Regional patterns of local diversity of trees: associations with anthropogenic disturbance

Abstract

We used a probability-based sampling scheme to survey the forested lands of 14 states in five regions in the US (California, Colorado, and parts of the Southeast, Mid-Atlantic, and Northeast) from 1990 to 1993. Using a nationally consistent plot design, we evaluated the local diversity of trees over 2.5 cm in diameter at breast height (dbh) at 780 1 15 - ha plots nationwide by measuring the plot-level species richness ( R). Visually evident anthropogenic disturbances (e.g. artificial regeneration, logging, grazing by livestock, and prescribed burning), if any, were recorded on each plot. We classified plots with visually evident anthropogenic disturbance as ‘disturbed’ and the remaining plots as ‘undisturbed’. In each of the five geopolitical regions, we quantified the difference in mean R between disturbed and undisturbed plots. With the exception of Colorado (5%), between 34 and 55% of forested lands in each region had recorded anthropogenic disturbances. Mean R was significantly higher for undisturbed areas than for disturbed areas in the Northeast and Southeast, with the largest differences occurring in the Southeast. Mean R was greater in undisturbed areas than in disturbed areas in most forest cover types for all regions. These differences were greatest in the loblolly pine ( Pinus taeda), oak ( Quercus spp.)-hickory ( Carya spp.), and oak-pine forest types of the Southeast. The only group for which mean R was significantly greater in disturbed areas was the mixed western hardwoods in California. As expected from previous studies, significant differences between regions in mean R were observed, in both disturbed and undisturbed areas. This study bridges an important gap between site-specific forest studies and remote-sensing studies of the forests of a region. We discuss (10 why combining site-specific studies is not appropriate in most cases for rigorous testing at the regional level and (2) how data for some important site-specific variables are not available from most remotely-sensed data sets. The widespread presence of anthropogenic disturbances in most regions, notably the cutting and planting of pine plantations in the Southeast, is associated with generally lower local species richness of trees. The results warrant further investigation at the regional level in light of recent empirical studies on diversity and ecosystem stability.