This study explores how deforestation relates to biophysical and socioeconomic variables in a quantitative, spatially explicit analysis. Both patterns of historical (approximately 4000 B.C. – 2000 A.D.) and recent (1990–2005) deforestation were explored and compared. The study uses location analysis, combining spatially explicit global databases of deforestation with variables hypothesized to relate to deforestation (e.g. rural livestock density, cost-distance). The results show that historical and recent deforestation show similar patterns when plotted against the selected variables suggesting the relations to be empirical. The strongest relations were between deforestation and rural population density, cost-distance, and crop suitability respectively. These findings support the hypothesis that agricultural expansion was the largest direct cause of deforestation historically and it continues to be the largest direct cause today. While crop suitability showed a strong correlation with deforestation, pasture suitability showed no correlation. The study accordingly compares contemporary deforestation in the tropical domain against economic development (approximated by a combination of GDP value and growth rate) which resulted in a reproduction of the deforestation pattern predicted by the forest transition curve. In view of the apparent empirical relation, the variables which relate to deforestation in this study are expected to have predictive power to identify locations of likely future deforestation. To illustrate the application value of this analysis, the deforestation risk in an African protected forest was assessed. This assessment revealed that protected areas are found in locations with below-average pressure and the pressures vary greatly in the different regions. The findings underline the importance of considering national circumstances to understand the government’s efforts to manage its forests sustainably and contribute to international mitigation by reducing forest-related emissions.