Freshwater mussels (Unionidae) are among the most imperiled groups of organisms in the world, and the lack of information regarding species distributions, life-history characteristics, and ecological and biological requirements may limit the protection of remaining mussel populations. We examined the influence of hydrologic factors on the occurrence of the Suwannee Moccasinshell Medionidus walkeri, a federally threatened freshwater mussel species, endemic to the Suwannee River Basin in Georgia and Florida. We also evaluated the influence of survey effort on detection of Suwannee Moccasinshell during field surveys. We compiled all recent (2013–2016) mussel survey records in the Suwannee River Basin. We calculated cumulative discharge contributed by upstream springs for each of 220 survey locations. We combined the spring discharge predictor variable with Suwannee Moccasinshell detection and nondetection data from each survey location to develop a suite of occupancy models. Modeling results indicated that detection of Suwannee Moccasinshell during surveys was strongly and positively related to survey effort. Modeling results also indicated that sites with cumulative spring discharge inputs exceeding ∼28 cubic meters per second were most likely (i.e., predicted occupancy probabilities >0.5) to support Suwannee Moccasinshell populations. However, occupancy declined in the lowermost reaches of the Suwannee mainstem despite high spring discharge inputs, presumably due to greater tidal influences and differences in physicochemical habitat conditions. Historical localities where Suwannee Moccasinshell has presumably been extirpated are all devoid of springs in their upstream watersheds. We hypothesize that springs may buffer extremely tannic, and at times polluted, surface waters, in addition to maintaining adequate flows during periods of drought, thereby promoting the persistence of Suwannee Moccasinshell populations. Our study suggests that springs are a critical resource for Suwannee Moccasinshell and may be more important for conservation planning than was previously recognized.

The Suwannee River Basin is an unusual system, with dendritic (Withlacoochee and Alapaha rivers) and poorly drained and swampy (Okefenokee Swamp) headwaters in Georgia. It eventually transitions to a karst system downstream of Cody Scarp in north-central Florida, where numerous mainstem springs replace discharge contributions of first- to third-order tributaries typically found in most river systems (Figure 1; Williams et al. 2014). Because of this distinctive geology, up to 90% of the average median flow of the lower Suwannee River is derived entirely from spring and groundwater sources (Grubbs and Crandall 2007), making springs an important component of the hydrology, ecology, and biology of the Suwannee River Basin. Springs serve as thermal refugia (e.g., manatees Family Trichechidae; Laist et al. 2013), buffer and clarify the extremely tannic and black Okefenokee Swamp water (Hornsby et al. 1998; Williams et al. 2014), are economically important, generating as much as US$45 million in recreation annually (Wynn et al. 2014), and contribute to the abundant aquatic habitats necessary for many taxa, including the Ichetucknee Siltsnail Floridobia mica, which occupies a single spring in the Ichetucknee River (Warren and Bernatis 2015). Taken together, these properties illustrate the importance of both individual springs and the overall contributions of springs to the Suwannee River system. Despite their immense economic and ecological value, the quantity and quality of water derived from springs are greatly affected by groundwater withdrawals (e.g., municipal use, agricultural irrigation, etc.; Torak et al. 1996; Grubbs and Crandall 2007; Shea et al. 2013) and eutrophication from excessive agricultural fertilization and runoff, delivering nitrates and ammonia to groundwater (Andrews 1994; Pittman et al. 1997). During drought, evapotranspiration rates and groundwater pumping can exceed aquifer recharge rates, drawing down the aquifer and reducing spring and stream flows (Grubbs and Crandall 2007). These effects are more pronounced in the smaller streams in the headwaters of the Suwannee River Basin where recharge rates are lowest as a result of confinement of the aquifer by less permeable sediments (Grubbs 1998), potentially resulting in the drying of streams receiving little to no spring discharge. Freshwater mussel communities in smaller streams are more likely to be affected by droughts due to these scenarios (Shea et al. 2013).

Figure 1.

Locations of freshwater mussel surveys and other physical attributes in the Suwannee River Basin in Florida and Georgia. Positive (black circles) and negative (gray circles) Suwannee Moccasinshell Medionidus walkeri survey locations during 2013–2016, historical (before 1996) Suwannee Moccasinshell localities (white circles), springs (yellow circles), area of upstream most tidal influence (dashed rectangle), and the Cody Scarp (green line) are shown. Numbers indicate subbasins: 1—Upper Withlacoochee River, 2—Lower Withlacoochee River, 3—Alapaha River, 4—Upper Suwannee River, 5—Middle Suwannee River, 6—Lower Suwannee River, 7—Upper Santa Fe River, 8—Lower Santa Fe River.

Figure 1.

Locations of freshwater mussel surveys and other physical attributes in the Suwannee River Basin in Florida and Georgia. Positive (black circles) and negative (gray circles) Suwannee Moccasinshell Medionidus walkeri survey locations during 2013–2016, historical (before 1996) Suwannee Moccasinshell localities (white circles), springs (yellow circles), area of upstream most tidal influence (dashed rectangle), and the Cody Scarp (green line) are shown. Numbers indicate subbasins: 1—Upper Withlacoochee River, 2—Lower Withlacoochee River, 3—Alapaha River, 4—Upper Suwannee River, 5—Middle Suwannee River, 6—Lower Suwannee River, 7—Upper Santa Fe River, 8—Lower Santa Fe River.

Close modal

Freshwater mussels (Family Unionidae) are among the most imperiled organisms globally (Haag and Williams 2014), and >25% of mussel species in Florida are listed as federally threatened or endangered pursuant to the U.S. Endangered Species Act (ESA 1973, as amended; Williams et al. 2014). One of those, Suwannee Moccasinshell Medionidus walkeri (Figure 2), is listed as federally threatened under the U.S. ESA and was rediscovered in 2012 after a 16-y gap in detection (Johnson et al. 2016). Suwannee Moccasinshell is endemic to the Suwannee River Basin and historically occupied both large river and small creek habitats in the middle and lower Suwannee River and two of its major tributaries, the Withlacoochee and Santa Fe rivers in south-central Georgia and north-central Florida (Figure 1; Johnson et al. 2016). Recent survey efforts have failed to detect the species at many formerly occupied sites around the periphery of the species' range in headwater systems that are more susceptible to drought and degraded water quality (Johnson et al. 2016). Although a recent study characterized many aspects regarding life-history requirements and distribution, including period of gravidity, age to reproductive maturity, fecundity, and host fish requirements (Johnson et al. 2016), information regarding ecological and biological requirements are lacking and may limit the protection of remaining Suwannee Moccasinshell populations.

Figure 2.

Photograph of Suwannee Moccansinshell Medionidus walkeri collected during mussel surveys in the Suwannee River Basin in Florida and Georgia from 2013 to 2016.

Figure 2.

Photograph of Suwannee Moccansinshell Medionidus walkeri collected during mussel surveys in the Suwannee River Basin in Florida and Georgia from 2013 to 2016.

Close modal

Previous efforts to monitor the status of Suwannee Moccasinshell throughout the Suwannee River Basin, though extensive, have generally lacked consideration for potential effects of incomplete detection on our understanding of the species' distribution. Like most freshwater mussel species, Suwannee Moccasinshell can be difficult to detect during sampling given their benthic nature and patchy distribution (Strayer and Smith 2003), which may be compounded by their relatively low abundance compared with other co-occurring mussel species. Thus, resource managers charged with conserving Suwannee Moccasinshell populations may consider that species nondetection does not necessarily mean species absence, and that inferences based on presence-only data may be susceptible to biases induced by incomplete detection of species during sampling (Yoccoz et al. 2001). Currently, managers lack a comprehensive monitoring approach for assessing the contemporary distribution of Suwannee Moccasinshell that also accounts for incomplete survey detection and provides a means for determining how broad-scale hydrologic factors within the Suwannee River Basin influence occupancy and detection.

Personnel charged with the management and conservation of the Suwannee Moccasinshell could benefit from a monitoring and modeling framework that efficiently and effectively enables them to assess the status of the species throughout its range. Toward this end, the primary objective of this research was twofold: 1) to determine how spring contributions to stream discharge influence Suwannee Moccasinshell occurrence throughout the Suwannee River Basin; and 2) to estimate detection probabilities for Suwannee Moccasinshell and quantify the relationship between survey effort and detection. Accomplishing these objectives will improve understanding of factors influencing the distribution of Suwannee Moccasinshell, provide managers with a basis for determining the amount of survey effort required to reliably detect extant populations, and provide a standardized survey methodology that is useful for long-term monitoring.

Site selection

Surveys were primarily conducted throughout the Suwannee River Basin in Florida, but several historical localities monitored for Suwannee Moccasinshell in Georgia were also surveyed (Figure 1). For site selection, our goal was to achieve an approximately even distribution of sites along the mainstem Suwannee River (1 km between sites) and its major tributaries (Alapaha, Withlacoochee, and Santa Fe rivers; 500–750 m between sites). During sampling, once the desired distance from a previously sampled site was covered, we selected a new site based on whether the site could be effectively surveyed with snorkeling gear. For example, we generally excluded sites with steep banks that immediately dropped off into depths >2.5 m. We excluded these sites primarily for safety considerations, because Suwannee River mussels are known to occupy deeper water habitats. Despite focusing on sites that could be sampled using snorkel gear, we generally found suitable survey locations at the desired 1-km and 500–750-m intervals. Additionally, we used Self-Contained Underwater Breathing Apparatus at a small number of sites (15 of 220) to search a subset of deep water habitats. We note that the Suwannee River mussel fauna does not consist of species such as Threeridge Amblema plicata or Washboard Megalonaias nervosa that are commonly associated with deep, large river habitat. As such, we do not believe that excluding exceptionally deep habitats from our study was problematic with respect to biasing our understanding of Suwannee Moccasinshell occurrence.

Mussel surveys

At each site, two to eight surveyors sampled all available habitats to a depth of ∼2.5 m using snorkeling and tactile methods along 30–50 m of shoreline. During sampling, individual surveyors started ≥10 m apart at the downstream-most boundary of the site. Surveyors then sampled while moving upstream, keeping track of other surveyors to ensure that each surveyor sampled a unique area. Each surveyor attempted to cover all available habitats encountered within the site, including backwaters, riffles, runs, and pools, to ensure that the area surveyed by each person was representative of the site with respect to the composition of physical habitat. Although total survey effort was generally targeted for 1 person-hour at each site, total effort varied among locations because of differences in habitat complexity, the number of mussels collected (either extremely high or low), and survey methods (e.g., Self-Contained Underwater Breathing Apparatus vs. snorkel; Data S1, Supplemental Material). The respective catches of each surveyor (number of individuals, by species) were bagged and recorded separately (Data S1, Supplemental Material). The same, highly experienced 3-person crew conducted most surveys in this study, and all additional surveyors had ≥5 years of experience at searching for and identifying freshwater mussels. The extensive experience of field crews combined with the distinctive morphological characteristics of Suwannee Moccasinshell (Figure 2) led us to conclude that the risk of species misidentification in this study was low (Shea et al. 2011).

Cumulative spring discharge

We obtained springs data from the Suwannee River Water Management District (http://www.srwmd.state.fl.us/index.aspx?NID=319) and categorized them by magnitude (i.e., first to fourth). We delineated watersheds for each survey location (i.e., contributing watershed area upstream from each survey site) from a one-third arc-second (∼10 m) National Elevation Dataset digital elevation model (data available from the U.S. Geological Survey) using the AGREE method in ArcHydro tools in ArcMap 10.3 (Environmental Systems Research Institute, Redlands, CA 2014;). We then extracted springs data by watershed and obtained the total spring discharge contributing to each survey location by summing the minimum discharge, measured as cubic meters per second (cms), across all springs of a given magnitude within the watershed (Data S1, Supplemental Material) following an adaptation of Meinzer's magnitude of discharge system that only considers first to fourth magnitude (i.e., first magnitude [≥2.83 cms]; second magnitude [0.283–2.82 cms]; third magnitude [0.0283–0.282 cms]; fourth magnitude [0.00283–0.0282 cms]; Rosenau et al. 1977). For example, a site with 6 magnitude one, 7 magnitude two, 8 magnitude three, and 20 magnitude four springs would have a cumulative minimum spring discharge (hereafter, SpringQ) of 19.24 cms. For each study site, we opted for the most conservative estimate of SpringQ possible by using the minimum discharge associated with a spring of a given magnitude because reliable discharge data are not available for most springs. Thus, our estimation of SpringQ is smaller than that observed at our study sites. For example, Manatee Spring is a first-magnitude spring with a mean annual discharge of 5.35 cms (Suwannee River Water Management District 2005) but is represented in our data set by the minimum value for a first-magnitude spring of 2.83 cms. In addition to SpringQ, we also calculated a quadratic term, SpringQ2, because we suspected a priori that Suwannee Moccasinshell occupancy may decline in the lower Suwannee River mainstem due to intensified tidal influences, changes in physical habitat, or both. However, the lowermost reaches of the Suwannee River mainstem also have the highest SpringQ discharge values of anywhere in the Suwannee River Basin, which meant that a covariate associated with tidal influence (i.e., a binary variable indicating tidal influence) was highly and positively correlated with SpringQ, which precluded its use as a covariate in the occupancy model. Hence, we opted for using SpringQ and SpringQ2 (Data S1, Supplemental Material) as covariates in the occupancy model to assess the relative support for our underlying hypotheses (see Occupancy modeling, below).

Occupancy modeling

We developed a single-season occupancy model (MacKenzie et al. 2002) to assess the influence of SpringQ on Suwannee Moccasinshell occupancy as well as SpringQ and survey effort on Suwannee Moccasinshell detection. We note that although the 220 sites were surveyed from 2013 to 2016, each site was surveyed only once during this time period (i.e., during only one of those years). Given the presumably slow dynamics of Suwannee Moccasinshell populations and that colonization and extinction events were unlikely during the 3-y study period, we deemed it reasonable to combine collections across all locations over the entire time period and model the system using a single-season occupancy model.

Conventional occupancy studies are usually conducted by repeatedly sampling the same area to generate detection and nondetection data required for estimation of species detection probabilities (MacKenzie et al. 2006). Given the logistical constraints associated with revisiting our study sites, we opted for a spatial replication approach using multiple surveyors to generate repeat survey data. During sampling, each surveyor searched for Suwannee Moccasinshell in all available habitats at a site, and we considered the total area covered by each surveyor to be an independent spatial replicate that was representative of the physical habitat conditions present at a site. Hence, we assumed that if Suwannee Moccasinshell was present at a site, then the species was available for detection by each surveyor (i.e., in each spatial replicate), and we defined detection as the probability of a surveyor detecting Suwannee Moccasinshell (or, equivalently, the probability of detecting the species in a spatial replicate). Under this study design, we note that the number of independent surveyors was equivalent to the number of spatial replicates. During all surveys, each surveyor spent approximately the same amount of time searching for Suwannee Moccasinshell, and we calculated the survey effort associated with each spatial replicate by dividing the total combined search time by the number of surveyors present. For example, if three surveyors collectively spent 3 person-hours at a given location, we assumed the effort per spatial replicate to be 1 h.

Model fitting and selection

Our modeling objective was twofold: first, we sought to determine the influence of SpringQ on Suwannee Moccasinshell occupancy; second, we sought to assess relations between SpringQ and survey effort on Suwannee Moccasinshell detection. We first developed a global model that contained two occupancy-model predictor variables, SpringQ and SpringQ2, and two detection-model predictor variables, Effort and SpringQ. We included Spring Q in detection models because it corresponds to stream size and we suspected detection may decline in higher order stream reaches because of reduced survey efficiency and generally more difficult survey conditions. To facilitate model-fitting, we standardized all continuous predictor variables with a mean of zero and standard deviation of one. We then fit 12 candidate models representing all possible combinations of these predictor variables, including the global model (Table 1). We used Akaike's Information Criterion (AIC; Akaike 1973) with a small-sample bias adjustment (AICc; Hurvich and Tsai 1989) to identify the best-approximating model, and we based all inferences on parameter estimates from the best-approximating model. To assess the influence of increasing the number of surveyors on Suwannee Moccasinshell detection probability, we used parameter estimates from the best-approximating model to calculate cumulative detection probability (p*) using the following formula, p* = 1 − (1 − p)N, where p represented the per-surveyor Suwannee Moccasinshell detection probability and N represented the number of independent surveyors (or, equivalently, the number of spatial replicates). We assessed goodness-of-fit for the global (all predictors) model by conducting 5,000 bootstrapped samples (simulated detection histories), comparing the bootstrapped detection-history frequencies with the frequency of observed detection histories, and calculating ĉ to test for overdispersion following MacKenzie and Bailey (2004). We implemented all models in Program R version 3.2.4 (R Core Development Team 2016) using the package unmarked (Fiske and Chandler 2011), and assessed goodness-of-fit using the package AICcmodavg (Mazerolle 2016).

Table 1.

Number of parameters (K), Akaike's Information Criterion adjusted for small sample size (AICc), ΔAICc, and model weights (wi) for the candidate set (i) of Suwannee Moccasinshell Medionidus walkeri occupancy models derived from surveys in the Suwannee River Basin in Georgia and Florida during 2013–2016. Spring Q is cumulative spring discharge, SpringQ2 is a quadratic cumulative spring discharge term, and effort is the amount of per-surveyor effort in hours.

Number of parameters (K), Akaike's Information Criterion adjusted for small sample size (AICc), ΔAICc, and model weights (wi) for the candidate set (i) of Suwannee Moccasinshell Medionidus walkeri occupancy models derived from surveys in the Suwannee River Basin in Georgia and Florida during 2013–2016. Spring Q is cumulative spring discharge, SpringQ2 is a quadratic cumulative spring discharge term, and effort is the amount of per-surveyor effort in hours.
Number of parameters (K), Akaike's Information Criterion adjusted for small sample size (AICc), ΔAICc, and model weights (wi) for the candidate set (i) of Suwannee Moccasinshell Medionidus walkeri occupancy models derived from surveys in the Suwannee River Basin in Georgia and Florida during 2013–2016. Spring Q is cumulative spring discharge, SpringQ2 is a quadratic cumulative spring discharge term, and effort is the amount of per-surveyor effort in hours.

Suwannee Moccasinshell was successfully detected at 26 of 220 sites surveyed throughout the Suwannee River Basin from 2013 to 2016 (Figure 1). Across all sites, the number of surveyors (i.e., spatial replicates) averaged three and ranged from two to eight. Per-surveyor effort varied among sites, averaging 0.38 person-hours and ranged from 0.034 to 1.5 person-hours. Across all 220 study sites, the number of upstream springs averaged 47 and ranged from 0 to 195, and 69 study sites had no upstream spring influence. Spring Q averaged 17.81 (SD = 22.59) cms across all study sites and ranged from 0 to 74.48 cms.

The best-approximating model contained Effort in the detection component and both SpringQ and SpringQ2 in the occupancy component (Table 2). This model was 1.84 times more plausible than the second best-approximating model, and there was little to no support for the remaining 10 candidate models (Table 1). The assessment of goodness-of-fit for the global model indicated that the observed frequency of detection histories agreed reasonably well with expected frequencies, and there was no evidence of overdispersion (ĉ = 0.3).

Table 2.

Parameter estimates and corresponding standard errors (SE) and 95% lower (LCL) and upper (UCL) confidence limits from the best-approximating Suwannee Moccasinshell Medionidus walkeri occupancy model derived from surveys in the Suwannee River Basin in Georgia and Florida during 2013–2016. Spring Q is cumulative spring discharge, SpringQ2 is a quadratic cumulative spring discharge term, and Effort is the amount of per-surveyor effort in hours.

Parameter estimates and corresponding standard errors (SE) and 95% lower (LCL) and upper (UCL) confidence limits from the best-approximating Suwannee Moccasinshell Medionidus walkeri occupancy model derived from surveys in the Suwannee River Basin in Georgia and Florida during 2013–2016. Spring Q is cumulative spring discharge, SpringQ2 is a quadratic cumulative spring discharge term, and Effort is the amount of per-surveyor effort in hours.
Parameter estimates and corresponding standard errors (SE) and 95% lower (LCL) and upper (UCL) confidence limits from the best-approximating Suwannee Moccasinshell Medionidus walkeri occupancy model derived from surveys in the Suwannee River Basin in Georgia and Florida during 2013–2016. Spring Q is cumulative spring discharge, SpringQ2 is a quadratic cumulative spring discharge term, and Effort is the amount of per-surveyor effort in hours.

Parameter estimates from the best-approximating model revealed that although Suwannee Moccasinshell occupancy was strongly and positively related to SpringQ, the probability of Suwannee Moccasinshell occurrence declined when contributing SpringQ exceeded ∼48 cms (Table 2; Figure 3). Parameter estimates also indicated that detection of Suwannee Moccasinshell was strongly and positively related to survey effort. Odds ratios suggested that for every 10-min increase in per-surveyor effort, the probability of detecting Suwannee Moccasinshell doubled. Our assessment of cumulative detection probability (p*) based on parameter estimates from the best-approximating model indicated that increasing the number of independent surveyors strongly and positively influenced the ability of at least one surveyor to detect Suwannee Moccasinshell (Figure 4). The assessment of cumulative detection probabilities revealed that to ensure a cumulative detection probability of ≥0.95 (i.e., a 95% chance of at least one surveyor detecting Suwannee Moccasinshell), a single surveyor would need to search for 1.31 h, whereas two, three, four, and five surveyors would each need to search for 0.88, 0.71, 0.60, and 0.52 h, respectively (Figure 4). Similarly, to ensure a cumulative detection probability of ≥0.90, a single surveyor would need to search for 1.13 h, whereas two, three, four, and five surveyors would each need to search for 0.77, 0.61, 0.51, and 0.44 h, respectively (Figure 4).

Figure 3.

Relationship between Suwannee Moccasinshell Medionidus walkeri occupancy probability and cumulative spring discharge (SpringQ) during 2013–2016 mussel surveys in the Suwannee River Basin in Georgia and Florida. Survey sites with SpringQ greater than or equal to 37 cubic m/s were only present in the mainstem Suwannee River, whereas sites that were heavily influenced by tidal forces were only present in main-stem Suwannee River sites with SpringQ ≥67 cubic m/s.

Figure 3.

Relationship between Suwannee Moccasinshell Medionidus walkeri occupancy probability and cumulative spring discharge (SpringQ) during 2013–2016 mussel surveys in the Suwannee River Basin in Georgia and Florida. Survey sites with SpringQ greater than or equal to 37 cubic m/s were only present in the mainstem Suwannee River, whereas sites that were heavily influenced by tidal forces were only present in main-stem Suwannee River sites with SpringQ ≥67 cubic m/s.

Close modal
Figure 4.

Relationship between per-surveyor effort and the probability of at least one surveyor detecting Suwannee Moccasinshell Medionidus walkeri during mussel surveys conducted in the Suwannee River Basin in Florida and Georgia from 2013 to 2016. Each curve represents the estimated cumulative species-detection probability assuming 1 (thinnest black line) to 5 (thickest black line) surveyors are conducting independent surveys at the same site.

Figure 4.

Relationship between per-surveyor effort and the probability of at least one surveyor detecting Suwannee Moccasinshell Medionidus walkeri during mussel surveys conducted in the Suwannee River Basin in Florida and Georgia from 2013 to 2016. Each curve represents the estimated cumulative species-detection probability assuming 1 (thinnest black line) to 5 (thickest black line) surveyors are conducting independent surveys at the same site.

Close modal

Our study suggests that springs have a strong influence on the contemporary distribution of Suwannee Moccasinshell. The strong positive relationship between Suwannee Moccasinshell occurrence and SpringQ begins to decline when SpringQ reaches ∼48 cms (Figure 3), for which there are several possible explanations: 1) a nonlinear relationship between occupancy and spring discharge exists because exceptionally high spring inputs are detrimental to Suwannee Moccasinshell populations; or 2) additional unmeasured factors that were not explicitly included in our model, such as tidal influence or changes in physical habitats, inhibit the ability of Suwannee Moccasinshell populations to become established and persist in downstream-most portions of the Suwannee River Basin with exceptionally high cumulative spring-discharge areas. We suspect that the latter explanation is more likely, although it is possible that both factors play a role in determining the distribution of Suwannee Moccasinshell throughout the Suwannee River Basin.

Effects of springs on occupancy

We hypothesized that proximity to springs may positively influence the occurrence of Suwannee Moccasinshell. However, preliminary modeling efforts incorporating the number and magnitude of discharge of respective springs within 100 m and 500 m from survey locations (Data S1, Supplemental Material) indicated little support for their inclusion as covariates in the occupancy model. We suspect the data were too sparse to estimate the influence of these covariates, because very few sites having springs within 100 m and 500 m were also occupied by Suwannee Moccasinshell. This seems to indicate a given spring's contribution to discharge and potentially water chemistry is more important to Suwannee Moccasinshell occupancy than proximity to that spring.

The high probability of Suwannee Moccasinshell occupancy with increasing SpringQ is not surprising given the headwaters of the Suwannee River proper, Withlacoochee River, and Santa Fe River derive their discharge predominantly from surface runoff via poorly drained floodplains yielding highly tannic water. Historically, Suwannee Moccasinshell and the federally endangered Oval Pigtoe Pleurobema pyriforme occurred in good numbers in small streams and medium-sized rivers in the Withlacoochee River and upper Santa Fe River system in areas with no or little spring input (Johnson et al. 2016). Today, Suwannee Moccasinshell appears to be extirpated from these areas and Oval Pigtoe occur at extremely low abundances. Droughts have affected this region in recent years, causing several of these historical headwater localities to stagnate or dry up altogether. The effect of drought has likely been exacerbated by water withdrawals in these areas because groundwater recharge is limited due to confinement of the aquifer (Grubbs 1998). Prolonged drought in small tributaries can depress freshwater mussel occupancy (Shea et al. 2013). The absence of springs and low aquifer recharge in these systems may be responsible for the extirpation of some Suwannee Moccasinshell populations (historical localities, Figure 1). Additionally, the upper Withlacoochee River in Georgia suffers from impaired water quality largely due to wastewater-treatment-plant discharges directly from Valdosta and indirectly from Tifton and Quitman. Wastewater treatment effluent more profoundly affects water quality in smaller rivers during droughts than at normal flow levels (Andersen et al. 2004); hence, wastewater effluent may have played a role in addition to drought in the decline of Suwannee Moccasinshell in the Withlacoochee River.

In contrast to the upper Withlacoochee and Santa Fe rivers, tributaries are relatively uncommon in the middle and lower Suwannee River, where flows are dominated by spring input that effectively replaces the discharge contributed by larger tributaries in typical alluvial river systems. In general, springs provide less polluted, sediment-free, and overall cleaner water, diluting tannic and polluted surface waters (Hornsby et al. 1998), and may also contribute more water quantity. The increase in water quality and quantity in areas dominated by spring influence may be responsible for the continued persistence of Suwannee Moccasinshell populations in these areas. Refugia provided by spring flows may be critical during drought conditions, when groundwater has the greatest influence on water quality (Hull et al. 1981) and effects of surface runoff and treated wastewater are greatest (Andersen et al. 2004).

Other watershed variables

Little research has focused on the effects of tidal influence on freshwater mussels. Strayer (1993) documented greater mussel species richness in tidal freshwater vs. nontidal reaches of the Susquehanna, Delaware, and Hudson river drainages. Although not quantified, Florida rivers appear to show opposite relationships, with greater species richness in nontidal vs. tidal reaches in the Escambia, St. John's, and St. Mary's rivers (Williams et al. 2014). It is possible that many species do not tolerate more lentic conditions (Haag and Williams 2014) associated with rising daily tides, thus reducing occupancy for some taxa. Additionally, extreme low drought flows could allow saltwater intrusion farther upstream than normal, which can negatively affect freshwater benthic macroinvertebrate communities (Attrill et al. 1996). In our study, the probability of Suwannee Moccasinshell occupancy dropped below 0.5 when SpringQ reached ∼68 cms (Figure 3), which is near the most upstream point at which tidal influence is present in the Suwannee River (Figure 1). Despite increasing SpringQ, we suspect the change in hydrology precludes occupancy of Suwannee Moccasinshell in the more tidally influenced reaches further downstream. It is uncertain whether drought-induced low spring flows and subsequent increases in salinity in tidal areas are a major factor influencing the distribution of Suwannee Moccasinshell. However, changes in flows and salinity concentrations may become a more likely scenario in the event of sea level rise associated with climate change. Lastly, tidal influences on Suwannee Moccasinshell occupancy in the lower Suwannee River may be compounded by cumulative effects of altered upstream nitrogen concentrations.

Highly elevated nitrogen levels have existed in the mid to lower main-stem Suwannee River Basin for some time (Pittman et al. 1997; Katz and Bohlke 2000; Katz et al. 2009). Springs have been shown to be primary vectors of this nitrogen. Pittman et al. (1997) documented a 160% increase in nitrogen in a 33-km reach of the Suwannee River receiving additional discharge only from springs and groundwater sources. The spring with the largest discharge (but not the highest nitrogen concentration) in that study contributed 550 kg of nitrates daily. Ammonia levels in springs are typically lower than U.S. Environmental Protection Agency Aquatic Life Ambient Water Quality Criteria, but ammonia is usually present in most groundwater wells and springs tested (Katz and Bohlke 2000; Katz et al. 2009; EPA 2013). Ammonia is extremely toxic to larval and juvenile freshwater mussels (Wang et al. 2007), and may persist longer in springs than in other systems because of low dissolved oxygen and eutrophic conditions limiting bacterial nitrification (Haag 2012). In areas where spring water has limited opportunity to mix with atmospheric oxygen (e.g., short spring runs, midchannel springs), the potential exists for ammonia to reach levels toxic to juvenile freshwater mussels. Hypoxic conditions or elevated levels of ammonia in spring outflows may explain why we observed minimal occurrences of Suwannee Moccasinshell within 100 m and 500 m of springs. Agricultural fertilizers, septic tanks, and animal waste are believed to be the primary sources of nitrogen in springs (Andrews 1994; Pittman et al. 1997; Katz and Bohlke 2000; Katz et al. 2009). The sites monitored by Pittman et al. (1997) were ∼32 km upstream of any tidal influence with a maximum SpringQ of 32 cms. If nitrate levels increase concomitantly with spring influence downstream to the tidal reach, the combination of altered water quality in the form of elevated nitrate levels plus effects of tidal influence may severely limit occupancy of Suwannee Moccasinshell in the lower Suwannee River. Despite multiple surveys (some including Self-Contained Underwater Breathing Apparatus), we were unable to detect live Suwannee Moccasinshell at historical localities in moderate to high tidal influence (Figure 1).

Detection

Suwannee Moccasinshell appears to be relatively easy to detect with a reasonable amount of survey effort. Our results indicate that one person searching for 1.31 h would have a 95% chance of detecting Suwannee Moccasinshell, given the species is present at a site (Figure 4). A multiple-surveyor study design where sites are surveyed by at least two people (preferably three or more) provides the added benefit that detection probabilities can be estimated directly from the data instead of assuming that detection was maximized during a single-person survey. Cumulative species detection (i.e., the probability of at least one surveyor detecting Suwannee Moccasinshell or, equivalently, the probability of detecting the species in at least one spatial replicate) can be increased by increasing 1) the amount of effort per-surveyor, 2) the number of independent surveyors, or 3) both.

From a multiple surveyor perspective, increasing either per-surveyor effort or the number of surveyors comes at a cost, so managers and biologists would have to design monitoring programs by balancing budget and time constraints with the cost of failing to detect a species when it is present at a location. We suspect that most field crews would almost always have at least two crew members, and three was the average across all surveys in the present study. Regardless, implementing a multiple surveyor approach reduces the total number of visits to the same locations by taking advantage of the fact that most survey teams can collect and record surveyor data independently. We therefore believe that the multiple surveyor approach for estimating species detection probabilities is a highly effective and efficient survey design, particularly for freshwater mussel species. We note that such an approach could easily be expanded to include multiple species in cases where managers are interested in assessing entire mussel assemblages or conducting watershed scale assessments of mussel communities (Shea et al. 2013; Pandolfo et al. 2016). The use of multiple surveyors during freshwater mussel surveys, however, remains relatively uncommon (but see Wisniewski et al. 2014 and Pandolfo et al. 2016). Our broad-scale study did not account for more detailed site-level factors, such as physical habitat conditions, that could have contributed to among site differences in detection of Suwannee Moccasinshell. Ongoing work in the Suwannee River Basin should incorporate local-scale substrate and flow attributes at survey sites, which allow for the development of more complex detection and occupancy models and further improve understanding of the predominant factors influencing detection and occupancy of Suwannee Moccasinshell.

Conclusions and management implications

Springs are an important component of the Suwannee River Basin: they buffer tannic waters draining the floodplains of the upper Basin, maintain flow during times of drought, provide habitats for endangered taxa, and are an important economic resource for the region. The contemporary distribution of Suwannee Moccasinshell appears to be strongly influenced by spring input, because recent surveys only found the species in portions of the Suwannee River Basin that receive significant contributions of spring discharge. Furthermore, most localities from which Suwannee Moccasinshell appears to be extirpated have no spring influence in their upstream watersheds (Johnson et al. 2016). Taken together, the ecological importance of springs to the Suwannee River Basin, and the aquatic biodiversity it supports, is remarkable. However, more research is needed to understand which factors most strongly influence the quantity and quality of Florida's groundwater, as well as the influence of groundwater resources on wildlife. In this study, we improved our understanding of the survey effort required to detect Suwannee Moccasinshell with a high degree of certainty, as well as broad-scale associations between springs and the species' distribution. Future work that includes more site-specific habitat characteristics coupled with watershed-scale variables and this multiple surveyor approach to occupancy and detection will provide an effective way to maximize data collected during valuable field time for management entities responsible for monitoring rare taxa.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.

Data S1. This is the data set used to evaluate the influence of spring discharge and effort on occupancy and detection of Suwannee Moccasinshell Medionidus walkeri in the Suwannee River Basin during 2013–2016. The data set contains two sheets: 1) a spreadsheet titled Holcomb et al. Data populated with all data used, and 2) a Data Key spreadsheet containing all 78 column headings from the Holcomb et al. Data spreadsheet and their definitions.

Found at DOI: http://dx.doi.org/10.3996/052017-JFWM-042.S1; also available at https://doi.org/10.5066/F7VX0DPM; Holcomb_et_al-JFWM-5-10-2017-Supplementary_Data.xlsx (83 KB XLSX).

Reference S1. Andrews WJ. 1994. Nitrate in groundwater and spring water near four dairy farms in north Florida, 1990–93: U.S. Geological Survey water-resources investigations report 94-4162.

Found at DOI: http://dx.doi.org/10.3996/052017-JFWM-042.S2 (561 KB PDF).

Reference S2. Grubbs JW. 1998. Recharge rates to the upper Floridan Aquifer in the Suwannee River Water Management District, Florida. U.S. Geological Survey water resources investigations report 97-4283.

Found at DOI: http://dx.doi.org/10.3996/052017-JFWM-042.S3 (9423 KB PDF).

Reference S3. Grubbs JW, Crandall CA. 2007. Exchanges of water between the upper Floridan Aquifer and the lower Suwannee and Lower Santa Fe rivers, Florida. U.S. Geological Survey professional paper 1656-C.

Found at DOI: http://dx.doi.org/10.3996/052017-JFWM-042.S4 (2021 KB PDF).

Reference S4. Hornsby D, Ceryak R, Zwanka W. 1998. Groundwater quality report. Suwannee River Water Management District, Live Oak, FL. Water resources report WR99-03.

Found at DOI: http://dx.doi.org/10.3996/052017-JFWM-042.S5 (3392 KB PDF).

Reference S5. Hull RW, Dysart JE Dysart, Mann IV WB. 1981. Quality of surface water in the Suwannee River Basin, Florida, 1968 through December 1977. U.S. Geological Survey water-resources investigations report 80-110.

Found at DOI: http://dx.doi.org/10.3996/052017-JFWM-042.S6 (39591 KB PDF).

Reference S6. Katz BG, Bohlke JK. 2000. Monthly variability and possible point sources of nitrate in ground water beneath mixed agricultural land use, Suwannee and Lafayette counties, Florida. U.S. Geological Survey water-resources investigations report 00–4219.

Found at DOI: https://dx.doi.org/10.3996/052017-JFWM-042.S7 (1965 KB PDF).

Reference S7. Pittman JR, Hatzell HH, Oaksford ET. 1997. Spring contributions to water quantity and nitrate loads in the Suwannee River during base flow in July 1995. U.S. Geological Survey water-resources investigations report 97-4152.

Found at DOI: http://dx.doi.org/10.3996/052017-JFWM-042.S8 (175 KB PDF).

Reference S8. Rosenau JC, Faulkner GL, Hendry CW Jr, Hull RW. 1977. Springs of Florida. Florida Bureau of Geology bulletin no. 31.

Found at DOI: http://dx.doi.org/10.3996/052017-JFWM-042.S9 (26934 KB PDF).

Reference S9. Suwannee River Water Management District. 2005. MFL establishment for the Lower Suwannee River and Estuary, Little Fanning, Fanning, and Manatee springs. Technical report.

Found at DOI: http://dx.doi.org/10.3996/052017-JFWM-042.S10 (34738 KB PDF).

Reference S10. Torak LJ, Davis GS, Strain GA, Herndon JG. 1996. Geohydrology and evaluation of stream–aquifer relations in the Apalachicola–Chattahoochee–Flint River Basin, Southeastern Alabama, Northwestern Florida, and Southwestern Georgia. U.S. Geological Survey water-supply paper 2460.

Found at DOI: http://dx.doi.org/10.3996/052017-JFWM-042.S11 (40918 KB PDF).

Reference S11. Warren GL, Bernatis JL. 2015. Status of the Ichetucknee Siltsnail (Floridobia mica) in Coffee Spring, Ichetucknee Springs State Park, Suwannee County, Florida, November 2015. Project report for Florida Department of Parks and Wildlife and U.S. Fish and Wildlife Services. Florida Fish and Wildlife Conservation Commission, Fish and Wildlife Research Institute.

Found at DOI: http://dx.doi.org/10.3996/052017-JFWM-042.S12 (786 KB PDF).

Reference S12. Wynn S, Borisova T, Hodges A. 2014. Economic value of the services provided by Florida springs and other water bodies: a summary of existing studies.

Found at DOI: http://dx.doi.org/10.3996/052017-JFWM-042.S13 (3380 KB PDF).

This research would not have been possible without the support of Florida Fish and Wildlife Commission personnel: M. Rowe, C. Morningstar, and J. Williams provided tremendous assistance with fieldwork; G. Warren was instrumental in creation, implementation, and funding of the Statewide Mussel Monitoring Program supporting this research; E. Leone and P. Schueller provided crucial insight on analyses used. Our resurveys of historical localities in Georgia would not have been possible without assistance from J. Wisniewski (GADNR) and S. Pursifull (USFWS). Additionally, this journal's Associate Editor, an anonymous reviewer, and D. Smith all provided fair and insightful critiques of our work that greatly improved this manuscript.

Any use of trade, product, website, or firm names in this publication is for descriptive purposes only and does not imply endorsement by the U.S. Government.

Akaike
H.
1973
.
Information theory and an extension of the maximum likelihood principle
.
Pages
267
281
in
Petrov
BN.
Csaki
F.
editors
.
Second international symposium on information theory
.
Budapest
:
Akademiai Kiado
.
Andersen
CB.
Lewis
GP.
Sargent
KA.
2004
.
Influence of wastewater treatment effluent on concentrations and fluxes of solutes in the Bush River, South Carolina, during extreme drought conditions
.
Environmental Geosciences
11
:
28
41
.
Andrews
WJ.
1994
.
Nitrate in groundwater and spring water near four dairy farms in north Florida, 1990–93
:
U.S. Geological Survey water-resources investigations report 94-4162
.
Attrill
MJ.
Rundle
SD.
Thomas
RM.
1996
.
The influence of drought-induced low freshwater flow on an upper-estuarine macroinvertebrate community
.
Water Research
30
:
261
268
.
Fiske
I.
Chandler
R.
2011
.
Unmarked: an R package for fitting hierarchical models of wildlife occurrence and abundance
.
Journal of Statistical Software
43
:
1
23
.
Grubbs
JW.
1998
.
Recharge rates to the upper Floridan Aquifer in the Suwannee River Water Management District, Florida
.
U.S. Geological Survey Water Resources Investigations Report 97-4283 (see Supplemental Material, Reference S2); also available: https://pubs.usgs.gov/wri/1997/4283/report.pdf (January 2018)
.
Grubbs
JW.
Crandall
CA.
2007
.
Exchanges of water between the upper Floridan Aquifer and the lower Suwannee and Lower Santa Fe rivers, Florida
.
U.S. Geological Survey professional paper 1656-C (see Supplemental Material, Reference S3); also available: https://pubs.er.usgs.gov/publication/pp1656C (January 2018)
.
Haag
WR.
2012
.
North American freshwater mussels; natural history, ecology, and conservation
.
New York
:
Cambridge University Press
.
Haag
WR.
Williams
JD
2014
.
Biodiversity on the brink: an assessment of conservation strategies for North American freshwater mussels
.
Hydrobiologia
735
:
45
60
.
Hornsby
D.
Ceryak
R.
Zwanka
W.
1998
.
Groundwater quality report. Suwannee River Water Management District, Live Oak, FL
.
Water resources report WR99-03 (see Supplemental Material, Reference S4); also available: http://dx.doi.org/10.3996/052017-JFWM-042.S5 (January 2018)
Hull
RW.
Dysart
JE.
Mann
WB.
IV
1981
.
Quality of surface water in the Suwannee River Basin, Florida, 1968 through December 1977
.
U.S. Geological Survey water-resources investigations report 80-110 (see Supplemental Material, Reference S5)); also available: https://pubs.usgs.gov/wri/1980/0110/report.pdf (January 2018)
.
Hurvich
CM.
Tsai
C.
1989
.
Regression and time series model selection in small samples
.
Biometrika
76
:
297
307
.
Johnson
NA.
Mcleod
JM.
Holcomb
J.
Rowe
M.
Williams
JD.
2016
.
Early life history and spatiotemporal changes in distribution of the rediscovered Suwannee Moccasinshell Medionidus walkeri (Bivalvia: Unionidae)
.
Endangered Species Research
31
:
163
175
.
Katz
BG.
Bohlke
JK.
2000
.
Monthly variability and possible point sources of nitrate in ground water beneath mixed agricultural land use, Suwannee and Lafayette Counties, Florida
.
U.S. Geological Survey water-resources investigations report 00-4219 (see Supplemental Material, Reference S6); also available: https://pubs.er.usgs.gov/publication/wri004219 (January 2018)
.
Katz
BG.
Sepulveda
AA.
Verdi
RJ.
2009
.
Estimating nitrogen loading to ground water and assessing vulnerability to nitrate contamination in a Large Karstic Springs Basin, Florida
.
Journal of the American Water Resources Association
49
:
607
627
.
Laist
DW.
Taylor
C.
Reynolds
JE.
2013
.
Winter habitat preferences for Florida manatees and vulnerability to cold
.
PLoS ONE
8
(
3
):
e58978
. doi.org/.
MacKenzie
DI.
Bailey
LL.
2004
.
Assessing the fit of site-occupancy models
.
Journal of Agricultural, Biological, and Environmental Statistics
9
:
300
318
.
MacKenzie
DI.
Nichols
JD.
Lachman
GB.
Droege
S.
Royle
JA.
Langtimm
CA.
2002
.
Estimating site occupancy rates when detection probabilities are less than one
.
Ecology
83
:
2248
2255
.
MacKenzie
DI.
Nichols
JD.
Royle
JA.
Pollock
KH.
Bailey
LL.
Hines
JE.
2006
.
Occupancy estimation and modeling: inferring patterns and dynamics of species occurrence
.
San Diego, California
:
Academic Press
.
Mazerolle
MJ.
2016
.
AICcmodavg: model selection and multimodel inference based on (Q)AIC(c). R package version 2.0-4
. .
Pandolfo
TJ.
Kwak
TJ.
Cope
WG.
Heise
RJ.
Nichols
RB.
Pacifici
K.
2016
.
Species traits and catchment-scale habitat factors influence the occurrence of freshwater mussel populations and assemblages
.
Freshwater Biology
61
:
1671
1684
.
Pittman
JR.
Hatzell
HH.
Oaksford
ET.
1997
.
Spring contributions to water quantity and nitrate loads in the Suwannee River during base flow in July 1995
.
U.S. Geological Survey Water-resources investigations report 97-4152 (see Supplemental Material, Reference S7); also available: https://fl.water.usgs.gov/PDF_files/wri97_4152_pittman.pdf (January 2018)
.
R Core Development Team
.
2016
.
R: a language and environment for statistical computing
.
Vienna
:
R Foundation for Statistical Computing
.
Rosenau
JC.
Faulkner
GL.
Hendry
CW
Jr,
Hull
RW.
1977
.
Springs of Florida
.
Florida Bureau of Geology bulletin no. 31 (see Supplemental Material, Reference S8); also available: http://ufdc.ufl.edu//UF00000232/00002 (January 2018)
.
Shea
CP.
Peterson
JT.
Conroy
MJ.
Wisniewski
JM.
2013
.
Evaluating the influence of land use, drought and reach isolation on the occurrence of freshwater mussel species in the Lower Flint River Basin, Georgia (U.S.A)
.
Freshwater Biology
58
:
382
395
.
Shea
CP.
Peterson
JT.
Wisniewski
JM.
Johnson
NA.
2011
.
Misidentification of freshwater mussel species: contributing factors, management implications, and potential solutions
.
Journal of the North American Benthological Society
30
:
446
458
.
Strayer
DL.
1993
.
Macrohabitats of freshwater mussels (Bivalvia:Unionacea) in streams of the Northern Atlantic Slope
.
Journal of the North American Benthological Society
12
:
236
246
.
Strayer
DL.
Smith
DR.
2003
.
A guide to sampling freshwater mussel populations
.
Bethesda, Maryland
:
American Fisheries Society
.
Monograph 8
.
Suwannee River Water Management District
.
2005
.
MFL establishment for the Lower Suwannee River and Estuary, Little Fanning, Fanning, and Manatee springs
.
Technical report (see Supplemental Material, Reference S9); also available: http://www.srwmd.state.fl.us/DocumentCenter/View/9471 (January 2018)
.
Torak
LJ.
Davis
GS.
Strain
GA.
Herndon
JG.
1996
.
Geohydrology and evaluation of stream–aquifer relations in the Apalachicola–Chattahoochee–Flint River Basin, Southeastern Alabama, Northwestern Florida, and Southwestern Georgia
.
U.S. Geological Survey water-supply paper 2460 (see Supplemental Material, Reference S10); also available: https://pubs.usgs.gov/wsp/2460/report.pdf (January 2018)
.
[ESA] U.S. Endangered Species Act of 1973, as amended, Pub. L. No. 93-205, 87 Stat. 884 (Dec. 28, 1973)
. .
[EPA] U.S. Environmental Protection Agency
.
2013
.
Final aquatic life ambient water quality criteria for ammonia—freshwater 2013
.
Federal Register
78
(
163
):
52192
52194
.
Wang
N.
Ingersoll
CG.
Hardesty
DK.
Ivey
CD.
Kunz
JL.
May
TW.
Dwyer
FJ.
Roberts
AD.
Augsperger
T.
Kane
CM.
Neves
RJ.
Barnhart
MC.
2007
.
Acute toxicity of copper, ammonia, and chlorine to Glochidia and juveniles of freshwater mussels (Unionidae)
.
Environmental Toxicology and Chemistry
26
:
2036
2047
.
Warren
GL.
Bernatis
JL.
2015
.
Status of the Ichetucknee Siltsnail (Floridobia mica) in Coffee Spring, Ichetucknee Springs State Park, Suwannee County, Florida, November 2015. Project report for Florida Department of Parks and Wildlife and U.S. Fish and Wildlife Services
.
Florida Fish and Wildlife Conservation Commission, Fish and Wildlife Research Institute (see Supplemental Material, Reference S11); also available
: http://static1.squarespace.com/static/5645105de4b0ba0f65eaa14a/t/57508c1286db43fc15d43cf4/1464896531563/Status + of + Floridobia + mica + in + Coffee + Spring + 11-2015.pdf (January 2018).
Williams
JD.
Butler
RS.
Warren
GL.
Johnson
NA.
2014
.
Freshwater mussels of Florida
.
Tuscaloosa
:
University of Alabama Press
.
Wisniewski
JM.
Rankin
NM.
Weiler
DA.
Strickland
BA.
Chandler
HA.
2014
.
Use of occupancy modeling to assess the status and habitat relationships of freshwater mussels in the lower Flint River, Georgia, USA
.
Walkerana
17
:
24
40
.
Wynn
S.
Borisova
T.
Hodges
A.
2014
.
Economic value of the services provided by Florida springs and other water bodies: a summary of existing studies (see Supplemental Material, Reference S12); also available: http://edis.ifas.ufl.edu/fe959 (January 2018)
.
Yoccoz
NG.
Nichols
JD.
Boulinier
T.
2001
.
Monitoring of biological diversity in space and time
.
Trends in Ecology and Evolution
16
:
446
453
.

Author notes

Citation: Holcomb JM, Shea CP, Johnson NA. 2018. Cumulative spring discharge and survey effort influence occupancy and detection of a threatened freshwater mussel, the Suwannee Moccasinshell. Journal of Fish and Wildlife Management 9(1):95–105; e1944-687X. doi:10.3996/052017-JFWM-042

The findings and conclusions in this article are those of the author(s) and do not necessarily represent the views of the U.S. Fish and Wildlife Service.

Supplemental Material