| |
Presented at the NABS Annual meeting, Vancouver, British Columbia, 2004
in Disturbance Gradients
Use of an integrated flow model to estimate ecologically relevant hydrologic characteristics at biomonitoring sites in New Jersey
J.G. Kennen1, L.J. Kauffman1, M.A. Ayers2, and D.M. Wolock3. 1U.S. Geological Survey, 810 Bear Tavern Rd., Suite 206, W. Trenton, New Jersey, 08628, 29818 Bluegrass Parkway, Louisville, Kentucky, 40299, 34821 Quail Crest Place, Lawrence, Kansas, 66049
A physically based hydrologic model was developed to provide values for a comprehensive set of hydrologic measures representing the major features of the flow regime at 856 biomonitoring sites in New Jersey. More than 78 parameters accounting for the frequency, magnitude, duration, and timing of flow events were modeled. Redundant variables were filtered with principal components and correlation analyses, and a non-redundant subset was used to determine which variables significantly affected aquatic invertebrate assemblage structure. Variation in assemblage structure was examined using non-metric multidimensional scaling (NMS). Multiple linear regression (MLR) analysis was used to build exploratory models that identify the most significant environmental and hydrologic variables. The first axis of the NMS ordination (MLR response variable) was interpreted as a gradient of watershed disturbance; its site scores were significantly correlated with hydrologic, habitat, and land use/land cover variables (R2=0.56) such as the average number of annual storms producing runoff, stream flashiness, patterns of stream velocity and depth, and percentage of land near the stream channel composed of forest (forest buffer). Site scores along this disturbance gradient were also highly correlated with the New Jersey impairment index and many of its component metrics (for example, number of EPT taxa; R2=−0.71, p<0.001).
|
|
