Reversal of competitive dominance between invasive and native freshwater crayfish species under near-future elevated water temperature

Biological invasions are a major cause of biodiversity loss and, coupled with climate change, will likely have detrimental impacts for native species and the functioning of ecosystems. To mitigate such impacts, it is important to elucidate the behavioural mechanisms underpinning interactions between invasive and native species. Here we examined how competitive interactions between invasive and native species are modified under conditions of near-future elevated water temperature using freshwater crayfish as a model system. Contest experiments between the native Euastacus spinifer and invasive Cherax destructor revealed that the competitive advantage of E. spinifer at current maximum temperatures (22 °C) was reversed at elevated near-future temperatures (26 °C), after controlling for relative body size. In addition, the native crayfish spent twice as long motionless at 26 °C than C. destructor, consistent with physiological challenges underpinning this competitive reversal. Most alarmingly, E. spinifer experienced significant mortality after fighting C. destructor, particularly at 26 °C. Mortality usually ensued two days post-contest even when E. spinifer had won. Mortality was rare when fighting conspecifics. Together, these results suggest that while E. spinifer is more likely to win contests under current conditions, it could suffer considerable impacts if climate change and the spread of invasive C. destructor continue unabated.


Introduction
Across the globe, it has become increasingly apparent that extinction rates of freshwater taxa parallel and often exceed those of terrestrial taxa (Ricciardi and Rasmussen 1999;Jenkins 2003;Sala et al. 2000;Pimm et al. 2014). Declines in freshwater biodiversity have been associated with multiple factors, including the rapid spread of invasive species leading to declines in native biota and ecosystem functioning (Dudgeon et al. 2006). Although the mechanisms underlying the impacts of invasive species on natives are varied, competitive exclusion of natives from key resources by aggressive invasive species appears to be a common theme (Holway 1999; Co-Editors' Note: This paper is a contribution to the Behaviour in Aquatic Invasions Special Issue of Aquatic Invasions. Papers in this Special Issue explore how behaviour contributes to invasion success; native species' behavioural strategies that reduce the impacts of invasions; how knowledge of behaviour can enhance management of invasive species; and potential effects of climate change on the behavioural impacts of aquatic invasive species. Juette et al. 2014) alongside direct predation (Case and Bolger 1991;Worthington and Lowe-McConnell 1994), habitat modification (Koehn 2004;Kimbro et al. 2009) and disease transmission (Du Preez and Smit 2013).
Freshwater systems are proving particularly vulnerable to climate change (Woodward et al. 2010) as well as to invasive species. Climate change is expected to negatively impact freshwater fauna as a result of fluctuations in water temperature, salinity and flow (Palmer et al. 2008;Whitehead et al. 2009;Pratchett et al. 2011). Additionally, climate change is likely to contribute to the successful establishment of invasive species in freshwater ecosystems (Hellmann et al. 2008;Rahel and Olden 2008). However, empirical studies investigating the interactive effects of climate change and invasive species on native species are rare, and as such, we are unable to predict how native freshwater fauna will fare in the future. This is especially alarming from a conservation standpoint, because many of these native freshwater taxa are already considered vulnerable, endangered or critically endangered (e.g. Gallardo and Aldridge 2013;Muhlfeld et al. 2014).
To examine how climate change may modulate the behavioural interactions between invasive and native freshwater fauna, we used freshwater crayfish as our model system. Freshwater crayfish are often the largest-bodied invertebrates in freshwater ecosystems representing the bulk of the biomass, and are ecosystem engineers (Hale et al. 2016;Hudina et al. 2016). Some species have been widely introduced into non-native catchments worldwide, including the common yabby, Cherax destructor (McCormack 2014;Scalici et al. 2009;Capinha et al. 2012) (Supplementary material Figure S1). In Australia, this species is native to the Murray-Darling river system in NSW but has since expanded beyond its native range largely owing to the stocking of farm dams to provide recreational fishing opportunities (Coughran et al. 2009;McCormack 2014; Figure S1). This species' high fecundity, rapid growth, protracted spawning period and wide thermal tolerance range (1-35 °C) contribute to its invasive capability (Beatty et al. 2005;Withnall 2000;Veselý et al. 2015). Although largely speculative, this species has been designated a Key Threatening Process to members of the native crayfish genus, Euastacus, of which 34 of over 50 taxa are considered endangered or critically endangered (McCormack 2015;IUCN 2018). To date, there has only been one study investigating the interactions between C. destructor and a member of the genus Euastacus (Lopez et al. 2019), which contrary to expectations demonstrated that the critically endangered native species, Euastacus dharawalus (Morgan, 1997) won more contests and was more aggressive than the invasive C. destructor. However, the simultaneous impact of climate change on the outcome of resource contests has yet to be assessed and given that many Euastacus taxa are cold adapted (Horwitz 1990), may indicate greater vulnerability of Euastacus species than is currently expected.
Here we tested the hypothesis that the outcome and intensity of interspecific contests between Cherax destructor and Euastacus spinifer (Heller, 1865), would be influenced by water temperature. We note that E. spinifer is currently listed as "Least Concern" on the IUCN Red List but investigating its interactions with C. destructor is important for providing insights into the potential role of climate change on Euastacus spp. as its range currently overlaps with that of C. destructor ( Figure S1). Based on previous studies on crayfish contests, we predicted that i) E. spinifer would win more interspecific contests over resources and be more aggressive than C. destructor under current temperature regimes, but that ii) C. destructor would win more interspecific contests over resources and be more aggressive under near-future temperatures owing to its reported capacity to tolerate elevated temperatures (Seebacher and Wilson 2006). Given the unexpected deaths of crayfish following contests observed in the current study, we were able to further investigate the direct fitness consequences of interspecific contests under both temperatures. We compared the mortality rate of E. spinifer following contests with conspecifics versus C. destructor under current and near-future temperatures. We predicted that i) E. spinifer would suffer higher mortality following contests with C. destructor compared to contests with conspecifics, and ii) rates of E. spinifer mortality would be higher if they had contested under near-future temperatures than at current temperatures.

Materials and methods
The investigation was conducted between May 2016-January 2017 using Euastacus spinifer and Cherax destructor collected from farm dams and freshwater creeks in the Illawarra region of NSW, Australia. All specimens were captured using a combination of baited opera house traps, baited shrimp traps, baited rope techniques and by handheld nets. Once collected, individuals (C. destructor, N = 12; E. spinifer N = 12) were transported back to the Ecological Research Centre (ERC), University of Wollongong, NSW, where they were housed individually in recirculating freshwater aquaria (40 × 30 × 30 cm). The diet of both species while held in captivity consisted of vegetables and frozen fish meat. Each aquarium was lined with gravel with a PVC pipe for shelter. The occipital carapace length (OCL) of each crayfish was measured with callipers to the nearest 0.1 mm.
After 2 weeks acclimation, twelve E. spinifer -C. destructor pairs were created and randomly assigned to one of two temperature treatments (22 °C or 26 °C), creating 6 pairs per treatment. Temperatures were changed at a rate of 1 °C per day. A temperature of 22 °C corresponds with the current maximum annual temperature based on 2006-2016 mean hourly water temperature measurements recorded at the adjacent Cataract River gauge (Station number 2122323), and 26 °C corresponds to a 4 C increase in the maximum annual temperature expected at these latitudes by 2100 according to the B2 and A1B greenhouse gas emission scenarios (van Vliet et al. 2011).
Contests were then staged in a specialised observation aquarium (120 × 50 × 35 cm) that was divided into equal thirds using two opaque plastic dividers. A piece of shrimp was placed in the central compartment and the paired contestants introduced into the two end compartments for a 10-minute acclimation period. The dividers were then simultaneously lifted allowing the individuals to interact for 20-minutes. For each trial, contest outcome (defined by determining the first individual to tail-flick or flee from its opponent (Seebacher and Wilson 2006), contest intensity (number of aggressive acts) (Table S1 for ethogram) and the amount of time spent motionless were recorded (Gherardi et al. 2013). Individuals were placed back into their original housing aquarium following each trial and any subsequent mortality recorded.
During at least 12 days of rest, each individual was then gradually acclimated to the alternate temperature it had yet to be trialled at (at increments of ± 1 °C per 3 days) and each individual was trialled a second time at this alternate temperature paired with an unfamiliar individual as its new contestant. At the conclusion of the experiment we completed 12 interspecific contest trials conducted at 22 °C and 26 °C. In total, 5 new E. spinifer had to be collected during the course of the trials to replace individuals that had died following contests (making a total of 17 E. spinifer individuals used for inter-specific trials). Following the interspecific trials, intraspecific trials were conducted in the same manner but testing only pairs of E. spinifer under both temperatures. Twelve additional E. spinifer were caught for these trials and assigned into 6 contestant pairs, each fighting at 22 °C and 26 °C.

Statistical analyses
Normality of data was checked via visual inspections of QQ plots using RStudio. A Generalized Linear Mixed Model (GLMM) with binomial distribution and logit link function (lme4 package) was used to investigate the effect of species (categorical fixed effect) on outcome (win versus loss) of interspecific contests (binary response variable) under i) current (22 °C) and ii) near future (26 °C) temperatures. Trial ID was added as a random effect to control for the non-independence of individuals within a given pair of contestants. Another GLM with binomial distribution and logit link function was used to investigate the effect of relative body size of the individuals in each pair (categorical fixed effect) on the outcome (win versus loss) of interspecific contests (response variable) for E. spinifer and C. destructor separately, under i) current (22 °C) and ii) near future (26 °C) temperatures. Sample sizes did not permit the inclusion of both species and relative size in a global model. Note that temperature was not included as a fixed effect in the models as the total number of wins and losses is identical under both temperatures.
For contest intensity and activity, General Linear Mixed Models (GLMM) were used to investigate the effects of temperature, species, relative size and an interaction between temperature and species. Contest intensity was expressed as the number of aggressive acts performed (log transformed) and activity was expressed as the number of seconds motionless during a trial. Crayfish ID and Trial were included as random factors in both models to account for multiple testing of individuals. The same tests were conducted to analyse intraspecific contests but without species as a fixed effect.
To investigate mortality, a Generalized Linear Model was used with temperature and type of contest (inter-or intraspecific) as fixed effects and mortality (yes or no) as binomial response variable.

Results
In interspecific trials, the average size of Euastacus spinifer was 122.1 mm (± 6.3 SE) and Cherax destructor was 121.2 mm (± 4.9 SE). In intraspecific trials, the average size of E. spinifer was 110.6 mm (± 4.48 SE). In interspecific trials, E. spinifer was the larger contestant in N = 15 trials and the smaller contestant in N = 9 trials. In all trials (N = 24), all crayfish of both species engaged in aggressive acts.

Mortality
Unexpectedly we observed deaths of Euastacus spinifer following interspecific contests. A total of 10 out of 17 individuals died (Table 1) Table 1. Euastacus spinifer deaths following interspecific contest trials for first and second trials (randomised order of temperature presentation). Shown are the crayID numbers, temperature at which they had fought prior to death, date of trial and death and associated number of days to death post-trial and the winner/loser status of the deceased individual. For the second trials, one individual (#8) was a new crayfish that had been collected to replace an individual that previously died after the first contest at 26°C. The remaining individuals in the second trial had all initially contested at 22°C. there were no deaths of Cherax destructor following inter-specific contests. Since one of the E. spinifer that died did so 12 days post-trial, we excluded this death from analyses as it may have occurred for reasons other than the trial. The remaining deaths occurred 1 day (N = 3), 2 days (N = 5) or 4 days (N = 1) after a trial (Table 1). Given that these crayfish had already been acclimated to laboratory conditions under stringent husbandry protocols, these deaths were therefore deemed to have occurred as a direct result of the trials themselves. Following intra-specific contests, only 1 E. spinifer died (1/12 individuals) (Table 1). In summary, E. spinifer were significantly more likely to die if they had fought Cherax destructor rather than a conspecific (Type of contest,  2 1,34 = 4.831, P = 0.028) (Figure 3a) and if they had fought at high temperature rather than low (Temperature,  2 1,33 = 13.63, P < 0.0002) (Figure 3b). There was no interaction between type of contest and temperature (Type*Temperature,  2 1,32 = 0.069, P = 0.792).

Discussion
Examining the behavioural mechanisms underpinning the interactions between native and invasive species under projected conditions of climate change is central to the management of species of conservation concern. Here, we show that while the native Euastacus spinifer was more likely to win a contest over food under current temperatures, it lost its competitive advantage against the warm-water tolerant Cherax destructor under the elevated temperatures anticipated in the near future. Further, contests appeared to exert a greater cost on E. spinifer, with higher mortality of the native after contests with C. destructor than after contests with conspecifics, even if E. spinifer had won the contest. Importantly, this was particularly apparent after contests at high temperature. These interesting results suggest that the alteration of aquatic habitats by climate change could yield non-lethal as well as lethal impacts on native Euastacus species, through the modification of the outcomes and costs of contests. The increased time spent motionless by E. spinifer at high temperature during interspecific contests may represent an adaptive response enabling individuals to tolerate elevated temperature (Gherardi et al. 2013). However, E. spinifer did not spend more time motionless at high temperature during intraspecific contests, suggesting that contesting with C. destructor at high temperatures rather than contesting at high temperatures per se elicited the reduced activity. It is also unlikely that E. spinifer perceived C. destructor as a predator and hence reduced activity as a predator avoidance strategy (Stein and Magnusson 1976), since in many instances E. spinifer were actually larger than their opponents. Therefore, the results suggest that E. spinifer experienced unique physiological challenges when competing with C. destructor at high temperatures, and that the reduced activity that ensued likely resulted in the reversal of competitive dominance at high temperature. In contrast to the native species, C. destructor spent similar amounts of time motionless at both temperatures indicating that it experienced no particular physiological challenge from contesting with E. spinifer under elevated temperatures. These findings are in line with the fact that E. spinifer occupies freshwater systems colder than 24 °C (Coughran and Furse 2010) whereas the optimal thermal range for C. destructor has been reported to lie between 24-25 °C (Veselý et al. 2015), although it has been recorded in waters up to 35 °C (Withnall 2000). Interestingly however, individuals of E. spinifer did not show a reduction in the intensity with which they contested at high temperature despite the increase in time they spent motionless, demonstrating that they still had the capacity to fight. We speculate that this persistence in contest behaviour coupled with increased thermal stress produced the subsequent elevated mortality after contests with C. destructor.
Relative body size was a key predictor of contest outcome regardless of temperature, with larger individuals being more likely to win in both interand intraspecific trials. The key role of size asymmetry is not surprising and has been well documented in a range of decapod crustaceans (e.g. Vorburger and Ribi 1999;Nakata and Goshima 2003). In addition, body size was related to time spent motionless, with smaller crayfish spending more time motionless than larger crayfish during interspecific (but not intraspecific) contests. The fact that time spent motionless only varied between small and large crayfish during interspecific contests again highlights a contest dynamic that is unique to when E. spinifer engages wth C. destructor (rather than with conspecifics). In terms of competitive interactions, this finding does not bode well for the success of the other Euastacus species which reach smaller average body sizes than E. spinifer, which is one of the larger members of this genus (McCormack 2012). As such, investigating the behavioural and physiological responses of other smaller-bodied members of this genus would therefore represent an invaluable focus for future research.
Given that more E. spinifer died following contests with C. destructor and contests at high temperature, our study demonstrates the potentially lethal effects of contesting with heterospecifics when under thermal stress; which we had not anticipated. Deaths occurred typically within two days post-contest and are unlikely to be due to any other stressor as E. spinifer were acclimated to laboratory conditions without any mortality prior to the start of trials. Therefore, native crayfish species experienced both nonlethal (loss of resource) and lethal (mortality) effects of competition at high temperatures. Future work could involve monitoring stress levels (either from blood or water samples) before and after pair-wise contests under both temperatures, at low and high temperatures without contests, as well as between intra-and inter-specific contests at both temperatures, to more accurately determine whether the ultimate causes of mortality in E. spinifer were directly related to the cost of contesting versus temperature.
The wide thermal tolerance of C. destructor suggests that under nearfuture expected water temperatures this species is likely to continue spreading beyond its native range and into that occupied by Euastacus species. In addition, its apparent competitive dominance under higher temperatures as revealed in this study suggests that Euastacus species currently listed as endangered or critically endangered could face a real risk of extinction when the impacts of invasive C. destructor and climate warming act in concert. Further contest trials under altered temperatures that incorporate a greater range of Euastacus species and larger sample sizes are clearly needed, however, to reinforce the findings proposed here. Nevertheless, the important next steps would be to consider strategies that reduce the influx of this invasive species, such as the creation of barriers, education of the public and implementation of early detection. Further, maintenance of dispersal corridors, such as protected bushland, could allow Euastacus species to move and settle new habitats, as they will forage out of water, rather than being forced to compete with the invasive species, if colonisation of their waterways by C. destructor does occur. More generally, since climate warming is likely to magnify the effects of invasive species (Rahel and Olden 2008;Mainka and Howard 2010), there is an urgent need to consider ways to mitigate the impacts of climate change in order to ensure the persistence of native freshwater fauna with restricted thermal tolerances.