INTRODUCTION

Freshwater mussels (Bivalvia: Unionoida) are a diverse, important component of freshwater ecosystems in North America and globally, and only recently has their ecological importance been well documented (Vaughn and Hakenkamp 2001; Howard and Cuffey 2006; Vaughn et al. 2008; Haag 2012; Lopes-Lima et al. 2014; Vaughn 2017). Their cultural importance in North America dates back more than 10,000 yr (reviewed in Haag 2012), including in the Pacific Northwest (Osborne 1951; Lyman 1984), where they remain culturally significant today (Brim Box et al. 2006; Norgaard et al. 2013; CTUIR 2015). Despite their ecological and cultural significance, freshwater mussels are among the most imperiled faunal groups worldwide (Bogan 1993; Williams et al. 1993; Lydeard et al. 2004).

North America has the greatest freshwater mussel diversity in the world, with more than 300 species currently recognized (Haag and Williams 2014). Much of this diversity is concentrated in the eastern (i.e., east of the Continental Divide), and specifically southeastern, USA (Graf and Cummings 2007; Haag 2012). The western freshwater mussel fauna from the Pacific region, which includes drainages flowing into the Pacific Ocean, Arctic Ocean, and the endorheic Great Basin, is composed of three genera (Anodonta, Gonidea, and Margaritifera). Gonidea angulata (Lea, 1838) is monotypic among North American freshwater mussels, being the only extant member of the genus. Both G. angulata and Margaritifera falcata (Gould, 1850) are easily identified and have well-defined distributions across western states in comparison with species comprising the genus Anodonta, for which the number and identity of species is a continuing source of confusion. Diagnostic shell characters are lacking in Anodonta. As a result, identification of specimens can be challenging, and misidentification is common, further complicating the interpretation of ranges of western Anodonta. Misidentification is also common, which further complicates the interpretation of ranges in western Anodonta.

Western species of Anodonta recognized by Turgeon et al. (1998) include Anodonta beringiana Middendorff, 1851; Anodonta dejecta Lewis, 1875; Anodonta nuttalliana I. Lea, 1838; Anodonta oregonensis I. Lea, 1838; Anodonta californiensis Lea, 1852; and Anodonta kennerlyi Lea, 1860. Recent genetic research by Chong et al. (2008; mitochondrial markers) and Mock et al. (2010; nuclear and mitochondrial markers) suggested that western Anodonta are composed of three distinct clades: A. nuttalliana/ A. californiensis, A. oregonensis/ A. kennerlyi, and A. beringiana. Furthermore, Lopes-Lima et al. (2017) advocate for reassigning A. beringiana to the genus Sinanodonta. Within the A. nuttalliana/californiensis clade, Chong et al. (2008) and Mock et al. (2010) found that shell morphology (including degree of inflation and wing prominence, characteristics historically used to identity individual species) was incongruous with genetic identity and relationships. In combination with the evident relatedness of populations and lack of interspecific differentiation, these findings indicate that there is only one species in that clade (properly named A. nuttalliana according to the rules of the ICZN Code [1999]). Because the geographic sampling was not very extensive for the oregonensis/kennerlyi clade, and because nuclear markers were not included in the study by Chong et al. (2008), the number of species within that clade remains unresolved.

The validity of an additional western Anodonta species, A. dejecta, also remains unresolved. Its validity was questioned by Bequaert and Miller (1973), although the Turgeon et al. (1998) and Graf and Cummings (2007) checklists include this species. Genetic analysis of Anodonta sampled from multiple basins in the southwest, within what has historically been considered the range (Simpson 1897, 1914), has only confirmed the presence of A. nuttalliana sensu lato (Mock et al. 2010; Culver et al. 2012, Arizona Game and Fish Department, unpublished report). Lewis' (1875) original type locality has long been considered in error, and Simpson redefined the type locality of A. dejecta on the basis of limited evidence (1897, 1914). Given the failure to confirm the presence of any Anodonta species distinct from A. nuttalliana in the region, we consider A. dejecta a nomen dubium.

Declines of North American freshwater mussels over the past century have been well documented, with 74% of species considered imperiled (FMCS 2016). However, compared with their eastern counterparts, less is known about western freshwater mussels, and detailed information on life history, conservation status, and management priorities remains incomplete. Although local or regional status assessments have been developed for western freshwater mussels in the past few decades (e.g., Bequaert and Miller 1973; Taylor 1981; Frest and Johannes 1995; COSEWIC 2003; Hovingh 2004; Howard et al. 2015), range-wide assessments based on detailed occurrence data have not been completed (but see reviews by Jepsen and LaBar 2012; Jepsen et al. 2012a, 2012b). Such occurrence data have now been compiled for western freshwater mussels (Xerces/CTUIR 2015), with the exception of Sinanodonta beringiana, for which fewer historic and recent records exist. With this new database, it has become possible to assess the extinction risk of western freshwater mussels using the categories and criteria of the International Union for the Conservation of Nature (IUCN) Red List. In this study we conducted assessments of the extinction risk for G. angulata, M. falcata, A. nuttalliana, and the A. oregonensis/ kennerlyi clade, and reviewed relevant threats and conservation considerations for western freshwater mussels.

METHODS

The IUCN Red List ( http://www.iucnredlist.org/) ranks organisms according to seven categories of extinction risk, ranging from Extinct to Least Concern (Table 1). We assessed extinction risk for the Winged Floater (A. nuttalliana), the Western Ridged Mussel (G. angulata), the Western Pearlshell (M. falcata), and the A. oregonensis/kennerlyi clade by assigning them to one of the seven categories based on the IUCN criterion A, which assesses population size reduction. Specifically, we used subcriterion A2, and assessed population size reductions for each species or clade on the basis of a decline in extent of occurrence (EOO) (IUCN 2012). Our analysis relied on occurrence data, and our estimates of population trends were informed only by the presence of individuals or populations, which in turn may be based on evidence of live animals or empty shells. This method of analysis has the potential to under- or overestimate population size trends if existing populations differ in abundance from historic populations or if abundance varies among populations. Because such information is not generally available, we also incorporated relevant research or anecdotal observations to inform and support the extinction risk assessments (IUCN 2017).

Table 1.

International Union for Conservation of Nature (IUCN) Red List categories and criteria based on subcriterion A2c: “An observed, estimated, inferred or suspected population size reduction … over the last 10 years or three generations, whichever is the longer, where the reduction or its causes may not have ceased OR may not be understood OR may not be reversible, based on… a decline in area of occupancy, extent of occurrence and/or quality of habitat” (IUCN 2012).

We used a data set composed of nearly 7,300 occurrence records (observations or collections of shells or live animals) from 10 western U.S. states, three Canadian provinces, and two Mexican states (Figs. 1, 2; Xerces/CTUIR 2015). Data sources included state and federal wildlife agencies, tribes, university and nongovernmental organization biologists, and mussel enthusiasts. Data were also sourced through museum databases, published literature, unpublished reports, and incidental observations (Xerces/CTUIR 2015). More than 850 specimens from historical museum collections were also physically inventoried, measured, or photographed between 2003 and 2015 from the Smithsonian Institution (USNM), Natural History Museum of Los Angeles County (LACM), California Academy of Sciences (CAS), the Academy of Natural Sciences of Drexel University (ANSP), the Utah Museum of Natural History (UMNH), the Carnegie Museum of Natural History (CMNH), the Field Museum (FMNH), the Museum of Comparative Zoology-Harvard University (MCZ), the North Carolina Museum of Natural Sciences (NCMNS), the Illinois Natural History Museum (INHS), and the University of Michigan Museum of Zoology (UMMZ).

Only records with sufficient locality (at least county-level accuracy) and temporal (confident assignment to either the “historic” or “recent” time period) information were included. We sought to evaluate recent search effort across each species' or clades' entire range, and to reduce the number of false negatives (i.e., a freshwater mussel is not currently detected but is present at a site where it also historically occurred). Therefore, we combined our data set with an additional ∼4.200 records from recent aquatic invertebrate surveys (targeting other faunal groups in addition to freshwater mussels) to document search effort. All records used in this analysis are depicted in Figure 3.

For the A. nuttalliana data set, we included records for A. nuttalliana, A. wahlamatensis (synonymized under A. nuttalliana by Call 1884), and A. californiensis. For the A. oregonensis/kennerlyi clade, we included records for A. oregonensis and A. kennerlyi. Given the confusion regarding identification of Anodonta species, many recent Anodonta records in our database (more than 450 in total) were only identified to genus, and in multiple instances, these were the only records for a watershed from the recent time period, providing important information regarding the recent distribution of this genus. Western Anodonta largely overlap in range, so when recent Anodonta sp. records fell within overlapping historic ranges, those records were included in each of the two Anodonta assessments. When recent records identified as Anodonta sp. fell within the historic range of only one species or clade, those records were assumed to correspond to that species or clade. Although there are several historic records of A. oregonensis from Utah, Nevada and southern California, previous studies (Mock et al. 2010; Culver et al. 2012, Arizona Game and Fish Department, unpublished report) and a re-examination of historical shells in museum collections (E. Blevins et al., 2016, unpublished data) suggest that only A. nuttalliana is known from the arid western states of Utah, Nevada, and Arizona, and from southern California

Records were divided into historic (1842–1989, but also including archeological records) and recent (1990–2015) time periods. The demarcation of historic and recent time periods was based on IUCN (2017) guidelines, which indicate that organisms should be categorized on the basis of an assessment of “the last 10 years or three generations (whichever is longer)”. Three generations would correspond to 24, 27, and 45 years for Anodonta, Margaritifera, and Gonidea respectively (Heard 1975; Dudgeon and Morton 1983; Toy 1998; COSEWIC 2010; Allard et al. 2015; CTUIR, 2016, unpublished data). However, we tried to reach a balance between the limitations of our data set and the necessity of conducting the analysis over an adequate time span. For example, if we had considered all records dating to 1970 or later as “recent,” which would correspond to ∼3 generations for G. angulata, only 30% of the records would be considered historic. The spatial distribution of these records also excludes known occurrences at range boundaries, including far-eastern Idaho and southwestern Oregon. For all western freshwater mussels, the number of records and the spatial distribution of records since 1990 provide a more complete picture of recent freshwater mussel occurrences and enable consideration of concurrent changes in mussel richness.

Figure 1.

Occurrence records for four western North American freshwater mussel species/clades.

Figure 2.

Number of records for freshwater mussels by year in the data set used for this analysis. Pre-1850s records are pooled across multiple years and include archeological evidence of mussel occurrences.

We compared historic and recent occurrences on the basis of occupancy of standard level 8 HydroBASINS (Lehner and Grill 2013) in the IUCN's Fresh Water Mapping Application tool, which creates convex hull polygons around selected watersheds. We selected basins on the basis of historic and recent occurrence records within watershed networks and assigned an occupancy status according to IUCN guidelines (2014). Watersheds were classified as Extant (occurrence record in recent time period) or Possibly Extinct (occurrence record in historic but not recent time period although recently searched). We calculated the EOO for each species or clade in each time period and determined percent change in area. To better depict the historical ranges of species, we also mapped watersheds that have historical records but have not been revisited as Presence Uncertain. These records were not otherwise included in our analysis based on IUCN guidelines (2014).

We also calculated a second measure: percent change in watershed area for each species or clade in each time period. This approach was based on a revised definition of EOO that incorporates hydrologic boundaries more relevant to aquatic organisms, accounting for the spatial distribution of aquatic organisms through networks of catchments (watersheds; Simaika and Samways 2010). The same measure of watershed decline was calculated using a combined data set of all records to assess general changes in freshwater mussel richness across the West.

RESULTS

The historic range of western mussels as a whole (watersheds having at least one species or clade) totaled 708 watersheds, whereas only 580 watersheds were found to be recently occupied, equaling an 18% decrease. Additionally, mussel richness has declined by 35% (Figs. 4, 5). When watersheds with higher past mussel richness (containing three or four species or clades) were considered independently, 48% of these historic “hot spots” have declined in richness in the recent time period.

Anodonta nuttalliana has declined in both EOO and watershed area (9% and 33% respectively; Table 2; Fig. 6) across Arizona, Southern California, western Nevada, and elsewhere (Blevins et al. 2016a). According to the IUCN subcriterion A2c for extinction risk (Table 1), the decline in watershed area qualifies A. nuttalliana for Vulnerable status. This status is also supported by recent research and observations (see Discussion). In contrast, although mussels of the A. oregonensis/kennerlyi clade have declined in both EOO and watershed area (9% and 26% respectively; Table 2; Fig. 7; Blevins et al. 2016b), they are still present in watersheds across the historic range, from Northern California to Alaska and east to Idaho. According to the IUCN subcriterion A2c for extinction risk (Table 1), mussels of this clade qualify as Least Concern.

In comparison, G. angulata has declined in both EOO and watershed area (28% and 43% respectively; Table 2; Fig. 8; Blevins et al. 2016c). According to the IUCN subcriterion A2c for extinction risk (Table 1), G. angulata qualifies as Vulnerable on the basis of decline in watershed area, a conclusion also supported by recent research and observations (see Discussion).

Figure 3.

Extent of recent (1990–2015) “search effort” in western states.

Figure 4.

Historic (pre-1990) western freshwater mussel presence and richness by level 8 HydroBASIN.

Figure 5.

Change in western freshwater mussel richness by level 8 HydroBASIN between historic (pre-1990) and recent (1990–2015) time periods.

Table 2.

Extinction risk assessment results for four western North American freshwater mussels.

Margaritifera falcata has declined in watershed area by 17% but just 1% in EOO (Table 2; Fig. 9; Blevins et al. 2016d). According to the IUCN subcriterion A2c for extinction risk (Table 1), the species does not qualify for Vulnerable on the basis of quantitative criteria. However, because declines in occupancy are thought to underestimate declines in abundance of this species, and because population extirpations have been reported since 1990 (see Discussion), this species meets qualitative criteria for extinction risk equaling Near Threatened according to the IUCN Red List criteria (IUCN 2012).

DISCUSSION

Extinction Risk

We applied IUCN categories and criteria to assess extinction risk in four freshwater mussel species or clades on the basis of multiple lines of evidence, including changes in historic and recent spatial EOO, changes in watershed area occupied, research by others, and anecdotal observations across western North America. We found that although these species or clades remain relatively widespread across the West as measured by EOO (ranging from 855,618 to 2,643,316 km²), range as measured by watershed area is considerably smaller (ranging from 103,096 to 409,966 km²). Additionally, freshwater mussel distribution maps also depict some level of range thinning (sensu Strayer 2008). Western mussels are found in multiple types of western freshwater ecoregions, including coastal, glaciated, unglaciated, and endorheic. Given the diverse hydrology and history of western watersheds, populations in specific watershed networks may be affected by threats independently of those at the range edges. For example, G. angulata has not recently been reported from watersheds in several Oregon basins in the interior of its range, though the species has been documented from watersheds at the edge of its range, like the Okanagan Basin in British Columbia. Freshwater mussel richness across watersheds has also declined by 35%, and 48% of watersheds that historically had higher mussel richness (three or four species) have since lost one or more species or clades. These declines were evident despite having twice as many recent observations as historic (Figure 2).

Our analysis found that A. nuttalliana has declined in occurrence by as much as 33%. Historically the species occurred from Southern California north to British Columbia and east to Wyoming, but recent surveys of historic sites by Howard et al. (2015) indicated that Southern California populations are extirpated (though the species was found as far south as the Bishop Creek Canal in Inyo County, California). Observations in Arizona in the 1990s and again in the 2000s indicate that the species is probably now extant only in the Black River drainage, where populations continue to decline (Myers 2009). Thus, “recent” occupancy as measured by this analysis may overestimate the species' current distribution, with some records now more than 25 yr old. Recent surveys in western states have also indicated that, even where the species has not been extirpated from a watershed, both the number and size of populations have declined (California: Howard et al. 2015; Wyoming: Mathias and Edwards 2014; Arizona: T. Myers, unpublished data, 2008; Myers 2009; Oregon and Washington: reviewed in Jepsen et al. 2012a; Mexico: T. Myers, unpublished data, 2008). For example, research by Brim Box et al. (2006) documented sites occupied by Anodonta in the Middle Fork John Day River of Oregon. In 2015, only 7 of 10 sites previously inhabited were still occupied. Among occupied sites, fewer mussels were observed overall (Maine et al. 2017). Recent research has also suggested that some populations may be at greater risk of local extinctions on the basis of low genetic diversity and isolation (Mock et al. 2004, 2010). Genetic structuring was also evident among populations spanning major drainage basins of the West and are considered evolutionarily significant units, many of which are also distinct management units (sensu Moritz 1994; Mock et al. 2010).

Figure 6.

Anodonta nuttalliana status by level 8 HydroBASIN. Basins were used to calculate changes in extent of occurrence and watershed area between historic (pre-1990) and recent (1990–2015) time periods.

Figure 7.

Anodanta oregonensis/kennerlyi clade status by level 8 HydroBASIN. Basins were used to calculate changes in extent of occurrence and watershed area between historic (pre-1990) and recent (1990–2015) time periods.

Figure 8.

Gonidea angulata status by level 8 HydroBASIN. Basins were used to calculate changes in extent of occurrence and watershed area between historic (pre-1990) and recent (1990–2015) time periods.

Figure 9.

Margaritifera falcata status by level 8 HydroBASIN. Basins were used to calculate changes in extent of occurrence and watershed area between historic (pre-1990) and recent (1990–2015) time periods.

Decline in occurrence by watershed was only marginally less for members of the A. oregonensis/kennerlyi clade. However, the more dramatic declines reported for A. nuttalliana have not been observed in this group, and a decline of 26% only corresponds to an IUCN ranking of Least Concern. Still, taxonomic and identification issues in Anodonta species complicate the analysis of extinction risk.

Gonidea angulata has declined in occurrence by as much as 43%, and though the species historically occurred from Southern California north to Canada and east to Nevada and Idaho, populations were reported as extirpated from Southern California and much of the Central Valley by Taylor (1981) and Coney (1993). Recent surveys have not located the species in any historic Southern California sites and few California sites in general (Howard 2008; Howard 2010; Howard et al. 2015), although the species does still occur in large beds in some Northern California sites (Howard 2010; Davis et al. 2013). Declines in Oregon, Washington, and Idaho have also been reported (Brim Box et al. 2006; Frest and Johannes 1995; reviewed in Jepsen and LaBar, 2012). A study by Brim Box et al. (2006) documented sites occupied by G. angulata in the Middle Fork John Day River of Oregon (as with Anodonta-, see above). Several of these sites were revisited in 2015, by which time one of the eight sites was extirpated and observed abundance of mussels in occupied sites had decreased (Maine et al. 2017). The species has been reported in the Humboldt Basin of Nevada since 1990, but its status should be evaluated given that more recent surveys did not identify any extant populations (A. Smith, unpublished data, 2009). COSEWIC (2010) ranked the species as endangered in Canada, citing observations of declines, limited distribution, and historic habitat alteration, as well as concerns regarding the likelihood of future introduction of zebra mussels (COSEWIC 2010; BCCDC 2015).

In comparison, M. falcata has declined in occurrence by as much as 17%, but populations in some parts of the range are considered stable (British Columbia: NatureServe 2015; Wyoming: Mathias and Edwards 2014) or are not well understood (Alaska and Nevada: Smith et al. 2005; Jepsen et al. 2012b). However, recent continuing declines have been observed in Montana, where less than a quarter of surveyed populations have been classified as viable, and another quarter of nonviable populations surveyed in 2010 were extirpated just 4 yr later (Stagliano 2015). Maine et al. (2017) similarly found that 2 of 13 previously surveyed occupied sites in the Middle Fork John Day River (Brim Box et al. 2006) were extirpated just 9 yr later. Though the species still occurs from California to Alaska and east to Montana and Wyoming, surveys in other states also reported recent extirpations, declining populations, and populations that appeared to lack recruitment (Utah: Hovingh 2004; Richards 2015; California: Furnish 2010; Southern California Edison Company 2010, unpublished report; Howard et al. 2015; May and Pryor 2016; Idaho: Lysne and Krouse 2011; Oregon: Brim Box et al. 2006; Nevada: Hovingh 2004; Washington: Hastie and Toy 2008; Wyoming and other states: reviewed in Jepsen et al. 2012b).

In this analysis, decline in M. falcata is underestimated where population abundance has decreased but the population is still extant, as with the Truckee River in California (∼20,000 individuals in a 0.8-km stretch in 1941 down to ∼120 individuals in a 2-km stretch in 2006: Murphy 1942; Howard 2008; Howard et al. 2015) and Battle Creek in Washington (1,372 individuals in 17 m² in 1995 down to 334 individuals in 25 m² in 2006: Hastie and Toy 2008). Population genetic research has also revealed “extreme inbreeding” in multiple populations, which may result from hermaphroditism and selfing (Mock et al. 2013) and could reduce fitness in already fragmented populations (Keyghobadi 2007).

Because our data set was composed of occurrence records, we were not able to more generally quantify trends in population abundance. However, at sites where abundance has been assessed over time for western mussels, a decreasing trend has typically been reported (Hastie and Toy 2008; Howard 2008; Jepsen and LaBar 2012; Jepsen et al. 2012a, 2012b; Stagliano 2015; Maine et al. 2017). The loss of equilibrium species (i.e., those typically long lived and reaching sexual maturity at older ages, such as G. angulata and M. falcata) may go unnoticed after habitat alteration or destruction. In eastern North America, equilibrium species persisted in reservoirs for as long as 40 yr before disappearing (Haag 2012). Additionally, our study was restricted to declines between historic and recent time periods and was unable to quantitatively incorporate more recent extirpations (i.e., if a watershed was occupied in 1995 but populations were extirpated by 2014, the watershed would still be classified as “Extant”), yet our analysis demonstrated that multiple western species still qualified as Near Threatened or Vulnerable. It is therefore important to note that these estimates of decline may underestimate true species declines and extinction risk.