Simulating global agricultural land-use change as local economic decisions in a multi-model framework: strengths, limitations and ways forward

We present a multi-model land-use modelling framework simulating global agricultural land-use patterns driven by local economic decisions on land use. The proposed framework coherently integrates a series of outputs from various specialised models (climate, gridded crop growth, computable general equilibrium) to account for the complex cross-scale interplay between land-use drivers operating in socio-ecological systems. In our framework, land-use changes emerge from decisions farmers make in managing agricultural land while seeking to maximise their utility. In turn, the utility that farmers are able to derive in each location depends on local opportunities and constraints resulting from a combination of biophysical, socioeconomic, technological and political factors that jointly determine local and regional competitive advantages of alternative land-use options. Hence, rather than extrapolating observed land-use patterns into the future, the framework aims at capturing the causality between land-use change drivers and processes, and the behaviour leading to land-use decisions. The modelling framework was operationalised and validated in MagnetGrid, a newly developed spatially-explicit land-use module of the MAGNET modelling platform. MagnetGrid combines MAGNET’s regional projections of agricultural sectors with scenario-based agro-ecological crop suitability maps to project future local utility of different production options, allocating agricultural land accordingly. We demonstrate the ability of MagnetGrid to explicitly simulate the effects of a large number of scenarios, policy alternatives and land-use discontinuities: climate change impacts in agricultural production; shifts in societal demands (e.g. dietary preferences); the emergence of new land-use options (e.g. second generation biofuels); the effects of non-spatially-explicit drivers on local land-use decisions (e.g. market prices, production costs, subsidies, taxes); and the adoption of innovative practices and adaptation measures. Finally, we critically discuss the strengths and limitations of the modelling framework for both global and local assessments of land-use change, and propose ways to address current challenges in capturing feedbacks and farmer behaviour in land-use projections.