Do invaders always perform better? Comparing the response of native and invasive shrimps to temperature and salinity gradients in south-west Spain Lejeusne C.a*, Latchere O.a, Petit N.a, Rico C.a, Green A. J.a a Doñana Biological Station-CSIC, EBD-CSIC, Wetland Ecology Department, Avenida Américo Vespucio s/n, 41092 Sevilla, Spain * Corresponding author:
Abstract Invasive species are often thought to benefit from climate change, outcompeting native species as temperatures increase. However, the physiological tolerance has been little explored as a potential mechanism explaining biological invasion success. In this study, we used empirical data from both invasive and native estuarine species as a case study to address the hypotheses that (1) invasive species show a better resistance to acute thermal stress, (2) invasive species present lower oxygen consumption rates owing to greater resistance to environmental stressors, and (3) native species have lower survival rates under chronic temperature and salinity stress. We conducted various comparative experiments on three sympatric and syntopic closely related shrimp species (one invasive Palaemon macrodactylus, and two natives Palaemon longirostris and Palaemonetes varians). We evaluated their critical temperature maxima, their oxygen consumption rates under different salinities and temperatures, and their survival rates under chronic salinity and temperature. We found that the invasive species was the most tolerant to rapid increase in temperature, and consistently consumed less oxygen over a broad range of temperatures and salinities. P. macrodactylus also had lower mortality rates at high temperatures than P. longirostris. These results support previously reported differences in physiological tolerance between native and invasive species, with the invasive species always performing better. The consistently higher tolerance of the non-indigenous species to temperature variation suggests that climate change will increase the success of invaders.
Keywords:Introduced species, Estuarine organisms, Environmental factors, Biological Stress, Palaemon macrodactylus
Regional index terms: Europe, Spain, Andalusia, Guadalquivir River
Invasive species often have tremendous ecological impacts on invaded ecosystems and native species . They also have huge economic impacts estimated at more than five per cent of the global economy . Together with climate change, they constitute a “deadly duo” threatening worldwide biodiversity . Both factors can act individually on species abundances, distributions and biotic interactions, inducing local and regional extinctions , but they also can act synergistically
To become established then invasive, a non-indigenous species (NIS) has to successfully pass through a series of biotic and abiotic filters acting as barriers between the different steps of the invasion process . However, the mechanisms leading to a successful invasion are poorly understood in most cases. The numerous non-exclusive hypotheses proposed to explain invasion mechanisms, include evolutionary hypotheses (e.g. hybridisation) and ecological hypotheses (e.g. enemy release) . Another potential mechanism, the physiological tolerance hypothesis, is as yet relatively unexplored . This hypothesis predicts that invasive species have a greater and/or broader physiological tolerance than native species occupying the same habitat. Predictions of this hypothesis have been verified in a large taxonomical panel of species and stress factors . However, owing to the importance of climate change, most of the studies dealing with this hypothesis have focused on temperature effects and eurythermality of invasives compared to a more stenothermal tolerance of natives . In the present study, we address tolerance to two major environmental factors (salinity and temperature) as potential contributors to the success of an invasive estuarine species.
Estuaries are very productive ecosystems providing nursery habitats to many marine and commercial species. These marine-freshwater ecotones show strong fluctuations of physical and chemical parameters at both spatial and temporal scales (e.g. tidal-based salinity fluctuations with a decreasing spatial gradient from the inner mouth). Estuaries are particularly impacted by climate change but are also especially susceptible to biological invasions . In the San Francisco estuary, one new NIS is recorded every 14 weeks, and in Europe one fifth of estuarine species are NIS . One key question is whether NIS have more resistance to environmental stressors than native estuarine species, being better adapted to strong fluctuations in temperature and salinity.
The oriental shrimp (also known as migrant prawn, or grass shrimp) Palaemon macrodactylus is an estuarine caridean shrimp native to China, Japan and Korea. It was initially introduced to San Francisco Bay, CA in the 1960s, before spreading northward along the US coast. Since 1992, it has reached Europe, Argentina and the north-eastern USA coast . In European estuaries, the species has spread rapidly and extensively since its first introduction. It is now present from SW Spain to Germany and England, and in the western Black Sea. On the Atlantic coast, the species can interact with two other commercially exploited native species: the Atlantic ditch shrimp Palaemonetes varians (a brackish water species found mainly in non-tidal ponds, marshes and canals with hydrological connections to estuaries) and the delta prawn Palaemon longirostris (an estuarine species). Despite its relatively small size, P. varians is often captured for human consumption, use as fishing bait, or use as live diet for aquaculture , while traditional fishing of P. longirostris has local economic importance . Both native Palaemonidae can be very abundant and they occupy a central position in the estuarine trophic network , being prey of many European native and commercial fishes (e.g. the European sea bass Dicentrarchus labrax for P. longirostris) .
Competitive interactions between the NIS P. macrodactylus and the native P. longirostris may be strong, especially for space and food. Both species are estuarine with strong overlap in habitat and trophic preferences . In the Guadalquivir estuary (SW Spain), this habitat overlap is maximal in autumn during low abundance of their shared mysid prey Mesopodopsis slabberi . Since the NIS was first recorded, an increase in P. macrodactylus densities recorded in some European estuaries has coincided with a decrease in density of the native P. longirostris . A previous study comparing the osmoregulatory capacities of P. macrodactylus with the two natives P. longirostris and P. varians indicates that the three species have similar osmoregulatory capacities . However, oxygen consumption rates measured under different salinities and dissolved oxygen concentrations suggested that the NIS has a more efficient metabolism and higher tolerance to hypoxic conditions . However, despite field surveys showing the salinity-related and spatial distribution patterns of these estuarine species , little is known of the ecophysiology of the NIS P. macrodactylus compared to the natives P. longirostris and P. varians, in particular regarding the influence of temperature variations. Taking into account the climate change expected in the Euro-Mediterranean area, the interaction between temperature and salinity might be central to the success of NIS and to changes in status of native species .
Studying the relative performance of NIS and natives under a range of environmental conditions allows evaluation of the likely mechanisms of a successful invasion, and testing of the physiological tolerance hypothesis. We therefore conducted a series of three experiments to test if P. macrodactylus performs better under extreme conditions of temperature and salinity, the two main abiotic stress factors found in estuaries. We evaluated behavioural activity and the critical temperature maxima of different shrimp species under an acute short-term thermal stress. We hypothesized that the NIS would show greater resistance to acute thermal stress, reflected in a higher critical thermal maximum. We also measured oxygen consumption under different conditions of temperature and salinity to test whether the NIS species present lower consumption rates owing to greater resistance to environmental stressors. Finally, we quantified survival under different chronic thermal and salinity stress to test whether the native species had lower survival rates.
Material and methods Shrimp collection and laboratory acclimation
The oriental shrimp P. macrodactylus and the delta prawn P. longirostris were collected in the Guadalquivir estuary, SW Spain (see Figure 1) at three distinct, tidal sites S1-S3, with P. macrodactylus was only found at site S2 (environmental parameters at each site are described in Appendix A). The Atlantic ditch shrimp P. varians was sampled in Veta La Palma (S4 and S5), a complex of fish ponds connected to and supplied with water from the Guadalquivir estuary (Figure 1 and Appendix A) and protected within Doñana Natural Park, where it is abundant and harvested commercially . Living individuals were collected in 2011 using shrimp keep-nets (mesh size 4mm) placed at low tide for S1-S3 and recuperated 24h later. Size of the shrimps was estimated by measuring the carapace length from the orbital edge of the eye to the edge of the cephalothorax under a stereomicroscope SteREO Discovery V8 (Zeiss) using the AxioVision Rel 4.8.2 (Zeiss) software. In order to reduce catching and manipulation stress, living shrimps were acclimated during at least 48h before any experiment in aerated aquariums with artificial saltwater at 20°C and a salinity of 5, obtained by dissolving dry sea-salt Instant Ocean (Aquarium Systems, Mentor, Ohio) in distilled water. Salinity was measured using the Practical Salinity Scale. Aquariums were placed in a climatic chamber (Fitoclima 10000EHHF, Aralab) on a 12h:12h dark:light photoperiod. Shrimps were fed daily ad libitum with commercial aquarium food (gammarids) before and during all the experiments. In order to reduce stress and injury associated with its determination, sex was characterized after the experiments by looking for the presence or absence of the masculine appendix on the endopodite of the second pleopod . A summary of size and sex ratio of the specimens used in each experiment is given in Table 1.
Experiment 1: critical Thermal maximum (CTmax) experiment
In order to compare thermal stress resistance between the shrimp species, Critical Thermal maximum (CTmax) experiments were conducted in May and August 2011. Carapace lengths were measured before the experiment. The experiment was not started until at least 24h after measurements of length. Acclimated shrimps were placed individually in a beaker filled with 200 ml of artificial water (salinity 5 and initial temperature 20 ºC) and capped with a transparent lid to allow observation throughout the experiment. The beaker was placed in a water bath with a magnetic stirrer allowing rapid homogenisation of surrounding water. Temperature was monitored every minute with an electronic thermometer (model SA880SSX, Oregon Scientific) and a temperature ramp of 1°C.min-1 was applied as in .
During the experiment, behavioral activity of each individual shrimp was continuously monitored over 30 s periods until reaching the end-point when the shrimp lay on its side or its posterior face for more than 30 s. We subdivided behaviour into four categories based on previous literature : ‘Movement’: any kind of motion of the animal except for active movement (see below): pereopods or pleopods movements, antennal lateral sweeping on the dorsal side, cleaning of mouth parts by rubbing them together; ‘Active movement’: when shrimps moved a distance (either by walking or swimming) exceeding their own length in less than 30 seconds; ‘Loss Of Equilibrium’ (LOE): shrimp on the bottom of the beaker in either an ‘upside-down' or a 'sideways' position for more than 2 seconds; ‘Spasmodic motions or spasms’: vibrations of the pleopods and/or sudden contraction of the abdomen.
The CTmax was determined as the temperature at which coordinated movements were lost, using LOE as the reference parameter. The CTmax was calculated for a total number of 20 individuals per species for P. macrodactylus and P. varians, and 18 individuals for P. longirostris.
Experiment 2: oxygen consumption rate
In order to compare oxygen consumption under different temperatures and salinities, we performed two series of treatments with varying temperatures (20°C, 25°C and 30°C at a constant salinity of 5) and salinities (salinity of 5, 15, 25, 35 and 45, at constant 20°C) respectively in May 2011 (9-10 shrimps per treatment for P. longirostris and P. varians, and 5 shrimps per treatment for P. macrodactylus). Each shrimp was weighted 24h before experiment with a Voyager analytical balance (Ohaus) after removal of excess water using blotting paper. In order to avoid any heat shock when moving shrimps from their original aquarium (20°C at salinity 5) to aquariums with higher temperatures, they were acclimated overnight prior to the experiments by gradually increasing temperature (2°C.h-1) or salinity (10 salinity units.h-1) depending on the treatment. Temperature within each treatment was maintained within ± 0.2°C using a heater (Jäger 300; Eheim) controlled by a Biotherm Pro (Hobby) temperature regulator.
To measure oxygen consumption rate (OCR), shrimps were put into cylindrical flasks (12.3 mL) and the flow rate of water circulating in each flask was measured. The difference between oxygen concentrations in water at the entrance and exit of the flasks was recorded using a 10-channel OXY-10 mini (PreSens) fiber optic oxygen transmitter connected to a computer with the OXY10v3_33 software. OCR was calculated according to the formula: OCR= F × ([O2]in – [O2]out)/BW, where OCR is oxygen consumption rate (mg O2.gwwt-1.h-1), F is water flow rate (L.h-1) circulating in each flask, [O2]in is oxygen content of the water inflow (mg O2.L-1), [O2]out is oxygen content of the water outflow (mg O2.L-1), and BW is wet mass (g).
Experiment 3: comparative survival of P. longirostris and P. macrodactylus under chronic stress
Shrimps were placed individually in small, closed plastic aquaria (0.35L) with a 1mm mesh sieve at the bottom and placed within 91L experimental aquaria.
Acclimatised shrimps were reared in different 91L aquaria at three different temperatures (20°C, 24°C and 28°C, with a constant salinity of 5) and three different salinities (5, 25, and 45, with a constant temperature of 20°C) during 28 days. As environmental parameters are constantly varying in an estuary, especially salinity that has tide-based regime, submitting estuarine organisms to constant salinity or temperature as here represents thus a chronic stress. Shrimp size and weight were measured twice a week and survival was checked daily.
Experiments were conducted in October 2011 when both species were caught at the same time and place. The experiment was repeated twice (15 day interval between sampling) on both species (8 individuals per treatment for both P. longirostris and P. macrodactylus and per sampling time).
All statistical analyses were performed using R 2.15.2 . The CTmax data were not normally distributed, even after transformations (Shapiro-Wilk’s test, p<0.001) and had unequal variances (Bartlett test, p<0.001). These data were therefore analysed using the non-parametric Kruskal-Wallis ANOVA and Wilcoxon-Mann-Whitney tests. However, for oxygen consumption data, a parametric ANOVA was performed, after data transformation when necessary. The post-hoc Tukey HSD test was used to compare treatments. Survival analysis were performed with the Survival package in R using Kaplan-Maier estimates and log-rank tests.
Results Experiment 1: behavioral analysis and Critical Thermal maximum (CTmax) experiment
P. macrodactylus was collected at only one site (S2) in the estuary (Fig. 1). In contrast, P. longirostris was collected at the three sampling sites in the Guadalquivir River (Fig. 1), representing a decreasing salinity gradient from S1 to S3.
Due to density variations at each sampling site, the experiment was conducted on 9, 7 and 2 individuals of P. longirostris. sampled the same day at sites S1, S2 and S3 respectively. We pooled the different sites into one group for further comparative analysis as no statistical differences were found among sites (Kruskal-Wallis test, p=0.37 and p=0.82 for the CTmax and temperature at first spasmodic motion respectively).
In the same manner, P. varians was sampled at two sites (S4 and S5; Fig. 1) differing in salinity. The experiment was conducted on 10 individuals from each site. No statistical differences were found for the CTmax and for the temperature at first spasmodic motion between the two sites (Wilcoxon-Mann-Whitney test, p=0.15 and p=0.17 respectively). Hence, we pooled the two sites for further analysis.
Despite identical pre-experimental acclimation conditions, at the start of the experiment, no P. longirostris individual was presenting an active motion, whereas 60% and 20% of P. varians and P. macrodactylus, respectively were presenting an active motion (Fig. 2). The temperatures at which 50% of individuals were actively moving were reached earlier for P. varians (20.0 ºC) and P. macrodactylus (22.0 ºC) than for P. longirostris (24.6 ºC). The maximal number of individuals presenting active movements was reached as early as ca. 28 ºC for P. longirostris, compared to ca. 33 ºC and ca. 32.0 ºC for P. varians and P. macrodactylus respectively. Moreover, the active moving curves closely preceded the loss of equilibrium (LOE) curves (Fig. 2). In the case of P. varians and P. macrodactylus, the LOE and spasms curves were closer together. As a consequence, LOE was observed at much lower temperature for P. longirostris (mean CTmax ± SE = 27.24 ºC ± 2.16) compared to P. varians (mean CTmax ± SE = 31.71 ºC ± 2.21) and P. macrodactylus (mean CTmax ± SE = 33.0 ºC ± 1.11)(Fig. 2 and 3). The CTmax values were significantly different between species (Kruskal-Wallis test, H=33.3276, df=2, p<0.001), with all pairwise comparisons between the three species being significantly different (Wilcoxon-Mann-Whitney test, p<0.001).
For individuals used in the above experiments, information about size and sex ratio of the different samples have been gathered into Table 1. When considering each species separately, there was no effect of size (Spearman rank correlation, p>0.15) or sex on LOE values (Kruskal-Wallis test, p>0.05).
Experiment 2: oxygen consumption rate
In the salinity experiment, OCR values were significantly different between species and salinity treatments (four-way ANOVA, F8,109=15.58, p<0.001) without effects of sex or size (see detailed value of these parameters in Table 1). The significant differences between species were due to lower OCRs of P. macrodactylus compared with the two native species (post hoc Tukey HSD test, p<0.001), which did not differ between them (post hoc Tukey HSD test, p=0.075). Whatever the salinity treatment, P. varians had the highest OCR values and P. macrodactylus the lowest (Fig. 4).
The significant effect of salinity treatment reflected an increase in OCR with salinity (Fig. 4), for all species (134%, 186% and 236% increase for Palaemonetes varians, P. longirostris, and P. macrodactylus respectively, between the lowest and highest treatments). In the case of P. longirostris, which presented intermediate OCR values, OCR was significantly lower at the lowest salinity treatment (salinity 5) than at all others. Low salinity effects were more gradual for the other two species (Fig. 4).
In the temperature experiment, OCRs were significantly different between species and temperature treatments (four-way ANOVA, F6,65=21.30, p<0.001) without effects of sex or size (see Table 1 for details on values of these parameters). All pairwise species differences were significant (post hoc Tukey HSD tests, p<0.015). As for the salinity experiment, P. macrodactylus had the lowest OCR, and P. varians the highest. OCR increased significantly with increasing temperature for all species (Fig. 4; 183%, 257% and 200% increase for Palaemonetes varians, P. longirostris, and P. macrodactylus respectively, between the lowest and highest treatments).
Experiment 3: comparative survival of P. longirostris and P. macrodactylus under chronic stress
We did not find any significant effect of sampling date for P. macrodactylus (log-rank test, χ2=0, df=1, p=0.923) or P. longirostris (log-rank test, χ2=0, df=1, p=0.989) on survival rate at different salinities. Likewise, no significant effect of the date of sampling was found in the temperature experiment (P. macrodactylus, log-rank test, χ2=0, df=1, p=0.946; P. longirostris, log-rank test, χ2=0, df=1, p=0.993).Thus, samples from different dates were pooled for further analyses. In the treatment 20 ºC-salinity 5 that provided the least stressful conditions, only one individual (P. macrodactylus) showed a premature death (two weeks before the end of the experiment). No significant size differences was found between the two shrimp species in each of the treatment (see Table 1).
Survival curves of both species according to salinity are shown in Fig. 5. There was a significant effect of salinity treatment on the survival of P. macrodactylus (log-rank test, χ2=54.0, df=2, p<0.001) and of P. longirostris (log-rank test, χ2=41.3, df=2, p<0.001). No significant difference was found between the two species when comparing their general survival trend according to salinity (log-rank test, χ2=1, df=1, p=0.308): the higher the salinity, the lower the survival (Fig. 5). However, when comparing survival of both species in detail at each salinity treatment, significant interspecific differences were noted only at the high salinity of 45 (log-rank test, χ2=8.6, df=1, p=0.00338).
In the temperature experiment (Fig. 5), no significant effect of temperature was found on survival of P. macrodactylus (log-rank test, χ2=1.2, df=2, p=0.555), despite a slight increase of mortality with increasing temperature (Fig. 5). However for P. longirostris, temperature had a significant effect (log-rank test, χ2=19.4, df=2, p<0.001), the highest temperature (28 ºC) being the only one to induce mortality (50% of individuals dead by the end of the experiment; Fig. 5). The general trend for survival with temperature was not significantly different between the two species (log-rank test, χ2=0.3, df=1, p=0.570). However, a marginally significant difference in survival could be noted between the two species at 28 ºC (log-rank test, χ2=3.8, df=1, p=0.051).