Modelling is essential for the analysis, and especially for the predictions, of the land use dynamics. Urban and regional evolution models usually internalize general and known characteristics for an area such as employment, population growth, GDP, roads constructions, etc. The model structure incorporates the these general characteristics as well as the local variations expressed by numerical values, which allow to reproducing more accurate the real patterns and reuse from region to region (Birkin et al., 1996). For a specific region context the urban model has to build a general-purpose model and use a technique in modelling called calibration. The key element of the modelling process is to understand the calibrated weight in order to obtain the best statistical model fit of land use based on assessments of economic, political, social, technological, and environmental trends. Calibration plays an important role, not only by allowing an improved fit between the simulation and the actual data by refining the control values, but also it allows us to see clearly how these self-modification rules contribute to the cellular automata (CA) behaviour (Silva and Clarke, 2002; Clarke et al., 1998). The calibration process is computationally intensive, and has led to experiments and methods to speed up the routines, including parallel processing and supercomputing.
A successful model application is also to understand what kind of information can be expected from the model and evaluating the strengths and weaknesses of different modelling approaches for sustainable development policy-making (Boulanger and Brechet, 2005).
Land use transformation models can therefore generate data of meaningful representations of the region’s characteristics and allow the processing of different data sets. The models contribute to understand the landscape changes and drivers of the dynamics in the development conditions of each study area. Its also help to answer where and at which intensity land-take for urbanization occurs and how spatial growth patters alter over time; how urbanization (e.g. sprawl) affects large areas overruling local and regional decision.
The MOLAND model for urban and regional scenario simulation (Barredo et al., 2004; Lavalle et al.,2004) is used to evaluate spatial planning for sustainable urban and regional development; the model predicts the likely future development of land use, for each year usually over the next ten to twenty-five years. MOLAND model is operating at both the micro-and macro-geographical levels. At the macro level are integrated several component sub-models, representing the natural, social, and economic sub-systems typifying the area studied. These are all linked to each other in a network of mutual reciprocal influence. At the micro level, cellular automata based models determine the fate of individual parcels of land based on their individual institutional and environmental characteristics as well as on the type of activities in their neighborhoods. The approach permits the straightforward integration of detailed physical, environmental, institutional variables, and the particulars of the transportation infrastructure.
This technical note describes the preparatory work of data collection need for the definition of scenarios of development for the Algarve Province in Portugal.
In a successive phase, these scenarios will be implemented by mean of the MOLAND Model.