By Rohan Chandralal Wickramasuriya
Abstract:
The spatial policy decision making in Sri Lanka is still paper based, and not supported by any scientific means. However, land use models have been successful elsewhere in the world as Spatial Decision Support Systems (SDSS). METRONAMICA is the modeling framework behind some well-known land use models which serve as SDSS. The backbone of the METRONAMICA framework is a Cellular Automata (CA) model. The objective of this study was to find out how a land use model developed using METRONAMICA can enhance the spatial decision making in Sri Lanka. In order to achieve the objective, a land use model was set up (the RUHUNUPURA model) using the METRONAMICA framework, for the Ruhunupura area of Sri Lanka. The land use classification adopted includes eleven land use classes altogether, out of which only seven are dynamic. Four classes are static over the simulation period. There were three predefined algorithms available in METRONAMICA to calculate the transition potential of cells. All three predefined algorithms were unable to simulate the dynamics of the land use class chena properly. Chena is a characteristic cultivation method followed by the local farmers and is also known as the shifting cultivation. Therefore, a new algorithm was developed by incorporating a time component, especially to capture the dynamics of chena. The RUHUNUPURA model was then manually calibrated for the period 1985 to 2001. The Fuzzy Kappa statistic, visual interpretation, and wavelet verification were used to assess the results of calibration. The calibration results reveal that the model can be safely used to study the behaviour of the land use classes shrub and other uncultivated area, homesteads, chena, forest, and paddy. However, the model could not handle very well the dynamics of the land use class other crops. The calibrated model was then used to run four scenarios. The first scenario was a validation test for the model. The model could predict lengthening cultivation periods and shortening fallow cycles of chena, which is a well observed, distinct character of chena dynamics. The second and the third scenario were formulated to try out two alternative zoning policies for the modelled region. The aim of the fourth scenario was to foresee what might happen if another tsunami hits the coasts of Sri Lanka. The final task of the study was to assess the usability of the RUHUNUPURA model for the user organization, the Urban Development Authority of Sri Lanka. The usability assessment was conducted for a period of 3 weeks at the user organization with 10 participants. An extended workshop was the protocol for the usability assessment, which consisted of an introductory session, one week of guided training, non-guided practical exercise, group discussion, and an interview with the Director of the UDA’s GIS centre. During the assessment two questionnaires were used; the first questionnaire checked participants’ knowledge in land use modeling concepts, while the second questionnaire estimated the user satisfaction and users’ attitude towards the model. The second questionnaire was evaluated twice; once in the beginning and then at the end of the assessment to study the attitude change of the participants. The average time spent by the participants to complete the practical exercise was 24 minutes. The participants were very effective in handling the model where the average number of tasks completed by a participant was 7 (total number of tasks were 8). The user satisfaction was found fairly high by evaluating the questionnaire and through the comments obtained during group discussions. The user organization feels that the freedom the model offers to analyze alternative policies through scenario as the most important function of the model. The larger cell size (500 m) of the model was considered as the biggest disadvantage by the user. Overall, the user was satisfied about the RUHUNUPURA model.
Keywords: Spatial Decision Support System, Land use model, Cellular Automata, Chena, Calibration, Scenario, Usability