Understanding the impacts of changing climate on human systems and decision making is complicated by differences in spatial scales among economic and biophysical models. Earth System Models (ESMs) operate on gridded representations of the globe while Integrated Assessment Models (IAMs) generally operate on coarser, geographically delineated land units. This study is a first step toward addressing the uncertainties associated with this mismatch in spatial scale by comparing the effects of different spatial initial conditions for the Global Change Assessment Model (GCAM) on land use projections. Three initial conditions-agricultural production, harvested area, and land rent-have different global distributions when calculated for two different sets of land units: one based on current climate (151 land units) and one based on year 2100 climate (184 land units). These different land unit sets will likely result in different land use trajectories through 2100, mainly due to changes in within-land unit climate heterogeneity over time. The current climate land units become more heterogeneous and the future climate land units become less heterogeneous as time progresses. Ongoing work in a coupled model framework will examine how these two land use trajectories differentially affect climate and vegetation productivity estimates in the Community ESM (CESM) and the associated feedbacks to GCAM energy and land use projections. As different land use trajectories may generate significant differences in global carbon cycling and climate, the presented results have broader implications for understanding policy scenarios and potential impacts and also for evaluating and inter-comparing IAMs and ESMs.