Department of Earth & Environmental Sciences
Long Island Sound: foraminifera
in Long Island Sound:
foraminifera and geochemistry
PUBLICATIONS and ABSTRACTS
Benthic Foraminifera and Environmental Changes
in Long Island Sound
Thomas, E., Gapotchenko, T., Varekamp, J. C., Mecray, E. L., and Buchholtz ten Brink, M. R., 2000, Benthic foraminifera and environmental changes in Long Island Sound, Journal of Coastal Research, 16: 641-655.
1 Department of Earth and Environmental Sciences
265 Church Street
Middletown, CT 06459-0139
2Center for Coastal and Marine Geology
U.S. Geological Survey
384 Woods Hole Road
Woods Hole, MA 02543
March 14, 2000
THOMAS, E.; GAPOTCHENKO, T.; VAREKAMP, J. C.; MECRAY, E. L., and BUCHHOLTZ TEN BRINK, M. R., 2000. Benthic Foraminifera and Environmental Changes in Long Island Sound. Journal of Coastal Research, 16, 641&endash; 655. Fort Lauderdale (Florida) ISSN 0749-0208.
Benthic foraminiferal faunas in Long Island Sound (LIS) in the 1940s and 1960s were of low diversity, and dominated by species of the genus Elphidium, mainly Elphidium excavatum clavatum, with common Buccella frigida and Eggerella advena. The distribution of these species was dominantly correlated with depth, but it was not clear which depth-related environmental variable was most important. Differences between faunas collected in 1996 and 1997, and in the 1940s and 1960s include a strong decrease in relative abundance of Eggerella advena over all LIS, an increase in relative abundance of Ammonia beccarii in western LIS, and a decrease in species diversity.
The decreased diversity suggests that environmental stress caused the faunal changes. Oxygen isotope data for E. excavatum clavatum indicate that a change in salinity is not a probable cause. Carbon isotope data suggest that the supply of organic matter to the benthos increased since the early 1960s, with a stronger increase in western LIS where algal blooms have occurred since the early 1970s, possibly as a result of nutrient input by waste water treatment plants. These blooms or the resulting episodes of anoxia/hypoxia may have played a role in the increased relative abundance of A. beccarii. There is no clear explanation for the decreased abundance of E. advena, but changes in the phytoplankton composition (thus food supply) are a possible cause. Benthic foraminiferal faunal and stable isotope data have excellent potential as indicators of physicochemical environmental changes and their effects on the biota in LIS.
ADDITIONAL INDEX WORDS: Elphidium excavatum, Buccella frigida, Eggerella advena, pollution, anoxia, hypoxia, oxygen isotopes, carbon isotopes.
Long Island Sound (LIS) is a large, urban estuary with a reduced salinity and a pronounced seasonal stratification in its western part (e.g., RILEY, 1956, 1959; KOPPELMAN et al., 1976; GORDON, 1980). Its tidal range increases from about 0.7 m in the east to about 2.2 m in the west, and its circulation is dominated by tidal currents (KOPPELMAN et al., 1976). The temperature of LIS waters fluctuates between ~ 0 oC in the winter to more than 20 oC in the summer. Bottom-water salinity ranges from 32 in the east to 22 in the west. Surface-water salinity ranges from about 30 in the east to 20 in the west, with values down to 15 close to rivers (CONNECTICUT DEPARTMENT OF ENVIRONMENTAL PROTECTION, 1977). The Connecticut River contributes more than 70% of the fresh water influx, the Housatonic 12%, and the Thames 9%. Precipitation (110 cm/y.) is distributed evenly throughout the year, slightly exceeding the annual evaporation (93 cm/y.). The LIS surface water salinity is strongly linked to the river discharges, with a time lag of about 2 months (KOPPELMAN et al., 1976).
The Sound traps sediment efficiently, and areas of deposition and erosion are distributed patchily throughout LIS (GORDON, 1980; BOKUNOWICZ and GORDON, 1980). Sidescan studies documented an area of sediment erosion and non deposition east of the mouth of the Connecticut River, whereas environments of coarse-grained bedload transport (with sand-ribbons and sand waves) are present between the Connecticut River and Hammonasset. Deposition of fine-grained sediment occurs in the central and western basins (KNEBEL et al., 1999; KNEBEL and POPPE, this volume).
The "quiet water" regions in LIS, especially in the western basins, experience periodic, seasonal bottom water hypoxia (dissolved O2<3 ml/L) or anoxia (dissolved O2<1 ml/L), with severe consequences to the biota (e.g., EPA, 1994). Salinity differences between bottom and surface waters are usually less than 1 , while temperature differences reach up to 8 oC. The density stratification is thus primarily thermally controlled (WELSH and ELLER, 1991), and low-oxygen conditions usually reach their peak in July-August, when thermal stratification is most pronounced (NEW YORK STATE DEPARTMENT OF ENVIRONMENTAL CONSERVATION and CONNECTICUT DEPARTMENT OF ENVIRONMENTAL PROTECTION, 1999).
Historical data since the beginning of the 20th century show occasional anoxia close to New York City (East River) in the 1920s and mid 1950s, but pervasive anoxia in the western basin of LIS was first observed in 1971 (PARKER and O'REILLY, 1991), and was more severe in the 1970s than in the following decades. There is no evidence for hypoxia in the eastern basin, but moderate hypoxia occurred in the 1980s in the central basin, where waters were well oxygenated in the 1950s (RILEY, 1956, 1959; PARKER and O'REILLY, 1991).
When nutrients are abundant, algal blooms occur during the periods of stratification. The algal blooms are more intense in western LIS (SUN et al., 1994). The top 5-10 m of water is supersaturated in oxygen produced by photosynthesis (WELSH and ELLER, 1991). Lower in the water column, oxygen is consumed through the decay of the abundantly produced organic matter, and hypoxia occurs even though some dissolved oxygen is transported laterally into western LIS (TORGERSON et al., 1997). Small density differences create a stable stratification and the pycnocline always coincides with the oxycline, at a depth of 10-11 m (e.g., WELSH and ELLER, 1991). The complex interactive process of oxygen loss in the bottom waters, including the temporal and spatial variations in circulation, mixing, and biological processes, is not well understood.
The health of the LIS ecosystem is further affected by sediment pollution. Atmospheric deposition and river-borne fluxes of pollutants, both organic and metallic, led to the pollution of the top section of sediments (e.g., ROBERTSON et al., 1991; TURGEON and O'CONNOR, 1991; WOLFE et al., 1991; BENOIT et al., 1999; MECRAY and BUCHHOLTZ TEN BRINK, this volume; VAREKAMP et al., this volume). In addition, some locations became contaminated by dredge-spoil dumping (e.g., KOPPELMAN et al., 1976; SCHUBEL et al., 1979; TUREKIAN et al., 1980) or outflow from sewage treatment plants (BUCHHOLTZ TEN BRINK et al., this volume).
Records of abundant biota which easily fossilize may reflect the effects of anthropogenic and natural environmental changes in LIS. Benthic foraminifera are such biota and are abundant in LIS. Changes in foraminiferal faunas occurred over the last few decades, possibly in response to anthropogenic changes, in European coastal waters (e.g., the Baltic region; ALVE and MURRAY, 1995; ALVE, 1996) and in Chesapeake Bay as a result of climatic variations (CRONIN et al., 2000).
The earliest description of foraminifera from western LIS and New York Harbor dates from 1934 (SHUPACK, 1934). A more detailed study was conducted in the late 1940s (PARKER, 1952; faunas collected in 1948), followed by an exhaustive study in the 1960s (BUZAS, 1965; faunas collected in 1961-1962). BUZAS' (1965) data were further analyzed and discussed by MURRAY (1976). The morphology of the dominant Elphidium species in LIS was studied by BUZAS (1966), MILLER et al. (1982), and BUZAS et al. (1985).
LIS faunas are classified as 'marginally marine' (e.g., MURRAY, 1991), and typically have a low species diversity. The most abundant species have high tolerances for fluctuations in temperature and salinity, as well as for low oxygen conditions and environmental pollution (e.g., ALVE, 1995; CULVER and BUZAS, 1995). There is no correlation between substrate and species relative abundance, but foraminifera are absent in coarse sand, and thus rare or absent in eastern LIS.
BUZAS (1965) recognized three benthic foraminiferal assemblages in both the living and total (living plus dead) populations. In shallow waters (ranging from 3-19 m; average depth 11 m), he recognized the Elphidium clavatum assemblage (= Elphidium excavatum clavatum in MILLER et al., 1982), dominated by the nominate species at >60%. At depths ranging from 15-33 m (average depth 26 m), E. excavatum clavatum was common to abundant, but Buccella frigida was present at >9% and Eggerella advena at <19%. At depths between 19-39 m (average depth 30 m) Eggerella advena was present at >19%. Ammonia beccarii was rare (<5%) in all samples. BUZAS' (1965) data agree well with PARKER' s (1952) data, but the relative abundance of E. advena was somewhat higher in 1961, and the diversity was somewhat lower.
The LIS foraminiferal assemblages were thus depth-zoned, but the actual controlling, depth-linked parameter(s) were not known. BUZAS (1965) suggested that the foraminiferal populations might be largely controlled by the food influx from primary producers, with the E. excavatum-dominated faunas in the shallower waters feeding more on benthic algae and diatoms. MURRAY (1991) described E. excavatum clavatum as a herbivore and E. advena as a detritivore, using more degraded and refractory organic matter, as demonstrated by its high abundance close to sewage outfalls (ALVE, 1995).
Elphidium excavatum sequesters chloroplasts, mainly from diatoms (LOPEZ, 1979; BERNHARD and BOWSER, 1999), and can survive in polluted and low oxygen environments (CATO et al., 1980; MOODLEY, 1990; MOODLEY and HESS, 1992; ALVE, 1995). The sequestered chloroplasts might help in survival during low oxygen conditions (CEDHAGEN, 1991), as well as in increasing feeding efficiency (e.g., BERNHARD and BOWSER, 1999). E. excavatum dwells infaunally in marginal marine regions (LANGER et al., 1989; MOODLEY, 1990), and it lives in the top 4 cm of sediment in LIS (BUZAS, 1965).
Shallow-water foraminifera have been used widely to monitor environmental pollution, specifically the influx of organic waste (e.g., ALVE, 1995; SCOTT et al., 1995). Benthic foraminiferal faunas are characterized by high specimen abundance and low species diversity close to sewage outfalls, and they react to different types of pollution (sewage, nutrient enrichment resulting from aquaculture, metals) by secreting deformed tests (e.g., GESLIN et al., 1998; YANKO et al., 1998). The pollution levels necessary to cause deformation have not been quantified. It has been suggested that foraminiferal faunas could also be used to monitor heavy-metal pollution (e.g., YANKO et al., 1998; ALVE and OLSGARD, 1999). Extremely high levels of Cu pollution (>700 ppm in the sediment) have been reported to affect benthic foraminiferal faunas (ALVE and OLSGARD, 1999), but data on the specific levels of pollution required to influence the faunas are contradictory.
The oxygen isotopic composition of calcareous foraminiferal tests can be used to estimate the water temperature during the time of calcification of the organisms, as is commonly done on specimens from open marine environments (e.g., ANDERSON and ARTHUR, 1983). In brackish environments, however, both salinity and temperature influence the d18O composition of carbonate tests.
The carbon isotopic composition of foraminiferal tests in coastal waters also depends on salinity, because riverine bicarbonate has lower d 13C values than open ocean bicarbonate (e.g., ANDERSON and ARTHUR 1983). In addition to the salinity effect, local primary productivity may cause more positive d 13C values in carbonates deposited at the location of productivity in surface waters, and oxidation of organic matter in bottom waters will create more negative d 13C values in carbonates. Oxidation of terrestrial organic carbon would create lighter d 13C values in bottom waters without the associated heavy values in surface waters. Oxidation of isotopically light organic carbon from local productivity and from terrigenous sources would result in oxygen consumption, and thus hypoxia or anoxia. According to the NEW YORK STATE DEPARTMENT OF ENVIRONMENTAL CONSERVATION and CONNECTICUT DEPARTMENT OF ENVIRONMENTAL PROTECTION (1999), during periods of anoxia/hypoxia up to about 25% of oxygen in western LIS may be consumed by oxidation of terrigenous organic material, the remainder by oxidation of locally produced organic matter. Negative carbon isotopic values in benthic foraminifera can indicate the occurrence of hypoxia or anoxia independent of the source of the organic carbon (e.g., McCORKLE and EMERSON, 1988).
We compare data on the composition of benthic foraminiferal assemblages in LIS in 1996/1997 with observations made before the first recognized episode of severe anoxia in 1971 (PARKER and O'REILLY, 1991), and present oxygen and carbon isotope data on benthic foraminiferal tests collected in 1961 and 1996/1997. The identification of changes in benthic foraminiferal distributions, linked to specific variations in environmental parameters, will be of great value for the reconstruction of past environments in LIS from core data.
MATERIALS AND METHODS
Surface sediments (0-2 cm below the sediment-water interface) were collected using a Van Veen grab sampler equipped with a video and still camera system by the U. S. Geological Survey (POPPE and POLLONI, 1998). The R/V Seaward Explorer was used on two cruises in 1996 (MECRAY and BUCHHOLTZ TEN BRINK, this volume), and the R/V John Dempsey on a cruise in 1997 (POPPE and POLLONI, 1998).
We selected 42 samples for foraminiferal analysis (Figure 1, Table 1) to provide coverage along the east-west axis of LIS, as well as along north-south transects which cross its bathymetric features (e.g., BUCHHOLTZ TEN BRINK et al., this volume). Our samples cover a depth range between 5 and 35 m, and a longitudinal range between 72.407 and 73.747 (decimal) degrees W (Figs. 2a, 2b). PARKER's (1952) data were collected in the eastern part of LIS, between 72.492 and 73.098 (decimal) degrees W, and between 6.5 and 29 m depth. BUZAS (1965) presents the largest data set and covers the largest depth range (3 &endash; 43 m). His samples range in longitude between 72.852 and 73.700 (decimal) degrees W. There are few samples in westernmost LIS, and many samples close to longitude 73.0 oW (Figure 2).
In order to compare our benthic foraminiferal data with those in PARKER (1952) and BUZAS (1965) we needed a consistent taxonomy. The main taxonomic problems are with the species of the genus Elphidium. The taxonomy of this genus is complex because its species show great morphological variability (e.g., MILLER et al., 1982; BUZAS et al., 1985). We used the taxonomic concept of MILLER et al., (1982), but BUZAS (1965) and PARKER (1952) used a different concept.
We compared our specimens with specimens in BUZAS' (1965) collection in the Peabody Museum of Yale University and concluded that E. clavatum in BUZAS (1965) is the same morphological group that we name E. excavatum forma clavatum . We consider PARKER's (1952) informal taxonomic group 'Elphidium incertum and variants' equal to our E. excavatum clavatum, conform MURRAY (1991). We consider her E. subarcticum equal to our E. excavatum forma excavatum.
The grab samples were processed as described in MECRAY and BUCHHOLTZ TEN BRINK (this volume); part of the freeze-dried samples was wet-sieved over a 63 mm screen for foraminiferal analysis. The complete fraction larger than 63 mm was used. BUZAS (1965) also used the >63 mm size fraction, but no information was given on the size fraction used for foraminiferal study by PARKER (1952).
The grab samples were not stained to detect living specimens, so we could only study total faunas (living and dead specimens), and compare these with similar data in BUZAS (1965). At several locations a sub-sample of the grab samples was placed in formaldehyde. These samples were not buffered for long-term storage, and all carbonate material was dissolved at the time of analysis. We processed and studied several of these samples in order to determine whether agglutinant specimens had been affected by freeze-drying. We found no differences between agglutinant taxa abundance in the formaldehyde and the freeze-dried samples.
We prepared splits of the samples to contain at least 100 specimens of foraminifera. Specimens were picked from all samples, identified to species level, and stored in foraminiferal slides. Relative abundances of the most common species of foraminifera are shown in Table 1. Locations of the sample stations of BUZAS (1965), who did not give latitude and longitude for his stations, were read from his original field map (Table 2), kept with BUZAS' (1965) collections of foraminifera at the Peabody Museum of Yale University. The locations of BUZAS' (1965) samples can not be determined with precision, because dead reckoning was used in navigation. PARKER (1952) listed latitude and longitude for her sample stations.
We analyzed specimens of Elphidium excavatum clavatum from our and BUZAS' (1965) samples for carbon and oxygen isotopic composition at the mass spectrometer facility, Department of Geology and Geophysics, Yale University. The values are reported with respect to the general standard of isotope analyses of carbonates, PDB (the isotopic value of a belemnite from the PeeDee Formation); the analytical precision is + 0.1 for both oxygen and carbon isotopic values.
Foraminifera were common in most samples (Table 1), with the exception of a few samples taken east of 72.6 oW (Hammonassett), where the sediment is predominantly sandy, and sidescan data indicate sediment erosion, non deposition and transport (KNEBEL et al., 1999; KNEBEL and POPPE, this volume). Samples in this region have only a few, badly corroded specimens of Elphidium spp. (no foraminifera listed in Table 1). These specimens can not be determined to the species level because they have been transported and eroded. Elphidium excavatum preferentially survives extensive transport, as documented in the north-west European shallow-water environments of the Kattegat (NORDBERG et al., 1999).
In addition to foraminifera, carbonate shells of bivalves and gastropods are common in the samples. This appears to be in contrast with previous work indicating that carbonate in LIS sediments is largely dissolved (ALLER, 1982; GREEN et al., 1992, 1993; 1998; ALLER and ALLER, 1998). In samples in western LIS (west of 73o30'W) some foraminifera show incipient dissolution (starting at the periphery), but in most samples even thin-walled specimens are well-preserved with no sign of dissolution (Plate 1). In BUZAS' (1965) samples, most specimens were also well-preserved, again with the exception of a few samples in western LIS (MURRAY, 1976). We do not understand this discrepancy between our observations and published chemical models. Possibly there is strong local variability in LIS in such factors as intensity of bioturbation and diagenesis (e.g., KROM and BENNET, 1985; BUCHHOLTZ TEN BRINK et al., this volume).
The LIS faunas are low-diversity as compared to open marine shelf faunas, and the most abundant species are those observed by PARKER (1952) and BUZAS (1965) (Figure 3a-d, Figure 4 a-d). By far the most abundant species belong to the genus Elphidium (Figures 3a, 4a). We identified most of the Elphidium species in LIS as E. excavatum s.l., following the species concepts of MILLER et al. (1982), with the forma E. excavatum clavatum most common. The second most common forma is E. excavatum excavatum; E. incertum (Williamson) is also common in many samples (BUZAS, 1966).
E. excavatum clavatum is most abundant in LIS at depths shallower than about 15 m, in agreement with BUZAS (1965). This trend is very broad, however, with numerous exceptions (Figure 3a). In our data, the relative abundance of E. excavatum clavatum is significantly correlated with longitude, with the species being less abundant in western LIS (Figure 4a). This correlation with longitude was more significant than the correlation with depth, whereas in BUZAS' (1965) the correlation with depth was significant, and the correlation with longitude was absent or non-significant (Table 3).
After Elphidium species, Buccella frigida is most common (Figures 3 b, 4b), in agreement with observations by PARKER (1952) and BUZAS (1965). This species tends to be more common at the deeper locations, particularly below about 15m, but this is a broad trend with numerous exceptions. The species is more common in the western and central basins of LIS than in the eastern basin, resulting in a weak positive correlation of its relative abundance with longitude (Table 3, Figure 4b).
The third most common species in the studies by PARKER (1952) and BUZAS (1965) was the agglutinant Eggerella advena, especially in the deeper parts of LIS (Figure 3c). This taxon occurred at less than a few percent in relative abundance in all our samples. The lack of abundant specimens of E. advena even in samples close to locations where it was common to abundant in BUZAS (1965) and PARKER (1952), was probably not caused by sampling artifacts (see section METHODS). PARKER' s (1952) study shows a lower maximum abundance of E. advena than BUZAS' (1965) study; east of 73 oW, PARKER' s (1952) values are slightly higher than those of BUZAS (1965), but her values for stations at about 73 oW are lower. The calcareous taxon Ammonia beccarii (Figures 3d, 4d) was more common in our samples than in the older data sets. This species was rare (<5% ) in all samples described by PARKER (1952) and BUZAS (1965), but it reaches peak abundances of about 27% in our samples from the westernmost region of LIS.
Evaluation of the number of species (normalized to the number of specimens counted) shows that the diversity of benthic foraminifera decreased somewhat since the 1960s, and that the diversity may also have been somewhat lower in the 1960s than in the 1940s (Figure 5). Alpha-diversity was below 1 for almost all our samples, and varied between 1 and 3 in the 1960s (BUZAS, 1965). The drop in diversity is caused by a decrease in abundance of the uncommon to rare species, which occur in a few samples only, as single or a few specimens. They include Rheophax dentaliniformis, Rheophax nana, Quinqueloculina seminula, Trochammina squammata, and Cornuspira planorbis. In a few samples towards the eastern end of LIS (E- and F transects, Fig. 1) specimens of Cibicides lobatulus, Nonion commune and Dentalina spp. occurred as rare individuals. In samples from close to the Connecticut shoreline, a few specimens of the marsh foraminiferal species Trochammina macrescens, Haplophragmoides manilaensis and Trochammina inflata occur, which are probably transported into LIS from the coastal marshes. Neither we, nor PARKER (1952), nor BUZAS (1965) did find planktonic foraminifera, which do not occur in such marginally marine environments as LIS and are apparently not transported into LIS from the Atlantic Ocean.
Stable Isotope Data
We determined carbon and oxygen isotope compositions of Elphidium excavatum clavatum in twelve LIS surface sediment samples from the 1996/1997 cruises and four samples from the 1961 collection of Buzas (1965) (Table 4, Fig. 6). Most of the d 18O values are lighter than those in the open ocean, reflecting the mixing of sea water with isotopically light river water in LIS. The d 18O values are significantly lighter in western LIS (Figure 6), as expected from the overall lower salinities in western LIS (e.g., KOPPELMAN et al., 1976). If the few studied samples may be considered representative for LIS, there are no significant differences in d 18O values of samples from 1961 and from 1996.
We made a preliminary mixing model of river water and sea water for LIS, and calculated the d 18O values in water and the d 13C values in dissolved bicarbonate as a function of salinity (Figure 7; Table 5). According to this model, the d 18O values in LIS waters should vary between minus 1 and minus 4 . We compiled average bottom water salinities for each LIS sampling site (after KOPPELMAN et al., 1976) and calculated d 18O values for water with the mixing model. From the calculated water and measured carbonate d 18O values, we obtained water-calcite equilibration temperatures (ANDERSON and ARTHUR, 1983), which range from 7.0 to 10.6 oC, with one anomalously low value of 5.4 oC (Figure 6). Such bottom water temperatures occur from late spring through summer (KOPPELMAN et al., 1976), at which time the foraminifera show a major population increase (BUZAS, 1965). The life span of E. excavatum clavatum probably ranges from a few months to a year (MURRAY, 1991). The calculated temperatures provide an average water temperature over the life span of the organisms, because foraminifera calcify their tests incrementally by chamber addition.
The oxygen isotope values change with longitude, so we need to look at data from a fixed longitude to see whether a depth effect exists (Figure 6). Four data points on the approximately north-south trending A-transect in west LIS (Figure 1) show no relationship between depth and d 18O. In these preliminary data we therefore do not find evidence for thermal or salinity-related stratification (e.g., RILEY, 1956, 1959; KOPPELMAN et al., 1976; GORDON, 1980), probably because all samples are from below 10-11 m, the average depth of the pycnocline (e.g., WELSH and ELLER, 1991).
The d 13C values also show an east-west trend, with lighter values in western LIS. There is no correlation between d 13C values and depth (Figure 6); we do not see an effect of photosynthesis in shallow waters, nor an effect of symbiotic algae. This may result from the fact that we do not have samples from shallower depths than about 10 m, which is about the depth of light penetration (e.g., BUZAS, 1965).
The range in observed d 13C values (Table 4; between minus 4.21 and minus 1.89) is considerably larger than would be expected from the mixing of fresh and salt water alone (Figure 7). Most likely, the oxidation of isotopically light organic matter is reflected in the carbon isotopic data. Because oxidation of organic matter consumes oxygen, the carbon isotope data thus provide an indirect estimate of the level of oxygenation. In order to evaluate this oxidation effect, we removed the salinity effect from the d13C data by calculating water salinities at an assumed water temperature from the measured d18O data, using 10 oC as a time-averaged LIS-wide bottom water temperature.
Our calculations involved the following steps: from the measured d18O values in carbonate, we calculated the d18O of the water, and then used the LIS mixing model to estimate the salinity. From the calculated salinity we estimated the d13C value of dissolved inorganic carbon in LIS water, again using the mixing model. We subtracted the calculated values from the observed d13C value, which we label the d13C &endash;excess data (d13C *) or:
under the assumption that little or no carbon isotope fractionation takes place during calcification. We do not know the 'vital effect' of the calcification of E. excavatum clavatum, but this is not important for the present discussion because we are looking at differences in d13C between individuals of the same species. The use of higher temperatures results in more negative calculated excess d13C values.
The d13C * values show a strong east-west trend, with lighter d13C * data at the west end of LIS (Figure 8). The "salinity-corrected" light carbon isotope values in western LIS are thus probably caused by the oxidation of organic matter, which adds isotopically light carbon to the dissolved inorganic carbon pool (e.g., ANDERSON and ARTHUR, 1983; McCORKLE and EMERSON, 1988; McCORKLE et al., 1990; CHANDLER et al., 1996). Elphidium excavatum clavatum lives in the upper 1-4 cm of sediment (Buzas, 1965). The carbon isotopic composition of its test can thus be expected to reflect the isotopic value of dissolved carbonate in the pore waters of the upper part of the sediment column or in the overlying bottom waters.
In western LIS, the planktonic primary productivity is highest (SUN et al., 1994), and the oxidized organic carbon could thus have been derived from local productivity. In that case, we would expect the dissolved bicarbonate in the surface waters (where the primary productivity occurs) to be enriched in d13C. We have no evidence for this, because we have no data on samples shallower than 10 m, where we would expect the surface d13C enrichment to occur (WELSH and ELLER, 1991). Alternatively, the oxidized organic carbon in western LIS could have been terrigenous material, natural or anthropogenic in origin. Additional data on the carbon isotopic composition of foraminifera from shallow water (5 - 10 m depth) are needed before we can decide whether the oxidized organic carbon was mainly derived from local productivity or from terrigenous sources.
The foraminifera in the 1961 samples time had slightly heavier d13C values than in 1996/1997 (Figure 8), indicating that a smaller amount of organic matter was oxidized in LIS in the early 1960s than in the late 1990s. Our data suggest not only that the amount of organic carbon that could be oxidized in LIS increased since the early 1960s, but also that the difference in the amount of organic matter from east to west in LIS increased, as indicated by the steeper slope of the correlation line between d13C * and longitude (Figure 8). Although based on few data points, these preliminary results are in agreement with the hypothesis that anoxic/hypoxic conditions in LIS did not occur in the early 1960s, in agreement with e.g., PARKER and O'REILLY (1991). If further research confirms our observations on the carbon isotopic composition of E. excavatum clavatum, the magnitude of d13C * in foraminiferal tests in LIS cores can be used to detect the occurrence and severity of anoxic periods of the past.
The relative abundances of common benthic foraminifera in LIS have changed since the early 1960s. The changes include: (1) an increase in relative abundance of Ammonia beccarii in west LIS; (2) a decrease in relative abundance of Eggerella advena in the deeper parts of the whole Sound; (3) an increase in correlation between longitude and the relative abundance of species, and a decrease in correlation with depth; and (4) a decrease in the already low species diversity. Our preliminary isotope data suggest that organic productivity in LIS has increased since the early 1960s, with a stronger increase in western LIS. Both faunal and isotope data thus indicate increased differences between eastern and western LIS since the early 1960s. The decrease in species diversity coeval with changes in relative species abundances strongly suggests that there has been an increase in environmental stress, from pollution, seasonal anoxia, or other causes (e.g., MURRAY, 1991).
These conclusions are preliminary. In order to interpret the isotopic data fully we must establish the isotope systematics of living E. excavatum clavatum and the magnitude of its possible deviation from isotopic equilibrium. In order to demonstrate unequivocally that the present benthic foraminiferal fauna in LIS differs from that in 1961, we need to establish which foraminifera were alive at the time of sampling (WALTON, 1952), and what the effects are of seasonal changes in foraminiferal distribution.
To understand the causes of benthic foraminiferal faunal change over the last few decades in LIS, we must understand which environmental variables affect the abundance of the most common species. BUZAS (1965) suggested that the overall zonation with depth (with Elphidium excavatum clavatum most dominant at shallower levels) might be explained by the fact that foraminifera are selective feeders, and that their depth zonation is related to the distribution of the material upon which they feed. There appears to be no physicochemical factor in the water column to cause a foraminiferal depth zonation; LIS waters are thoroughly mixed every fall through spring, with stratification only present during the summer, and all the common species have been reported to survive low oxygen conditions (e.g., ALVE, 1995). Oxygen isotope values measured on E. excavatum clavatum show no correlation with depth (Figure 6).
Elphidium excavatum sequesters chloroplasts, mainly from diatoms, and these chloroplasts might actively photosynthesize within the foraminiferal protoplasm (LOPEZ, 1979; BERNHARD, 1996; BERNHARD and BOWSER, 1999). The lower limit of benthic microflora in LIS is about 11 m (BUZAS 1965); the highest abundance of this species might occur in the photic zone where the sequestered chloroplasts can perform photosynthesis.
The weak correlation between the relative abundance of Buccella frigida and longitude (Table 3) is probably caused by the occurrence of this species in the deeper parts of the depositional basins of LIS (KNEBEL et al., 1999; KNEBEL and POPPE, this volume), where sediment is dominantly fine-grained and total organic carbon levels in the sediment are high (Figure 9). The species is less common in the eastern basins, where sediment erosion, non deposition, and bedload transport occur (KNEBEL et al., 1999; KNEBEL and POPPE, this volume). This difference between eastern and western basins fully explains the weak correlation with longitude in the relative abundance of Buccella frigida. This species is probably a detritivore, which becomes dominant below depths where E. excavatum can dominate by using the sequestered chloroplasts, but only in regions where there is abundant fine sediment with organic carbon.
The decrease in relative abundance of Elphidium excavatum clavatum towards western LIS in the 1996/1997 data is linked to the strong increase in relative abundance of A. beccarii. Both E. excavatum clavatum and A. beccarii are able to survive extreme variability in oxygenation (MOODLEY and HESS 1992), in salinity, and in temperature (MURRAY, 1991), as well as high levels of pollution (e.g., ALVE, 1995). E. excavatum has a slightly smaller range in optimal temperature and salinity than A. beccarii, but our preliminary oxygen isotope data suggest that there were no significant changes in these parameters since 1961.
The east-west gradient in the abundance of A. beccarii in 1996/1997 suggests that its increase in relative abundance since 1961 is linked to an environmental variable which shows an east-west gradient. Metal pollution increases from the east to the west (e.g., TURGEON and O'CONNOR, 1991; WOLFE et al., 1991; WILLIAMS et al., 1997; MECRAY and BUCHHOLTZ TEN BRINK, this volume; VAREKAMP et al., this volume), and foraminiferal populations may be affected by metal pollution (YANKO et al., 1998; ALVE and OLSGARD, 1999). Copper levels in LIS sediments (MECRAY and BUCHHOLTZ TEN BRINK, this volume) do not reach the high levels reported to affect foraminifera (ALVE and OLSGARD, 1999). Neither we, nor BUZAS (1965), nor PARKER (1952), observed high percentages of deformed tests, and metal pollution is therefore probably not implicated.
The seasonal episodes of anoxia in western LIS which have occurred since 1971 (PARKER and O'REILLY, 1991) might be implicated, although both E. excavatum and A. beccarii can survive very low oxygen concentrations (MOODLEY and HESS, 1992). A comparison of our relative abundance data with the data on abundance of the bacterial spore Clostridium perfringens, an indicator of sewage pollution (BUCHHOLTZ TEN BRINK et al., this volume), shows that the highest abundance of A. beccarii occurs in samples with the highest sewage input (Figure 10). Clostridium perfringens has become more abundant over the last few decades (BUCHHOLTZ TEN BRINK et al., this volume), and the increased abundance of A. beccarii thus might result from increased sewage inputs into western LIS.
The linkage between sewage input and abundance of A. beccarii could very well be indirect. In Chesapeake Bay, planktonic centric diatoms increased in relative abundance over the last few hundreds of years, while benthic diatoms decreased, reflecting increased turbidity and eutrophication (COOPER and BRUSH, 1991, 1993; COOPER, 1999a, b). In Narragansett Bay, just to the east of LIS, such eutrophication has also occurred (NIXON, 1997). Similar processes probably also occurred in LIS, but no record is available. Centric diatoms are very abundant in our samples, especially in western LIS, where algal blooms are most intense (SUN et al., 1994). In addition, our carbon isotope data (Figure 8) indicate that the amount of organic matter that was oxidized in LIS increased, and that the difference in the amount of organic matter available from east to west LIS increased. Possibly, A. beccarii became more successful than E. excavatum clavatum under increasingly eutrophic conditions resulting from increased sewage input. Additional data on d13C values of E. excavatum clavatum in BUZAS' (1965) material, and in d13C values of E. excavatum clavatum and A. beccarii in our samples, might help to decide whether A. beccarii uses sewage directly or feeds upon algal production fueled by the nutrients in sewage influx.
It is more difficult to find an explanation for the strong decrease in relative abundance of E. advena since the 1960s. Eggerella advena appears to have varied in relative abundance in LIS over short time scales. BUZAS (1965) documented a variable relative abundance of this species between the years 1961 and 1962. The species was more abundant in 1961 and 1962 than it was in PARKER's (1952) data. Eggerella advena has less tolerance for low salinity than the other LIS species (MURRAY, 1991). The year 1948, as well as the years 1961-1962, were drought years, and the diminished fresh-water discharge could have resulted in higher-than-average salinities (KOPPELMAN et al., 1976). Our preliminary oxygen isotopic data, however, suggest no apparent salinity differences between 1961 and 1996 (Fig. 6).
Eggerella advena decreased in abundance over the full east-west extent of LIS, suggesting that this decrease can not be linked to an environmental variable with an east-west gradient, such as anoxia, pollution, or eutrophication. This species survives in even highly polluted and oxygen depleted waters, such as close to sewage outfalls, where it is said to out-compete E. excavatum and to use refractory organic matter efficiently (e.g., ALVE, 1995; CULVER and BUZAS, 1995). We speculate that the relative abundance of E. advena might have been influenced by changes in the phytoplankton community of LIS. As suggested above, centric diatoms could have increased in abundance in LIS as a result of increased eutrophication. The carbon isotope data (Figure 8) indicate that local productivity might have increased over the full length of LIS since the early 1960s. If that was the case, the decrease in abundance of the detritivorous E. advena may have been caused by a change in the composition of the food supply, with a larger part of the total made up by fresh phytoplankton, less by more refractory (possibly terrigenous) organic carbon. More research is needed to establish what caused the abundance fluctuations of this species, which may turn out to be a sensitive environmental indicator species on short time scales.
Benthic foraminifera (dead plus live) in LIS in 1996/1997 were low-diversity, and dominated by Elphidium excavatum clavatum, as they were in 1948 (PARKER, 1952) and 1961 (BUZAS, 1965). The second most common species in 1996 was Buccella frigida, also as in 1948 and 1961. The third most common species in 1948 and 1961, Eggerella advena, was rare in all samples in 1996. Ammonia beccarii was rare in 1948 and 1961, but formed up to 27% of the assemblage in western LIS in 1996/1997. Species diversity decreased since the 1960s, suggesting that environmental stress from hypoxia, pollution or other causes increased and caused the changes in species' relative abundance.
We can not explain the decrease in relative abundance of E. advena that occurred since the early 1960s over the full east-west length of LIS. Our preliminary data on oxygen isotopes in E. excavatum clavatum suggest that there were no significant changes in salinity over this time period, but our carbon isotope data suggest that the rate of oxidation of organic matter in LIS increased between the early 1960s and the late 1990s. Therefore, we speculate that possible changes in composition of the phytoplankton and/or in the total primary productivity may have caused the decline in abundance of E. advena. The increase in relative abundance of A. beccarii in western LIS since the early 1960s might have been caused by increased sewage input, either directly, or indirectly by algal blooms fueled by the nutrients in the sewage. Our carbon isotopic data indicate that more organic matter was oxidized in western LIS than further to the east, and that this gradient in oxidation of organic matter increased between the early 1960s and the late 1990s.
More research is needed to fully understand the isotope systematics of the LIS benthic foraminifera and the changes in benthic foraminiferal faunas over the last 35 years, but the observation that their assemblages changed over this time period suggests that they can be valuable environmental indicator species. The information obtained from stable isotope data is important in understanding which environmental changes could have caused these faunal changes.
We thank Zobeida Monserrate-Cruz and Polina Rabinovich for their help in picking and counting benthic foraminifera and in making the SEM micrographs, and Dan McCorkle for sampling assistance. We are also grateful to Tim White and Copeland McClintock at the Yale University Peabody Museum for their help in access to the foraminiferal and map collection of Buzas. We thank John Kingston at the Mass Spectrometer facility of the Yale University Department of Geology and Geophysics for the carbon and oxygen isotope analyses. We thank H. J. Knebel, C. Wylie Poag, M. A. Buzas and J. W. Murray for their helpful reviews. ET and JCV acknowledge financial support from the Connecticut Sea Grant College Program (grant nr.DSG443 R/ER-2), from Wesleyan University, and from the U. S. Geological Survey Coastal and Marine Geology Program.