Brackish Tidal Marsh Management and the Ecology of a Declining Freshwater Turtle.

Water management practices in tidal marshes of the San Francisco Bay Estuary, California are often aimed at increasing suitable habitat for threatened fish species and sport fishes. However, little is known about how best to manage habitat for other sensitive status species like the semiaquatic freshwater Western Pond Turtle (Actinemys marmorata) that is declining throughout much of its range. Here, we examined the basking activity, abundance, survival, and growth of Western Pond Turtles at two brackish water study sites in Suisun Marsh, California that differed in how they were managed, with one having passive management (i.e., no active water regulation) and another having active management (i.e., water regulated for seasonal hunting). Our results revealed that basking activity was greatest when salinity, water stage, and air temperatures were low, shortwave radiation was high, and wind levels were intermediate. These preferred habitat characteristics often reflected conditions that were naturally maintained at the passively managed, muted tidal site. We also found that turtles were more abundant and had higher survival rates in the passively managed habitat compared to the actively managed habitat (201-323 turtles/km2 and 96% survival versus 11-135 turtles/km2 and 77% survival, respectively). Finally, characteristic growth constants from von Bertalanffy models showed that turtles grew more quickly in passively managed habitat compared to the actively managed habitat. Our results suggest that management strategies for this sensitive status species may be more effective if they protect passively managed muted tidal systems that limit or delay extreme cycles of salinity and water levels and conserve elevated terrestrial buffer zones adjacent to muted and full tidal systems.

Physiological consequences of rising water salinity for a declining freshwater turtle.

Sea-level rise, drought and water diversion can all lead to rapid salinization of freshwater habitats, especially in coastal areas. Increased water salinities can in turn alter the geographic distribution and ecology of freshwater species including turtles. The physiological consequences of salinization for freshwater turtles, however, are poorly known. Here, we compared the osmoregulatory response of two geographically separate populations of the freshwater Western Pond Turtle (Actinemys marmorata)-a species declining across its range in western North America-to three constant salinities: 0.4 ppt, 10 ppt and 15 ppt over 2 weeks. We found that turtles from a coastal estuarine marsh population regulated their plasma osmolality at lower levels than their conspecifics from an inland freshwater creek population 45 km away. Plasma osmolalities were consistently lower in estuarine marsh turtles than the freshwater creek turtles over the entire 2-week exposure to 10 ppt and 15 ppt water. Furthermore, estuarine marsh turtles maintained plasma osmolalities within 1 SD of their mean field osmolalities over the 2-week exposure, whereas freshwater creek turtles exceeded their field values within the first few days after exposure to elevated salinities. However, individuals from both populations exhibited body mass loss in 15 ppt water, with significantly greater loss in estuarine turtles. We speculate that the greater ability to osmoregulate by the estuarine marsh turtles may be explained by their reduced feeding and drinking in elevated salinities that was not exhibited by the freshwater creek population. However, due to mass loss in both populations, physiological and behavioural responses exhibited by estuarine marsh turtles may only be effective adaptations for short-term exposures to elevated salinities, such as those from tides and when traversing saline habitats, and are unlikely to be effective for long-term exposure to elevated salinity as is expected under sea-level rise.

Salinity tolerances and use of saline environments by freshwater turtles: implications of sea level rise.

The projected rise in global mean sea levels places many freshwater turtle species at risk of saltwater intrusion into freshwater habitats. Freshwater turtles are disproportionately more threatened than other taxa; thus, understanding the role of salinity in determining their contemporary distribution and evolution should be a research priority. Freshwater turtles are a slowly evolving lineage; however, they can adapt physiologically or behaviourally to various levels of salinity and, therefore, temporarily occur in marine or brackish environments. Here, we provide the first comprehensive global review on freshwater turtle use and tolerance of brackish water ecosystems. We link together current knowledge of geographic occurrence, salinity tolerance, phylogenetic relationships, and physiological and behavioural mechanisms to generate a baseline understanding of the response of freshwater turtles to changing saline environments. We also review the potential origins of salinity tolerance in freshwater turtles. Finally, we integrate 2100 sea level rise (SLR) projections, species distribution maps, literature gathered on brackish water use, and a phylogeny to predict the exposure of freshwater turtles to projected SLR globally. From our synthesis of published literature and available data, we build a framework for spatial and phylogenetic conservation prioritization of coastal freshwater turtles. Based on our literature review, 70 species (∼30% of coastal freshwater turtle species) from 10 of the 11 freshwater turtle families have been reported in brackish water ecosystems. Most anecdotal records, observations, and descriptions do not imply long-term salinity tolerance among freshwater turtles. Rather, experiments show that some species exhibit potential for adaptation and plasticity in physiological, behavioural, and life-history traits that enable them to endure varying periods (e.g. days or months) and levels of saltwater exposure. Species that specialize on brackish water habitats are likely to be vulnerable to SLR because of their exclusive coastal distributions and adaptations to a narrow range of salinities. Most species, however, have not been documented in brackish water habitats but may also be highly vulnerable to projected SLR. Our analysis suggests that approximately 90% of coastal freshwater turtle species assessed in our study will be affected by a 1-m increase in global mean SLR by 2100. Most at risk are freshwater turtles found in New Guinea, Southeast Asia, Australia, and North and South America that may lose more than 10% of their present geographic range. In addition, turtle species in the families Chelidae, Emydidae, and Trionychidae may experience the greatest exposure to projected SLR in their present geographic ranges. Better understanding of survival, growth, reproductive and population-level responses to SLR will improve region-specific population viability predictions of freshwater turtles that are increasingly exposed to SLR. Integrating phylogenetic, physiological, and spatial frameworks to assess the effects of projected SLR may improve identification of vulnerable species, guilds, and geographic regions in need of conservation prioritization. We conclude that the use of brackish and marine environments by freshwater turtles provides clues about the evolutionary processes that have prolonged their existence, shaped their unique coastal distributions, and may prove useful in predicting their response to a changing world.

The evolution of different maternal investment strategies in two closely related desert vertebrates.

We compared egg size phenotypes and tested several predictions from the optimal egg size (OES) and bet-hedging theories in two North American desert-dwelling sister tortoise taxa, Gopherus agassizii and G. morafkai, that inhabit different climate spaces: relatively unpredictable and more predictable climate spaces, respectively. Observed patterns in both species differed from the predictions of OES in several ways. Mean egg size increased with maternal body size in both species. Mean egg size was inversely related to clutch order in G. agassizii, a strategy more consistent with the within-generation hypothesis arising out of bet-hedging theory or a constraint in egg investment due to resource availability, and contrary to theories of density dependence, which posit that increasing hatchling competition from later season clutches should drive selection for larger eggs. We provide empirical evidence that one species, G. agassizii, employs a bet-hedging strategy that is a combination of two different bet-hedging hypotheses. Additionally, we found some evidence for G. morafkai employing a conservative bet-hedging strategy. (e.g., lack of intra- and interclutch variation in egg size relative to body size). Our novel adaptive hypothesis suggests the possibility that natural selection favors smaller offspring in late-season clutches because they experience a more benign environment or less energetically challenging environmental conditions (i.e., winter) than early clutch progeny, that emerge under harsher and more energetically challenging environmental conditions (i.e., summer). We also discuss alternative hypotheses of sexually antagonistic selection, which arise from the trade-offs of son versus daughter production that might have different optima depending on clutch order and variation in temperature-dependent sex determination (TSD) among clutches. Resolution of these hypotheses will require long-term data on fitness of sons versus daughters as a function of incubation environment, data as yet unavailable for any species with TSD.

Mass mortality of eastern box turtles with upper respiratory disease following atypical cold weather.

Emerging infectious diseases cause population declines in many ectotherms, with outbreaks frequently punctuated by periods of mass mortality. It remains unclear, however, whether thermoregulation by ectotherms and variation in environmental temperature is associated with mortality risk and disease progression, especially in wild populations. Here, we examined environmental and body temperatures of free-ranging eastern box turtles Terrapene carolina during a mass die-off coincident with upper respiratory disease. We recorded deaths of 17 turtles that showed clinical signs of upper respiratory disease among 76 adult turtles encountered in Berea, Kentucky (USA), in 2014. Of the 17 mortalities, 11 occurred approximately 14 d after mean environmental temperature dropped 2.5 SD below the 3 mo mean. Partial genomic sequencing of the major capsid protein from 1 sick turtle identified a ranavirus isolate similar to frog virus 3. Turtles that lacked clinical signs of disease had significantly higher body temperatures (23°C) than sick turtles (21°C) during the mass mortality, but sick turtles that survived and recovered eventually warmed (measured by temperature loggers). Finally, there was a significant negative effect of daily environmental temperature deviation from the 3 mo mean on survival, suggesting that rapid decreases in environmental temperature were correlated with mortality. Our results point to a potential role for environmental temperature variation and body temperature in disease progression and mortality risk of eastern box turtles affected by upper respiratory disease. Given our findings, it is possible that colder or more variable environmental temperatures and an inability to effectively thermoregulate are associated with poorer disease outcomes in eastern box turtles.

Turbines and Terrestrial Vertebrates: Variation in Tortoise Survivorship Between a Wind Energy Facility and an Adjacent Undisturbed Wildland Area in the Desert Southwest (USA).

With the recent increase in utility-scale wind energy development, researchers have become increasingly concerned how this activity will affect wildlife and their habitat. To understand the potential impacts of wind energy facilities (WEF) post-construction (i.e., operation and maintenance) on wildlife, we compared differences in activity centers and survivorship of Agassiz's desert tortoises (Gopherus agassizii) inside or near a WEF to neighboring tortoises living near a wilderness area (NWA) and farther from the WEF. We found that the size of tortoise activity centers varied, but not significantly so, between the WEF (6.25 ± 2.13 ha) and adjacent NWA (4.13 ± 1.23 ha). However, apparent survival did differ significantly between the habitat types: over the 18-year study period apparent annual survival estimates were 0.96 ± 0.01 for WEF tortoises and 0.92 ± 0.02 for tortoises in the NWA. High annual survival suggests that operation and maintenance of the WEF has not caused considerable declines in the adult population over the past two decades. Low traffic volume, enhanced resource availability, and decreased predator populations may influence annual survivorship at this WEF. Further research on these proximate mechanisms and population recruitment would be useful for mitigating and managing post-development impacts of utility-scale wind energy on long-lived terrestrial vertebrates.

Using motion-sensor camera technology to infer seasonal activity and thermal niche of the desert tortoise (Gopherus agassizii).

Understanding the relationships between environmental variables and wildlife activity is an important part of effective management. The desert tortoise (Gopherus agassizii), an imperiled species of arid environments in the southwest US, may have increasingly restricted windows for activity due to current warming trends. In summer 2013, we deployed 48 motion sensor cameras at the entrances of tortoise burrows to investigate the effects of temperature, sex, and day of the year on the activity of desert tortoises. Using generalized estimating equations, we found that the relative probability of activity was associated with temperature (linear and quadratic), sex, and day of the year. Sex effects showed that male tortoises are generally more active than female tortoises. Temperature had a quadratic effect, indicating that tortoise activity was heightened at a range of temperatures. In addition, we found significant support for interactions between sex and day of the year, and sex and temperature as predictors of the probability of activity. Using our models, we were able to estimate air temperatures and times (days and hours) that were associated with maximum activity during the study. Because tortoise activity is constrained by environmental conditions such as temperature, it is increasingly vital to conduct studies on how tortoises vary their activity throughout the Sonoran Desert to better understand the effects of a changing climate.