Natural and Anthropogenic Influences on the Mount Hope Bay Ecosystem 2006 Northeastern Naturalist 13(Special Issue 4):95–116 Trends in Fish Abundance in Mount Hope Bay: Is the Brayton Point Power Station Affecting Fish Stocks? Joseph T. DeAlteris1,*, Thomas L. Englert2, and John A.D. Burnett2 Abstract - Trends in abundance for winter flounder (Pseudopleuronectes americanus), windowpane (Scophthalmus aquosus), hogchoker (Trinectes maculatus), tautog (Tautoga onitis), and scup (Stenotomus chrysops) in upper and lower Mount Hope Bay were compared to trends in Narragansett Bay to assess the effect of natural and anthropogenic stressors, including Brayton Point Power Station, on Mount Hope Bay fishes from 1972 to 2001. Sources of data included the Rhode Island Division of Fish and Wildlife trawl survey for Narragansett Bay and lower Mount Hope Bay, the University of Rhode Island Graduate School of Oceanography trawl survey for Narragansett Bay, and the Marine Research, Inc. trawl and Brayton Point Station impingement surveys for upper Mount Hope Bay. Analysis of covariance and Tukey-Kramer multiple comparison tests were used to evaluate differences in the slopes of transformed abundance indices from 1972–2001 and for two subsets of years, 1972 to 1985 and 1986 to 2001, periods of lower and higher power plant cooling water withdrawals, respectively. Trends in abundance of these species in both upper and lower Mount Hope Bay are not substantively different from those in Narragansett Bay during any of the three time periods evaluated. This is evident through either a highlevel visual inspection of the slopes measured for each species, time period, and area or a more detailed inspection of the analysis of covariance results and Tukey-Kramer confidence intervals associated with each slope estimate. Natural and anthropogenic stressors unique to Mount Hope Bay, including Brayton Point Station, have not caused Mount Hope Bay fish stocks to change at rates different from those observed for the same stocks in Narragansett Bay. This supports the conclusion that large-scale factors such as overfishing, climate change, and increased predator abundance are more likely to be the cause of the observed declines in important species such as winter flounder in Mount Hope Bay and Narragansett Bay. Introduction Fishery resource abundance indices in Rhode Island estuarine and coastal waters have undergone dramatic changes in the last three decades primarily due to the effects of overfishing, global climate change, increases in predator abundance and habitat degradation (DeAlteris et al. 2000). Quantifying the relative importance of individual causal factors associated with changes in the abundance of individual fish stocks is very difficult when multiple stressors are simultaneously acting on the fish stock (Hilborn and Walters 1992). In recent years, questions have been raised regarding the cumulative effect of natural and anthropogenic stressors including Brayton Point Station (BPS), a 1University of Rhode Island, Building 50, East Farm, Kingston, RI 02881. 2Lawler, Matusky, and Skelly Engineers, LLP, One Blue Hill Plaza, Pearl River, NY 10965. *Corresponding author - jdealteris@uri.edu. 96 Northeastern Naturalist Vol. 13, Special Issue 4 fossil fuel power generating facility located in upper Mount Hope Bay, on the abundance of certain fish populations in Mount Hope Bay. Mount Hope Bay is a smaller sub-estuary of Narragansett Bay, a major estuary located in the northeastern United States (Fig. 1). Fish populations in Narragansett Bay have been intensively investigated for the last 40 years by scientists at Rhode Island Division of Fish and Wildlife (RIDFW), University of Rhode Island Graduate School of Oceanography (URIGSO), US Environmental Protection Agency, National Marine Fisheries Service, and others. The results of these investigations have documented the ever-changing nature of Narragansett Bay fish stocks (Jefferies 2000). In general, the abundance trends of winter flounder (Pseudopleuronectes americanus Walbaum, 1792), windowpane (Scophthalmus aquosus Mitchill, 1815), hogchoker (Trinectes maculatus Bloch & Schneider, 1801), tautog (Tautoga onitis Linnaeus, 1758) and scup (Stenotomus chrysops Linnaeus, 1766) have all oscillated over the last 40 years (DeAlteris et al. 2000). Beginning in the 1980s, winter flounder, windowpane, hogchoker, and tautog Figure 1. Map of the study area showing the trawl and impingement sampling locations used to calculate the fish abundance indices for Narragansett Bay exclusive of Mount Hope Bay, for lower Mount Hope Bay, and for upper Mount Hope Bay. 2006 J.T. DeAlteris, T.L. Englert, and J.A.D. Burnett 97 all declined dramatically, while scup increased markedly (Jefferies 2000). Large-scale changes in fish abundance patterns, which have been attributed to the combined effects of changing climate and selective over-harvesting by the fisheries, have been observed in Narragansett Bay, southern New England and Georges Bank (Sullivan et al. 2001). Gibson (2002) analyzed trawl and impingement data collected in upper Mount Hope Bay from 1972 to 1985 and concluded that winter flounder abundance in all of Mount Hope Bay was declining at a greater rate than in Narragansett Bay and adjacent coastal waters. DeAlteris (2003) conducted a similar analysis for five fish species (i.e., winter flounder, windowpane, hogchoker, tautog, and scup) over the 1972 to 2001 period using data representative of upper and lower Mount Hope Bay and Narragansett Bay with the exclusion of data from coastal waters and found that all species fared as well in lower Mount Hope Bay as in Narragansett Bay over the entire observation period. Windowpane, hogchoker, and tautog fared as well or better in upper Mount Hope Bay than in Narragansett Bay. For scup and winter flounder, the results of the analyses for upper Mount Hope Bay were equivocal: some analyses indicated no difference in fish abundance trends between upper Mount Hope Bay and Narragansett Bay, while other analyses indicated both positive and negative trends in upper Mount Hope Bay as compared to Narragansett Bay. In this paper, we apply an alternative analytical method to that previously used by Gibson (2002) and DeAlteris (2003) to investigate and compare trends in the abundance of five fish species in Mount Hope Bay and areas of Narragansett Bay, with the goal of determining whether fish abundance trends in Mount Hope Bay have been differentially influenced by operation of the Brayton Point Power Station and other non-specific anthropogenic and natural stressors. Methods Brayton Point Station has four operating units that combust low-sulfur coal (units 1, 2, and 3) and oil and natural gas (unit 4) to produce electricity. Cooling water withdrawn from Mount Hope Bay for all units combined has fluctuated from year to year but averaged 0.74 BGD from 1972 to 1985 and 1.00 BGD from 1986 to 2001 (Fig. 2). Because of the difference in average flows between the pre- and post-1985 periods, the analysis contained herein was performed for the 1972 to 1985, 1986 to 2001, and 1972 to 2001 periods. Indices of fish abundance There are five primary sampling programs conducted within Narragansett Bay that can be used to calculate long-term indices of fish abundance (Table 1). Data collected in these programs were used to generate indices of abundance for the five fish species that Lawton and Corriea (1996) identified as being the only species captured in the Marine Research, Inc. (MRI) standard trawl survey at levels sufficient for documenting abundance trends: winter flounder, windowpane, hogchoker, scup, and tautog. 98 Northeastern Naturalist Vol. 13, Special Issue 4 The MRI trawl abundance index (1972–2001) was calculated by MRI (M. Scherer, MRI, pers. comm.) using all fixed-station data collected within a given year except tows conducted at the Spar Island Station located in Rhode Island waters of lower Mount Hope Bay and tows of 3-minute duration at the discharge station (i.e., discharge station tows prior to 1979). The sampling program is described in detail in USGen NE (2004). The Spar Island Station was excluded from the MRI trawl index because the intent was to represent fish abundance trends in upper Mount Hope Bay separately from lower Mount Hope Bay. The annual catch-per-unit-effort (CPUE) values for the MRI trawl are the delta mean (Pennington 1983, 1986, 1996; Smith 1988) catch per tow. The MRI fixed station CPUE indexes fish abundance in approximately the upper one-third of Mount Hope Bay (Fig. 1). The BPS impingement index (1973–2001) was calculated from individual impingement collection data recorded at Brayton Point Station’s units 1, 2, and 3 intake screens, as provided by MRI (M. Scherer, pers. comm). The sampling program is described in detail in USGen NE (2004). The index was calculated by dividing the total number of fish of a given species captured during each year in the impingement sampling program by the total intake volume sampled during that year. Annual values were converted to the number per million m3 of intake water. The BPS impingement index documents fish abundance in the vicinity of the Station’s units 1, 2, and 3 intakes in upper Mount Hope Bay (Fig. 1). The RIDFW fixed and random-trawl program’s individual tow data, provided by RIDFW (Tim Lynch, RIDWF, pers. comm.), were combined to calculate two different indices of abundance (1979–2001) for each species: Figure 2. Combined Brayton Point Station cooling water (BGD) withdrawal from Mount Hope Bay for units 1, 2, 3, and 4 during the period 1972 to 2001. 2006 J.T. DeAlteris, T.L. Englert, and J.A.D. Burnett 99 one for lower Mount Hope Bay and another for Narragansett Bay exclusive of Mount Hope Bay (Fig. 1). The RIDFW trawl sampling program is described in detail in Lynch (2000). Lower Mount Hope Bay was defined to the north by the Rhode Island/Massachusetts state border and to the south by the Mount Hope and Sakonnet Bridges. The Narragansett Bay exclusive of Mount Hope Bay index was calculated with data from all Narragansett Bay tows, except those defined as being in Lower Mount Hope Bay. To generate the annual index for each area, all fish captured in the random and fixed station tows made in that area were summed and divided by the total number of tows. In generating the indices, only April to May and September to October fixed station tows were used because these time periods were sampled consistently throughout the duration of the random trawl program. The URIGSO trawl index (1972–2001) was provided by J. Collie URIGSO, pers. comm.). The URIGSO trawl sampling program is described in detail in Jefferies and Johnson (1974). The index is the annual geometric Table 1. Specification for the various sampling programs used to calculate fish abundance indices for upper Mount Hope Bay (MHB), lower Mount Hope Bay, and Narragansett Bay (NB). Sampling program RIDFW RIDFW trawl trawl Gear and program MRI BPS (fixed (random URIGSO information trawl impingement stations) stations) trawl Head rope (m) 7.62 NA 11.89 11.89 10.06 Foot rope (m) 10.97 NA 16.46 16.46 10.06 Rise (m) 1.23–1.52 NA 2.44–3.05 2.44–3.05 1.22 Body mesh (cm, stretch) 12.07 NA 5.08–11.43 5.08–11.43 7.62 Cod end mesh (cm, stretch) 3.81 NA 3.81 3.81 5.08 Liner (cm) None NA 1.02 1.02 None Doors (m) 0.43 x 0.76 NA 0.61 x 1.22 0.61 x 1.22 0.76 x 1.52 Tow speed (m/sec) 1.28 NA 1.28 1.28 1.28 Tow duration (minutes) 15 NA 20 20 30 Program start 1972 19721 1990 1979 1967 Sampling frequency Monthly > 3x / week Monthly Spring, fall Weekly Current number of stations 62 1 13 26 24 Abundance index5 MRI trawl—upper MHB 67 (36, 96) BPS impingement—upper MHB 156 RIDFW trawl—lower MHB2 6 (4,8) 6 (2,9) RIDFW trawl—NB3 31 (19,40) 50 (40,77) URIGSO trawl—NB 104 NA = Not applicable. 1Data available starting in 1973. 2Number of fixed stations sampled has varied from four to seven, excluding the Spar Island Station. 3Both RIDFW fixed (April–May and September–October only) and random stations were used to calculate abundance indices. 4Only data from the Fox Island Station were used in this analysis, as the Whale Rock Station is located in Rhode Island Sound, just outside of Narragansett. 5Approximate number of samples used to calculate annual abundance index values (min, max provided where possible). 100 Northeastern Naturalist Vol. 13, Special Issue 4 mean catch per tow at the Fox Island Station, calculated from all tows made within a given year for all species except scup and tautog. For these two species, the annual geometric mean catch per tow is based on May through October and May through November tows, respectively. A hogchoker index was not available. The URIGSO Fox Island index documents fish abundance in the West Passage of Narragansett Bay (Fig. 1). Statistical analysis of fish abundance trends Analysis of covariance (ANCOVA) and Tukey-Kramer multiple comparison tests were used to test for difference among trends in abundance for each species in each of the five indices of abundance during each of the three time periods studied (1972 to 1985, 1986 to 2001, and 1972 to 2001). Prior to running the statistical analyses on the abundance indices, the following transformation was applied to each index so as to meet the assumption of linearity in the ANCOVA model: CPUEi,j,transformed = ln(CPUEi,j + CPUEj,mean1972–2001 x m), (1) where CPUE = catch per unit effort, i = year, j = index, and m.= 0.01. This particular transformation was used because the more typical transformations of ln(CPUE) and ln(CPUE + 1) resulted in, respectively, the loss of information when CPUE was equal to zero (i.e., the natural log of zero is undefined) and unreasonable minimization of the variance among annual abundance values where abundance values were small relative to a value of 1. Alternative values of m in Equation 1 were found to either unreasonably minimize variance (e.g., 0.1) or create unreasonable outliers from the zero catch values (e.g., 0.001). The transformation performed in Equation 1 linearizes the relationship between fish abundance and year, Table 2. ANCOVA equality of slopes test results (􀁟 = 0.05) for five indices of fish abundance (MRI trawl and BPS impingement for upper Mount Hope Bay, RIDFW trawl for lower Mount Hope Bay, and RIDFW trawl and URIGSO trawl for Narragansett Bay) for each of five species during each of three time periods. Species Years Type III SS Mean square F value p value* Winter flounder 1972–1985 5.07 1.27 2.74 0.040 1986–2001 0.54 0.14 0.49 0.742 1972–2001 16.24 4.06 8.57 < 0.001 Windowpane 1972–1985 2.21 0.55 0.92 0.463 1986–2001 8.20 2.05 1.79 0.141 1972–2001 6.50 1.62 1.71 0.151 Hogchoker 1972–1985 12.40 4.13 5.71 0.003 1986–2001 5.25 1.75 0.88 0.455 1972–2001 26.52 8.84 5.10 0.003 Scup 1972–1985 0.58 0.14 0.10 0.980 1986–2001 13.18 3.30 1.65 0.171 1972–2001 38.52 9.63 5.18 0.001 Tautog 1972–1985 5.17 1.29 2.18 0.087 1986–2001 2.39 0.60 0.52 0.719 1972–2001 30.57 7.64 7.51 < 0.001 *p values less than 0.05 indicate that at least one slope was significantly different from one other for the specified species and time period. 2006 J.T. DeAlteris, T.L. Englert, and J.A.D. Burnett 101 while maintaining variance among the low CPUE values and not creating unreasonable outliers when CPUE values were zero. Analysis of covariance (􀁟 = 0.05) was used to evaluate differences among the slopes of the various transformed indices of abundance for the five species of fish during each cooling water regime, and for the entire time series (Sokal and Rohlf 1995). Each ANCOVA was implemented in SAS software using Proc GLM (SAS 2000). Significance of the type III sum of squares of the interaction term (i.e., year × abundance index) indicated whether any one index demonstrated a slope significantly different from any other. Given the large number of ANCOVA results concluding significant differences among at least one pair of slopes (6 of 15 tests), differences among individual slopes were evaluated using Gabriel’s approximate method for the Tukey-Kramer multiple comparison test (Sokal and Rolf 1995; 􀁟 = 0.05) implemented in Excel. Gabriel’s approximate method for the Tukey-Kramer multiple comparison test produced 95% confidence intervals around each calculated slope. Statistical difference among individual slopes was determined based on non-overlap of the 95% confidence intervals. Results Winter flounder For the early time period, 1972 to 1985, there is one significant difference among the slopes of the five indices (p = 0.040): the Brayton Point Station impingement index slope for upper Mount Hope Bay has a significantly steeper negative slope than the RIDFW trawl index slope for lower Mount Hope Bay (Table 2, Fig. 3). For the later time period, 1986 to 2001, there is no significant difference in the slopes of the five indices (p = 0.742). For the entire time series, 1972 to 2001, there are significant differences among the slopes (p < 0.001). The slope of the MRI trawl index is significantly steeper than the slopes of all the other indices except for the RIDFW trawl index slope for Narragansett Bay. Windowpane Windowpane is the only species examined for which all slopes are negative and not significantly different from one another within each of the three time periods examined (1972 to 1985, p = 0.463; 1986 to 2001, p = 0.141; 1972–2001, p = 0.151) (Table 2, Fig. 4). Hogchoker For the early time period, 1972 to 1985, there are significant differences in the slopes among the four indices (p = 0.003), with the BPS impingement index slope being significantly less steep than both the RIDFW trawl lower Mount Hope Bay and Narragansett Bay index slopes (Table 2, Fig. 5). The MRI trawl index slope for this period was also found to be significantly less steep than the RIDFW lower Mount Hope Bay index slope. There is no significant difference among slopes during the later time period, 1986 to 102 Northeastern Naturalist Vol. 13, Special Issue 4 2001 (p = 0.455). For the entire time series, 1972 to 2001, there are significant differences in slopes among the four indices (p < 0.003). The estimated Figure 3.Top. Winter flounder abundance indices for the MRI trawl survey, the BPS impingement survey, the RIDFW trawl survey in lower Mount Hope Bay, and the URIGSO and RIDFW trawl surveys in Narragansett Bay (*post-analysis constants have been applied to the transformed indices from equation 1 to achieve separation of the indices for visual presentation). Bottom. Slopes with Tukey-Kramer 95% confidence intervals for the periods 1972 to 1985, 1986 to 2001, and 1972 to 2001. 2006 J.T. DeAlteris, T.L. Englert, and J.A.D. Burnett 103 slope for the MRI trawl index and BPS impingement index are significantly less steep than the slope of the RIDFW trawl Narragansett Bay index. Figure 4. Top. Windowpane flounder abundance indices for the MRI trawl survey, the BPS impingement survey, the RIDFW trawl survey in lower Mount Hope Bay, and the URIGSO and RIDFW trawl surveys in Narragansett Bay (*post-analysis constants have been applied to the transformed indices from equation 1 to achieve separation of the indices for visual presentation). Bottom. Slopes with Tukey-Kramer 95% confidence intervals for the periods 1972 to 1985, 1986 to 2001, and 1972 to 2001. 104 Northeastern Naturalist Vol. 13, Special Issue 4 Figure 5. Top. Hogchoker abundance indices for the MRI trawl survey, the BPS impingement survey, the RIDFW trawl survey in lower Mount Hope Bay, and RIDFW trawl surveys in Narragansett Bay (*post-analysis constants have been applied to the transformed indices from equation 1 to achieve separation of the indices for visual presentation). Bottom. Slopes with Tukey-Kramer 95% confidence intervals for the periods 1972 to 1985, 1986 to 2001, and 1972 to 2001. 2006 J.T. DeAlteris, T.L. Englert, and J.A.D. Burnett 105 Figure 6. Top. Scup abundance indices for the MRI trawl survey, the BPS impingement survey, the RIDFW trawl survey in lower Mount Hope Bay, and the URIGSO and RIDFW trawl surveys in Narragansett Bay (*post-analysis constants have been applied to the transformed indices from equation 1 to achieve separation of the indices for visual presentation). Bottom. Slopes with Tukey-Kramer 95% confidence intervals for the periods 1972 to 1985, 1986 to 2001, and 1972 to 2001. 106 Northeastern Naturalist Vol. 13, Special Issue 4 Scup There is no significant difference among slopes within either of the 1972 to 1985 (p = 0.980) or 1986 to 2001 (p = 0.171) time periods for scup (Table 2, Fig. 6). For the long time series, 1972 to 2001, there are significant differences among the slopes of the five indices (p = 0.001). While the estimated slopes for the MRI trawl and the BPS impingement indices are not significantly different for the long time period, both were found to have significantly steeper slopes than the RIDFW lower Mount Hope Bay and RIDFW Narragansett Bay indices, while only the MRI trawl index slope is significantly steeper than the URIGSO trawl index slope. Tautog Similar to the findings for scup, tautog were found to have no significant differences among slopes within either the 1972 to 1985 (p = 0.087) or 1986 to 2001 (p = 0.719) time periods (Table 2, Fig. 7). However, while the ANCOVA result reported above shows no significant difference among the slopes for the 1972 to 1985 period, the Tukey-Kramer multiple comparison test results suggest that there was in fact a significant difference between the slopes for the BPS impingement and the URIGSO trawl indices. This inconsistency in the results from ANCOVA and the multiple comparison tests, which is the only inconsistency of this sort encountered in the analysis, can be attributed to subtle differences in the two analytical methods and the truly borderline statistical significance in this particular case (i.e., p = 0.087, while 􀁟 = 0.05). If a significant difference was concluded for this case, it would indicate that, based on these indices, tautog faired better in upper Mount Hope Bay than Narragansett Bay during this period. However, we conclude that results are equivocal as to whether a difference in these two trends in fact exists. For the long time series, 1972 to 2001, there are significant differences among the slopes of the five indices (p < 0.001). The estimated slope of the BPS impingement index is significantly less steep than the URIGSO trawl index slope, while the MRI trawl index slope is significantly steeper than all other index slopes except the URIGSO trawl slope. Discussion and Conclusions The weight of evidence from this analysis suggests that trends in abundance of fish in both upper and lower Mount Hope Bay are not different from those in Narragansett Bay for both the long time period of 1972 to 2001, and the two shorter time periods of 1972 to 1985 and 1986 to 2001, periods of lower and higher power plant cooling water withdrawals, respectively. This is evident through either a high-level visual inspection of the slopes measured for each species, time period, and area or a more detailed inspection of the ANCOVA and Tukey-Kramer confidence intervals associated with each slope estimate. Specifically, with regard to upper Mount Hope Bay, results for the two shorter time periods show that the one significant difference between upper Mount Hope Bay and Narragansett Bay (i.e., 1972–1985 BPS impingement and RIDFW trawl index slopes for hogchoker) indicates that 2006 J.T. DeAlteris, T.L. Englert, and J.A.D. Burnett 107 Figure 7. Top. Tautog abundance indices for the MRI trawl survey, the BPS impingement survey, the RIDFW trawl survey in lower Mount Hope Bay, and the URIGSO and RIDFW trawl surveys in Narragansett Bay (*post-analysis constants have been applied to the transformed indices from equation 1 to achieve separation of the indices for visual presentation). Bottom. Slopes with Tukey-Kramer 95% confidence intervals for the periods 1972 to 1985, 1986 to 2001, and 1972 to 2001. 108 Northeastern Naturalist Vol. 13, Special Issue 4 the decline was steeper in Narragansett Bay. Analyses for the long time period show that the five species studied followed similar trends in upper Mount Hope Bay to those measured for Narragansett Bay when compared to at least one of the two indices for Narragansett Bay. The exceptions were for hogchoker, where declines were found to be significantly steeper in Narragansett Bay, and for scup, where only the MRI trawl showed significantly steeper declines in upper Mount Hope Bay than in either Narragansett Bay index. Lack of significant difference among slopes for the two shorter time periods may be attributable in part to the fact that sample sizes were relatively small. This is particularly true for comparisons involving the RIDFW trawl indices which started in 1979. Nonetheless, visual inspection of the calculated slopes, irrespective of the confidence interval, shows no clear and consistent pattern of steeper declines in Mount Hope Bay. Winter flounder, a highly valued commercially and recreationally fished species, was once one of the dominant fish species throughout New England, including Mount Hope and Narragansett Bay. The decline of regional and local winter flounder populations has been attributed to the combined effects of overfishing, global climate change, habitat degradation, and power plant impacts, among other sources (Collie and Delong 2002, DeAlteris et al. 2000, Gibson 2002). The presence of Brayton Point Station in upper Mount Hope Bay and the declines evident in the winter flounder abundance indices from the MRI trawl survey and the BPS impingement screens that represent upper Mount Hope Bay have caused concern for environmental regulators and fishery resource managers. Results from this analysis show that slopes for winter flounder abundance trends as measured in the two indices for upper Mount Hope Bay bound the slopes measured for both lower Mount Hope Bay and Narragansett Bay. That is, the MRI trawl index displayed the steepest slope, while the Brayton Point Station impingement index displayed the least-steep slope of all five long time period slopes. Additionally, each of these upper Mount Hope Bay index slopes was found not to be significantly different from at least one of the Narragansett Bay indices. Overall, there is no consistent evidence of a steeper rate of decline in winter flounder abundance in any part of Mount Hope Bay relative to Narragansett Bay. It is unclear why the MRI trawl and BPS impingement indices should display such differing trends in abundance for winter flounder, particularly over the long time period. DeAlteris (2003) investigated this inconsistency and found at least two potential factors that could have contributed to a decrease in catch efficiency of the MRI trawl: significant dredging of the intake channel during late 1985 (the intake channel typically contributed a large proportion of the MRI trawl catch) and replacement of the MRI trawl survey net in early 1986. As for Brayton Point Station impingement sampling, the deepening of the intake channel during late 1985 as well as an increase in the sampling frequency from three times per week to every day since 1997 are potential sources of change in the catch potential of this survey. Considering the factors that could have influenced the catch potential in the two surveys and the trends displayed by the RIDFW trawl in lower 2006 J.T. DeAlteris, T.L. Englert, and J.A.D. Burnett 109 Mount Hope Bay, DeAlteris (2003) concluded that the more likely explanation for the trends discrepancy in the upper Mount Hope Bay indices was a change in the catch potential of the MRI trawl. The results of the analyses for scup are equivocal with respect to upper Mount Hope Bay. The shorter-term indices indicate that the abundance trends are similar to those in Narragansett Bay, while the long-term MRI trawl index and BPS impingement index slopes suggest that scup abundance has declined in upper Mount Hope Bay, and the lower Mount Hope Bay index and both Narragansett Bay indices suggest it has increased in these areas (although only very slightly according to the URIGSO trawl). In fact, the BPS impingement index slope is not significantly different from the URIGSO index slope. The analyses of scup in lower Mount Hope Bay indicate that the abundance trend for this area is virtually identical to that measured by the RIDFW trawl index for Narragansett Bay and not significantly different from that measured by the URIGSO trawl in Narragansett Bay. As a result, there is only limited evidence of a steeper rate of decline in scup abundance in upper Mount Hope Bay and no evidence of a faster decline in lower Mount Hope Bay relative to Narragansett Bay. A possible explanation for the difference in abundance trends for scup in the MRI trawl survey relative to the other trawl and impingement indices is that scup, a pelagic species, are not sampled with this small, low-rise bottom trawl in sufficient numbers to accurately document abundance trends. The largest MRI trawl annual CPUE value for scup during the 1972 to 2001 period is 36 fish. This value, which is over three times larger than the next largest CPUE value of 11 fish, is considerably lower than catches in the other trawl gears. The time series high annual CPUE values in the RIDFW trawl in lower Mount Hope Bay and Narragansett Bay and the URIGSO trawl in Narragansett Bay were 428, 713, and 331, respectively. Despite the differences in analytical approaches applied in this study and that of DeAlteris (2003), results from the two studies closely agree. However, there is one important exception for winter flounder. While DeAlteris (2003) found evidence of a significantly steeper rate of decline in the MRI trawl index than in that of the RIDFW trawl index for Narragansett Bay, the current analysis found the slopes of the two indices to be relatively similar and not significantly different. Given that the p value for this comparison in DeAlteris (2003) was very close to 0.05, this difference in conclusion can be attributed to differences in analytical techniques applied in the two analyses. While the results and conclusions drawn from the current analysis are similar to those of DeAlteris (2003), both are different from those reported by Gibson (2002). Gibson suggested that declines in abundance of winter flounder in Mount Hope Bay were unequivocally steeper than those in Narragansett Bay during 1972 to 1985. The differences between the conclusions of the current and previous DeAlteris analyses and those of Gibson (2002) are most likely attributable to Gibson’s use of upper Mount Hope Bay indices characterized as Mount Hope Bay indices, and the inclusion of coastal data in the “control” indices incorporated in his analysis. We believe that there is no reason to assume that trends in fish abundance in Mount Hope Bay would be the same as those at coastal stations, given differences in 110 Northeastern Naturalist Vol. 13, Special Issue 4 habitat type and the high potential for inclusion of fish from other estuaries and the ocean in the catches at these stations. For these reasons, these coastal stations should not be included in the control indices. In summary, the weight of the evidence from this analysis suggests that trends in fish abundance in upper and lower Mount Hope Bay are not substantively different from those in Narragansett Bay over the long time period of 1972 to 2001 and the two shorter time periods of 1972 to 1985 and 1986 to 2001. Natural and anthropogenic stressors unique to Mount Hope Bay, including Brayton Point Station, have not caused Mount Hope Bay fish stocks to decline at rates different from those observed for the same stocks in Narragansett Bay. This supports the conclusion that more large-scale factors such as overfishing, climate change, and increased predator abundance are the cause of the observed declines in important species such as winter flounder in Mount Hope Bay. Acknowledgments The authors acknowledge the owners and operators of the Brayton Point Power Station for their support of this research. The authors thank John Young, guest editor of our manuscript, and the anonymous reviewers of the original manuscript for their thoughtful editorial comments and suggestions of the alternative statistical analyses presented in this revised manuscript. Literature Cited Collie, J.S., and A.K. Delong. 2002. Examining the decline of Narragansett Bay winter flounder. Final Report to Rhode Island Department of Environmental Management, Division of Fish and Wildlife. July 23, 2002. DeAlteris, J. 2003. An assessment of trends in fish abundance in Mount Hope Bay relative to Narragansett Bay. A report from DeAlteris Associates, Inc. to Brayton Point Station, MA., dated 1 May 2003. DeAlteris, J., M. Gibson, and L. Skrobe. 2000. Fisheries of Rhode Island. White Paper, Narragansett Bay Summit 2000. 48 pp. Gibson, M. 2002. Winter flounder abundance near Brayton Point Station, Mount Hope Bay revisited: Separating local from regional impact using long-term abundance data. RI Division of Fish and Wildlife, Research Reference Document 02/1. 28 pp. Hilborn R., and C. Walters. 1992. Quantitative Fish Stock Assessment. Chapman and Hall. NY. 570 pp. Jefferies, P. 2000. Rhode Island’s ever-changing Narragansett Bay. Maritimes 42: 4–6. Jefferies, H.P., and W.C. Johnson. 1974. Seasonal distributions of bottom fishes in the Narragansett Bay area: Seven-year variations in the abundance of winter flounder (Pseudopleuronectes americanus). Journal of the Fisheries Research Board of Canada 31(6):1057–1066. Lawton, R.P., and S. Corriea. 1996. Division position regarding RIDFW report on fish declines in Mount Hope Bay in relation to the operation of Brayton Point Power Station. Memorandum to W.L. Bridges, Massachusetts Division of Marine Fisheries, Boston MA. November 19, 1996. Lynch, T.R. 2000. Assessment of recreationally important finfish stock in Rhode Island waters. Coastal fishery resource assessment trawl survey. Rhode Island Division of Fish and Wildlife. 2006 J.T. DeAlteris, T.L. Englert, and J.A.D. Burnett 111 Pennington, M. 1983. Efficient estimators of abundance for fish and plankton surveys. Biometrics 39:281–286. Pennington, M. 1986. Some statistical techniques for estimating abundance indices from trawl surveys. Fishery Bulletin 84(3):519–521. Pennington, M. 1996. Estimating the mean and variance from highly skewed marina data. Fishery Bulletin 94(3):498–505. SAS 2000. Statistical Analysis System for Windows. Version 8.2. SAS Institute Inc. Cary, NC. Smith, S.J. 1988. Evaluating the efficiency of the 􀂨-distribution mean estimator. Biometrics 44:485–493. Sokal, R.R., and J.F. Rohlf. 1995. Biometry. Third Edition. W.H. Freeman and Company. New York, NY. Sullivan, B.K., D. Van Kueren, and M. Clancy. 2001. Timing and size of blooms of the ctenophore Mnemiopsis leidyi in relation to temperature in Narragansett Bay, RI. Hydrobiologia 451:113–120. USGen New England, Inc. (USGen NE). 2004. Brayton Point Station hydrological and biological monitoring program. 2003 Annual Report. USGen New England, Inc., Somerset, MA. September 1, 2004. 112 Northeastern Naturalist Vol. 13, Special Issue 4 Appendix 1. Winter flounder abundance indices for Narragansett Bay and Mount Hope Bay, 1972–2001. Upper MHB Lower MHB Narragansett Bay MRI BPS RIDFW RIDFW URIGSO Year trawl1 impingement2 trawl1 trawl1 trawl1 1972 45.3 75 1973 58.7 16.3 81 1974 40.7 22.0 65 1975 18.7 2.9 44 1976 10.9 4.9 36 1977 15.7 11.4 50 1978 55.8 13.7 88 1979 86.9 4.8 12.0 154.6 284 1980 16.1 7.2 61.4 64.7 202 1981 19.2 3.7 88.5 77.7 120 1982 54.8 1.8 28.8 37.0 107 1982 46.2 1.4 17.0 55.3 174 1984 19.9 3.8 50.1 40.2 63 1985 16.5 6.6 193.2 31.4 28 1986 6.6 8.5 99.6 43.5 24 1987 2.4 5.9 19.5 56.5 66 1988 0.9 1.6 26.5 27.3 50 1989 0.4 4.8 8.6 15.0 30 1990 1.2 1.3 24.5 16.9 17 1991 0.7 2.2 27.2 27.8 15 1992 0.5 1.1 15.1 11.4 8 1993 1.0 4.1 11.6 7.6 10 1994 0.5 2.5 5.2 9.0 5 1995 0.9 3.9 19.7 16.4 37 1996 0.5 1.5 14.6 11.7 17 1997 0.7 1.0 12.1 8.5 18 1998 0.7 0.4 6.8 2.9 13 1999 0.3 1.3 5.5 3.7 7 2000 0.1 0.4 3.6 6.7 8 2001 0.2 0.6 4.3 4.3 6 1average catch per trawl tow 2number per million cubic meters of intake water 2006 J.T. DeAlteris, T.L. Englert, and J.A.D. Burnett 113 Appendix 2. Windowpane abundance indices for Narragansett Bay and Mount Hope Bay, 1972–2001. Upper MHB Lower MHB Narragansett Bay MRI BPS RIDFW RIDFW URIGSO Year trawl1 impingement2 trawl1 trawl1 trawl1 1972 8.0 17.0 1973 7.2 0.88 16.0 1974 8.0 0.22 13.0 1975 4.5 0.13 7.5 1976 1.7 0.24 6.4 1977 2.3 1.20 10.0 1978 11.0 1.44 18.0 1979 9.5 0.85 26.0 36.9 24.0 1980 5.1 1.32 26.6 27.6 16.0 1981 2.8 0.84 31.5 23.4 10.0 1982 5.8 0.47 13.8 11.9 5.9 1982 3.1 0.11 6.3 15.5 9.2 1984 1.7 0.02 6.4 4.8 2.5 1985 2.0 0.20 13.6 5.1 1.7 1986 0.7 0.21 15.4 9.1 2.1 1987 0.4 0.10 4.0 21.4 9.1 1988 0.5 0.04 8.3 6.7 4.9 1989 0.1 0.03 0.6 2.6 1.4 1990 0.3 0.01 2.8 4.2 0.8 1991 0.1 0.01 1.5 3.3 0.5 1992 0.0 0.06 12.8 0.7 0.4 1993 0.2 0.18 2.4 0.8 0.7 1994 0.3 0.13 0.4 1.5 0.5 1995 0.3 0.09 0.3 2.4 2.4 1996 1.1 0.11 43.6 6.3 3.5 1997 0.8 0.13 1.8 1.5 2.5 1998 0.4 0.04 1.2 1.0 1.7 1999 0.3 0.06 0.0 2.0 0.4 2000 0.2 0.03 0.2 0.6 0.6 2001 0.0 0.00 0.0 0.5 0.3 1average catch per trawl tow 2number per million cubic meters of intake water 114 Northeastern Naturalist Vol. 13, Special Issue 4 Appendix 3. Hogchoker abundance indices for Narragansett Bay and Mount Hope Bay, 1972–2001. Upper MHB Lower MHB Narragansett Bay MRI BPS RIDFW RIDFW URIGSO Year trawl1 impingement2 trawl1 trawl1 trawl3 1972 0.9 1973 0.6 1.1 1974 3.4 1.2 1975 1.8 1.7 1976 1.2 5.3 1977 0.2 22.4 1978 1.1 3.0 1979 1.9 1.8 14.3 3.8 1980 0.9 3.4 9.6 6.5 1981 1.0 0.4 19.0 6.5 1982 1.8 4.4 12.8 5.0 1982 1.0 2.9 6.5 3.1 1984 0.5 6.0 3.9 1.1 1985 0.7 6.6 0.2 0.5 1986 0.6 3.2 5.3 0.7 1987 0.4 1.5 0.5 0.2 1988 0.0 0.8 0.5 0.2 1989 0.0 4.9 0.2 0.1 1990 0.1 0.4 0.0 0.3 1991 0.0 0.3 0.3 0.0 1992 0.0 0.1 1.3 0.0 1993 0.1 0.3 0.4 0.0 1994 0.0 0.7 0.0 0.0 1995 0.4 0.2 0.1 0.2 1996 0.6 0.2 1.3 0.0 1997 0.4 0.1 0.0 0.0 1998 0.1 0.4 0.0 0.0 1999 0.1 0.1 0.2 0.1 2000 0.0 0.1 0.3 0.0 2001 0.9 0.1 0.1 0.0 1average catch per trawl tow 2number per million cubic meters of intake water 3index not available 2006 J.T. DeAlteris, T.L. Englert, and J.A.D. Burnett 115 Appendix 4. Scup abundance indices for Narragansett Bay and Mount Hope Bay, 1972–2001. Upper MHB Lower MHB Narragansett Bay MRI BPS RIDFW RIDFW URIGSO Year trawl1 impingement2 trawl1 trawl1 trawl1 1972 2.6 18 1973 0.2 0.00 94 1974 5.3 0.04 41 1975 3.6 0.05 130 1976 61.9 0.05 331 1977 0.6 0.03 64 1978 2.7 0.08 39 1979 7.8 0.08 7 71 64 1980 36.2 0.03 51 61 47 1981 10.7 0.01 15 71 96 1982 3.2 0.10 58 18 40 1982 7.5 0.01 15 79 108 1984 2.3 0.04 16 286 51 1985 3.9 0.01 28 25 86 1986 0.3 0.01 75 89 62 1987 0.3 0.01 19 113 79 1988 0.0 0.03 1 396 42 1989 1.4 0.02 63 225 176 1990 0.0 0.05 44 387 124 1991 3.2 0.28 414 389 286 1992 1.8 0.06 243 289 75 1993 0.3 0.02 10 248 31 1994 0.0 0.00 22 34 21 1995 0.4 0.02 26 64 48 1996 0.0 0.00 17 70 58 1997 1.4 0.02 13 143 54 1998 0.2 0.00 9 23 45 1999 0.1 0.02 215 461 108 2000 0.3 0.02 428 468 176 2001 3.4 0.00 85 713 105 1average catch per trawl tow 2number per million cubic meters of intake water 116 Northeastern Naturalist Vol. 13, Special Issue 4 Appendix 5. Tautog abundance indices for Narragansett Bay and Mount Hope Bay, 1972–2001. Upper MHB Lower MHB Narragansett Bay MRI BPS RIDFW RIDFW URIGSO Year trawl1 impingement2 trawl1 trawl1 trawl1 1972 0.9 2.5 1973 1.0 0.2 3.2 1974 0.9 0.2 2.5 1975 1.2 0.2 3.1 1976 1.2 0.3 4.1 1977 1.0 1.7 1.5 1978 1.7 1.3 1.0 1979 1.0 0.2 0.8 4.3 1.1 1980 0.9 0.4 7.7 2.5 0.4 1981 1.0 0.2 2.5 0.7 0.5 1982 1.8 0.7 0.2 0.3 0.9 1982 1.0 0.3 1.0 0.8 1.4 1984 0.5 0.7 11.9 1.9 1.5 1985 0.8 0.8 1.0 1.4 0.4 1986 0.5 0.6 4.0 2.5 0.3 1987 0.2 1.6 0.5 2.5 0.2 1988 0.1 0.8 1.3 1.1 0.1 1989 0.1 0.6 0.2 1.0 0.4 1990 0.0 0.1 0.4 1.0 0.1 1991 0.0 0.1 0.2 1.5 0.1 1992 0.1 0.1 0.4 0.6 0.1 1993 0.2 0.4 0.6 0.6 0.2 1994 0.0 0.2 0.2 0.3 0.1 1995 0.0 0.2 0.1 0.4 0.1 1996 0.2 0.1 0.9 0.4 0.2 1997 0.0 0.6 0.3 0.2 0.1 1998 0.0 0.2 0.2 0.3 0.0 1999 0.1 0.4 0.7 0.5 0.3 2000 0.0 0.3 2.4 0.5 0.2 2001 0.2 0.2 0.5 1.0 0.3 1average catch per trawl tow 2number per million cubic meters of intake water