This study seeks assess how the uncertainties associated with the un- derlying biophysical processes in uence the optimal profile of land use over the next century, in light of potential irreversibility in these deci- sions. Our analysis is based on a dynamic stochastic model of global land use, and employs 3 modeling scenarios constructed using global crop simulation and climate models. The results of the deterministic model show that climate impacts appear to have mixed effects on yields - higher temperatures hurt food production but this effect is partially offset by greater CO2 fertilization effect. Declining food crop yields result in rela- tively small expansion of cropland and accumulated GHG emissions from land use change. We then contrast this optimal path to that obtained when the uncertainty is not ignored, thereby demonstrating significance of factoring uncertainty in the optimization stage.