The patterns and processes underlying the distribution of trait values within plant communities are currently poorly understood. One cause of the non-random trait distributions seen in nature is the environmental restriction of viable persistence strategies. Under environmental land-use change, trait value distributions are likely to shift resulting from changes in the range, variance, and/or kurtosis of trait states that are viable in the disturbed system. Such changes may have important implications for the ecosystem functioning of the location. Species diversity usually changes under land-use alteration and is accompanied by shifts in the distributions of key traits. These may be changes in the whole range of trait values (neutral loss) or in selected portions of a trait’s distribution (non-random loss). Trait changes may reflect species turnover whereby new colonizers share traits with the existing species pool, or introduce novel trait values. The conditions under which these scenarios occur are poorly understood.  In this study, we use a synthetic approach to explore how trait distributions respond to common types of land-use change. To do this we examine trait data from plant communities in 18 landscapes from around the globe. These allow for multiple comparisons of undisturbed forest with three common land-use categories: abandoned pastures regenerating to forest (“Old-field”), forests regenerating immediately after logging (“Legacy”) and forests permanently converted into pasture (“Pasture”). The traits that we examine are: Height, specific leaf area, seed mass, wood density, growth form, and dispersal mechanism.

Results/Conclusions

The trait distributions found in individual comparisons between undisturbed forest and land-use categories were quite variable, though across landscapes some significant average patterns emerged. For example, median height class shifted toward smaller species in “Pasture,” driven by a loss of larger size classes (upper quantile significantly decreased), did not change in “Old field” systems and significantly increased in tropical but not temperate “Legacy” systems. For seed mass, however, we found no median change in “Pasture” and “Legacy” systems but there were significant losses of the largest seed classes in these land-use categories; the reverse was true (more larger seeds) in “Old field” systems.  Our study provides evidence for specific functional changes occurring in a diversity of forest systems in response to common land-use changes.  We explore the ecological mechanisms likely involved in the observed patterns.  This is a major advancement in our mechanistic understanding of changes in community species and functional diversity following land-use alteration.