Habitat Restoration As a Means of Controlling Non-Native Fish in a Mojave Desert Oasis G

Home , Desert pupfish, Mosquitofish, Sailfin molly

Habitat Restoration as a Means of Controlling Non-Native Fish in a Mojave Desert Oasis G. Gary Scoppettone,1,2 Peter H. Rissler,1 Chad Gourley,3 and Cynthia Martinez4

Abstract depth (TD) than non-native Sailfin molly (Poecilia lati- Non-native fish generally cause native fish decline, and pinna) and Mosquitofish (Gambusia affinis) in warm once non-natives are established, control or elimination is water stream habitat, and Ash Meadows speckled dace in- usually problematic. Because non-native fish colonization habited significantly faster water than non-natives in cool has been greatest in anthropogenically altered habitats, water stream habitat. Modification of the outflow of Kings restoring habitat similar to predisturbance conditions may Pool Spring from marsh to warm water stream, with offer a viable means of non-native fish control. In this MWCV, TD, and temperature favoring native fish, investigation we identified habitats favoring native over changed the fish composition from predominantly non- non-native fish in a Mojave Desert oasis (Ash Meadows) native Sailfin molly and Mosquitofish to predominantly and used this information to restore one of its major warm Ash Meadows pupfish. This result supports the hypoth- water spring systems (Kings Pool Spring). Prior to restora- esis that restoring spring systems to a semblance of pre- tion, native fishes predominated in warm water (25–32°C) disturbance conditions would promote recolonization stream and spring-pool habitat, whereas non-natives pre- of native fishes and deter non-native fish invasion and dominated in cool water (•23°C) spring-pool and marsh/ proliferation. slack water habitat. Native Amargosa pupfish (Cyprino- don nevadensis) and Ash Meadows speckled dace (Rhi- Key words: Ash Meadows, Cyprinodon nevadensis, habi- nichthys osculus nevadensis) inhabited significantly faster tat manipulation, Mojave Desert, non-native fish control, mean water column velocities (MWCV) and greater total Rhinichthys osculus nevadensis, thermo springs.

Introduction non-natives. Because non-native fish invasions are often Non-native fish alter native aquatic communities, are an associated with anthropogenically disturbed environments agent of native fish decline and extirpation (Taylor et al. (Moyle & Nichols 1974), restoration of the aquatic system 1984; Moyle et al. 1986; Miller et al. 1989; Minckley & may counteract the effects of human disturbance. In un- Deacon 1991), and are difficult to control or eliminate. disturbed or more natural habitat, native fishes may have Chemical treatment is often unsuccessful (Meffe 1983; a better chance of tolerating a non-native fish invasion Rinne & Turner 1991; Meronek et al. 1996) and is lethal because such conditions contributed to evolution of the to aquatic invertebrates (Morrison 1987; Mangum & native species (Southwood 1988; Ricklefs 1991). Thus, Madrigal 1999), and physical removal has usually met with restoring aquatic habitat to predisturbance conditions limited success (Meronek et al. 1996; Knapp & Mathews may serve to promote native fishes (Baltz & Moyle 1993; 1998). Thus, alternate and innovative strategies need to be Moyle & Light 1996) if environmental conditions that developed to enhance population expansion of natives at favor natives over non-native species are emphasized in the expense of non-natives. the restoration. Physical changes in aquatic ecosystems can alter fish Ash Meadows National Wildlife Refuge (AMNWR) in community structure, population demographics, and rela- Northern Mojave Desert offers an opportunity to study tive abundance of species (Bain et al. 1988; Rabeni & the feasibility of non-native fish control through habitat Jacobson 1993; Gido & Propst 1999). Therefore, non- manipulation associated with spring and stream restora- native fish control may be possible through the creation or tion. Ash Meadows’ water resources are a series of ther- re-creation of habitat that promotes native fishes over mal springs with discharge sufficiently low that flows are manipulable. The spring habitats have been significantly altered and invaded by non-native fishes. Typical of spring 1 Biological Resources Division, U.S. Geological Survey, 1340 Financial systems of the southwestern United States, there are few Boulevard, Suite 161, Reno, NV 89502, U.S.A. native species present (Miller 1961). In addition, Ash 2 Address correspondence to G. G. Scoppettone, email gary_scoppettone@ usgs.gov Meadows has few non-native fish species, thus simplifying 3 Otis Bay, Incorporated, 110 Mule Deer Drive, Reno, NV 87523, U.S.A. the task of identifying habitat conditions favoring the 4 U.S. Fish and Wildlife Service, 4701 N. Torrey Pines Drive, Las Vegas, NV native fishes over non-natives. The natives in spring sys- 89130, U.S.A. tems of most of AMNWR had Amargosa pupfish Ó 2005 Society for Ecological Restoration International (Cyprinodon nevadensis), Ash Meadows speckled dace

JUNE 2005 Restoration Ecology Vol. 13, No. 2, pp. 247–256 247 Habitat Restoration as a Means of Controlling Non-Native Fish

(Rhinichthys osculus nevadensis), or both species. The Meadows speckled dace, and the now extinct Ash Mead- primary non-native fishes are Mosquitofish (Gambusia ows poolfish (Empetrichthys merriami) (Miller 1948, 1961). affinis) and Sailfin molly (Poecilia latipinna). Centrally located on the eastern edge of AMNWR is a com- Like most aquarium fish invaders of Southwestern ther- plex of higher-elevation (710 m msl) springs that are suffi- mal springs, Sailfin molly evolved in lentic or slack water ciently isolated physiographically to harbor another endemic habitat (Harrington & Harrington 1961). This is also true fish, Warm Springs pupfish (C. n. pectoralis) (Miller 1948). of the Mosquitofish, which has successfully invaded slack Each spring harboring Warm Springs pupfish is character- and lentic temperate to warm water throughout the west- ized by low water discharge (<0.01 m3/second) and warm ern United States (Swanson et al. 1996). Our impression water (32–33°C). The most physiographically isolated of from initial observations in Ash Meadows was that pupfish Ash Meadows fishes is the Devils Hole pupfish (C. diabolis), predominated over non-native fishes in warmer water, which occurs at 730 m msl in a 15-m depression on a hillside. especially lotic warm water and that non-native fishes pre- At the northeast edge of AMNWR, Devils Hole has been dominated in cool water, especially lentic habitat. Further- part of the Death Valley National Park system since 1952 more, Ash Meadows speckled dace do occur in warm (Deacon & Williams 1991). water but flourish in cool fast water. We hypothesized that Prior to its acquisition by the U.S. Fish and Wildlife non-native lentic and slack water species can be controlled Service for the preservation of its endemic species, Ash through spring system restoration so outflow channels Meadows’ landscape had been greatly altered. Carson retain their warmer temperature with velocities conducive Slough was mined for peat and surrounding areas cleared to pupfish and speckled dace but detrimental to non- and leveled for agricultural use. Several springheads were native lentic forms. In this study we investigated habitat fitted with pumps, eliminating surface flow (Deacon & favoring native Ash Meadows pupfish and Ash Meadows Bunnell 1970; Pister 1974; Deacon & Williams 1991), and speckled dace over non-native Sailfin molly and Mosqui- spring-pools were enlarged. Water was diverted from nat- tofish. We also tracked species composition of a spring ural stream courses to a few earthen and concrete ditches outflow before and after it was restored to promote native and either stored in reservoirs or used directly for crop or fishes over non-natives. pasture irrigation. Along with loss of natural channel, there was loss of native riparian corridors, and non-native vegetation became established along several of the new or Study Site altered stream courses. Massive levees constructed to pro- The Mojave Desert is the driest region in North America, tect agricultural fields and irrigation ditches from flash with Ash Meadows its largest oasis, and it harbors one of floods eliminated these intermittent events in most spring the greatest numbers of endemic species, for its area, in outflows. Mosquitofish became established in the lower- North America (Sada 1990). Ash Meadows is situated elevation springs by the 1930s (Miller 1961), followed by within the Amargosa River Drainage, subdrainage of the Sailfin molly in most of the lower-elevation springs in the Death Valley System at the southwestern edge of Nevada, 1960s (Deacon & Bunnell 1970). Also in the 1960s Large- just east of Death Valley, California (Fig. 1) (Hubbs & mouth bass (Micropterus salmoides) were stocked in Ash Miller 1948; Miller 1948). Ash Meadows’ primary water Meadows’ largest reservoir, Crystal Reservoir, and Brown sources are approximately 24 thermal springs within a 7-km bullhead (Ameiurus nebulosus) inhabited Davis Spring radius and with cumulative discharge of 0.65–0.68 m3/ until removal through chemical treatment in 1996. Other second (Dudley & Larson 1976). Garside and Schilling introduced aquatic species include Bullfrog (Rana (1979) reported near-source water temperatures from 18.0 catesbeiana), Crayfish (Procambarus clarki), and Oriental to 33.0°C, with highly mineralized water and dissolved snail (Melonoides tuberculata). Since the summer of 1993, oxygen well below saturation. AMNWR staff has annually removed Sailfin molly, Historically, Ash Meadows spring water emerged from Mosquitofish, and Crayfish from the larger spring-pools carbonated aquifers into small spring-pools and then (St. George, unpublished data). through narrow outflow channels to discharge into Carson Ash Meadows native fishes are all federally listed as Slough. Using 1:24,000 scale 10-m digital elevation models endangered (U.S. Fish and Wildlife Service 1983), and and hydrologic modeling in ArcView GIS (ESRI, Red- their recovery is predicated on habitat restoration and lands, CA, U.S.A.), we modeled the historic direction of elimination of non-natives (Sada 1990). The once wide- the larger spring outflows and determined historic loca- spread Ash Meadows speckled dace was relegated to three tions of over 45 km of perennial stream channel (Fig. 1). spring complexes, a small fraction of the range described The simulated stream courses do not incorporate stream by Miller (1948). Warm Springs pupfish has been extir- sinuosity, which would increase absolute channel length. pated from one of the six springs and its range reduced in Most Ash Meadows springs lie within an elevation of the others (Williams et al. 1985). Ash Meadows pupfish 655–700 m above mean sea level (msl) and were histori- has remained widespread even though much of its former cally connected via Carson Slough. Until 50 years ago, habitat had been eliminated (Pister 1974; Sada 1990). three endemic fishes were known from these springs: Ash When we began monitoring Ash Meadows fish popula- Meadows pupfish (Cyprinodon nevadensis mionectes), Ash tions in 1989, there were less than 10 km of spring lotic

248 Restoration Ecology JUNE 2005 Habitat Restoration as a Means of Controlling Non-Native Fish

Figure 1. Map of AMNWR showing major spring systems as they existed in 1989 along with modeled historic connections of lower-elevation spring outﬂow channels preanthropogenic disturbance. Inset is AMNWR in relation to the state of Nevada.

habitat, and this included concrete and earthen-lined irri- (C. Gourley & E. Ammon 1997, Otis Bay, Incorporated, gation ditches. Of this habitat, less than 5 km flowed in personal communication). It traveled southwest from the a well-incised channel with sufficient slope that water tem- spring-pool, draining Kings Pool Spring Marsh before perature was near constant and offered a variety of water heading south (Fig. 2). In this article we refer to the con- velocities. In 1989 the source of one of these thermal crete outflow channel as Kings Pool Spring Concrete Out- streams (Kings Pool Spring) partially broke from its con- flow and the excavated meander channel as Kings Pool crete channel approximately 50 m from the source pool to Spring Restored Outflow. form a headwater marsh, and by 1994 almost the entire Discharge of individual Ash Meadows springs ranged 0.07 m3/second flow contributed to this proximal marsh from 0.04 to 0.18 m3/second. Emerging water typically is- (Fig. 2). In 1996 a 1-km meander channel was excavated sued into spring-pools ranging from 7 to 30 m in diameter along what was determined to be a historic outflow course and 0.5 to 10 m in depth and fed into stream channels

JUNE 2005 Restoration Ecology 249 Habitat Restoration as a Means of Controlling Non-Native Fish

Figure 2. Map showing Kings Pool Spring Concrete Outflow system as it was in 1989 and Kings Pool Spring Restored Outflow as it was in 2003. ranging from 0.1 to 2.5 km in length before discharging 1-mm mesh and baited with dog food. It was assumed that into a marsh, pasture, or reservoir. Channels sampled both methods would yield relative species abundance, and ranged from 0.7 to 1.5 m in width and 0.2 to 0.8 m in we did not mix methods within our five macrohabitats. depth. Warm water spring-pools, warm water stream, and cool water stream were sufficiently open and clear that the snorkel method was used. Representative reaches along Materials and Methods the outflow length (of warm water and cool stream) con- stituting 10–40% of the available habitat were snorkeled, Macrohabitat and Relative Species Composition whereas in spring-pools all fish were counted. In cool To determine habitat that favored natives over non- water pool and cool water marsh, water was often shallow, natives, we monitored fish species composition in five turbid, or heavily vegetated and could not be effectively macrohabitat categories: warm water spring-pool, warm snorkeled so we used ‘‘Gee’’ traps to obtain representative water stream, cool water spring-pool, cool water stream, samples. and marsh-like habitat. Warm water spring-pool and We used a paired t test to determine significant predom- stream habitats had near-constant year-round tempera- inance of native fishes in four of our five habitat categories ture from 25 to 33°C; this included six warm spring-pools (Sokal & Rohlf 1995). Cool water stream habitat was not and seven near-source outflows. Six cool water spring- tested because there was only one. For each of the four pools also had near-constant temperature but the temper- habitat types percentage of natives was paired against ature ranged from 18 to 21°C. Cool water stream habitat percentage of non-natives. had daily or seasonal water temperature falling below 22°C, but because of alteration of Ash Meadows’ spring systems, there was only one spring outflow (Jackrabbit) of Mean Water Column Velocity and this habitat type. Thermal springs cool in a downstream Total Depth Used and Available direction, and only Jackrabbit had a sufficiently long out- We quantified mean water column velocity (MWCV) and flow to have cool water stream habitat. Marshy habitats, total depth (TD) used by native and non-native fishes in of which seven were sampled, exhibited a wide range in warm and cool water stream habitat. MWCV and TD water temperature and included old irrigation channels were studied because they can be integrated into stream clogged with Cattails (Typha) and Bulrush (Scripus), restoration. Using mask and snorkel, fish were sighted and water seeping from man-made channels, and water that location was marked with a numbered washer to be revis- had originally been spread to irrigate pasture. ited after completing a 10-m reach. Before placing num- Monitoring was by direct fish count with mask and snor- bered washers, fish species, size, and washer number were kel or using standard ‘‘Gee’’ minnow traps lined with recorded. Warm water stream habitat use was studied in

250 Restoration Ecology JUNE 2005 Habitat Restoration as a Means of Controlling Non-Native Fish the Crystal Spring outflow where Ash Meadows pupfish the same years because flows of both had been manipu- occurred with Sailfin molly and Mosquitofish and in Jack- lated during the course of the study and the two systems rabbit Spring outflow, which supported Ash Meadows were only similar during these two time periods. speckled dace along with the three species found in Crys- tal Spring outflow. Measurements in Crystal Spring were taken at 10 stations 10 m in length and spaced 250–500 m Kings Pool Spring Marsh versus apart, extending 2,400 m downstream, and including con- Kings Pool Spring Restored Outflow crete side channels. Sampling took place in winter and In 1997 Kings Pool Spring Marsh was drained and the summer of 1990, 1991, and 1992; a change in refuge water water routed into an excavated channel configured to management curtailed sampling after summer 1992. In simulate the historic outflow stream (C. Gourley & Jackrabbit Spring outflow, warm water stream conditions E. Ammon 1997, Otis Bay, Incorporated, personal com- extended from the spring-pool downstream 750 m and cool munication). The channel design incorporated MWCV water stream conditions from 1,100 m from the spring-pool and TD used by Ash Meadows pupfish and Ash Meadows to 3,750 m downstream. There were three 10-m-long sta- speckled dace but used only infrequently by Sailfin molly tions spaced 250 m apart in the warm water lotic reach and and Mosquitofish. We compared fish species composition four 10-m-long stations in the cool water stream reach of Kings Pool Spring Marsh with Kings Pool Spring spaced 500–1,500 m apart. Snorkeling was conducted sea- Restored Outflow to gauge the efficacy of habitat manipu- sonally from fall 1990 to summer 1992 and then in winter lation for the control of Sailfin molly and Mosquitofish. and summer of 1993. Kings Pool Spring Marsh was sampled in winter 1989 by Available MWCV and TD were quantified by conduct- placing 12 minnow traps in representative areas. In Kings ing a cross-sectional profile at each station’s downstream Pool Spring Restored Outflow, three 50-m-long stations end, middle, and upstream end. Measurements were taken were snorkeled in winter 2002, 5 years postrestoration. at nine evenly spaced points along the cross section. Out- Stations were at the downstream end, middle, and upper flow from Crystal and Jackrabbit springs varied little dur- end of the restored 1-km-long channel. Chi-square test ing the study, and measurements were taken in winter and was used to determine if there was a significant change in summer 1991 and 1992. proportion of native to non-native fish after the change Velocity and depth measurements were made with from marsh to stream habitat. a Marsh–McBirney model 201D digital flow meter In April 2002 we quantified MWCV and TD used by (Marsh–McBirney, Incorporated, Frederick, MD, U.S.A.) Ash Meadows pupfish, Sailfin molly, and Mosquitofish at mounted on a calibrated rod. One-factor analysis of three 50-m-long stations along the restored stream reach variance (ANOVA) was used to test whether pupfish, to determine if they showed the same habitat use as in speckled dace, molly, and Mosquitofish used significantly Crystal Spring and Jackrabbit Spring systems. Available different velocity and depth in warm water stream habitat. MWCV and TD were quantified by conducting cross- Also tested were species’ MWCV and TD use in relation sectional profiles at three equally spaced transects at the to available MWCV and TD among stations. Significant upper, middle, and lower 10 m of each 50-m-long station. differences in MWCV and TD use were compared using MWCV and TD were measured at nine points equally a modified Bonferroni (a ¼ 0.05) (Keppel 1982). dividing each cross-sectional transect. One-factor ANOVA was used to test whether pupfish used significantly different velocity and depth than Sailfin molly and Mos- Test Shift in Habitat Use quitofish in Kings Pool Restored Outflow. We also tested Crystal Spring outflow (with Sailfin molly) and Fairbanks species MWCV and TD use in relation to available Spring (without Sailfin molly) were used to test the hypo- MWCV and TD. Significant differences in MWCV and TD thesis that presence of Sailfin molly did not cause a shift in use were compared using a modified Bonferroni (a ¼ 0.05) Ash Meadows pupfish MWCV and TD use. This was the (Keppel 1982). only native–non-native combination that presented the opportunity for testing. We used an ANOVA to test the null hypothesis that Ash Meadows pupfish use different MWCV and TD in the presence of Sailfin molly. Although Results Crystal Spring has substantially greater discharge than Fairbanks Spring, stream flow was split among its channels Habitat and Relative Species Composition and we used monitoring stations that had flow comparable Native fishes predominated in warm water spring-pools to Fairbanks and had the same fish complement (except and streams, whereas non-natives predominated in cool for Sailfin molly). We took Ash Meadows pupfish habitat water spring-pools and marshy habitat. An average of use information for Fairbanks outflow in winter 1999 and 80% of the fish in warm water streams and 65% in warm used winter habitat use information for Ash Meadows water spring-pools were native (Fig. 3). The difference pupfish in Crystal Spring warm water stream habitat taken was significant in stream habitat (t6 ¼ 26.44, p ¼ 0.001), in 1990, 1991, and 1992. Test between systems was not for but variance was very high among spring-pools and the

JUNE 2005 Restoration Ecology 251 Habitat Restoration as a Means of Controlling Non-Native Fish

In the cool water segment of this system pupfish and Sail- 100 Native 6 Non-Native fin molly did not use significantly (F[1] ¼ 2.72, p ¼ 0.099) 7 different MWCV (Fig. 4); both used significantly (F[1] ¼ 7 75 28.93, p < 0.001) faster water than the available MWCV 6 and than that used by Mosquitofish, but the absolute differences were relatively small. Speckled dace used sig- 1 50 nificantly (F[1] ¼ 15.27, p < 0.001) faster MWCV than Sailfin molly and pupfish. Pupfish and speckled dace occu-

Frequency (%) pied the deepest water, but the depth was not signiﬁcantly

25 (F[1] ¼ 0.16, p ¼ 0.692) greater than that used by Sailfin molly. Mosquitofish occupied significantly (F[1] ¼ 38.85, p < 0.001) shallower water than the other species.

0 Stream Stream Spring-pool Spring-pool Marsh Warm Cool Warm Cool Cool Test of Microhabitat Shift Habitat Sailfin molly did not cause pupfish to shift MWCV or TD Figure 3. Mean percent and standard deviation of native and use. In Crystal Spring outflow pupfish used an MWCV of non-native fishes in five aquatic habitat categories in Ash Meadows, 21.5 cm/second and mean TD of 35.2 cm compared to an Nevada. Numbers denote each habitat type sample size. MWCV of 21.6 cm/second and mean TD of 32.1 cm for Fairbanks Spring outflow. These differences were not significantly different for either MWCV (F[1] ¼ 0.001, p ¼ difference was not significant (t5 ¼ 21.16, p ¼ 0.299). 0.976) or mean TD (F[1] ¼ 1.20, p ¼ 0.275). There was a significantly greater percentage (92%) of non-natives in cool water spring-pools (t ¼ 10.44, p ¼ 5 Kings Pool Spring Marsh versus 0.001) and cool water marshes (75%; t6 ¼ 6.07, p ¼ 0.001) Kings Pool Spring Restored Outflow than natives. Native and non-native composition was similar for the single cool water stream sampled. After the conversion of Kings Pool Spring outflow from marsh to stream warm water habitat, there was a signifi- 2 cant shift (v 1 ¼ 68.63, p < 0.001) in species composition, MWCV and TD Used and Available from 23 to 91% native fish (Table 1). In addition, most of In warm water, native fishes occupied faster water than the Sailfin molly and Mosquitofish counted were in the non-native fishes (Fig. 4). In Crystal Spring stream, warm downstream station in the restored channel, which was water MWCV used by pupfish (21.7 cm/second) was not only 30 m upstream of marsh-like habitat. The fish found in the two upstream stations were 99.5% pupfish. significantly different (F[1] ¼ 2.62, p ¼ 0.106) from the stream’s available MWCV (23.1 cm/second) but was over In the restored channel, pupfish occupied faster and twice as fast as that used by Sailfin molly and Mosquito- deeper water than did non-natives (Table 2). They oc- fish (<9.0 cm/second), and this difference was significant curred in MWCV of 27 cm/second compared to 18 and 15 cm/second for Sailfin molly and Mosquitofish, respec- (F[1] ¼ 182.80, p < 0.001). Pupfish occupied the deepest tively. MWCV used by pupfish was not significantly (F[1] ¼ water but not significantly (F[1] ¼ 1.97, p ¼ 0.160) deeper than that occupied by Sailfin molly. Pupfish and Sailfin 2.61, p ¼ 0.107) different than the 30 cm/second available but was significantly (F[1] ¼ 6.53, p ¼ 0.011) greater than molly occupied mean TD significantly (F[1] ¼ 13.77, p < 0.001) greater than the mean available TD, whereas Mos- that used by Sailfin molly and Mosquitofish. There was no significant difference (F[1] ¼ 0.13, p ¼ 0.721) in stream quitofish occupied mean TD significantly (F[1] ¼ 51.04, p < 0.001) less than mean available depth. In Jackrabbit Spring mean TD and the mean depth used by Sailfin molly, but stream, pupfish and speckled dace tended to use signifi- pupfish used significantly (F[1] ¼ 37.95, p < 0.001) greater TD than the mean stream depth and that used by Sailfin cantly (F[1] ¼ 24.91, p < 0.001) greater MWCV in warm water than the stream’s MWCV, whereas Sailfin molly molly. Mosquitofish used significantly (F[1] ¼ 10.95, p ¼ 0.001) shallower water than the mean TD used by pupfish and Mosquitofish used MWCV significantly (F[1] ¼ 4.43, p ¼ 0.036) less than the available MWCV. Speckled dace and Sailfin molly. used the fastest MWCV at 30 cm/second, and this was significantly (F[1] ¼ 24.91, p ¼ 0.001) faster than the 20 cm/ second used by pupfish. Jackrabbit Spring pupfish and Discussion speckled dace used significantly (F[1] ¼ 11.18, p ¼ 0.001) Water flow has been employed to promote native over deeper water than Sailfin molly and Mosquitofish, but the non-native fishes in water systems that are both altered difference was small. and highly regulated (Tyus 1992; Marchetti & Moyle 2001; Jackrabbit Spring outflow had the only substantial cool Propst & Gido 2004). The spring systems of Ash Meadows water stream habitat but with a limited range in MWCV. are not highly regulated but are very much altered, and

252 Restoration Ecology JUNE 2005 Habitat Restoration as a Means of Controlling Non-Native Fish

Warm Stream 735 113 60 50 Crystal Spring Crystal Spring 50 792 743 40 735 412 40 30 30 20 20 113 412 10 10

0 0

60 50 Jackrabbit Spring Jackrabbit Spring 50 55 40 138 55 40 50

Total Depth (cm) 89 Velocity (cm/sec) 138 30

Mean Water Column 324 324 30 50 89 20 20 10 10

0 0 Stream Pupfish Molly Mosquito Dace Stream Pupfish Molly Mosquito Dace fish fish

Cool Stream

60 50 Jackrabbit Spring Jackrabbit Spring 50 40 223 81 267 98 40 81 756 30 30 756 223 267 98 20 20

10 10 Total Depth (cm) Velocity (cm/sec)

Mean Water Column 0 0 Stream Pupfish Molly Mosquito Dace Stream Pupfish Molly Mosquito Dace fish fish

Figure 4. MWCV and TD used by Ash Meadows fishes in warm and cool water lotic habitats in Crystal Spring and Jackrabbit Spring outflows. Numbers denote sample size, bars the mean use, and lines the standard deviation. there is thus ample opportunity for near-complete restora- That native fishes did not predominate in cool water tion of physical habitat with its ensuing benefits to native lotic habitat does not imply that they are not suited to, or fishes. Our study suggests that a channel configuration that are competitively excluded from, relatively cool flowing retains the spring’s high stable temperature at a mean flow water. Speckled dace flourish in the cool water stream velocity of about 30 cm/second favors native Ash Mead- conditions of the upper Amargosa River (Soltz & Naiman ows fishes, especially Ash Meadows pupfish. Like Sailfin 1978). Ash Meadows speckled dace appear to require molly, many of the successful invaders of thermal springs cooler water. They have been found to reproduce in water of the American Southwest are tropical aquarium fish that temperatures ranging from 17.5 to 24.0°C (Scoppettone, prefer lentic conditions (Courtenay et al. 1984). Although they can endure the chronically warm, high-salinity water, Table 2. MWCV and TD of Kings Pool Restored Outflow (KPRO), they tend to avoid the fast water used by Ash Meadows and MWCV and TD used by Ash Meadows pupfish, Sailfin molly, pupfish. and Mosquitofish.

Mean ± SD Table 1. Fish species by percent composition in Kings Pool outﬂow before and after conversion from marsh to stream habitat. MWCV (cm/sec) TD (cm) n

Native Non-Native KPRO 30.3 ± 17.6 a 21.9 ± 6.9 a 216 Pupfish 26.9 ± 8.3 a 28.1 ± 5.3 b 63 Pupfish Molly Mosquitofish Molly 18.5 ± 4.9 b 21.4 ± 8.8 a 29 Mosquitofish 14.7 ± 4.5 b 17.2 ± 2.8 c 31 Marsh before restoration 23% (54) 61% (144) 16% (37) Stream after restoration 91% (814) 8% (71) 1% (9) Species and Kings Pool Restored Outflow followed by the same letter do not differ significantly using one-factor ANOVA and modified Bonferroni (a ¼ Numbers in parentheses represent the number observed. 0.05).

JUNE 2005 Restoration Ecology 253 Habitat Restoration as a Means of Controlling Non-Native Fish unpublished data), and two of the springs (Bradford 1 In the design and construction of the restored outflow and Bradford 2) in which they occur are cool water. channel, emphasis was placed on creating salutary TD and Speckled dace are known to persist in a broad array of MWCV condition for natives; Kings Pool Spring Restored habitats and has been described as a bottom browser on Outflow had one of the highest ratios of native to non- invertebrates (Moyle 2002). In lotic habitat in Ash Mead- native fishes. We anticipate that this ratio will increase ows, they have been associated with faster water feeding further once the channel is completed by eliminating the on invertebrate drift (Scoppettone, unpublished data). source of non-native fishes from the marshy habitat down- The loss of cool water lotic habitat associated with habitat stream. When the outflow channel of Kings Pool Spring alteration may have contributed to the loss of speckled Restored Outflow is completed, its length will extend to dace in spring systems throughout Ash Meadows. about 11 km and its lower reaches will be a cool water We focused on MWCV in stream restoration for stream. Speckled dace will then be reintroduced into the controlling non-native fishes. However, enhancement of system, and we can better define their habitat preference. natives in streams influences adjacent spring-pools by sup- Without a proactive restoration program, return of an plying a source of native fishes rather than non-natives; aquatic ecosystem to its historic equilibrium may take cen- pupfish especially are known to expand their range (Baugh turies (Poff et al. 1997) if it returns at all. In the interim, et al. 1986). Although pupfish tended to be the predomi- there could be shifts in habitat features such as the for- nant species in warm water spring-pools, their abundance mation of headwater marshes as occurred in Kings Pool was extremely low in cool water spring-pools and marshes, Spring Concrete Outflow with habitat favoring non-native even though Amargosa pupfish are known to be tolerant (poecilids) over native fishes. Headwater marshes pose an of temperature extremes (Gerking et al. 1979; Gerking & additional threat by providing excellent habitat for Large- Lee 1983). In fact, pupfish occurred in only one of six cool mouth bass and other predators. Restored lotic habitat water spring-pools and composed only 11% of the popula- increases the likelihood of native fishes predominating in tion in the remaining spring (Forest Spring). All six sys- Ash Meadows. The lotic habitat can be designed to favor tems had been disconnected from surrounding habitats the natives, which should further reduce the relative num- during the period of agricultural development, and this ber of non-natives. Furthermore, connecting the stream isolation probably led to pupfish extirpation. Although channel with historic washes and their associated flash flood Amargosa pupfish can feed and survive in these cool events may further serve to promote native fishes over non- springs, they have only been found to reproduce in water natives (Meffe 1984; Minckley & Meffe 1987). Restoration temperatures from 25 to 31°C (Gerkin & Lee 1983). The of preagricultural aquatic habitat is a management objec- chronically cool spring water (18–21°C) may not allow tive of AMNWR, and this study has shown that restoration reproduction, leading to pupfish extirpation. Forest Spring can be used to reduce populations of non-native fishes. now has pupfish, but they colonized after the partial restoration of Kings Pool Spring outflow reconnected to Forest Spring. There were no pupfish in this system in 1995 when Acknowledgments it was isolated from warm water spring systems. Similarly, Numerous individuals contributed to this study. Sean lack of connectivity caused the extirpation of Preston Shea, James Harvey, Stephanie Byers Mark Buettner, White River springfish (Crenichthys baileyi albivallis) from Peter Tuttle, James Call, Linda Hallock, and James Lund Town Spring, a spring cooler than four others har- Heinrich contributed in determining fish population den- boring the species (Scoppettone & Rissler 2002). Thus, sity and habitat use. Sean Shea and Scott Cecchi assisted connectivity is another important element to evaluate for with graphics. The study was funded by Nevada Division habitat restoration to promote native species. of Wildlife, U.S. Fish and Wildlife Service, and the Biolog- Non-native fishes may cause a shift in native fish habitat ical Resources Division of the U.S. Geological Survey. use (Brown & Moyle 1991; Douglas et al. 1994), thus We also thank Tom Strekal, Jim Deacon, Jerry Smith, obscuring the actual preference of natives. Sailfin molly Kristin Swaim, and Mark Fabes for reviewing this manu- did not cause a habitat shift in Ash Meadows pupfish in script and for making useful suggestions. warm water lotic habitat. Mosquitofish is a sufficiently specialized surface dweller that it overlaps little with ben- thic natives and, consequently, is unlikely to cause a habitat shift. Therefore, the habitat in Kings Pool Spring LITERATURE CITED Restored Outflow controls non-native Sailfin molly and Bain, M. B., J. T. Finn, and H. E. Booke. 1988. Stream regulation and fish Mosquitofish without compromising the quality of adult community structure. Ecology 69:382–392. pupfish habitat. The influence of Sailfin molly and Mos- Baltz, D. M., and P. B. Moyle. 1993. Invasion resistance to introduced quitofish on Ash Meadows speckled dace habitat shift species by a native assemblage of California stream fishes. Ecologi- cal Applications 3:246–255. requires further study. Baugh, T., J. E. Williams, D. A. Buck, and J. E. Deacon. 1986. New distri- The transition of Kings Pool outflow from a marsh to bution records for Cyprinodon nevadensis mionectes, an endangered warm water stream demonstrated the potential impor- pupfish from Ash Meadows, Nevada. The Southwestern Naturalist tance of lotic habitat for control of some non-native fish. 31:544–546.

254 Restoration Ecology JUNE 2005 Habitat Restoration as a Means of Controlling Non-Native Fish

Brown, L. R., and P. B. Moyle. 1991. Changes in habitat and microhabitat Miller, R. R. 1961. Man and the changing fish fauna of the American with an assemblage of stream fishes in response to predation by southwest. Papers of the Michigan Academy of Science, Arts and Sacramento squawfish (Ptychocheilus grandis). Canadian Journal Letters 46:365–404. of Fisheries and Aquatic Sciences 48:849–856. Miller, R. R., J. D. Williams, and J. E. Williams. 1989. Extinction of North Courtenay, W. R. Jr, D. A. Hensley, J. N. Taylor, and J. A. McCann. American fishes during the past century. Fisheries (Bethesda) 1984. Distribution of exotic fishes in the Continental United States. 14:22–38. Pages 41–77 in W. R. Courtenay Jr and J. R. Stauffer Jr, editors. Minckley, W. L., and J. E. Deacon. 1991. Battle against extinction: native Distribution, biology, and management of exotic fishes. The Johns fish management in the American West. The University of Arizona Hopkins University Press, Baltimore, Maryland. Press, Tucson. Deacon, J. E., and S. Bunnell. 1970. Man and pupfish, a process of Minckley, W. L., and G. K. Meffe. 1987. Differential selection by flooding destruction. Cry California 5:14–21. in stream fish communities of the arid American Southwest. Pages Deacon, J. E., and C. D. Williams. 1991. Ash Meadows and the legacy of 96–104 in W. I. Matthews, and D. C. Heins, editors. Community and the Devils Hole Pupfish. Pages 69–87 in W. L. Minckley and J. E. evolutionary ecology of North American stream fishes. University Deacon, editors. Battle against extinction: native fish management of Oklahoma Press, Norman. in the American West. The University of Arizona Press, Tucson. Morrison, B. R. S. 1987. Use and effects of piscicides. Pages 47–52 in P. S. Douglas, M. E., P. C. Marsh, and W. L. Minckley. 1994. Indigenous fishes of Maitlands, and A. K. Turner, editors. Angling and Wildlife in Fresh Western North America and the hypothesis of competitive displace- Waters. ITE Symposium No. 19. Institute of Terrestial Ecology, ment: Meda fulgida (Cyprinidae) as a case study. Copeia 1994:9–19. Cumbria, United Kingdom. Dudley, W. W. Jr, and J. D. Larson. 1976. Effect of irrigation pumping on Moyle, P. B. 2002. Inland fishes of California. University of California desert pupfish habitat in Ash Meadows, Nye County, Nevada. U.S. Press, Berkeley. Geological Survey Professional Paper 927. Moyle, P. B., H. W. Li, and B. A. Barton. 1986. The Frankenstein effect: Garside, L. I., and J. H. Schilling. 1979. Thermal waters of Nevada. Nevada impact of introduced fishes on native fishes in North America. Pages Bureau of Mines and Geology, Bulletin 91. Mackay School of Mines, 415–426 in R. H. Stroud, editor. Fish culture in fisheries man- University of Nevada, Reno. agement. American Fisheries Society, Fish Culture Section and Gerking, S. N., and R. M. Lee. 1983. Thermal limits for growth and repro- Fisheries Management Section, Bethesda, Maryland. duction in the desert pupfish. Cyprinodon n. nevadensis. Physiologi- Moyle, P. B., and T. Light. 1996. Fish invasions in California: do abiotic cal Zoology 56:1–9. factors determine success? Ecology 77:1666–1670. Gerking, S. N., R. M. Lee, and J. B. Shrode. 1979. Effects of generation- Moyle, P. B., and R. Nichols. 1974. Decline of the native fish fauna of the long temperature acclimation on reproductive performance of the Sierra-Nevada foothills, central California. American Midland desert pupfish, Cyprinodon n. nevadensis. Physiological Zoology Naturalist 92:72–83. 52:113–121. Pister, E. P. 1974. Desert fishes and their habitats. Transactions of the Gido, K. B., and D. L. Propst. 1999. Habitat use and association of native American Fisheries Society 103:531–540. and non-native fishes in the San Juan River, New Mexico and Utah. Poff, N. L., J. D. Allan, M. B. Bain, J. R. Karr, K. L. Prestegarrd, B. D. Copeia 1999:321–332. Richter, R. E. Sparks, and J. C. Stromberg. 1997. The Natural Flow Harrington, R. W. Jr, and E. S. Harrington. 1961. Food selection among Regime; a paradigm for river conservation and restoration. Bio- fishes invading a high subtropical salt marsh: from onset of flooding Science 47:769–784. through the progress of a mosquito brood. Ecology 42:649–665. Propst, D. L., and K. B. Gido. 2004. Response of native and non-native Hubbs, C. L., and R. R. Miller. 1948. The zoological evidence: correlation fishes to natural flow regime mimicry in the San Juan River. Trans- between fish distribution and hydrographic history in the desert basins actions of the American Fisheries Society 133:922–931. of western United States. Pages 17–166 in The Great Basin with Rabeni, C. F., and R. B. Jacobson. 1993. The importance of fluvial emphasis on glacial and postglacial times. Bulletin of the University of hydraulics to fish-habitat restoration in low-gradient alluvial Utah 38, Biological Series 10. The University of Utah, Salt Lake City. streams. Freshwater Biology 29:211–220. Keppel, G. 1982. Design and analysis: a researcher’s handbook. 2nd Ricklefs, R. E. 1991. Structures and transformation of life histories. Func- edition. Prentice-Hall, Englewood Cliffs, New Jersey. tional Ecology 5:174–183. Knapp, R. A., and K. R. Matthews. 1998. Eradication of non-native fish Rinne, J. N., and P. R. Turner. 1991. Reclamation and alteration as man- by gill netting from a small mountain lake in California. Restoration agement techniques, and a review of methodology in stream renova- Ecology 6:207–213. tion. Pages 219–244 in W. L. Minckley and J. E. Deacon, editors. Mangum, F. A., and J. L. Madrigal. 1999. Rotenone effects on aquatic Battle against extinction: native fish management in the American macroinvertebrates of the Strawberry River, Utah: a five-year West. University of Arizona Press, Tucson. summary. Journal of Freshwater Ecology 14:125–135. Sada, D. W. 1990. Recovery plan for the endangered and threatened spe- Marchetti, M. P., and P. B. Moyle. 2001. Effects of flow regime on fish cies of Ash Meadows, Nevada. U.S. Fish and Wildlife Service, Reno, assemblages in a regulated California stream. Ecological Applica- Nevada. tions 11:530–539. Scoppettone, G. G., and P. H. Rissler. 2002. Status of the Preston White Meffe, G. K. 1983. Attempted chemical renovation of an Arizona spring- River springfish (Crenichthys baileyi albivallis). Western North brook for management of the endangered Sonoran Topminnow. American Naturalist 62:82–87. North American Journal of Fisheries Management 3:315–321. Sokal, R., and F. J. Rohlf. 1995. Biometry. 3rd edition. Freeman, New York. Meffe, G. K. 1984. Effects of abiotic disturbance on coexistence of Soltz, D. L., and R. J. Naiman. 1978. The natural history of native fishes predator–prey fish species. Ecology 65:1525–1534. in the Death Valley system. Natural History Museum of Los Meronek, T. G., P. M. Bouchard, E. R. Buckner, T. M. Burri, K. K. Angeles County, California, Science Series 30:1–76. Demmerly, D. C. Hatheli, R. A. Klumb, S. H. Schmidt, and D. W. Southwood, T. R. E. 1988. Tactics, strategies, and templets. Oikos 52: Coble. 1996. A review of fish control projects. North American 3–18. Journal of Fisheries Management 16:63–74. Swanson, C., J. J. Cech Jr, and R. H. Piedrahita. 1996. Mosquitofish– Miller, R. R. 1948. The cyprinodont fishes of the Death Valley system of biology, culture and use in mosquito control. Mosquito and Vector eastern California and southeastern Nevada. Miscellaneous publica- Control Association of California and The University of California tions of the Museum of Zoology, University of Michigan 68:1–155. Mosquito Research Program.

JUNE 2005 Restoration Ecology 255 Habitat Restoration as a Means of Controlling Non-Native Fish

Taylor, J. N., W. R. Courtenay Jr, and J. A. McCann. 1984. Known im- U.S. Fish and Wildlife Service. 1983. Endangered and threatened wildlife and pacts of exotic fishes in the continental United States. Pages 322–373 plants; determination of endangered status and critical habitats for two in W. R. Courtenay Jr and J. R. Stauffer Jr, editors. Distribution, fish species in Ash Meadows, Nevada. Federal Register 48:40178–40186. biology, and management of exotic fishes. Johns Hopkins University Williams, J. E., D. B. Bowman, J. E. Brooks, A. A. Echelle, R. J. Edwards, Press, Baltimore, Maryland. D. A. Hendrickson, and J. J. Landye. 1985. Endangered aquatic eco- Tyus, H. M. 1992. An instream flow philosophy for recovering endan- systems in North America deserts with a list of vanishing fishes of the gered Colorado River fishes. Rivers 3:27–36. region. Journal of the Arizona–Nevada Academy of Science 20:1–62.

256 Restoration Ecology JUNE 2005