Applying Occam's razor to global agricultural land use change
Reference: Engström, K., Rounsevell, M.D.A., Murray-Rust, D., Hardacre, C., Alexander, P., Cui, X., Palmer, P.I. & Arneth, A.
Environmental Modelling & Software 75: 212-229.
We present a parsimonious agricultural land-use model that is designed to replicate global land-use change while allowing the exploration of uncertainties in input parameters. At the global scale, the modelled uncertainty range of agricultural land-use change covers observed land-use change. Spatial patterns of cropland change at the country level are simulated less satisfactorily, but temporal trends of cropland change in large agricultural nations were replicated by the model. A variance-based global sensitivity analysis showed that uncertainties in the input parameters representing to consumption preferences are important for changes in global agricultural areas. However, uncertainties in technological change had the largest effect on cereal yields and changes in global agricultural area. Uncertainties related to technological change in developing countries were most important for modelling the extent of cropland. The performance of the model suggests that highly generalised representations of socio-economic processes can be used to replicate global land-use change.