Item

Abstract

The global integrated land use model, LandSHIFT, is evaluated by testing its performance against available data sets, by analyzing the sensitivity of model parameters and structure, and by conducting a scenario analysis of future land use change in Africa. Despite the paucity of suitable data sets, a range of different tests were designed to make best use of available data and to examine the model's ability to compute cropland suitability, extent of cropland area, and location of deforestation. The model showed more ability to calculate the spatial distribution of cropland suitability and continental average deforestation rates than to compute the spatial distribution of deforestation. LandSHIFT was found to be particularly sensitive to assumptions about future climate change for simulations extending over several decades through the influence of climate on cropland and grassland productivity. With regards to the scenario analysis, the model was applied to two scenarios for Africa that cover a wide range of assumptions about future driving forces. Results showed that cropland land may expand greatly up to 2050 (34-40%, depending on the scenario) because of increasing food demand and despite expected increases in crop yield. This expansion comes largely at the expense of forested land, although the average continental deforestation rate computed from 2000 to 2050 is lower than the computed rate for the 1990s. The testing and scenario analysis showed the ability of the model to develop consistent scenarios of land use change on the continental scale by combining the effects of driving forces and competition between land uses in a single spatially-explicit framework. (C) 2011 Elsevier Ltd. All rights reserved.