Simulating urban growth dynamics of Dhaka metropolitan area: a cellular automata based approach

Abstract

Fore8eemg urban growth and accordingly predicting fllture urban extent is ah,ays accepted as it vital issue lor any urban planning decision making process, Diffcrent leclmigue,> and models have alr{;d.liyhecn developed for Simulating the dynamics of urban growth that can bc employed for this prediction process. This study applicd the concept of Cellular Automata (CA) to simLilalc the historical growth paUem of Dhaka Metropolitan Area (DMA), It has calibrated the SLEUTH CA based model for Ihis and predictcd the f,ducc growth scenario for dlITcrent policy decisions in this urban agglomcration. The theory of Cel1ular ALilomaton (CA) has been wlde]y uscd to Slm\Llatc urban gro\\1h for pasl few ye"rs, It requires that the space shoLild hc represented as a grid of cells that can change stale as thc model,terates. These changes are regulated by rules that specify a set of neighborhood conditions to be mCI before" change in state can occur. Urban landscape can be lessellated according to its land Liseor spatial extent 01' development, Th,s cell-based framework can be utilized 10 apply the theory of automata and therehy to model its complex struclmc using simple lransition rules. Tilis study found that CA applies the samc transition niles throughoutlhe space. As a result same growth rate happens tnroughout the cell space irreRpcctive of their 10ca[lOn in relation 10 the exisling urban areas. This study mtroduced a travel eost approach 10 the model which assumes Ihat development probability wlll decrease ibrough space with increased dlStanec (and incrcased travel dliTiculty due 10 lopography) from the existing urban centers. lt introduced Ihe travel cost a, a pan of the eXCI\l~lOn layer or Ihe model. Through cahbTalion of the two modding fi:a1n~works (one wlth travel co,t and one withoLit travel cost) It was found that the model with travel CQRthas better capahility to Sil]1\Llatcthe historical growth pattem of the silidy area. As u result, it employcd the model with travel cost for future gro\\,th prediction of the study area for different poliey scenarios. VIl Through observation of the best fit coefficients of the model it was fmInd that the ~tLLdyarea is expericncing high in-fill and edge growth throughout the urban centers. Diffused growth is also occurring; significantly at the viClmty of the urban centers. The~e diffused ;etllcmcnts have some significant probability to evolve further as urhan centers, Existing road networks are also significantly influencing the overall groVv1hprocess of the study arca. The best fit coefficicnts were applied for future growth predic\lon ofthc study area Lor the period 2008-1030. D,rrcrent policy scenarios were developed based on the de\,elopmcnt plans and policies for the study area (i.e. DMDP and STP) It was found that throllgh Introduction of new roads, the study area will experience an accelerated growth throughout the prediction time period. Through impOSItion or environmental restrictions, the study area will experience decelerated growth in future, lmposition of development rcstriction on agricultural land also will significantly retard thc overall growth of the study area. Imposition of all restriction and mtroduction of the new roads also shows a reduced growth trend throughout the prediction time period. Tluollgh analysis of the calibration results and thc predIction capability of the model, this sllldy has sho",ed that CA is an effectivc approach to simulatc urban growth dynamics of DMA. $ustainability of urban growth can hc cnsured tluough propcr apr lication of this kind of model in the planning dccision making of this urban region