10.1371/journal.pone.0177552.g005 Jonas Nüchel Jonas Nüchel Jens-Christian Svenning Jens-Christian Svenning Partial dependence plots of the variables in the random forest model on county scale. Public Library of Science 2017 results show afforestation programs land abandonment dynamics anthropogenic predictor variables 2017-06-07 17:27:40 Figure https://plos.figshare.com/articles/figure/Partial_dependence_plots_of_the_variables_in_the_random_forest_model_on_county_scale_/5087494 <p><b>a)</b> slope, <b>b)</b> elevation, <b>c)</b> GDP per square kilometer, <b>d)</b> actual evapotranspiration, and <b>e)</b> change in population density between 2000 and 2010. The ticks inside the graphs indicate the deciles for the data. The x-axis in partial dependence plot <b>c)</b> has been cut off so big outliners above the 9<sup>th</sup> decile are not shown. See supporting information <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0177552#pone.0177552.s007" target="_blank">S6D Fig</a> for the complete partial dependence plot of GDP per square kilometer. The x-axis in partial dependence plot <b>e)</b> has been cut off at the 1<sup>st</sup> and 9<sup>th</sup> decile, so the graph only shows the mid 80 percent of the data. See supporting information <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0177552#pone.0177552.s007" target="_blank">S6C Fig</a> for the complete partial dependence plot of change in population density between 2000 and 2010.</p>