Freshwater fish are ecologically important in stream ecosystems, and they provide people with significant food, recreation, and conservation value as biological indicator of freshwater streams. Historically, the streams and rivers of southern New England supported moderately diverse and abundant assemblages of native fishes. Climate change and land development are strongly affecting freshwater habitats by altering flow patterns and warming the water, which can reduce the abundance of fish species that need cold or cool (referred through this chapter as cold-coolwater fish species) and/or flowing water for some or all of their life cycle. Consequently, the relative abundance of these cold-coolwater and fluvial fish serves as an indicator of ecosystem condition. For example, brook trout (Salvelinus fontinalis) is a cold-water, fluvial species that is highly valued for conservation and recreation. The presence of brook trout and other cold-coolwater, fluvial species is indicative of good environmental health, and their distribution within the watershed provides a benchmark for the development of a robust indicator such as an index of biological integrity.
In the Narragansett Bay watershed, the Palmer River had the lowest percent relative abundance of cold-coolwater (11 percent) and fluvial (9 percent) fish species. The Upper Blackstone River had the highest at 69 percent (cold-coolwater) and 67 percent (fluvial). In the Little Narragansett Bay watershed, the percent relative abundances were lowest in the Wood River, which had 30 percent cold-coolwater and 18 percent fluvial, and highest in the Lower Pawcatuck River (60 percent, 57 percent). The Southwest Coastal Ponds watershed had 21 percent cold-coolwater fish and 15 percent fluvial fish. Brook trout habitat is largely concentrated in the Lower Blackstone River and the Pawtuxet River, along catchment areas covering over 25 percent of those two watersheds.
Freshwater fish are ecologically important in stream ecosystems and provide significant value to humans. Ecosystem services supported by fish include the regulation of food web dynamics and recycling of nutrients, linkages within aquatic systems and to terrestrial ecosystems, food production, recreational activities, and information services, including indicators of ecosystem stress and resilience (Holmlund and Hammer 1999). As indicators of stress, fish assemblages may provide evidence of impairment that differs from the signals provided by algae or benthic invertebrates (Carlisle et al. 2008) and is therefore important information for managing watersheds.
Historically, the streams and rivers of southern New England supported moderately diverse and abundant assemblages of native fish communities. Enser and Lundgren (2005) noted that in Rhode Island, the upstream reaches of streams and rivers are fairly steep and cold, with narrow, shallow streambeds characterized by well-defined riffles and pools, and mixed bedrock, boulder, and cobble substrate; characteristic fish include brook trout, longnose dace (Rhinichthys cataractae), and blacknose dace (Rhinichthys atratulus). Downstream reaches of streams and rivers tend to be wider and sluggish, with sand and silt substrate, and native fish including pumpkinseed (Lepomis gibbosus), chain pickerel (Esox niger), and yellow perch (Perca flavescens).
Stream and river habitats are affected by stressors including climate change and land use, both of which can alter flow and increase water temperatures (Kanno and Vokou 2010, Beauchene et al. 2014, Bain and Meixler 2000, Kashiwagi and Richards 2009, Bain 2011, Meixler 2011). Among the land use stressors affecting streams and rivers are nutrient enrichment, toxic chemicals, and road salts. Hydrologic alterations from dams, channel modifications, shoreline stabilization, and other in-stream activities can alter flow, causing shifts in fish species composition and abundance (Carlisle et al. 2010, Poff and Zimmerman 2010).
Freshwater fish assemblages are known to change in characteristic ways in response to stressors, and those changes can be used as indicators of stream and river health and integrity. Managers may use an index of biotic integrity (IBI) or a multi-metric index (MMI) to quantify aspects of the fish assemblage that are expected to respond to stressors (Whittier et al. 2007, Stoddard et al. 2008). In both approaches, metrics are calculated and then combined into a final score, which can then be related to a class of impairment. Rhode Island has not adopted an index approach but does have information on the sensitivity of fish communities to temperature and flow (RIDEM 2012, Rashleigh et al. 2013). Massachusetts and Connecticut have applied multi-metric indices to understand relationships between fish assemblage characteristics and anthropogenic factors, including temperature, flow, and other environmental, land use, and physical watershed characteristics (Armstrong et al. 2011, Bassar et al. 2016, Kanno et al. 2010, Kanno and Vokoum 2010). The states’ approaches differ, and there is no clear choice for adopting a multi-metric indicator for the entire Narragansett Bay watershed. For this report, we used two metrics that are used in all three states: percent relative abundance of cold-cool water fish and percent relative abundance of fluvial fish. Fluvial fish require flowing water for some or all of their life cycle. Coordination among the three states in the watershed could result in the development of a multi-metric index for the watershed.
Brook Trout Habitat
Eastern brook trout, in the salmon family, has received special attention in the region due to its economic, recreational, ecological and cultural value. Brook trout primarily eat insects, and they spawn in cold-water streams in the fall. Then adults migrate to deeper waters to overwinter, and eggs hatch in the spring. Sea-run brook trout have been documented in New England, but it is unclear whether any remain in the Narragansett Bay watershed. In pre-colonial times, brook trout were present in nearly every cold-water stream and river in the eastern United States. As land clearing took place for agriculture and timber harvesting, the amount of sediment entering streams increased, and wild brook trout began to disappear (Eastern Brook Trout Joint Venture, undated). The Narragansett Bay watershed’s main streams and rivers, such as the Blackstone River, were channelized and dammed, reducing brook trout habitat through altered hydrology and pollution from industrialization activities. Many of these threats to brook trout persist today.
Brook trout are known to be highly sensitive to multiple stressors, including increased temperatures, land use, and flow alteration. DeWeber and Wagner (2015) found a negative response of eastern brook trout to increasing temperature and agricultural land cover. Brook trout are found in perennial streams with temperatures between 34 and 72 °F (1 to 22°C) (Xu. et al. 2010). Bassar et al. (2016) demonstrated that declines in brook trout in western Massachusetts were related to changes in temperature and stream flow.
Brook trout are particularly important ecological indicators for the region, as the Narragansett Bay watershed is near the limit of their thermal tolerance, and climate change may have a significant effect on their distribution and abundance. In addition to brook trout, other fish species common in the Narragansett Bay watershed rely on cold-water stream and riparian habitats to meet one or more of their life history requirements (Massachusetts Division of Fisheries and Wildlife 2017). Thus, protecting these habitats from current and future anthropogenic and climate change stressors is critical. A decline in the brook trout population indicates negative changes in the habitat and overall ecosystem integrity.
In Massachusetts, the Division of Fisheries and Wildlife maintains a list of waters that are identified as Coldwater Fishery Resources. At the regional level, brook trout status has been assessed across its range from Maine to Georgia by the Eastern Brook Trout Joint Venture (EBTJV), a partnership among state and federal agencies, regional and local governments, businesses, conservation organizations, academia, scientific societies, and private citizens working toward protecting, restoring, and enhancing brook trout populations and their habitats across their native range. Hudy et al. (2008) and the EBTJV (undated) described an array of analyses comparing current distributions to the native range, as defined by MacCrimmon and Campbell (1969). In Connecticut and Rhode Island, researchers concluded that remaining populations are small and fragmented. We used data from the EBTJV to calculate total extent, percent, and total square miles of brook trout habitat within the watershed. The information presented in this chapter represents a coordinated effort across the states within the Narragansett Bay watershed and also represents a consensus of the best professional judgment accepted by scientists and managers.
The Narragansett Bay Estuary Program and the U.S. Environmental Protection Agency (Narragansett Lab) gathered, diagnosed, and reconciled fish sampling data from Massachusetts, Rhode Island, and Connecticut for this analysis. Other data sources were considered, but issues related to inconsistencies, geographical coverage, or availability made them unsuitable for this analysis. Localized studies have been conducted within the watershed, but those data were not readily available. National datasets such as the National Aquatic Resources Survey were also available, but only a limited number of the sampling sites were located within the Narragansett Bay watershed and the other two study areas, the watersheds for Little Narragansett Bay and Southwest Coastal Ponds.
The fish data from Rhode Island, Massachusetts, and Connecticut were assumed to be comparable in terms of methods, as all three states used single-pass backpack shocking in the upstream direction, sampled all habitats, covered similar lengths of stream (at least 100 meters), and had similar sampling schedules (typically late spring to fall).
Data were obtained from the Rhode Island Department of Environmental Management (RIDEM) Division of Fish and Wildlife (Libby 2004, Libby 2013), Massachusetts Division of Fish and Wildlife (MDFW 2014), and Connecticut Department of Energy and Environmental Protection (Hagstrom et al. 1996, Machowski and Hagstrom 2015). Data from Connecticut were included to analyze this indicator within Little Narragansett Bay watershed. For Rhode Island, samples ranged in date from April 29, 1992, to September 30, 2009, with a median date of August 26, 1999; most of the data were collected as part of a multi-year statewide survey of fish communities in rivers and streams conducted between 1992 and 2002, and survey work continued on a less frequent basis after 2002. For Massachusetts, sample dates ranged from September 21, 1994, to October 14, 2014, with a median date of August 8, 2006. For Connecticut, sample dates ranged from June 14, 1993, to August 6, 2014, with a median date of February 28, 2006. In all three states’ datasets, fish collected in the field had been identified to species and abundance recorded. We used data from 2002 to 2014 in our analysis; we limited the range of years used for analysis for consistency with other reports and to enhance comparability. Due to limited repeat sampling in the dataset, it was not possible to conduct a trend analysis to identify changes in the fish communities over time.
In accordance with standard practice, we adjusted the compiled dataset by removing stocked salmonids—rainbow trout (Oncorhynchus mykiss), Atlantic salmon (Salmo salar), and brown trout (Salmo trutta)—because their presence is due exclusively to stocking (Kanno and Volkun 2010, Bellucci et al. 2011, Libby 2013). Species predominantly associated with marine habitats such as mummichog (Fundulus heteroclitus) and anadromous fishes such as alewife (Alosa pseudoharengus), blueback herring (Alosa aestivalis), shad (Alosa sapidissima, Alosa mediocris), striped bass (Morone saxatilis), and sea lamprey (Petromyzon marinus) were also removed (Meixler 2011, Armstrong et al. 2011), although their abundances were very low in the samples so their inclusion would not likely have changed the results. Any hybrids or specimens not identified to species were also dropped.
We calculated two metrics of the fish community that are used by both Massachusetts and Connecticut: (a) the percent relative abundance of cold-coolwater fish and (b) the percent relative abundance of fluvial fish (Table 1). This analytical approach was also supported by RIDEM. Higher values of the metrics indicate better water quality and aquatic habitat. Cold and cool-water designations for species were obtained from Kanno et al. (2010, Appendix 1). Fluvial designations for species were derived from Armstrong et al. (2011), where fluvial dependents and specialists are fish species that require flowing water for some or all, respectively, of their life cycle. In our analyses, cold- and cold-water fish were combined into a single group termed cold-coolwater fish, and fluvial dependents and specialists were combined into a single group termed fluvial fish (Table 1). In some cases, Kanno et al. (2010) presented different information regarding fluvial designations of fish species. However, we followed the designations used by Armstrong et al. (2011) in Massachusetts because most of the Narragansett Bay watershed is located within Massachusetts.
The metrics were calculated as:
(Abundance of Cold-coolwater Fish/Total Fish Abundance) 100
(Abundance of Fluvial Fish/Total Fish Abundance) 100
Abundance was defined as the number of individuals. We note that Armstrong et al. (2011) reported relative abundance as counts per hour, so these methods are not directly comparable.
We used geographical information systems (GIS) tools in ArcGIS desktop (ESRI 2016) to associate the latitude and longitude reported by the states for each sample site were used with a HUC10 watershed. For the Southwest Coastal Ponds, metrics were summarized at the HUC12 subwatershed level, as the entire area is encompassed by one HUC10 watershed. Within each HUC10 watershed or HUC12 subwatershed, we calculated the average of percent relative abundance for each group: cold-coolwater and fluvial fish.
Table 1. Freshwater fish species of the Narragansett Bay watershed and their requirements for cold/cool water temperatures, flowing water (fluvial), or both.
1 Brook lamprey was not categorized by Armstrong et al. (2011) but was considered a fluvial specialist by Kanno (2010).
2 These species are considered tolerant to pollution (Halliwell et al. 1999), so they may be present in areas of low quality despite their specialized habitat requirements. Kanno et al. (2010) developed a metric that omitted blacknose dace. However, Armstrong et al. (2010) demonstrated strong relationships between flow and dace, so we included this species. Also, Kanno et al. (2010) included a metric for white sucker that is expected to be higher in more degraded areas.
3 The dataset included fourspine stickleback, which is primarily a saltwater or brackish-water fish but is often found in freshwater. The species is very rare in the Narragansett Bay watershed and did not strongly affect results.
4 These species are considered nonnative. Some assessment approaches exclude nonnative species, as their presence may not be desirable. The abundance of these species was extremely low and did not strongly affect results.
5 Fluvial dependents and specialists are fish species that require flowing water for some or all, respectively, of their life cycle. In analyses, the two categories are combined into a single group termed fluvial fish.
Brook Trout Habitat
We report on the amount of brook trout habitat because this habitat is indicative of sensitive areas for watershed protection to sustain cold-coolwater aquatic ecosystems. To this end, we obtained data on brook trout habitat patches from the Eastern Brook Trout Joint Venture (EBTJV 2015). In 2015, the EBTJV performed a habitat patch analysis coordinated across states at a regional scale, including the Narragansett Bay watershed, to identify catchment areas of rivers and streams with presence of brook trout in the previous two decades. We used these data as a proxy to identify potential presence of brook trout and thus available habitat.
The EBTJV methodology for the salmonid catchment assessment and habitat patch layers was described in Coombs and Nislow (2015). The salmonid catchment assessment and habitat patch layers were created using algorithms for spatial data analysis performed on an ArcGIS platform. The geospatial analysis used fish data from different Joint Venture partners from each state, the National Hydrography Dataset (NHD+ version 2, McKay et al. 2012), and the National Anthropogenic Barrier Dataset to define areas where brook trout was at least once observed.
The Estuary Program utilized the EBTJV’s brook trout patch layers within the Narragansett Bay watershed as well as within the Little Narragansett Bay and Southwest Coastal Ponds watersheds. In order to exclude introduced or stocked species from the analysis, we included brook trout only and removed others (brown, rainbow, and stocked trout) from the dataset. We used all habitat patches, both observed and predicted, that had been identified by the EBTJV.
For each HUC10 watershed or HUC12 subwatershed, we calculated brook trout habitat as the total area in square miles and percent of each watershed or subwatershed. It is important to note that habitat patch areas encompassed the entire extent of a catchment area of streams and rivers where fish sampling took place and brook trout was observed, and not only natural buffers. In the context of habitat protection, these catchment areas are important for watershed characterization, and while it is expected that sampling of brook trout from decades ago might not represent today’s conditions, the areas defined by EBTJV are of great importance to target restoration activities for protection of potential habitat and conservation of brook trout communities. Moreover, it is assumed that current floodplains, freshwater wetlands, and overall riparian buffers along the streams within the brook trout habitat patches are ecologically significant for protecting cold-water habitat.