Abstract:
Land-water interactions were examined in three regions in the Virginian Biogeographic Province; the southern shore of Cape Cod, Massachusetts; the Hudson/Raritan region of New York; and the eastern shore of the Delmarva (Delaware/Maryland/Virginia) Peninsula. Cumulative distribution functions were used to evaluate similarity in environmental condition among estuaries. Spatial-setting variables (location in a river, coastal lagoon, or in open waters) were associated with variation for some measures of estuarine condition. Patterns of coastal urban and agriculture gradients were measured and their relationship with indicators of estuarine condition was modeled statistically. When estuaries were pooled, the highest variation explained by spatial-setting variables was found for dissolved oxygen (DO, R 2 = 0.44) and salinity (R 2 = 0.58), with DO decreasing in river locations and salinity decreasing with rainfall and sampling locations near rivers. The explanatory power for the other indicator variables was low and varied from 6% to 27%. Rainfall explained some of the variation (R 2 = 0.23) in total suspended solids. Moderate (0.4 < | r | < 0.7) to strong (| r | = 0.7) linear associations were found between total urban area and measures of estuarine condition. Within regions, total urban area was positively associated with Silver (r = 0.59), Cadmium (r = 0.65), and Mercury (r = 0.47) in Cape Cod, and inversely related to DO (r = -0.65) in the Hudson/Raritan region. No associations were found in the Delmarva Peninsula study area. Total area of agriculture showed a moderate association with Arsenic in Cape Cod, but no other associations were found in the other two regions. Our analyses show a measurable impact of urban land use on coastal ecosystem condition over large areas of the northeastern United States. This pattern was most evident when many different landscapes were considered simultaneously. The relationship between urban development and estuarine condition were weaker within the individual regions studied. The use of land use/cover models for predicting estuarine condition is a challenging task that warrants enhancements in the type, quantity, and quality of data to improve our ability to discern relationships between anthropogenic activities on land and the condition of coastal environments.
