Baltimore Ecosystem Study Institute of Ecosystem Studies

2011 BES Annual Meeting Presentation and Poster Abstracts



 
What’s missing from Baltimore’s urban fish communities and why?: A preliminary comparative analysis of fish assemblage data
 
Kemp, Stanley
Co-Authors:

 
Abstract: Understanding the impact of urbanization on aquatic biological communities is a central part of evaluating and guiding ecosystem restoration efforts. Assessments of ecosystem health using fish assemblage data are frequently accomplished using metrics such as the Index of Biotic Integrity (IBI), which are developed regionally using correlation of fish assemblages and degree of stream ecosystem impact. Alternative approaches (e.g. Target Fish Community) construct an expected fish assemblage of a particular waterway using data from less impacted reference streams, and compare these with observed species assemblages. Results of a conceptually similar comparison on fish species assemblages from two urbanized Baltimore waterways (Gwynns Falls and Jones Falls) and area reference streams are presented here. A discriminant analysis was performed using Canonical Analysis of Principal coordinates (CAP) between fish assemblages of urbanized and reference sites. The analysis produced a significant discrimination between target and reference sites (P= .00002, Trace statistic permutation test), and had a classification accuracy rate of 89% (Leave-one-out Allocation test). Species strongly associated with reference sites showed some general life history similarities, including the fact that all but one were lithophilic spawners. Of the five species most associated with urbanized sites, only one was a lithophilic spawner. General pollution tolerances tended to match expectations though there was variation in each of these two groups, suggesting that site specific effects were important. Detailed analysis of fish communities combined with species specific life history information can provide insight into the progress and needs of ecosystem restoration programs.