Introduction

The processes controlling sediment deposition on vegetated tidal salt marshes are not well understood, yet rates of sediment deposition are of fundamental importance to marsh ecology and hydrology. On the Atlantic coast of the United States, relative sea level is rising at a rate of approximately 2 mm per year. It is therefore important to understand the mechanisms by which marshes are maintained at an elevation that is adequate for sustaining the marsh ecology. In addition, most wetlands are sinks for fine grained sediments. In polluted environments, contaminants may be carried into the wetlands with the fine grained sediment, and it becomes important to have an accurate understanding of where the sediment primarily deposits.

Several researchers have attempted to determine whether marshes on average import or export sediment by measuring sediment flux in tidal creeks (Stevenson et. al. , 1988). Both landward and seaward fluxes are large, however, and the uncertainty related to determining both numbers makes it difficult to determine accurately the rate and direction of the net transport. Furthermore, tidal creek measurements only describe net marsh flux as long as the flow is within the creek banks. As soon as the marsh surface floods, the majority of the flow occurs outside of the creeks. In a review, Stevenson et. al. , 1988 concluded that marshes on the Atlantic coast of the United States in general were eroding and that sea level rise would eventually lead to large areal losses of coastal salt marshes. In contrast, on the Eastern Shore of Virginia, long term deposition rates of 1-2 mm per year have been quantified using 210_Pb dating techniques, indicating that the marshes in this area are accreting at a rate comparable to the rate of sea level rise. The mechanisms that govern sediment deposition on these vegetated surfaces and their relative importance remain unclear.

The main objective of this study is to determine the primary controls on marsh sedimentary processes. Marsh sediment dynamics are expected to be closely related to the advective and diffusive properties of the flow on the marsh surface during flooding tides. Friction exerted on the flow by the marsh vegetation has a significant effect on the flow, and thus on sediment transport and deposition. Through measurements of flow velocity, turbidity, water surface elevation, and sediment properties, and combining this information with a mathematical description of the physical processes, the dominant sediment transporting processes on the marsh surface are being addressed. The secondary objective of the study is to quantify the relative importance of sediment deposition occurring during the regular tidal cycle versus during less frequent storm events. During storms, the marsh surface is often inundated for a longer time than during fair weather periods. For this purpose, a longer-term (~1-2 month) time series of measured suspended concentration on the marsh and multi-year records of water level and wind-speed are being used to determine the relationships among weather patterns, inundation period, sediment concentration, and deposition.

Experimental Setup

Our study site is a low salt marsh on the mainland side of Hog Island Lagoon, in the Phillips Creek area near Brownsville, VA, located on the Atlantic side of the Delmarva Peninsula (Figure 1). The tidal range in this area is 1-1.5 meters. The marsh, which is vegetated with Spartina Alterniflora , is flooded on most high tides. The marsh was surveyed to produce a detailed topographic map of the study site. We installed pressure transducers to measure water level on the marsh during flooding tides. These water level measurements were supplemented with a long term record from a nearby tide gauge maintained by the VCR/LTER and with data from a tide gauge maintained by NOAA at Wachapreague, VA, 20 km North of the site. A transect of 5 stations has been set up perpendicular to the adjacent tidal creek. (Figure 2). The transect extends 50 meters from the bank of the tidal creek to the marsh interior. Three stations were established within close proximity of one another on the creek bank where the largest gradients are expected. The remaining two stations were set up in the marsh interior, away from the creek bank. At each station we measured time series of turbidity, flow speed and flow direction during inundation events. Turbidity was measured using optical back scatter (OBS) sensors. Prior to applying the sensors in the field, the sensor response was calibrated to the sediment found at Phillips Creek Marsh and compared to measured sediment concentrations in water samples taken on the marsh during sampling. Velocity was measured in three dimensions using an Acoustic Doppler Velocimeter (ADV). Both instruments were connected to dataloggers which allowed sampling at high frequencies. Turbidity was measured every 3 seconds whereas velocity was sampled at 10 hz to allow resolution of the turbulent properties of the flow. The ADV measures velocity 5 cm below the position of its sensor head, which must be submerged. To sample 10 cm above the bottom (cmab), the instrument must be located 15 cmab. Thus sampling does not begin before the water level is 15 cm above the boundary. Measurements were made at one station at a time. We could not measure at several stations simultaneously owing to the number of instruments available. Meteorological conditions characterized by wind speed, wind direction and barometric pressure at the water surface have been obtained from NOAA's NDBC Buoy 44014. The buoy is located 50 km south of the study site, off shore from the mouth of the Chesapeake Bay.

Preliminary Results

Circulation on the Marsh Surface

Early in the study, the general circulation patterns on the marsh were determined by tracing the flow using fluorescent dye. These observations indicated that the flow on the marsh surface is perpendicular to the flow direction in the tidal creek. The flooding of the marsh occurs first from the small tidal creek on the north side of the marsh. As water level increases, a larger part of the flow enters from Phillips Creek. The direction of this flow is predominantly perpendicular to the creek bank, along the study transect. On the rising tide, water flows on to the marsh surface; on the falling tide flow direction reverses and water is drained from the marsh surface, still moving perpendicular to the creek bank.

Measurements of flow velocity made using an Acoustic Doppler Velocimeter (ADV) are shown in Figure 3. Figure 3 shows time series of velocity at three different locations, Stations 1, 3 and 5 along the transect. Because the stations are located at different elevations, peak water depth varies among stations: 0.55 m at Station 1, 0.4 m at Station 3, and 0.7 m and at Station 5. The different ranges in depth means that each station is flooded for a different period of time. While Stations 1 and 3 are flooded for approximately 4 hours, Station 5 is flooded for more than 5 hours.

All the measured velocities are low (less than 1 cm/s). In this environment, the largest velocities occur when the flow depths are lowest, at the beginning and end of the inundation period. The largest velocities are observed on the creek bank during the falling tide. In general velocities on the bank are higher than in the marsh interior. The velocities are lowest when the depth is greatest (slack tide). The velocities are larger during falling tide than during the rising tides. At both Station 1 and 3, more water passes the sampling point during the falling tide than on the rising tide. At Station 5 this comparison could not be made because the time series is incomplete. Tidal asymmetries suggests that water does not enter and leave the marsh along the same flow path.

Turbulent intensities (mean(u'^2),mean(v'^2) and mean(w'^2)) reflect the ability of the flow to maintain sediment in suspension. The presence of vegetation dampens turbulence, and the turbulent intensities decrease with distance from the tidal creek. The most dramatic decrease is between Station 1 and 3, within 7 meters from the tidal creek. The two horizontal components of turbulent intensitys are of similar magnitude but the vertical intensity is consistently lower (Figure 4).

Sediment Transport

A survey of the marsh surface has been used to compile a topographic map of the area. On the banks of the marsh, levees have developed to an elevation that is approximately 50 cm higher than the interior of the marsh. A 9-month tidal record from the LTER tide gauge has been used to determine the inundation frequency of different elevations on the marsh surface (Figure 5). During each tidal cycle the maximum tidal elevation was selected and a histogram and cumulative distribution was compiled of the 382 maximum tidal elevations. The cumulative distribution was superimposed on the marsh topography (Figure 6). The 50-cm elevation difference between the marsh interior and the banks becomes significant when inundation frequencies are considered. The lowest parts of the marsh were flooded on 95% of the high tides whereas the levees were only flooded on 45% of the high tides. In addition, the banks are flooded for a much shorter period of time than the marsh interior. This relationship between topography and inundation suggests that deposition on a marsh is an event-driven process where only the highest tides contribute new sediment to the marsh surface.

Time series of sediment concentration and water levels have been measured on the marsh. The variation of suspended sediment concentration at five stations along the transect is shown in Figure 7. The high frequency variation in concentration has been filtered out using a lowpass filter with a cut-off frequency of 15 minutes, so only the low frequency variation is included in the figure. The measurements were made during five tidal cycles, all of which were sufficiently high to occur on less than 15% of the tidal cycles. The highest concentrations are measured at Stations 1 and 2, but decrease to much lower levels at Stations 3, 4 and 5.

At Stations 1 and 2 the concentrations are high while the tide is rising. As the tide approaches slack water the sediment concentrations decrease and remain low for the rest of the tide. The concentration time series have been centered around the time of maximum tidal elevation. The creek bank measurements show that more sediment is brought to the marsh on the rising tide than leaves the marsh on the falling tide, indicating that the marsh is a depositional environment. It appears that concentrations measured at Station 1 decrease sooner than they do at Station 2, which may indicate that sediment is suspended on the creek bank and advected with the flow past Station 2. The dramatic decrease in concentration over the 4 meters between Stations 2 and 3 suggest that much of the sediment is deposited in this region.

In the marsh interior (Stations 4 and 5), the concentrations are constant and low throughout the tidal cycle. Although no large changes in sediment concentration occur, these stations are lower and are flooded for a longer period of time than the stations associated with the creek bank. We expect that vegetation plays a greater role in trapping sediment in this region.

Our measurements show that there is a relationship between suspended sediment concentration on the creek bank and maximum tidal elevation. The time series shown in Figure 7 are all measured during very high tides. If the tides are not high enough to produce overbank flow (70% of the tides are in this category), no significant amount of sediment is suspended. The period of a tide is always 12.5 hours and the higher a tide, the more rapidly tidal elevation changes, producing larger stresses on the marsh surface. It is therefore likely that sediment concentrations remain low during the lowest tides and only become significant when the tidal range exceeds a certain level.

Regular Tidally Driven Processes Versus Storm Processes

In Figure 8 an example is given of how tidal elevation is modified by wind direction. In the left column, the astronomical tide is compared to the actual tide during time when the wind direction is from the northeast, i.e. water is blowing from the Atlantic Ocean into the lagoon. During this time, water level is elevated by approximately 10 centimeters relative to the water level expected from the astronomical tides. In the right column the same comparison is made during a time when the wind direction is from the southwest and in this case the water level is 10 cm lower than the water level predicted by the astronomical tides.

It is clear that sediment deposition on the marsh surface is related to the frequency of tides that produce overbank flow. Tidal elevation is modified by meteorological conditions, and storms can both increase and decrease water surface elevations relative to astronomical tides. We have compared the astronomical (predicted) tides to measured water levels for a 5 year period between 1990 and 1995. An event has been defined to occur when the difference between predicted and measured tidal elevation exceeds 15 cm. Each event has been characterized in terms of duration, water level above or below predicted level, wind speed and direction and barometric pressure. Water level differing from astronomical tide has been explained by wind speed, direction and barometric pressure. Multiple regression analysis in which astronomical tide, east and north wind velocity components and barometric pressure are used to predict the measured tide show that only the astronomical tide and wind direction has statistical relevance; barometric pressure does not.

Summary

Identification of the hydrodynamical environment and its effect on sediment transport

We have determined that the flow direction on the marsh surface is predominantly perpendicular to the creek bank when water level exceeds creek bank elevation and that asymmetries between rising and falling tide exist. Because we are only able to make measurements at one point in the water column, we do not know the vertical structure of the flow, and it is possible that some of the observed asymmetries are due to differences in flow velocity between the bottom and the surface. In general, the effect of vegetation is to modify the velocity profile to one that is uniform with depth. Leonard and Luther (1995) measured velocity profiles in canopies. They found the largest gradients in the bottom 25% of the flow. In the remainder of the flow, velocity structure is approximately uniform. Our measurements are made in the bottom 20-30% of the flow, in the region of uniform gradient, assuming that the profile structure is similar to the one suggested by Leonard and Luther (1995).

At the lowest elevations in the marsh interior, we hypothesize that the marsh topography controls the direction of flow when water levels are lower than 0.5 meters. While the marsh is flooding from the creek on the northern side, the marsh can be thought of as a basin being filled, as water gradually spreads out on the marsh. This process, however, only lasts 20 minutes at the beginning and end of the tidal cycle.

We have used the high frequency velocity measurements to determine the relative changes in the turbulent intensities as a function of distance from the creek. In addition we are examining the spectra to determine whether the observed fluctuations are related to turbulence created by the inertial forces in the flow or to separation of the flow around the plant stems. Leonard and Luther (1995) suggest that an appropriate scaling the Reynolds number is the diameter of the plant stems rather than the flow depth. In either case the Reynolds numbers are small (less than 2000), but the high frequency fluctuations suggest that the flow can not be characterized as laminar.

Identification of transport dynamics on the marsh

Sediment is transported to the marsh on the rising part of the highest tides. However, we have not yet been able to determine the source of the sediment. It is likely that the sediment transferred to the marsh surface is a combination of local creek bank sediment and material moving as suspended load in the tidal creek. Sediment concentrations have been measured at the water surface of the tidal creek, but not during tides sufficiently high to fully inundate the marsh surface. These measurements indicate that on the rising tide, concentrations in the tidal creek follow a similar pattern to the one observed on the marsh surface. Sediment concentrations are high on the rising tide and decrease during slack tide.

To identify sediment dynamics on the marsh surface we have designed two different types of experiments. The first type of experiment is directed towards determining how much sediment deposits at a particular location during a particular tide. The second type of experiments are directed towards determining the processes that control deposition.

To determine the relationship between measured concentration levels and sediment deposition, we collected sediment deposited on sediment plates during periods when sediment concentration was measured at a particular location. We have hypothesized that the sediment deposition at a particular location is a function of sediment settling as a consequence of decreased turbulence levels and that this process may further be enhanced by flocculation of particles suspended in the water column. To address this issue we have designed a settling experiment where we use OBS sensors to record the changes of turbidity in a 60 cm long tube at two different levels. By observing the clearing rate in still water we can derive characteristic settling rates. If the clearing rate at the bottom of the column is greater than at the top, it is an indication that flocculation occurs. Rather than conducting this experiment at the field site, we repeat the measurements at different known concentration levels, but using water and sediment from the study area. Further there may be a net downward flux of sediment due to turbulent fluctuations. To determine the effect of vertical mass flux on the observed rates of deposition, we have combined high frequency measurements of velocity and concentration.

It is possible that one of the mechanisms controlling the difference in clearing rate between stations on the transect is related to sediment adhering to plants. The vegetation density of Spartina Alterniflora is 120-140 stems/m^2 at the creek bank and 90-200 stems/m^2 in the interior. The plants on the creek bank have larger biomass (3-6 grams/stem) than the plants in the interior (2-4 grams/stem). Both vegetation density and biomass are highly variable. In general the vegetation density on the creek bank is lower, but each plant is larger whereas the vegetation density in the marsh interior is higher, but each plant is smaller. At each point along the transect three 0.25 m^2 plots were selected. Sediment was carefully washed off all plants and two months later the plants were harvested and cleaned in the lab. For each square, sediment mass, dry plant mass and number of stems has been determined. While a full analysis has not yet been completed the data suggest that more sediment adheres to plants in the lower areas of the marsh than on the creek bank, which corresponds well with the longer inundation time in the lower area.

The study was in part designed to explore whether resuspension of sediment occurs on the marsh surface during any part of the tidal cycle. None of the time series of sediment concentration that we measured at any location show dramatic increases in sediment concentration after maximum tidal elevation has been reached and flow direction reverses. In Figure 7, there is a slight increase in concentration at Station 1 during the last 20 minutes of flooding and it is possible that this increase reflects resuspension of sediment. The last 20 minutes of flooding is also when velocities are highest and hence when sediment is most likely to be suspended. The velocities on the marsh are, however, very low, and one of the effects of the growth of Spartina Alterniflora because it is a rigid plant and it protrudes through the surface of the flow, is to modify the shape of the velocity profile to one that is uniform with flow depth (Tsujimoto et al., 1991). The stress on the boundary is proportional to the gradient of the near-bottom velocity profile, and when the velocities are low, it is likely that this gradient is small.

Determine whether storms produce significant sediment transporting events

It is unclear whether wind stress and increased wave activity related to storms produce increased sediment loads in the lagoon and thus in the tidal creeks, or if the stresses related to a greater range in water levels are sufficient to suspend the material available for transport in the tidal creeks. Further we wish to investigate whether the material in suspension is derived primarily from the creek bank (a local source) or from the channel bottom, in which case the source may be remote. We are addressing this issue through a combination of sediment concentration measurements made simultaneously on the marsh and in the tidal creek during a range of water level and wind conditions. A more accurate understanding of sediment transporting events in this environment will aid in making predictions of its long term fate.

Leonard, L. and M. Luther(1995) Flow hydrodynamics in tidal marsh canopies. Limnology and Oceanography 40(8), 1474-1484

Stevenson, J.C., L.G. Ward and M.S. Kearney (1988) Sediment Transport and Trapping in Marsh Systems: Implications of Tidal Flux Studies. Marine Geology, 80, 37-59

Tsujimoto, T., T. Kitamura and T. Okada (1991) Turbulent Structure of Flow over Rigid Vegetation covered bed in open channels. KHL Progressive report, Hydraulics Laboratory, Kanazava University.

Return to table of contents.