Trajectory analyses of past human land use as a tool to understand present terrestrial mammals' distribution

Abstract

Apart from the known environmental factors influencing species distribution at the global scale, additional anthropogenic factors need to be considered in order to understand their current extents of occurrence. Namely, threatened species should particularly be influenced by human activities, since they are known to be in decline under their influence. The role of present human activities in modeling species distribution seems clear, but there is still a facet of land use that has not yet been sufficiently explored: land-use history. In the present work we show how to summarize overall land-use trajectories based on available data ranging from c.B.C.6000 to c.A.D.2000 by using a clustering method. Then, we explore how indicators of total land use area at different time spans, rates of land-use change, or the occurrence of remarkable land-use shifts associate to the current distribution of total and threatened mammal richness by means of boosted regression trees. We find that we can separate the Earth surface into three groups, according to their trajectory of past land use: low-, recently- and steadily-used areas. Differences in total or threatened mammalian richness among clusters are not significant, however differences in the shape of the past land use- species richness relationship are relevant. This finding warn us about the risk of interpreting certain patterns as global rules, when they may actually depend on the past land-use history of an area. On the other hand, past land-use metrics are little relevant in predicting their distribution compared to latitude or total mammalian richness. Nevertheless, past land-use indicators are more informative than present land-use values, which may be pointing at the existence of a generalized extinction debt in the whole planet Earth regarding mammals. In conclusion, understanding general past land-use changes improves our ability to interpret current species distribution patterns.