Development and application of an online tool to quantify nitrogen removal associated with harvest of cultivated eastern oysters

Data compilation

The Aquaculture Nutrient Removal Calculator (ANRC) is based on a publicly-available dataset of eastern oyster (Crassostrea virginica) morphometrics and nutrient concentration of tissue and shell measured on individual animals across the Northeast region of the United States, defined here as spanning the US states of North Carolina to Maine [31]. The ANRC was constrained to this geographic range for several reasons: 1) C. virginica is the sole oyster species cultivated across the entire region; 2) similar cultivation practices are employed across this region, including bottom cultivation without gear, off-bottom and/or floating aquaculture gear; 3) environmental conditions on farms are broadly similar, typically involving temperate estuaries and nearshore coastal locations in state waters.

The data compilation effort focused on assembling a dataset with geographic representation across the full region, and inclusion of all major cultivation practices (Fig 1; Table 1). Identification of data from each state was conducted to assemble a dataset most closely reflecting on-farm conditions as possible: Highest priority was given to identifying data collected from oysters that were grown on farms. If no data from farms were found, the second priority was identifying data collected from oysters that were grown by scientists employing common cultivation practices (e.g. hatchery-produced oysters, use of aquaculture gear, etc.). Most of the data fit one of these two categories, but for one state within the region (Rhode Island), data collected on wild oysters from waterbodies that had active oyster farms were used instead.

Data were identified using Web of Science searches related to oyster aquaculture in the Northeast United States, and by directly contacting scientists conducting oyster research within each state across the region. A comprehensive and exhaustive literature review of all published eastern oyster nutrient and morphometrics data was beyond the scope of this project, and most of the available peer-reviewed literature for eastern oysters has been collected on wild populations, which was not the focus of this tool. Most of the data in this dataset were previously published solely as summary statistics, and the raw data for individual animals was contributed by project leads to the new publicly available dataset (Fig 1; Table 1; [31]).

Calculation of nitrogen removed at harvest

The approach employed here is an adaptation of methodology used by the Chesapeake Bay Program (CBP), within their nutrient management program, to calculate nitrogen reduction by farmed eastern oysters in Chesapeake Bay, USA [27, 32]. Briefly, the CBP convened an expert panel to conduct a literature review and synthesis to determine the mean nitrogen concentration of eastern oyster tissue (% of dry weight), as well as robust regressions for the relationship between eastern oyster shell height and oyster tissue dry weight. The panel compiled data on nitrogen concentration in oyster tissue across the US Northeast region (Virginia to New Hampshire). Data on oyster morphometrics were constrained to the Chesapeake Bay only. Both nitrogen concentration and oyster morphometrics datasets included a mix of wild, restored, and farmed oysters. The CBP only considered nitrogen reduction associated with oyster tissue, and did not include nitrogen from oyster shells in their calculation. Farm-scale nitrogen removal at harvest was calculated by multiplying the mean nitrogen concentration of eastern oyster tissue by the tissue dry weight at the mean harvest size to calculate the weight of nitrogen in an individual oyster, then multiplying this value by the number of oysters harvested.

We employed the same methodology to our farm-based, region-wide data set of oyster tissue, and we extended the analysis here to include oyster shell, for calculation of whole-animal weight of nitrogen (g). Data were compiled as described above for shell nitrogen concentration (%), oyster tissue and shell dry weight (g), and oyster shell height (mm). Oyster shell height was defined as the longest distance (parallel to the long axis) between the hinge and the lip of the oyster; it is worth noting that some studies included in the regional data set referred to this measurement as oyster shell length. Information on oyster ploidy (diploid/triploid), harvest/sampling location, and cultivation practice (grouped into three categories: no gear, bottom gear, floating gear) were also recorded for each individual oyster in the dataset.

Hypothesis testing was conducted using the statistical software program R version 4.3.1 (www.r-project.org). Nutrient concentration of eastern oyster tissue and shell was compared across states, ploidy, and cultivation practice. Due to data gaps in cross-factor datasets, 1-way hypothesis testing was chosen over a two-way analysis for main effects and interactions. A robust analog of ANOVA was employed that used a percentile bootstrap method with 20% trimmed means to compare independent groups [33]. The mean nitrogen concentration of oyster tissue and shell was calculated for the full regional dataset, and this mean value was applied in calculations of nitrogen removal across all locations, cultivation practices, and ploidy. The extreme curvature in the relationship between oyster dry weight and shell height was addressed in hypothesis tests by natural log transformation of shell height and oyster dry weight (tissue and shell) to approximate linearity, an approach previously employed in the shellfish aquaculture literature (e.g. [34, 35]). Robust regression analysis was conducted on the transformed data using the Theil-Sen regression estimator and a percentile bootstrap method [36]. While a transformation was applied in order to facilitate hypothesis testing, nonlinear methodology could be used within the tool itself to describe the morphometrics relationships on the original untransformed data. Nonlinear quantile regressions were generated using the R statistical package quantreg [37, 38], based on the previous literature indicating that the relationship between oyster dry weight and shell height is described by a power function [27]. The 50th quantile was used as an estimate of the median of the dataset, as 50% of the tissue dry weight values lie above each value of shell height using this approach. The use of nonlinear quantile regression also facilitates direct comparison to the previous Chesapeake Bay oyster nutrient best management practice [27].

Tool development

The ANRC is a tool designed for growers and resource managers to inform shellfish aquaculture permitting. Resource managers have expressed interest in easy-to-use tools that produce location and operation-appropriate values for beneficial services, and these values need to be produced in a format that aligns with the permitting process. We have synthesized available literature for eastern oyster farms across the US Northeast region (North Carolina to Maine), and applied methodology used by the Chesapeake Bay Program to calculate nutrient removal at harvest. Variability in oyster tissue and shell nutrient concentration was low, and an assessment of farm location, ploidy, and cultivation practice (with vs. without gear) suggested that a single average value could reasonably be applied across all farms.

The nutrient removal calculations are based on relationships of oyster dry weight-to-length and the average nitrogen (N) concentrations in shell and tissue material. First, we estimate the weight of the oysters based on the typical size of oysters harvested on a farm. The weight estimates are based on non-linear quantile regressions of oyster shell height and dry-weight for both tissue and shell material. Next, the N portion of total oyster weight is calculated using the average N concentration value for both shell and tissue. Adding the tissue and shell nitrogen yields the total weight of N per oyster. This result is scaled to the total number of oysters harvested, as input by the user.

The ANRC can be used for new permit applications based on estimated production value or to provide information on existing farms from actual harvest numbers. The tool was developed as an R Shiny application and is hosted on the NOAA Posit Connect server, hosted here: https://connect.fisheries.noaa.gov/ANRC/

The application (version 1.0) is divided into three tabs, the `Calculator`tab is the landing page, where the user is able to input their information and generate a report, a `Reverse Calculator`tab, where a user can calculate the number of oysters required to mitigate a given nitrogen load, and an `About`tab that explains the project background and user inputs, and contains details on data contributors, sample locations, and references.

The Calculator inputs are separated into three categories: Farm Practices, Farm Location, and Harvest Details. Farm practices include the name of the farm, oyster ploidy, and culture method primarily used for growing oysters. These values currently do not affect the calculation of nitrogen removal, and are simply incorporated into the report.

Farm Location input includes a text box for the harvest location, and a Leaflet-based map interface where the user can drop a location pin on the waterbody where the oysters were harvested. The map interface records the latitude and longitude of the dropped marker.

Harvest Details include the average oyster size at harvest, the total number of oysters harvested, and the period of harvest. Our analysis of the region-wide values of nutrient concentration and morphometric relationships for oyster shell and tissue suggest that just the number and size of the oysters harvested are sufficient to provide a robust region-wide estimate of nutrient removal associated with oyster harvest. The period of harvest (1 day to 5 years) is included for use in generating the report, but does not affect the calculation of nitrogen removed.

Oyster tissue and shell nitrogen concentration

The mean nitrogen concentration of oyster tissue was 7.70% across the full regional dataset (Table 2; n = 1,339, sd = 1.34, range 2.59–14.08). The nitrogen concentration of oyster shell was 0.19% across the full regional dataset (Table 2; n = 494, sd = 0.10, range 0.04–0.86). These observations were similar to the previously reported mean of oyster tissue nitrogen concentration by the CBP (mean 8.2%, range 5.64–9.27%, n = 17 [27], and by [23] mean 8.22%, range 7.0–11.8%, n = 14). The overall range of observed values was slightly higher here than in these previous studies, but this was not surprising given the two orders of magnitude increase in sample size over previous work. Similarly, the mean nitrogen concentration of oyster shell reported here was very close to that reported by [23], (mean 0.23%, range 0.13–0.32%, n = 13), with a much larger sample size and overall larger range in observed values. While the CBP does not include oyster shell in their calculation of oyster aquaculture nitrogen reduction, they have included oyster shell in their calculation of nitrogen reduction associated with restored oyster reefs, and the data compiled here show the same trend of very similar mean, slightly greater range, much larger sample size ([39] mean 0.2%, range 0.08–0.32%, n = 15). The similarities in overall mean, with a much larger and geographically extensive dataset, strengthens previous evidence indicating low overall variation in nitrogen concentration of eastern oyster tissue and shell.

Diploid and triploid oysters exhibited small but statistically significant differences in tissue and shell nitrogen concentration (Table 3; both p ≤ 0.001). The effect sizes were small relative to the magnitude of nitrogen concentration for both tissue and shell, as well as the overall range of nitrogen concentration observed here and in the literature (Figs 2C and 3C; [23, 27]; as summarized above). Mean tissue nitrogen concentration was 7.76% for diploid oysters, and 7.52% for triploids. Mean shell nitrogen concentration was 0.18% for diploids and 0.22% for triploids. In light of the small effect sizes observed between diploid and triploid oysters, we opted to use a single mean nitrogen concentration in the ANRC, regardless of the ploidy of harvested oysters.

Small but statistically significant differences in oyster tissue and shell nitrogen concentration were also observed across cultivation practices (p < 0.001) (Table 3), and effect sizes were also small relative to the magnitude of nitrogen concentration and the overall range of nitrogen concentration observed here and in the literature (Figs 2C and 3C; [23, 27]). Mean tissue nitrogen concentration was 7.66% for oysters cultivated in bottom gear, 9.18% for oysters cultivated in floating gear, and 7.62% for oysters cultivated on the seafloor with no gear. Mean shell nitrogen concentration was 0.21% for oyster cultivated in bottom gear, and 0.18% for oysters cultivated on the seafloor without gear. We were unable to find data on shell nitrogen concentration for oysters grown in floating gear in the published literature. The absence of shell nitrogen concentration data from floating gear does represent a data gap, and the number of data points available for oyster tissue nitrogen concentration was relatively low compared to bottom gear and no gear (floating gear n = 47, bottom gear n = 874, no gear n = 418). However, given the overall low variation in shell nitrogen concentration across all other factors considered, we believe available information indicates that the regional tool can be used to accurately predict the nitrogen removal by farmed eastern oysters grown in floating gear. Additionally, the ANRC can be easily updated should new data become available in the future that indicates oysters grown in floating gear contain substantially different tissue or shell nitrogen concentrations. Adaptive management based on an evolving scientific literature is a common practice in nutrient management (e.g., [40, 41]), and can easily be accommodated by the ANRC.

Small but statistically significant differences in oyster tissue and shell nitrogen concentration were observed based on a global test of location (both p < 0.001) (Table 3), with small effect sizes relative to the magnitude of nitrogen concentration and the overall range of nitrogen concentration observed here and in the literature (Figs 2C and 3C; [23, 27]). We opted not to perform pairwise comparisons given the low statistical power associated with controlling for familywise error across 21 pairs. The range in mean tissue nitrogen concentration across states was 7.25–9.24% and the range in mean shell nitrogen concentration across states was 0.14–0.24. Shell nitrogen concentration is more rarely reported in the literature, and data were only available for 4 states (VA, MD, CT, MA). Given that these states span both the mid-Atlantic and New England sub-regions of the US Northeast, and given the low observed variation, we believe a region-wide mean shell nitrogen concentration is supported by the available data.

Oyster morphometrics

The 50^th quantile nonlinear power function provided a good fit to both the tissue and shell datasets (Fig 4). The regression equation for tissue dry weight vs. shell height was y = 1.42e⁻⁵x^2.607 for the full regional dataset. The regression equation for shell dry weight vs. shell height was y = 0.00039x^2.58 for the full regional data set.

Diploid and triploid oysters exhibited small but statistically significant differences (all p < 0.05) in the relationship between transformed shell height and tissue/shell dry weight (Table 4; Fig 5). The effect sizes were small: 0.2 (tissue slope), 0.3 (tissue intercept), 0.13 (shell slope), and 0.79 (shell intercept) and the differences in tissue and shell dry weight were particularly small at the typical size range that eastern oysters are harvested in the Northeastern US (2.5–3.5 inches, 63.5–88.9 mm; Fig 5 vertical lines). These results are inconsistent with the current nutrient management practice by the CBP that assigns a higher tissue dry weight to triploid oysters across the range of oyster shell heights [27]. The analysis conducted by the CBP in support of their nutrient management program did not contain sufficient data to allow direct comparison of diploids and triploids grown under identical conditions, and all of the triploid data in the CBP dataset came from a single study [27, 32]. A more recent study that directly compared morphometrics of diploid and triploid eastern oysters grown in bottom cages at two farms in Maryland and Virginia, USA found no significant difference in the relationship between shell height vs. tissue dry weight between diploids and triploids at either farm individually or when the farm data was combined [35]. The data from [35] are included in the ANRC dataset, as well as data from two farms in Maine, USA, that grew diploids and triploids in floating gear, and had very small effect sizes in the relationship between shell height vs. tissue dry weight by ploidy. Due to the small ploidy-based difference in oyster morphometrics when oysters were grown in either bottom or floating gear, as well as the absence of data supporting a difference in diploids and triploid when grown under identical conditions, we have opted to use a single regression for both diploid and triploid oysters in the ANRC tool.

Small differences in oyster morphometrics were also observed based on cultivation practices, with a combination of significant and not significant statistical results (Table 4; Fig 6). There were no significant differences in oyster morphometrics between the bottom gear and no gear cultivation practices, for either tissue or shell datasets (all p > 0.05). Small but significant differences (all p < 0.05) were observed between floating gear vs. no gear cultivation practices, with effect sizes 0.80 (tissue slope), 3.6 (tissue intercept), 0.21 (shell slope), and 1.1 (shell intercept). Small but significant differences (all p < 0.05) were also observed between floating gear vs. bottom gear cultivation practices, with effect sizes 0.69 (tissue slope), 2.9 (tissue intercept), 0.18 (shell slope), and 0.84 (shell intercept). The absence of a significant result between bottom gear and no gear cultivation practices clearly supports the use of a single regression to represent these two cultivation practices. While there were significant differences between floating gear and the other two cultivation practices, the size of this effect was minimal, particularly at the typical size range that eastern oysters are harvested in the Northeastern US (2.5–3.5 inches, 63.5–88.9 mm; Fig 6 vertical lines). Differences were more pronounced at oyster shell heights exceeding 125 mm (~5 inches), but as eastern oysters are rarely harvested at this size, we did not feel this justified the development of a more complex tool. As observed with the nitrogen concentration data, the ANRC can be easily updated should new data become available in the future that indicates oysters grown in floating gear exhibit a substantially different relationship between tissue/shell dry weight vs. shell height.

While the full regional dataset had good spatial coverage, containing morphometrics data from all states, within-state data was more limited (Table 1). Of the 11 states with oyster morphometrics data available, only 4 had sufficiently large datasets to generate a statistically significant quantile regression for the relationship between oyster tissue dry weight and oyster shell height (North Carolina, New York, Connecticut, Maine), and only 2 had sufficient data to generate a statistically significant quantile regression for the relationship between oyster shell dry weight and oyster shell height (North Carolina, Maine). Within these states, geographic coverage was limited, with 3 farms sampled across 2 estuaries in North Carolina, 4 locations sampled across 2 estuaries in New York, 1 location sampled in Connecticut, and 4 locations sampled across 3 estuaries in Maine. None of the state-specific farmed oyster datasets contained the within-estuary spatial resolution achieved by the CBP in their mixed wild/restored/farmed oyster dataset, which had samples from 22 locations across the upper, mid, and lower Bay, and included the full range of salinity conditions across which oysters are grown in this estuary [27]. It is unclear whether oysters sampled from a small number of locations within a single coastal waterbody adequately capture the potential variability in this relationship within the waterbody/state as a whole.

While state-specific comparisons in oyster morphometrics were not feasible, we did assess the potential for a general north-south trend by dividing the dataset into two geographic sub-regions (Fig 7). Oyster morphometrics data from the states of Maine, New Hampshire, Massachusetts, Rhode Island, and Connecticut were pooled into a “New England” dataset, and compared to data from the “Mid-Atlantic”, which included the states of New York, New Jersey, Delaware, Maryland, Virginia, and North Carolina. The comparison was statistically significant, but the effect sizes were small: 0.35 (tissue slope), 1.4 (tissue intercept), 0.28 (shell slope), and 1.0 (shell intercept) and the differences in tissue and shell dry weight were particularly small at the typical size range that eastern oysters are harvested in the Northeastern US (2.5–3.5 inches, 63.5–88.9 mm; Fig 7 vertical lines).

At the same time, the range in observed values for oyster tissue dry weight across the US Northeast region was similar to the previously reported range for oyster tissue dry weight within Chesapeake Bay alone, across all oyster shell heights measured (Fig 8A). This result indicates that the variability in oyster tissue dry weight across the region as a whole was similar to the variability in oyster tissue dry weight within a single, well-sampled estuary. The similarity in the ranges of these datasets suggests that the sample size of each may be sufficient to reflect the physiological limits of the relationship between tissue dry weight vs. shell height for eastern oysters given environmental conditions in northern temperate estuaries.

The range of tissue dry weight observed within the regional, farm-based dataset was similar to that observed for the CBP dataset’s mixture of wild, farmed, and restored oysters. This similarity was a little surprising, given the efforts by many farmers to produce oysters with higher meat weight [42]. While the overall range is visually comparable, unfortunately, the central tendency in the data could not be easily compared, as the Chesapeake Bay Program separated their reporting of 50th quantiles by ploidy, and approximately 20% of the data used in the Chesapeake Bay analysis has not been made publicly available [32]. It would be worthwhile to compare the average tissue dry weight of the regional, farmed dataset to the average tissue dry weight of the CBP data if all of the data become available in the future.

The range in observed values for oyster shell dry weight across the US Northeast region was less than the previously reported range for oyster tissue dry weight within the Chesapeake Bay, across all oyster shell heights measured (Fig 8B). In general, it appears that the oysters within the regional dataset had lighter shells at a given size than oysters from the CBP, although again, direct comparison of 50th quantile regressions was not possible due to reporting differences and incomplete data. Reduced shell weights of farmed oysters have been reported previously in the literature, particularly for oysters grown in gear [42], which would be consistent with the trend observed here. While the range of tissue dry weights were similar between our farm-centric, regional data and the mixed wild/restored/farmed oyster data from Chesapeake Bay, the difference in shell weights between the two datasets highlights the value of the farm-specific approach taken here. It appears that the use of shell data from a dataset containing wild/restored oysters to calculate nitrogen removal on farms could result in an overestimate of the nitrogen removal service provided.

After consideration of the similarity in range of reported oyster tissue/shell dry weights between the Chesapeake-only and the full regional datasets, the small effect size of the comparison between New England vs. Mid-Atlantic geographic sub-regions, and the geographic limitations of the within-state data, we opted for a single tissue regression and a single shell regression across the full region. We believe this choice is a conservative approach that is well supported by available data, and this approach can be easily revisited in the future as new data is collected.

Calculation of nitrogen removal

As described above, low variation was observed in eastern oyster nitrogen concentration and morphometrics related to ploidy, cultivation practice, and location within the regional data set. Our analysis indicated that the magnitude of these differences was not sufficient to justify a more complex tool. Furthermore, discussions with industry stakeholders at a regional aquaculture conference (Northeast Aquaculture Conference and Expo 2024) and at a national aquaculture conference (Aquaculture America 2024) indicated industry preference for the simplest possible tool, in order to facilitate the broadest implementation for farms across the region. A single regional value for nitrogen concentration of oyster tissue and shell, a single regional regression describing the relationship between oyster tissue dry weight vs. shell height, and a single regional regression describing the relationship between oyster shell dry weight vs. shell height have been employed here to calculate mean nitrogen removal by farms across the US Northeast Region (North Carolina to Maine).

In the ANRC online tool, the weight of nitrogen for an individual oyster was calculated by multiplying the 50th quantile of tissue dry weight (given the mean oyster shell height at harvest; an input variable) by the regional mean tissue nitrogen concentration to obtain the tissue nitrogen (g). This process was repeated with shell dry weight and mean shell nitrogen concentration to calculate the weight of shell nitrogen (g). The two values were added together to obtain whole animal nitrogen (g). Whole animal nitrogen was then multiplied by number of individuals harvested, as input by the user, to obtain farm-scale nitrogen removal, which can be reported in either pounds or kilograms.

Nitrogen removal outputs for the harvest of 1 million oysters were calculated using the ANRC online tool, for three common sizes of harvested eastern oysters (2.5, 3, 3.5 in; Table 5). Nitrogen removal was 194 pounds for 2.5 in oysters (interquartile range 143–274), 311 pounds for 3 in oysters (IQR 234–274), and 465 pounds for 3.5 in oysters (IQR 354–611). The use of a single regression equation for oyster morphometrics on nitrogen removal was evaluated by calculating nitrogen removal using different regressions for cultivation practice (bottom vs. no gear vs. floating gear) and for ploidy (diploid vs. triploid; Table 5). When cultivation practice was varied for 3 inch oysters, the differences in nitrogen removal were very small (<4%) among the overall regression (311 pounds), floating gear (310 pounds), and bottom gear (307 pounds). Use of the overall regression resulted in a smaller calculation of nitrogen removal than would have been predicted by the regression for oysters grown without gear (311 vs. 375 pounds), but all of these values were well within the interquartile range for the full regional dataset (234–423 pounds). When ploidy was varied for 3 inch oysters, the differences in nitrogen removal were larger: 311 pounds for the overall regression, 330 pounds for diploids, and 236 pounds for triploids, however, the ploidy-based differences were again within the interquartile range for the full regional regression. We do not believe that these differences are large enough to justify the development of a more complex tool at this time.

Integration with shellfish aquaculture permit review

Due to the results of the analysis supporting the use of a single nitrogen concentration across all locations and farming practices, as well as a single regression to describe the relationship between oyster shell height and tissue/shell dry weight, the development of a simple tool to calculate nitrogen removal at the farm scale is complementary to the ease of integration into the permit review process for shellfish aquaculture in the US Northeast. The comprehensive scale of the analysis provides support for defensible integration of consideration of nitrogen removal into permitting review decisions. To further enhance ease of integration into the permitting process, we have designed the ANRC tool to both generate nitrogen removal outputs that display online in numerical and graphical format, and also in a downloadable report that can be attached to permit application materials. The report is tailored for integration with the US Army Corps of Engineers public interest review process and similar state-level permitting processes, and provides a succinct summary of the ecological services associated with nutrient removal in eutrophic locations, project-specific values, and citations supporting the calculation of those values. In discussion with US Army Corps of Engineers project managers, this information can be specifically integrated into their public interest review related to water quality effects from a proposed project. Future iterations of the tool may include information related to the eutrophication status of project areas, further emphasizing the nutrient removal service of shellfish aquaculture in impaired waters.

Project-specific considerations in the report include farm location (general description as well as GIS coordinates), harvest period, number of oysters harvested, and size of harvested oysters. This allows growers to account for nitrogen removal between different-sized oysters at harvest. As nitrogen removal is tied to the size of oysters, and oyster size at harvest may vary within and across farms, based on either market preferences and/or state regulations on minimum harvest size, the tool allows for growers to account for these differences and run individual calculations based on varying oyster harvest sizes across and/or within different harvest periods. The adjustable timeframe in the tool also provides information for managers that may want to equate nitrogen removal to a particular timeframe for an authorized permit or lease.

Comparison to other shellfish ecosystem services calculators

The ANRC joins a growing number of shellfish ecosystem services calculators available online, while providing a unique combination of functionality and geographic range. The Nature Conservancy maintains an online Oyster Calculator (https://oceanwealth.org/tools/oyster-calculator/) for restored oyster reefs, based on peer-reviewed syntheses of the scientific literature. The Oyster Calculator provides the volume of water filtered for locations around the United States using methods described in [43], and subsequently expanded with environmental data reported in [44]. Custom locations outside the existing database can also be used if data on bay volume, residence time, and temperature are provided by the user. The Oyster Calculator also outputs enhanced production of wild fish and invertebrates from habitat provided by restored oyster reefs for the Northern Gulf of Mexico, Floridian, Carolinian, and Virginian ecoregions, using methods described in [44]. This calculator is focused on restored oyster reefs, and does not currently calculate oyster nitrogen removal.

Rutgers University maintains an online Oyster Eco-Serve Calculator (https://itsappserver.sebs.rutgers.edu/Oysterecoservecalc/) for eastern oyster farms from Massachusetts to Virginia, USA. The Oyster Eco-Serve Calculator provides the volume of water filtered, the total mass of particles removed through oyster filtration, and the total mass of particles added to the ecosystem through oyster waste, on an annual, seasonal, and daily basis. The data and methods supporting the calculator are described in [7]. All of the data supporting this calculator were collected in the field on working oyster farms, but the calculator does not currently calculate oyster nitrogen removal at harvest.

The University of Florida maintains an online Florida Shellfish Farm Nitrogen Calculator (https://ufl864.outgrow.us/ufl864-2) for eastern oyster and hard clam (Mercenaria mercenaria) farms within the state of Florida, USA. The calculator outputs nitrogen removal from shellfish harvest, from enhancement of denitrification in sediments underneath farms, and a potential monetary value of the combined nitrogen removal service. All of the data for oyster harvest nitrogen were collected from farmed oysters in Florida, the denitrification enhancement was measured in the field on working oyster and clam farms, and the value of this service was calculated based on a willingness to pay survey of Florida wastewater treatment plants. This calculator is currently intended only for use by Florida oyster and clam farms.

All of these calculators provide different and valuable information on shellfish ecosystem services, based on robust scientific data collection. The ANRC adds to this existing body of work by providing a robust calculation of the typical expected harvest-based nitrogen removal for any eastern oyster farm from North Carolina to Maine, USA, and by generating outputs that are intended to directly integrate into state and federal shellfish aquaculture permitting processes that allow for consideration of environmental benefits in the permit application review. We compared our results to the Florida Shellfish Farm Nitrogen Calculator using inputs of 1,000,000 oysters harvested at a shell height of 3 inches. The Florida Calculator predicted nitrogen removal of 265 pounds, and the ANRC tool predicted 311.6 pounds of nitrogen. We also compared our tissue-only nitrogen removal results to the Chesapeake Bay diploid regression. The Chesapeake Bay diploid tissue regression predicted nitrogen removal of 198 pounds, and the ANRC tool predicted 194.5 pounds.

The approach described here is highly transferable to other locations and other shellfish species for which sufficient data exists to predict nitrogen removal by shellfish farms with high confidence. Given the low observed variation in oyster tissue and shell nitrogen concentration across the geographic region, ploidy, and common cultivation practices in this study, the data needed to accurately quantify whole animal nitrogen concentration for other species and locations should be relatively modest. While data associated with oyster morphometrics were more variable, other shellfish species may exhibit lower variability in weight-to-length ratios than oysters, potentially reducing data needs. Furthermore, measurements of dry weight and length are much less expensive and less technologically challenging than measurements of nitrogen concentration, making these data easier to obtain.

Conclusions

The ANRC is a valuable new tool for the shellfish aquaculture industry and for the shellfish aquaculture permit review process, and is directly responsive to management needs. The size and breadth of the dataset compiled in this study gives high confidence in the farm-scale harvest nitrogen removal calculations, and the report generated is designed to be easily integrated into the permit application for a new or expanding farm. The high transferability of the approach employed here increases the potential future application of this work to a variety of bivalve shellfish species farmed in eutrophic locations around the world.

• 1.Alleway HK, Gillies CL, Bishop MJ, Gentry RR, Theuerkauf SJ, Jones R. The Ecosystem Services of Marine Aquaculture: Valuing Benefits to People and Nature. BioScience. 2019;69(1):59–68. doi: 10.1093/biosci/biy137 [DOI] [Google Scholar]
• 2.Gentry RR, Alleway HK, Bishop MJ, Gillies CL, Waters T, Jones R. Exploring the potential for marine aquaculture to contribute to ecosystem services. Reviews in Aquaculture. 2020;12(2):499–512. doi: 10.1111/raq.12328 [DOI] [Google Scholar]
• 3.Petrolia DR, Nyanzu F, Cebrian J, Harri A, Amato J, Walton WC. Eliciting expert judgment to inform management of diverse oyster resources for multiple ecosystem services. Journal of Environmental Management. 2020;268:110676. doi: 10.1016/j.jenvman.2020.110676 [DOI] [PubMed] [Google Scholar]
• 4.Lindahl O, Hart R, Hernroth B, Kollberg S, Loo L-O, Olrog L, et al. Improving marine water quality by mussel farming—a profitable solution for Swedish society. Ambio. 2005;34(2):129–36. [PubMed] [Google Scholar]
• 5.Petersen JK, Hasler B, Timmermann K, Nielsen P, Tørring DB, Larsen MM, et al. Mussels as a tool for mitigation of nutrients in the marine environment. Marine Pollution Bulletin. 2014;82(1–2):137–43. doi: 10.1016/j.marpolbul.2014.03.006 [DOI] [PubMed] [Google Scholar]
• 6.Rose JM, Bricker SB, Tedesco MA, Wikfors GH. A Role for Shellfish Aquaculture in Coastal Nitrogen Management. Environmental Science & Technology. 2014;48(5):2519–25. doi: 10.1021/es4041336 [DOI] [PubMed] [Google Scholar]
• 7.Barr JM, Munroe D, Rose JM, Calvo L, Cheng KM, Bayer S, et al. Seasonal Feeding Behavior of Aquaculture Eastern Oysters (Crassostrea virginica) in the Mid-Atlantic. Estuaries and Coasts. 2023. doi: 10.1007/s12237-023-01293-9 [DOI] [Google Scholar]
• 8.Hanley LT. Wave attenuation through submerged oyster aquaculture cages: University of Maine; 2023. [Google Scholar]
• 9.Zhu L, Huguenard K, Zou Q-P, Fredriksson DW, Xie D. Aquaculture farms as nature-based coastal protection: Random wave attenuation by suspended and submerged canopies. Coastal Engineering. 2020;160:103737. 10.1016/j.coastaleng.2020.103737. [DOI] [Google Scholar]
• 10.Theuerkauf SJ, Barrett LT, Alleway HK, Costa-Pierce BA, St. Gelais A, Jones RC. Habitat value of bivalve shellfish and seaweed aquaculture for fish and invertebrates: Pathways, synthesis and next steps. Reviews in Aquaculture. 2021;n/a(n/a). 10.1111/raq.12584. [DOI] [Google Scholar]
• 11.Barrett LT, Theuerkauf SJ, Rose JM, Alleway HK, Bricker SB, Parker M, et al. Sustainable growth of non-fed aquaculture can generate valuable ecosystem benefits. Ecosystem Services. 2022;53:101396. 10.1016/j.ecoser.2021.101396. [DOI] [Google Scholar]
• 12.Michaelis AK, Walton WC, Webster DW, Shaffer LJ. Cultural ecosystem services enabled through work with shellfish. Marine Policy. 2021;132:104689. 10.1016/j.marpol.2021.104689. [DOI] [Google Scholar]
• 13.Kotta J, Futter M, Kaasik A, Liversage K, Rätsep M, Barboza FR, et al. Cleaning up seas using blue growth initiatives: Mussel farming for eutrophication control in the Baltic Sea. Science of The Total Environment. 2020;709:136144. doi: 10.1016/j.scitotenv.2019.136144 [DOI] [PubMed] [Google Scholar]
• 14.Rose JM, Bricker SB, Deonarine S, Ferreira JG, Getchis T, Grant J, et al. Nutrient Bioextraction. In: Meyers RA, editor. Encyclopedia of Sustainability Science and Technology. New York, NY: Springer New York; 2015. p. 1–33. [Google Scholar]
• 15.Taylor D, Saurel C, Nielsen P, Petersen JK. Production Characteristics and Optimization of Mitigation Mussel Culture. Frontiers in Marine Science. 2019;6:698. [Google Scholar]
• 16.Breitburg D, Levin LA, Oschlies A, Grégoire M, Chavez FP, Conley DJ, et al. Declining oxygen in the global ocean and coastal waters. Science. 2018;359(6371):eaam7240. doi: 10.1126/science.aam7240 [DOI] [PubMed] [Google Scholar]
• 17.Bricker SB, Longstaff B, Dennison W, Jones A, Boicourt K, Wicks C, et al. Effects of nutrient enrichment in the nation’s estuaries: A decade of change. Harmful Algae. 2008;8(1):21–32. doi: 10.1016/j.hal.2008.08.028 [DOI] [Google Scholar]
• 18.Greening H, Janicki A. Toward reversal of eutrophic conditions in a subtropical estuary: water quality and seagrass response to nitrogen loading reductions in Tampa Bay, Florida, USA. Environmental Management. 2006;38(2):163–78. doi: 10.1007/s00267-005-0079-4 [DOI] [PubMed] [Google Scholar]
• 19.Boesch DF. Barriers and Bridges in Abating Coastal Eutrophication. Frontiers in Marine Science. 2019;6. doi: 10.3389/fmars.2019.00123 [DOI] [Google Scholar]
• 20.Beseres Pollack J, Yoskowitz D, Kim H-C, Montagna PA. Role and value of nitrogen regulation provided by oysters (Crassostrea virginica) in the Mission-Aransas estuary, Texas, USA. PLoS One. 2013;8(6):e65314. doi: 10.1371/journal.pone.0065314 [DOI] [PMC free article] [PubMed] [Google Scholar]
• 21.Kellogg ML, Cornwell JC, Owens MS, Paynter KT. Denitrification and nutrient assimilation on a restored oyster reef. Marine Ecology Progress Series. 2013;480:1–19. [Google Scholar]
• 22.Ray NE, Fulweiler RW. Meta-analysis of oyster impacts on coastal biogeochemistry. Nature Sustainability. 2021;4(3):261–9. doi: 10.1038/s41893-020-00644-9 [DOI] [Google Scholar]
• 23.Clements JC, Comeau LA. Nitrogen removal potential of shellfish aquaculture harvests in eastern Canada: A comparison of culture methods. Aquaculture Reports. 2019;13:100183. 10.1016/j.aqrep.2019.100183. [DOI] [Google Scholar]
• 24.Nielsen P, Cranford PJ, Maar M, Petersen JK. Magnitude, spatial scale and optimization of ecosystem services from a nutrient extraction mussel farm in the eutrophic Skive Fjord, Denmark. Aquaculture Environment Interactions. 2016;8:311–29. [Google Scholar]
• 25.Reitsma J, Murphy DC, Archer AF, York RH. Nitrogen extraction potential of wild and cultured bivalves harvested from nearshore waters of Cape Cod, USA. Marine Pollution Bulletin. 2017;116(1):175–81. doi: 10.1016/j.marpolbul.2016.12.072 [DOI] [PubMed] [Google Scholar]
• 26.Sebastiano D, Levinton JS, Doall M, Kamath S. Using a Shellfish Harvest Strategy to Extract High Nitrogen Inputs in Urban and Suburban Coastal Bays: Practical and Economic Implications. Journal of Shellfish Research. 2015;34(2):573–83, 11. [Google Scholar]
• 27.Reichert-Nguyen J, Cornwell J, Rose J, Kellogg L, Luckenbach M, Bricker S, et al. Panel recommendations on the oyster BMP nutrient and suspended sediment reduction effectiveness determination decision framework and nitrogen and phosphorus assimilation in oyster tissue reduction effectiveness for oyster aquaculture practices. Report to the Chesapeake Bay Program. Available online at https://www.oysterrecovery.org/wp-content/uploads/2017/01/Oyster-BMP-1st-Report_Final_Approved_2016-12-19.pdf: 2016. [Google Scholar]
• 28.Town of Mashpee Sewer Commission. Comprehensive watershed nitrogen management plan, Town of Mashpee. Available online at http://www.mashpeewaters.com/documents.html: 2015. [Google Scholar]
• 29.Geist G. Falmouth looks to vastly expand public-private aquaculture program. CapeNewsnet. 2023. Oct 3;Sect. Falmouth News. [Google Scholar]
• 30.Rogers D, Szalay T, Edgcomb V, Mara P, Martinsen C, Lovely C, et al. Best practices guide for nitrogen remediation using oyster aquaculture. Report to the Waquoit Bay NERR. Available online at https://nerrssciencecollaborative.org/resource/best-practices-guide-nitrogen-remediation-using-oyster-aquaculture: 2021. [Google Scholar]
• 31.Morse R, Rose JM, Ayvazian SG, Barr JM, Bayer S, Brady DC, et al. Morphometrics and nutrient concentration of farmed eastern oysters (Crassostrea virginica) sampled from the Northeastern United States Zenodo. 2024. Available from: doi: 10.5281/zenodo.11966672 [DOI] [Google Scholar]
• 32.Morse R, Rose JM, Reichert-Nguyen J, Bricker SB, Brown BL, Higgins CB, et al. "Morphometrics and nutrient concentration of eastern oysters (Crassostrea virginica) sampled from Chesapeake Bay, USA" in Panel Recommendations on the Oyster BMP Nutrient and Suspended Sediment Reduction Effectiveness Determination Decision Framework and Nitrogen and Phosphorus Assimilation in Oyster Tissue Reduction Effectiveness for Oyster Aquaculture Practices. Zenodo. 2024. Available from: doi: 10.5281/zenodo.11205266 [DOI] [Google Scholar]
• 33.Wilcox RR. Introduction to robust estimation and hypothesis testing London: Elsevier; 2022. 928 p. [Google Scholar]
• 34.Higgins CB, Stephenson K, Brown BL. Nutrient bioassimilation capacity of aquacultured oysters: quantification of an ecosystem service. Journal of Environmental Quality. 2011;40:271–7. doi: 10.2134/jeq2010.0203 [DOI] [PubMed] [Google Scholar]
• 35.Poach M, Morse R, Meseck SL, Alvarado A, Reichert-Nguyen J, McFarland K, et al. Nutrient reduction by eastern oysters exhibits low variability associated with reproduction, ploidy, and farm location. Marine Pollution Bulletin. 2024;202:116286. doi: 10.1016/j.marpolbul.2024.116286 [DOI] [PubMed] [Google Scholar]
• 36.Wilcox RR. A guide to robust statistical methods: Springer; 2023. 343 p. [DOI] [PubMed] [Google Scholar]
• 37.Koenker R. Quantile regression in R: a vignette. Available online at http://www.econ.uiuc.edu/~roger/research/rq/vig.pdf2006. [Google Scholar]
• 38.Koenker R. Package ’quantreg’: Quantile Regression. Available online from the R Development Core Team: https://cran.r-project.org/web/packages/quantreg/quantreg.pdf; 2016. [Google Scholar]
• 39.Cornwell J, Bricker S, Lacatell A, Luckenbach M, Marenghi F, Moore C, et al. Oyster BMP Expert Panel Second Incremental Report. Report to the Chesapeake Bay Program. Available online at https://d18lev1ok5leia.cloudfront.net/chesapeakebay/documents/Oyster-BMP-Expert-Panel-Second-Report.pdf: 2023. [Google Scholar]
• 40.Sherman K. Adaptive management institutions at the regional level: The case of Large Marine Ecosystems. Ocean & Coastal Management. 2014;90:38–49. 10.1016/j.ocecoaman.2013.06.008. [DOI] [Google Scholar]
• 41.Stow CA, Glassner-Shwayder K, Lee D, Wang L, Arhonditsis G, DePinto JV, et al. Lake Erie phosphorus targets: An imperative for active adaptive management. Journal of Great Lakes Research. 2020;46(3):672–6. 10.1016/j.jglr.2020.02.005. [DOI] [Google Scholar]
• 42.Mizuta DD, Wikfors GH. Seeking the perfect oyster shell: a brief review of current knowledge. Reviews in Aquaculture. 2019;11(3):586–602. 10.1111/raq.12247. [DOI] [Google Scholar]
• 43.zu Ermgassen PSE, Spalding MD, Grizzle RE, Brumbaugh RD. Quantifying the Loss of a Marine Ecosystem Service: Filtration by the Eastern Oyster in US Estuaries. Estuaries and Coasts. 2013;36(1):36–43. doi: 10.1007/s12237-012-9559-y [DOI] [Google Scholar]
• 44.zu Ermgassen PSE, Grabowski JH, Gair JR, Powers SP. Quantifying fish and mobile invertebrate production from a threatened nursery habitat. Journal of Applied Ecology. 2016;53(2):596–606. doi: 10.1111/1365-2664.12576 [DOI] [Google Scholar]