Thursday, May 29, 2008 - 11:15 AM
544

Stream integrity in the chesapeake bay watershed: A classification of biological condition using benthic macroinvertebrates and fish

Kelly O. Maloney1, Donald E. Weller1, and Marc J. Russell2. (1) Smithsonian Environmental Research Center (SERC), PO Box 28, Edgewater, MD 21037, (2) Gulf Ecology Division, US Environmental Protection Agency, 1 Sabine Island Dr., Gulf Breeze, FL 32561

The Chesapeake Bay is the largest U.S. estuary and has been a focus of much research on anthropogenic stress, yet detailed landscape-scale classifications of stream conditions within the basin are limited.  We used Federal (US EPA) and Maryland indicies of biotic integrity derived from stream benthic macroinvertebrate (BIBI) and fish (FIBI) surveys to develop Classification and Regression Tree (CART) models for predicting stream biological condition from watershed characteristics for all 1st-3rd order streams of the Chesapeake basin.  Independent variables included catchment-scale measures of anthropogenic (population density, land cover), topographic (slope, elevation), and climatic (precipitation) characteristics.  One quarter of sites were left out of model development as an independent validation set.  CART models for the BIBI classified Poor validation sites best (78% correct), followed by Good (44%) and Fair sites (32%). FIBI CART models predicted Good sites best (65%), followed by Poor (58%) and Fair (34%).  For the entire Chesapeake basin, BIBI CART predicted 49% as Poor, 13% as Fair, and 38% as Good, while the FIBI CART predicted 37% as Poor, 17% as Fair, and 46% as Good.  Agreement between predictions from BIBI and FIBI were strong for Poor (68%) and Good (62%) sites but weak for Fair sites (8%).


Web Page: assessment, landscape, integrity

See more of Landscapes II
See more of Contributed Sessions

See more of The Nabs 56th Annual Meeting (25-28 May, 2008)