Fig. 3
Most important drivers of woody plant cover change. a-e Boosted regression tree partial dependence of fractional woody cover change on selected explanatory variables, when accounting for the average effect of all other driver variables (see Supplementary Fig. 7). The red lines are smoothed representations of the responses, with fitted values (model predictions based on the original data) for each 0.5° grid cell over sub-Saharan Africa. The trend of the line, rather than the actual values, describes the nature of the dependence between response and explanatory variables. Small bats on the x-axis represent data deciles, and grey bands indicate data between the 25th and 75th percentile. The x-axis was clipped to the 5th and 95th percentile to highlight trends in the bulk of the data. The full model explained 51% of the total deviance in woody cover change, and the relative contribution (%) of each explanatory variable is indicated
