available to help convert between NAD27 and NADS3 (see, for example, the directory of sofware on the US National Geodetic Survey's website and their online conversion, tool’), Details about the ellipsoid and projection system used in che construction of ly printed inthe comer orate contained in an associated metatata ata (ee Chapter 13 for further discussion of metadata elemests) ‘Many GIS programs use geographic coordinates of latiude and longitude as the ional maps (most often as decimal degrees whece minutes and seconds 100.508 of degree). Whil thatare able to manage the corections for spatial measurement, Cartesian systt ‘based on metric units, soch as UTM or national (military) grid system, is often a systems provide metric planar coordinates al longitude, Metic planar coordinates are used in the global UTM projection, the Tonnage noaa.gov. T wne.age.noas.gov/ogiebin/nadson pr 2.3 Cartographic principles 3 ans tnorthings) US State Plane system, British National Grid and in most other national grids. as the US State Plane system, are often better choices ‘area covered by that particular system. In pars of o tary grids are unavailable, then UTM is um excellent choice. We must emphasise again oor warning from the previous section regarding {Oe inevitable and significant spatial errors that will result from combining data different projections and easy calculation of distance and area, For example: linear distance measurements can be caleulated using Pythagoras’ theorem (Fig. 2.9); polygon areas can be2 trapezia and then total zea; and the goometic cent of a ‘aking the mean ofthe coordinates of all smatives, sec Jones 1997, p. 66; Burrough 24 Data models and data structurs of spatial phenomena represent spatial data? The roots of GIS originate with the devel- idle ofthe last century. Inthe late 19508 8 for handling geogrephic information were ‘he digital representation other systems Die of astomated mapping (Foresman 1998, Tomlinson 1 “offered considerable time savings over raditional paper methods by prov tr ‘ancl more accurate facilities forthe management and updating of spt Syste lcd on point, ine and plygon‘gengrpbic primi representa only hasafnite set of resources with which complex world and, as a consequence, the digital ly schematic and genet this way is referred to as a da very si representations of reality, although as we shall 3c0 in later chapters, simple models ‘may become the building blocks for more nels that are designed t0 ‘quantify relationships between diferent. hapter 1, GIS represent spatial data using one or both of the Id data models. These are usually implemented as vector ing users the freedom thout necessarily using a different program, The following sections outline the differences between the vector and es, and provide some examples of ways that different sorts of daa are handled by each format, 24 Data models and data structures 25 Points Lines a od (or they have no length or and polylines (often also referred to as ares or edges) are one-dimensional vectors (having the property of length, but not breadth) defined by two ot more coordinate pairs. Polygons, or areas, are two-dimensional objects defined by three or mare coordinate pais. On the bass ofthis unigue toaset of additional 1 tiuibutes that desde addons roperes of tat object. These propectis most often consist of real-world quantitative and/or qualitative variables that give the vector object meaning within the GIS (Fig. 2.10). Vector topology An extremtely important concept that undetlis the vector structure isthe geomet- cal ips between vector objects, referred to as ‘of topotogical relationships is explored mare fa sufficient to note a few basic conceptsf the roads do share @ node, they aze lation of spatial relationships and is dependent both on the data structure and accuracy of properly maintained and defined. Many ines that ean be used to create topologies 24 Dara models and data structures 4 ig 2.12 Three topology elated plygons, Folygos 1,2 3 share ats (dees) ee ty des em, and Where this ocours are identified ‘and the appropriate action taken (eg. the over- polygon defined by the jent storage of vector data as ge ofthis when dfnitions by only recording an arc (and its vertices) once, ‘ad then defining its relationship o polygons. In Fig. 2. rj and mg need only be stored once instead of ti Chapter 4 2.42 The raster data structure use coordinate geametry to def Isohas a value associated. that location. Ina digital eleva 1g in comparison tothe vector ucture lies is suength, RasterPolygon ‘chapter, the raster data stmenute can be used to model some extremely complex. spatial phenomena. ce they define require a grid of 5000 columns by 5000 rows and therefore storage of informa for 25 000 000 pixels). Decisions relating to resolution need to be made Ver ‘on in the mod ‘can have important con for making sense of spat Where vector systems provide the only sensible means of answering 2 specific set of questions, or handling the sorts of lar and precisely defined data that one might be intoested in exploring. On the other hand, tial modelling, sable suitable for aerial photography and satellite imagery. Many modem GIS systems 2a Data models and data structures 2 ace toa large exter ‘Advantages and disadvantages of the vector structure ‘A major advantage of the vector required for the analy vel of precision 3 eset forthe study ofthe debt of chen fakes on inding the balance between spatial pre scale of analysis is cn tem recognition, In practic ase integration, attnbute queries can be performed wi veetormapmay therefore act asa window into database, in which each object is described in great etal Disadvantages of the vector: ing. Every vertex and node ofa vec ing vector objects ity. The most important imposition is "boundedness'. Although id objects do indeed have precise and discrete boundaries, certain yes tase moe zy and do olen theses ote ad sos ‘of vector objects. As vector data cannot readily deal with fuzziness ox in in mtfcil precision in some scenarios. An issue relatedFor example, tribute value thot represe This implies a continuous dist rarely the case (eg. Fi Advantages ofthe raster structure offers advantages for some ap ie i 5 e 5 realistically by er than a single value, When in relation othe abject being using raster mrespondingly increased, curved lines always polygon that arbitrarily defines the rganisation of culture ‘of GIS only define the starting point for exploring .y of the human use of space with GIS, In fact, many ofthese fist principles are being constantly challenged by research {hat is pushing beyond the constr vo-diqmensional mapping to use GIS to ‘model space-time relationships more adequately than the basic vector and raster this tye of information is excl, then the restr Fi advantage. Secondly | bined mor csi th j spatial modeling. A j consected by aaa Les that vector-based GIS programs provide. The raster in ql itations for the management and querying of multiscalar 4 4 ooo that offer &"32 Fist principles building blocks presented in this chapter. Nevertbeless, GIS is ~ for the time being atl lint on cartographic principles and a reductionist tendency that ites for representing and interpreting the ceal world, ver, there ill very broad range of ways that GIS tnd the next chapter provides some real-world examples of how GIS ean end does work in archaeology. Putting GIS to work in archaeology ‘This chapter reviews four typical applications of GIS in archacology: management ‘ofarchacological resources, excavation, andseape archaeology andthe spatial mod- elling of past human behaviour. For each app followed by a present ‘case sudy that illustrates the contribution that GIS has made. though ples are in no way exhaustive, they do provide @ good overview of the ‘and potential contributions that GIS ean make to archaeological managemet research, 3.1 Management of archaeological resources not our intention to discuss the objectives of cultural resource management nor the appropriate structure of aspatial database for managing the archaeo- cord as these decisions are most appropriately madeby goveramentbodies ‘hacologists charged withthe tasks ofzecording and managing the arcnae— al resource, However, we note that archaeological and historic databases have incseasingly been subject buildings, parks and gardens, historic aircraft crash sites, ete, Moreover, the role of HERS as essen- tial vehicles both for the management of the archasological record and for public education was emphasised, ‘maintenance by hindering standardisation (APPAG 2003), Several additional ‘ecommendations made fo the govemment to improve SMR include the need to strcamline systems and centralise access via the Internet in order to increase thei availability to the public (APPAG 2003, p. 19). 33