Razorback sucker recruitment in Lake Mead, Nevada–Arizona, why here?

Abstract

Populations of the endangered razorback sucker (Xyrauchen texanus) have been reduced in the Colorado River during much of the last century. The inability of razorback sucker to recruit in the presence of nonnative fishes and altered flow regimes is thought to be the major factor contributing to their decline. Through funding from the Southern Nevada Water Authority and the US Bureau of Reclamation, we have conducted an ongoing razorback sucker research project on Lake Mead, Arizona and Nevada, since 1996. A major emphasis of this research has been to determine if natural recruitment was occurring in Lake Mead and identify reasons for that recruitment. Ages calculated using a nonlethal aging technique for 186 individual razorback sucker indicate the Lake Mead population is relatively young and that natural, wild recruitment has regularly occurred since the late 1970s. Comparisons of back-calculated ages of captured fish with historical Lake Mead water elevations provide evidence that a change in annual lake level fluctuations is the most likely mechanism that initiated this recruitment phenomenon. Lake level changes along with inundated terrestrial vegetation and turbidity in specific sites in Lake Mead may provide littoral nursery cover for larval and juvenile razorback sucker, allowing them to avoid predation.

Key words:

• Lake Mead
• littoral zone
• razorback sucker (Xyrauchen texanus)
• recruitment
• reservoir
• spawning

The razorback sucker (Xyrauchen texanus [Abbott]) was historically widespread and common throughout the larger rivers of the Colorado River Basin (Minckley et al. 1991), but its distribution and abundance are greatly reduced from historic levels. It is 1 of 4 endemic large-river fish species (Colorado pikeminnow [Ptychocheilus lucius], bonytail [Gila elegans], humpback chub [Gila cypha]) currently considered endangered by the US Department of the Interior (USFWS 1991). One major factor causing the decline of razorback sucker and other large-river fishes has been the construction of mainstem dams and the resultant cool tailwaters and reservoir habitats that replaced a once warm, riverine environment (Holden and Stalnaker 1975, Joseph et al. 1977, Wick et al. 1982, Minckley et al. 1991). Competition and predation from nonnative fishes that are successfully established in the Colorado River and its reservoirs have also contributed to their decline (Minckley et al. 1991, Mueller and Marsh 2002, Clarkson et al. 2005, Mueller 2005).

Razorback sucker persisted in several of the reservoirs that were constructed in the lower Colorado River Basin; however, these populations comprised primarily adult fish that were apparently recruited during the first few years of reservoir formation. The population of long-lived adults then disappeared 40–50 years following reservoir creation and the initial recruitment period (Minckley 1983). The largest reservoir population, estimated at 75,000 in the 1980s, occurred in Lake Mohave, Arizona and Nevada, but declined to approximately 30 wild individuals by 2008 (Marsh et al. 2003; C. Pacey, Marsh and Associates, LLC., May 2009, pers. comm.). Predation by bass (Micropterus spp.), common carp (Cyprinus carpio), channel catfish (Ictalurus punctatus), sunfish (Lepomis spp.) and other nonnative species seems to be the primary reason for lack of razorback sucker recruitment (Minckley et al. 1991, Marsh et al. 2003).

The Lake Mead razorback sucker population seemed to follow the trend of populations in other lower Colorado River Basin reservoirs. Lake Mead was formed in 1935 when Hoover Dam was closed, and razorback sucker were relatively common lakewide throughout the 1950s and 1960s, apparently from reproduction soon after the lake was formed. Their numbers became noticeably reduced in the 1970s, approximately 40 years after closure of the dam (Minckley 1973, McCall 1980, Minckley et al. 1991, Sjoberg 1995). From 1980 through 1989, neither the Nevada Department of Wildlife (NDOW) nor the Arizona Game and Fish Department collected razorback sucker from Lake Mead (Sjoberg 1995). This may be due in part to changes in their lake sampling programs; however, there was a considerable decline from the more than 30 razorback sucker collected during sportfish surveys in the 1970s. These results are not surprising and fit well within the pattern of razorback sucker dying off approximately 40–50 years following reservoir development, as was seen in other lower Colorado River Basin reservoirs.

In 1990 local anglers reported that razorback sucker were still found in 2 areas of Lake Mead (Las Vegas Bay and Echo Bay). Based on these reports, NDOW initiated limited sampling. From 1990 through 1996, 61 razorback sucker were collected; 34 from the Blackbird Point area of Las Vegas Bay and 27 from Echo Bay in the Overton Arm. Two razorback sucker larvae were collected by an NDOW biologist in 1995 near Blackbird Point, confirming suspected spawning in this area. A study was initiated in 1996 to answer questions about the Lake Mead razorback sucker population. This paper discusses specific methods and results related to the determination of, and suspected reasons for, recruitment of razorback sucker in portions of Lake Mead.

Study site

The primary study areas were the Echo Bay and Las Vegas Bay areas of Lake Mead, Nevada with the addition of the Muddy River/Virgin River Inflow in 2005 (Fig. 1). Limited effort was also expended near the Colorado River inflow area of Lake Mead in Arizona and in portions of Lake Mohave for physicochemical comparison.

Figure 1 Lake Mead and general study area locations.

Materials and methods

Netting

Trammel nets (91.4 m long by 1.8 m deep with an internal panel of 2.54, 3.81 or 5.08 cm mesh and external panels of 30.48 cm mesh) were used to collect subadult and adult razorback sucker. Nets were generally set with one end near shore in approximately 3–9 m of water, with the remainder of the net stretched out into deeper areas. All trammel nets were set in the late afternoon just before sundown and pulled the next morning shortly after sunrise. Sampling was generally conducted monthly within each study area from September through May, with variable effort within months. Netting locations were selected based on habitat usage patterns displayed by sonic-telemetered fish and/or prior capture locations of larval razorback sucker. Allowing the fish to dictate the netting locations helped to reduce sampler bias and maintain consistency throughout the study.

Upon capture, fish were removed from the nets as quickly as possible, and live fish were placed in large, water-filled containers for processing. Razorback sucker were scanned for passive integrated transponder (PIT) tags, tagged with PIT tags if they were not recaptures, measured (including standard length [SL]), weighed and released at the point of capture. Destron/Fearing Model TX1400 (400 kHz) PIT tags were primarily used to be consistent with other fishery research along the Colorado River.

Age determination

Otoliths have typically been used to age razorback sucker, but other boney parts have shown promise for aging (McCarthy and Minckley 1987). During the early years of razorback sucker studies on Lake Mead, 2 razorback sucker carcasses recovered from Lake Mead were aged using both otoliths and pectoral fin rays to evaluate the potential of using pectoral fin rays as a nonlethal technique for reliably aging razorback populations. Ages estimated from pectoral fin rays agreed with those obtained from sectioned otoliths for both carcasses. Surprisingly, both fish proved to be relatively young (ages 5 and 8). In addition, we extracted fin ray sections from 8 age-3 razorback sucker held by the NDOW at the Lake Mead hatchery and correctly aged them at 3 years. Additional age validation has taken place using multiple known-age fish stocked in Lake Mead and subsequently captured, where ages identified via fin rays correctly identified proper ages.

Aging of selected razorback sucker collected in adult trammel netting surveys began in earnest in 1999 after pectoral fins rays were shown to be a reliable structure for aging razorback sucker. In addition to the standard processing described previously, razorback sucker selected for aging were anesthetized, and a single, approximately 6-mm-long segment of the second left pectoral fin ray was surgically removed. Before the surgery, fish were anesthetized with a lake water bath containing MS-222, NaCl and slime coat protectant to avoid accidental injury from surgical tools, reduce surgery-related stresses and speed recovery. During surgery, standard processing was accomplished (weighing, measuring, PIT-tagging) and then, using sterilized bone snips, a sample was collected surgically. The connecting membrane between fin rays was cut using a scalpel blade, and the section was placed in a labeled envelope for drying. Personnel at BIO-WEST developed the surgical tool used to remove fin rays; it is essentially a matched pair of finely sharpened chisels welded to a set of Vise-Grips^TM. All surgical equipment was sterilized before use, and subsequent wounds were packed with antibiotic ointment to minimize postsurgical bacterial infestations and promote rapid healing. All razorback sucker were immediately placed in fresh lake water, allowed to recover and released as soon as they regained equilibrium and appeared recovered from the anesthesia. Vigilant monitoring of the fish was conducted during all phases of the procedure (Fig. 2).

Figure 2 Collection of pectoral fin ray and recaptured razorback sucker with healed fin.

In the laboratory, fin ray segments were processed with methods similar to McCarthy and Minckley 1987. The fin ray segments were embedded in thermoplastic epoxy resin and heat cured. This technique allowed the fin rays to be perpendicularly sectioned using a Buhler IsoMet®low-speed saw. Resultant sections were then mounted on microscope slides, sanded, polished and subsequently examined under a stereo-zoom microscope. Each sectioned fin ray was independently aged by at least 2 readers. Sections were then read a second time (by the readers) in instances where the assigned age was not under agreement. If discrepancies remained between the 2 ages after the second reading, the readers viewed the structure together and assigned an age.

Physicochemical investigations

A study was initiated in spring 2000 with the goal of identifying habitat parameters that may be contributing to the successful recruitment of razorback sucker in specific portions of Lake Mead but not in other portions of the lake or in Lake Mohave. Five study locations were established: Las Vegas Bay, Echo Bay, and Trail Rapids Bay in Lake Mead; and Arizona Bay and Tequila Cove in Lake Mohave. All 5 sites have razorback sucker spawning except Trail Rapid Bay, and recruitment is occurring in Las Vegas Bay and Echo Bay as discussed in this paper. Within each location, 5 sampling sites were chosen to capture parameter variability throughout the bay. Each site was defined as an approximate 200 m × 20 m rectangle that encompassed a portion of shoreline.

From 2000–2002, sampling occurred during March and May when larval razorback sucker are generally present. Within each 200 × 20 m site within the 5 study locations, water quality and plankton measurements were collected in 3 locations. Dissolved oxygen, nitrate, ammonium, total dissolved solids and turbidity (during the 2001 and 2002 study years), were taken with a HydroLab Datasonde 4a and Surveyor 4. A one-grab water sample was also collected, preserved with hydrochloric acid and stored on ice until it could be delivered to NEL Laboratories in Las Vegas for analysis of phosphorus. Plankton tows were conducted using a 30 × 90 cm plankton net with 153 μm mesh at each of the 3 locations. Tows were pulled horizontally for approximately 6 m at 0–0.5 m from the surface. Ten percent formalin was used to preserve samples that were later analyzed in a laboratory (Wetzel and Likens 1979). Finally, aquatic cover within each 200 × 20 m site was mapped using one or more of the following methods to assess vegetation in deeper areas: visual observation from boat, underwater video camera, and/or snorkeling.

Zooplankton samples were examined and enumerated under a 10–45× magnification using a Ward Counting wheel. Specimens were identified to order and, in some cases, family (Cladocera) level. Samples with large numbers of zooplankton were subsampled with a Henson Stemple pipette. Subsamples were taken by suspending the entire sample from a site in 50 mL of water and extracting two 2 mL subsamples with the pipette. The 2 subsamples were compared to determine the efficiency and accuracy of the subsampling method. After determining that little variation existed between the 2 subsamples, they were combined for analysis. The goal was to measure a subsample of 30 individuals from each taxonomic group in each sample. Zooplankton density was calculated as number of organisms per liter.

Results

Lake elevation

Month-end lake elevations were measured in feet above mean sea level (AMSL) and obtained from US Bureau of Reclamation's Lower Colorado Regional Office (Reclamation 2008). At the beginning of the study (Fig. 3), Lake Mead water surface elevation was approximately 1200 ft AMSL. After reaching a peak in 1998 and 1999 at 1215–1220 ft AMSL, Lake Mead declined more than 100 ft, reaching a month-end low of 1105 ft AMSL in May 2008.

Figure 3 Lake Mead month-end elevations, January 1980 to May 2008.

Age determination

BIO-WEST personnel captured 517 individual adult and subadult razorback sucker in Lake Mead between 1996 and 2008. In total, 218 razorback sucker have been captured at Las Vegas Bay, 248 at Echo Bay and 51 near the Muddy/Virgin River inflow areas. Fin ray sections were taken from 186 newly captured, wild razorback sucker.

Following the 2005 sampling season, back-calculated razorback spawning years were plotted against historical Lake Mead elevations from January 1935 to June 2005 to elucidate patterns of recruitment (Fig. 4). Back-calculation techniques showed that most of the razorback sucker aged were spawned between 1974 and 1998. The oldest fish aged was spawned around 1966. The majority of these fish originated in association with high lake elevations from 1978 to 1989, with another strong year class originating during the 1997–1998 period (Fig. 4).

Figure 4 Lake Mead hydrograph from January 1935 to June 2008 with the number of aged razorback sucker that were spawned each year through 2005 (gray bars) and through 2008 (black bars).

Through 2005, the ages of 78 fish identified that recruitment occurred fairly regularly from 1974 through 1998. Pulses in recruitment occurred from 1985 to 1987 and in 1998. At that time 35% of the razorback sucker aged in the study were aged at 10 years or younger, indicating recent (1990s) wild recruitment to the Lake Mead population. Interestingly, 4 of the 6 fish aged in 2005 (67%), were recruited in the 1997–1998 time period, coinciding with elevated lake levels present during the early years of this study (Fig. 4). Similarly, one of the fish collected in 2005 at the new Muddy/Virgin Inflow spawning area in the Overton Arm of Lake Mead was determined to be 6 years of age, further strengthening the 1998 cohort of razorback sucker. All of the 6-year-old razorback sucker collected in 2005 were expressing gametes, indicating that the cohort produced in 1998 was sexually mature. Subadult fish collected in 2004 (again, fish from the 1998 cohort) did not appear sexually mature at 5 years of age, suggesting that razorback sucker in Lake Mead mature near 6 years of age.

More recent aging data through 2008 depict razorback recruitment through 2006 (Fig. 4). When comparing year class numbers to the lake hydrograph, spikes in recruitment no longer coincide with high or increasing lake levels. Some of the strongest year classes (2000–2004) are now associated with declining lake levels (Fig. 4). Ages of razorback sucker through 2008 ranged from 2 to 36 years old. Based on these data, recruitment has occurred in some capacity regardless of associated lake levels during the spawning season.

Physicochemical investigations

Although zooplankton, temperature, dissolved oxygen, nitrates, ammonium, total phosphorus and total dissolved solids were all measured, little to no discernable differences or noteworthy trends were discovered for any of these parameters despite rigorous data collection and analytical efforts. One interesting result was that zooplankton were found in variable, yet similar abundances across sites, and we concluded that young razorback sucker were likely not food limited at the sites evaluated. Vegetative cover, however, was 1 of 2 items that seemed to differentiate locations where natural recruitment occurred, compared with locations where no recent recruitment was observed (Fig. 5). Comparing data collected during March sampling from all years and locations sampled, we found that mean percent cover was significantly higher at Echo Bay than at Arizona Bay, Las Vegas Bay, Tequila Cove and Trail Rapids Bay in 2001 (ANOVA, p< 0.04) and higher than all locations in 2002 (ANOVA, p< 0.04). Additionally, the mean percent cover at Las Vegas Bay in 2000 was significantly higher than the mean percent cover at Las Vegas Bay and Trail Rapids Bay in 2002 (ANOVA, p< 0.05). In May 2000 Echo Bay had significantly higher mean percent cover than Las Vegas Bay, Tequila Cove and Trail Rapids Bay in 2001 (ANOVA, p< 0.02), and Echo Bay, Las Vegas Bay and Trail Rapids Bay in 2002 (ANOVA, p< 0.004). Furthermore, as expected, mean percent cover at locations in Lake Mead known to produce young razorback sucker was found to decline from 2000–2002 in response to lowering lake elevations during the course of those years. As a result, the idea of long-term, climate-driven, lake-elevation changes being an important factor in the establishment of vegetative cover and, ultimately, the interaction between vegetative cover, diminished predation and the roles of these factors leading to pulses of razorback sucker recruitment was conceived (Fig.5).

Figure 5 Turbidity (a) and percent cover (b) comparisons from Echo Bay (EB) and Las Vegas Bay (LVB) in Lake Mead and Arizona Bay (AZB), Tequila Cove (TQC), and Trail Rapids Bay in Lake Mohave during May 2000-2002, as applicable.

Turbidity (another form of cover) was the second item that separated Echo Bay and Las Vegas Bay from the other locations evaluated within Lake Mead and Lake Mohave (Fig. 5). Turbidity levels at Las Vegas Bay and Echo Bay were higher than at the other locations sampled. When data for both years were combined, Las Vegas Bay had significantly higher turbidity levels than all other locations in March (ANOVA, p< 0.001) and all locations but Echo Bay in May (ANOVA,p< 0.001). While Echo Bay had higher turbidity levels than all locations except Las Vegas Bay in March and May, the difference was only significant in May (ANOVA, p< 0.001; Fig. 5). However, in hindsight, turbidity did not seem to be influenced by declining lake levels, at least to the extent that vegetative cover seemed to be (Fig. 5).

Discussion

The development of a nonlethal, accurate aging technique, coupled with the physicochemical study efforts, has allowed us to look more closely at razorback sucker recruitment in Lake Mead. A comparison of Lake Mead's elevation with the timing of razorback sucker recruitment (Fig. 4) suggests the change in the management of Lake Mead might be responsible for the apparent, sudden recruitment of razorback sucker in the late 1970s. From the 1930s to 1963, Lake Mead was either filling (a time when initial recruitment likely occurred and created the original lake population of razorback sucker) or was operated with a sizable annual fluctuation. The lake was drawn down approximately 100 ft in the mid 1960s as Lake Powell filled. Since that time, Lake Mead has been operated with relatively small annual fluctuations but relatively large multiple-year fluctuations. We suspect that the drawdown for filling of Lake Powell and a subsequent drawdown in the 1990s allowed terrestrial vegetation to become well established around the lake shoreline. The vegetation was then inundated as the lake rose, providing cover in coves and other littoral habitats that young razorback sucker may have inhabited. With only small annual fluctuations in water level, these areas of cover could remain intact for many years. During the pre-1970 period, the relatively large annual fluctuations in water level probably prevented the establishment of large areas of terrestrial vegetation around the shoreline. The presence of an individual razorback sucker greater than 30 years of age indicates that limited recruitment may have occurred during the period of 1966 to 1978, a time when lake elevations slowly rose to their highest levels (1978–1987) and the maximum amount of intact inundated vegetation probably existed in the lake.

Furthermore we found that both turbidity and vegetation were substantially more abundant in Las Vegas Bay and Echo Bay than the other coves, suggesting that both cover factors may play a role in recruitment (Fig. 5). Our recent aging data from Lake Mead show that recruitment pulses can and do occur at lowered lake conditions when vegetative cover may be limited. Given this, turbidity may be an important driving factor allowing recruitment under low lake level conditions on Lake Mead.

The combination of the recent capture of juvenile fish in Lake Mead, along with information obtained using the nonlethal aging technique described herein, shows that Lake Mead razorback sucker populations are naturally recruiting. No other population of razorback sucker is currently showing continued or sustained recruitment, although a few juveniles have been recently found (Modde 1996, Modde et al. 2001, Brandenberg et al. 2005, Golden and Holden 2005, Jackson 2005). Mueller (2005) and Clarkson et al. (2005) recently opined that persistence and recruitment of razorback sucker and other native fishes of the Lower Colorado River are not possible where nonnative fishes dominate. Lake Mead is a well known sport fishery with good populations of largemouth, smallmouth and striped bass as well as channel catfish (NDOW 2009). The information collected in this study shows that natural recruitment is occurring in a razorback sucker population in the presence of large populations of nonnative predators, indicating that limited recruitment may be possible when razorback sucker and nonnative fishes co-occur, if certain environmental conditions prevail. If these conditions can be isolated through more exploration, it may be possible that active management can facilitate their simulation elsewhere in the basin to allow for additional recruitment of wild and stocked populations of razorback sucker.

Acknowledgments

We thank the Southern Nevada Water Authority, which provided funding and facilitated field crew help; the US Bureau of Reclamation, which provided funding, equipment, storage facilities and technical support; and the Nevada Department of Wildlife, which also provided equipment, technical support and field support for the project. We would also like to thank other agencies that joined as cooperators at the beginning of the study, including the National Park Service, which provided residence facilities in their campgrounds; and the US Fish and Wildlife Service, for its assistance with permitting. We also thank P.D. Abate, J.D. Ruppert and T.L. Welker along with other biologists and technicians who devoted their time, energy and ideas to our study.