Why are there vast differences in agricultural activity across space? How do these differences shape the aggregate impact of agricultural shocks? To address these questions, I build a quantitative general equilibrium model that accounts for rich spatial differences in agriculture and use comprehensive county-level data from Brazil to estimate the model. I find that differences in natural advantages and factor intensities are key causes of the spatial patterns of agricultural specialization but that differences in trade costs across crops play a minor role. In addition, I study two major shocks in Brazilian agriculture: the adaptation of soybeans to tropical regions and the rise in the Chinese demand for commodities. The results show that general equilibrium effects substantially shaped the returns to agricultural research, the impact of tropical soybeans on urbanization, and the gains from trade with China.