Species persistence and mean rank abundance in global Nutrient Network plots from 2007-2019

This dataset uses data from the NutNet dataset to examine how temporal and spatial rarity are related, how temporal and spatial rarity predict species loss in ambient and experimentally perturbed plots. We use data from all sites that had NPK and/or fencing treatments with a minimum of five years of cover data when data was downloaded on August 2, 2019; 49 sites in all were used. Perturbations were NPK treatments (nitrogen, phosphorus, potassium and micronutrients) and fencing (vertebrate herbivore exclusion). Temporal rarity was assessed as the percentage of years a species was found in a plot (calculated in R script) and spatial rarity was assessed as the mean rank percentile of a species across all years in a plot (included in data table). We found that persistence (i.e. temporal rarity) was a better predictor than local abundance (i.e. spatial rarity) of whether a species would be absent in a neighboring plot, despite the rarity axes being correlated. Additionally, perturbations reduced persistence most strongly in highly persistent species and low abundance species, further suggesting these are unique dimensions of rarity.