Dissolved Trace Element Concentrations in the East River−Long

(July 2000) and high (April 2001) river discharge conditions in surface waters of Long Island Sound (LIS). To evaluate the impact of fluvial sources t...
0 downloads 3 Views 336KB Size
Environ. Sci. Technol. 2005, 39, 3528-3537

Dissolved Trace Element Concentrations in the East River-Long Island Sound System: Relative Importance of Autochthonous versus Allochthonous Sources N A T H A N I E L J . B U C K , †,‡ CHRISTOPHER J. GOBLER,§ AND SERGIO A. SAN ˜ U D O - W I L H E L M Y * ,† Marine Sciences Research Center, Stony Brook University, Stony Brook, New York 11794-5000, and Marine Science Program, Southampton College of Long Island University, Southampton, New York 11968

Dissolved trace metal (Ag, Cd, Cu, Fe, Ni, Pb, and Zn), inorganic nutrient (NO3, NH4, PO4, H4SiO4), and DOC concentrations were measured at 43 stations during low (July 2000) and high (April 2001) river discharge conditions in surface waters of Long Island Sound (LIS). To evaluate the impact of fluvial sources to the total metal budget of the sound, samples were collected from major tributaries discharging into LIS (Thames, Quinnipiac, Housatonic, Connecticut, and East Rivers). To compare LIS with other coastal embayments, samples were also collected from five LIS coastal embayments (Manhassett Bay, Huntington Harbor, Oyster Bay, Hempstead Harbor, and Port Jefferson Harbor), which are monitored by the U.S. National Status and Trends Program. Metal and nutrient distributions identified two biogeochemical regimes within LIS: an area of relatively high nutrient and metal concentrations in the East River/Narrows region in western LIS and an area in the eastern region of the sound that had comparatively lower concentrations. Mass balance estimates indicated that, during low flow conditions, the East River was the dominant allochthonous source of most trace metals (Ag, Cd, Cu, Ni, Zn) and inorganic nutrients (NO3 and PO4); during high flow conditions, the most influential source of these constituents was the Connecticut River. Mass balance estimates also evidenced a large autochthonous source of Cu, Ni, and Zn, as their spatial distributions displayed elevated concentrations away from point sources such as the East River. Principal component analysis suggested that metal and nutrient distributions in the LIS system were influenced by different seasonal processes: remobilization from contaminated sediments, anthropogenic inputs from sewage discharges and phytoplankton scavenging during the spring freshet, and benthic remobilization during summer conditions. * Corresponding author phone: 631-632-8615; fax: 631-632-8820; e-mail: [email protected]. † Stony Brook University. ‡ Current address: International Arctic Research Center, University of AlaskasFairbanks, Fairbanks, Alaska. § Southampton College of Long Island University. 3528

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 10, 2005

Introduction Increased loadings of toxic metals and nutrients in response to changes in land use have significantly changed the water quality of coastal areas (1-3). An example of such effects can be found in Long Island Sound (LIS), a large estuary located along the northeast Atlantic coast. LIS receives inputs from 86 wastewater treatment plants, 225 industrial facilities, and 16 power plants (4) that have resulted in historically high concentrations of toxic metals in sediments, mussels, and fish (5, 6). Despite the obvious indications that the waters of LIS are heavily impacted by anthropogenic activities, concentrations of dissolved toxic metals in LIS are mostly unknown. Only three peer-reviewed articles that focus on dissolved metals in LIS exist. Studies conducted by Rolfhus and Fitzgerald (7) and Tseng et al. (8) examined the spatial and temporal distributions of Hg within the sound, while Sweeny and San ˜ udo-Wilhelmy (9) focused on concentrations of Ag, Cd, Cu, and Pb within the East River and the western extent of LIS during the summer of 1999. Therefore, despite being one of the most important estuarine systems of the United States, LIS is extremely understudied with respect to dissolved toxic metals. This study was designed to establish the relative importance of various dissolved metal and nutrient sources (e.g. riverine, sewage inputs) into LIS. To accomplish this, we measured concentrations of dissolved trace metals (Ag, Cd, Cu, Fe, Ni, Pb, Zn) and inorganic (NH4+, NO3+, PO4-4, H4SiO4) and organic (DOC) constituents in the surface waters of LIS under high (April 2001) and low (August 2000) river flow conditions. Additional samples were collected from the major rivers discharging into LIS (East, Thames, Housatonic, Quinnipiac, and Connecticut Rivers). Elevated concentrations of Ag, accompanied by high concentrations of inorganic nutrients (NO3 and PO4), were used to identify areas affected by sewage inputs (10, 11), while DOC and Fe were used to identify different biogeochemical carrier phases. Distributions of H4SiO4 and Ni were used as indicators of diagenetic remobilization from contaminated sediments (12, 13). A firstorder mass balance budget and principal component analysis were used to identify the major factors influencing the biogeochemical cycling of trace metals within LIS. Description of the Area of Study. Long Island Sound is roughly 93 km long and 34 km at its widest point extending from Hell Gate in the East River, adjacent to New York City, to The Race, which is situated between Plum and Fishers Islands (Figure 1). LIS is the sixth largest estuary in the United States (surface area of 3284 km2) and third largest in terms of volume (6.2 × 1010 m3). The Sound has a mean depth of 20 m and a maximum depth of 90 m (3, 14). The land drainage area of LIS exceeds 4.5 × 104 km2 and extends as far north as Canada. Long Island Sound is not a classic estuary, as it lacks a freshwater source at its head. Instead, the western headwaters of LIS mix with the lower salinity waters of New York Harbor through the East River, a 26 km long tidal straight (15). Freshwater flow enters into the sound from runoff and drainage along the coast of Long Island, New York, and Connecticut. The discharge of four major tributaries (Thames, Housatonic, Quinnipiac, and Connecticut Rivers) comprises most of the freshwater input to LIS, with the Connecticut River, which empties into the central basin of the LIS, being responsible for ∼70% of the total freshwater inflow. Three gyres reside within the waters of LIS: a clockwise gyre in the Central Basin and counterclockwise gyres in the Western and Eastern Basins. As a result, freshwater leaving 10.1021/es048860t CCC: $30.25

 2005 American Chemical Society Published on Web 04/08/2005

FIGURE 1. Map of Long Island Sound including sampling stations for this study. The dark lines denote the boundaries for each geographic region. From west to east, the East River, Narrows, and Western, Central, and Eastern Basins. the Connecticut River is advected westward and mixes within waters of the Central Basin (16). The exchange of deeper oceanic water (1.68 × 1012 ( 1.68 × 1011 L d-1 (17)) occurs on the eastern margin of LIS, where more saline ocean water exhibits a net westerly flow, while fresher surface waters generally moves eastward toward the Atlantic Ocean through The Race at an estimated flow of 7.9 × 1011 ( 1.37 × 1011 L d-1 (17, 18). The Sound also receives significant amounts of treated sewage from wastewater treatment in New York and southern Connecticut, mostly discharged within the East River-western LIS region (9). These arrays of geographic and oceanographic features make the hydrodynamics of LIS both complex and unique.

Materials and Methods Sampling cruises were conducted aboard the R/V Paumanok of Southampton College during July 2000 and April 2001. The average freshwater flux into LIS from the rivers in Connecticut between the years 1990 and 2000 was ∼5 × 1010 L d-1 (19). The July 2000 cruise was characterized by low river flow conditions with the combined flux of the rivers in Connecticut being measured at 2.0 × 1010 L d-1, which is 40% of the decadal average. In contrast, the April 2001 cruise was characterized by high discharge conditions with the flux of the rivers in Connecticut (2.4 × 1011 L d-1) being 485% higher than the decadal average. Hence, samples were collected during seasonal extremes to reflect as wide a range of conditions as possible on LIS water column constituents. A transect of 20 east/west stations beginning at the confluence of the East and Hudson Rivers at The Battery and progressing eastward to The Race in eastern LIS was sampled (Figure 1). Since a steep gradient in water quality exists in the western extent of LIS (9), sampling stations were closer together in this region. Additional samples were collected from five coastal embayments (Manhassett Bay, Huntington Harbor, Oyster Bay, Hempstead Harbor, and Port Jefferson Harbor) monitored by the U.S. National Status and Trends Program. Those results were used to determine whether metal concentrations measured throughout LIS were consistent with those measured in other coastal environments within the same watershed. To establish the impact of fluvial input on chemical constituents within the water column, north/ south transects in LIS’s largest rivers (Thames, Connecticut, Quinnapiac, and Housatonic Rivers) were sampled from the mouth of each river to the freshwater end-member. The

importance of each tributary as a source of trace metals, organic carbon, and nutrients was established via mass balance estimates (see below). Seawater samples were collected at 43 stations using a peristaltic pump equipped with trace-metal-clean Teflon tubing attached to a 7-m trace-metal-clean boom (12). The boom was submerged to a depth of 1 m and directed into approaching currents and prevailing winds. Dissolved samples were obtained by filtration through trace-metal-clean polypropylene capsule filters (0.2 µm) and analyzed for dissolved trace metals (Ag, Cd, Cu, Fe, Ni, Pb, Zn), inorganic nutrients (nitrate ) NO3, ammonium ) NH4, silicate ) H4SiO4, orthophosphate ) PO4), and DOC. The relative standard deviation of duplicate analyses was never greater than (15% and was characteristically around (5%. Triplicate chlorophyll a samples were collected on station using GF/F glass fiber filters (nominal pore size ) 0.7 µm) and stored frozen. At each sampling station a Hydrolab Quanta CTD was used to vertically profile salinity and temperature. All chemical protocols used in this study, as well as an explanation of the principal component analysis and mass balance calculations are presented in Appendix 1 (see the Supporting Information). Station locations, nutrient and DOC concentrations, salinity, temperature, and chlorophyll a levels for low and high flow conditions as well as procedural blank concentrations and instrumental detection limits are listed in Appendix 2 (see the Supporting Information). Mean values of duplicate measurements of dissolved trace metals from filtered seawater collected in July 2000 and April 2001 are also listed in Appendix 2.

Results and Discussion Spatial Gradients of Metal, Nutrient, and DOC Concentrations within LIS. Spatial differences in metals, nutrients, and DOC within the area of study were determined by comparing the distributions and mean concentrations measured within the East River, Narrows, Western Basin, Central Basin, and Eastern Basin of LIS (Figure 1). The spatial distributions of the dissolved constituents measured in the study were broadly categorized into two groups on the basis of their spatial distributions: (1) dissolved constituents with west to east gradients (e.g., Ag, Fe, Pb, Cu, Ni, Zn, PO4, and NO3) and (2) dissolved constituents with uniform distributions throughout LIS (e.g., Cd and DOC). VOL. 39, NO. 10, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3529

FIGURE 2. Geographical distributions of Ag, Pb, NO3, and PO4 along LIS. Kilometers indicate the distance east of The Battery (point 0), in the East River, New York City, to The Race in the Eastern Basin. Open circles represent samples taken during low-flow conditions, while filled circles represent samples taken during high-flow conditions. Open and filled triangles denote samples taken from the freshwater end members of the four rivers in Connecticut during low and high flow, respectively. Concentrations of Ag, Pb, PO4, and NO3 showed strong west to east gradients with all four constituents showing at least an order of magnitude decrease from the East River to the Eastern Basin during both low and high flow conditions (Figure 2). For example, concentrations of dissolved Ag (a tracer of sewage (10)) during low flow conditions decreased from 248 ( 88.2 pM (Mean ( SD) within the East River to 38.5 ( 14.1 pM within the Eastern Basin. Dissolved Pb, NO3, and PO4 also showed strikingly similar west to east gradients during both sampling periods. Concentrations of Pb during low flow conditions were 506 ( 170 pM within the East River 3530

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 10, 2005

and below the detection limits within the Central and Eastern Basins of LIS, while mean concentrations of PO4 and NO3 concurrently decreased from 3.32 ( 1.10 and 14.0 ( 7.45 µM within the East River to 0.64 ( 0.11 and 0.96 ( 0.93 µM within the Eastern Basin. Distributions of dissolved Cu, Fe, Ni, and Zn also declined from west to east (Figure 3) but to a lesser degree than Ag, Pb, PO4, and NO3. For example, during low flow conditions, average dissolved Cu and Ni decreased by about half from the East River to the Eastern Basin (Cu, from 28.6 ( 4.33 to 11.3 ( 2.12 nM; Ni, from 39.6 ( 11.5 to 18.9 ( 4.73 nM).

FIGURE 3. Geographical distributions of Cd, Cu, Fe, Ni, Zn, and DOC along LIS. Kilometers indicate the distance east of The Battery (point 0), in the East River, New York City, to The Race in the Eastern Basin. Open circles represent samples taken during low-flow conditions, while filled circles represent samples taken during high-flow conditions. Open and filled triangles denote samples taken from the freshwater end members of the four rivers in Connecticut during low and high flow, respectively. Due to the high levels of Fe in the freshwater end member, those values were excluded from the figure. Dissolved Zn concentrations were 46.5 ( 2.56 nM in the East River and 13.9 ( 13.9 nM in the Eastern Basin. The concentrations of dissolved Fe decreased from the East River (52.7 ( 21.5 nM during low flow and 46.2 ( 13.9 nM during high flow) to the Western Basin (14.0 ( 11.0 nM during low flow and 26.3 ( 16.2 nM during high flow). However, dissolved Fe concentrations remained relatively constant from the Central to the Eastern Basin. The relatively high concentrations of Fe detected in some areas of the Sound represent areas influenced by riverine inputs, as concentrations of this element ranged from 313 nm to 2.6 µM in the rivers of Connecticut (Supporting Information, Appendix 2). In contrast to constituents that displayed a west to east gradient, concentrations of Cd and DOC were uniformly

distributed throughout LIS (Figure 3). For example, during the summer sampling period, the highest average concentration of Cd (0.23 ( 0.05 nM) was detected in the East River and the lowest concentration (0.18 ( 0.07 nM) was found within the Narrows region in western LIS, while the Eastern Basin contained intermediate concentrations (0.21 ( 0.074 nM). Similarly, the highest concentrations of DOC (191 ( 6.67 µM) during the summer sampling period was detected within the Narrows while the lowest concentrations (148 ( 42.8 µM) were found in the Eastern Basin. Anthropogenic Perturbations of Trace Metal and Nutrient Distributions in LIS. Comparisons of metal and nutrient concentrations measured in LIS with those reported for other estuaries with different levels of urbanization are useful in VOL. 39, NO. 10, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3531

FIGURE 4. Box plots of surface concentrations of Ag, PO4, Cu, Ni, and Pb in Long Island Sound (ER, East River; Narrows; WB, Western Basin; CB, Central Basin; EB, Eastern Basin; NST, National Status and Trends Program sites sampled) as compared to levels measured in the Peconic Estuary System (PB), Great South Bay (GSB), southern San Francisco Bay (SFB), and the Hudson River (HR). Empty boxes represent low-flow conditions, while dashed boxes represent high-flow conditions. assessing the impact of sewage and fluvial inputs on chemical constituents within the area of study. Box plots of metal and nutrient concentrations for the different regions of LIS and the five National Status and Trends Program sites sampled in this study were compared with concentrations reported for the relatively pristine estuaries Peconic Bay (PB) (20) and Great South Bay (GSB) (21) and more urbanized systems San Francisco Bay (SFB) (12) and the Hudson River (HR) (22). Our comparison showed that concentrations of Ag within the East River are the highest values ever reported in the 3532

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 10, 2005

United States (Figure 4). Concentrations of Ag found in the Narrows were slightly lower than the East River concentrations but were similar to those reported for other urban estuaries, such as SFB and the HR. As with Ag, concentrations of dissolved PO4 (Figure 4) and NO3 (Supporting Information, Appendix 2) within the western extent of LIS were comparable to values reported for estuaries whose water quality is substantially affected by anthropogenic activities. For example, PO4 and NO3 concentrations in the East River and the Narrows were both comparable to the concentrations

TABLE 1. Relative Contributions of the Thames, Quinnipiac, Housatonic, Connecticut, and East Rivers to the Total Fluvial Input of Metals, Inorganic Nutrients, and DOC into the LIS during Low and High Flowa NO3

PO4

H4SiO4

DOC

Thames Quinnipiac Housatonic Connecticut East River

1.5 0.25 5.88 22.6 70.1

0.36 0.08 1.91 3.54 94.2

1.46 0.18 4.86 46.9 46.9

1.62 0.27 8.66 23.6 65.8

Thames Quinnipiac Housatonic Connecticut East River

1.19 1.48 16.9 61.6 18.8

0.96 0.28 15.8 35.9 47.1

1.51 0.29 2.62 76.9 18.7

1.64 1.72 3.11 69.8 23.7

Ag

Cd

Cu

Ni

Zn

Fe

Low Flow 0.14 0.01 0.26 0.51 99.0

2.37 2.38 9.33 20.2 65.7

0.86 0.34 4.7 9.63 84.4

0.84 0.22 11.3 11.2 76.3

0.89 0.09 7.27 6.9 84.8

8.95 0.63 22.2 51.2 17.0

High Flow 0.69 1.36 2.17 9.67 86.1

0.84 1.06 8.82 36.8 52.5

1.26 1.57 7.2 68.5 21.5

1.15 5.56 6.32 55.9 31.1

1.34 0.86 3.79 62.8 31.2

12.3 0.59 13.4 69.7 4.05

a The relative contributions of each river were calculated as percentages according to Q riverCriver/(QThamesCThames + QQuinnipiacCQuinnipiac + QHousatonicCHousatonic + QConnecticutCConnecticut + QERCER).

reported in SFB but markedly higher than the concentrations reported in less urbanized estuaries such as PB and GSB. The high concentrations of Ag, PO4, and NO3 measured in western LIS, coupled with significant correlations between NO3 (r2 ) 0.95, p < 0.05) and PO4 (r2 ) 0.79, p < 0.05) with dissolved Ag, indicate that sewage, which is discharged in high volumes in this region, is the likely source of these constituents (Supporting Information, Appendix 2). Although significant improvements have been made in sewage treatment and sewage input loads have decreased in the New York City watershed over the past several decades (23), the negative impact of these discharges on the water quality of this area is not surprising. For example, O’Shea and Brosnan (4) reported that six water pollution control plants discharge effluent into the East River at a rate of 36.8 m3 s-1. The daily volume of sewage discharged into the East River (3.18 × 109 L d-1) is higher than the volume of freshwater entering eastern LIS through the Thames River during low flow conditions (1.25 × 109 L d-1). In contrast to the western extent of LIS, concentrations of anthropogenic tracers such as Ag within the eastern region were comparable to concentrations measured in estuaries with less urbanization (i.e. PB and GSB, Figure 4). For example, concentrations of Ag within the Central and Eastern Basins were markedly lower than those reported for SFB and HR and similar to the values reported for PB and GSB. Dissolved PO4 and NO3 concentrations were also 1 order of magnitude lower in eastern LIS with respect to SFB but similar to concentrations found in GSB. Dissolved Cu and Ni (as well as Zn; not shown) concentrations within LIS were comparable to the concentrations reported for both the highly urbanized estuaries (SFB and HR) and for less urbanized estuaries, such as PB and GSB (Figure 4). The elevated concentrations of Cu and Ni within the Western and Central Basins indicate the existence of a source for these trace elements within LIS besides sewage. Ambient concentrations of Cu and Ni measured in the National Status and Trends Program sites also showed elevated concentrations when compared to other estuaries. This metal enrichment was more evident for Ni, which displayed markedly higher concentrations than that seen for both SFB and HR. The high concentrations of Cu in those embayments could be reflective of the abundance of recreational marinas found in those areas and with the restricted water exchange with LIS (14). Dissolved Pb concentrations within LIS were only detectable within the East River and Narrows regions of LIS (Figure 4), areas strongly influenced by anthropogenic inputs. Furthermore, Pb was found to be markedly higher during low flow conditions when compared to high flow conditions.

The high concentrations of dissolved Pb observed within the western extent of LIS during low flow conditions suggest that this metal is most likely derived from sewage effluents or benthic remobilization from contaminated sediments rather than runoff from the surrounding watershed (22). The high levels of trace metals we report for western LIS are also consistent with metal concentrations found in other highly urbanized and industrialized coastal environments, such as Galveston Bay (24), San Diego Bay (25), South San Francisco Bay (12), and the Scheldt estuary in Europe (26) (Supporting Information, Appendix 2). Apportionment of Dissolved Constituents Sources to LIS. Mass balance estimates were used to establish the firstorder metal and nutrient budget of the LIS. This relatively simple mass balance model provides an insight into the relative contributions of fluvial, sewage, internal, and oceanic sources of dissolved constituents to LIS. A description of the mass balance calculations is reported in Appendix 1 (see the Supporting Information). The mass balance results showed substantial temporal changes in the allochthonous sources of dissolved inorganic nutrients and metals into LIS (Table 1). During low fluvial flow, the Connecticut River flux (1.94 × 1010 ( 3.90 × 109 L d-1) into the LIS was slightly less than the subtidal exchange through the East River (9.50 × 1010 ( 8.60 × 109 L d-1). However, because of the markedly higher concentrations of metals and inorganic nutrients in the East River (with the exception of Fe; Appendix 2), this river is the most important source of NO3 (70%), PO4 (94%), and dissolved Ag (99%), Cu (85%), Ni (76%), and Zn (85%) into the LIS. In contrast, because concentrations of Fe were 1 order of magnitude higher within the Connecticut River when compared to the East River, the Connecticut River was responsible for the majority (51%) of the Fe input during the same sampling period. During high flow conditions, the main fluvial source of metals and nutrients to LIS appears to shift from the East River to the Connecticut River, with the exception of Ag, which is still mostly introduced into LIS by the East River (86%). Compared to the East River, the Connecticut River supplies LIS with a much greater volume of water (1 order of magnitude greater) during high discharge conditions (2.41 × 1011 L d-1 vs 9.50 × 1010 L d-1). The major fluvial input of dissolved constituents (Cu, 69%; Ni, 56%; Zn, 63%; Fe, 70%; NO3, 62%; and H4SiO4, 77%) to the LIS during the spring runoff is the Connecticut River. Our results also demonstrated that the different sources considered in the mass balance estimates (e.g., fluvial and oceanic supply) could not account for the total export of Cu, Ni, and Zn to the Atlantic Ocean from the LIS (Table 2). For VOL. 39, NO. 10, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3533

TABLE 2. Mass Balance Calculation Results for Long Island Sounda,b Friver

error (()

Focean

error (()

FLUXin

NO3 PO4 H4SiO4 Ag Cd Cu Ni Zn Fe

1.90 × 106 3.35 × 105 1.25 × 106 23.8 339 3220 4680 5210 2.93 × 104

1.04 × 105 9.59 × 104 3.16 × 105 8.8 48.7 80 1548 853 1.51 × 103

5.09 × 105 3.39 × 105 4.66 × 106 13.0 250.16 4876 6106 4113 8.17 × 104

3.18 × 104 9.59 × 104 1.32 × 105 2.8 68 982 1262 831 1.66 × 103

2.41 × 106 6.74 × 105 5.92 × 106 36.8 598 8.09 × 103 1.08 × 104 9318 1.11 × 105

NO3 PO4 H4SiO4 Ag Cd Cu Ni Zn Fe

5.86 × 106 3.44 × 105 3.83 × 107 5.8 43.90 11559 10524 13126 1.09 × 105

2.78 × 106 6.82 × 104 2.80 × 106 8.7 4.87 80 1774 1147 1.11 × 104

6.71 × 105 3.39 × 105 4.66 × 106 20.3 250.16 4876 43700 45800 1.34 × 104

3.18 × 105 6.82 × 104 9.37 × 105 1.8 68 147 339 144 5.14 × 102

6.53 × 106 6.83 × 105 4.30 × 107 26.0 2.94 × 102 1.64 × 104 5.42 × 104 58926 1.22 × 105

error (()

FLUXout

error (()

imbalance (X)

error (()

1.35 × 105 1.92 × 104 1.64 × 105 11.6 117 1.06 × 102 2.81 × 103 1684 3.17 × 104

7.73 × 105 8.58 × 105 7.00 × 106 61.7 311.76 3.2 × 104 4.67 × 104 3.86 × 104 1.92 × 104

1.55 × 104 1.75 × 104 1.43 × 105 5.9 13.71 1.4 × 103 3.03 × 103 1.86 × 103 9.74 × 102

-1.64 × 106 1.84 × 105 (21%) 1.08 × 106 (15%) 24.9 (40%) -277 2.34 × 104 (74%) 3.59 × 104 (77%) 2.93 × 104 (76%) -9.19 × 104

1.20 × 105 1.70 × 104 2.09 × 105 5.7 103 3.55 × 102 2.16 × 102 1.80 × 102 3.07 × 103

3.12 × 106 5.48 × 105 4.89 × 106 67.6 367.57 3.5 × 104 5.74 × 104 4.00 × 104 2.04 × 104

8.83 × 105 1.55 × 105 1.38 × 106 14.8 73.92 7.1 × 103 1.19 × 104 8.09 × 103 4.13 × 103

-3.40 × 106 -1.35 × 105 -3.81 × 107 41.6 (61%) 74 (20%) 1.88 × 104 (53%) 3.20 × 103 (6%) -1.89 × 104 -1.02 × 105

2.21 × 106 1.86 × 104 2.35 × 106 4.25 1.05 6.86 × 103 9.75 × 103 6.80 × 103 7.45 × 103

Low Flow

High Flow

3.10 × 106 1.36 × 105 3.73 × 106 10.5 7.29 × 101 2.27 × 102 2.11 × 103 1291 1.16 × 104

a Fluxes are reported in mole/day. The uncertainties reported in the flux calculations are the result of an error propagation analysis (Appendix 1, Supporting Information). The imbalance column shows the necessary input needed to balance the total output. Negative numbers indicate that those constituents are being removed within the LIS. Data in parentheses represent the percent of the imbalance relative to the total output.b The spring cruise was characterized by high discharge conditions (Quinnipiac, 4.90 × 109 L d-1; Housatonic, 1.23 × 1010 L d-1; Connecticut, 2.41 × 1011 L d-1; Thames 5.31 × 109 L d-1). The summer cruise was characterized by low river flow conditions (Quinnipiac, 2.32 × 108 L d-1, Housatonic, 5.01 × 109 L d-1, Connecticut, 1.94 × 1010 L d-1, Thames, 1.25 × 109 L d-1). The subtidal water exchange in the East River used in the calculations was 9.50 × 1010 L d-1. Friver ) river input; Focean ) oceanic input; FLUXin ) total flux in; FLUXout ) total flux out; imbalance ) FLUXout - FLUXin.

example, additional inputs of Cu (73%), Ni (77%), and Zn (78%) are needed in order to account for the export of these metals during low flow. These imbalances were larger than budget uncertainties and suggest that processes other than fluvial, sewage, and oceanic inputs control the concentrations of Cu, Ni, and Zn within LIS. The high concentrations of metals found within the LIS sediments could represent a significant source of contaminants to the Sound. During low flow, significant correlations were found for Cu and H4SiO4 (r2 ) 0.61, p < 0.05) and Zn and H4SiO4 (r2 ) 0.729, p < 0.05), suggesting that Cu, Zn, and H4SiO4 may have a similar benthic source (Supportive Information, Appendix 2). Aller and Benninger (13) reported fluxes of H4SiO4 from sediments within LIS that increased with temperature and thus had the highest flux rates in the summer. Diagenetic remobilization of H4SiO4 has also been reported in San Francisco Bay (12). While particulate metal concentrations were not measured for this study, their influence in the cycling of dissolved metals cannot be ignored. Sediment-bound metals represent a large reservoir that may reenter the water column in the dissolved form via a number of processes, including diffusion and sediment resuspension (27). A substantial sedimentary flux into LIS via diagenetic remobilization is consistent with the elevated concentrations observed for Cu, Ni, and Zn throughout the system and away from point sources located within the East River/Narrows region (Figure 3). Atmospheric deposition is another possible source of metals into LIS not accounted for in our mass balance calculations. However, Cochran et al. (28) estimated atmospheric fallout of Cu and Zn from six salt marsh cores located in New York City and the northern shore of Long Island and reported that nonatmospheric sources of these metals were of greater importance throughout the LIS estuarine system. Consistent with those results, Rozan and Benoit (29) quantified atmospheric inputs of Ag, Cd, and Cu near the Quinnipiac River using sediment cores from three local, pristine lakes and reported that atmospheric deposition was of minor importance. While atmospheric deposition may explain some of the imbalance found in the mass balance estimates, it is likely to represent only a small fraction for these trace metals. 3534

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 10, 2005

The mass balance results also indicated that, during spring conditions, NO3 (212%) and PO4 (131%) inputs greatly exceeded export, which suggests that inorganic nutrients were retained within LIS. This imbalance may reflect biological uptake of nutrients during the spring phytoplankton bloom, and this process could represent an additional sink for these constituents (30-32). In contrast, during summer, when concentrations of algal biomass (chlorophyll a) were lower, the inorganic nutrient imbalance was within the uncertainty of our estimates. Although our budget does lend insight into the major processes influencing metal and nutrient distributions within the LIS, our analysis should be considered a first-order evaluation, as major uncertainties have yet to be resolved. For example, the scarcity of metal data from rivers in Connecticut did not allow us to establish the impact of freshwater end-member variability on the riverine flux. Furthermore, particulate-bound metal analysis was not undertaken in this study, thus the contribution and cycling (e.g., desorption) of these metal forms to the budget of LIS was not directly quantified. Although our samples were collected during dry and wet seasons, high-energy events (e.g. storms) may strongly influence metal transport and our budget. The influence of watershed hydrology and land usage on metal and nutrient loadings to the sound was not established. Last, the transfer of metal and nutrients from the highly contaminated western basin to other areas of the LIS was not evaluated. Future studies should address all of these questions as well as a quantitative estimate of benthic remobilization from contaminated sediments. Biogeochemical and Anthropogenic Controls of Dissolved Constituents in the LIS. Principal component analysis (PCA) was used to identify the processes influencing the cycling of metals in LIS. This allowed for the formation of correlating variables that represent linear composites of the data without substantial loss of information. PCA explained 73.4% and 79.0% of the variance during the summer and spring sampling periods, respectively. The PCA also suggested that the sources, transports, and processes affecting the biogeochemical properties of LIS were very different between the two sampling periods. Results of PCA, including factor

FIGURE 5. Principal component model projections for the first and second principal components during the summer sampling period and the first, second, and third components during the spring sampling period (East River, ER; Narrows, Nar; Western Basin, WB; Central Basin, CB; Eastern Basin, EB). loadings of the extracted components and factor scores are reported in Appendix 2 (see the Supplementary Information). During the high river flow, the first component had high loadings of NO3 (0.774), PO4 (0.810), H4SiO4 (0.883, and Fe (0.712) that were inversely related to chlorophyll a (-0.788). This component most likely represents the uptake and scavenging of nutrients and Fe by phytoplankton (chlorophyll a) and accounted for 33.8% of the total variance. The relatively high contribution for factor 1 indicates that primary production is an important process influencing the distributions of inorganic nutrients and Fe during high discharge conditions. This is consistent with past observations that found that, during spring conditions, there is an increase in nutrient inputs from runoff, creating optimal conditions for spring phytoplankton blooms within LIS (32, 33). The populations of these blooms are dominated by diatoms, species that can exhaust the surface waters of nutrients within LIS in a matter of weeks (32, 33). This intense primary productivity is likely the process represented by the first factor extracted by PCA.

The second high-flow component is identified by the association of Cu (0.887), Ni (0.810), and temperature (0.769), and it is attributed to benthic remobilization. Increased temperature is known to exert a significant influence on the magnitude of benthic fluxes (13, 34). Thus, the second component reflects the importance of internal processes, such as remobilization from contaminated sediments. The third high-flow component is characterized by high loading of dissolved Ag (0.818) and Cd (0.801). PCA is a hierarchical analysis and this principal component associated with sewage has less weight than the sediment remobilization factor, suggesting that contaminated sediments could be the largest contributor of contaminants to the whole LIS as opposed to localized point sources. During the low-flow sampling period, principal component 1 was characterized by high loading of DOC (0.838) and Fe (0.883), elements that represent different fluvial geochemical carrier phases, which were inversely related to salinity (-0.919; mixing) and Cd (-0.723). The high loadings of Fe VOL. 39, NO. 10, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3535

and DOC in component 1 may reflect the importance of colloidal organic material influencing the cycling of those constituents (35). Principal component 2 was characterized by high loading of Cu (0.872), Ni (0.812), and Zn (0.840). As previously discussed, the concentrations of these three metals within LIS are believed to originate from benthic sources when water temperatures are elevated. Factor scores were plotted for each sampling period to identify chemically distinct areas within the LIS (Figure 5 and Appendix 2). Distributions of factor scores in relation to one another can indicate how the average water quality varies over the study area (36). During high flow, the sample separations shown by the two components reflect spatial changes in influences of primary productivity and benthic flux and resemble patterns reported in the past for LIS (13, 37-39). Specifically, the East River was found to have high nutrient stocks and low chlorophyll a, while the eastern section of the sound displayed a reverse trend. Principal component 2 denotes increased influence of benthic remobilization in the Narrows and western end of the sound, which agrees with observations of substantial benthic flux in this region despite lower water temperatures (39). The clustering of factor scores for principal component 1 with principal component 3 (Ag and Cd) also reflects the distinct geographic regions of LIS. Stations located within the East River and the Narrows represent principal component 3, suggesting that this region is distinct from the rest of the LIS. During low flow, the clustering of score plots for the first two components resemble the geographic areas of LIS. Principal component 1 showed stark separation of the freshwater end member stations from the stations within LIS (Figure 5). However, principal component 1 makes little contribution to the geographical differences observed within the Sound and suggests spatial homogeneity during low discharge conditions with respect to Fe, DOC, Cd, and salinity. Factor score plots for principal component 2 are consistent with observations of a decreasing west to east gradient of benthic flux within LIS as a result of increasing water temperatures during summer conditions (39). Factor score plots of principal components suggest that the biogeochemical cycling as well as the impact of anthropogenic inputs within LIS are relatively distinct within each geographic region of the Sound. Our metal/nutrient budget provides an initial estimate of the sources and processes controlling the cycling of dissolved constituents within the LIS as a whole. However, future studies should establish the transfer of dissolved constituents among the different regional basins as well as the main mechanisms influencing metal and nutrient levels in each basin. In summary, this study evidences the spatial and temporal differences of trace metal and nutrient sources to the LIS ecosystem. The East River contains the highest concentrations of some dissolved trace metals in the United States and is the primary source of many dissolved constituents to LIS during low freshwater flow conditions. In contrast, the rivers in Connecticut supply LIS with the majority of dissolved metals and nutrients during maximum, spring river flow conditions. While remobilization from contaminated sediments may be an important internal source of some dissolved constituents to LIS, phytoplankton may represent a substantial sink for bioactive substances, particularly during the spring bloom. Although substantial progress has been made reducing sewage-derived nitrogen loads to LIS over the past decade, our results indicate that anthropogenic loadings continue to strongly impact the concentrations of dissolved trace metals in some regions of the LIS.

Acknowledgments We acknowledge the EPA for funding. We thank Don Getz of the R/V Paumanok for captaining cruises and Richard 3536

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 10, 2005

MacIntyre for logistical support. We thank M. P. LaHaye Renaud, F. Koch, G. Boneillo, M. Renaghan, A. Kolker, B. Kentrup, J. Reimer, R. Wiggum, and A. Tovar-Sanchez for laboratory and cruise assistance. The U.S. Environmental Protection Agency (X982277-01) and NY Sea Grant supported this research.

Supporting Information Available Appendices 1 and 2. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Driscoll, C. T.; Whitall, D.; Aber, J.; Boyer, E.; Castro, M.; Cronan, C.; Goodale, C. L.; Groffman, P.; Hopkinson, C.; Lambert, K.; Lawrence, G.; Ollinger, S. Nitrogen pollution in the northeastern United States: Sources, effects, and management options. Bioscience 2003, 53, 357-374. (2) de Jonge, V. N.; Elliott, M.; Orive, E. Causes, historical development, effects and future challenges of a common environmental problem: Eutrophication. Hydrobiologia 2002, 475 (1), 1-19. (3) Wolfe, D. A.; Monahan R.; Stacey, P. E. Environmental quality of Long Island SoundsAssessment and management issues. Estuaries 1992, 14 (3), 224-236. (4) O’Shea, M. L.; Brosnan, T. M. Trends in indicators of eutrophication in western Long Island Sound and the Hudson-Raritan Estuary. Estuaries 2000, 6, 877-901. (5) Robertson, A.; Gottholm, B. W.; Turgeon, D. D.; Wolfe, D. A. A comparative study of contaminated levels in Long Island Sound. Estuaries 1991, 14, 290-298. (6) Turgeon, D. D.; O’Conner, T. P. Long Island Sound: Distributions, trends and effects of chemical contamination. Estuaries 1991, 14, 279-289. (7) Rolfhus, K. R.; Fitzgerald, W. F. The evasion and spatial/temporal distributions of mercury species in Long Island Sound, CT-NY. Geochim. Cosmochim. Acta 2001, 65(3), 407-418. (8) Tseng, C. M.; Balcom, P. H.; Lamborg, C. H.; Fitzgerald, W. F. Dissolved elemental mercury investigations in Long Island Sound using on-line amalgamation-flow injection analysis. Environ. Sci. Technol. 2003, 37, 1183-1188. (9) Sweeny A.; San ˜ udo-Wilhelmy S. A. Dissolved metal contamination in the East River-Long Island sound system: Potential biological effects. Mar. Pollut. Bull. 2004, 48 (7-8), 663-670. (10) San ˜ udo-Wilhemy, S. A.; Flegal, A. R. Anthropogenic silver in the Southern California Bight: A new tracer of sewage in coastal waters. Environ. Sci. Technol. 1992, 26, 2147-2151. (11) Ravizza, G. E.; Bothner, M. H. Ossium isotopes and silver as tracers of anthropogenic metals in sediments from Massachusetts and Cape Cod bays. Geochim. Cosmochim. Acta 1996, 60 (15), 2753-2763. (12) Flegal, A. R.; Smith, G. J.; Gill, G. A.; S. A. San ˜ udo-Wilhelmy. Dissolved trace element cycles in the San Francisco Bay Estuary. Mar. Chem. 1991, 36, 329-363. (13) Aller, R. C.; Benniger, L. K. Spatial and temporal patterns of dissolved ammonium, manganese, and silica fluxes from the bottom sediments of Long Island Sound, USA. J. Mar. Res. 1981, 39, 295-314. (14) EPA. The Long Island Sound Study: The Comprehensive Conservation management Plan; EPA: Washington, DC, 1994; pp 168. (15) Jay, D. A.; Bowman, M. J. The Physical Oceanography and water quality of New York Harbor and Western Long Island Sound; Technical report 23, Marine Sciences Research Center, State University of New YorksStony Brook: Stony Brook, New York, 1975. (16) Wilson, R. E. Gravitational circulation in Long Island Sound. New Estuar. Coast. Mar. Sci. 1976, 4, 443-453. (17) Riley, G. A. Aspects of oceanography of Long Island Sound. II. Transport and mixing processes in Long Island Sound. Bull. Bingham Oceanogr. Collect. 1967, 19, 35. (18) Valle-Levinson, A.; Wilson, R. E. Rotation and vertical mixing effects on volume exchange in eastern Long Island Sound. Estuar. Coast. Shelf Sci. 1998, 46 (4), 573-585. (19) www.usgs.gov. (20) Breuer, E.; San ˜ udo-Wilhelmy, S. A.; Aller, R. C. Trace metals and dissolved organic carbon in an estuary with restricted river flow and a brown tide bloom. Estuaries 1999, 22 (3a), 603-615. (21) Clark, L. B. MS thesis, State University of New YorksStony Brook, Stony Brook, NY, 2000.

(22) San ˜ udo-Wilhemy, S. A.; Gill, G. A. Impact of the Clean Water Act on the levels of toxic metals in urban estuaries: The Hudson River Estuary revisited. Environ. Sci. Technol. 1999, 33, 34773481. (23) O’Shea, M. L.; Brosnan, T. M. New Harbor York Water Quality Survey 1995; NTIS No PB97-100192; New York City Department of Environmental Protection, Marine Science Section: Wards Island, New York, 1997. (24) Wen, L. S.; Paternostro, C.; Santschi, P. H. Gill, G. Estuarine trace metal distributions in Galveston Bay: Importance of colloidal forms in speciation of the dissolved phase. Mar. Chem. 1999, 63, 185. (25) Flegal, A. R.; San ˜ udo-Wilhelmy, S. A. Comparable levels of trace metal contamination in two semienclosed embayments; San Diego Bay and South San Francisco Bay. Environ. Sci. Technol. 1993, 27 (9), 1934-1936. (26) Baeyens, W.; Goeyens, L.; Ducastel, G.; De Gieter, M.; Leermakers, M. The biogeochemical behavior of Cd, Cu, Pb and Zn in the Scheldt estuary: Results of the 1995 surveys. Hydrobiologia 1998, 366, 45-62. (27) Aller, R. C. The importance of relict burrow structures and burrow irrigation in controlling, sedimentary solute distributions. Geochim. Cosmochim. Acta 1984, 48 (10), 1929-1934. (28) Cochran, J. K.; Hirschberg, D. J.; Wang, J.; Dere, C. Atmospheric deposition of metals to coastal waters (Long Island Sound, New York USA): Evidence from saltmarsh deposits. Estuar. Coast. Shelf Sci. 1998, 46 (4), 503-522. (29) Rozan, T. F.; Benoit, G. Mass balance of heavy metals in New Haven Harbor, Connecticut: Predominance of nonpoint sources. Limnol. Oceanogr. 2001, 46 (8), 2032-2049. (30) Sieracki, M. E.; Verity, P. G.; Stoecker, D. K. Plankton community response to sequential silicate and nitrate depletion during the 1989 North Atlantic spring bloom. Deep-Sea Research Part IITopical Studies in Oceanography 1993, 40, 213.

(31) Townsend, D. W.; Thomas, M. Springtime nutrient and phytoplankton dynamics on Georges Bank. Mar. Ecol. Prog. Ser. 2002, 228, 57-74. (32) Gobler, C. J.; San ˜ udo-Wilhelmy, S. A.; Buck, N. J.; Sieracki, M. E. Estuaries Submitted. (33) Riley, G. A.; Conover, S. A. M. Oceanography of Long Island Sound, 1952-1954. Bull. Bingham Oceanogr. Collect. 1956, 15, 62. (34) Riedel, G. F.; Sanders, J. G.; Osman, R. W. Biogeochemical control on the flux of trace elements from estuarine sediments: Water column oxygen concentrations and benthic infauna. Estuar. Coast. Shelf Sci. 1997, 44, 23-38. (35) Guo, L. D.; Santschi, P. H.; Warnken, K. W. Trace metal composition of colloidal organic material in marine environments. Mar. Chem. 2000, 70(4), 257-275. (36) Boyer, J. N.; Fourqurean, J. W.; Jones, R. D. Spatial characterization of water quality in Florida Bay and Whitewater Bay by multivariate analyses: Zones of similar influence. Estuaries 1997, 20 (4), 743-758. (37) Riley, G. A. Hydrography of the Long Island and Block Island Sounds. Bull. Bingham Oceanogr. Collect. 1952, 13, 5. (38) Hunt, C. D. Variability in the benthic Mn flux in coastal marine ecosystems resulting from temperature and primary production. Limnol. Oceangr. 1983, 28(5), 913-923. (39) Aller, R. C. The sedimentary Mn cycle in Long Island SoundsIts role as intermediate oxidant and the influence of bioturbation, O2, and C (org) flux on diagenetic reaction balances. J. Mar. Res. 1994, 52 (2), 259-295.

Received for review July 22, 2004. Revised manuscript received January 27, 2005. Accepted March 10, 2005. ES048860T

VOL. 39, NO. 10, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3537