1US Geological Survey, Woods Hole, MA 02543
²Department of Biology, University of New Haven,300 Orange Ave.
West Haven, CT 06516
³Long Island Sound Resource Center, Connecticut Geological and Natural History Survey,
Department of Environmental Protection, 1084 Shennecossett Road, Avery Point, Groton, CT 06340

Ongoing research by the U.S. Geological Survey in Long Island Sound, a major U.S. east-coast estuary surrounded by the most densely populated region of the United States, is building upon cooperative research with the State of Connecticut that was initiated in 1982. During the initial phase of this cooperative program, geologic framework studies in Long Island Sound were completed (Lewis and Needell, 1987; Needell and others, 1987; Lewis and Stone, 1991). Emphasis of the present program in the Sound is shifting from framework studies toward studies of the sediment distribution, processes that control this sediment distribution, nearshore environmental concerns, and the relationship of benthic community structures to the sea-floor geology.

Because of the enormous surrounding population, large inputs of anthropogenic wastes (e.g. fertilizer and sewage) and toxic chemicals have produced stresses on the environment of the Sound, causing degradation and potential loss of benthic habitats (Long Island Sound Study, 1994). To examine this problem, we have completed a surficial geologic map of the western Long Island Sound sea floor off Norwalk, Conn., in an effort to define benthic habitats and how they are changing with time. This study area is of special interest because it is affected by seasonal hypoxia. It covers approximately 60 km² and is located 3-15 km south of the Norwalk Islands (location map) in approximately 16-48 m water depth. The U.S. Geological Survey, in cooperation with the University of New Haven and the Connecticut Geological and Natural History Survey, completed a bathymetric, sidescan sonar, and high-resolution seismic-reflection survey of the study area during November, 1993 as the first phase of this long-term program. The second phase, which involved bottom sampling and photography, was completed during April - August 1995.   This report presents surface and shallow subsurface geological information including the bathymetry, a sidescan sonar mosaic, a preliminary interpretation of the mosaic, and a surficial-sediment distribution map of the study area (fig. 1).

Long Island Sound is a large estuary that is underlain by Precambrian and Paleozoic rocks along its northern side and Mesozoic and Cenozoic strata along its southern side (Rodgers, 1985; Fuller, 1914; Lewis and Stone, 1991). Surficial sediments in the Sound were deposited primarily during the Wisconsinan and Holocene glacial and interglacial stages. This report deals only with the shallow stratigraphy, thus, this geologic summary is limited to the glacial and post-glacial history of the Sound. A more detailed discussion of the entire geological history of the Sound has been presented by Lewis and Stone (1991). A Wisconsinan glacier, the Laurentide ice sheet, advanced across what is now Long Island Sound and formed the Ronkonkoma and Harbor Hill end moraines on Long Island (Schafer and Hartshorn, 1965; Sirkin, 1982). Upon retreat of the Laurentide ice sheet from the Sound, a glacial lake formed behind the moraines. Thick deposits of varved clays accumulated in the central part of the Sound, and deltas formed along the northern edge of this lake. Deltaic facies presently underlie much of the northern part of the Sound (Lewis and Stone, 1991). The moraine that served as the dam for this lake was eroded at the spillway, probably between Fishers Island and the eastern tip of Long Island (the Race), and the lake drained (see location map). The exposed lake floor was cut by river channels which ran down from high ground on either side of the Sound to join a central river which ran eastward down the axis and exited the basin through the Race. As sea level rose, the Sound was flooded. The transgressing sea eroded the underlying strata and somewhat smoothed the fluvially cut surface. On top of this marine unconformity, sediments have accumulated during the Holocene.

Bathymetric data, sidescan sonar imagery, and 3.5-kHz subbottom profiles were collected during November 13-16, 1993 aboard the R/V John Dempsey along tracks spaced about 150 m apart (fig. 2). The bathymetric data were collected using an Odem 200-kHz echo sounder, and the data were logged digitally on an IBM personal computer. The sidescan sonar data were collected using a 100-kHz Klein sidescan sonar system set to a 100-m range to either side of the ship's track. These data were logged digitally on a QMIPS data logger (Danforth and others, 1991), and were written to tape at 2048 pixels per scan. The subbottom data were collected using an Ocean Research Equipment 3.5-kHz subbottom profiler transmitting at a 0.25-s repetition rate. These data were logged in analog form on a paper recorder. Ship's navigation was by differential GPS using the shore station at Montauk Point, Long Island. Navigation fixes were logged onto computer every 10 s. Ranges to the sidescan vehicle from the ship were logged during the cruise, and these ranges were used to compute the navigation of the sidescan vehicle after the cruise was completed.

Surficial sediment samples and bottom photography were attempted at 67 locations during April 27-29 and August 25-26, 1995 cruises aboard the RV John Dempsey using a Van Veen grab sampler equipped with video and still camera systems. These photographic systems were used to appraise intra-station bottom variability and to observe the benthic flora and fauna. Surficial sediments (sediments from the sediment-water interface down to 2 cm) were subsampled from the grab sampler; then frozen and stored for later analysis.

Post-cruise processing was performed on the digital data sets, and an interpretative geologic map was made of the analog 3.5-kHz subbottom profiles (figs. 3,4). The bathymetric data had to be corrected for the 2 m tidal range in the study area. This correction was done by adjusting the measured depth values to mean sea level based on tide levels measured at the Bridgeport, Conn. tide station during the study period.

The sidescan sonar data were processed, and the final products were a digital mosaic and an interpretation of the entire study area (figs. 5,6, respectively). The processing steps included compressing the raw sidescan-sonar data through a median filtering routine to suppress speckle noise (Malinverno and others, 1990; Danforth and others, 1991) and correcting for slant range distortion, signal attenuation, and dropped lines in the sonar data. These processing procedures are summarized by Danforth and others (1991), and Paskevich (1992a,b). The processed sidescan data were used to make the composite digital mosaic. The mosaicking procedure included placing each strip of sonar data in its proper geographic location at the appropriate scale and projection (Paskevich, 1992c), trimming off unwanted data where two sonar images overlapped, tone matching the adjacent sonar images if there are differences, and combining the individual strips together into a composite image. The final editing step and tone matching was completed using a software package developed by PCI (Nelson, 1991). The final digital mosaic was output to film and photographically enlarged to a scale of 1:15,000.

In the laboratory, the textural samples were corrected for salt contents, disaggregated, and wet sieved through a number 230, 62 micron (4 phi) sieve to separate the coarse and fine fractions. The fine fraction was analyzed by Coulter Counter (Shideler, 1976); the coarse fraction was analyzed by sieving (gravel) and by rapid sediment analyzer (sand; Schlee, 1966). Because biogenic carbonates commonly form in situ, they usually are not representative of the depositional environment from a textural standpoint. Therefore, bivalve shells and other biogenic carbonate debris were manually removed from the gravel fraction. Size classifications are based on the method proposed by Wentworth (1929). The verbal equivalents were calculated using the inclusive graphics statistical method (Folk, 1974) and are based on the nomenclature proposed by Shepard (1954). A detailed discussion of the laboratory methods employed are given in Poppe and others (1985); the raw grain-size data, associated statistics, and detailed descriptions of the bottom photography are reported in Poppe and others (1996). The map of the surficial sediment distribution shown in figure 7 is based on data from the sample and photographic stations, on tonal changes in backscatter on the sidescan sonar image, and on the bathymetry.

Bathymetry

The bathymetry was contoured at a 2 meter interval (fig. 2). It shows water depths that range from about 16 to about 48 m. A narrow trough that runs nearly east-west marks the axis of Long Island Sound in the study area. The sea floor south of the trough has a uniform gradient that shoals to approximately 26 m along the southern edge of the study area. North of the trough, the sea floor has a steeper gradient than the southern side up to approximately 28 m water depth. Shallower than 28 m, the gradient is much gentler. A subdued ridge along the western side of the study area extends south-southeast to the northern edge of the trough.

Near-surface Seismic Facies and Stratigraphy

Three distinctive seismic facies were identified on the 3.5-kHz subbottom profiles shown in figure 3: (1) a unit with a highly reflective and prolonged surface return with no subbottom reflections below it (shown as glacial drift on profiles A and B); (2) a unit with a moderately reflective, moderately prolonged surface return with no subbottom reflections below it (shown as glaciodeltaic deposits on profiles B and C); and (3) an acoustically transparent unit having a moderately reflective surface (shown as marine deposits on profiles A, B, and D). In places, the transparent unit is charged with gas and is acoustically opaque. The gas charged sediments blank out everything below them (profile D).

The facies with the highly reflective prolonged surface return represents the oldest of the units mapped in the study area. Based on the regional stratigraphy of Lewis and Stone (1991), this unit is interpreted to be Pleistocene glacial drift, or a lag deposit developed on the drift because fields of boulders are shown by the sidescan mosaic where it is exposed on the sea floor (figs. 5 and 6). Profile B in figure 2 shows that the drift underlies the two other seismically identified units. The drift crops out in a small area in the western part of the study area (fig. 4) on the southernmost part of the ridge shown in the bathymetry (fig. 2).

The facies with the moderately reflective surface overlies the acoustically prolonged unit (fig. 3, profile B), and crops out in the northwestern corner of the study area (figs. 3, profile C; fig. 4). Based on the regional stratigraphy (Lewis and Stone, 1991) we infer that this unit consists of Pleistocene-age sandy deltaic deposits that accumulated along the northern side of the Sound when it was occupied by Glacial Lake Connecticut. This facies can be traced under the acoustically transparent facies throughout the study area except where gas is present. The glaciodeltaic deposits are of unknown thickness because the 3.5-kHz subbottom profiling system did not have sufficient power to penetrate through them. The surface of the underlying drift could be traced down to approximately 30-40 m subbottom depth on the 3.5-kHz subbottom profiles (fig. 3, profile B) which suggests that, at least in places, the deltaic deposits are of this thickness.

The acoustically transparent facies (fig. 3, profiles A, B, and D) is the youngest of the three units and is interpreted to be a fine-grained Holocene marine deposit based on Lewis and Stone's (1991) regional stratigraphy. The thickness of the marine deposit is shown in figure 3. The bathymetry (fig. 2) and profiles (fig. 3, profiles A, B, and D) show that this deposit has a smooth surface, and the differences in thickness are due to its filling in irregularities in the late Pleistocene surface. The marine deposits are thickest in the central and southern parts of the study area, although its maximum thickness appears to be obscured by areas of gas charged sediment (fig. 4). This deposit is less than 5 m thick in the northern half of the area, and pinches out to the west where glaciodeltaic deposits are exposed on the sea floor.

Gas is concentrated within the marine deposits in the two areas where the unit is thickest (fig. 4). The largest patch is elongate, and coincides with the east-west trending bathymetric trough in the central part of the study area. Here, the thickness of the marine deposits surrounding the gassy sediments are mostly between 5 and 10 m. The second patch is to the south where the marine deposits surrounding the patch are greater than 10 m thick.

Sidescan Sonar Mosaic and Surficial Geologic Interpretation

The sidescan-sonar mosaic (fig. 5) reveals six distinctive sea floor patterns, and when combined with the seismic data can be used to make some inferences about the surficial geology (fig. 6). On figure 5 the six patterns include (1) high acoustic backscatter (light tones), (2) low backscatter (dark tones), (3) high-backscatter targets interpreted to be boulders, (4) depressions with high-backscatter floors, (5) east-west trending high-backscatter stripes with no apparent relief associated with them, and (6) linear and curvelinear narrow depressions interpreted to be trawl marks. Boat wakes and fish are two water column phenomenon that were observed as well.

The image can be divided into two general provinces; namely high and low acoustic backscatter. The other features tend to coincide with one or the other of these provinces. The boundary between high and low backscatter was set at a digital number of 50 (fig. 6). This number was picked because the distribution of values greater than 50 closely coincides with the extent of the glacial drift and glaciodeltaic deposits, while values less than 50 coincide with the distribution of the marine deposits shown in figure 4. The one exception is the high-backscatter area just south of the bathymetric trough through the center of the study area (fig. 6). Here grab samples show that the glaciodeltaic deposits are only covered by a thin veneer of the marine deposits. Perhaps the marine deposits are so thin that the sidescan signal penetrated them and imaged the underlying coarser material, or perhaps the marine deposits represent a seasonal accumulation that occurred between the time of the sidescan survey and the sampling.

The distribution of patches of boulders is limited to the high backscatter area (fig. 6). Boulders are found in the southern part of the area of high backscatter, and most patches coincide with the outcrop of glacial drift as shown in figure 3.

The area of low backscatter on the sidescan image coincides with the extent of the marine deposits of Holocene age. The boundaries between the high and low-backscatter regions are gradational, and backscatter intensity is not uniform throughout the low-backscatter area. In general, backscatter values tend to be lowest where the marine deposits are thickest, and the surficial sediments finest. Backscatter values are higher where the sediments are thinner. The areas of higher backscatter probably indicate mixing with the adjacent or underlying coarser-grained glaciodeltaic deposits. The gas observed in the marine deposits on the subbottom profiles does not have a surficial expression on the sidescan image, because its distribution coincides with areas of varying backscatter intensity (fig. 5).

Shallow depressions with high-backscatter floors occur only in areas of low acoustic backscatter, but also are not associated with the distribution of gas-charged sediments (fig. 6). Most of these depressions are less than 40 m in diameter, and most of them occur in the northeastern part of the study area where the marine deposits are less than 5 m thick. The origin of these features is unknown.

Linear high-backscatter stripes occur along some of the boundaries between areas of high and low backscatter, but always in areas where there is a thin cover of marine deposits. They are most abundant around the high-backscatter area south of the bathymetric trough. These features are less than 10 m wide and up to 150 m in length. They all are oriented east-west, roughly parallel with the major axis of the tidal currents (Haight, 1942), and do not appear to have any relief associated with them. Their distribution near the edges of the areas of high backscatter suggests that they may be places where the tidal currents have scoured the thin cover of marine deposits, exposed the underlying coarser-grained material, and created patchy lithologic distributions (fig. 7).

Surficial Sediment Distribution

The locations of the sediment sample and bottom video stations and surficial sediment distribution are shown in figure 7. The southern and eastern portions of the study area are characterized by clayey silts. These fine-grained sediments are typically poorly to very poorly sorted (standard deviations are usually greater than 2.0), and have unimodal, coarsely-skewed to nearly-symmetrical distributions. Although data on the total organic carbon content of the samples is unavailable, petrographic observation of smear slides suggests that more than 20 percent of these samples may be biogenic in nature.

The fine-grained sediments are also extensively bioturbated; shrimp and lobster burrows, worm and amphipod tubes, mollusk shell hash, and tracks were commonly present at most of the stations. Most of the bedforms observed on the bottom video from this part of the study area appear to also have originated from the biological activity. For example, crustaceans mound the sediments while feeding and burrowing. These mounds, which extend above the surrounding sediments, are then modified into the rippled and undulating bottoms by tidal currents.

Gravelly sands are predominant in the northwestern and west-central portions of the study area. Relatively small patches of boulders and gravel occur within the gravelly sands, and bands of sand, silty sand and sand-silt-clay surround the gravelly areas, separating them from the lower energy environments dominated by clayey silt. These coarser sediments are also very poorly sorted, but are commonly finely skewed and have bimodal distributions. In the shallow northwestern part of the study area, scour and shadow effects, probably a result of the strong tidal currents, were observed around most of the isolated gravel-sized rocks. Gravel concentrations within the gravelly sands decrease with depth in the grab sampler, especially in the northwestern corner of the study area, suggesting that the finer-grained sand and mud have been winnowed away leaving the gravel to partially armor the surface. Oyster and razor clam shells litter the bottom, hydrozoans grow on the gravel and shell debris, and spider and rock crabs are common in this area. A relatively small (< 1km in diameter) patch of gravelly sediment, which coincides with an isolated area of high backscatter on the sidescan sonar image, occurs in the south-central part of the study area, but is covered by a thin (less than 2 cm), possibly seasonal, veneer of muddy sediment.

The detailed view of the surficial geology of this part of Long Island Sound shows that it is closely related to the shallow subsurface stratigraphy. Areas of high acoustic backscatter mostly coincide with outcrops of glaciodeltaic deposits and drift, and boulders are mostly limited to where the drift is exposed on the seafloor. The area of low acoustic backscatter on the sidescan mosaic coincides with the extent of the marine deposits.The high-backscatter stripes and high-backscatter depressions on the seafloor all occur where underlain by marine deposits, but their origin still needs to be resolved. This interpretation of the surficial geology based on the bathymetry, acoustic information, sampling, and photography provides a valuable base map for future studies of sedimentary processes and the interaction of biological, physical, and chemical activity with the sea floor in Long Island Sound.

We acknowledge the help of a number of people for the successful completion of this project. The skilled handling of the RV John Dempsey by Peter Simpson, Miles Peterle and Kurt Gottschall resulted in an extremely successful field program. Dr. Anthony Calabrese, director of the NOAA, National Marine Fisheries Lab in Milford, Connecticut is thanked for making space available at that lab for processing of the sidescan sonar data. V. Paskevich aided in reformatting and transferring the sidescan sonar data to the image processing computer. Steve Gill of the Tidal Analysis Branch of NOAA provided the tide measurements from the tide station at Bridgeport, CT that were used to correct the bathymetry. Drafting was done by Jeff Zwinakis, and photography by Dann Blackwood. Robert N. Oldale and William C. Schwab provided constructive reviews of the manuscript.  This survey was originally published as US Geological Survey Geologic Investigations Map I-2589.

FIGURE CAPTIONS

Figure 1. Location map showing study area in Long Island Sound off Norwalk, Conn. And major onshore moraines.. Also shown are the locations of other sidescan sonar surveys completed to date as part of this series[Milford--Twichell and others (1995); Hammonasset--Poppe and others (1997); New London--Poppe and others (1992) and Moffett and others (1994)]

Figure 2. Bathymetric map of the study area. Contour interval 2 m. Depths have been corrected for tides, and are adjusted to mean sea level. Fine dashed lines are the tracks along which bathymetry and geophysical data were collected.

Figure 3. 3.5-kHz subbottom profiles showing the geologic units mapped in figure 3. Profiles A and B show the acoustically prolonged facies inferred to be glacial drift; profiles B and C show the moderately prolonged facies inferred to be glaciodeltaic deposits; profiles A, B, and D show the acoustically transparent facies inferred to are marine deposits of Holocene age, and profile D shows the gas-charged marine deposits. Profile locations are shown in figure 3. VE, vertical exaggeration.

Figure 4. Geologic map compiled from the 3.5-kHz subbottom profiles showing the distribution of the geologic units defined by the subbottom profiles. Contour lines show the thickness of the marine deposits of Holocene age; contour interval is 5 m. Heavy lines indicate locations of profiles A-D shown in figure 2.

Figure 5. Sidescan sonar mosaic of the study area. Dark tones on the image represent areas of low acoustic backscatter and light tones represent areas of high acoustic backscatter.

Figure 6. Interpretation of the sidescan sonar mosaic. The patterns defined in the explanation are discussed further in the text.

Figure 7. Map showing the distribution of surficial sediments and the locations of the sample sampling and photographic stations . Interpretation of the sediment distribution is based on data from these stations, on tonal changes in backscatter on the sidescan sonar image, and on the bathymetry. The different patterns are discussed in the text.

SIDESCAN-SONAR MOSAIC

The sidescan-sonar mosaic is a digital image with a 1-m pixel size ( a pixel represents one meter on the sea floor).  The 15Mbyte digital image is available for viewing as a graphic in "tif" format with its associated Arcview geo-reference world file.