Use of a Fish Index to Assess Habitat Quality in Narragansett Bay, Rhode Island

Abstract

We developed an estuarine index of biotic integrity to assess habitat quality in Narragansett Bay, Rhode Island. Fish were collected at 18 fixed stations with a 61-m × 3.05-m beach seine once per month in July and August from 1988 to 1999. Stations were designated high or low quality depending on the total nitrogen concentration, number of low dissolved oxygen events, extent of human disturbance, abundance of macroalgae, and presence or absence of eelgrass Zostera marina. We used stepwise discriminant analysis on 13 candidate metrics based on fish community data to determine those most important in distinguishing between high and low habitat categories. We then used discriminant analysis and canonical discriminant analysis with the resulting significant metrics. The six metrics that were able to discriminate between sites were the number of estuarine spawner species, proportion of killifish, number of individuals, proportion of flounder, Shannon's diversity index, and proportion of benthic-associated species. Discriminant analysis with these metrics distinguished among sites with an overall error rate of 22%. Surprisingly, when an index was calculated from the metrics, sites designated as low quality based on environmental data had the highest scores whereas high-quality sites had low scores. Degraded sites in the upper estuary may have scored high due to the availability of shallow-water refuge from predation, abundant food, and warmer water with less current. Our results indicate that upper estuaries are important to juvenile fish and are worth protecting despite apparent human impacts. Further, an index of biotic integrity may not be useful in estuaries because of their characteristically harsh environmental conditions, which result in fish communities that are dominated by tolerant species that are habitat and feeding generalists.