Species density per grid cell no longer predicts the local abundance of woody plants

Abstract

Species abundance calculated from the number of individuals per hectare in phytosociological surveys is a key parameter to measure plant population size. The species density calculated by the occurrences per grid cell provides a proxy of population size when the species abundance is lacking, but its efficiency as a population parameter is poorly understood. Here, we quantified the species density of 79 woody species of the Cerrado. We tested to what extent species density quantified from grid cells could predict species local abundance. Species density per grid cell was calculated using herbarium data and data from published papers for six spatial resolutions, ranging from 30 s to 1°. General additive models and Spearman's correlation coefficient were used to test the species abundance-density relationship. Although we found a correlation between abundance and density, the model fit ranged from poor to moderate. Herbarium and paper data together improved the species abundance-density relationship relative to herbarium data only. Resolutions between 30-min and 1° produced better fits than fine resolution. We concluded that species density per grid cell no longer predicts the local abundance of woody species satisfactorily, multiple data sources are better than a single one, and the coarse resolutions are more suitable for the predicted species abundance-density relationship.