(Reviewers' comments received 4/98. Revisions in progress.)

Authors:

G.R. Cutter, Jr

R.J. Diaz

Virginia Institute of Marine Science

Project Manager

Carl H. Hobbs, III

Virginia Institute of Marine Science
 
 

Prepared under MMS Cooperative

Agreement 14-35-0001-3087 through

Virginia Institute of Marine Science of the

College of William & Mary
 
 

Introduction

The issues of coastline protection have become increasingly critical as erosion and coastal sediment transport have significantly altered or even eliminated ecologically and recreationally important coastal habitats. Increased public use of beaches, development of coastal lands, and preservation of the limited and sensitive coastal ecosystems have all lead to the need for beach nourishment as a means of stabilization and protection. The sand resources suitable for economical beach nourishment are usually located in the near shore coastal zones adjacent to the project areas. However, ongoing and planned beach nourishment activities along the coast of Virginia required sands from federal waters beyond the three mile line. The U. S. Mineral Management Service controls these sand resources and formed Cooperative Agreement (# 14-35-001-30807), "Environmental Studies Relative to Potential Sand Mining in the Vicinity of the City of Virginia Beach, Virginia" with the Virginia Institute of Marine Science and Old Dominion University in order to assess impacts of sandmining. Task 1 of this Cooperative Agreement ("Benthic Habitat Mapping and Evaluation of Existing Benthic Resources") involved benthic surveys of the region conducted by V. I. M. S. using sediment profile imaging and bottom grab samples.

Environmental concerns which arise in connection with proposals to excavate or mine sand from those areas identified as suitable for beach nourishment focus on potential ecological impacts associated either directly or indirectly with:

1. Removal or dredging of the sand from near coastal areas

2. Placement of the sand on the beach.

The configuration and location of the borrow site and the methods of handling the dredged material can be an important determinant in the level of impact. The level of potential impact would vary as a function of the characteristics of the material to be dredged, the exposure to currents and wave action, and the benthic resources (Thompson, 1973; Tuberville and March 1982; Hobbs, et al. 1985; Schaffner and Hobbs, 1992). Of the two geographic areas of concern associated with any beach-nourishment project, the source or borrow area and the beach area being nourished, we are focusing on the offshore source or borrow sites. In particular, the three sites off of Virginia Beach (Fig. 1) that will be used to nourish Virginia coastal beaches will be studied as a model project for evaluating environmental concerns.

Benthic habitats and non-commercial biological communities offshore Virginia were surveyed 1996 and 1997 in the vicinity of potential sandmining activities, where borrow areas had been identified and in regions of possible future interest . Benthic surveys were conducted semi-annually, during which sediment profile imaging (SPI) and standard bottom photographic camera systems and Smith-MacIntyre grabs were deployed.

SPI and standard photographs allow relatively rapid determination and assessment of benthic habitat characteristics and capability for broad areal sampling coverage. Grabs allow detailed determination of benthic biological community characteristics. Together, SPI and grab sampling provide complementary data which are capable of forming the basis for resource maps. Grab data may serve as the basis for confirming inferences made about biological and physical habitat characteristics using SPI data, and SPI data may be used to produce habitat coverage maps which should represent the potential limits of biological community manifestations.

SPI and grab data allow mapping of substrate types, biological community characteristics and functional aspects of the communities, delineation of habitat spaces, and determination of spatial heterogeneity of habitats and resources (Bonsdorff et al., 1996). Spatial and temporal patterns of habitats and community characteristics and the local and regional water flow patterns determine benthic community response to disturbance events such as sandmining, and therefore are important to the activities proposed off Virginia. This study provides determinations of benthic habitat types, spatial extent of substrate properties and habitats, benthic secondary production and biological community characteristics off the Virginia coast in and around areas where sandmining is proposed.
 
 

Study Area

The study area offshore Virginia extended from just inside the three-mile line to approximately 10 miles offshore, and from the latitude of the southern shore of Chesapeake Bay mouth (36.925° N) to a few miles south of Sandbridge, VA (36.675° N) (Figures 1 and 2). Within this broad region, smaller regions of interest were sampled at higher spatial densities during spring and fall 1996 and fall 1997 (Figure 2). Spring and fall 1996 sampling was done using three sample grids, one off Virginia Beach (northwest grid), one to the northeast of that (northeast grid) off the Chesapeake Bay mouth, and one off Sandbridge (southern or Sandbridge grid) (Figure 3: SPI , and Figures 4 and 5: grab). The entire region was sampled at lower spatial density during the spring 1997 deployment. The 1996 study areas covered approximately 30 NM², the spring 1997 study area covered approximately 60 NM², and the fall 1997 study area covered approximately 10 NM². The fall 1997 deployment involved a very high density sample coverage in the vicinity of Sandbridge shoal and the proposed borrow areas (Table 1).

Previous descriptions of the study region include geological and geotechnical descriptions of the Virginia inner continental shelf have been done by Williams (1987) Berquist and Hobbs (1988); Kimball and Dame (1989); and biological descriptions by Ranasinghe, et al. (1985), and Dauer (1981). The study areas generally encompassed various sand substrates typical of the inner continental shelf, and substrates composed of finer grain-size materials delivered to the shelf by the Chesapeake Bay excurrent plume.
 
 

Methods

Sediment Profile Imagery (SPI) and Smith-MacIntyre grabs sampling

Spatial mapping of benthic habitats was accomplished using sediment profile imaging and standard bottom photography. Three different spatial supports for the data were utilized. During 1996, three study grids were defined and divided into cells which were approximately 0.2 NM on a side. One grid extended offshore Virginia Beach and Dam Neck and was composed of 200 cells (numbered 1 - 200), one grid composed of 100 cells was located to the northeast of the Virginia Beach grid, and cells were numbered 201 - 300, and one grid was located offshore Sandbridge, in the vicinity of Sandbridge shoal and was composed of 100 cells (numbered 301 - 400) (Figure 3). (and figure 4). The prefix 96 was prepended to station numbers from 1996 (96001 - 96400) for presentation of the SPI data to provide consistency with the spring 1997 sample labels (Figures 3, 4, and 5 , and Table 1). Cells called 1 - 400 from spring and fall 1996 are synonymous with cells 96001 - 96400. Although it was unfeasible to revisit the exact point sampled upon subsequent sampling, revisited sample site numbers were maintained and positions recorded. Standard deviations of positions between seasons or years at a given sample site typically were on the order of 0.0005 degree latitude or longitude.

Grids were located in regions where sandmining activities were either planned (Sandbridge shoal area) or likely because of nearby localities (Virginia Beach, VA). Grids were composed of regularly spaced cell rows whose nearshore boundaries paralleled the Virginia Coast. Ten cells were defined for each cell row west to east across each 1996 grid. Because of the shape and arrangement of the grids, 1996 SPI samples were separated east to west neighbor by approximately 0.4 NM, northwest to southeast neighbor by approximately 0.25 NM, and northeast to southwest neighbor by approximately 0.2 NM.

The spring 1996 sampling cruise began May 14, 1996 upon the R/V Bay Eagle (V.I.M.S) and was halted by weather May 15, 1996. Completion of the spring 1996 sampling effort occurred aboard the R/V Bay Eagle (V.I.M.S), June 4 - 7, 1996. The fall 1996 cruise began October 21-22 aboard the R/V Langley (V.I.M.S) and November 4 - 6, 1996 aboard the R/V Bay Eagle.

A broadly spaced sample coverage was implemented for the spring 1997 deployment. A single staggered grid extending from Cape Henry to south of Sandbridge, Virginia was divided into grid cells measuring 1 NM per side. SPI samples were taken at each grid cell centroid (Figure 6) , and grab samples were taken at a random subset of the stations, in addition to revisited locations sampled 1996 (Figure 7). Several stations from the 1996 survey were revisited during the spring 1997 deployment, and SPI and grab samples were taken. The spring 1997 cruise occurred June 16 - 19, 1997 aboard the R/V Bay Eagle.

Fall 1997 sampling was concentrated on the proposed borrow areas in the vicinity of Sandbridge shoal. Borrow Areas A to the south and B to the north coincided with the crest of Sandbridge shoal (Figure 8). SPI samples were taken at six stations from eight north to south transects crossing or nearby Borrow Area A, and from several points along three transects which crossed into Borrow Area B perpendicular to each of the sides defining B (Figure 9). Grab samples were taken at a random subset of the stations, in addition to revisited locations sampled 1996 and spring 1997 (Figure 10). Fall 1997 sampling occurred October 6 and 7, 1997 aboard the R/V Bay Eagle. Table 1 lists all positions (in decimal degrees latitude and longitude) occupied during sampling and the type of sample acquired during the two year study, and Table 2 lists coordinates for the corners of the proposed borrow areas off Sandbridge. All samples taken in the vicinity of the proposed borrow areas 1996 and 1997 are labeled, along with the boundaries of the borrow areas, in Figure 11.

SPI samples were taken at points defining by the centroids of grid cells. SPI samples were taken at every other cell centroid point and were staggered from row to row, odd numbered cells visited one row, even numbered cells the next. At each station (grid cell centroid), the SPI camera system was deployed twice. If there was uncertainty about camera function or success, additional drops were made. During the spring 1996 cruise a Benthos model 3731 sediment profiling camera was used until it malfunctioned. From then on and throughout the next three cruises, a Hulcher model Minnie sediment profiling camera was used attached to a Benthos stainless steel frame. After deployment, depth of prism penetration and frame count was recorded. Camera tests were done periodically to ensure proper function. 350 pounds of lead and steel were used to weight the camera prism during May 14 and 15, and 450 pounds of lead and steel were used for all of the subsequent deployments. Color slide film (Fujichrome 100 professional) was used in the cameras.

Smith-MacIntyre grabs were taken from randomly selected cells from each grid, following a stratified random sampling design. 150 pounds of lead were used to weight the grab in order to produce consistently deep bites into the bottom. Grabs were only accepted if over half the volume of the grab was filled with sediment and the sediment surface was preserved intact. Grab success rate was high, confounded only when large gravel or intact shells wedged the grab jaws open. The sediment surface of these inner shelf sands and muds was preserved very well by the grab.

The top millimeter of sediment in grab samples from spring and fall 1996 were scraped off using a flat blade, and stained and refrigerated for analysis of living foraminifera and ostracoda (Cronin, et al. 1998). Subcores (circular 10 cm diameter and 10 to 15 cm deep) were also taken from 1996 grab samples, and preserved in formalin for later analysis of meiofaunal and macrofaunal communities. The rest of the volume from 1996 grab samples and all the volume from 1997 grab samples were washed upon 500 mm sieves aboard the vessel just after they were collected. Residue upon each sieve was stored in a cloth bag and preserved in 10% formalin solution and contained in 5 gal. buckets for transport to the laboratory.
 
 

Laboratory processing and analyses

Sediment profile images were analyzed by visual counts and measures of sedimentary and biogenic features in images projected upon a calibrated grid. Measurements were made of SPI prism penetration depth (PEN), average depth of the apparent color redox potential discontinuity (RPD), sediment surface relief (Sed. Rel.), relief type, sediment type, epifaunal presence and type, presence of tubes at the sediment-water interface, amount of biogenic shell material present; number, type and depth of infauna visible; number, depth and type of water filled infaunal feeding voids present (whether surrounded by anoxic or oxidized sediments); number and depth of gas voids present; and number and type of infaunal burrow structures present Tables 3, 4, 5, and 6. Notes were also made concerning any unusual features encountered during SPI analysis. Details on analyses and more extensive description of SPI parameters may be found in Rhoads and Germano (1982) and in Diaz and Schaffner (1988).

SPI parameter determination and delineation of habitats have been shown effective and comparable to sediment core sample data habitat delineation (Bonsdorff, et al., 1996) when samples have come from widely varying habitats and distinct geomorphic regions. Mapping the individual SPI parameters as well as biological community characteristics and production for this study was done for the inner shelf, allowing comparisons of SPI - grab delineation capabilities within a relatively uniform geomorphologic region. For habitat mapping, SPI feature types were designated biological or physical if features from either origin composed most of the SPI image. If both feature types were present in approximately equal amounts, feature type was designated combination.

In the laboratory, grab samples were processed to obtain secondary production estimates and organismal densities and biomasses. Organisms were sorted into major taxa and enumerated. Samples which retained a large amount of sand which would not pass through 500 mm sieves were elutriated and the organisms then extracted for sorting. Processing for secondary production calculations involved resieving the sorted taxa through a series of sieve sizes (6.3, 3.35, 2.0, 1.0, and 0.5 mm), then counting and weighing the organisms in each size fraction. Counts and biomass were converted to m^-2 by multiplying by 11.1.

Production was calculated using the technique described in Brey (1990). Total wet weight per sample taxa per size class was converted to mean individual weight per size class using the number of individuals counted. Wet weights were converted to ash-free dry weights (AFDW) using the conversion factor of Waters (1977):

AFDW (g) = Wet Weight (g) * 0.152.

Combined weights from all size classes allowed determination of biomass and production for each taxa per sample. Mean individual weight and total biomass per square meter for each major taxa and size fraction allowed estimation of secondary production using the multiple regression models of Brey (1990) including different coefficients for each taxa. When very small biomasses were present, production estimates were unrealistically high due to the limited range for which Brey's (1990) model applies. Therefore if mean annual biomass was estimated to be less than 0.02 g m², the value was excluded and production was not calculated for that observation. Production estimates reported are from 1996. Calculation of production for 1997 is in progress.
 
 

Maps

Maps representative of SPI parameters, habitat types, and biological densities and secondary production estimates were produced using the SPI and grab data. For each parameter mapped, the legends are consistent between maps from different sample dates, easing comparison. Interpolation and contouring of SPI and grab data was done by an inverse distance weighted (IDW) squared, nearest 12 neighbor method using Arcview for Windows NT. For habitat maps, the IDW squared method was used, but with only one neighbor to prevent generation of apparent intermediate habitat classes by interpolation.
 
 

Results

SPI analysis

SPI analyses data for spring and fall 1996 and spring and fall 1997 sampling efforts are presented in Tables 3, 4, 5, and 6. Maps constructed using these data are addressed hereon. While most parameters for 1997 were not plotted as maps, the data are available in Tables 5 and 6. SPI prism penetration patterns were similar between the seasons and years. Deepest penetration depths occurred in the northwestern part of the study area. Shallowest penetrations occurred in the northeastern part of the study area. Patches of shallow penetration were observed in the vicinity of Sandbridge shoal, and patches of deep penetration were observed just to the east and west of the shoal (Figures 12, 13, and 14). Spring 1996 SPI prism penetration depth ranged from < 5 to > 20 cm. Deepest penetrations (>10 to 20 cm) in the spring of 1996 were in the northwest portion of the study area (Figure 12). Fall 1996 penetration was also deepest in the northwest part of the study region, some exceeding 20 cm (Figure 13), and also in parts of the study area off Sandbridge. Spring 1997 sampling revealed a similar pattern, with deep penetration in the northwest, shallow penetration in the northeast, and patches of deep and shallow penetration off Sandbridge (Figure 14). SPI prism penetration within the proposed borrow areas ranged was generally within the 5 to 10 cm range (Figures 13 and 14).

Sediment-water interface (SWI) relief (surface relief) ranged from < 1 to 10 cm from spring 1996 through spring 1997 (Figures 15, 16, and 17). Lowest relief (smoothest surfaces) was observed in the study area off Sandbridge, just inshore from Sandbridge shoal and the proposed borrow areas, and in the northeastern part of the study area, as well as in patches within the study region off Virginia Beach. Highest sediment-water interface relief was observed in the northwest portion of the study area and off Sandbridge within borrow area A (Figure 16). Most of the images in the study area from spring and fall 1996 and spring 1997 revealed SWI relief of 1 to 2 cm. The area between the northwest and northeast parts of the 1996 study area was interpolated to have relief from 1 to 2 cm using spring 1996 data, and 2 to 4 cm using fall 1996 data (Figures 15, 16, and 17). The points sampled spring 1996 and 1997 within this zone revealed lower relief (< 1 cm) suggesting that the high relief observed from cell 281 fall 1996 was a temporary artifact.

The apparent color redox potential discontinuity (RPD) layer depth ranged from < 1 to > 10 cm during spring and fall 1996 (Figures 18 - 19). RPD depth measurements from 1997 samples has not yet been completed. Spring 1996 sampling revealed a large area off Virginia Beach and a point within borrow area B with RPD depths of 5 to 10 cm. In the northeastern part of the study area, RPD depths were nearly all between 1 to 3 cm. Shallowest RPD depths were observed off Sandbridge, just inshore of Sandbridge shoal and the proposed borrow areas (Figure 18). Fall 1996 sampling revealed slightly lower RPD depths in the northwestern part of the study area, off Virginia Beach, and similar RPD depths in the northeast. RPD depths in the study area off Sandbridge were deepest fall 1996, most between 5 to 10 cm and one > 10 cm (Figure 19).

Infaunal tubes were present in many of the SPI images during spring 1996, with highest numbers of tubes in the northwestern part of the study area (> 25), and very few in most of the northeastern-most part of the study area (Figure 20), except at a few stations where there were high numbers (10 to 25). Spring 1997 SPI images revealed lower numbers of tubes over most of the study area. Most of the images had no tubes visible at the sediment-water interface (Figure 21). There were some tubes visible in images from the northwestern part of the 1997 study area, as seen also in spring 1996. However, the high numbers observed in the northwestern part of the 1996 study area were not apparent in spring 1997 images. The absence of tubes in numbers > 25 per SPI image from spring 1997 may indicate changes in community composition or limited spatial patches dominated by tube-dwelling infauna, perhaps not as extensive in coverage as suggested by the interpolation displayed in Figure 20. Most of the tube building infauna were polychaetes, from the families Maldanidae (Clymenella, Asychis, Euclymene, Maldanopsis), Ampharetidae (Asabellides oculata), or Onuphidae (Diopatra cuprea). Species were confirmed by grab samples, but tube structures were usually distinct enough to discern tube types in SPI images. Grab samples produced many other species as well which were not detected in SPI images (See section on grab sample analysis below).

Sediment types in the study region were primarily sands from -1 to 4 phi, but some muds (4 to 8 phi) were also present. Muds were prevalent in the northwestern part of the study area and in patches across the region (Figure 22). The muds were typically silt to clayey silt, and sands ranged from very fine sands to coarse sands and granule (Figure 23). Fine sands (2 to 3 phi) were most common throughout the study area. The spring 1997 sampling grid did not encounter as many silty sediment patches as the 1996 sampling. Sediment grain size and alkalinity determined from spring 1997 grab samples are listed in Table 7.
 
 

Habitats classifications from SPI images

The variety of sediment types and habitat types were apparent in images from spring and fall 1996 when more dense sampling grids were employed. Nine gross habitat classes were identified from a set of sixteen habitats initially identified using 1996 SPI images. Habitat classes were labeled A - I, and are summarized in Table 8. SPI images representative of the nine habitat types are displayed in Figures 24 - 32. Habitat determinations from 1997 data are not yet complete. Habitat maps made using the nine habitat type classes reveal the diversity of bottom types across the study area encountered spring 1996 (Figure 33) and fall 1996 (Figure 34).

Seven of the habitat types were identified in SPI images from spring 1996. The most common and extensive habitat types spring and fall 1996 were (class C) silty-sand to very fine sand sediments with both biological and physical characteristics, and (class D) fine sand sediments with primarily physical features (Figure 26). Overall, across the entire study region, 73 of the 157 cells from which determinations were made using SPI images from spring 1996 were habitat class C, and 31 were class D. Eight of the nine habitat types were identified from fall 1996 SPI images. For the entire study region, 100 of 191 cells revealed habitat class C, 39 were class D, and 28 were class F (Table 9). The entire grid block in the northeastern part of the study area (cells 96201 - 96300) was found to have only habitat class C (combined biological and physical fine sands) spring and fall 1996 (Figures 33 - 34).

Biologically dominated silts (class A) and fine sands (class E) were also present across much of the study area, though slightly fewer during fall 1996 sampling than spring 1996. Physically dominated silt sediment habitats (class B) were present only during fall 1996 at three stations off Virginia Beach. Physically dominated medium sand and shell sediments (class F) were present most in the study area off Sandbridge, and were more extensive fall 1996. Biologically dominated medium sand and shell sediment habitats (class G) were apparent at some of the sample stations off Sandbridge, just outside the proposed borrow areas, and also at several stations in the study area off Virginia Beach. Physically dominated coarse sands to gravel sediments (class H) were encountered only off Sandbridge, just inshore of the proposed borrow areas spring and fall 1996, and at one station off Virginia Beach spring 1996. Apparently transitional substrates, composed of coarser grain-size sediments layered over finer grain-size sediments (class I) were encountered fall 1997 at three stations off Virginia Beach, and at one station off Sandbridge (Figures 33 - 34).

Grouping the habitats in terms of dominance by biological or physical features, or combined interaction of biological and physical features, without respect to sediment type reveals broad and apparently continuous regions of either physically or biologically dominated substrates spring 1996 (Figure 35). During spring 1996, biologically dominated habitats were prevalent in the northwestern part of the study area off Virginia Beach, physically dominated habitats are prevalent in the vicinity of the proposed borrow areas off Sandbridge, and combination habitats were interspersed in those two regions and ubiquitous in the northeastern part of the study area (Figure 35). Similarly, fall 1996 SPI images revealed biologically dominated habitats in the northwest part of the study area and in the study area off Sandbridge, however in apparently non-continuous, smaller patches than during spring 1996 (Figure 36). Habitats with combined biological and physical interaction were present in the study areas off Virginia Beach and Sandbridge and pervasive across the northeastern grid. Table 10 lists positions, cell numbers and habitat classifications.

Of the seven habitat types identified using SPI images in the northwest sample grid off Virginia Beach spring and fall 1996, four were biologically dominated or combination biological/physical substrates, and three were biologically dominated or combination during fall 1996. In the northeastern sample grid 1996, only one habitat type was identified during spring or fall, and that had combination biological and physical features. In the sample grid off Sandbridge, five habitats were identified spring 1996, and of those two were biologically dominated or combination. Similarly, in fall 1996, in the sample grid off Sandbridge, two of seven habitats identified were biologically dominated or combination.

Secondary production

Total benthic secondary production calculated using the technique of Brey (1990) ranged from < 1 g m^-2 yr^-1 to > 50 g m^-2 yr^-1 (Figure 37). Low (< 5 g m^-2 yr^-1) to high (> 25 g m^-2 yr^-1) production was observed in the northwestern part of the 1996 study area off Virginia Beach. Low to moderate (5 to 25 g m^-2 yr^-1) production was found in the northeastern sample grid and in the study area off Sandbridge in the vicinity of the proposed borrow areas. The high total production values correspond to high combined production by molluscs (Figure 38), and annelids (Figure 39), and to a lesser degree by crustaceans (Figure 40). Mollusc production was high at one site in the northwestern sample grid off Virginia Beach, and low to moderate throughout the rest of the study area (Figure 38). Annelid production was high at a site just west of where mollusc production was highest, in the central part of the northwestern sample grid off Virginia Beach, and was low to moderate elsewhere (Figure 39). Crustacean production was low throughout the study area , but relatively higher in the northwest sample grid and at one site in the study area off Sandbridge, within proposed borrow area B (Figure 40). Miscellaneous taxa, composed primarily of echinoderms and cnidarians, had very low production across the area. However, some patches of relatively higher production by miscellaneous taxa corresponded with higher production by molluscs, annelids and crustaceans (Figure 41).
 
 

Sandbridge study area analysis, including proposed borrow areas

Habitat types in the study area off Sandbridge were mostly physically dominated fine to medium sands along Sandbridge shoal crest in spring 1996 (Figure 42) and combination biological-physical silty sands or biologically dominated silts around the shoal. Habitats were similarly distributed fall 1996 (Figure 43) but slightly fewer biologically dominated muds were encountered, one sample revealed silt in part of borrow area A. Another site had sediments which had apparently undergone recent transition, with a coarser grain-size sediment layer overlain upon clayey silt (Figure 43). Both spring and fall 1996 SPI images reveal primarily physically dominated habitats throughout most of borrow area B, and approximately half of borrow area A (Figures 44 - 45). Biologically moderated habitats were found to the west (inshore), east, and south of the proposed borrow area (Figures 44 - 45).

Total community secondary production from 1996 within the study area off Sandbridge was low to moderate, and relatively highest in at one site near the southern boundary of borrow area A and to the west of borrow area A (Figure 46 with cell numbers labelled, and Figure 47). Production by molluscs was very low to low throughout the study area off Sandbridge, except at a station to the west where production was moderate (Figure 48). Polychaete production in the study area off Sandbridge was very low to low. Lowest polychaete production was found in borrow area B (Figure 49). Crustacean production was very low (< 1 g m^-2 yr^-1) at all except three stations where production was low (1 - 5 g m^-2 yr^-1) in the study area off Sandbridge, 1996 (Figure 50).

Total infaunal densities ranged from < 100 to > 2000 m^-2 in the study area off Sandbridge. Highest overall densities (> 2000 m^-2) were found to the west and in the center of borrow area B (Figure 51). Lowest overall densities were observed in samples taken just to the south and west of the sample with high density in the center of borrow area B. Molluscs were found in highest densities at one cell to the west of the proposed borrow areas,. and were found in higher densities inshore of the borrow areas within the study area off Sandbridge (Figure 52). Polychaete densities were high inshore of the proposed borrow areas and at one cell in the center of borrow area B and one cell to the east of borrow area A (Figure 53). Crustacean densities in the study area off Sandbridge, 1996 were relatively high (> 1000 m^-2) in borrow area B, and low to moderate elsewhere (Figure 54). Comparisons of apparent general habitat type determined using SPI to benthic secondary production (Figure 55) and faunal densities (Figure 56) reveal that total community production and densities for all taxa agree fairly well with the habitat delineations made using SPI.
 
 
 
 

Grab sample analysis

Benthic Community Species Composition

At total of 119 taxa were identified from 13 of the Smith-MacIntrye grabs collected in 1996 (Tables 11 - 12). Half of the top 14 taxa in terms of occurance and abundance were polychaetes. The other half were one representitive each from the amphiods, decapods, bivalves, nemerteans, tanaids, echniderms, and chordates (Table 13). The distribution of species among the taxa was similar to other benthic studies from the region. Table 14 compares our study to that of Dauer (1981). While Dauer's data are for about 3.5 times as much bottom area as ours, the total numbers of species from both studies are close. Dauer's (1981) survey of the benthos at the proposed Norfolk District COE open water disposal site, about 15 miles East of the entrance to Chesapeake Bay is the closest historical data set to our study area. The Dauer study was at the same latitude as the northernmost extent of our study area, and about 8 miles to the east, but was at similar sediment types and depths.

Overall, the community composition within our study area was typical for sandy shallow continental shelf habitats. Detailed studies by Boesch (1979) off the coasts of New Jersey and Maryland, Maurer et al. (1976) off Delaware, and Day et al. (1971) off North Carolina reported simialar species composition for similiar depths and sediment types. While there were many differences in the species composition, several species and taxa were consistently reported by these studies, such as, Spiophanes bombyx, various Nephtys species, Tellina agilis, Magelona rosea, Aricidea spp., Spio setosa, Nassarius trivittatus, Ampelicsa verrilli, Unciola irrorata, and Mellita quinquiesperforata. All of these are know to be high salinity sand species.

Benthic Community Size Spectra

The size distribution of the benthos, both biomass and number of individuals, is an important factor in determining the potential food resources available to bottom feeding fish and crabs (Diaz and Schaffner 1990, Edgar 1990) and was the data used in calculation of secondary production. Size spectra of the benthos (Class intervals of 0.5 to 0.9, 1.0 to 1.9, 2.0 to 3.3, >6.3 mm) were determined from the grab samples and followed a pattern typical for marine communities, where the highest biomass was in the larger size classes and highest number of individuals was in the smaller size classses (Figures 57 - 58). Overall for June and November 1996 data, 15 % of all individuals and 81 % of the wet weight biomass were in the larger biomass size fractions, 3.35 and 6.3 mm size classes. The taxonomic composition of the larger biomass size fractions spanned a broad range with only the total biomass of anemones and amphipods being less than 50 % in the larger size fractions (Table 15). In terms of numbers, only echinoderms, bivalves, and chordates had 50 % or greater of their total abundance in the larger size fractions (Table 15).

Overall, total biomass about doubled between June to November, going from 4.1 to 7.7 g wet wt m² (Table 16). Annelids, the dominant taxonomic group in biomass, numbers and trophically, are typical of this trend and averaged about 13 g wet wt m² in June and 28 in November. The modal biomass size fraction for annelids in both June and November was 3.35 mm. Maldanid and Nephtid polychaetes were the predominant families that accounted for most of the 3.35 mm size fraction biomass.

Total abundance declined from June to November being 2350 and 1850 ind. m2, respectively. Again, annelids were typical of this trend declining from 960 to 910 ind. m2 with the modal size fraction being 1.0 mm (Table 16). This increase in biomass and decline in abundance are likely due to post settlement seasonal growth and mortality of macrofauna.
 
 

Discussion

Deep SPI prism penetrations in the northwestern portion of the study area coincided with the Chesapeake Bay plume deposits, composed of finer grain-size sediments. Deep prism penetration in other parts of the study area appeared to coincide with the depositional environments induced by large-scale bottom features such as Sandbridge shoal. The shoal consisted of coarsest sediments along the crest and was surrounded by a rim or at least patches of finer sediments, silts to clayey silts. Biological activity and numbers of macro-infaunal organisms and structures were high in these finer sediments, as apparent in the sediment profile images (Figure 45) and in the grab samples (Figure 51).

Relief was generally caused by sediment bedforms, primarily smooth-crested wave-orbital ripples. Bedform heights were observed to be 1 to 2 cm and as demonstrated by the surface relief, or SWI relief, measurements. Larger bedforms were observed in the study area off Sandbridge, within borrow area A, and in the northern part of the study region off Virginia Beach.

Habitat mapping using SPI and sediment grab sampling was very effective for covering large regions such as the study areas off Virginia Beach. Relative resource evaluations would then be made using parameters measured from SPI and grab samples. This is in contrast to indices, such as the organism-sediment index (OSI) (Rhoads and Germano 1982) which includes only certain SPI data, and benthic index of biotic integrity (B-IBI) (Weisberg, et al. 1997) derived only from grab data. Such indices are of questionable value for assessing resources over broad areas. The OSI, for instance, relies upon the RPD depth without considering the sediment type or grain size. RPD depth, however, may be influenced significantly by grain size. Thus OSI may be useful for comparing habitat properties within a relatively uniform sedimentary environment, but for comparison across sediment type regions, it is confounded by not accounting for collinearities in habitat parameters. Thus, interpretation of the SPI and grab data is necessary to provide a habitat resource value assessment, unless a new index is developed, or one is modified, which accounts for inter-regional parameter behavior.

Using only two parameters determined from SPI analysis, a rapid assessment of gross habitat type can be made. Using just an assessment of sediment grain size class (coarse to very fine sand, coarse to fine silt, and clay, and mixtures) and determination of prevalence of biological features or physical structures provides a simple method for initial delineation of benthic habitats. These may be further refined by determination of dominant fauna or surficial geological characteristics, both available using SPI data or inferences about features present in SPI images confirmed by grab data.

Benthic secondary production for 1996 was high (> 25 g m^-2 yr^-1) in the northern portion of the northwest sample grid off Virginia Beach and low (< 5 g m^-2 yr^-1) to moderate (5 - 25 g m^-2 yr^-1) throughout the rest of the study area (Figure 37). In the northwest sample grid, off Virginia Beach, the high production calculated using grab sample data corresponded to regions which were identified using SPI images as biologically dominated fine sand during spring 1996 (Figure 33), but as physically dominated fine sand fall 1996 (Figure 34). Habitats were identified as physically dominated at some locations using SPI images, but relatively high production was found using grab data. In some sediments, especially non-cohesive sands, biological features may not persist, and fauna may be inconspicuous.

Therefore, dependent upon substrate characteristics, more or less effort with the different sampling techniques may be necessary. Accurate habitat mapping then may require initial reconaissance and subsequent allocation of sampling effort. Some habitat determinations using SPI may not have agreed with observed high or low production or biological densities because of small spatial-scale variabilities since grabs were not necessarily acquired on the exact spot where images were taken. General agreement is good between the two sampling techniques for gross characterizations of habitat and for biological resource assessment. The agreement between the interpolations made using SPI and grab sampling is encouraging, especially considering that each was done upon a different support, the sample spatial structure.

Discriminant analysis using SPI and grab data should reveal whether the habitat determinations convey objective discretions. Evaluation of the pertinence of the 1996 determinations will be used to delineate habitats using 1997 data. In general, if SPI can be used to reliably relate habitat characteristics which allow prediction of benthic community attributes, then statistical spatial models may be constructed for habitats and resources with efficiency. Refined SPI and grab sampling strategies may accomplish regional benthic resource assessment and mapping with a reasonable effort.

Where persistence of SPI features would be confounded by lack of sediment cohesivity, we expect that habitat delineations will underestimate community attribute variability. However, SPI and grab data and maps for the region off the Virginia coast elucidate an overall more complex benthic system than was expected based upon the generalized descriptions of previous studies for the inner continental shelf off Virginia.
 
 

References