We build a model to study optimal land use, encompassing land use activities, pollution and climate change. This benchmark set-up allows us to identify the spatial drivers behind the interaction between land use and the environment.Pollution generates local and global damages since it flows across locations following a Gaussian Plume. In contrast to the previous literature on spatial dynamics, we prove that the social optimum problem is well-posed, i.e., the solution exists and is unique. We close the paper with a numerical analysis which illustrates the richness of our model, and its global dynamics. We study the different drivers of spatial heterogeneity. In particular, abatement technology stands out as a fundamental ingredient to achieve steady state solutions, which are compatible with the emergence of spatial patterns.