Back to NJ Grab sampling page

Does grain size determine acoustic backscatter on the New Jersey continental shelf? Sometimes!

J. A. Goff, H. Olson, C. L. Schuur, G. Moss, and J. A. Austin, Jr.

University of Texas Institute for Geophysics

In September, 1998, with JOI-USSAC Site Survey Augmentation funding, we collected ~300 sediment grab samples along the New Jersey "Mid-Atlantic Transect," in the vicinity of three proposed jack-up drilling sites that have been surveyed with both swath mapping (Goff et al., 1999) and multichannel seismic reflection profiles (survey by G. Mountain et al.) (Figure 1). The samples have been analyzed for grain-size distribution and their micropaleontology will be investigated this summer. Ultra-high resolution chirp seismic reflection data will be collected in June of this year. The purpose of the work is threefold: (1) to provide additional site characterization for future placement of jack-up rigs; (2) to "ground truth" the sonar backscatter data; and (3) to better our understanding of bedform evolution in inner- and mid-shelf settings. We have intermittently established an unprecedented level of correlation between grain size and backscatter intensity, generally supporting our hypothesis that bedforms in ~20-45 m water depth are responsive to the present day hydrodynamic environment, although the larger ridges may be partially relict from an interval of lower sea level. However, the observed correlation of grain size to backscatter can be degraded if the sediments include even a few extra weight percent of the largest grain sizes (>4 mm), typically shell hash.

Ground Truthing Backscatter

Sidescan backscatter has long been used as a qualitative seafloor mapping tool, a "sonic photograph" of the seafloor. Backscatter intensity variations can arise from bottom topography, from varying acoustic properties of seafloor material, or both. The holy grail of acoustic swath mapping is to use backscatter more quantitatively; that is, to transform the image into estimates of seafloor properties (e.g., grain size, density, composition). "Ground truthing" the sidescan data, i.e., sampling and analyzing seafloor material within the survey area to correlate backscatter properties to bottom properties, is the key to the quantitative approach. In a purely sedimentary seafloor environment, where variations in backscatter and sediment properties can be subtle, this is a difficult task because there are many sources of uncertainty: (1) navigational inaccuracies of sonar and sampling surveys, (2) noisiness of backscatter data, (3) difficulties in estimating in situ sediment properties from samples, and (4) the complexity inherent in ascertaining whether a sample is representative of a larger backscattering region (pixel size is ~5 m by 5 m).

Grain size is thought to be a principal determinant of backscatter intensity for sedimented seafloor. However, this assumption has only been demonstrated previously for sandy and coarser sediments, and then only fairly broadly (i.e., coarse vs. medium vs. fine sands; e.g., Davis et al., 1996). The correspondence between backscatter and mean grain size demonstrated in Figure 2 is remarkable because of the subtlety of the variations. Backscatter variations here fall within the pixel-to-pixel noise level; they are visible in profile only after careful filtering. Corresponding mean grain sizes range only from ~0.3 to 0.4 mm, entirely within the medium sand category. We can attribute part of this correlation to improvements in data acquisition techniques; i.e., consistent use of differential GPS navigation, large sample sizes (~300 g), and careful filtering of sidescan data. However, this level of correlation is not seen everywhere in our study area; in fact, in some locations on the New Jersey inner and middle shelf we can find no correspondence at all. Our best correlations appear when the grain-size distribution is unimodal, and particularly when the sediments are well sorted (low variance). Even a small extra percentage of large grain sizes (> 4 mm), commonly shell hash, degrades the correlation because, although backscatter is disproportionately affected by these larger grain sizes, their proportion is difficult to measure accurately without even larger and more areally extensive sample sizes. Nonetheless, our results indicate that backscatter from sandy shelf sediments is predominantly responsive to variations in grain-size distribution. This information will be important for future siting of jack-ups on the New Jersey inner shelf.

Sand Ridge Evolution

Sand ridges are among the largest and most pervasive bedforms on the mid-Atlantic continental shelf, yet they are also the most enigmatic. Their puzzle comes from the fact that they are oriented obliquely (~20°-40°) to the direction of formative bottom current flow. Sand ridges are most vigorously active along the shoreface (e.g., Swift and Field, 1981), although there are five competing models for their formation (see Goff et al., 1999, for summary). The fate of ridges when sea level rises is also poorly understood: do they become "moribund," or do they continue to evolve, and if so, how? The implications go beyond understanding bedform evolution. In this sediment-starved shelf environment, reworking of transgressive sand deposits could alter the paleobathymetry derived from the micropaleontologic record by mixing nearshore and offshore faunas; this has potentially important implications for the interpretation of sea-level history from drilling records.

In the portion of the New Jersey inner shelf surveyed (Figure 1), there are two clusters of ~NE-SW-oriented ridges, each several meters high and ~1-3 km wide. Numerous smaller dunes are oriented ~N-S, generally <1 m in amplitude and <200 m in width. Bottom current directions in the surveyed portion of the inner shelf are evidently ~W (Vincent et al., 1981; Figure 1), a deviation from the ~S to SW, contour parallel currents more typically observed in the Mid-Atlantic Bight. In contrast, when sea level was lower and the shoreline was closer, currents through this region were likely constrained to be shore parallel, or ~SSW (Figure 1) as determined by regional contours. This change in current direction over the past ~10,000 years is important because it helps us to distinguish modern from relict bedforms.

The grain-size pattern over the dunes, larger on the E flanks (Figure 2), is consistent with formation transverse to the modern current because the eroding, upcurrent flanks should have a coarser residue. But are larger ridges also responding to the modern current? Their ~NE-SW ridge orientation is what we might expect if these had been formed oblique to a ~SSW paleoshoreline. However, their slope asymmetry contradicts this interpretation. Nearshore, the seaward flanks, in the lee of a SSW-directed alongshore flow, are steeper (e.g., Swift and Field, 1981). Offshore in our study area, landward flanks tend to be steeper (Figure 3), suggesting a response to a current direction from seaward. The backscatter and grain-size pattern over these ridges is complicated. In the displayed example (Figure 3), backscatter and grain size are generally higher on the seaward flank, which would support its interpretation as an upcurrent slope, but there is also a large grain size and backscatter spike at the base of the landward slope, which is a common feature seen in nearshore ridges (Swift and Field, 1981).

The slope break on the seaward ridge flank (Figure 3), another common feature of these ridges which gives them a somewhat trapezoidal cross section, is an added complexity. Similar shapes have been observed in sand waves in tidal regimes, and interpreted as resulting from alternation between primary and secondary current directions (e.g., Swift et al., 1978). Like the tidal bedforms, these ridges may have formed under the influence of more than one current direction. We hypothesize that they formed originally in a nearshore paleoenvironment under a ~SSW-directed flow when sea level was lower, and have subsequently been heavily modified, but not entirely deconstructed, at their present water depth by the modern, ~W bottom currents. Forthcoming chirp seismic data, which will image the internal structure of the ridges, should test this hypothesis.

Davis, K. S, and others., 1996. Acoustic backscatter and sediment textural properties of inner shelf sands, northeastern Gulf of Mexico. Geo-Mar. Lett., 16: 273-278.

Goff, J. A., and others, 1999. High resolution swath sonar investigation of sand ridge, dune and ribbon morphology in the offshore environment of the New Jersey margin. Mar. Geol. in press.

Swift, D. J. P. and Field, M. E., 1981. Evolution of a classic sand ridge field: Maryland sector, North American inner shelf. Sedimentology, 28: 461-482.

Swift, D. J. P., and others, 1978. Evolution of a shoal retreat massif, North Carolina shelf: Inference from areal geology. Mar. Geol., 27: 19-42.

Vincent, C. E., Swift, D. J. P. and Hillard, B., 1981. Sediment transport in the New York Bight, North American Atlantic shelf. Mar. Geol., 32: 369-398.