Quantifying the Loss of a Marine Ecosystem Service: Filtration by the Eastern Oyster in US Estuaries
The oyster habitat in the USA is a valuable resource that has suffered significant declines over the past century. While this loss of habitat is well documented, the loss of associated ecosystem services remains poorly quantified. Meanwhile, ecosystem service recovery has become a major impetus for restoration. Here we propose a model for estimating the volume of water filtered by oyster populations under field conditions and make estimates of the contribution of past (c. 1880–1910) and present (c. 2000–2010) oyster populations to improving water quality in 13 US estuaries. We find that filtration capacity of oysters has declined almost universally (12 of the 13 estuaries examined) by a median of 85 %. Whereas historically, oyster populations achieved full estuary filtration (filtering a volume equivalent or larger than the entire estuary volume within the residence time of the water) in six of the eight estuaries in the Gulf of Mexico during summer months, this is now the case for only one estuary: Apalachicola Bay, Florida. By contrast, while all five estuaries on the North Atlantic coast showed large decreases in filtration capacity, none were achieving full estuary filtration at the time of our c. 1900 historic baseline. This apparent difference from the Gulf of Mexico is explained at least in part by our North Atlantic baseline representing a shifted baseline, as surveyed populations were already much reduced by exploitation in this region.
KeywordsCrassostrea virginica USA Restoration Estuarine habitat Historical ecology Water quality
The accelerated loss of many habitats since the industrial revolution is widely documented (Winslow 1887; Roberts 2002; Fearnside 2005); however, the ecological and social ramifications of this loss have only recently gained recognition (Millennium Ecosystem Assessment 2005). Quantifying losses and the impact of habitat alteration is challenging, as long-term data on habitat area or condition are rare. The vast majority of habitats represented in historic datasets lack detailed insights beyond areal extent (e.g., Sommer 1976; Bromberg and Bertness 2005), yet habitat degradation is frequently a threat in addition to habitat loss (Lambin 1999; Turner et al. 1999; zu Ermgassen et al. 2012). In many marine environments, it is difficult, if not impossible, to determine the nature of pristine habitats that have been lost (Roberts 2007). This is problematic in deriving estimates of ecosystem service provision historically, as many services are strongly dependent on species abundance or species richness (Diaz et al. 2006; Gibbs et al. 2007). In these cases, knowledge of the change in habitat quality may be necessary to determine the change in ecosystem service delivery as habitats degrade.
Oyster reefs are among the most threatened of marine habitats having suffered substantial declines globally over the past century (Beck et al. 2011), primarily due to overfishing, hydrological changes, pollution, and disease (Winslow 1887; Mackenzie 2007; Powell et al. 2008; Wilberg et al. 2011). These losses have been quantified for the USA, where zu Ermgassen et al. (2012) utilized data on oyster size and density alongside spatial extent measures to derive estimates of a 64 % decline in oyster extent and 88 % loss of oyster biomass between the early 1900s and the early 2000s. Such measures underpin efforts to formulate estimates of the loss of a critical coastal ecosystem service—water filtration.
Filtration by suspension feeders such as oysters impacts water quality by directly removing particulate matter from the water column, with oysters ingesting the edible particles and binding rejected particles with mucus, then depositing this as pseudofeces onto the sediment surface. Through this action, both phytoplankton and suspended sediment that would otherwise reduce water clarity are drawn from the water column to the benthos. The eastern oyster, Crassostrea virginica (Gmelin, 1791), is known to filter particles >5 μm in size with high efficiency (Riisgaard 1988) and can have a marked effect on light penetration in shallow, calm waters, which in turn may have positive impacts on other important adjacent habitats, such as seagrass beds (Newell and Koch 2004).
The physiology of oyster feeding and filtration is well studied, and filtration rate is known from laboratory studies to be influenced by a variety of environmental factors such as temperature, flow rate, salinity, seston concentration, and particle size, as well as oyster size (Loosanoff 1958; Walne 1972; Shumway et al. 1985; Riisgaard 1988). Although it is recognized that conditions in situ may significantly affect filtration rate (Doering and Oviatt 1986; Powell et al. 1992), few studies have sought to quantify rates under field conditions. To estimate the contribution of oyster filtration within an estuary, one can model field-based filtration rates by summing the effect of environmental variables, for which the relationships have been derived and for which field measurements are available. This has been done effectively for Chesapeake Bay (Cerco and Noel 2005; Fulford et al. 2007); however, such methods are expensive and require the input of a large amount of environmental data and computational capacity (Cerco and Noel 2005). While the required environmental data are available for well-studied estuaries such as Chesapeake Bay, they are not available for most other estuaries, and therefore, such high-resolution models are not widely applicable.
In addition to the challenges posed by the lack of data, there remain concerns that oysters in situ may not respond as predicted by models primarily based on filtration by few oysters in the laboratory. Oyster populations in situ may spend a different proportion of time with their valves shut, there is the potential for synergistic population-level influences and re-filtration, and physical attributes of the reef may influence flow dynamics and hence the uptake of particles (Dame et al. 1984; Harsh and Luckenbach 1999). Despite these concerns, it seems that models may be the only means available to estimate filtration rates at large scales.
In order to make broad, estuary-scale estimates of the volume of water filtered by oyster populations nationally, we derived a model of oyster filtration based on the only in situ measurements of oyster filtration currently available (Grizzle et al. 2006, 2008). We use this model to explore the estimated delivery of this ecosystem service historically versus presently across 13 US estuaries.
Overview of variables effecting oyster filtration rate
Effect on filtration rate
Unimodal with optimum filtration at ~27 °C
Newell and Langdon (1996)
Mean available nationally
Steep decline in filtration rate between 14 and 10 ppt
Varies dramatically spatially within estuaries
Unknown. Modeled as strong decrease <2 mg/l
Cerco and Noel (2005)
Rare, variable at a small spatial scale
Retain particles >5 μm at high efficiency
Rare, varies seasonally
Maximal when seston concentration >5 and <10 mg/l
Rare, varies seasonally
Effect poorly understood
Rare, variable at a small spatial scale
Increases as a function of dry mass by an exponent of 0.58
Mean available nationally
In order to account for field conditions in our model, we used the field measurements of seston uptake over natural oyster reefs reported in Grizzle et al. (2006, 2008) to estimate filtration rates. Grizzle et al. measured seston drawdown over reefs with a known mean oyster SH and density, and under optimal temperature conditions. We assumed that all seston drawdown was the result of filtration by oysters and estimated the mean filtration rate per oyster for each reef surveyed. We subsequently fitted a standard filtration model (Eq. 1) to these field data. SH to dry tissue mass conversions from the respective regions (South Carolina and Florida) were applied. Field data were collected on intertidal reefs during both ebb and flood tides (Grizzle et al. 2006, 2008) and therefore captured the impact of the tidal cycle (Dame et al. 1984, 1992). Negative values caused by waves or other disturbances were excluded.
Filtration by historic and present oyster populations, and for 1 ha of oyster ground at either historic or present oyster densities
Historic volume filtered (1,000 m3 h−1)
Present volume filtered (1,000 m3 h−1)
Historic area (ha) and density (ind m−2)
Present area (ha) and density (ind m−2)
Volume filtered by 1 ha at historic density (1,000 m3 h−1)
Volume filtered by 1 ha at present density (1,000 m3 h−1)
Hudson River/Raritan Bay
Tangier and Pocomoke Sound
West Mississippi Sound
San Antonio Bay
Corpus Christi Bay
We collated monthly mean water temperature data for all 13 estuaries from NOAA, USGS, and the National Estuarine Research Reserve network. We then applied our field-based filtration model to the present and historic extent, and the mean density of both size classes of oyster in each estuary, to determine the change in this ecosystem service over time. As we know of no published data illustrating differences in filtration rate between intertidal and subtidal oysters of the same size, we made no alteration to the model in order to represent subtidal reefs. We assumed no change in mean monthly water temperature or SH–biomass relationship between time periods. For each estuary, we estimated the filtration capacity of the population historically and presently, across all seasons. We summarized and examined the change in the total volume filtered and the volume filtered per unit area over time. We used the Sharipo–Wilk test to determine whether data were non-normally distributed. All statistics were run in R version 2.13.1 (2011-07-08).
We then estimated the potential ecological impact of the change, by relating the volume filtered to the estuary volume and residence time listed in Bricker et al. (2007). Residence time is defined as the mean transit time of freshwater through the estuary. We use the term full estuary filtration to describe the situation where the oyster population filters more than the entire volume of the estuary within the residence time of water in that estuary. We recognize that this does not equate to complete filtration of all estuarine waters and does not account for phytoplankton production, but believe that this number nonetheless provides a useful indicator of the volume of filtration relative to water flow (Smaal and Prins 1993; Dame 2011).
We estimate that reefs with oyster densities typical of modern oyster populations would filter a median of 0.15 × 103 m3 ha−1 h−1 (range, 0.06 × 103 to 6.47 × 103 m3 ha−1 h−1) in summer months, as compared to 0.92 × 103 m3 ha−1 h−1 historically (range, 0.1 × 103 to 5.74 × 103 m3 ha−1 h−1). An overview of estuary-specific mean filtration rates is provided in Table 2.
Proportion of the volume of each estuary that may be filtered by the historic and current populations of oysters, across seasons
Historic proportion of bay filtered with the residence time
Present proportion of bay filtered with the residence time
Volume (1,000 m3)
Residence time (days)
Mean % change
Hudson River/Raritan Bay
Tangier and Pocomoke Sounds
West Mississippi Sound
San Antonio Bay
Corpus Christi Bay
Coastal systems have undergone unprecedented change over the past century (Jackson et al. 2001). While the role that oysters play in improving water quality is increasingly valuable given the high incidence of eutrophication (Bricker et al. 2007), the provision of this ecosystem service has declined precipitously over the past century (Table 3, Fig. 2).
We report a near universal decline of the filtering capacity of oyster habitats by more than 80 %. Our results suggest that 100 years ago, filtration by oysters was likely to have been a major ecological function, achieving full estuary filtration in many estuaries (Table 3, Fig. 2). This reduction in filtration capacity is likely to have caused substantial changes to the ecosystem function of estuaries (Newell 1988; Dame et al. 2002). While the ecological importance of a healthy oyster population has been the focus of much attention in the Chesapeake (e.g., Ulanowicz and Tuttle 1992; Fulford et al. 2007), the decline in the ecosystem services provided by healthy oyster habitats in US estuaries more generally remains poorly appreciated and understudied. Although increases in oyster filtration alone are unlikely to resolve the water quality concerns of many US estuaries (Cerco and Noel 2007), it may be possible to restore this ecosystem service to levels at which it will have some beneficial ecological impact locally (Newell and Koch 2004).
The decline in filtration capacity in estuaries spanning such a large area is striking. While the decline in filtration capacity appears to have been greater in the Gulf of Mexico than in the North Atlantic (Table 3, Fig. 2), this is certainly at least in part because the surveys conducted along the North Atlantic coast were undertaken after exploitation rates had peaked and thus reflected an already shifted baseline with low oyster densities (zu Ermgassen et al. 2012, Table 2). For example, our 1887 historic baseline for Tangier and Pocomoke Sound (Chesapeake Bay, Atlantic coast) documented oyster densities 2–7-fold lower than observations from 30 years earlier (Winslow 1887), while even this earlier 1850s estimate is likely to have been dramatically lower than precolonial densities (Newell 1988; Kirby 2004). By contrast, the historic surveys undertaken in the Gulf of Mexico were conducted with the caveat that many oyster reefs remained undiscovered (Moore 1907). As such, our estimates of the historic level of services in the northeast likely represent a significantly shifted baseline.
All historic baselines should be considered relative to their date of origin, as they may be shifted from pristine conditions (Roberts 2007). This non-pristine baseline is further supported by observations that mean SH in precolonial shell deposits in South Carolina were found to be 62 % greater than mean harvested SH in 1938 (Lunz 1938). Such age and size truncation is typical of harvested species (Hutchings and Reynolds 2004) and would have a marked impact on population-level filtration rates (Mann et al. 2009a, b). If a moderate correction of assuming four-times-higher densities (6.5 oysters/m2) and precolonial size distributions is applied to the historic extent in Tangier and Pocomoke Sound, this would result in approximately a 7-fold increase in filtration capacity and the estuary being filtered multiple times within its residence time. As such, it is reasonable to expect that oysters in many of the northeastern US estuaries would have exerted full estuary filtration capacity prior to the industrial exploitation of oysters.
We concur with a number of authors who have previously asserted that oysters would historically have been dominant filter feeders with significant ecological impacts through filtration in many estuaries (Newell 1988; Cerco and Noel 2007; Mann et al. 2009a, b). Our historical data represent a shifted baseline, particularly in the northeastern USA, but such quantitative historical data help to avoid further shifts in baselines and ensure that modern management goals are not misguided (Swetnam et al. 1999). Modern management goals should utilize such historical data, but in a broader context. For example, ecologically relevant restoration goals might best focus on delivery of the desired ecosystem services (Jackson and Hobbs 2009), aided by our understanding of their relative importance in the face of a changing environmental landscape (Swetnam et al. 1999). Our model and bay-specific data provide the basis for such comprehensive goal setting.
In contrast to the situation on the Atlantic Coast, it is noteworthy that for Apalachicola Bay in the Gulf of Mexico, we estimate present day levels of filtration that are greater than historic estimates. Apalachicola Bay has undergone intensive management of the oyster resource including extensive shell planting since the historic survey efforts which has led to increased areal extent of oysters, in addition to the recorded densities being higher (zu Ermgassen et al. 2012). It is noteworthy that this is one of the few estuaries in which harvesting is primarily by tonging, and dredging is not allowed. It is widely agreed that tonging is a less-destructive harvest method than dredging (Lenihan and Peterson 2004), and further work may well reveal that this has been a critical factor in ensuring the long-term sustainable benefits that appear to be delivered in these estuaries.
Our model of filtration rates is the first to incorporate field measurements from in situ oyster populations, and therefore to represent whole-reef filtration. When converted to the same unit, the filtration rate estimated by our model is lower than the maximum filtration rate of 11.5 l h−1 used by Cerco and Noel (2007) in their model for the Chesapeake. This is as would be expected if the use of field data were, as we have assumed, a useful surrogate where the environmental data required to modify maximum filtration rate are absent.
Considering filtration relative to residence time can be a useful indicator of the potential for oysters to have an ecologically significant impact on an estuary (Dame 2011). However, even when undisturbed, estuaries exhibit high variability in sediment load, planktonic productivity, depth, residence time, and natural abundance of oysters (Bricker et al. 2007; zu Ermgassen et al. 2012), all of which influence the extent to which oyster filtration may impact water quality (Officer et al. 1982; Pomeroy et al. 2006; Cerco and Noel 2007). Temporal mismatching between phytoplankton production and peak oyster filtration may also limit the potential for oyster populations to have a regulating influence. Additionally, the impact of filtration on seston drawdown on large scales cannot be directly inferred from filtration rates due to the influences of wave action (Porter et al. 2004), the unequal distribution of oysters (Cerco and Noel 2007), and imperfect mixing within the estuary (Pomeroy et al. 2006). This in turn may lead to variable impacts of oyster filtration on nutrient cycling within the bay, as the biodeposition of seston may stimulate enhanced denitrification and anammox in the sediments (Dame 2011). In this context, the point at which the population filtration rate matches the residence time simply represents a point on a continuum, albeit one that may provide a useful guide for estimating the ecological impact.
While our estimates of change over time stress the extreme nature of the losses of filtration by oysters, our results also highlight that changes in management of oyster reefs, particularly in their diminished condition, can have a significant influence on the amount of filtration provided in the future. In particular, management decisions that lead to higher average densities on existing reef footprint, or expansions of reef area through restoration coupled with increasing oyster size or density, could move a number of estuaries toward a state where oysters can once more play a role in supporting water quality (Table 3).
Currently, oyster restoration efforts and goals are accounted for by their areal extent (e.g., NOAA 2012); however, the volume of water filtered by oysters is not solely a function of the area of oyster habitat, but also of the density and size frequency of the oysters. Therefore, as restoration of oyster habitat for ecosystem services gains momentum, it will be critical to devise appropriate metrics to assess the contribution of restoration projects toward those target ecosystem services. Any restoration undertaken with the goal of water filtration needs to account for the density and size distribution of oysters in addition to the area restored. Without these data, the contribution of restoration projects and their progress toward their ecosystem service goals will remain unknown.
This study provides a unique numerical insight into the dramatic functional changes that can accompany the degradation of an estuarine habitat. Even from “non-pristine” historical baselines, it is clear that a significant and nearly universal loss of ecosystem services has occurred in US estuaries, which has gone hand in hand with the loss and degradation of oyster reef habitat.
This work was supported by a grant from the National Fish and Wildlife Foundation (award no. 2009-0078-000) and by the National Partnership between The Nature Conservancy and the National Oceanic and Atmospheric Administration Community-based Restoration Program (award nos. NA07NMF4630136 and NA10NMF463008). Additional funding support for the project was provided by The Nature Conservancy-Shell Partnership and The Turner Foundation, Inc. The authors would like to thank D. Bushek, S. Geiger, and M. LaPeyre for preparing and sharing shell height–biomass relationships and J.R. Gair for assistance with model development.
This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.
- Beck, M.W., R.D. Brumbaugh, L. Airoldi, A. Carranza, L.D. Coen, C. Crawford, O. Defeo, G.J. Edgar, B. Hancock, M. Kay, H.S. Lenihan, M.W. Luckenbach, C.L. Toropova, G. Zhang, and X. Guo. 2011. Oyster reefs at risk and recommendations for conservation, restoration and management. BioScience 61: 107–116.CrossRefGoogle Scholar
- Bricker, S., B. Longstaff, W. Dennison, A. Jones, K. Boicourt, C. Wicks, and J. Woerner. 2007. Effects of nutrient enrichment in the nation's estuaries: a decade of change. Silver Spring: National Centers for Coastal Ocean Science.Google Scholar
- Cerco, C.F., and M.R. Noel. 2005. Assessing a ten-fold increase in the Chesapeake Bay native oyster population. A report to the EPA Chesapeake Bay Program. Vicksburg: US Army Engineer Research and Development Center.Google Scholar
- Cerco, C.F., and M.R. Noel. 2007. Can oyster restoration reverse cultural eutrophication in Chesapeake Bay? Estuaries and Coasts 30: 331–343.Google Scholar
- Churchill Jr., E.P. 1920. The oyster and the oyster industry of the Atlantic and Gulf coasts. Appendix VIII to the report of the U. S. Commissioner of Fisheries for 1919. Washington D.C.: Department of Commerce, Government Printing Office.Google Scholar
- Harsh, D.A. and M.W. Luckenbach. 1999. Materials processing by oysters in patches: interactive roles of current speed and seston composition. In: M.W. Luckenbach, R. Mann and J.A. Wesson (eds) Oyster reef habiatat restoration: a synopsis and synthesis of approaches. Virginia Institute of Marine Science, School of Marine Science, College of William and Mary, VIMS Press, pp. 251–265.Google Scholar
- Jackson, J.B.C., M.X. Kirby, W.H. Berger, K.A. Bjorndal, L.W. Botsford, B.J. Bourque, R.H. Bradbury, R. Cooke, J. Erlandson, J.A. Estes, T.P. Hughes, S. Kidwell, C.B. Lange, H.S. Lenihan, J.M. Pandolfi, C.H. Peterson, R.S. Steneck, M.J. Tegner, and R.R. Warner. 2001. Historical overfishing and the recent collapse of coastal ecosystems. Science 293: 629–637.CrossRefGoogle Scholar
- Lenihan, H.S., and C.H. Peterson. 2004. Conserving oyster reef habitat by switching from dredging and tonging to diver hand-harvesting. Fishery Bulletin 102: 298–305.Google Scholar
- Liddel, M. K. 2008. A von Bertalanffy based model for the estimation of oyster (Crassostrea virginica) growth on restored oyster reefs in Chesapeake Bay. PhD thesis, Marine, Estuarine and Environmental Sciences, University of Maryland. p. 159.Google Scholar
- Millennium Ecosystem Assessment. 2005. Ecosystems and human well-being: current state and trends. Washington D.C.: Island Press.Google Scholar
- Moore, H.F. 1907. Survey of oyster bottoms in Matagorda Bay. Washington D.C.: GPO.Google Scholar
- Moore, H.F. 1910. Condition and extent of the oyster beds of James River, Virginia. Washington, D.C.: GPO.Google Scholar
- Newell, R.I.E. 1988. Ecological changes in Chesapeake Bay: are they the result of overharvesting the American oyster, Crassostrea virginica? Understanding the estuary: advances in Chesapeake Bay Research. Baltimore: Chesapeake Research Consortium Publication.Google Scholar
- Newell, R.I.E., and C.J. Langdon. 1996. Mechanisms and physiology of larval and adult feeding. In The eastern oyster Crassostrea virginica, ed. V.S. Kennedy, R.I.E. Newell, and A.F. Eble, 185–229. A Maryland Sea Grant Book: College Park.Google Scholar
- NOAA. 2012. National Oceanic and Atmospheric Administration restoration atlas. . http://seahorse2.nmfs.noaa.gov/restoration_atlas/src/html/index.html. Accessed Feb 3, 2012.
- Powell, E.N., E.E. Hofmann, J.M. Klinck, and S.M. Ray. 1992. Modeling oyster populations I. A commentary on filtration rate. Is faster always better? Journal of Shellfish Research 11: 387–398.Google Scholar
- Press, W.H., S.A. Teukolsky, W.T. Vetterling, and B.P. Flannery. 2007. Numerical recipes: the art of scientific computing. New York: Cambridge University Press.Google Scholar
- Roberts, C.M. 2007. The unnatural history of the sea. The past and future of humanity and fishing. London: Gaia.Google Scholar
- Smaal, A.C., and T.C. Prins. 1993. The uptake of organic matter and the release of inorganic nutrients by bivalve suspension feeder beds. In Bivalve filter feeders in estuarine and coastal ecosystem processes, ed. R.F. Dame, 273–298. Heidelberg: Springer.Google Scholar
- Sommer, A. 1976. Attempt at an assessment of the world's tropical moist forests. Unasylva 28: 5–24.Google Scholar
- Spalding, M.D., H.E. Fox, G.R. Allen, N. Davidson, Z.A. Ferdana, M.A.X. Finlayson, B.S. Halpern, M.A. Jorge, A.L. Lombana, S.A. Lourie, K.D. Martin, E. McManus, J. Molnar, C.A. Recchia, and J. Robertson. 2007. Marine ecoregions of the world: a bioregionalization of coastal and shelf areas. BioScience 57: 573–583.CrossRefGoogle Scholar
- Winslow, F. 1887. Deterioration of American oyster-beds. The Popular Science Monthly 20: 29–43.Google Scholar
- zu Ermgassen, P.S.E., M.D. Spalding, B. Blake, L.D. Coen, B. Dumbauld, S. Geiger, J.H. Grabowski, R. Grizzle, M. Luckenbach, K.A. McGraw, B. Rodney, J.L. Ruesink, S.P. Powers, and R.D. Brumbaugh. 2012. Historical ecology with real numbers: past and present extent and biomass of an imperilled estuarine ecosystem. Proceedings of the Royal Society B 279: 3393–3400.CrossRefGoogle Scholar