Dynamic simulation modeling of the land use, economy and environment in Chiang Mai, Thailand using GIS and remote sensing

Land use in Chiang Mai, northern Thailand is changing rapidly. These changes are associated with considerable degradation of the environment there. Remote sensing, geographic information systems (GIS) and spatial modeling such as (GEOMOD2) are important tools for analyzing, simplifying, simulating, communicating and presenting spatial and/or temporal trends in land use change. The main objective of my study is to analyze and integrate the socioeconomic and land use data in order to assess rate and pattern drivers for land use change. Then I link these drivers with a spatial model to predict the rates of forest area change and where the changes in deforestation are most likely to occur spatially. I used the GEOMOD2 model to simulate forest area change and geographic information systems (GIS) for spatial analysis. I validated the model by deriving a 1995 land use map from Landsat TM5 digital images. My analysis of the Landsat TM5 images showed that in 1995 Chiang Mai had a total forest cover area of 13,948 sq. km. (62.4 percent of the total province area). The overall accuracy of this image when judged against aerial photographs was 88 percent.; GEOMOD2 was able to simulate 73.6 percent of the 1990 forest area in Chiang Mai correctly based on lubrication values derived from one map (1973 forest cover map) and the rate of deforestation predicted from change in population (Kappa = 0.0946). GEOMOD2 predicted only 66.3 percent of the forest area in 1995 correctly using only one map. At the district level, GEOMOD2 was able to predict 80.6 to 88.5 percent of forest area correctly in 3 of 22 districts of Chiang Mai (Kappa between 0.4593 and 0.7113).; GEOMOD2 predicted 72.5 percent of the 1995 forest area in Chiang Mai correctly based on maps at two points in time (1973 and 1990 forest cover maps) when the rate of deforestation was extrapolated from 1973 and 1990 forest cover maps (Kappa = 0.3635). At the district level, GEOMOD2 simulated 80.3 to 92.4 percent of 1995 forest area correctly in 11 of 22 districts of Chiang Mai (Kappa between 0.6038 to 0.7632).; In conclusion, GEOMOD2 was reasonably able to predict more than 70 percent of the spatial patterns of forest area change correctly in Chiang Mai, northern Thailand.