The Delaware Water Gap National Recreation Area (DEWA) is a long, thin park encompassing 28,000-ha of land that straddles the Delaware River in northeastern Pennsylvania and northwestern New Jersey. On the Pennsylvania side of the river, only 14% of the drainage area of the 17 tributary watersheds is within park boundaries, and dramatic increases in urban land cover and human population have been observed adjacent to the park. Consequently, stream communities may be adversely affected by increased pollution, sediments, and altered temperature and flow regimes. We used benthic macroinvertebrate (BMI) communities as biological surrogates for the broader stream community to evaluate changes in stream condition and biological diversity associated with an urbanization gradient in contributing watersheds. We sampled BMI communities from 33 sites in 17 watersheds and used the data to inform stress-response models of abundance and distribution of component BMI taxa. Subsequently, we used response patterns derived from these statistical models to establish two utility thresholds of urban development explicitly designed to link urbanization patterns to the primary management goal of protecting aquatic diversity. First, an early-warning utility threshold of urban development was identified where a significant fraction of the BMI community exhibited changes in density across the urban gradient, but prior to evidence of taxa loss. Second, an extinction utility threshold of urban development was identified where a significant fraction of the BMI community showed evidence of local extinction. We then used these utility thresholds to estimate the current distribution of stream reaches in the park exhibiting decline.
Results/Conclusions
We found that stress-response models based upon spatially-explicit summaries of urbanization (i.e. urban patches weighted by flow-distance to sample sites) were substantially more sensitive in detecting change in BMI taxa than models based upon traditional summaries of urban land cover (i.e., cumulative total of land use above sample site). We identified an early-warning utility threshold of 1.5% urban where 22% of BMI taxa pool (24/109 taxa) showed changes in density; and an extirpation utility threshold of 9% urban where 6% of BMI community showed strong evidence of local extinction. Using these utility thresholds, we estimated that over 23% of the stream reaches in DEWA already show early-warning evidence of decline and slightly more than 1% of stream reaches show evidence of local extirpation. However, observed declines were not spatially uniform, varying widely among watersheds. Our analyses provide a spatially-explicit assessment of urbanization effects useful for prioritizing conservation and restoration efforts.