Re-evaluating invasive species in degraded ecosystems: a case study of red-eared slider turtles as partial ecological analogs

Exotic species are often vilified as “bad” without consideration of the potential they have for contributing to ecological functions in degraded ecosystems. The red-eared slider turtle (RES) has been disparaged as one of the worst invasive species. Based on this review, we suggest that RES contribute some ecosystem functions in urban wetlands comparable to those provided by the native turtles they sometimes dominate or replace. While we do not advocate for releases outside their native range, or into natural environments, in this review, we examine the case for the RES to be considered potentially beneficial in heavily human-altered and degraded ecosystems where native turtles struggle or fail to persist. After reviewing the ecosystem functions RESs are known to provide, we conclude that in many modified environments the RES is a partial ecological analog to native turtles and removing them may obviate the ecological benefits they provide. We also suggest research avenues to better understand the role of RESs in heavily modified wetlands.


Introduction
Invasive exotic species have long been recognized for having negative effects on the biodiversity of native ecosystems. As early as 1958, Charles Elton stated "We are living in a period of the world's history when the mingling of thousands of kinds of organisms from different parts of the world is setting up terrific dislocations in nature …" Modern ecologists are now even more aware of the problems caused by the invasion of exotic species into natural areas and the attendant effects on local, regional, and global patterns of biodiversity [1,2], now and in the future [3]. The spread of exotic species into novel habitats is considered the greatest threat to biodiversity after habitat loss [4] and/or climate change [5]. Once established, some exotic species can displace or even permanently replace native plant and animal species, as well as disrupt ecosystem processes including nutrient and fire cycles, and cause changes in patterns of plant succession [6,7].
The documented negative effects of invasive species have produced a deeply entrenched dichotomy of opinions among conservation biologists. Exotic species are often labelled as "bad" and native species are considered to be "good" [8], with little if any middle ground [9]. The gulf is further widened in the literature by reference to native species as "friends," and non-native species as "foes" [10][11][12]. We agree that preservation of native biodiversity in natural habitats is an important goal of effective conservation efforts. However, might some exotic species perform compensatory and potentially valuable ecological functions in degraded urban habitat where native species have already lost their 1 3 have been introduced to Europe, reproducing in the warmer Mediterranean climate as well as in more temperate climates [33], and their presence is reported to have negative effects on the European pond turtle, Emys orbicularis [34,35], but evidence is speculative or based on small-scale experiments under artificial conditions that may not reflect interactions in complex natural environments [31].
Wetlands altered and degraded by human activities are the rule rather than the exception globally, providing novel habitats that can challenge native species, including turtles. Disentangling the detrimental impacts of invasive species dominance from other causes of species' decline is difficult due to the confounding and interacting effects of human impacts to aquatic systems. As an omnivore and habitat generalist [36], RESs appear to be well adapted to benefit from these changes and often dominate in degraded urban settings where they have been introduced.

Why are RESs so successful?
Against the many negative obstacles facing most modern turtles, we find RESs successfully filling niches where many native turtle populations are declining. Their success in areas outside their native geographic range has resulted primarily from continuing human-mediated pet releases into favorable biotic and abiotic conditions in new environments. RESs are now widely distributed [37] in all 48 US contiguous states and Hawaii as well as in 64 different countries (mostly in urban or peri-urban locations).
RESs appear to tolerate human activity better than other turtles [38] as well as abiding the presence of heterospecifics better than others, like the Mediterranean pond turtle, Mauremys leprosa [39]. Released from competition with many sympatric turtle species in their native range, RESs might have an advantage in new environments with fewer competitors [40,41]. Non-native RESs are slower to respond to predator threats in California [42] so will bask longer and, because of a smaller surface-to-volume ratio than the native western pond turtle (Actinemys marmorata), RESs might gain a thermal advantage. As climate change has the potential to warm temperate areas where RESs do not yet reproduce [43,44], future range expansion is a foreseeable possibility [45,46]. Introduced RESs may have an advantage in that native prey do not always recognize the invaders as threats [39,47]. RESs might also be faster at handling prey [48], permitting greater prey consumption. Juvenile RESs also outcompete other juveniles in limited-resources experiments against red-bellied turtles (Pseudemys rubriventris) [49] and can outcompete European pond turtles (Emys orbicularis) and the Mediterranean pond turtle (Mauremys leprosa) in experimentally controlled conditions [34,35,39].
The potential invasiveness of the RES has been attributed to a variety of impacts, including their ability to outcompete native turtles for basking sites in experimental ponds [34] or to affect native turtle basking behaviour in human-altered closed waterways [50]. These results have rarely been documented in natural wetlands as basking sites are not a limiting factor in most wetland environments [46]. However, Drost et al. [31] found that the presence of RESs at a natural albeit unusual wetland (a small collapsed travertine spring) with limited basking sites inhibited basking in Sonoran mud turtles (Kinosternon sonoriensis). In certain cases, invasiveness is presumed without a quantifiable ecological threat. For example, in Queensland, Australia, RESs are declared a Class-1 pest [51] because they have the "potential" (although not demonstrated in the local habitat) to cause adverse economic, environmental, or social impacts. That assessment is based on presenting a range of attributes that confer significant pest potential such as: climatic suitability, a history of being a pest in non-native countries, a broad geographic range, a general diet, and high fecundity.
In Spain, in a study of two coastal lagoons, researchers found that RESs thrived on a generalist omnivorous diet of leaves, seeds and weeds and that their preferred prey was an invasive species, the red swamp crayfish (Procambarus clarkia) [52]. RESs could also potentially outcompete some native turtles by maturing earlier and having greater fecundity [36,53]. Similarly, captive breeding stock of RESs in South Africa have shown an ability to adapt their reproductive cycles to match the Southern hemisphere summer months [54], thereby showing advantageous reproductive plasticity.
Turtles are relatively long-lived and feed at various trophic levels and as such become reservoirs for various toxins and heavy metals [55] and pass these toxins to their eggs [56,57]. Degraded urban wetlands are often recipients of heavy metal runoff [58] and turtles can bioaccumulate various contaminants [59]. Bioaccumulation of toxins can have deleterious effects including suppression of the immune system and the prevalence of shell disease [60]. We suspect that many RESs are released into wetlands as adults after they have outgrown their owner's abilities or desires to care for them. An online investigation of the age classes of turtles put up for adoption in Ontario (www. littl eresq. net) found a vast majority of the available RES turtles were given up as adult (32 of 36). Therefore, if the same holds true for wild-released turtles, these have benefitted from a captive upbringing therefore not being exposed to the same toxin concentrations, microplastics, and other pollutants during their captive lives. Ironically, a partial captive life might give late-released RES founder populations an additional advantage over local turtles that have been bioaccumulating toxins throughout their lives.

Threats from RESs
Genetic admixture is a serious threat that RESs pose to other members of their genus as well as other genera. Sixteen species, and numerous subspecies, are currently recognized in Trachemys [61] throughout North and South America. When RESs are introduced into habitats occupied by other taxa in the genus, hybridization or introgression (among subspecies) can occur [62], threatening the survival of unique lineages such as the Rio Grande slider turtle (T. gaigeae). Hybrids have also been reported between RESs and northern map turtles (Graptemys geographica- [63]), suggesting that other intergeneric hybrids are possible.
The introduction of RESs may also include the transport and introduction of novel diseases and parasites. For example, RESs are suspected to have transmitted non-native blood flukes, Spirorchis elegans, to a population of European pond turtles (Emys orbicularis) in a natural wetland in Spain in a parasite "spillover" event [64]. Cross-transmission of helminth parasites from RESs to native Spanish turtles was also reported by Hidalgo-Vila et al. [65]. However, populations of western pond turtles (Actinemys spp.) in California, USA, sympatric with non-native RESs did not have increased risk of Mycoplasma spp. infection compared to populations that did not have RESs [66]. Shell disease has been documented in various turtle species and populations, including slider turtles, but the etiology of the disease and cross-species transmission is not wellunderstood in sliders [67] or other turtle species [68]. RESs as well as fish are also confirmed carriers of various diseases and can transmit some (e.g., Ranavirus) to amphibians in experimental settings, without being affected themselves [69].

Turtles perform important ecosystem functions
Turtles perform crucial ecosystem functions that fall into the following six categories biomass contributions, energy flow and scavenging value, mineral cycling and bioaccumulation, trophic status, seed dispersal and germination enhancement, and bioturbation in soil dynamics (for an in-depth review see Lovich et al. [70]). These general categories show the important breadth of ecosystem services that turtles perform, and we can infer that with greater numbers and densities in the past, these services were greatly enhanced. However, even in low densities and in temporary mesocosms some turtles have shown significant positive effects on ecosystem services [71]. Because of the challenges that face turtles worldwide, conservation biologists have tried to find ways to protect turtles and preserve the important ecosystem services they provide. A small percentage of turtles have become naturalized outside of their native range. In a review of introduced reptiles and amphibians, Lever [72] listed 20 turtles of the 356 species recognized in the most current taxonomic checklist [73]. Almost nothing is known about the potential effects turtle introductions have on native turtles, other than research with RESs.
RESs, like most turtles, perform all six crucial ecosystem services to some degree, yet ironically are considered undesirable species even where native turtles are no longer available in numbers high enough to do so.
1. Biomass contribution. In non-native environments, slider biomass can become an important proportion of a wetland ecosystem with RES densities in Japan ranging from 89.6 to 299.6 turtles/ha and 47.8-186 kg/ha [41]. These biomass estimates surpass the average biomass of RESs in their native range of 41.8-61.6 turtles/ha and 33.6-37.1 kg/ha. RESs share their native range with many heterospecifics, including snapping turtles (Chelydra serpentina), chicken turtles (Deirochelys reticularia), Eastern mud turtles (Kinosternon subrubrum), Florida cooters (Pseudemys floridana) and stinkpot turtles (Sternotherus odoratus). In these particular communities, RESs can be the dominant turtle species in some habitats, sometimes accounting for over 60% of the turtles present [74]. One of the largest known modern biomasses attained by a vertebrate was for the closely related yellow-bellied slider (T. s. scripta) at 877 kg/ha [75] and up to 2200 individual turtles/ha in a disturbed nutrient-enriched habitat [76]. In the Sonoran Desert, the Mexican mud turtle (Kinosternon integrum) reached a standing crop biomass of 2860 kg/ha with a density of 20,000 individuals/ ha. [77]. Not all habitats can sustain high turtle biomass, but high biomass implies an important role in ecosystem food webs, both as predators and prey [75]. Late Jurassic records of an estimated 1800 turtles in an area less than a hectare, probably concentrated due to drought [78], nonetheless support the contention that turtles have held an important role in aquatic ecosystems for millions of years. 2. Energy transfer and scavenging value. Turtle eggs and hatchlings, and to a lesser extent larger juveniles and adults, contribute to a transfer in energy flow from aquatic to terrestrial environments as prey. Eggs are rich in lipids and proteins and are an important source of food for many native and introduced predators [29]. RESs in their natural range often have the largest biomass of eggs, due to the sheer number of females in the community, even though the mean number of eggs per clutch is not as large as other sympatric species (e.g., C. serpentina averages 35.2 eggs (range 4-109 eggs) versus RESs 10.5 eggs (range 1-30 eggs), [29]). However, RESs in non-native habitats may have greater reproductive output than local turtles; for example, in Japan, RES turtles produce 1-19 eggs per clutch compared to the Japanese pond turtle (Mauremys japonica) clutch sizes of 5-8 eggs, [79]. RESs are also known to scavenge [29,80] and survive in effluent ponds with no emergent vegetation, surviving on detritus that falls into the water [81]. 3. Mineral cycling and bioaccumulation. Because turtles are long-lived and accumulate and store minerals (calcium, phosphorous) in their shells, as well as the calcium required for the annual egg production, they may play an important role in biogeochemical cycles [70]. RESs and other Trachemys relatives also accumulate several environmental pollutants (coal ash residue and copper- [82], mercury- [83,84], cadmium- [84,85], other metals- [86], radioactive compounds- [87]. Amazingly, studies on T. s. scripta, showed tolerance to low doses of gamma and beta radiation [88]. 4. Trophic status. RESs are omnivorous and opportunistic, starting out as mostly carnivorous as hatchlings and becoming progressively more herbivorous as they age [89], although adults become preferential carnivores when animal food is available [90]. RESs are also an important prey as hatchlings, which suffer predation by a wide variety of predators including ants [91], frogs, snapping turtles, snakes, alligators, mammals, and birds. Ironically, T. scripta hatchlings appear to be rejected as a prey item by largemouth bass (Micropterus salmoides) as behavioural adaptations or shell architecture make them unpalatable [92]. Adult slider turtles fall prey to a variety of carnivorous mammals, e.g., raccoons (Procyon lotor) and river otters (Lontra canadensis) and alligators (Alligator mississippiensis). These predator/prey roles presumably continue in novel environments, although with a slightly different cast of predators in many cases. In a study of experimental ponds stocked with RESs or devoid of turtles, researchers found that turtle-filled ponds showed significantly improved pH levels, conductivity, sediment accumulation, leaf litter decomposition rates, and a higher abundance of invertebrates [93]. 5. Seed dispersal. Turtles are well known fruit and seed dispersers and passage through turtle digestive tracts enhance germination of many seeds and spread them across the landscape [28,94,95]. RESs have been shown to disperse seeds [96] similarly to midland painted and snapping turtles [97], including Nymphaea seeds [96,98]. However, passage through the gut of RESs is not beneficial to all seed germination [96] since mulberry (Morus sp.) seeds are often damaged in the process. 6. Bioturbation in soil dynamics. Unlike tortoises, where the digging of burrows can alter the landscape and benefit multiple species including invertebrates, snakes, lizards, mammals, and birds, we know little about the role of RESs in soil dynamics. Females have been documented to move up to 1.6 km away from water to find suitable nesting sites [81]. Certainly, in large densities, RESs will have cumulative bioturbation effects on land where they nest by digging

Ecological functions are diminished when turtle populations decline
Turtle populations worldwide have declined precipitously because of anthropogenic pressures [101,102], especially overutilization and habitat destruction [103]. Although we often do not have rigorous historical estimates of turtle population abundance or densities, convincing evidence exists that past densities of some turtle species were orders of magnitude greater than they are today. As a result, current densities of both freshwater and marine turtles are likely a mere fraction of what they once were [70,104,105]. This has significant ecological consequences, since turtles often represent the major vertebrate portion of an aquatic community by virtue of their potential for high population density and thus biomass [75,106]. Depleted turtle populations continue to be negatively affected by a multitude of anthropogenic factors including habitat loss [107][108][109][110], road mortality [111], altered sex ratios due to climate change [112,113], subsidized predators [114,115], disease [116], boat activity and related injuries/mortalities [117,118], over-collection [119], microplastic exposure [120], and invasive species [121,122]. Even when habitat is putatively protected, studies have documented declines in freshwater turtle population densities due to subsidized predators, habitat succession, and poaching [123][124][125]. Even in undisturbed non-urban areas, climate change, including changes to precipitation patterns and the accompanying risk of drought and flood, could have far-reaching consequences on native turtles unable to adapt to or escape from new climatic conditions [45,126,127].

Are RES a potential solution in degraded ecosystems?
RESs are not newcomers to New World turtle assemblages, which have fluctuated over time due to glacial movements, changes in climate, and the connection of North and South America by the Isthmus of Panama [128]. At a fossil-rich site in South Carolina, late Pleistocene Trachemys fossils were found in a diverse fourteen-member turtle community that has no modern analog. Ancestors of the modern sliders were sympatric with turtles of the genera Kinosternon, Sternotherus, Chelydra, Macrochelys, Chrysemys, Clemmys, Glyptemys, Terrapene, Deirochelys, Emydoidea, Pseudemys, and Apalone [129], some of which have not been found in sympatry since then. This suggests that sliders have a long history of coexistence with many other turtle species and that modern assemblages might have some degree of tolerance for new combinations of species. In modern times, RESs live in varying degrees of sympatry with as many as nine other freshwater turtle species in some areas [74,75].
We are not advocating for further releases or maintenance of RES populations in natural wetlands. However, we do suggest delving more deeply into the impacts, both positive and negative, of this introduced species as contributors to ecosystem functions in degraded wetlands and advocate for managing the species based on these impacts [130]. We recognize that some fragile native turtle populations in natural wetlands can be negatively affected by the introduction of RESs [31,50] or that genetic admixture can potentially threaten native turtles inhabiting small ranges. When does the balance of ecological functions tip in favour of not removing RES in degraded isolated wetlands versus removing RES from wetlands that are connected to more pristine waterways and native turtle populations?
RESs in novel, degraded ecosystems pose an ecological conundrum (see Fig. 2). On the one hand, RESs are non-native, but on the other hand they provide important ecological functions, especially in areas where native turtle populations are much reduced or absent. In addition, RESs in urban and or degraded environments are often conspicuous to potentially large numbers of wildlife watchers [131] and increase turtle appreciation that could further awareness and effective conservation action at the local level [132], with proper interpretative programs. However, to date mixed evidence exists that general awareness of turtles translates into more research on the most critically endangered species [133] or that turtle eco-tourism results in effective conservation action [134]. Also, evidence has been provided that the conspicuousness and pervasiveness of RESs in non-native environments can cause a form of generational amnesia where the RES is accepted as a native species and creeps into local culture [135].
Conservation biologists have raised alarms about the diminishing numbers and species of turtles and the fact that most known turtle species are either extinct or threatened by extinction worldwide [70,101,102]. By vilifying RESs instead of admonishing the poor behaviour by pet owners, scientists risk confusing and even alienating the public from the broader issue of the global need for turtle conservation [102] when RES are targeted for removal from urban and peri-urban wetlands where they are visible to large numbers of people. Would it be better to let RES persists in these environments, knowing that they contribute to ecological functions in wetlands? Is it time to rethink the entrenched dogma that RESs always have negative effects? We suggest that there is scientific merit in discussing and debating the potential ecological value of non-native RESs in novel and degraded ecosystems while reiterating our position that future releases are not condoned or promoted. As Davis et al. [136] argue: "don't judge species on their origins" but on the ecological functions they provide. We suggest that the "guilty until proven innocent" construct [137] that pervades invasion biology overlooks the ecological functions non-native RESs might bring to impoverished urban wetlands. In degraded anthropogenic environments, where urbanisation has introduced hundreds of non-native species that have colonized, naturalized, or invaded the host ecosystem, RESs as generalist turtles show adaptability to these novel ecosystems [138].

Future research avenues
Although RESs are probably the most studied turtle species in the world [133], their ecological role in novel ecosystems has not been investigated in depth or detail. More research is needed to understand the ecological carrying capacity for turtles, and how a plethora of anthropogenic issues suppress populations below that density. A major consideration for field experiments examining the role of RESs in non-native habitat is that comparison with control systems in which RESs have not been introduced is essential. Other research questions that deserve attention include.
• Do non-native RES turtles have beneficial, detrimental, or neutral impacts on native turtle populations in degraded habitats/environments? • How do non-native RESs turtles compare with other introduced turtle species in terms of beneficial, detrimental, or neutral impacts? • Is adding non-native RESs to an existing turtle community an ecological zero-sum game? • Which localities have the greatest danger of genetic admixture and what are the likely outcomes? • In degraded urban and peri-urban ecosystems where native turtles are in decline or absent, are there potential benefits of adding RESs or maintaining existing RES populations? • How should conservationists manage RESs if these turtles are truly better adapted to the no-analog urban wetland ecosystems? • Is there value to having a large colourful and conspicuous non-native turtle species like RESs that can lead to an increase in public awareness of less conspicuous and endangered native species?

Conclusion
Obviously, conservationists care deeply about native turtles and preserving a pristine environment for them. However, many of the urban and peri-urban wetlands around the world have been highly disturbed, and the novel communities of organisms living in these degraded wetlands no longer resemble the wetland communities in which native turtles evolved. Native turtles everywhere are assaulted by varying degrees of anthropogenic stressors that limit their ability to reach carrying capacity and fulfill their ecological roles [71]. Based on the discussion above we suggest that: 1. RESs can be partial ecological analogs for native turtle species; 2. RESs persist or even prosper because of human introduction, propagule pressure and modification of wetlands that are detrimental to native turtles. Thus, in many cases RESs are "passengers" and not "drivers" of the decline in local turtle populations, i.e., RES are not directly interacting with subordinate native species and limiting native species survival but are more likely "along for the ride" [139]; 3. Preservation of native species and habitats is a primary goal of conservation but in urban and peri-urban wetlands dominated by non-native species, RESs might be useful ecological surrogates for populations of native turtles that are reduced or eliminated by anthropogenic pressures.
Considering these conclusions, the adaptability of RES turtles may have value to the ecological functioning of novel ecosystems. The often-reflexive response of conservation biologists who view RESs and other invasive species as villainous, when instead they might be an asset under some circumstances, should be reconsidered.