The Combined Effects of Multiple Invasive Species on Persistence of Imperiled Pahrump Poolfish

Abstract

Many ecosystems have been invaded by more than one non-native species, but research evaluating the combined effects of multiple invasive species has been limited. In the southwest USA, many aquatic systems have been invaded by multiple species such as non-native crayfish and non-native fishes. We document the rapid decline of a population of the endangered Pahrump Poolfish, Empetrichthys latos, which occurred following the sequential introduction of Red Swamp Crayfish, Procambarus clarkii, and Western Mosquitofish, Gambusia affinis. We also report the results of mesocosm experiment where we tested individual and combined effects of invasive Red Swamp Crayfish and Western Mosquitofish on experimental populations of Pahrump Poolfish. Survival rates were near 100% for adult poolfish in allopatry but were significantly lower for the other two treatments; when poolfish were sympatric with crayfish (~53%), and when poolfish were sympatric with both crayfish and mosquitofish (~55%). In allopatry, poolfish produced over 90 juveniles per mesocosm, while approximately 65 juveniles per mesocosm when poolfish were sympatric with crayfish, but this difference was not significant. However, juvenile production plummeted to near zero when poolfish were sympatric with both crayfish and mosquitofish. This study demonstrates that Pahrump Poolfish must be actively managed to detect and control invasive species, otherwise extinction will likely occur. This study also provides an example of the compounding effects of multiple invasive species.

1. Introduction

The introduction of non-native species has become so widespread that many systems have been invaded by multiple non-native species [1,2,3,4]. Once established, invasive species can facilitate the colonization of additional non-native species, an ecological process referred to invasional meltdown [5,6,7]. Thus, work evaluating the combined effects of multiple species introductions on native species is a topic of concern for conservation biologists [8,9].

Some studies have shown that multiple invasive species can have substantial impacts on endemic species [10]. For example, ballast water exchange introduced Eurasian Zebra Mussel, Dreissena polymorpha, and the Round Goby, Neogobius melanostomus [11,12]. Zebra Mussels altered the planktonic community structure that allowed the Round Goby to obtain competitive superiority over endemic species such as the Mottled Sculpin, Cottus bairdii [12].

Emergent effects of multiple predators can reduce risk due to interactions among predators or increase risk if prey responses to one predator increases risk to another predator [1]. For example, Palacios et al. [13] discussed the potential implications of introducing novel predators with a piscivorous prey species and found that predator identity determined if there were any positive or negative interactions on the multiple predator effect on prey species persistence. By contrast, Porter-Whitaker et al. [3] reported that prey responses to multiple predators were intermediate to the sole effects of each predator.

A wide variety of invasive predators have been directly associated with the extirpation of numerous fish populations in the southwestern United States [14,15,16,17,18]. These reports, while widespread, are largely correlative, with the decline of native species co-occurring with the invasion of non-native species [14,15,16,17]. These impacts have been attributed to the predator naïveté of endemic fishes, which evolved in simple communities, often as the only fish species in that particular habitat. Specifically, endemic fishes are hypothesized to have lost anti-predator traits as they evolved in species-poor systems with limited predation, thus making them vulnerable to invasive predators [14,15,16,19,20].

Non-native predators include both crayfish and small-bodied fish such as the Western Mosquitofish, Gambusia affinis. Crayfish prey on the adults and larvae of benthic fishes [21,22], while mosquitofish are voracious predators of fish eggs and larvae [23,24,25,26,27,28]. The direct effects of both Red Swamp Crayfish and Western Mosquitofish have been independently evaluated amongst numerous experimental studies [21,22,23,25,26,27,28], however, the combined effects of crayfish and mosquitofish have not yet been studied empirically.

Understanding the solitary and combined effects of both Western Mosquitofish and Red Swamp Crayfish is critical for resource management because both of these non-native predators are listed as threats to various endemic fishes [16,17,18,29]. Invasive species have played a significant role in the decimation of the poolfishes of southern Nevada. In fact, Miller et al. [17] attributed extinction of the Ash Meadows Poolfish, E. merriami, to crayfish, while Minckley and Deacon [16] inferred that the co-invasion of both crayfish and mosquitofish caused the extinction of the Ash Meadows Poolfish.

Here, we report a case study where the largest population of endangered Pahrump Poolfish, E. latos, (Lake Harriet) rapidly declined following the sequential colonization of the lake by Red Swamp Crayfish and Western Mosquitofish. Further, we report the results of a mesocosm experiment designed to test the synergistic effects of dual species invasion on experimental populations of Pahrump Poolfish.

Background

The poolfishes, along with various species of springfish, Crenichthys sp., evolved as the only oviparous genera within the Goodeidae family of livebearing fishes [30]. While the livebearing species of this family occur in central Mexico, the poolfish and springfish evolved in Nevada. The springfishes include two species and numerous sub-species that occur in Eastern Nevada, while the poolfishes historically occurred in two watersheds in southern Nevada. As previously stated, the Ash Meadows Poolfish went extinct in the 1950s [17]. The other species of poolfish, E. latos, was comprised of three subspecies that were isolated in three distinct springs following the desiccation of Pleistocene Lake Pahrump [31,32]. Two of these subspecies went extinct by 1960 due to spring failure associated with groundwater pumping [17], leaving the Pahrump Poolfish as the only extant member of its species and of its genus. The Pahrump Poolfish evolved in Manse Spring, Nye County, Nevada, and was among the first fishes listed as endangered in 1967 [33].

Habitat modifications in the late 1960s including the establishment of goldish, Carassius auratus, led managers to translocate Pahrump Poolfish as a hedge against extinction [30,34]. Originally, 76 Pahrump Poolfish from Manse Spring were introduced among three refuge habitats in Nevada between 1970 and 1972. Manse Spring failed in 1975 [34], and thus Pahrump Poolfish have been managed as an artificial meta-population with a number of introduction, extirpation and re-colonization events among refuge habitats. As of 2025, four refuge populations occur in Nevada at the following sites: (1) Corn Creek on the Desert Game Refuge (36°26′19.55″ N; 115°21′32.37″ W) approximately 35 km northwest of Las Vegas; (2) Shoshone Ponds (38°56′9.59″ N; 114°25′8.32″ W), approximately 51 km southeast of Ely Nevada; (3) Springs Preserve, Las Vegas; and (4) Lake Harriet at Spring Mountain Ranch State Park (36° 4′4.33″ N; 115°27′41.74″ W), approximately 31 km west of Las Vegas. This paper focuses on the Lake Harriet population at Spring Mountain State Park which was established in 1983 [30].

Poolfish are a small-bodied fish (up to 79 mm standard length) but are relatively long lived. Although sexually monomorphic in body shape, poolfish are sexually dimorphic in size by the end of their first year [35]. Poolfish males can reach 44 mm standard length and live up to 7 years, while poolfish females can reach nearly 80 mm standard length and live up to 10 years [35,36]. Pahrump Poolfish females lay eggs singly, but the mating system of poolfish has not been described. Previous work has shown that experimental populations of poolfish can be maintained with high survivorship and high juvenile production when poolfish are maintained in allopatry but perform poorly in the presence of western mosquitofish [36,37].

2. Materials and Methods

2.1. Lake Harriet Population Estimates

Lake Harriet Refuge population of Pahrump Poolfish was established in 1983, and Mark-recapture sampling has been conducted on an annual basis since 1997. The mark-recapture procedures evolved over time, but we report the entire data set as these data were used by managers to assess the status of this population and to detect if non-native species were introduced. Mark-recapture sampling typically consisted of 1–2 marking sessions and one final recapture session. During the first four years (1997–2000), 24 to 25 traps were used, with most traps deployed along the shoreline (Table S1). From 2001–2010, trapping effort included about 35–50 traps and was standardized starting in 2011 as follows. A total of 53 minnow traps were used to capture poolfish (Table S1). Fifty G-1 minnow traps with an opening of 2.54 cm and a wire mesh of 0.64 cm were set approximately every 10 m along the perimeter of the Lake Harriet shore and allowed to fish for 3 h (Figure 1). Another 10 traps were set around the perimeter of an island, and these traps were set for 5 h. Further, three fine mesh traps were set with an opening of 2.54 cm and a wire mesh of 0.32 cm were set for 5 h (Figure 1). These latter traps could sample smaller bodied fish, thus only fish that were at least 30 mm in total length were marked. Fish were marked by removing a small portion of the caudal fish and then released. A second sampling session was typically conducted two to 27 days later (Table S1). The same trapping protocol was used and the number of marked and un-marked fish recorded. The Lincoln–Peterson index was used to estimate population size along with 95% confidence intervals [38].

2.2. Mesocosm Experiment

Poolfish used in this experiment included a mixture of wild poolfish and lab-reared poolfish. The wild poolfish were collected on 13 June 2017 from Shoshone Stock Pond (White Pine County, NV, USA) while the lab-reared poolfish were descendants from poolfish originally collected in 2014 from Spring Mountain Ranch State Park, Clark County [28]. Western Mosquitofish were obtained from Sutter-Yuba Mosquito and Vector Control district in Yuba City, CA. Red Swamp Crayfish were sourced from Carolina Biological Supply^® Burlington, NC.

Three experimental communities were used in this experiment, forming a single block within a randomized block design. Each of the seven blocks contained a total of three mesocosms including the following treatments: (I.) allopatric poolfish, (II.) poolfish sympatric with crayfish, and (III.) poolfish sympatric with both mosquitofish and crayfish (

Table 1. Mesocosm treatments and their respective species composition.

Mesocosm Treatment 1	Species Composition in Each Mesocosm
Allopatric Poolfish	7 adult poolfish 2
Poolfish and Crayfish	7 adult poolfish 2 4 adult crayfish 3
Poolfish, Crayfish, and Mosquitofish	7 adult poolfish 2 4 adult crayfish 3 7 adult mosquitofish 5 females & 2 males

¹ Seven replicates for each mesocosm treatment. ² Pahrump Poolfish are sexually monomorphic. ³ Dead crayfish were replaced.

). We did not include a poolfish + mosquitofish treatment because of the limited number of poolfish available for this experiment. Further, two previous experiments consistently showed that mosquitofish effectively eliminated juvenile poolfish [28,37].

Each block of three tanks was replicated seven times for a total of 21 experimental tanks, arranged in a linear sequence. All 21 tanks received seven adult poolfish of indeterminate sex and of indeterminate population of origin (Shoshone Stock Pond 2017 or Spring Mountain Ranch 2014) (Table 1). Four individual crayfish were introduced into two randomly selected mesocosms per block (Table 1). One of the two crayfish mesocosms within each block was randomly selected to receive mosquitofish, including five gravid females and two males (Table 1). Crayfish density was maintained by replacing any crayfish that died. Crayfish deaths were likely due to aggression among conspecifics as inferred from missing appendages on dead crayfish, and by active feeding by live crayfish on dead conspecifics (Table 1).

All mesocosms were provided with reclaimed PVC vinyl Fishiding^® structures to simulate aquatic plants and to provide spatial structure along with ~57 L of river rock. Supplemental food was provided by hand every day to each tank at rates of ~2–3% of total fish biomass, while crayfish were fed twice weekly at a rate of ~5% of total crayfish biomass. Food consisted of Tetra tropical flake^® for fish, and Aquatic Arts (Fish, Inverts, and Aquatic Plant) sinking pellets^® for the crayfish. Water quality was assessed weekly for ammonia and nitrates. All tanks were checked daily for mortalities, and to ensure air flow was constant from air stones. After ten weeks, the tanks were drained, and the poolfish adults and juveniles counted. The number of adults provided a measure of adult survival, while the number of juveniles provided a measure of juvenile production. Data were analyzed using JMP Pro 17^® software. ANOVAs were used to test for treatment effects on (a) adult survival, (b) juvenile production, and (c) the number of juveniles produced per surviving adult poolfish. We then used a Tukey HSD to test for pairwise differences among the treatments. Experiment-wise, alpha was maintained at 0.05. For each test, the effect of block was not significant and was eliminated from the final model. Non-parametric analyses produced the same significance levels among treatments for all tests, but here we report the parametric ANOVA results.

3. Results

3.1. Lake Harriet Population Estimates

The Pahrump Poolfish population size has varied considerably from year to year but typically was over 8000 fish (Figure 2). There were three notable drops in population size in 2001, 2008, and 2016. In 2001, population sampling was conducted relatively early, taking place entirely in June, when adult poolfish are less active and thus less likely to be trapped (personal observation). From this point on, sampling was typically conducted in July and/or August (Table S1). The population decline in 2008 occurred when the water level of Lake Harriet had been lowered for construction [39]. It is notable that for both of these cases, the population rapidly recovered the following year (Figure 2).

The third decline followed the sequential introduction of Red Swamp Crayfish and Western Mosquitofish. Red Swamp Crayfish were first detected in 2012, when the poolfish population was estimated at 31,570 fish (25,738–37,862; 95% confidence interval). The following year, the poolfish population declined by about 50% (16,813 fish; 15,118–18,508 95% CI). The poolfish population size in 2014 and 2015 was about 30% lower at about 12,000 fish (Figure 2). Western mosquitofish were detected in 2015 (Figure 2), when the poolfish population was estimated at 12,286 poolfish (10,791–13,988, 95% confidence interval). However, within one year only 104 fish were trapped across three sampling sessions, leading managers to abandon mark-recapture efforts and salvage the entire population. Over the following year, 688 poolfish were captured during 14 total trapping events and relocated to a fish hatchery (644 fish) and the Corn Creek Refuge (44 fish). Lake Harriet was desiccated by redirecting the flow away from the lake starting in 2016. All non-native fishes were eliminated, and, as of 2 May 2018, no crayfish or active crayfish boroughs were observed in Lake Harriet. Lake Harriet was refilled, and Poolfish were reintroduced in 2020.

3.2. Mesocosm Experiment

Adult survival (percentage) significantly differed among the three treatments (F_2,18 = 16.65, p < 0.001; Figure 3). In allopatry, adult poolfish survival rates were near 100% (95.9 ± 2.6%; mean ± one standard error of the mean) and significantly higher compared to adult poolfish survival when sympatric with crayfish (53.1 ± 6.0%; t = 5.11, p < 0.001) and when poolfish were sympatric with both crayfish and mosquitofish (55.1 ± 7.9%; t = 4.87, p < 0.001; Figure 3). The latter two treatments did not significantly differ from each other (t = 0.24, p = 0.969).

Poolfish juvenile production differed significantly among the three treatments (F_2,18 = 12.56, p < 0.001; Figure 4). In allopatry, poolfish produced over 90 juveniles per mesocosm (91.4 ± 12.0; juveniles per tank), which was 41% higher than for mesocosms hosting poolfish and crayfish (64.9 ± 19.0), but this difference was not significant (t = 1.45; p = 0.339). However, juvenile production, which plummeted to near zero (1.9 ± 0.5) when poolfish were sympatric with both crayfish and mosquitofish, was significantly different from allopatric poolfish (t = 4.88, p < 0.001) and when compared to the poolfish sympatric with only crayfish (t = 3.43, p = 0.008; Figure 4).

Due to the high variation in adult survival among treatments, we also examined the number of juveniles per surviving adult. After adjusting for the number of surviving adults, relative juvenile production was significantly different among the three treatments (F_2,18 = 11.94; p < 0.001; Figure 5). Relative juvenile production for mesocosms with allopatric poolfish (13.6 ± 1.7 juveniles/surviving adult) did not differ from mesocosms with poolfish sympatric with crayfish (17.5 ± 4.1 juveniles/surviving adult; t = 1.07, p = 0.545). However, relative juvenile production was significantly higher for both of these treatments compared to Poolfish sympatric with both crayfish and mosquitofish (0.5 ± 0.1 juvenile/surviving adult; t = 3.60,, p < 0.01 and t = 4.66, p < 0.001, respectively).

4. Discussion

The Pahrump Poolfish has been managed by the establishment and maintenance of many refuge populations since the 1970s [28,30,34]. The use of numerous refuge populations has been motivated in part by the potential risk of un-regulated introduction of non-native species [40]. This risk averse strategy has proven a good investment due to the periodic loss of refuge populations that coincided with the introduction of non-native species [30].

Here, we document sequential introduction of Red Swamp Crayfish and Western Mosquitofish and subsequent population crash of the Lake Harriet poolfish population that occurred 3 and 1 year(s) later, respectively. The poolfish population estimates varied widely among years, but the population estimates were typically above 8000 fish. On two occasions the population declined to approximately 3500 fish, but in both cases the population increased 7-fold to 16-fold the following year. The first decline was observed when the population was sampled early in the year when adult poolfish are less active and less likely to be trapped (personal observation). Subsequent sampling sessions were conducted later in the year to manage this sampling limitation. The second population decline happened after the water level at Lake Harriet was drawn down for maintenance work during February 2008 [39].

Following the discovery of Red Swamp Crayfish in 2012, the poolfish population declined from about 31,500 fish by about 50% in 2013, with an additional reduction of 30% the following years to about 12,000 fish (Figure 2). The poolfish population was estimated at over 12,000 fish in 2015, the year that Western mosquitofish were detected. However, the following year, the poolfish population experienced a severe decline, with only about 100 fish captured over three trapping sessions. This decline, which was much more severe than the two previous population declines, motivated managers to salvage this population by transplanting the remaining fish to a fish hatchery (644 fish) and the Corn Creek refuge (44 fish).

The field data inspired us to conduct a mesocosm experiment to better understand the impacts of crayfish and mosquitofish on experimental poolfish populations. Two earlier experiments demonstrated that poolfish juvenile production was eliminated when mosquitofish were present [28,37]. However, the additional effects of crayfish were unknown. In the mesocosm study, crayfish caused an approximately 40% increase in adult poolfish mortality, but the addition of mosquitofish did not have any additional effects on poolfish adult mortality. Such impacts on adult survival would be expected to decrease juvenile productivity. In fact, there was a notable, but non-significant, reduction in poolfish juvenile production for the poolfish and crayfish mesocosms. It is notable, however, that relative juvenile production (juveniles/surviving adult poolfish) was not significantly impacted by the sole presence of crayfish.

The combined effects of both crayfish and mosquitofish effectively eliminated the production of poolfish juveniles. It is noteworthy that similar work has shown mosquitofish to virtually eliminate poolfish juvenile production [28,37]. Thus, this experiment did not reveal emergent effects [1] of the combination of these two predators.

Our findings suggest that the sole introduction of crayfish may have notable impacts on the survival of poolfish adults, and these, in turn, reduce the number of juveniles produced. However, poolfish mesocosm populations grew by nearly 10-fold when in the presence of crayfish. Thus, crayfish are unlikely to have immediate acute impacts. Indeed, the Corn Creek refuge population of poolfish co-persisted with Red Swamp Crayfish for five years. However, this population collapsed shortly after non-native fish were detected (K.G. unpublished data). Furthermore, while crayfish appeared to reduce the population size of Lake Harriet Pahrump Poolfish, the population had a much more severe decline within one year of the introduction of mosquitofish.

We recognize that mesocosms have limitations in mimicking the complexity of natural systems [41,42]. However, understanding ecological interactions involving protected species is typically challenging. Fortunately, for fish species that are locally abundant, ex-situ experiments are feasible [22,26,28], and such findings can be useful for guiding management decisions. For instance, a mesocosm experiment demonstrated that endangered Mohave tui chub, Siphateles bicolor mohavensis, could co-persist with western mosquitofish, and this finding led managers to establish a refuge population of Mohave tui chub in a habitat occupied by mosquitofish [26]. In this current case, the negative impacts of non-native predators are consistent with the observations from the wild population at Lake Harriet.

Overall, this study, combined with previous mesocosm experiments [28,37], demonstrates that non-native predators can have negative effects on Pahrump Poolfish populations. The vulnerability of Pahrump Poolfish to non-native predators is consistent with the predator naiveté hypothesis [20]. In contrast to other small-bodied fishes, poolfish do not respond to conspecific alarm cues, suggesting limited anti-predator competence [20]. Thus, the current approach of managing Pahrump Poolfish in single species refugia is clearly warranted. Poolfish may be able to co-persist with invasive crayfish, but immediate intervention should be taken if Western Mosquitofish invade any of the poolfish refuge habitats. Our study shows the value of evaluating the effects of multiple invasive species on native species, but additional work should be undertaken to evaluate other combinations of invasive species.