Fig. 1: Quantifying resistance, return rates and recovery times of biodiversity in a tropical rainforest.
a, For any recovering system attribute (for example, the diversity of a certain animal group), the resistance, defined as the amount remaining during perturbation (clear cut and agriculture) and the return rate towards the predisturbance reference state (old-growth forests) determine the recovery time. We fit a negative exponential function to the empirical data (a) to interpolate between the values in the agricultural plots \({\varPsi }_{0}\) and the old-growth forest plots \({\varPsi }_{{\rm{OG}}}\) with the return rate λ (Methods). The recovery time (Trec) is calculated as the time difference between the intercept of this function and the value at 90% of the median of old-growth forest plots, represented by dotted lines. Blue dots represent species composition data for bees (independent samples from n = 62 plots). Boxplots mark active agriculture (n = 12) and old-growth forest (n = 17) plots (orange line shows median, boxes show data in 25th and 75th quartile and whiskers indicate 1.5× the interquartile range). An example calculation for recovery time, resistance and the return rate for the bee dataset can be found in Supplementary Note 1. b, The resistance and the return rate independently decrease the time until recovery to 90% of reference conditions (recovery time Trec). We predict that communities of mobile animals recover faster, particularly as a result of a high return rate, than most trees, which have a much later age at first reproduction. a, Forest image reproduced with permissions from ref. 53, Ecological Society of America, under a CC BY-NC 4.0 licence. a,b, Silhouettes were reproduced from PhyloPic (https://www.phylopic.org/): frugivorous bird, created by E. Price under a CC BY 4.0 licence; tree, created by T. M. Keesey under a CC0 1.0 Universal Public Domain licence.
