Quantifying the historic contribution of Olympia oysters to filtration in Pacific Coast (USA) estuaries and the implications for restoration objectives

Oysters are filter-feeding bivalves that historically formed reefs or beds in estuaries throughout temperate latitudes (Stenzel 1971). Both oysters and the complex physical habitats they build may provide a number of important ecosystem services, including the direct provision of food, filtration of the water column (Grizzle et al. 2008), enhanced denitrification rates (Newell et al. 2002), the enhancement of non-oyster fish stocks (Peterson et al. 2003), and coastal protection (Scyphers et al. 2011). Most detailed studies on these services have focused on Crassostrea virginica (Gmelin, 1791) reefs on the Atlantic and Gulf coasts of the United States (see Coen et al. 2007); yet, similar benefits are attributed to other oyster species, such as the Olympia oyster, Ostrea lurida Carpenter 1864 on the Pacific coast (White et al. 2009). These analogies are drawn despite numerous studies highlighting differences between C. virginica and O. lurida as regards their morphology (Kellogg 1915), physiology (Elsey 1935), environmental tolerance (Stenzel 1971), and the structure of formed habitats (Stafford 1915), all of which may affect the delivery of those ecosystem services. As ecosystem service delivery becomes an increasingly important concept in the management of marine resources, assumptions of functional equivalency should be evaluated and additional data collected as necessary.

O. lurida was formerly abundant on the Pacific coast of the United States (Bancroft 1890), where it was harvested at low intensity for thousands of years (Barrett 1963). Populations of O. lurida collapsed, however, soon after commercialization of the fishery and rising demand in the mid 1800s (Kirby 2004). Initially, the decline in landings was met with concern and the propagation of this species was the focus of research and numerous investigations (see Baker 1995 for a full bibliography), but interest in the native oyster waned soon after the successful propagation of the faster growing introduced Pacific oyster, Crassostrea gigas (Thunberg, 1793). While O. lurida persists in a number of Pacific coast estuaries, it is generally a rare component of the biota, represented by scattered individuals (Polson and Zacherl 2009), such that the habitat is widely believed to be functionally extinct in estuaries south of Puget Sound, Washington, USA (Beck et al. 2011; zu Ermgassen et al. 2012). The species’ current range stretches from northern British Columbia, Canada, to Baja California, Mexico in the South (Polson and Zacherl 2009).

Since 2000, there have been renewed efforts to restore O. lurida on the Pacific coast (Cook et al. 2000), involving a broad range of stakeholders including government agencies, industry, tribes, and conservation organizations (Peter-Contesse and Peabody 2005). The motivations to restore are also diverse, including brood stock rehabilitation and the return of various ecosystem services such as filtration, enhanced nutrient cycling, and provision of habitat for other species (Peabody and Griffin 2008; White et al. 2009). With the expansion and increasing success of restoration efforts (McGraw 2009), there is a growing need to calculate the ecosystem service benefits of O. lurida, such that restoration planning and advocacy can take place with greater transparency and a stronger knowledge base.

Filtration of the water column by filter-feeding organisms such as oysters is an important ecological function in aquatic systems (Gili and Coma 1998; Prins et al. 1998). The potential to enhance this ecosystem service is one of the key motivations to restore oysters on the Atlantic coast (e.g., Rossi-Snook et al. 2010). Filtration results in the deposition of organic and inorganic particles to the benthos, which may lead to enhanced denitrification rates in the sediments (Newell et al. 2002), and the enhanced richness and abundance of benthic macrofauna and microbes (Norkko et al. 2001; Nocker et al. 2004). Furthermore, sediment nutrient enrichment (Booth and Heck 2009) and the increased irradiance resulting from the removal of seston may promote the growth of submerged aquatic vegetation (Newell and Koch 2004), which is a critical nursery habitat for a wide range of marine species (Heck et al. 2003). As such, filtration can be considered a critical ecosystem service.

Filtration rates (FR) can vary widely among bivalve species (Moehlenberg and Riisgaard 1979; Riisgaard 1988) and have not previously been determined for O. lurida. Here, we present the results of the first laboratory-based measurements of FR by O. lurida and propose a model of FR as it varies with oyster size and temperature. FR can be defined as the volume of water completely cleared of particles per unit time (Newell and Langdon 1996; Dame 2011), which we adopt here. We apply our FR model to historic abundance (ca. 1850–1935) and a higher estimate of potential abundance before commercial exploitation in five Pacific coast estuaries. From these, we estimate the historic contribution of O. lurida to estuary filtration relative to estuarine residence times. This measure can be considered to be an indicator of the impact of filtration by oysters on the estuary more broadly (Dame 2011). We consider the implications of our results for habitat restoration and management where filtration of estuarine waters is a restoration objective.

The FR of 3.08 l g⁻¹ h⁻¹ calculated for O. lurida at optimum temperature falls close to the mean of 3.47 ± 0.49 l g⁻¹ h⁻¹ (± 2 s.e.) for oysters with a DTW of 1 g suggested by Cranford et al. (2011); however, it is substantially lower than the rates of 5 l g⁻¹ h⁻¹ (Newell 1988) and 6.79 l g⁻¹ h⁻¹ (Riisgaard 1988) frequently cited for C. virginica, and 4.83 l g⁻¹ h⁻¹ cited for C. gigas (Bougrier et al. 1995). While there has been significant interest in developing a generalized model for filtration by bivalve filter feeders (e.g., Powell et al. 1992; Cranford et al. 2011), which is useful in the absence of species-specific data, it is also clear that differences in FR between species may be substantial (Moehlenberg and Riisgaard 1979; Riisgaard 1988; Powell et al. 1992). Early work comparing the gill morphology of Crassostrea species with Ostrea species highlighted the likelihood of differences in their feeding ecology (Elsey 1935). This hypothesis was substantiated in a number of studies comparing C. gigas with O. edulis, which found that C. gigas had filtration rates approximately two times higher per unit DTW (Walne 1972; Mathers 1974; Rodhouse and O’Kelly 1981). Although this is not necessarily the case when comparing O. edulis with C. virginica (Shumway et al. 1985), our results support the findings that Ostrea species have lower FR compared to those reported for Crassostrea species at optimum temperature (Fig. 5). Similarly, while the b value determined from our experiment is lower than the mean suggested by Cranford et al. (2011), it is not dissimilar to some b values included in their meta-analysis; for example, Kesarcodi-Watson et al. (2001) reported that Saccostrea commercialis had a b value of 0.32. While it is unsurprising that at the low temperatures typical of the Pacific coast, C. virginica has the lowest FR of the three species examined (Fig. 5) and it does help to illustrate the importance of inter-species differences in biology in determining their impact on the ecosystem. These differences include FR, growth rates, mean size, mean densities, and temperature tolerance. We have attempted in this study to tease apart some of these factors.

Open image in new window

Ostrea species generally grow more slowly and have a smaller maximum size than Crassostrea species (Stafford 1913; Walne 1972). Furthermore, while very little is known about the density, patch size, and structure of O. lurida beds pre-commercial exploitation, descriptions of intact habitat in British Columbia suggest that O. lurida typically did not reach as high densities within oyster beds as those formed by C. virginica on the Atlantic coast (Stafford 1915). The lower densities and smaller size implicit in these descriptions suggest that O. lurida beds have a lower rate of filtration per unit area than C. virginica reefs.

Whereas oyster populations on the Atlantic and Gulf coasts were often capable of achieving full estuary filtration historically (zu Ermgassen et al. 2013), our estimates for the Pacific coast suggest that O. lurida populations were capable of filtering only a fraction of the volume within the residence time (Table 4). For example, the filtration capacity of the historic O. lurida population in Willapa Bay has been previously conservatively estimated, using oyster abundance extrapolated from harvested biomass statistics, to filter 0.8 % of the estuary within its residence time (Ruesink et al. 2006). We estimate an order of magnitude greater filtration in this study, with the historic population estimated to be capable of achieving ~12 % of full estuary filtration, and the estimated pre-commercial exploitation population ~36 % of full estuary filtration. Full estuary filtration does not equate to the whole volume of the estuary being cleared of all particles due to the high probability of re-filtration within dense beds, re-suspension of particles, selective feeding, and the patchy distribution of oysters (Cranford et al. 2011); however, full estuary filtration values can be used to assess the potential role of bivalves in regulating phytoplankton abundance (Dame 2011).

The lack of evidence of large-scale regulation of seston does not exclude the possibility that filtration may still be an important process in some areas within an estuary. Banas and Hickey (2005) show that the residence time in Willapa Bay is highly uneven throughout the estuary, with water in the upper third of the estuary being retained for 3–5 weeks, while there is near full exchange with every tide near the mouth of the estuary. Oysters are also known to recruit best to the regions in the bay with the greatest retention times (Banas et al. 2007); therefore, it is possible that filtration by O. lurida could have been a structuring process in the upper portion of the bay historically.

The differences in the proportion of the estuary volumes filtered by oysters between the Pacific and Atlantic/Gulf of Mexico coasts cannot simply be explained by the differences between the FR of O. lurida and Crassostrea species, as is illustrated in Fig. 4. The physical attributes of the estuaries, in particular the low residence time of water in Pacific coast estuaries and the low water temperatures (Fig. 5), also play an important role in explaining the difference in the ecological role of native oysters between US coasts. Atlantic coast estuaries are typically coastal plain estuaries, with high nutrient and sediment loads, and relatively long residence times (Uncles et al. 2002). By contrast, Pacific coast estuaries are typically drowned river valleys, with small, high elevation catchments resulting in relatively low nutrient inputs, low sediment retention, and short residence times (Inman and Nordstrom 1971; Bricker et al. 2007). Previous work has illustrated that full estuary filtration values may be appropriate indicators of large-scale impacts on seston within estuaries with long residence times which are dominated by autochthonous primary productivity, but less appropriate in estuaries with short residence times (Dame and Prins 1998; Dame 2011). While this is the case for many Atlantic coast estuaries, the potential impact of full estuary filtration is less well understood for Pacific coast systems, where autochthonous primary productivity is less important and oceanic imports of phytoplankton may dominate (Banas et al. 2007).

The past century has seen the introduction and establishment of C. gigas on the Pacific coast. C. gigas now has a significant presence in some estuaries both in aquaculture facilities and in newly established wild populations (Dumbauld et al. 2011). Given this increased abundance, it is likely that the filtration capacity provided by oysters, albeit a different species, may be returning toward historic values in some estuaries. For example, if we consider the estimated population level filtration rate for cultured C. gigas in Willapa Bay from Banas et al. (2007), we find that ~ 15 % of the estuary could be filtered within the estuary residence time, which lies between our estimated historic and pristine values for the Olympia oyster (Table 4).

While O. lurida restoration is unlikely to lead to large-scale regulation of seston at whole-estuary scales, restoring oyster beds may nevertheless result in significant local impacts on sea grasses. At a local scale, studies using both C. gigas and C. virginica have illustrated that oysters may facilitate sea grass via improved sediment stability (Smith et al. 2009), water clarity (Wall et al. 2008), or nutrient availability (Booth and Heck 2009). The potential positive impact of oyster presence on sea grass appears, however, to be strongly mediated by oyster density, with the impact tending to become negative as oyster or shell density increases (Booth and Heck 2009, Wagner et al. 2012), as well as near reefs (Kelly and Volpe 2007). In contrast to the reef habitat commonly formed by C. gigas and C. virginica, O. lurida beds were historically described as agglomerations of loosely associated individuals (Stafford 1915) that may result in a mixed habitat structure. Indeed, O. lurida beds in British Columbia were described historically as mixed beds of oysters and sea grass, and possible facilitation between the two habitat building species was suggested (Stafford 1915). Therefore, here, as with other ecosystem services, the potential implications of differences between the species should be taken into account.

Where it may not be possible or appropriate to set habitat restoration goals on the basis of the historic status due to, for example, irreversible changes in the abiotic environment (Hobbs et al. 2009), it may be appropriate to set goals on the basis of potential ecosystem services (Hughes et al. 2011). In order to do so, it is necessary to have knowledge of the services potentially provided by the restored habitat. This study is a first step toward quantifying the services that may be provided by O. lurida beds. Further research is needed to determine to what degree our laboratory filtration rates are representative of in situ rates, including how FR is impacted by seston concentrations, as well as to determine the contribution of O. lurida beds to other services of interest, such as denitrification and the provision of fish habitat. While structured habitat is widely accepted to be beneficial to fish, the degree of complexity of the habitat may be important in determining the value of those benefits (Soniat et al. 2004). Historic O. lurida beds had less vertical relief and a lower habitat complexity than C. virginica reefs (Stafford 1915) and are therefore likely to differ in their function as fish habitat. That said, O. lurida is commonly likened to the European oyster O. edulis, due to its similar life history, physiology, and habitat (Elsey 1935; Sherwood 1931), and O. edulis has been observed to have significant habitat value for species in the Wadden Sea (Reise 1982).

In summary, while previous work has documented the decline of O. lurida on the Pacific coast (Steele 1957; Kirby 2004; zu Ermgassen et al. 2012), this study is the first to estimate historic levels of filtration across numerous Pacific coast estuaries, using species-specific FR. We show that filtration by O. lurida may not have played an important role in regulating phytoplankton in estuaries on a large scale, in contrast to C. virginica on the Atlantic and Gulf coasts. However, there is still potential for significant local impacts of filtration by this species. Although parallels are often drawn between oyster habitats in different locations, there is a growing body of evidence that the structures and functions may vary. As efforts are undertaken to restore species and their associated functions in a large number of US estuaries, it is increasingly important that restoration objectives should reflect differences among species and that research is undertaken to better define potential ecosystem services that may be obtained through restoration.