Farm FARM Model for Production and Nitrogen Removal Estimates

The FARM model combines physical and biogeochemical models, bivalve growth models, and screening models for determining shellfish harvest and for eutrophication assessment at the farm scale (Ferreira et al. 2007). The FARM model is well described and has been tested in the United States, European Union, China, and elsewhere (Bricker et al. 2014, 2015, 2018, 2020, Ferreira et al. 2007, 2009, 2011, 2012, Nobre et al. 2010, Nunes et al. 2011, Saurel et al. 2014;  www.farmscale.org). Briefly, the model calculates the phytoplankton and detrital carbon removed by shellfish as water and “food” passes through the lease area, and then converts those values to N and deducts losses due to pseudofeces, feces, excretion, mortality, and spawning. The mass balance, for a single cultivation cycle, provides a value for net removal of N from the water column by the population of cultivated oysters, which effectively equates to a drawdown of phytoplankton (represented as CHL in the model), that is, of one of the primary symptoms of eutrophication. The model also provides an estimate of production (harvest) for one cultivation cycle (e.g., 18–24 mo; Ferreira et al. 2007). The model is useful for decision support for aquaculture siting (Silva et al. 2011, Ferreira et al. 2012, Bricker et al. 2016) because it evaluates both harvest and ecological carrying capacity (e.g., changes in dissolved and particulate concentrations due to aquaculture activity) without the financial cost (e.g., for seed, cages, lease fees, insurance, boat and boat maintenance, and fuel) and time (i.e., one cultivation cycle is 18–24 mo) required for implementation of a shellfish farm.

The FARM model has recently been calibrated to Chesapeake Bay oyster farms, which is the model used for this study (Cubillo et al. 2018). An individual model for eastern oysters was developed for Maryland Chesapeake Bay from measured oyster data from the four cage operation farms in this study and was incorporated into the FARM model to simulate oyster population growth. Additional modifications were made to account for the high particulate matter environment within the bay, and an option was included that can turn off spawning so that growth of both diploid and triploid oysters can be simulated (Cubillo et al. 2018).

Data inputs required for simulation of oyster growth using the FARM model include monthly measures of temperature, salinity, concentrations of total suspended solids (TSS), total volatile solids (TVS; which represent the organic part of suspended solids, also called particulate organic matter), and phytoplankton (represented as CHL; Ferreira et al. 2007). Other measures, DO, nitrite nitrate (NO₂₃), and ammonia (NH₄) are not required for growth simulations but rather are used to determine changes in water quality that occur that are attributable to the oysters as the water moves across the oyster farm. These are also indicators of nutrient enrichment, and it was of interest to evaluate the changes in these indicators that could be attributed to the oyster farms. Additional inputs needed for the simulations are current speeds (taken from the NOAA buoy closest to each study site), oyster seeding density, size of oyster seed and harvestable oysters (3 inches, 7.6 cm), and typical mortality over the cultivation cycle (reported by the growers). The model outputs of interest to this study were oyster production and the net mass of N removed through uptake of phytoplankton and detritus (i.e., food for the oysters). The removal of N estimated by the FARM model is a result of the balance of shellfish filtration, assimilation into tissue and shell through physiological growth processes, and return of the remaining N to the environment through pseudofeces and feces, natural mortality, and excretion (Ferreira et al. 2007). Eight model simulations were made at Site 1 in Chester River using different mortalities (20%, 30%, 40%, 50%, 60%, 70%, 80%, and 90%) to represent the potential range of production and N removal at a farm, given the range of mortalities reported by all growers (20%–90%). This was the only site for which harvest data were made available (320 × 10³ oysters harvested in 2016) and, thus, was the only site for which a comparison of reported harvest and model results was possible.

Water Sampling and Analysis

Data to drive FARM model simulations were collected monthly, from bottom-culture production operations (Sites 4 and 5) from May 2016 to August 2017, and from water-column (cage) production sites (Sites 1, 2a, 2b, and 3) from May 2016 to August 2018, as producer schedules allowed. The difference in sampling schedules was determined based on the growing time (i.e., time to reach harvest size) for oysters produced by each method, and the available information in the literature on the growth of oysters in restored oyster reefs, which are representative of bottom culture (Kellogg et al. 2014, Cornwell et al. 2016). Sampling efforts were focused on water-column production methods to capture water quality parameters throughout the entire culture period.

At each site, monthly measures of DO, water temperature, and salinity were measured in the field with a YSI Pro2030 and recorded. Water samples were collected monthly from each site and prepped in the field for analysis by the Nutrient Analytical Services Laboratory located at the University of Maryland Center for Environmental Science Chesapeake Biological Laboratory, Solomons, MD. Water samples were filtered through GF/F glass microfiber filters (0.7 micron pore size) in the field, wrapped in aluminum foil, put on ice, and transported to the laboratory. The samples were analyzed for TSS, TVS, NH₄, NO₂₃, and CHL components using standard laboratory practices (Table 1).

Oyster Sampling and Analysis

In addition to water samples, oyster samples were collected from a single cohort from each water-column production location to provide accurate growth estimates for calibration of the FARM model to Chesapeake Bay (Ferreira et al. 2007, Cubillo et al. 2018). Oysters were not collected from bottom-culture operations because of the small size of oysters at deployment and the difficulty of obtaining accurate measurements from clusters of oysters. In addition, it was assumed diploid bottom-culture spat-on-shell oysters would have similar growth patterns to wild diploid oysters (Kellogg et al. 2014, Cornwell et al. 2016). Therefore, growth data from restored reef sites were used as a proxy for spat-on-shell aquaculture oyster growth.

Each month, 50–100 live oysters were collected from the four water-column production sites and transported to the Quantitative Fisheries Ecology Laboratory at the University of Maryland Center for Environmental Science Chesapeake Biological Laboratory for analysis. Oyster measurements included shell height, live whole weight, wet and dry shell weight, wet and dry tissue weight, and the volume of water inside the oyster shell via standard laboratory practices. Once shell height and live whole weight were recorded, the oysters were shucked and the volume of water inside the shell was measured. Tissue was separated from the shell to determine wet shell weight and wet tissue weight. After wet weights were recorded, both tissue and shell were placed in an oven to dry for 72 h at 60°C, and dry weight recorded. These data were used to compare growth at the different sites and to calibrate the FARM model to Chesapeake Bay (Cubillo et al. 2018).

TABLE 1.

Methods used for water quality parameter analysis by Chesapeake Biological Laboratory Nutrient Analytical Services Laboratory.

Economic Avoided Costs Analysis: Estimated Compensation for Nitrogen Removal Ecosystem Services

An avoided, or replacement, cost approach was used to estimate the value of the nutrients removed by oyster farms in this study. This approach assumes there is equivalency of N removal services by oysters and by alternative removal methods, that the avoided cost good is the least cost for N removal, and that there is willingness to pay because of the inclusion of parts of Chesapeake Bay on the 303d list for N impairment and the requirement of N load reductions to meet water quality goals (Freeman et al. 2014). In this approach, the costs of alternative nutrient management measures (e.g., wastewater treatment plant upgrades, and agricultural and urban stormwater BMPs) are used to represent the potential value of N removed by oysters (e.g., Bricker et al. 2018, 2020, Rose et al. 2015). There continues to be great interest in determining the economic value of oyster-mediated nutrient removal services, given the need for continued nutrient management in most United States coastal waterbodies and worldwide. Harvest of cultivated bivalve shellfish is already being used to fulfill mandated nutrient reductions in several United States jurisdictions and parts of Europe (e.g., Massachusetts, oysters and clams; Town of Mashpee Sewer Commission 2015, Reitsma et al. 2017; Denmark, mussels, Nielsen et al. 2016; Chesapeake Bay region, oyster tissue, Cornwell et al. 2016). Compensation to shellfish growers for the nutrient removal ecosystem service provided by their oysters through a nutrient credit trading program is gaining momentum (Lindahl et al. 2005, Profeta & Daniels 2005, Lal et al. 2009, Miller 2009, Jones et al. 2010, Ferreira & Bricker 2016, Weber et al. 2018).

This is evidenced in the Chesapeake region by the inclusion by the State of Virginia of shellfish aquaculture in their nutrient trading program (§10.1–603.15:2; VA DEQ 2017; VNCEA 2017), and the recent application of the Oyster Company of Virginia to be a credit aggregator under the nutrient trading program of Virginia (VA DEQ 2018). In July 2018, Maryland adopted its Nutrient Trading Program regulations for the agricultural, stormwater, wastewater, and on-site sewage disposal sectors (COMAR 26.08.11.01). Within this program, harvested oyster tissue from aquaculture practices are eligible for consideration as nutrient reduction BMPs where the N (and P) removed in harvested oyster tissue can be credited toward fulfillment of nutrient reduction requirements. The framework has been developed for inclusion of sales of oyster aquaculture credits, but very few sales have been made (Wheeler 2020). The Maryland Department of Environment has developed a marketplace that includes trades of N removed in aquaculture oyster tissue, based on the approved oyster aquaculture BMP (Cornwell et al. 2016). The MD DNR has been designated as the agency that will verify harvests from shellfish aquaculture leases that will be used to generate credits through the Maryland Nutrient Trading Program.

In anticipation of this, one Maryland company has been formed to act as a third-party verifier and aggregator of credits produced from aquacultured oysters (Blue Oyster Environmental n.d.; East Coast Shellfish Growers Newsletter, June 2019; Shockley 2019). This company, in late 2019, was able to broker a voluntary payment of $1,600 (4 lb of N) to an oyster grower for nutrient reduction credits to shrink the environmental footprint of events held at the Baltimore Convention Center, highlighting the Baltimore Convention Center commitment to sustainability and environmental protection. In the second payment to oyster growers in Maryland Chesapeake Bay, Anne Arundel County paid $4,950 (107 pounds of nitrogen and 12 pounds of phosphorus) for nutrient credits to satisfy a regulatory requirement to meet state mandates for treating stormwater runoff—the credits were equivalent to treating runoff from three acres of impervious surface (Wheeler 2020). The two recent trades provide optimism regarding the recognized value of aquaculture-related nutrient removal and the possibility of full inclusion of oyster growers in a comprehensive nutrient management program. In addition, these first payments enhance public awareness of the beneficial impacts of oyster aquaculture on water quality and might help shift attitudes to allow expansion of shellfish aquaculture that will also support domestic sustainable sources of seafood.

Quantification and valuation of N removed through shellfish harvest can be used by policy-makers to assess the available set of nutrient management tools to design a comprehensive nutrient management plan that will most efficiently and cost effectively achieve water quality goals. Assigning an economic value to the N removed via oyster bioextraction is difficult (Peterson & Lipcius 2003) and is further complicated by the lack of functioning N trading programs (Stephenson & Shabman 2017). Previously published avoided cost values determined for Virginia Chesapeake Bay (Jones et al. 2010, Stephenson et al. 2010, Rose et al. 2015) were used to calculate economic benefits for oyster harvest in Maryland Chesapeake Bay (Table 2). Those estimates were compared with recent results from Weber et al.'s (2018) evaluation of potential effects of economic compensation for nutrients removed via oyster aquaculture. Weber et al. (2018) estimated upper and lower N credit values ($22 and $418 per kg of N removed) that would increase aquaculture production by a small or large amount. This range of costs is representative of the range of costs of providing nutrient removal by oyster aquaculture. Note that the range of costs reported by Weber et al. (2018) overlaps with ranges reported by Bricker et al. (2018) and Rose et al. (2015), although at the lower end of those ranges. In addition, published payment rates from the Virginia Nutrient Credit Exchange Association (VA DEQ 2017, VNCEA 2017; Table 2) were also included for comparison. Although the value of the N removed may be provided to growers as compensation for the N removal service their oysters provide within a fully developed nutrient credit trading program, the upper-level values do not represent what expected compensation will be, as noted earlier. In a competitive functioning nutrient market, the expected payments will be closer to the lowest cost alternative (Freeman et al. 2014; e.g., wetlands or agricultural BMPs). If there is a limitation on the lowest cost BMPs available for purchase, other slightly more expensive BMPs will be available, but not the highest avoided cost value. Values of N from recent trades were not evaluated because the value of N from the Anne Arundel County trade was not identified, and the Maryland Department of Environment has not recorded enough trades to determine an average value for N removed via oyster aquaculture in the Maryland Nutrient Trading Program.

TABLE 2.

Estimated value of nitrogen removal as an avoided or replacement cost analysis using costs of alternative nutrient management measures to estimate value for Maryland Chesapeake Bay oyster bioextraction.

Water Quality Data Model Inputs

Annual averages of monthly measures of water quality parameters at the six study sites from May 2016 to August 2018 are shown in Table 3. Parameter concentrations vary from farm to farm, but annual patterns are similar at all sites. Salinities are lower in summer, and overall concentrations are lower in the river sites, i.e., Chester, Wicomico, and West rivers (Sites 1, 4, and 5) than at farms located closer to the mainstem of the bay (i.e., Sites 2b and 3), highlighting the influence of freshwater and estuarine water masses. Annual average salinities at all sites are within the range of tolerable salinities for Crassostrea virginica (5–40, with an optimum of 15–25; Shumway 1996).

Average annual concentrations of CHL for all stations are considered moderate according to the range of values in the Assessment of Estuarine Tropic Status (ASSETS, Bricker et al. 2003) model (5–20 µg L^–1; Table 3). The 90th percentile of CHL, an ASSETS criterion representing worst-case (highest) concentrations during an annual cycle (Bricker et al. 2003), is in the high category for three river farm sites (Sites 1, 4, and 5) and moderate for the other three farm sites (Sites 2a, 2b, and 6). Chlorophyll concentrations at all sites are greater than the threshold supportive of desired growth rates shown in an analysis in Long Island Sound (4.5 µg L^–1, Bricker et al. 2016). Dissolved oxygen concentrations indicate no problems with depletion in neither annual average nor 10th percentile concentrations, a criterion representing worst-case (lowest) concentrations during an annual cycle (Bricker et al. 2003).

TABLE 3.

Annual averages and standard deviations of water quality parameters from monthly measures at the six oyster farm study sites in Maryland Chesapeake Bay from May 2016 to August 2018 (Fig. 1 for locations).

FARM Model Production and N Removal Estimates

Measured data from the farm sites were used as inputs to the FARM model to estimate production and N removal at each site; results are shown in Table 4. Production, or harvest, ranges from total farm production of 20.1–194 metric tons of harvestable oysters y^–1 or 1.78–25 metric tons of harvestable oysters acre^–1 y^–1. Oyster filtration–related N removal was estimated to be at a range of 28–457 kg N acre^–1 y^–1, which, put into a water management perspective based on 3.3 kg N person^–1 y^–1, represents water treatment for people equivalents (PEQ) of 8–139 PEQ acre^–1 y^–1. Results showed that CHL, DO, and NH₄ concentrations did not change at any site.

Ecosystem Service Valuation

Annualized cost estimates for removal of 1 kg N via several different nutrient management methods (Table 2) were applied to the estimated N removed by the oyster farm operations (Table 5). The annual cost to replace the bioextractive removal of N in tissue and shell is estimated to range from $0.56 × 10³–$12,466 × 10³ y^–1 among the six Maryland farm study sites depending on the cost of the alternative management measure used for the analysis.

Production and N Removal at Oyster Farms in Maryland Chesapeake Bay

Results of modeling of the study farms show variable annual harvest from a low of 14 metric tons at Site 1 to a high of 194 metric tons of oysters at Site 2b. These are total harvests for each farm, which vary in size and culture practices, and are thus somewhat difficult to compare. The range of annual harvests normalized per acre is presented to provide a more standardized comparison, which vary from a low of 1.78 metric tons of oysters acre^–1 y^–1 at Site 3 to a high of 25.1 metric tons of oysters acre^–1 y^–1 at Site 2a (Table 4). That DO and NH₄ did not change at any site indicates that the farm operations did not have a negative impact on water quality at the farm sites. This means that there is margin for expansion of oyster operations via increased seeding densities at these sites, although this must be performed carefully so that other unintended consequences are not created, i.e., drawdown of DO due to overabundance of biodeposits (Lindahl et al. 2005).

The results for these Maryland Chesapeake Bay oyster farms are comparable to FARM-estimated production of Crassostrea virginica and other oyster species in other locations, although production at Site 2b is highest among sites in all studies (Table 6). Although a more comparable means of contrast, it is important to note that farm production on a per acre basis may not be appropriate because producers typically refer to the number of oysters they harvest per year, rather than an average number per acre. Also, growers may use portions of their lease for nursery production or may have obtained a larger lease than physically necessary to accommodate future production increases. This is important because to standardize the farm practices of the six study sites for use in this study, the total amount of seed was divided by the total lease area, which may not reflect their actual practices. The “ideal” lease size and number of cages placed on a lease are a matter of operator preference and desired production level, as evidenced by the range of lease sizes and stocking densities when measured on a per acre basis. These results, however, still provide insight about the range in potential harvest among Maryland oyster farms.

Differences in harvest are a result of a combination of factors including size of oyster farm, aquaculture practice (cage or bottom spat-on-shell), seeding density, food quantity and quality, salinity, and natural mortality. A principal component analysis indicated that among all variables, salinity is the strongest identifiable driver of differences in harvest among sites. An example of the different factors that confound the development of any correlations is illustrated by the high mortalities of bottom-grown oysters (75%–80% over the culture cycle) compared with cage-grown oysters (25%–70% over the culture cycle), which suggests that cage-grown oysters might be expected to show greater production. Among these sites, however, seeding density is much higher for bottom spat-on-shell (247 oysters m^–2 y^–1) than for cage-grown oysters (20–50 oysters meter^–2 y^–1), which partially counterbalances the difference in mortality. For example, at Site 1, the food quality (measured as TVS/TSS; Cubillo et al. 2018) and CHL concentrations are highest among the four cage-culture sites, and the 90th percentile CHL (a measure of highest concentrations within the year) is highest among all sites (Table 3). Thus, it might be expected that Site 1 would show the highest production. The seeding density at Site 1, however, is the lowest of all sites, and this location in upper Chesapeake Bay periodically experiences salinities that are below the tolerance of the oysters, which limits their growth. Periodic low salinity may explain why mortality is the highest at this site among the four cage operations (70% compared with 30%–50%).

A series of eight FARM simulations was performed with data from Site 1 to highlight the potential range of harvest that might be expected from year to year, given the range of mortalities reported by growers (20%–90%) and for comparison with reported harvest. This site was selected because it was the only site for which harvest was available for comparison with modeled harvest (see below). The estimated harvest ranged from 4.6 to 36 metric tons y^–1, which is the range that might be expected at any site, given variability in mortality due to salinity variations, predation, poor food quality, or other influences on mortality and oyster growth in any given year (e.g., high sedimentation from storms). The corresponding range of oyster-associated N removal in tissue and shell at Site 1 is 147 to 445 (average 308) kg N removed y^–1.

TABLE 6.

Comparative FARM model estimates of oyster harvest and oyster-related N removal via sequestration into tissue and shell for Chesapeake Bay (this study), Great Bay Piscataqua River Estuary (Bricker et al. 2020), and Long Island Sound (Bricker et al. 2018).

Oyster-associated removal of N via filtration and assimilation into tissue and shell estimated by the FARM model also shows a wide range among all farm sites. Removal varies among farms from a low of 253 kg N to a high of 2,554 kg N removed per farm per year. On a per acre basis, the range in removal varies from 28 to 457 kg acre^–1 y^–1. The corresponding PEQ represented by total N removal by a farm ranges from a low of eight to a high of 139 PEQ acre^–1 y^–1. As for production, the N removal by Chesapeake Bay oyster farms is within the range of removals reported for oysters in other United States and international study sites (Table 6).

Comparison of FARM Production and N Removal Results to Reported Harvest and Chesapeake Bay Program Nitrogen BMP Estimates (i.e., based on tissue only)

The multiple simulations with differing mortalities at Site 1 also afforded us an opportunity to evaluate the accuracy of the FARM model at predicting harvest. Although previous studies have shown that the FARM model estimates harvest of cultivated oyster and other bivalves reasonably well (Ferreira et al. 2009, Bricker et al. 2020), it would be useful to have reported harvest at all Maryland farm study sites for confirmation of that for the Chesapeake Bay calibrated FARM model. The Maryland industry is young and growing, and there was only one site for which harvest numbers were made available for comparison with model results. As noted earlier, harvest reported for Site 1 was 320 × 10³ triploid oysters, which were assumed to be 3 inches in size and weigh 35 g each. There are two results for FARM model simulations for Site 1 in Table 4 for comparison with reported harvest. The original estimate of mortality given by the grower was 33%; however, using this as a model input resulted in model estimated harvest (converted from metric tons) of 860 × 10³ oysters y^–1, much greater than the reported harvest. The grower reviewed records of seed and harvest to confirm that the actual mortality was 70%. Model simulations using eight mortalities representing the range observed by all growers (20%–90%) were performed to estimate the potential range of harvest and N removal at this site, as an example of the range that might be expected at any site. The range of oyster harvest at Site 1 was 4.6 to 36 (average 20.3) metric ton oysters y^–1. The FARM model estimate of harvest at 70% mortality (388 × 10³ oysters y^–1) was closest of all simulation results to the reported harvest, suggesting that the Chesapeake Bay FARM model reasonably well estimates harvest (Table 4). This highlights that model inputs must be as accurate as possible (e.g., mortality estimates) to make an estimate that is representative of reported harvest.

These simulations also allow for comparison of the FARM model estimates of N removal, which are calculated by a mass balance approach, with an estimate using a per oyster N content as prescribed by the Chesapeake Bay Oyster BMP (Cornwell et al. 2016). The FARM model estimates the total amount of N in shell and tissue removed via oysters produced on aquaculture operations, whereas the approved BMP currently includes only the N removed in oyster tissue. To compare results, the FARM estimates were normalized to N removal by tissue only. FARM N removal estimates were normalized to tissue only using results from Cornwell et al. (2016), showing measured tissue N content as 72% of total triploid oyster N content and 47% of total diploid oyster N content. The BMP-based oyster N removal was calculated from the number of oysters (converted from FARM harvest results) and prescribed BMP credit values for harvested tissue of 90 kg N removed with harvest of one million diploid oysters and 130 kg N removed per million triploid oysters (Table 7).

The FARM model resulted in higher estimates of N removal via harvested aquaculture oyster tissue (e.g., 182 kg y^–1 at Site 1; Table 7) when compared with the approved BMP (e.g., 50 kg y^–1 at Site 1), similar to higher estimates by FARM noted in a previous study (Bricker et al. 2020). This in part reflects the difference between the oysters that are included in the calculations. For example, FARM includes all oysters in the water and uses a mass balance approach where removal is based on filtration and assimilation into tissue and shell minus mortality, excretion, feces, and pseudofeces (values were normalized to tissue as described earlier but is still based on the whole population). The BMP calculation includes only tissue in harvested oysters, although this is unlikely to account for the entire difference; however, the conservative estimate used by the BMP to avoid over crediting does add to the magnitude of the difference. It is interesting to note that the overestimate of removal by triploid oysters is less than the overestimate by diploid oysters. This is likely an artifact of the Chesapeake Bay calibration being based on cage-grown oysters, which grow faster and are lighter than bottom-grown oysters. Ultimately, nutrient credits and economic compensation for N removal by oyster aquaculture will be determined by estimates based on numbers of oysters harvested using the Chesapeake Bay oyster BMP (Cornwell et al. 2016), rather than from modeled estimates. It is useful, however, to have a way to estimate production scenarios, for example, under expanded aquaculture either by higher density planting or added lease acres, and the FARM model can do that without the resources and time that would be required to actually expand the industry. Production estimates from those scenarios could be used with the BMP-based N removal to estimate potential N removal with changes in aquaculture harvest.

TABLE 7.

Estimated nitrogen removal by harvested tissue only per farm compared with approved Chesapeake Bay Program oyster BMP credits for harvested tissue.

Valuation of the Nitrogen Removal Service Provided by the Oyster Aquaculture Industry

Several previous studies have highlighted the economic value of nutrients removed by oysters and other bivalves and the potential economic benefit that bivalves might contribute to a comprehensive nutrient management program (e.g., Lindahl et al. 2005, Miller 2009, Rose et al. 2015, Bricker et al. 2018, 2020). Currently, there are nutrient credit trading programs that allow credits for nutrient removal by oysters and clams. Those credits are being used to help fulfill required nutrient reductions; however, no sales/payments have been documented to growers within any state management program outside of Maryland (e.g., Town of Mashpee Sewer Commission 2015, Cornwell et al. 2016, Reitsma et al. 2017).

Results here, as elsewhere, show that there is a range of costs for removal of N ($2.2–$4,873 per kg N removed; Table 2), which leads to a range of values depending on the alternative management practice used to assign the value to the removed N (Table 5). The range among all the farm study sites is $0.56 × 10³–$12,445 × 10³ y^–1 if the total FARM-related N removal is considered, but would be reduced to $0.40 × 10³–$6,687 × 10³ if only tissue-related removal was included in the calculation. The value of N removed, based on removal calculated by the BMP-related method, is a range of $0.11 × 10³–$3,518 × 10³ y^–1 among all farms. These ranges describe the value based on the costs of different management alternatives; however, the upper-level values do not represent expected compensation. In a competitive functioning nutrient market, the expected payments will be closer to the lowest cost alternative (e.g., wetland and agricultural BMPs; Table 2). If there is a limitation on the BMP credits at the lower range of costs available for purchase, then other slightly more expensive BMPs will be available, but it is highly unlikely that there will be any purchases of urban BMP credits at the $4,783 per kg N cost.

How Much Nitrogen Is Being Removed by Current Maryland Aquaculture Oyster Harvest and What Is the Value of the Removal?

Knowing the current harvest for Maryland Chesapeake Bay, and using the Oyster BMP method as described earlier, the total N removed and the economic value of the removed N were estimated assuming that bottom-culture cultivates diploid and water-column culture cultivates triploid oysters. In 2017, the MD DNR reported harvest for bottom-grown oysters of 44,805 bushels, and for water-column–grown oysters, harvest was 29,296 bushels. Assuming there are 250 oysters per bushel, the total harvest was 11.2 million diploid (bottom grown) oysters and 7.32 million triploid (water-column–grown) oysters. Based on the Chesapeake Bay oyster tissue BMP crediting guidelines (Cornwell et al. 2016), the estimated N removal by oysters in 2017 was 1,008 kg by bottom-grown oysters and 952 kg by water-column–grown oysters, a total removal of 1,960 kg N. The value of the oyster-related N removal in tissue only, based on the range in costs of alternative management methods, is $4.3 × 10³–$9,551 × 10³. This provides insight about the economic value of the N removal ecosystem service that the Maryland oyster aquaculture industry provides and about potential payments that might be made to oyster growers. The compensation to growers will be at the lower end of this estimated range of values, as noted. Growers can already receive credits toward fulfillment of required N reductions for their waterbody through the Maryland Nutrient Trading Program, which has developed a framework that is now allowing sales of credits. This is an optimistic and encouraging development for both water quality concerns and domestic seafood needs. The Maryland Nutrient Trading Program will be an example for the development of nutrient trading programs elsewhere.