CICEET progress report for period 08/01/2000 through 01/31/2001

Project Title:

Modeling the effects of changes in turbidity on light available for submerged aquatic vegetation

Project Coordinator:

Roger I. E. Newell
Professor, Horn Point Laboratory, University of Maryland Center for Environmental Science
PO Box 775, Cambridge, MD 21613
Phone: 410-221-8410, Fax:410-221-8490
Email: newell@hpl.umces.edu

Additional Principal Investigators:

Raleigh R. Hood
Assistant Professor, Horn Point Laboratory, University of Maryland Center for Environmental Science
PO Box 775, Cambridge, MD 21613
Phone:410-221-8434, Fax:410-221-8490
Email: raleigh@hpl.umces.edu

Evamaria W. Koch
Assistant Professor, Horn Point Laboratory, University of Maryland Center for Environmental Science
PO Box 775, Cambridge, MD 21613
Phone:410-221-8418, Fax:410-221-8490
Email: koch@hpl.umces.edu

Raymond E. Grizzle
Associate Professor
Jackson Estuarine Laboratory
85 Adams Point Road
Durham, NH 03824
Phone: 603-862-2175, Fax: 603-862-1101

I. Accomplishments

A. Scheduled Tasks

Undertake in October 2000 the second of the two proposed field studies to quantify sediment resuspension in SAV beds when, due to normal senescence, the seagrass shoot density will be lower than during the June study period.

Complete the experimental work in the laboratory to measure changes in light extinction coefficients associated with oyster and clam feeding over a range of seston concentrations at 15, 20, and 25^oC water temperatures.

Continue the development of our mathematical model in STELLA. Our first goal is to predict, on theoretical grounds, the interactions between seagrasses and suspension feeders through their effects upon suspended sediment concentrations and light penetration. Start to parameterize the model using our own field and laboratory data.

B. Progress on Tasks.

1. The influence of seagrasses on sediment resuspension has been measured in June and October in Duck Point Cove at Bishop’s Head Point at the entrance to Monie Bay, Maryland. In this same cove we selected two sites, one densely vegetated by the seagrass, Ruppia maritima, and one unvegetated. At each site, two small platforms were mounted to hold the necessary equipment to monitor environmental conditions in a vegetated and an unvegetated area over two 10-d periods (Summer when the vegetation occupied the entire water column and Fall when the plants only occupied a small fraction of the water column). We used an ISCO automated water sampler at each site to collect 1 l of water every 2 h to determine total suspended solids (TSS), grain size distribution, and salinity. Additional abiotic parameters recorded included water temperature, light availability (4  LiCor sensors) and wave characteristics. At the end of the experiment, seagrass density, canopy height, and photosynthesis versus irradiance curves were measured. Samples are in the final stages of being analyzed. Additionally, the critical erosional threshold of the sediments at each site was quantified using microcosms under controlled conditions

2. We used adult oysters (Crassostrea virginica) (8 to 10 cm shell height) collected from the Sandy Hill oyster bar in the Choptank River. Hard clams, Mercenaria mercenaria ,(shell height of 5 to 6 cm) were grown in Plantation Creeks, VA and provided by Dr. Mike Pierson of Cherrystone Aquafarms. Both species of bivalve were collected in March and August 2000 and held in the lab for 14 to 20 d acclimation prior to use in the feeding studies. In March one group of each species was held at the field ambient water temperature of 15^oC and one group was acclimated to 20^oC.

In August animals were maintained only at the field ambient water temperature of 25^oC. Oysters were held in flowing ambient estuarine water salinity (salinity 12 to 15 ppt) with a natural seston complement. Clams were held in non-flow through aerated tanks with 20 ppt water made by mixing estuarine water with full strength seawater with 20% of the water exchanged every second day. Clams were fed algal paste (Chgra) at 2% of dry body weight per day as a maintenance ration. Clams were put into 12 cm deep plastic beakers containing coarse sand into which the deeply buried during the acclimation period.

The influence of eastern oysters and hard clams on turbidity and light penetration was evaluated in 1 m deep 1000 l tanks filled with estuarine water. These tanks have a continuous mixing systems consisting of rotating, reversing paddles, with speed, direction, and duration controlled by computer to simulate mixing in nature. This system ensures homogeneous mixing of the water column without resuspending sediment from the bottom as described by Sanford (1997). Light was provided by 4 ft high output flourescent tubes which produced irradiance of ca 200 ( µmol photons m^-2 s^-1) just below the water surface

In separate experiments oysters and clams were placed on the bottom of the tanks and allowed to feed undisturbed for 2 h. Oysters were placed in 2 cm deep plastic pots to facilitate retention of both feces and pseudofeces and clams were placed directly in the tank while still buried in the sand-filled plastic beakers. At intervals for between 10 and 24 h the diffuse attenuation coefficient for photosynthetically available light, K_par was measured using a 4  Li-Cor light meter positioned just beneath the water surface (O) and at 0.5 m (Z) beneath the surface. The flourescent lights were only switched on for the 10 min period required to make readings and were then turned off to reduce any phytoplankton growth during the experimental runs. Due to slight variations in light output from the high intensity flourescent tubes we measured light levels (µmol photons m^-2 s^-1) haphazardly 4 times in each of the 3 tanks. The 4 light readings were then used to calculate the extinction coefficient

K_d = [ln (light Z /Light O )] / Z from which an average K_d was calculated.

In order to estimate the loss in light intensity associated with distance from the light source, rather than attenuation by material in the water, we filled these same tanks with tap water containing no suspended particles and, as before, measured light just under the water surface and 0.5 m further into the tank. This K_d value was used to correct all of the experimental K_d values. Regression equations of Kd against time were calculated for each data set and used to calculate the absolute change in Kd (increase in light penetration) for a 24 h period. The change in Kd for the control tank associated with particle settlement was then subtracted away from the experimental values.

Water was collected concurrently with the light measurements using a siphon hose close to the middle of the tank. Particle size distribution and abundance of particles > 2 µm (size above which clams and oysters can retain a large percentage of suspended particles) was counted on subsamples using a Coulter Multisizer II. Water samples were filtered through 47 mm Whatman GF/C filters for analysis of seston dry weight (mg l^-1) in order to quantify the relationship between seston and K_par. Filters were rinsed with isotonic Ammonium formate to remove salt (Armstrong 1958, Berg and Newell 1984). Changes in chlorophyll concentration due to bivalve feeding was measured at each water collection by filtering water and extracted for chlorophyll a analysis (µg l^-1). At the end of each experiment, oyster and clam tissues were individually dissected from shells into pre weighed aluminum pans and dried at 80 ^oC for 2 d for determination of each animal’s total dry tissue weight.

Aggregate clearance rates for the known biomass of bivalves in each tank (liters water cleared h^-1 g^-1) were calculated based on the rate of particle disappearance in tanks. Particle concentrations measured using the Coulter Multisizer at intervals in each tank were transformed by logarithm and used to prepare a linear regression of particle disappearance over time for each tank. These equations were used to predict the initial (time 0; C_i) and final (time 24 h; C_f) log transformed particle concentrations in each tank, including control tanks without bivalves (time 0; C_ci and time 24 h; C_cf) in order to correct for particle settlement during the feeding period.

These values were then inserted into the following equation (Coughlan 1969) to calculate aggregate clearance rate (CR):

CR (L h^-1) ={ [ (log_e C_i- log_e C_f)]- [(log_e C_ci - log_e C_cf)]} x (V/t)

where V = Tank volume (1000 L) and t = duration of feeding period. Clearance rates were corrected to a dry tissue weight of 1 g by dividing by the total dry tissue weight of all animals placed in that tank.

3. We have continued to make substantial progress in the development of our mathematical model predicting the interrelationships between seagrasses and suspension feeders on light penetration.

 

C. Difficulties Encountered

No major difficulties have been encountered so far in the development of this project.

D. Anticipated Success in Meeting Project Objectives in Scheduled Project Period.

We have met all the objectives we set out for the first 18 months of this 24 month project

E. Preliminary data

The field studies in the seagrass beds and adjacent unvegetated areas are showing the contribution of SAV to decreasing total suspended solids (TSS) loads in the water column. This is a process driven by physical processes (waves) as well as the reproductive state of the plants. The concentration of TSS increases with increasing wave height in both the vegetated and unvegetated areas (Figs 1 and 2).

Attenuation of waves is a function of the amount of the water column that is occupied by the vegetation. In the Summer, when plants are in a reproductive state, and hence occupy the entire water column, wave attenuation is high. In the Fall, when plants are in a smaller vegetative state, and hence occupy only a small fraction of the water column, wave attenuation is low (Fig. 3). Based on this we might expect that the overall TSS levels within the beds would be higher in the Fall than in the Summer. Our data suggest that in fact the highest loads are in the Summer most likely due to the high contribution of particulate organic matter from phytoplankton during the warmer months. In addition, because there are more seagrass leaves in the summer there is a greater area upon which material from the water column can settle and later be resuspended. Independent of the source of suspended particles, light attenuation in the water column increases with increasing TSS concentrations (Fig. 4).

Figure 1. Wave dependent resuspension of particles in a vegetated (black symbols) and adjacent unvegetated (white symbols) area when the vegetation occupied the entire water column (Summer).

Figure 2. Wave dependent resuspension of particles in a vegetated (black symbols) and adjacent unvegetated (white symbols) area when the vegetation occupied only a small portion of the water column (Fall).

Figure 3. Wave attenuation of Ruppia maritima when reproductive (shoots occupy the entire water column — black symbols) and when in its vegetative state (shoots occupying only a portion of the water column — white symbols).

Figure 4. Light attenuation (Kd) as a function of total suspended solids concentration (TSS) in a SAV habitat.

 

2. The data from the completed oyster and clam feeding studies (Table 1) show that there were clear and measurable changes in the light attenuation coefficient (K_d) over the course of the feeding period. Our study provides the first quantification of how the removal of suspended particles by bivalves increases the amount of light that can penetrate through the water column, with rates being greatest for both species at 25^oC. As anticipated, the weight-specific feeding activity of oysters was much greater than the clams. This is due to the fact that oysters are better adapted to maintaining high rates of particle clearance in turbid waters. The oyster's feeding strategy relies on the highly efficient sorting capacity of their gills and labial palps that allows them to reject non-nutritious and excess particles as pseudofeces and hence ingest a constant amount of nutritious particles. In contrast, clams produce fewer pseudofeces and regulate ingestion by reducing clearance rates. These data will then be used to parameterize the oyster and clam filtration component in the simulation model described below.

F. Continued Model Development:

We have essentially completed the formulation and parameterization of the predictive model and implemented it in Stella. As described in our previous reports, the basic model is composed of two equations, one which describes the rate of change in SAV biomass (B_sav):

And a second which describes the rate of change of suspended seston concentration (S):

In (1) u_m is the maximum growth rate of SAV, I_avg is the average irradiance over the length of the shoot, I_k is the light saturation parameter for SAV photosynthesis and r is coefficient characterizing respiratory losses. In (2) M is the erosion rate, _b and _c are the bottom shear stress and the critical shear stress for resuspension, respectively. In addition, in (2) w_s is the sinking rate of seston, C_b is the filtration rate of bivalves, and B_b is the biomass of bivalves (the latter is specified, i.e., not dynamically modeled). Z_bot is the distance from the water surface to the bottom. The model is cast in a vertically integrated form and with spatial units of m². Equation (2) feeds back on (1) through S, which partly determines I_avg in (1). S is determined in 2) by the balance between sedimentation, bivalve filtering, and sediment resuspension.

In order to calculate I_avg we include light attenuation due to water and dissolved substances, K_x, seston, K_s, and self-shading by SAV, K_sav. And we split this attenuation into two parts, attenuation above the shoots, K₁, which is due only to water, dissolved substances and seston:

and attenuation below the tips of the shoots, K₂, which include all of the above plus self-shading:

For simplicity we assume that K_x is a constant. We estimate K_s = m₁S from seston concentration as described above, and we estimate K_sav = a_savB_sav using an SAV biomass specific attenuation coefficient, a_sav. The incorporation of a self-shading effect in the model allows us to dynamically model maximum bed density as a function of depth and light attenuation in the overlying water column.

The average light over the shoot length is determined by first calculating the irradiance at the top of the shoots, Z_top, where light is attenuated by K_x and K_s:

and then averaging the light from Z_top to the bottom, Z_bot:

where I_o is the irradiance at the water surface.

For this application we assume that the critical bottom stress for resuspension, ô_c, is a constant, and that the bottom stress, ô_b is determined primarily by waves:

In 7) H_r is the wave height, ñ is the density of the water, õ is the kinematic viscosity of the water, T is the wave period, and L is the wavelength. The relationship between wavelength and period is calculated using standard shallow water formulae.

Equation 7 does not include wave damping effects due to the presence of SAV. Yet a key aspect of this model is the specification of these effects, i.e., how do changes in the density and height of SAV modify waves and how does this, in turn, change the bottom stress, ô_b. To our knowledge, there are no published quantitative relationships that describe these effects for SAV beds in shallow water environments. We are currently working on deriving relationships that can be used to modify H_r as a function of bed density and SAV height using observations collected by Koch as part of another study.

The model has been implemented in STELLA and preliminary runs (with Z_top = 0 and ô_b = 0, and the SAV model parameterized to simulate Rupia) give realistic depth distributions, i.e., equilibrium SAV bed densities declining exponentially from about 3000 shoots m^-2 at Z_top = 0 m, to 0 shoots m^-2 at Z_top = 5 m depth.

 

 

II. Tasks and activities for next reporting period

A. Tasks for the next reporting period

1. Formulate and implement a relationship which describes wave height as a function of SAV height and bed density.
2. Complete the development of the predictive mathematical model incorporating all functional relationships and parameters derived from the field and laboratory studies. Carry out a full sensitivity analysis on the model.
3. Characterize the model-generated interactions between wave height, SAV growth and bivalve feeding.
4. Test the predictions of light penetration and seagrass growth from our STELLA model at clam aquaculture farms on the lower eastern shore of Chesapeake Bay where turbidities may have been reduced sufficiently to allow seagrass beds to become reestablished.
5. Develop the user interface of the model. Make verbal presentations at scientific meetings and hands-on demonstrations of the model to resource managers. Make any needed improvements and refine the user interface.
6. Distribute the model via the internet.
7. Write paper describing the model and the results of the field research in scientific journals.

B. Work plan to accomplish tasks

We will complete the development and testing of our mathematical model in STELLA. The model will incorporate the data from the field studies of the seagrass beds at Bishops head and the laboratory bivalve feeding studies described above. Our first goal will be to predict the interactions between seagrasses and suspension feeders through their effects upon suspended sediment concentrations and light penetration. This model will be completed before we start our second summer of field work in May 2001.

We will test our model predictions of the influence of bivalve feeding on light penetration and seagrass growth at Cherrystone hard clam aquaculture cooperative on the lower eastern shore of Virginia on the Chesapeake Bay. We have been in contact with the manager, Dr. Mike Pierson, and have selected two clam grow-out sites with different sediment characteristics and exposure to wave action. Mattawoman creek has quite a sandy bottom and is exposed to NW waves from Chesapeake Bay. In the upper part of Cherrystone inlet clams are grown in a sheltered location with a muddy bottom.

We have made a strong start in disseminating the results of this work to the management and research community. We have been invited to make the following scientific presentations concerning our research:

Newell, R.I.E. Potential for N, P, and Sediment Removal by Oysters. Invited presentation at an AExploratory meeting on nutrient/sediment removal by oysters@ organized by EPA Chesapeake Bay Program. Annapolis MD March, 2001

Newell, R.I.E., E. Koch and R.R. Hood. Modeling influence of populations of suspension-feeding bivalves and seagrasses on suspended particulate load and consequent changes in light attenuation. Invited presentation Estuarine Research Federation Biennial Meeting. November2001.

C. Concerns or difficulties

At this point we anticipate no major difficulties regarding the future progress of this project.

III. Expenditures

In no budget categories are expenditures exceeding estimates. Apparently invoicing from the University of Maryland to the University of New Hampshire is behind schedule. Our business office realizes the problem and told me that they would take care of it immediately.