Implications of land use change in tropical northern Africa under global warming

Abstract. A major link between climate and humans in tropical northern Africa, and the Sahel in particular, is land use and associated land cover change, mainly where subsistence farming prevails. Here we assess possible feedbacks between the type of land use and harvest intensity and climate by analysing a series of idealized GCM experiments using the Max Planck Institute Earth System Model (MPI-ESM). The baseline for these experiments is a simulation forced by the RCP8.5 (radiation concentration pathway) scenario, which includes strong greenhouse gas emissions and anthropogenic land cover changes. The anthropogenic land cover changes in the RCP8.5 scenario include a mixture of pasture and agriculture. In subsequent simulations, we replace the entire area affected by anthropogenic land cover change in the region between the Sahara in the north and the Guinean Coast in the south (4 to 20\degree N) with either pasture or agriculture. In a second set-up we vary the amount of harvest in the case of agriculture. The RCP8.5 baseline simulation reveals strong changes in the area mean agriculture and monsoon rainfall. In comparison with these changes, any variation of the type of land use in the study area leads to very small, mostly insignificantly small, additional differences in mean temperature and annual precipitation change in this region. These findings are only based on the specific set-up of our experiments, which only focuses on variations in the kind of land use, and not the increase in land use, over the 21st century, nor whether land use is considered at all. Within the uncertainty of the representation of land use in current ESMs, our study suggests marginal feedback between land use changes and climate changes triggered by strong greenhouse gas emissions. Hence as a good approximation, climate can be considered as an external forcing: models investigating land-use–conflict dynamics can run offline by prescribing seasonal or mean values of climate as a boundary condition for climate.