Surrounding land cover types as predictors of palustrine wetland vegetation quality in conterminous USA

Highlights: • We calculated 2 indices of wetland vegetation quality at 380 sites in 4 regions of USA. • 8 land cover types in 4 zones surrounding wetlands were used to predict the indices. • Forest, followed by wetland, had the greatest overall positive effect on the indices. • Agriculture had the greatest overall negative effect on the indices. • Forest buffers and wetland contiguity should increase regional vegetation quality. The loss of wetland habitats and their often-unique biological communities is a major environmental concern. We examined vegetation data obtained from 380 wetlands sampled in a statistical survey of wetlands in the USA. Our goal was to identify which surrounding land cover types best predict two indices of vegetation quality in wetlands at the regional scale. We considered palustrine wetlands in four regions (Coastal Plains, North Central East, Interior Plains, and West) in which the dominant vegetation was emergent, forested, or scrub-shrub. For each wetland, we calculated weighted proportions of eight land cover types surrounding the area in which vegetation was assessed, in four zones radiating from the edge of the assessment area to 2 km. Using Akaike's Information Criterion, we determined the best 1-, 2- and 3-predictor models of the two indices, using the weighted proportions of the land cover types as potential predictors. Mean values of the two indices were generally higher in the North Central East and Coastal Plains than the other regions for forested and emergent wetlands. In nearly all cases, the best predictors of the indices were not the dominant surrounding land cover types. Overall, proportions of forest (positive effect) and agriculture (negative effect) surrounding the assessment area were the best predictors of the two indices. One or both of these variables were included as predictors in 65 of the 72 models supported by the data. Wetlands surrounding the assessment area had a positive effect on the indices, and ranked third (33%) among the predictors included in supported models. Development had a negative effect on the indices and was included in only 28% of supported models. These results can be used to develop regional management plans for wetlands, such as creating forest buffers around wetlands, or to conserve zones between wetlands to increase habitat connectivity.