Biophysical and socioeconomic drivers of global land-use change: A global-local-global linkages approach

Abstract:

Sustainable use of the world’s land resources depends on national, regional or local circumstances, requiring fine-scale analyses of land-use/cover change (LUCC). The key modeling challenge lies with the linkage/transfer functions between global drivers and local responses, which is capable of embodying economic information stimulating local responses to global changes and could be directly embedded into a global model. There is a large body of literature on projections of LUCC based on general- or partial-equilibrium impact assessment models. The projections are typically long-run, the local LUCC is typically at the national or sub-national level, and validation of such models is difficult to implement. The predictions of LUCC at the fine grid-cell level are typically implemented using the cellular automata models and earth observation datasets. This type of models is suitable for individual cities and small regions, but unable to serve as a linkage function between global drivers and local responses so as to be directly embedded into a global model. The aim of this research is to establish a linkage/transfer model between global drivers and local responses at the fine grid-cell level. The model can be validated at the grid-cell level for a region, country, and the world, and is able to predict LUCC at the grid-cell level on an annual time-step. Its basic setting is a Logit-linear transformed Share Model. We estimate the gross economic benefits of land-use conversion using the estimated annual value of potential gross revenues over the areas of rival land use types. This is done with the help of the global agro-ecological zone (GAEZ) model. We proxy for fixed and variable costs of land-use conversion using a constant term and a linear combination of biophysical variables which characterize the biophysical features of the grid-cell. The model was trained using 2000-2010 Land-Use Harmonization (LUH2) consolidated observation data on land-use. The fitted model is then used to predict land-use shares at the grid cell (0.25 degree) level in 2011-2014 for each of the 18 global agroecological zones as employed in GTAP-AEZ model. Results indicate the grid-cell level regressions with 4 land-use types of cropland, forest, grassland, and built-up land have strong prediction power.