Using multisource data and the V-I-S model in assessing the urban expansion of Riyadh city, Saudi Arabia

This paper examines the application of remote sensing, based on the Vegetation-Impervious surface-Soil (V-I-S) model and spatial metrics, in an urban analysis for promoting sustainability and understanding urban growth theory. In order to improve the accuracy of land-cover classification, spectral angle mapping (SAM), spectral mixture analysis (SMA) and band ratioing were applied on satellite images for land-cover classification and comparison of the discrimination efficiency of these techniques. For the SMA, subsets of the Landsat (2, 3, 4, 5, and 7) and ASTER (1, 2, and 3N) images were selected. After endmember extration and purification, the Bayesian probability of each component was computed and used for the spectral unmixing. The classified images of different years were compared to analyze the changes in land use and spatial pattern using V-I-S, a form of percentage of landscape (PLAND) and annualized urban sprawl index (AUSI). The result indicates that the performance of band ratioing (69% accuracy) is not as good as that of SAM (75%) and SMA (86%) in discriminating between vegetation and agricultural land. The land-use analysis results denote that urban growth management strategies in Riyadh have not been completely successful and the growth pattern substantiates the urban theory of diffusion and coalescence.