Phytoplankton Primary Production and Fish Community Structure in New Bedford Harbor: A Comparison Study to Evaluate Human Impacts

This study fulfilled a portion of an ongoing program at the Environmental Protection Agency (Narragansett, RD to examine cumulative human impacts in New Bedford Harbor, Massachusetts. One of the many changes New Bedford Harbor has experienced since the 1600s has been an increasing urban population along with an increased volume of poorly treated wastewater and combined sewer overflows. In addition to the threat of increased production and anoxia of harbor waters, industrial toxic wastes deposited in the harbor over the last three to four decades have raised concern for the health of estuarine inhabitants. These toxic wastes include the industrial discharge of polychlorinated biphenyls from electrical component manufacturing plants, and the discharge of heavy metals into the Acushnet River, the primary tributary into the harbor. The Environmental Protection Agency has approached the problem by examining anthropogenic impacts on various ecosystem components. A relatively pristine estuary, the Slocums River, Massachusetts, was selected to compare parameters which may indicate differences in the state of estuarine health between the two sites. The objects of this portion of .the study was to examine differences in rates of phytoplankton primary production between the two sites to detect eutrophication in New Bedford Harbor. This was measured by performing in situ incubations for production and respiration of oxygen over the summer of 1994. The top trophic consumers, estuarine fish, were also studied at both sites to examine differences in fish biomass,

The study area, New Bedford Harbor, and comparison 28 . site, the Slocums River, are both located on the southeast coast of Massachusetts. Locations of deep stations, A, B, and C as well as shoreline stations 1, 2, 3 and 4 are indicated. Transects for macroalgae biomass are indicated by a line with arrowheads.          x Figure 6. Growth rates of Menidia species based on length frequency data for the periods of June 1994-July 1994, July 1994-August 1994and August 1994-83 October 1994. Closed circles represent New Bedford Harbor, open circles represent the Slocums River. The range lines represent standard deviation calculated by the bootstrapping program.   Table 1. Station locations, depths and salinity ranges for the study site, New Bedford Harbor, and comparison 38 site, the Slocums River. Stations A, B,.c were in deeper ranges of the estuary while stations 1, 2, 3 and 4 were shoreline stations where surface water was sampled. Table 2. Integrated net phytoplankton primary production and respiration values to lm depth in New Bedford Harbor and the Slocums River, and to 4m depth in New Bedford 39-40 Harbor alone. Production rates were measured by using light and dark incubation bottles at depth. See Appendix A for production and respiration values at each depth. Table 3. T-tests performed on station averages of net phytoplankton primary production rate, respiration rate, light extinction coefficient k, chlorophyll a, Dissolved organic nitrogen, NH4, N03 + N02, P04, and Si04. Station averages represent the mean of the entire study period, June 1994-41 May 1995. A nonpooled t-test for two normal populations using independent samples (standard deviations not assumed equal) was used to determine if populations were significantly different at the 953 confidence limit. The null hypothesis was that the two populations were equal (NB=SR), if the test statistic fell outside of the t-critical range, then the alternative hypothesis was accepted (NB~R). Table 4. Production estimates for phytoplankton for entire estuary at both sites. Production estimates include areas of depth (see Appendix C for .area of different depths and phytoplankton production rates integrated to depths; see Appendix D for production and respiration integrations 42 at each depth). These production rates are based on net primary production. Summertime estimates were based on the assumption that a four hour incubation (from 10 A. M. to 2 P. M.) represents 553 of daytime production, and the summer consists of 90 days, from June to August. Table 5. The VERSAR/URI criteria for assessment of eutrophication in an estuary. The corresponding scores for indicators in New Bedford Harbor and the Slocums River are listed below (See Appendix E for concentrations). The scores were given for average concentration for each indicator over the study period. Sediment organic carbon and oxygen 43-44 saturation were not measured at every station, consequently an average over the entire estuary was used to develop a score for these two indicators. The overall score at each station was calculated by the summation of all indicator scores, and divided by the number of indicators, which in this case was five. Table 6. Dennison index criteria  for growth of submerged aquatic vegetation. Light attenuation 45 coefficient, chlorophyll a, Dissolved Inorganic Nitrogen (DIN) and Dissolved Inorganic Phosphorous (DIP) were measured in New Bedford Harbor and the Slocums River.
Fish Biomass, Abundance and Community Structure in New Bedford Harbor  (Table 2a) and the Slocums River 87-90 (Table 2b) described by abundance and biomass throughout the entire study period. Table 3a and 3b. Data used in analyzing community similarity . Species counts at stations 1, 2, 3 and 4 during July and August in New Bedford Harbor (Table 3a) 91-92 and the Slocums River (Table 3b). A total of sixteen replicates from each location were used. (J. Heltshe, pers. comm.). Table 4. Nonpooled t-test for independent samples assuming unequal variances (NB=New Bedford, SR=Slocums River) for testing significant differences at the 5% level of 93 significance for monthly average biomass (based on wet weight and dry weight) and monthly abundance from Xlll Table 5. Nonpooled t-te~t for independent samples assuming unequal variances (NB=New Bedford, SR=Slocums River) for testing significant differences at the 5% level of significance for estimates of diversity 94 during the summer, fall/winter and spring seasons. The indices used were the Shannon-Wiener index, the Simpson index and species richness. Table 6. The Estuarine Biotic Integrity index (EBO criteria based on wet weight biomass. Scores developed for each station at each site (NB=New Bedford, SR=Slocums River) 95 were averaged over four observations taken over July and August, with each observation being the average of three seines. Table 7. The Estuarine Biotic Integrity index (EBO criteria based on abundance. Scores developed for each station at each site (NB=New Bedford, SR=Slocums River) 96 were averaged over four observations taken over July and August, with each observation being the average of three seines. Table 8. Nonpooled t-test for independent samples assuming unequal variances (NB=New Bedford, SR=Slocums River) for testing significant differences 97 at the 5% level of significance for EBI scores. Average EBI scores at each site were tested for c:Ufferences. Table 9. Nonpooled t-test for independent samples assuming unequal variances (NB=New Bedford, SR= Slocums River) for testing significant differences at the 5% level of significance for condition factor k. Differences 98 between the two sites for condition of Menidia species, Fundulus heteroclitus and Fundulus majalis were tested. Table 10. Comparison of species found in New Bedford · Harbor and the Slocums River with species assemblages 99-102 in Waquoit Bay . XIV

INTRODUCTION
Anthropogenic activities are causing eutrophication of many coastal waterways and estuaries adjacent to large urbanized areas. There is a relationship between eutrophication of these marine environments and nitrogen enrichment, and this relationship results in modifications to ecosystem structure Nixon 1995). A recent definition describes eutrophication as an increase in the rate of supply of organic matter to an ecosystem (Nixon 1995).
Therefore, eutrophication is a process resulting from human impacts that redefines the trophic status of an ecosystem, and is mediated by a change in metabolic activity or organic inputs. Metabolic activity by primary producers is from phytoplankton, sediment microalgae and macroalgae in coastal environments.
The relationship between eutrophication and nitrogen enrichment has been examined by cross-estuary approaches, controlled experiments, historical approaches and comparison studies. A wide variety of marine systems ranging from the open ocean to heavily nutrient-loaded estuaries showed a positive correlation of primary production with dissolved inorganic nitrogen (DIN) input . Examination of the estuaries showed that these systems have the highest production of all marine systems. Plots of primary production with DIN input show that estuaries have large variation . This high variation is due to variation in flushing times, nutrient regeneration by the water column and the benthos, and spatial or temporal variations of both nitrogen input, primary production and light attenuation. Experiments in marine enclosures also reflected nitrogen limitation of marine environments Oviatt et al. 1994). Controlled experiments exhibited an increase in phytoplankton production and abundance (measured as chlorophyll a) with a gradient of nitrogen loading. Changes in phytoplankton abundance or production and average DIN concentration or nitrogen loading have been recorded in some coastal ecosystems over time Smith et al. 1986;Johanssen and Lewis 1992;Wetsteyn and Kremkamp 1994;Wienhuis and Small 1994). Most of these systems had an elevated phytoplankton abundance or production rate which was consistent with higher nitrogen availability. However, the magnitude of the response to nitrogen was not consistent, and the rate of increase was not parallel in each system . A comparison of three subwatersheds in Waquoit Bay revealed that the watershed receiving the highest groundwater concentration of nitrogen consequently had the highest rate of phytoplankton production and macroalgae biomass .
We have examined in situ nutrient concentrations and phytoplankton production in New Bedford Harbor, Massachusetts, to measure the effect of cumulative human impacts. This study was part of a larger project to analyze cumulative human impacts on New Bedford Harbor using a less disturbed reference site, the Slocums River, MA, in a comparison study (Environmental Protection Agency, Narragansett, RI). Multiple anthropogenic influences have accumulated in New Bedford Harbor since the late 1600's when agriculture was predominant. From 1755 to 1875, New Bedford Harbor was characterized as the world's largest whaling port. During the industrial revolution, the textile industry was dominant in New Bedford, and with it came many mills and industrial plants which were built along the Acushnet River from 1846-1890 (Richard Voyer, U.S. E. P. A., personal communication). Today, the city of New Bedford continues to develop and encourage the movement of new industry to the area. Municipal wastewater from the neighboring town of Fairhaven is directed into the harbor, and combined sewer overflows from the city of New Bedford contribute high volumes of untreated wastewater in times of heavy rainfall. Industrial discharges include the discharge of PCBs into the Acushnet River by electronic capacitor manufacturers between 1950 and 1970; and runoff of heavy metals, primarily Cu, from industries located on the Acushnet River.
The construction of the hurricane barrier, which separates New Bedford from Buzzard's Bay, created a sediment trap for these industrial contaminants . The control area for the study, the Slocums River, has experienced fewer alterations due to human activities over the past three centuries. There are few homes, which contain individual septic systems, and there are no industrial activities in this rural area.
Those involved in water quality management are interested in measuring eutrophication and describing estuarine health. The most direct parameter to measure eutrophication is carbon production. Unfortunately, the methods for measuring carbon production are labor intensive and require an intensive time series of measurements. There is interest. in the development of an index of water. quality criteria which indicates degradation of estuarine health. In the Chesapeake Bay, researchers have established the water quality criteria necessary to support submersed aquatic vegetation, a major resource for fish habitat Batiuk et al, in press) .
Another index ranks the state of eutrophication using water quality parameters based on parameters measured in Delaware coastal bays (Frithsen et al. 1995). This eutrophication index was based on parameters measured in Delaware coastal bays. We have measured some of these parameters in this comparison study in addition to phytoplankton production; these indices have been applied in this study to determine the state of eutrophication in New Bedford Harbor.

STUDY LOCATIONS
New Bedford Harbor (Fig. 1), an estuary with an area of 390 ha, is located on the southeast coast of Massachusetts, the Acushnet River is the major tributary to the harbor. The harbor opens into Buzzard's Bay through a hurricane barrier located at the mouth. The maximum depth of the harbor is 10 min the dredged shipping channel. Average depth from the Coggeshell Street bridge to the hurricane barrier is 6 m. However, there are areas outside of the shipping channel with depths reaching to 10 m (Fig 2a) . The New Bedford side of the harbor, with population 100,000 (1990), consists of many industries and fish processing plants, while the City of Fairhaven, population 16,000, on the east side of the harbor consists of residential homes, marinas and fishing ports. Spartina marshes border the Acushnet River on the Fairhaven side, however there are few marsh areas in the middle and lower harbor. Areas of Ulva growth are evident along the shoreline in association with-storm drains.
The Slocums River (Fig. 1), area of 180 ha, was selected as a comparison site. The estuary is much shallower with few areas where depth is beyond 3 m, and an average depth from the Gafnee Street boat ramp to the mouth of the river of 2 m . The Slocums River is bordered by a salt marsh with extensive shoreline stretches of Spartina, within the embayment extensive mats of Ulva growth are on the bottom in shallow areas. There are few homes in the area, and some agricultural activity is present.

Phytoplankton Primary Production
Phytoplankton primary production was measured at stations A, B, Cat both sites from June 3, 1994 through September 10, 1994 (Tables 1 and 2) .
Phytoplankton production and respiration were measured by changes of oxygen concentration during incubations of light and dark bottles at different depths. Three 300 mL light BOD bottles and one 300 mL dark BOD bottle were incubated at four depths, 0.2 m, 0.5 m, 1 m and 3 or 4 min New Bedford; the same number of bottles were incubated at three depths in the Slocums River, 0.2 m, 0.5 m and 1 m. The water samples were collected by dropping a Niskin bottle at one depth, the sample was brought back up to the boat and incubation bottles for that depth were filled and overflowed to remove oxygen bubbles. Three additional light bottles were filled with water from that depth and fixed immediately with chemicals for the Winkler titration , to determine the initial 0 2 concentration. The bottles were hooked onto cross-bars attached to a line, so that one line had one crossbar (accommodating four bottles) at each depth. The incubations were set afloat with a buoy and weight at each station location for four hours, approximately from 10 am to 2 pm. After the incubation, the line was pulled up and bottles removed from crossbars. The incubation bottles were fixed with the chemicals for the Winkler titration to determine production of 0 2 from the light bottles and respiration in the dark bottles. The initial samples, light and dark bottles were transported back to the laboratory where titrations were performed on an automatic burette, a Radiometer ABU 91.
Oxygen production was converted to carbon production by use of an average photosynthetic quotient (0 2 / C0 2 ) for natural populations of 1.2, this PQ was recommended by John H. Ryther and reflects an average of photosynthetic quotients from many studies of marine algae .
Daily production rates were calculated by assuming that the four hour incubation period represents 55% of daytime production ).
Summer production rates were calculated by assuming that 90 days were in the summer period. Carbon production to depth was calculated by integrating the area under the curve representing production as a function of depth. The areas of different depth ranges were measured in each estuary, and total production was calculated by adding the products of area of depth x integrated production at that depth.
One transect for measuring macroalgal abundance was performed in the Slocums River on August 12, 1994 and in New Bedford Harbor on August 17, 1994 (Fig. 1). The transect in the Slocums River was across a shallow embayment which was typical of the many shallow embayments in the Slocums River. New Bedford Harbor was too deep to perform a transect across the width, and the shallow areas upstream were highly contaminated, so a transect was performed along the shoreline adjacent to two storm drains.
An Ekman grab with area of 0.05 m 2 was employed, twelve quadrats were sampled in each transect. The macroalgae from each quadrat ~as transported back to the laboratory where the sample was cleaned and wet weight measured. Dry weight was measured after drying the sample in a 60° C.
During phytoplankton production incubations, profiles of light attenuation, temperature and salinity were made at stations A, B, and C.
Temperature and salinity profiles were performed using a Beckman salinometer. Water column light measurements and ambient light were measured simultaneously with a LICOR LI-1000 data logger; water column incident light at depth to ambient incident light at surface ratios were recorded. Light attenuation coefficient, k, was calculated by

Nutrient and chlorophyll a concentrations
Nutrients and chlorophyll a concentrations were measured by 10 sampling at 1 m with a Niskin bottle at stations A, B and C during phytoplankton production incubations from June 3, 1994 to September 10, 1994. Nutrient samples were also taken at stations A, B, C, 1, 2, 3 and 4 from June 8, 1994 to January 1995. Samples at stations 1, 2, 3 and 4 were surface water samples collected at shoreline locations by inverting a sample bottle.
Stations A, B, C, 1, 2, 3 and 4 were sampled from September 1994 to May 1995 by collecting surface samples.
Nutrient samples were filtered with a 0.4 uM polycarbonate membrane filter and preserved with chloroform in the field . Nutrient samples were analyzed for concentrations of ammonium, nitrate + nitrite, phosphate and silicate on a Technicon Autoanalyzer II  .
Chlorophyll samples were filtered on 0.7uM Whatman glass fiber filters in the laboratory. The filters were ground, chlorophyll was extracted with acetone, and fluorescence was determined on a Turner Designs 10 fluorometer .

Application of indices 11
The criteria for the eutrophication index (Frithsen et al, 1995) involved ten parameters, however, only five parameters were measured in this study.
These parameters measured were nitrate + nitrite, phosphate, chlorophyll a, percent sediment organic carbon and oxygen saturation. A mean score was developed by assigning a score from 1 (oligotrophic) to 5 (eutrophic) based on criteria for each parameter; the scores were added and divided by the number of parameters used to get an average score. A replicate for each station (yearly mean) was used to obtain an average score in the eutrophication index, so that each parameter had seven replicates representing seven stations. Sediment organic carbon (%) data required for the eutrophication index was obtained from Skip Nelson (US EPA, Narragansett, RI). Averages of percent sediment organic carbon were obtained from many samples taken on a polygon grid south of the Coggeshell street bridge in New Bedford, and south of the Gafnee bOat ramp in the Slocums River, one average percent sediment organic carbon for each site was used for all replicates. Oxygen saturation data was converted from g/m 3 0 2 measurements taken for primary production estimates in the summertime taken at stations A, B and C, the average number at stations A, B and C was used for each replicate. Parameters in the Dennison water quality index used in this study were the light attenuation coefficient, chlorophyll a, dissolved inorganic nitrogen and dissolved inorganic phosphor':1s. Index criteria were given for different salinity regimes, the salinity of all stations in this study ranged from 15-32, therefore the criteria for the polyhaline range was used. An average value of all stations throughout the study period was used in the Dennison index.

12
A non-pooled t-test for two population means (assuming unequal standard deviations) was performed for nutrients, chl a, light attenuation and phytoplankton production to determine significant differences between the test site and control site . A hypothesis test, with the null hypothesis being Ho: u 1 = u 2 where Ho = null hypothesis, u 1 = population one, u 2 = population two, for two population means with normal distribution, but not necessarily equal standard deviations involved the calculation of a test statistic. The critical values of the t-test were decided upon by the significance level, alpha = 0.05, and the degrees of freedom. Rejection of the null hypothesis occurred if the test statistic, t, fell outside of the critical values.

Phytoplankton Primary Production
Phytoplankton production incubations showed higher phytoplankton production rates in New Bedford Harbor. An example of one incubation from each site was typical of the difference between net production rates measured at the two sites where respiration rates were similar (Fig. 3).
Production rates integrated to 1 m were higher in New Bedford with the majority of sampling dates (Table. 2). Box plots of production rates measured in stations A, Band C of both estuaries showed the average phytoplankton production to be higher in New Bedford, in addition the variability was greater at the New Bedford stations (Fig. 4). The average of all net production rates integrated to 1 m was 0.58 + /-0.12 (standard deviation) g 0 2 m-2 hr· 1 in New Bedford while the average net production to 1 min the Slocums River was 0.11 +/-0.03 (standard deviation) g 0 2 m-2 hr· 1 (Table 3). Respiration rates to 1 min New Bedford Harbor averaged 0.09 +/-0.06 (standard deviation) g 0 2 m-2 hr· 1 , and to 1 m in the Slocums River was 0.03 + /-0.01 (standard deviation) g 0 2 m·-2 hr· 1 (Table 3). Station averages of phytoplankton production and respiration rates were significantly higher in New Bedford.
High variation in respiration rates and small sample size made site to site differences not significant at the 5% level of significance (Table 3).
The summation of production per areas of 1 m depth, 2 m depth, and 4 m depth (Appendix D) in New Bedford was used to calculate total production across the estuary. In New Bedford Harbor, total production was 1.13 x 10 6 +/-1.65 x lc>5(standard deviation) g C hr-1 (Table 4, Appendix C); and divided by the area of New Bedford Harbor gives a normalized value of 0.29 +/-0.04 g C m-2 hr-1 (Table 4). In the Slocums River, the summation of plankton production at 0.2 m, to 0.5 m and 1 m (Appendix C) gave a total of 7.05 x 10 4 +/-1.87 x 10 4 gC hr-1 (Table 4, Appendix C). Unfortunately, this is an underestimate of total production in the Slocums River because there are areas of the Slocums River deeper than 1 m which light enters and where production occurs. This underestimate is not likely to be great because the phytoplankton production rate was very low at 1 m, and further depths are likely to have even less production. Normalization to area in the Slocums River gives an estimate of 0.04 +/-0.01 g C m-2 hr-1 (Table 4). An estimate for New Bedford production rate over the summer season (June, July and August) is 165 +/-26 g C m-2 summer 1 , and much higher than the estimated summer rate for the Slocums River, 20 +/-3 g C m-2 summer-1 .

Macroalgae biomass was higher in almost all quadrats in the Slocums
River transect when compared to the New Bedford macroalgae transect (Fig. 5 There was high variability in the light extinction coefficient, k, in both New Bedford Harbor and the Slocums River. The average k measured at stations A, Band C was 0.90 in New Bedford and 0.98 in the Slocums River. The t-test showed that there were no significant differences between light attenuation coefficients in the estuaries (Table 3). Temperature and salinity profiles at all stations indicated that the water column was well mixed throughout the summer.
Chlorophyll a

15
The yearly cycle of chlorophyll a concentrations indicated that mean phytoplankton abundance was higher in New Bedford in July and August and in the springtime during January, February and March (Fig. 6). Mean chlorophyll a concentrations across the year was 7.64 mg/m 3 for New Bedford and 3.28 mg/m 3 for the Slocums River. The difference was significant at the 5% level of significance with the nonpooled t-test for independent population means. (Table 3).

Nutrients
Nutrient samples analyzed from June 8, 1994 to January 28, 1995 showed significant differences between the two sites. Mean dissolved inorganic nitrogen concentrations increased over the fall and early winter at both sites. However, New Bedford showed a pattern of higher mean dissolved inorganic nitrogen concentration over the Slocurns River, especially in the fall and winter (Fig. 7 that the average of ammonium over the study period in New Bedford Harbor was higher than station averages in the Slocums River (Fig. 9). These differences in ammonium, dissolved inorganic nitrogen and phosphate concentrations at the two sites were significant at the 5% level of significance (Table 3). In addition, station averages of phosphate were significantly different over the study period; however, mean nitrite + nitrate and silicate were not significantly different when station to station variations were taken into account.

Application of indices
The application of the eutrophication index (Frithsen et al. 1995) resulted in a score of 1.09 +I -0.37 (standard deviation) for the Slocums River and 2.25 +I-0.3 (standard deviation) for New Bedford Harbor (

DISCUSSION
Higher phytoplankton primary production and abundance indicated that New Bedford Harbor was eutrophied at least in relation to the comparison site, the Slocums River. This conclusion was based on the definition of eutrophication being "an increase in the rate of supply of organic carbon" (Nixon, 1995). Integral primary production per unit area was higher in New Bedford because it is a deeper system, but also primary production per unit volume was higher in New Bedford Harbor. According to Nixon (1995), another source of eutrophication besides phytoplankton production is an increase in carbon production from macroalgae. The carbon production estimates in this study are based on phytoplankton production, and do not include production from macroalgae, seaweeds, or benthic microalgae. A transect across a shallow embayment in the Slocums river showed higher macroalgae biomass than along the shoreline in New Bedford Harbor, where shallow embayments are limited. The greater percentage of shallow embayments in the Slocums River which allow light to reach the bottom along with higher macroalgal biomass suggested that prod~ction from macroalgae may have been higher in the Slocums River than in New Bedford Harbor. A study of eutrophication in many different estuaries and lagoons has shown that the shallow salt marshes and lagoons have a higher contribution of production from macrophytes over phytoplankton .
seasonal phytoplankton production rate. The west passage of Narragansett Bay had a summertime average daily production rate of 0.7 gC/m 2 I day in 1971 , while the daily production rate in this study was 2.10 +/-.14 gC/m 2 /day in New Bedford and 0.29 +/-0.07 gC/m 2 /day in the Slocums River. Light availability was also restricted in both New Bedford and the Slocums River when compared to Narragansett Bay, where in 1971 light extinction coefficient in Narragansett Bay varied from 0.5 to 0.7.
The coefficient in New Bedford Harbor was 0.90 +/-0.22 and 0.98 +/-0.20 in the Slocums River. The decreased light availability in New Bedford Harbor may be due to increased phytoplankton biomass. However, in the Slocums River, decreased light transmission may be due to sediments being disturbed in shallow water regions.
Higher phytoplankton production in New Bedford Harbor, compared to the Slocums River, is likely due to increased sewage effluent directed into the harbor, resulting from high population density. In 1775, the population of New Bedford was only 500, and by the 1920's had increased to a maximum of 125,000 in the era of the textile mills. Since the 1920's, the population has slowly been decreasing to a p~esent day estimate of 98,000 (Richard Voyer, EPA, personal communication). In comparison, the town of Dartmouth at the head of the Slocums River has a present day population of less than 1,000.
Ammonium and phosphate concentrations are typically increased in sewage effluent, and the in situ concentrations of ammonium and phosphate are significantly higher throughout the year in New Bedford Harbor. It is likely effluent, and the in situ concentrations of ammonium and phosphate are significantly higher throughout the year in New Bedford Harbor. It is likely that these higher in situ concentrations in New Bedford Harbor paralleled higher nutrient loadings. Higher phytoplankton production has been positively correlated with higher nitrogen loading rates in estuarine systems Nixon 1995).
The eutrophication index, based on the average of five water quality parameters measured throughout the year, showed the Slocums River as being oligotrophic, and New Bedford was only slightly above oligotrophic, but not in the mesotrophic range (Frithsen et al, 1995). This contrasts direct comparison of phytoplankton primary production and chlorophyll concentrations, which show a large difference in phytoplankton biomass and production between the two sites. However, only using five of the ten parameters in the index may impede the utility of the index in showing differences in the extent of eutrophication. The only parameter which consistently showed New Bedford as being eutrophic at every station was percent sediment. organic carbon, which was probably a result of the high primary production in New Bedford. However, the nutrient concentrations, oxygen concentration and chlorophyll a were not high enough to score in the mesotrophic range (score of 3). The mean parameters measured in the Slocums River throughout the year showed that at each station, water quality parameters met criteria for growth of submerged aquatic vegetation (SAV) .
However, yearly mean parameters in New Bedford showed that none of the stations meet the criteria for dissolved inorganic phosphate, and some of the stations fail to meet criteria for dissolved inorganic nitrogen. Therefore the Dennison water quality index does show a difference between the two sites, but does not attempt to rank how strong the difference is.
There are several speculations as to why the index did not recognize a strong difference in eutrophication between the two estuaries, and did not classify New Bedford Harbor as eutrophic. Only five out of ten parameters in the index were measured in this study, the metrics which were not measured were total dissolved nitrogen and phosphorus, total particulate nitrogen and phosphorus, and total particulate carbon. The parameters measured only involved inorganic nutrient concentrations. In situ, inorganic, nutrient concentrations may not always give an accurate picture of the rate of primary production for several reasons: 1) high primary production may use nutrients at a high rate, leaving a low in situ nutrient concentration, 2) differences in the source of primary production, for instance a phytoplankton dominated system and a system dominated by seagrasses and benthic macroalgae may have different rates of nutrient uptake, 3) differences in benthic remineralization of nutrients, 4) different residence times, S) high spatial and temporal variability of nutrient concentrations. There are also problems with using chlorophyll a concentrations as an indicator of eutrophication. Carbon to chlorophyll a ratios may not be constant; and chlorophyll a is not an indication of production arising from seagrasses a· nd benthic algae. Oxygen saturation must be measured completely throughout the estuary to find either shallow or deep areas with anoxia. The light attenuation coefficient may not always be a good indicator because high variability may arise from stirred up sediments in shallow regions during windy weather.   ""         Table 4. Production estimates for phytoplankton for entire estuary at both sites. Production estimates include areas of depth (see Appendix C for area of different depths and phytoplankton production rates integrated to depths; see Appendix D for production and respiration integrations at each depth). These production rates are based on net primary production. Summertime estimates were based on the assumption that a four hour incubation (from 10 A. M. to 2 P. M.) represents 55% of daytime production, and the summer consists of 90 days, from June to August.  for eelgrass habitats, since the best correlation between poor water quality characteristics and poor fish parameters has been in eelgrass habitats. The EBI may have not been applicable to this study, since there were no eelgrass beds in the two sites.

INTRODUCTION
The observation that five of the six most important commercial fishery species in the United States are somehow dependent on estuaries, and that 75% of the nation's commercial fishery landings of fish and shellfish (by weight, 1985) are composed of estuarine dependent species (Chambers 1991) demonstrates that estuaries play a vital role in fisheries. These numbers translate into an annual economic value to society approaching $14 billion, demonstrating that estuarine ecosystems are essential to the fishing industry.
Fish and shellfish depend on estuaries for reproduction, nursery areas and food sources.
Estuaries are commonly adjacent to centers of population and therefore subjected to environmental stresses attributed to anthropogenic activities.
Examples of the types of human impacts which disturb fisheries are overfishing, eutrophication, industrial waste additions, dredging, introduction of artificial reefs, the filling of wetlands, and const~uction of barriers. Chemical stress on aquatic ecosystems affects the occurrence of some Species, and encourages colonization by others. Rapid changes in overall community composition beyond the simple appearance or disappearance of indicator species are found in a.Ssociation with chemical stresses. These stresses include acidification, herbicides, pesticides, heavy metals, oil, pulp mill effluents and organic enrichment. In addition, heavily stressed systems tend to have reduced biomass, abundance, species richness and species diversity relative to pristine ecosystems . The alteration of habitat due to dredging and the filling of wetlands will lead to declines in fish spawning, food sources, and hiding areas.
Eutrophication of coastal waterways has many impacts. In Swedish waters, the Kattegat and the Belt Sea, increased inputs of nutrients in combination with relatively low water exchange and high macro-algal biomass have led to minimal oxygen concentrations in bottom-water . These events have been responsible for mortalities of benthic animals and decreased fish catches around Sweden. Eutrophication may change the food web structure in an ecosystem. Decreased cod stocks in the Baltic Sea are blamed in part on the disappearance of amphipods and other benthic fauna from deeper Baltic bottoms when an oxygen deficiency has developed . Eutrophication increased the amount of a filamentous algae in the Baltic Sea which has been responsible for mortality of Baltic Herring eggs . One hypothesis for the decline of Striped Bass tMorone saxatilis) in Chesapeake Bay is that nutrient enrichment and greater planktonic production have decreased concentrations of dissolved oxygen and thus impacted deep-water habitat for adults .
Although many reports in the literature show degradation of fish populations due to eutrophication; there is also evidence of increased fisheries yield with increased primary production, or eutrophication (Aleem 19 n; Cross 1975;Hansson 51 and Rudstam 1990). The concept of the sea being a farm, where input of dissolved organic nutrients regulates the level of primary production and eventually the yield of fish from marine ecosystems, was initiated about a century ago . Predictions of fish production from primary production have shown that a significant portion of fish production occurs in coastal waters where primary production is the highest . The relationship between fisheries yield and measured primary production in a variety of marine systems, including estuaries, shows a positive correlation . However, few studies have examined individual estuarine systems for a relationship between primary production and fish production, perhaps because there are few studies reporting both primary production and fish yield for near coastal waters. Estuarine fish yield has a large variation of response to primary production due to differences between estuaries. These are differences in residence time and circulation which influence anoxic events. Differences may also be evident in transfer efficiency between trophic levels and foodweb structure, habitat structure and fishing pressure.
An Estuarine Biotic Integrity Index (EBI) has been developed to indicate estuarine ecosystem health in New England estuaries ).
The EBI is designed to indicate the quality of habitat by using fish biological parameters. The indicators of ecosystem health in the index are top trophic level consumers, estuarine fish. Estuarine fish are higher trophic levels assumed to be sensitive to degradation since they require a wide diversity of ecosystem functions to be sustained. The EBI was modified from the original Index of Biotic Integrity (IBD for evaluating water quality and ecosystem conditions in freshwater streams (Karr, 1981(Karr, , 1991 Oviatt, 1995). Industrial activities have also increased sediment concentrations of PCBs to 100,000 ppm in some areas  and have increased the concentration of toxic heavy metals to more than 1% dry weight of sediments   (Fig. 2a). A hurricane barrier separates the outer harbor from Buzzard's Bay.
The Slocums River, (Fig. 1), a smaller estuary than New Bedford Harbor, with an area of 180 ha, is located in a rural area of southeastern Massachusetts with few homes, and no industrial activity. The average depth of the Slocums River, from the Gaffnee Street boat ramp to the mouth is 2 m.
The estuary is much shallower than New Bedford Harbor, with few areas where depth is beyond 3 m, and a maximum depth is 6 m (Fig. 2b). There are extensive areas of Spartina marsh along the river, and mats of Ulva on the bottom of shallow areas.

Collection of samples
Four stati~ns were selected for sampling in each estuary; two stations were in the middle harbor and two stations were in the lower harbor of New Bedford (Fig. 1). Accordingly, the stations in the Slocums River were selected in the middle and lower part of the river (Fig. 1). The station habitats were either Spartina marsh or sandy or cobbly beach; the salinity ranges were similar between the two sites (Table 1). No eelgrass (Zostera marina) was located at either site. There was more beach habitat in New Bedford in the middle and lower harbor than saltmarsh compared to the Slocums River.
Marsh stations upstream of the Acushnet River were not sampled due to the PCB toxicity of the sediments. The efficiency of the seine was measured by seining a section of Marsh Meadows Wildlife Preserve in Jamestown, RI. An area of 382.5 m 2 was closed off by stretching a beach seine from shoreline to shoreline. An initial seine of the enclosed area was performed using the beach seine employed in the study, the fish in this seine were identified and counted. Seining was continued in the enclosed area until all fish were captured. The fish abundance in the enclosed area was compared to fish abundance counted in the first seine to determine the efficiency of the seine. This procedure was performed five times at the same location to obtain a variance.

Analysis of samples
Each seine catch was analyzed for species and numbers. Fish were identified to the genus and species level, except for Menidia menidia and Menidia berrylina which were classified to the genus. If the number of a species exceeded 100, total numbers were estimated by volume displacement.
Subsamples of major species were taken back to the laboratory along with minor species to measure wet weight and dry weight. Dry weights were measured after drying in a 60"C oven for 4-7 days, depending on size.
Samples were preserved in buffered formalin in the field for transport back to the laboratory. Length of all species were measured on subsamples of 100, or total numbers if less than 100, either in the field or in the laboratory. On a few subsamples, length and dry weight and wet weight were measured concurrently to perform length-weight regressions.
Length-wet weight and length-dry weight regressions were performed for each species using a sample size of at least 60 individuals for dry weight, and 200 for wet weight. Some minor species collections throughout the study were not abundant enough to perform regressions, consequently length and weight were recorded on these minor species continuously throughout the study. A log (length) versus a log (wet or dry weight) plot was used to derive an equation for a straight line. These regressions were used to estimate biomass from length measurements on the rest of the samples.
Pondera! index, or condition factor, k, was calculated by using the where k =condition factor, W =weight (g), L = length (cm), (Weatherley and Gill 1987). Condition factors were calculated for major species present at each site: Menidia species, Fundulus heteroclitus and Fundulus majalis. A subsample of fish from many different seines and stations in each estuary was compiled to calculate k for each species.

Application of the EBI
The guidelines for using the EBI that were met were: 1) sample in late July and August; 2) sample at least three replicates at each station each time; 3) sample at least twice during July and August. It was also recommended to sample eelgrass habitats exclusively, however only marsh and beach habitats were available to be sampled in this study. The EBI was applied using criteria developed for either numbers or wet weight biomass of fish ). The criteria for were: number of species< 6, number of estuarine spawners <3, number of nursery species <3, number of resident species <4, abundance <4 g/m 2 (biomass) or 3.8 individuals/ m 2 (numbers), dominance <3 comprising 90% of total (biomass or numbers), abnormality >0.01, and % benthic fishes <0.9 (based on biomass or numbers). A score was developed for each station in New Bedford and the Slocums River using the average of each parameter across July and August. If the metric did not meet minimum criteria, the parameter in that observation was given a zero, otherwise the raw value of the parameter was used as a score. The scores for each station were added to obtain an overall score which was compared to the score considered low quality habitat, which was <30. all measurements in a pool. The program was designed to perform bootstrap sampling 100 times. Periodic growth rates were calculated by taking the mean and standard error of the 100 bootstrap samples ).

Species Diversity
Two diversity indices were used, the Shannon-Wiener index and the Simpson index. The Shannon-Wiener index is an example of a type I index, Which is more sensitive to changes in the rare species in the community sample. The Simpson index is a type II index, which is more sensitive to changes in abundant species. The diversity was estimated by using a jacknife procedure for both indices (Zahl 1977  ~ommunity Similarity Similarity of fish species composition between New Bedford Harbor and the Slocums River was analyzed by Dr. James Heltshe by using a program designed to test community similarity as described by .
This procedure generates a similarity matrix between the two populations of species by using a distance metric. A permutation test is run to determine if between similarities are significantly differen. t from within similarities.
Sixteen observations were used from July and August at each location in the analysis. The observations were the compilation of three seines at each station, each station was sampled twice during July and August.

T-test
A nonpooled t-test for two population means, assuming that standard deviations were not equal, was performed to detect significant differences between the two sites for condition factor k, biomass m· 2 , numbers m· 2 , diversity indices, species richness and EBI scores. Rejection of the null hypothesis occurred if the test statistic, t, fell outside of the critical values.
A mean k, condition factor, was calculated from a subsample of fish of one particular species from different stations and seines. A sample standard deviation was calculated, and the t-test was performed using the overall mean and deviation of this subset.
Averages fish biomass m· 2 , based on wet weight and dry weight, and abundance (individuals m· 2 ) were calculated for each month in each location.
The t-test was performed using the mean of these monthly values and the .sample standard deviation was calculated between months.
EBI scores at each station over July and August were averaged in each site. The mean between the four stations, and the sample standard deviation were used in the t-test.
The jacknife estimates of diversity and species richness along with variances were used directly in the t-test. An individual t-test was performed for summer, fall/winter and spring to detect significant differences in separate seasons.

Species Assemblages
Fish species assemblages were less diverse in New Bedford Harbor compared to the Slocums River. A total of 27 species were collected between the two sites from June 1994 to May 1995 ( River. There was a greater number of minor species, particularly unique species, in the Slocums River. The community similarity analysis of species composition at both locations performed by Dr. James Heltshe showed that community structure was dissimilar at the 5% level of significance ) (Table 3).

Seine Efficiency
The test for seine efficiency was performed five times, with an average of 88% +/-8% (standard deviation). Total numbers of each seine was used, seine efficiency of individual species was not calculated. The location seined was a flat sandy area without any submerged marsh area, macroalgae or rocks.
The raw data for calculation of seine efficiency is in Appendix B.

Biomass and Abundance
Dry weight biomass throughout the study in New Bedford was 13, 862 g; dry weight biomass in the Slocums River was 7,710 g ( Table 2).The total number of fish collected in New Bedford Harbor throughout the entire study period was 32,027 individuals. In comparison, 20,864 individuals were collected in the Slocums River (Table 2).
Monthly average biomasses m· 2 calculated by wet weight and dry weight, were higher in New Bedford during August, 1994, through December, 1994, than in the Slocums River ( Fig. 3 and 4 the Slocums River average was 0.32 g dry weight/m 2 (corrected for seine efficiency) +/-0.19 (standard deviation) (Table 4). Biomass monthly means based on wet weight were not significantly different between the two sites at the 5% level of significance (Table 4), but were at the 10% level of significance.
Even though dry weight and wet weight data were roughly parallel, the wet weight data was more variable, leading to less stringent differences. New Bedford had a monthly average of 0.47 individuals/m 2 (corrected for seine efficiency)+/-0.46 (standard deviation); and the Slocums River average was 0.27 individuals/m 2 (corrected for seine efficiency) +/-0.22 (standard deviation) ( Table 4). Test of differences in monthly means of individuals m· 2 between the two sites by a t-test showed that there were no significant differences in individuals m· 2 between June 1994 to May 1995 (Table 4).

Growth Rate
Growth rates of Menidia species were higher in New Bedford Harbor from June to July, July to August and August to October than in the Slocums River (Fig. 6). The differences in Menidia growth rates were significant at the 5% level of significance (Appendix E). Fundulus heteroclitus growth rates were similar between New Bedford Harbor and the Slocums River (Fig. 7).
Fundulus majalis growth rates were higher in New Bedford Harbor at the 95% confidence limit from June to July and from July to August (Fig. 8). The growth rates generated by the bootstrapping program are in Appendix F, the length-frequency data pooled for each month used in the bootstrapping program are in Appendix G, and the bootstrapping program itself is in Appendix H.

Diversity Indices and Species Richness
The difference in species diversity and richness between the two sites was most evident in the summer season. In summer, significant differences  (Table 5) .

Application of the EBI
The resulting EBI scores for stations in both sites were below the minimum criteria for good quality eelgrass habitat. An overall score of less than 30 is considered low for the EBI; total EBI scores generated from both biomass and abundances failed to score above 30 (Tables 6, 7). The average EBI score based on wet weight biomass in New Bedford was 10.06 +/-5.18 (standard deviation) and 11.42 +/-3.39 in the Slocums River. Average EBI 65 scores based on numbers were 9.23 +/-4.02 in New Bedford Harbor and 11.38 +/-3.13 in the Slocums River. In comparison, application of the EBI using biomass in Buttermilk Bay gave a medium quality eelgrass station a score of 36, while three low quality stations scored 27, 29 and 30. Using numbers to estimate EBI from the same data gave an EBI score of 39 for the medium quality data and the low quality stations scored 23, 30 and 33 . At-test using the mean and sample standard deviation be~een stations in each estuary showed no significant differences in EBI scores between the two sites (  (Tables 6,7) .
'ondition factor The condition factor, k, was not significantly different between New Bedford Harbor and the Slocums River for Menidia species, or for Fundulus heteroclitus (Table 9). However, there appeared to be a significant difference at the 5% level of significance for Fundulus majalis, where the condition factor was higher for this this species in New Bedford, 0.0032, compared with k=0.0025 in the Slocurns River (Table 9) . The high density of macroalgae in the Slocums River, which was often pulled up in large amounts in the seine, may have interfered with seine efficiency. Throw trap samples are more efficient than beach seines in heavily vegetated sites, with catch efficiencies for epibenthic fishes from 70%

DISCUSSION
to nearly 100% . Beach seines are likely less efficient for active water column species if the seine is filled with macroalgae. The seine efficiency measured in this study, 88% +/-8%, was performed in a sandy area without any macroalgae present. Therefore, it is likely that the seine efficiency was much lower in the Slocums River due to high macroalgae biomass.
Fish condition, biomass and growth rates measured in this study indicated that New Bedford Harbor fish populations were not degraded in relation to a comparison site. The condition factor, k, which is an indicator of health did not differ between the two sites for Menidia species and Fundulus heteroclitus. Furthermore, the condition factor of Fundulus majalis was higher in New Bedford Harbor than the Slocums River. The growth rate of Menidia species was higher from June-October, and Fundulus majalis growth rate was higher from June-August, in New Bedford Harbor then the Slocums River.  has found that counting otilith daily growth rings for growth rates is more accurate than using length frequency data. The presence of cohorts in length frequency data may underestimate the growth rate. The implications of this are that there is a real difference in growth rates of Menidia and Fundulus majalis between the two sites, but these growth rates are likely underestimated. The higher condition factor and growth rate of .E.undulus majalis in New Bedford Harbor may be explained by the fact that Fundulus majalis prefers deeper water. Average monthly dry-weight biomass m-2 was significantly higher in New Bedford Harbor. The fact that average monthly individuals m-2 between the two sites were not significantly different while monthly biomass m-2 was, is an indication that fish were generally larger in New Bedford Harbor. Therefore, New Bedford fish were not only in good condition, but also seemed to have higher growth rates and size.
The increased growth rate and resulting higher biomass of Menidia species in New Bedford Harbor may be due to the higher phytoplankton primary production and phytoplankton biomass in New Bedford Harbor (Absher and Oviatt 1995). The increased primary production was a result of high nutrient loadings due to urbanization of the harbor area. Menidia menidia and Menidia berrylina are both plankton feeders, and feed in the pelagic zone . The higher rate of phytoplankton production in New Bedford Harbor could have led to the dominance of Menidia species in this system. In contrast, Fundulus heteroclitus, a benthic omnivore, was the dominant species in the Slocums River. Although high Primary production may limit fisheries yields in some cases by creating anoxic areas, the eutrophication in New Bedford Harbor does not appear to limit fish yield, at least in comparison to a less disturbed site.
More diverse community structure and higher species richness in the Sloeums River suggested that the Slocums River had more ecological niches available than New Bedford Harbor. A basic tenet of ecology is that each species belongs to a particular niche, and disturbance of habitat may narrow the number of niches available. It is difficult to surmise whether the 70 differences in diversity and species richness were due to natural variations or disruption of habitat by human impacts. The fact that the Slocums River supported many more Spartina marsh areas than New Bedford Harbor, and that more marsh areas were sampled in the s· locums River, suggests that increased diversity was due to more habitat with vegetation which typically supports a wider diversity of species than beach habitats .
The decreased diversity in New Bedford Harbor, relative to the Slocums River, was possibly due to human influence. A land usage map dated back to 1780 provided by Richard Voyer (EPA, Narragansett, RI) showed a larger extent of salt marsh around the harbor area than is present today.
Many of these wetlands were filled for development of wharves and industrial ports. In addition, dredging of the shipping channel, construction of the hurricane barrier and other structures in the harbor have disturbed the natural habitat. These events could have led to the decreased diversity of fish species in comparison to the Slocums River and Waquoit Bay.
The EBI score showed no differences in the quality of habitat in the two estuaries. However, examination of fish community parameters measured showed real differences in fish biomass, where the biomass m· 2 was higher in New Bedford. By contrast, diversity and species richness were higher in the Slocums River. The assumption of the index is that human disturbances will decrease both biomass and diversity. These inversely related parameters between the two sites balanced each other to give a similar total score. The EBI is not designed to detect habitat quality in non-eelgrass habitats; the index indicated that both the Slocums River and New Bedford Harbor have poor habitats for estuarine fish compared to eelgrass habitats. Poor habitat is indicated by low dissolved oxygen, high nutrients, high biomass of macroalgae and reduced circulation . Dissolved oxygen was measured at water quality stations in the middle of each estuary, the levels of oxygen were acceptable at the two sites (Absher and Oviatt 1995).

Nutrient concentrations measured at fish habitat stations were higher in New
Bedford than the Slocums River, but nutrients were not extraordinarily high at either site when compared to concentrations in Delaware Bay and the

June -July
July -August August -October       Table 3a and 3b. Data used in analyzing community s1muanry ~:::>nurn t:L cu, 1990). Species counts at stations 1, 2, 3 and 4 dur.ing July and August in New Bedford Harbor (Table 3a) and the Slocums River (        tPrimary production was measured down to 4m. where production is near zero.