Search

AbstractWe 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.

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.

Abstract: We used two methods and existing field survey data to link juvenile fish and their habitats. The first method used seine survey data collected monthly from July to October 1988–1996 at fixed stations in Narragansett Bay, Rhode Island. Thirteen fish species making up 1% or more of the catch were analyzed by principal components analysis for two time periods: July–August and September–October. The stations were then plotted by their principal component scores to identify station groupings and habitat types. The second method used environmental data collected in July and August 1996 at the established survey stations in a principal components analysis. The stations and 13 most abundant species were plotted by principal components scores resulting from the environmental data. For the environmental data, the first two principal components explained 59% of the variance. The first principal component described the amount of energy shaping the habitat and was positively correlated with salinity, dissolved oxygen, current flow, and slope, and negatively correlated with silt. The second component was positively correlated with depth and silt, and negatively correlated with dissolved oxygen. The environmental data grouped the stations according to their distance from the ocean and three habitat types emerged. The uppermost station was a silty barren having low salinities and dissolved oxygen. Three other stations grouped together as low energy, protected habitats with sandy substrates. Lower bay stations had higher salinities, higher dissolved oxygen, higher flow rates, greater slopes, and larger size substrates, mostly cobble and gravel. Results from the fish data grouped the stations similarly. Combining results from both datasets revealed the uppermost station had the highest catches, most species, and greatest number of winter flounder (Pseudopleuronectes americanus) juveniles. Plots of winter flounder catches with principal component scores from the environmental data indicated the winter flounder distribution in the bay has shrunk from baywide to mostly the upper estuary near their primary spawning grounds. Results illustrate the value of coupling historic fish survey data with environmental measurements for identifying previously undervalued habitats important to fish.