Linking Land Use Patterns to Water Quality in the Hudson Valley

In consultation with Hudson River Estuary Program staff, IRIS correlated landscape patterns within the Wappinger Creek watershed with water samples obtained along Wappinger Creek and its tributaries with the objective of determining how the scale of the land cover data influences the correlation between landscape patterns within a watershed and ecological indicators of water quality in Hudson River tributaries. Within-stream water quality indicator data were organized into three major water quality indicator categories: benthic macro-invertebrate, fish and nutrient/biological. Multiple regression models were used to statistically correlate water quality data to area of land cover at two levels of resolution. Results indicate that a quantifiable link exists between the area of land cover and within-stream water quality indicator data in the Wappinger Creek watershed exists, and the relationship is affected by the spatial properties of the land cover data sets. Important findings include the observation that the lower-resolution land cover map (i.e., larger minimum mapping unit) with a higher-resolution land cover classification scheme generates the strongest relationships between water quality variables and area of land cover, while a watershed level analysis produces stronger relationships than a sub-watershed level analysis. The regression analyses reveal a strong correlation between four land cover types (herbaceous, agriculture, water and barren) and water quality. This indicates that these land cover types are notable landscape-level ecological indicators of water quality and should be considered as important contributors to the ecosystem services of local landscapes.


Dr. Magdeline Laba, ml49 [at] (ml49[at]cornell[dot]edu)