Applied Ichthyology J. Appl. Ichthyol. 28 (2012), 161–167 2012 Blackwell Verlag, Berlin ISSN 0175–8659
Received: March 17, 2011 Accepted: November 16, 2011 doi: 10.1111/j.1439-0426.2011.01930.x
Historical population structure of White Sturgeon in the Upper Columbia River detected with combined analysis of capture, telemetry and genetics By R. J. Nelson1 and D. S. O. McAdam2 1 Department of Biology, Centre for Biomedical Research, University of Victoria, Victoria, BC, Canada; 2Ministry of Environment, University of British Columbia, Vancouver, BC, Canada
Summary A combined analysis of capture ⁄ recapture, sonic tracking, and genetics was carried out to examine population structure of White Sturgeon of the Kootenay and Upper Columbia rivers. Analysis of capture ⁄ recapture and sonic telemetry data identiﬁed spatial patterns in habitat use, with individuals showing a preference for one of four high-use zones. Mitochondrial DNA analysis of groups of ﬁsh deﬁned by zonal preference and age suggests there is genetic population structure across this region, even between groups frequenting diﬀerent zones in the contiguous river system. This population structure likely has its origin in patterns of habitat use established prior to the impacts of human development when White Sturgeon in the region existed as assemblage of semireproductively isolated populations. Identiﬁcation of this historical population structure reveals unrecognized complexity regarding the spatial extent of historic habitat use and breeding structure of White Sturgeon in the region, which must be considered in further research and recovery eﬀorts.
Introduction Many species of sturgeon, including the White Sturgeon (Acipenser transmontanus), have experienced signiﬁcant population declines due to overﬁshing (Semakula and Larkin, 1968; DeVore et al., 1995), and river regulation (Coutant, 2004; see also Pikitch et al., 2005). Within the Canadian portion of the speciesÕ range, Smith et al., 2002 identiﬁed six populations, four of which are listed as endangered under CanadaÕs Species at Risk Act, and one of which is also listed under the US Endangered Species Act (Duke et al., 1999). Without eﬀective intervention extirpation will occur within one generation (Wood et al., 2007). Upper Columbia Rivers (UCR) White Sturgeon reside within the transboundary (US–Canada) reach of the Columbia River and within Arrow Lakes Reservoir (Fig. 1). Prior to development, the riverine habitat was largely contiguous. At present, the transboundary reach is bounded downstream in Washington State by the Grand Coulee Dam (completed 1942) and upstream in British Columbia by the Hugh Keenleyside Dam (completed 1968). Movement of ﬁsh into the major tributaries of the Upper Columbia River is limited by the Brilliant Dam on the Kootenay River (completed 1944) and the Waneta Dam on the Pend dÕOreille River (completed 1954) (Fig. 1). Recent population estimates for White Sturgeon in this area show 52 (95% CI = 37–92) individuals in Arrow Lakes Reservoir (Golder Associates, 2006), 1157 (95% U.S. Copyright Clearance Centre Code Statement:
CI = 414–1900) ﬁsh in the Canadian section of the transboundary reach (Irvine et al., 2007), and 2037 (95% CI = 1093–3223) individuals in the American portion of the transboundary reach (Howell and McLellan, 2006). Due to recruitment failure in all regions of this area, a recovery program was initiated in 2001 (UCWSRI, 2002). Populations of White Sturgeon in the region are currently deﬁned by the location of dams rather than by biological or genetic criteria (see Coutant, 2004). Delineation of population units is fundamental to ﬁshery management and recovery (Begg et al., 1999; Cope and Punt, 2009). In the case of White Sturgeon in the UCR, identiﬁcation of biologically appropriate population units is crucial for uncovering the cause of recruitment failure and for the design of an eﬀective recovery strategy. Genetic analysis is useful for the delineation of population units especially in the absence of detailed information regarding demographics, life history and habitat use, such as is the case for UCR White Sturgeon. Genetic population structure is also a component of biodiversity warranting protection to allow for high productivity and longterm persistence. The presence of unrecognized population structure increases the likelihood of extirpation for less populous groups within the larger population (Ryman et al., 1995; Taylor and Dizon, 1999). Thus, the failure to consider intraspeciﬁc genetic heterogeneity can lead to ineﬀective or counter productive management and conservation actions (Begg et al., 1999; Laikre et al., 2005). Examination of genetic variation at nuclear loci (typically microsatellite DNA) and mitochondrial loci (typically the control region) have been used with roughly equal frequency for the examination of the population genetics of sturgeon. Allozymes have been used much less frequently. Studies of the sturgeon intraspeciﬁc population structure have shown that populations inhabiting diﬀerent rivers separated by ocean coastline have a high likelihood of diﬀerentiation; this has been observed with both nuclear (Green Sturgeon: Isreal et al., 2004) and mitochondrial markers (Shortnose Sturgeon: Grunwald et al., 2002; Wirgin et al., 2002; Wirgin et al. 2010; White Sturgeon: Brown et al., 1992). In contrast to diﬀerences detected between river basins, it is not possible to generalize regarding population structure within a given watershed or lake system. In this case nuclear and mitochondrial DNA markers have each failed and succeeded in detecting population structure. Among the nuclear DNA studies that detected population structure are those of Brown et al., 1992 and Smith et al., 2002 (White Sturgeon); Dugo et al., 2004 (Gulf Sturgeon); Zhu et al., 2006 (Chinese Sturgeon); DeHaan et al., 2006 and Welsh et al.,
R. J. Nelson and D. S. O. McAdam
Fig. 1. Map of Upper Columbia River with geographic location (A); Columbia River upstream of Grand Coulee Dam (B); the transboundary reach (C). Putative groups deﬁned in the text indicated in bold adjacent to their associated river zone. Note that samples from Arrow Lakes Reservoir or Kootenay Lake come from upstream areas. Only principle dam locations are identiﬁed
2008 (Lake Sturgeon). The studies of Smith et al. (2002) and DeHaan et al. (2006) also detected genetic diﬀerentiation with mitochondrial DNA analysis. Genetic diﬀerentiation was not detected in Shortnose (Wirgin et al., 2005) or Lake Sturgeon (Wozney et al., 2011) with mitochondrial and nuclear DNA markers, respectively. The need for genetic analysis of White Sturgeon in the UCR is compelled by the results of Brown et al. (1992) and Smith et al. (2002) who detected population diﬀerentiation within the Fraser River, and Setter and Brannon (1992) who described the population structure in the Columbia River. The contention that White Sturgeon within the contiguous river system of the UCR is a single genetically homogenous population is currently implicit in management and recovery programs for this species. This assumption has not been tested; in fact, studies suggest the existence of spatially segregated groups in the UCR (Irvine et al., 2007; Van Poorten and McAdam, 2010), which challenges the assumption of a single homogeneous population. To further examine spatial segregation in habitat use, we employed analysis of a combination of tagging and tracking data. To examine whether population structure exists across the region we employed analysis of mitochondrial DNA variation as Smith et al. (2002) found this approach provided greater resolution, with less ambiguity than analysis of microsatellite markers. In combination, these complementary lines of inquiry can be used to test whether any detected diﬀerences in habitat use have a basis in historical reproductive isolation. Our overall aim is to identify the appropriate population units for future study and to promote conservation and recovery eﬀorts. Materials and methods Analysis of contemporary geographic partitioning
Discrete geographic zones were designated based on previously identiﬁed sturgeon high use areas (Hildebrand et al., 1999) and inspection of capture and telemetry data. These zones are deﬁned as: Arrow Lakes Reservoir (AR), Columbia River within 6.5 km downstream of Hugh Keenleyside Dam (HLK), Kootenay River within the 3.4 km downstream of Brilliant Dam to the conﬂuence with the Columbia River (BRL),
Columbia River from Fort Shepard [river km (rkm) 52] to Waneta (rkm 57) (WAN), downstream of the Canada ⁄ U.S. border including Lake Roosevelt (ROOS), and Kootenay Lake (KL) (Fig. 1). Analysis of capture data considered only the HLK, BRL, and WAN groups, due to insuﬃcient data for the ROOS group. Four groups (HLK, BRL, WAN, ROOS) were included in the analysis of multiyear sonic tracking data. The AR and KL groups were excluded from the spatial analysis since they are separated from the other groups by dams. Analysis of capture ⁄ recapture locations used data drawn from a database (maintained by the Upper Columbia River Sturgeon Recovery Initiative) which documents all sturgeon captures between 1993 and 2006 in research ﬁsheries conducted between Hugh Keenleyside Dam and the Canada–US border and within Arrow Lakes Reservoir. The database contains records for 607 individuals identiﬁed either by Floy or PIT tags. In order to limit biases caused by single captures at one location, the analysis considered only those ﬁsh which had been captured two or more times (N = 270). The sonic tracking database included information for 77 tagged ﬁsh monitored throughout the UCR from Keenleyside Dam to Lake Roosevelt. Data ﬁltering to remove noise was undertaken as part of the data collection studies (i.e. not the present study), and was based on criteria such as detections prior to tag release, a low number of detections at a site, and lack of proximity of sequential detections (e.g. a single detection at a given location without any detection at four or more intermediately located receivers). Four ﬁsh were excluded due to a combination of a high proportion of noise (100, 48, 57, and 100%) and a low number of total detections (13, 23, 37 and 5, respectively). The remaining 73 ﬁsh were monitored for 1.5–4 years between 2002 and 2007 with an array of 26 sonic detectors in ﬁxed locations throughout 180 km of the UCR (study details in Golder Associates, 2006 and Howell and McLellan, 2006). The database contains unique detection locations for each day a ﬁsh was located and the number of detections for each day-location combination. Evaluation of an individualÕs presence in high use habitat zones provided a more detailed examination of ﬁdelity by area. With both capture (N = 270) and telemetry (N = 73) data, individual ﬁsh were assigned to a geographic group based on
Population structure of White Sturgeon in the Upper Columbia River
the zone where they were most frequently detected. Individuals that had equal numbers of capture detections in diﬀerent geographic zones were equally apportioned between groups; ﬁsh with detections that could not be evenly split were not placed into any group. The proportion of all detections within each geographic zone was then evaluated for capture data (three groups) and telemetry data (four groups). Statistical evaluation of geographic ﬁdelity was based on comparison of the proportion of individuals with initial and subsequent detections in the same zone, and evaluated against the conservative null hypotheses that habitat use is equally split between a ﬁshÕs preferred zone and all other zones combined (Ho = 0.5). Analysis of capture data compared the geographic zone of ﬁrst detection (capture) with a random subsequent detection that was at least 6 months later. In telemetry analysis the ﬁrst detection is not independent of the decision to apply a tag so a random initial detection was selected from the ﬁrst 20 detections after 2 months at large, and the subsequent detection was randomly selected from remaining detections which occurred at least 6 months later (only 59 of 73 ﬁsh had suﬃcient data to meet this criteria). For both capture and telemetry data this exercise was repeated 100 times to generate a stable mean for the proportion of subsequent detections from the same geographic area. Genetic analysis
Selection of ﬁsh for analysis. The genetic analysis considers only ﬁsh born prior to recruitment failure to allow for analysis of historic population genetic structure. Recruitment failure began around 1970. In order to account for aging imprecision, only ﬁsh with estimated birth years of 1968 or earlier were included in this analysis. Selection of ﬁsh born in 1968 or earlier was based on year of capture and estimated age determined from analysis of ﬁn rays (Brennan and Cailliet, 1989). If a ﬁn ray based age estimate was lacking, age was estimated from the following age length relationship developed with information from the Upper Columbia River capture database: 2
Age = 0.0006 * FL þ 0:0862* FL + 3.9914 For genetic analysis, the designation of individuals into groups was based on the zone in which the ﬁsh was most commonly captured. This approach included ﬁsh represented by a single capture in a particular zone. This criterion was more liberal than the analysis of capture data, but was required due to the need to ensure suﬃcient sample sizes. Fish with equal numbers of captures in one or more zones were excluded from genetic analysis. After application of geographic and temporal criteria, the number of individuals analysed genetically per each group ranged from 21 to 56 (Table 1). Tissue sampling. Tissue biopsy samples used in this analysis were collected during a variety of baseline surveys (e.g. R, L&L Environmental Services Ltd, 1994; Howell and McLellan, 2006) from the areas shown in Fig. 1 and maintained as archived samples by the BC Ministry of Environment. Samples were preserved in 95% ethanol prior to DNA extraction. Laboratory analysis. Genetic variation in the mitochondrial DNA control region was examined using restriction fragment length polymorphism analysis. Laboratory methods followed
163 Table 1 Mitochondrial DNA haplotype frequencies of White Sturgeon
AR HLK WAN BRL KL ROOS
24 56 21 28 28 50
0.04 0.09 0 0.04 0 0.06
0.33 0.29 0.43 0.71 0.96 0.52
0.5 0.43 0.38 0.14 0.04 0.16
0.04 0.09 0.14 0.04 0 0.26
0 0.02 0 0 0 0
0 0.02 0 0 0 0
0.04 0.07 0.05 0.04 0 0
0.04 0 0 0.04 0 0
Sample groups: AR, Arrow Lakes Reservoir; HLK, Hugh Keenleyside Dam; WAN, Waneta; BRL, Brilliant Dam; ROOS, Lake Roosevelt; KL, Kootenay Lake. Haplotype designations follow those in Smith et al. (2002). (see text and Fig. 1 for details regarding sample group designation).
those in Smith et al. (2002). Brieﬂy, a 598 base pair fragment of the mtDNA control region ampliﬁed with the PCR primers L185 and H740 (Brown et al., 1993) was digested with the restriction enzymes MseI, SfcI, and Hsp92II. Composite haplotypes generated by scoring the individual enzyme digestion patterns were assigned to each ﬁsh according to the scheme of Smith et al. (2002).
Pair-wise (between groups) and global (including more than two groups) tests of genetic homogeneity were carried out with the genetic analysis program ARLEQUIN (Schneider et al., 2000) with 10 000 steps in the Markov chain, 1000 de-memorization steps and an a value of 0.05. Global and pair wise FST values were also calculated with ARLEQUIN with 100 permutations and a value of 0.05. Construction of a neighbor-joining tree based on Cavalli–Sforza and Edwards genetic distances was accomplished with PHYLIP (Felsenstein, 2005). Results Geographic analysis
Of 607 ﬁsh in the capture database 270 had been captured at least twice and had a total of 802 detections. Sonic tracking data provided a total of 41 803 unique location-day detections. Overall, 3.6% of the total telemetry detections were identiﬁed as noise, with individuals ranging from zero to 41.9%, with a median value of 1.8%. The remaining 40 278 detections included in the analysis led to an average, median, and range for number of detections per ﬁsh of 537, 413, and 78–1585, respectively. Examination of zonal preference showed that initial and subsequent detections were in the same zone 84% of the time (P < 0.001) based on capture data and 90% of the time based on telemetry data (P < 0.001), indicating a high level of ﬁdelity to a particular high use zone. Analysis of capture data also showed that out of all 1145 capture records for the 607 ﬁsh in the database, only 20 captures occurred outside the three deﬁned high use zones. Group designation based on capture location showed that for ﬁsh caught at least twice (N = 270), 249 ﬁsh were caught exclusively within one zone (HLK-101, BRL-11, WAN-137). Equal captures in multiple locations occurred for 22 ﬁsh (8% of total), with the speciﬁc ﬁsh locations (and number of occurrences) being: HLK-BRL (8), BRL-WAN (4), HLKWAN (6), HLK-BRL-WAN (2), HLK-other (1), BRL-other (1). After applying the criteria regarding group assignment, 108, 16 and 142 ﬁsh were apportioned to the HLK, BRL and
R. J. Nelson and D. S. O. McAdam HLK Capture (n = 110)
Telemetry (n = 5)
80 100 60 40
ROOS Telemetry (n = 43)
Capture (n = 145)
WAN Telemetry (n = 7)
Capture (n = 15)
Telemetry (n = 18)
Fig. 2. Percent of detections for each high use area for ﬁsh designated to four geographic areas using telemetry and capture data. Points are values for individual ﬁsh, and have been varied slightly in the x-axis due to high frequency of equal values. Vertical lines = standard error and Ô+Õ indicates the mean. Note appearance of bars within the individual data is due to the high frequency of equal values and the addition of a small variation in x-values. Sample groups: HLK, Hugh Keenleyside Dam; WAN, Waneta; BRL, Brilliant Dam; ROOS, Lake Roosevelt
WAN groups, respectively. Evaluation of each of these groups separately shows that the mean ﬁdelity to the zone associated with each group was 92.0, 64.5 and 93.4% for the HLK, BRL and WAN groups, respectively. Group classiﬁcation based on the location of most frequent telemetric detection resulted in the assignment of 18, 5, 7 and 43 individuals to the HLK, BRL, WAN, and ROOS groups, respectively. Fidelity of individual ﬁsh to a single geographic area ranged from 52 to 100%, with a mean of 95% (Fig. 2). Group means of 94, 98, 95 and 95% for the HLK, BRL, WAN and ROOS groups, respectively, indicating high ﬁdelity to the high use zones (Fig. 2).
Pairwise homogeneity tests provide for the speciﬁc examination of genetic diﬀerentiation between groups (Table 2). The failure to reject the null hypothesis of genetic homogeneity between AR and HKD (P = 0.82), combined with the signiﬁcant diﬀerences between these two groups and BRL,
Haplotype frequencies. A subset of the data analysed in Smith et al., 2002 was incorporated into the analysis reported here after application of habitat use and age criteria. Haplotype 7, which Smith et al., 2002 detected only in the Fraser River and Nechako River (a tributary to the Fraser), was not seen in our samples. Haplotype frequencies are shown in Table 1 for each group. As was the case in the study of Smith et al., 2002; haplotypes 2 and 3 predominated. The haplotype frequencies are visually displayed as pie charts in Fig. 3. Tests of genetic homogeneity. The null hypothesis of genetic homogeneity was rejected when all groups were included (P < 0.001), and when the physically isolated KL and AR groups were excluded (P < 0.001). These results indicate that there is genetic heterogeneity within the UCR watershed (including Kootenay Lake) as well as among groups within the contiguous riverine habitat between Hugh Keenleyside Dam and Grand Coulee Dam.
Fig. 3. Haplotype frequencies for samples groups of White Sturgeon shown in pie chart form adjacent to the sample group
Population structure of White Sturgeon in the Upper Columbia River
FST calculated among BRL, WAN, and ROOS only the value calculated between WAN and BRL was statistically signiﬁcant (see Table 2). Neighbour joining tree. A neighbour joining tree was created to visually represent the overall relationship among the sample groups (Fig. 4). The tree corresponds to the geographical patterns discerned from analysis of genetic homogeneity and FST. Discussion
Fig. 4. Neighbour joining tree showing overall genetic relationship among sample groups. Sample groups: AR, Arrow Lakes Reservoir; HLK, Hugh Keenleyside Dam; WAN, Waneta; BRL, Brilliant Dam; ROOS, Lake Roosevelt; KL, Kootenay Lake. Cavalli–Sforza distance bar shown to allow for comparison of distance between groups Table 2 Pairwise genetic homogeneity test and FST Site
AR HLK WAN BRL KL ROOS
24 56 21 28 28 50
0.82 0.68 0.03 <0.01 <0.01
0.73 0.004 <0.01 <0.01
0.12 <0.01 0.12
0.17 0.16 0.09
0.46 0.37 0.38 0.10
0.11 0.09 0.02 0.05 0.23
FST values (above diagonal) in bold = statistically signiﬁcant. Sample groups: AR, Arrow Lakes Reservoir, HLK, Hugh Keenleyside Dam; WAN, Waneta; BRL, Brilliant Dam; ROOS, Lake Roosevelt; KL, Kootenay Lake. (see text and Fig. 1 for details regarding sample group designation).
KL and ROOS (P-values range from <0.001 to 0.03) but not WAN (P = 0.68 and 0.73, respectively), suggests that AR and HKD groups were historically one distinct breeding population. Genetic homogeneity is also rejected between KL and all other sample groups except BRL (although this test approaches rejection with P = 0.06). For the ROOS group the hypothesis of homogeneity is rejected for all pairwise comparisons, except with the WAN group (P = 0.12). The WAN group appears to be intermediate between upstream (AR ⁄ HLK, BRL) and downstream (ROOS) groups since it only shows a signiﬁcant diﬀerence with Kootenay Lake samples. Sample size (N = 21) may limit the power of tests applied to the WAN group. Calculation of FST. The FST calculated across all sample groups was 0.14 (P < 0.001), and was 0.07 (P < 0.001) when the physically isolated populations KL and AR were excluded. Pairwise FST = 0 between AR and HKD, which further supports a genetic association of these two sample groups. AR and HKD both have signiﬁcant FST values ranging from 0.09 to 0.46 (all P values <0.001) with all other sample groups with the exception of WAN (FST = 0). The KL group has signiﬁcant FST values with all other groups. Among pairwise
Identiﬁcation of geographic and genetic distinction between groups of ﬁsh in the UCR presents a new perspective on White Sturgeon population structure in this region. These ﬁndings are concordant with other studies indicating genetic heterogeneity within sturgeon, and highlight the challenges to conservation and recovery planning in highly altered riverine habitats. The existence of discrete high use areas of the UCR for White Sturgeon was established in previous studies (Hildebrand et al., 1999; Howell and McLellan, 2006; Irvine et al., 2007). The present analysis extends this by showing that individual ﬁsh have ﬁdelity to a particular high use area; this may have its basis in behavioural patterns established prior to river regulation. The tendency towards residency within restricted river zones agrees with an increasing number of studies that reveal habitat ﬁdelity for sturgeon (Dugo et al., 2004; Parsley et al., 2008). Analysis of mtDNA of ﬁsh born prior to 1968 suggests that the groups identiﬁed by their contemporary habitat preference have a basis in historical reproductive isolation. The examination of capture and telemetry data presented here are complementary modes of analysis. Telemetry data provides a highly detailed time series for a limited number of ﬁsh, while capture data provides data for a large number of ﬁsh, but with fewer detections per ﬁsh. Capture data may be prone to bias due to unequal distribution of eﬀort. Telemetry analysis does not suﬀer from such limitations due to the ﬁxed locations of the receivers. Concordance of the ﬁdelity estimates of the HLK and WAN groups given by examination of the capture and telemetry data, suggests that such biases are minimal in the capture data. Seasonal spawning movements (Golder Associates, 2006; Howell and McLellan, 2006) do not appear to limit the detection of area based ﬁdelity, as these are short-term movements undertaken by only a portion of the population. Fidelity to particular areas of the river seems to predominate throughout the year. The low ﬁdelity for the BRL group as judged by analysis of capture data is the one exception to this general observation. Fish in the BRL zone have been observed to move out of their primary area of occupancy (Kootenay Eddy) when water depths at that location decrease below 15 m (L. Hildebrand, pers. comm.). Lower water depths in Kootenay Eddy would have been more common during the 1990s (a period of lower snowpack conditions) when most of the capture data was collected. The collection of telemetry data between 2002 and 2007, after most of the capture data was collected, may therefore explain the diﬀerence between the site ﬁdelity estimates for the BRL group for the two techniques. The high degree of residency estimated for most areas is also supported by the 3–5% migration rate estimated in between two population strata which are roughly equivalent to the WAN group and the HLK and BRL groups combined (Irvine et al., 2007). The presence of genetic diﬀerences between geographically deﬁned groups suggests concordance between contemporary
geographic partitioning and historic reproductive isolation. The observation that the KL group is the most genetically depauperate and distinct from all other groups is consistent with prior ﬁndings (Smith et al., 2002); this is likely due to the isolation of the KL population upstream of Bonnington Falls over the last 12 000 years (Duke et al., 1999). Despite the apparent isolation of the KL group, analysis of FST and tests of genetic homogeneity (Table 2) suggest a genetic aﬃliation of KL and the BRL group. This genetic relationship is supported by observations of the occasional downstream movement of Kootenay Lake ﬁsh (UCWSRI, 2002). These ﬁsh may have ﬁdelity to habitat downstream of Kootenay Lake or persistently attempt to return to Kootenay Lake. Examination of the genetic relationship between the AR and HKD groups is of particular interest as these two groups are currently separated by the Hugh Keenleyside Dam. Pair-wise tests of genetic homogeneity show that AR and HLK groups are not distinct from each other, have the same genetic relationship with all other sample groups, and are proximal on the neighbour joining tree (Fig. 4). Taken together these observations suggest that prior to the completion of Hugh Keenleyside Dam in 1968, AR and HLK ﬁsh comprised one population, which was distinct from other groups except possibly the WAN group. At the southern end of the study area, the ROOS group shows an analogous pattern as it appears to be distinct from all groups with the exception of WAN. In summary, genetic analysis suggests that AR ⁄ HLK and ROOS comprised historically distinct groups within the mainstem Upper Columbia River, with the WAN group being intermediate between them. In addition, the BRL group appears to be intermediate between all UCR groups and the KL group. With respect to the intermediate classiﬁcation of the WAN group it is important to note that the small sample size limits the power of statistical tests regarding this group. Future studies could greatly beneﬁt from increasing sample size to reduce such potential ambiguity. Increasing the sample size also holds the possibility for detection of low frequency haplotypes that were not detected in some samples. This too may allow for increased insight into the population structure of White Sturgeon in the region. Our assertion that White Sturgeon in the UCR historically existed as a collection of semi-isolated sub-populations may have important management implications since this perspective diﬀers from the prevailing view. Setter and Brannon (1992) identiﬁed genetic heterogeneity within the Columbia River watershed, including Lake Roosevelt, but they attributed observed diﬀerences to diﬀerential trapping of anadromous migratory ﬁsh rather than to genetic heterogeneity between resident groups. Their assumption of widespread historic anadromy seems unlikely given that populations residing proximal to marine waters show low levels of anadromy (DeVore and Grimes, 1993; Veinott et al., 1999). Without this assumption of anadromy Setter and Brannon, 1992 ﬁndings suggest the presence of historic genetic heterogeneity similar to the ﬁndings of Smith et al. (2002) for White Sturgeon of the Fraser River. This re-interpretation of the results of Setter and Brannon (1992), combined with present ﬁndings suggests that reproductively isolated populations also existed in the Columbia Basin prior to industrial development. In conjunction with other studies (e.g. Welsh et al., 2008) the results reported here contribute to an increasing number of examples of intraspeciﬁc genetic heterogeneity within sturgeon. Grunwald et al. (2002) identiﬁed diﬀerences
R. J. Nelson and D. S. O. McAdam
between Shortnose Sturgeon in the Kennebec and Androscoggin rivers, which share a common mouth, and studies of both Gulf Sturgeon (Dugo et al., 2004) and Chinese Sturgeon (Zhu et al., 2006), suggest genetic diﬀerentiation within a single watershed. Recently Welsh et al. (2008) found that groups of Lake Sturgeon found at several spawning locations in the Great Lakes represented distinct populations. Our ﬁndings are compatible with an emerging outlook which suggests that despite their capacity for long range migration, sturgeon may have a propensity for habituation to a speciﬁc region as well as reproductively driven natal homing (see Welsh et al., 2008). With regard to the spatial proximity of distinct groups examined in the present study, it is important to consider the implications of the habitat modiﬁcations caused by impoundment. While the majority of ﬁsh in the AR ⁄ HLK group currently reside downstream of the Hugh Keenleyside Dam, prior to the completion of this dam in 1968 over 300 km of additional upstream habitat was accessible to White Sturgeon (Westslope Fisheries Ltd, 2001). Similarly, the creation of Lake Roosevelt by Grand Coulee Dam in 1942 likely caused migration to the ﬂuvial areas upstream of the 230 km long reservoir (see Howell and McLellan, 2006). The genetic heterogeneity of White Sturgeon that now occupy the transboundary reach of the UCR (now approximately 100 km), was therefore generated when the available area was substantially more expansive (>800 lineal kilometres) and more conducive to reproductive isolation and genetic drift. Such river fragmentation has deleterious eﬀects on the viability of White Sturgeon, as suggested by the simulations of Jager et al. (2001). Recognition that contemporary patterns of habitat use and genetic heterogeneity are based on historical conditions underscores the challenges faced by recovery planners to promote population viability under highly altered conditions. While vestigial behaviours remain, changes of river regulation make it unclear whether the biological phenomena (e.g. natal homing) can be maintained. For example, recent tracking studies (e.g. Golder Associates, 2006; Howell and McLellan, 2006) suggest a mixing of groups during spawning periods, which may diminish historical genetic diﬀerences if reproduction were successful. Identiﬁcation of genetic heterogeneity as presented here may suggest that individual groups should be treated as distinct management units. However, since all of the groups are currently undergoing recruitment failure, such an approach may not be useful (see Green, 2005). The core value of the results reported here is the identiﬁcation of historical genetic heterogeneity suggesting reproductive isolation. Future work aimed at promoting recovery of White Sturgeon in the UCR should therefore consider this historical perspective of habitat use and population structure while pursuing the dual goals of population recovery and maintenance of genetic diversity. Acknowledgements This study was originally funded by BC Hydro, which also provided data for the movement and genetic analysis. We would like to acknowledge Jason McLellan, who provided genetic samples and telemetry data for Lake Roosevelt. The authors would also like to acknowledge the support of the BC Ministry of Environment and thank Glenn Cooper, Emily Rubidge, Kathryn Clark, and Chantal Rajotte for expert technical assistance in the laboratory.
Population structure of White Sturgeon in the Upper Columbia River
References Begg, G. A.; Friedland, K. D.; Pearce, J. B., 1999: Stock identiﬁcation and its role in stock assessment and ﬁsheries management: an overview. Fish. Res. 43, 1–8. Brennan, J. S.; Cailliet, G. M., 1989: Comparative age-determination techniques for White Sturgeon in California. Trans. Am. Fish. Soc. 118, 296–311. Brown, J. R.; Beckenbach, A. T.; Smith, M. J., 1992: Inﬂuence of Pleistocene glaciations and human intervention upon mitochondrial DNA diversity in white sturgeon (Acipenser transmontanus) populations. Can. J. Fish. Aquat. Sci. 49, 358–367. Brown, J. R.; Beckenbach, A. T.; Smith, M. J., 1993: Intraspeciﬁc DNA sequence variation of the mitochondrial control region of white sturgeon (Acipenser transmontanus). Mol. Biol. Evol. 10, 326–341. Cope, J. M.; Punt, A. E., 2009: Drawing the lines: resolving ﬁsheries management units with simple ﬁsheries data. Can. J. Fish. Aquat. Sci. 66, 1256–1273. Coutant, C. C., 2004: A riparian habitat hypothesis for successful reproduction of White Sturgeon. Rev. Fish. Sci. 12, 23–73. DeHaan, P. W.; Libants, S. T.; Elliott, R. F.; Scribner, K. T., 2006: Genetic population structure of remnant lake sturgeon populations in the upper Great Lakes basin. Trans. Am. Fish. Soc. 135, 1478–1492. DeVore, J. D.; Grimes, J. G., 1993: Migration and distribution of White Sturgeon in the Columbia River downstream from Bonneville Dam and adjacent marine areas. In: Status and habitat requirements of white sturgeon populations in the Columbia River downstream of McNary Dam. Unpubl. R. C. Beamesderfer, A. A. Nigro (Eds). Report to U. S. Department of Energy, Bonneville Power Administration, Portland, OR, USA, pp. 83–100. DeVore, J. D.; James, B. W.; Tracy, C. A.; Hale, D. A., 1995: Dynamics and potential production of White Sturgeon in the unimpounded Lower Columbia River. Trans. Am. Fish. Soc. 124, 845–856. Dugo, M. A.; Kreiser, B. R.; Ross, S. T., 2004: Conservation and management implications of ﬁne-scale genetic structure of Gulf Sturgeon in the Pascagoula River, Mississippi. J. Appl. Ichthyol. 20, 243–251. Duke, S. P.; Anders, P.; Ennis, G.; Hallock, R.; Hammond, J.; Ireland, S.; Lauﬂe, J.; Lauzier, R.; Lockhard, L.; Marotz, B.; Paragamian, V. L.; Westerhof, R., 1999: Recovery for the Kootenai River White Sturgeon (Acipenser transmontanus). J. Appl. Ichthyol. 15, 157–163. Felsenstein, J., 2005: PHYLIP (Phylogeny Inference Package) version 3.6. Department of Genome Sciences, University of Washington, Seattle, WA, USA. Golder Associates, 2006: Upper Columbia White Sturgeon stock monitoring and data management program – Synthesis report, 1 November 2003–31 March 2006. Prepared for BC Ministry of Environment, Nelson, BC, Canada. Green, D., 2005: Designatable units for status assessment of endangered species. Cons. Biol. 19, 1813–1820. Grunwald, C.; Stabile, J.; Waldman, J. R.; Gross, R.; Wirgin, I., 2002: Population genetics of shortnose sturgeon Acipenser brevirostrum based on mitochondrial DNA control region sequences. Mol. Ecol. 11, 1885–1898. Hildebrand, L.; McLeod, C.; McKenzie, S., 1999: Status and management of White Sturgeon in the Columbia River in British Columbia, Canada: an overview. J. Appl. Ichthyol. 15, 164–172. Howell, M. D.; McLellan, J. G., 2006: Lake Roosevelt White Sturgeon recovery project annual progress report, January 2004–March 2005. Washington Department of Fish and Wildlife Region 1 (Northeast), Spokane, WA, USA. Irvine, R. L.; Schmidt, D. C.; Hildebrand, L. R., 2007: Population status of White Sturgeon in the Lower Columbia River within Canada. Trans. Am. Fish. Soc. 136, 1472–1479. Isreal, J. A.; Cordes, J. F.; Blumberg, M. A.; May, B., 2004: Geographic patterns of genetic diﬀerentiation among collections of Green Sturgeon. N. Am. J. Fish. Manag. 24, 922–931. Jager, H. I.; Chandler, J. A.; Lepla, K. B.; Van Winkle, W., 2001: A theoretical study of river fragmentation by dams and its eﬀects on white sturgeon populations. Environ. Biol. Fishes 60, 347– 361. Laikre, L.; Palm, S.; Ryman, N., 2005: Genetic population structure in ﬁshes: Implications for coastal zone management. Ambio. 34, 111–119.
167 Parsley, M. J. L.; Popoﬀ, N. D.; van der Leeuw, B. K., 2008: Seasonal and diel movement of white sturgeon in the lower Columbia River. Trans. Am. Fish. Soc. 137, 1007–1017. Pikitch, E. K.; Doukakis, P.; Lauck, L.; Chakrabarty, P.; Erickson, D. L., 2005: Status, trends and management of sturgeon and paddleﬁsh ﬁsheries. Fish Fish. 6, 233–265. R, L&L Environmental Services Ltd., 1994: Status of White Sturgeon in the Columbia River, BC. Prepared for BC Hydro, Vancouver, BC, Canada. Ryman, N.; Utter, F.; Laikre, L., 1995: Protection of intraspeciﬁc biodiversity of exploited ﬁshes. Rev. Fish Biol. Fish. 5, 417–466. Schneider, S.; Roessli, D.; Excoﬃer, L., 2000: Arlequin ver. 2.0: a software for population genetic data analysis. University of Geneva, Genetics and Biometry Laboratory, Geneva. Semakula, S. N.; Larkin, P. A., 1968: Age, growth, food and yield of the white sturgeon (Acipenser transmontanus) of the Fraser River, British Columbia. J. Fish. Res. Board Can. 25, 2589–2602. Setter, A.; Brannon, E., 1992: A summary of stock identiﬁcation research on white sturgeon of the Columbia River (1985–1990). Report DOE ⁄ BP-97298-1. Bonneville Power Administration Portland, Portland, OR, USA. Smith, C. T.; Nelson, R. J.; Pollard, S.; Rubidge, E.; McKay, S. J.; Rodzen, J.; May, B.; Koop, B., 2002: Population genetic analysis of White Sturgeon (Acipenser transmontanus) in the Fraser River. J. Appl. Ichthyol. 18, 307–312. Taylor, B. L.; Dizon, A. E., 1999: First policy then science: why a management unit based on genetic criteria cannot work. Mol. Ecol. 8, S11–S16. UCWSRI, 2002: Upper Columbia River White Sturgeon recovery and management plan. Available at: http://uppercolumbiasturgeon.org/RecoveryEﬀorts/RecoveryPlan.pdf (accessed on 30 December 2012). Van Poorten, B. T.; McAdam, S. O., 2010: Estimating diﬀerences in growth and metabolism in two spatially segregated groups of Columbia River white sturgeon using a ﬁeld based bioenergetics model. Open Fish Sci. J. 3, 132–141. Veinott, G.; Northcote, T.; Rosenau, M.; Evans, R. D., 1999: Concentrations of strontium in the pectoral ﬁn ray of the white sturgeon (Acipenser transmontanus) by laser ablation –inductively coupled plasma – Mass spectrometry as an indicator of marine migrations. Can. J. Fish. Aquat. Sci. 56, 1981–1990. Welsh, A.; Hill, T.; Quinlan, H.; Robinson, C.; May, B., 2008: Genetic assessment of lake sturgeon population structure in the Laurentian Great Lakes. N. Am. J. Fish. Manag. 28, 572–591. Westslope Fisheries Ltd., 2001: Local Knowledge of Columbia River ﬁsheries in British Columbia, Canada. Prepared for ColumbiaKootenay Fisheries Renewal Partnership, Cranbrook, BC, Canada. Wirgin, I.; Waldman, J.; Stabile, J.; Lubinsk, B.; King, T., 2002: Comparison of mitochondrial DNA control region sequence and microsatellite DNA analyses in estimating population structure and gene ﬂow rates in Atlantic Sturgeon Acipenser oxyrinchus. J. Appl. Ichthyol. 18, 313–319. Wirgin, I.; Grunwald, C.; Carlson, E.; Stabile, J.; Peterson, D. L.; Waldman, J., 2005: Range-wide population structure of shortnose sturgeon Acipenser brevirostrum based on sequence analysis of the mitochondrial DNA control region. Estuaries 28, 406–421. Wirgin, I.; Grunwald, C.; Stabile, J.; Waldman, J. R., 2010: Delineation of discrete population segments of shortnose sturgeon Acipenser brevirostrum based on mitochondrial DNA control region sequence analysis. Conserv. Genet. 11, 689–708. Wood, C. C.; Sneep, D.; McAdam, S.; Korman, J.; Hatﬁeld, T., 2007: Recovery potential assessment for white sturgeon populations listed under the Species at Risk Act. DFO Can. Sci. Advis. Sec. Res. Doc. 2007 ⁄ 003. Wozney, K. M.; Haxton, T. J.; Kjartanson, S.; Wilson, C. C., 2011: Genetic assessment of lake sturgeon (Acipenser fulvescens) population structure in the Ottawa River. Environ. Biol. Fish 90, 183– 195. Zhu, B.; Zhao, N.; Shao, Z. J.; Lek, S.; Chang, J. B., 2006: Genetic population structure of Chinese sturgeon (Acipenser sinensis) in the Yangtze River revealed by artiﬁcial neural network. J. Appl. Ichthyol. 22, 82–88. AuthorÕs address: R. J. Nelson, Department of Biology, Centre for Biomedical Research, University of Victoria, PO Box 3020, STN CSC, Victoria, BC V8W 3N5, Canada. E-mail: [email protected]