| |
Presented at the NABS Annual meeting, Athens, Georgia, 2003
in Large River Ecology
Habitat evaluation of non-wadeable rivers
J.G.O. Wilhelm1, J.D. Allan1, K.J. Wessell2, R.W. Merritt2, and K.W. Cummins3. 1School of Natural Resources and Environment, University of Michigan, Ann Arbor, MI 48109-1115 USA, 2Department of Entomology, Michigan State University, East Lansing, MI 48824 USA, 3California Cooperative Fisheries Unit, Department of Fisheries, Humboldt State University, Arcata, CA 95521 USA
We developed a habitat index for Michigan's non-wadeable rivers, selecting metrics based on professional judgment, habitat variation across the landscape, and predictive relationships with disturbance. By modifying existing methods we assessed habitat at reach, riparian, and landscape scales on fourteen non-wadeable rivers in Michigan during 2000, 2001 and 2002. The 194 habitat measures were reduced to ten using correlation and principal components analyses. Remaining variables were tested against riparian and human disturbance gradients derived from aerial photograph measurements, area land use within a buffer and the upstream basin, and measures of hydrological modification and point source pollution. Multiple linear regression models were used to predict the disturbance gradients from instream habitat, and best measures were given the greatest weight in the final index. Habitat variables explained 80% and 76% of the variation in the riparian and human disturbance gradient scores, respectively, among 22 sites used in model development. When used to predict riparian and human disturbance for an independent data set of 23 non-wadeable sites, the model explained 76% and 77% of the variation between observed and expected values. We conclude that the final habitat index provides an adequate index of overall site quality and is sensitive to anthropogenic impacts.
|
|
