Over the last decade we have seen increasing numbers of reports on catastrophic effects of climate change (warming) on turtles with temperature-dependent sex determination. With most species with TSD displaying a type whereby eggs experiencing higher temperatures producing female biased nests and lower temperatures producing predominantly males, rising temperatures, along with wider oscillations in temperature, could disrupt the ratio of males to females, eventually leading to population collapse and possibly extinction.
"Catastrophic Population Crashes?"
"What's Wrong with Feminization?"
"Could Global Warming Provide Greater Resilience for Turtles?"
"Catastrophic Population Crashes?"
A recently published article in the journal Scientific Reports, An Iowa State University biologist is sounding the alarm for the painted turtle, one of many reptiles for which climate change could prove particularly threatening. Nicole Valenzuela and her coauthors exposed eggs from Iowa to temperatures recorded in nests from three different painted turtle populations in Iowa, Nebraska and Canada from which the proportion of males and females was also recorded. Valenzuela said that allowed the experiments to compare the responses of multiple painted turtle populations, which revealed that not all populations exhibit the same sensitivity to temperature. The research showed that cooler temperature profiles that would tend to produce males trended toward females when the temperature fluctuations intensified. Embryos from warmer profiles, on the other hand, remained female or died when the fluctuations intensified.
"If what we found is generalizable to other species with temperature-dependent sex determination, this is bad news," she said. "If an average increase in temperature is accompanied by greater variance, we'll see populations becoming unisexual faster than anticipated. The greater oscillations add to the effect of just higher average temperature."
Valenzuela said loss of habitat and exploitation has already left many turtles vulnerable to extinction, and climate change only adds to the peril these species face.
"The whole message here is the potential effects climate change can have on these species and the importance of our findings for conservation," she said. "Turtles are the most vulnerable group of vertebrates, and many use temperature-dependent sex determination."
(Source: Iowa State University)
(Source: Iowa State University)
"What's Wrong with Feminization?"
Turtles do face significant threats. Of the 356 species of turtles worldwide, approximately 61% are threatened or already extinct. Turtles are among the most threatened of the major groups of vertebrates, in general, more so than birds, mammals, fishes or even the much besieged amphibians. Reasons for the dire situation of turtles worldwide include the familiar list of impacts to other species including habitat destruction, pest animals, unsustainable overexploitation for pets and food, and climate change. But will potential feminization add further pressure to their existence?
The potential risks of feminization to populations is not a new concept. a simple google search (news) reveals headlines such as
"Turtles Have Survived a Very Long Time"
The common link between all studies is an extrapolation of thermal profiles from a limited number of locations into population level impacts. This is problematic for several reasons. Firstly, populations extend beyond the local area where these thermal profiles were generated. Unless evaluated at landscape (or even global) levels, then habitat heterogeneity will ensure that full feminization across across meta-populations will not occur. If some nesting areas become too hot for embryos or hatchlings to survive, other areas, that were once too cool for incubation, will become available. Considered at a landscape level, current sources may become sinks and vice versa, however, all things being equal, the proportion of male and female producing nests may not significantly change even though sex ratios and survivability of local nesting areas may change considerably.
Secondly, females can adjust nesting behavior to accommodate even if nest site choice is heritable. With warming climates, turtles will start nesting earlier and the timing of the incubation period will change and potentially avoid the extremes that may occur during summer that may affect survival. Similarly, changing climates may significantly change post-nesting habitat, with both vegetation type and quantity, potentially mitigating increases in temperature. Schwanz and Janzen (2008) suggest that it seems unlikely that individual plasticity in the timing of nesting will offset the effects of climate change on sex ratios in painted turtles, however despite an impressive set of long-term data, it is data from a very limited nesting area of one population of painted turtles in Illinois.
"Perhaps Compare Apples with Apples?"
One discussion paper has taken a broader approach to the issue of TSD and the impacts of climate change. Kallimanis (2009) hypothesized that, under stable climatic conditions, populations of some TSD species at the edge of their range are regulated by reduced growth rate (due to skewed sex ratios or due to limited availability of suitable nesting sites). Under climate change, these populations face new climatic conditions that trigger fast population growth (e.g. by more balanced sex ratio, or greater availability of nesting sites). Increased population size may lead to increased dispersal, and thus efficient colonization of the newly created habitat patches. So, the species rapidly tracks the geographical position of its climatic niche. Frequency-dependent selection to maintain 50/50 sex ratios may see them leaving for cooler nesting areas beforehand anyway.
Because of the longevity of turtles, normal immigration/emigration rates may be more than enough to ensure population resilience under changing climate scenarios. The conceptual model proposed Kallimanis (2009) also assumes that population growth rates are maximal when sex ratios are close to 50/50, yet under most population growth scenarios, males are rarely the limiting sex and only likely to become a factor under extreme feminization scenarios- scenarios predicted by studies based on localized nest thermal profiles and warming temperatures. But the point where males become the limiting sex will depend on population life history parameters, not nest thermal temperature profiles, and the "tipping" point is likely to be very close to the upper limits of offspring feminization in long-lived organisms like turtles. ie. Males are generally surplus to population growth rate requirements. Arguments around inbreeding, genetic integrity and diversity are completely valid, but most turtle species have experienced severe bottlenecks in their evolutionary history and given that most turtles live (and breed) for well over 50 years, these impacts may be minimal.
'Population Viability With Increasing Feminization"
So with the worst case scenario that turtles have no capacity to respond to warming climates, what might increasing feminization mean for them? Where is the tipping point where feminization becomes detrimental to population growth?
I created some population models whereby six populations were connected through limited dispersal (1% per year to each population). The turtle life history patterns were loosely based on painted turtle (Chrysemys picta) parameters, which I have used previously (Spencer and Janzen 2010). Males mature at 6yo and females 8yo. For the basic scenarios, females can produce up to 15 eggs per clutch and and 80% of the population produce a second clutch each year. Annual mortality decreases with age with eggs/hatchling mortality rates set at 75%pa (nest predation) and adult mortality set at 5%pa. A standard deviation of 5-10% was imposed for each iteration. Initial population sizes were set at 1000 and carrying capacities were set at 5000 per population. Maximum lifespan was set at 75yo.
The scenarios we compared are shown in Figure 2. I used Vortex 10 PVA software to create each scenario, which had 100 iterations over 200 years. A second round of scenarios was also created where a catastrophe was imposed on each population. Survival rates for each age/stage group declined for one year by 80%. These catastrophes occurred randomly 4 times over 200y.
"Let's Look At The Obvious First"
Alright we have to do it. If we go from 50/50 sex ratios to 100% females (extreme feminization), will populations go extinct? Well yes of course they will. At least some sperm is required to fertilize eggs, although as Jeff Goldblum has told us "Life, uh, finds a way".
But for this purpose, I am going to assume males have some use. As predicted, extreme feminization does lead to population extinction within 75-100 years (Fig. 3).
"Males or Females? What's the Better Sex (for populations)?"
I ran two scenarios to simulate global cooling and global warming. In each of these scenarios, all populations produced 100% offspring of either sex, except for one population that produced 10% offspring of the opposite sex (scenario 2 in Fig. 2). Clearly, females win hands down in terms of population stability. Male dominated offspring scenarios lead to a 50% probability of extinction, which suggests that cooling climates may be more of an issue for species with this form of TSD (Type 1A)- a possible reason why dinosaurs may have died out. It would certainly be informative to evaluate whether periods of cooling led to increased turtle extinctions or population crashes in species with TSD compared to species with GSD.
An Australian Broad Shell Turtle from Australia (GSD Turtle) nesting in March 2019. Original video found here
"Perhaps Compare Apples with Apples?"
One discussion paper has taken a broader approach to the issue of TSD and the impacts of climate change. Kallimanis (2009) hypothesized that, under stable climatic conditions, populations of some TSD species at the edge of their range are regulated by reduced growth rate (due to skewed sex ratios or due to limited availability of suitable nesting sites). Under climate change, these populations face new climatic conditions that trigger fast population growth (e.g. by more balanced sex ratio, or greater availability of nesting sites). Increased population size may lead to increased dispersal, and thus efficient colonization of the newly created habitat patches. So, the species rapidly tracks the geographical position of its climatic niche. Frequency-dependent selection to maintain 50/50 sex ratios may see them leaving for cooler nesting areas beforehand anyway.
Because of the longevity of turtles, normal immigration/emigration rates may be more than enough to ensure population resilience under changing climate scenarios. The conceptual model proposed Kallimanis (2009) also assumes that population growth rates are maximal when sex ratios are close to 50/50, yet under most population growth scenarios, males are rarely the limiting sex and only likely to become a factor under extreme feminization scenarios- scenarios predicted by studies based on localized nest thermal profiles and warming temperatures. But the point where males become the limiting sex will depend on population life history parameters, not nest thermal temperature profiles, and the "tipping" point is likely to be very close to the upper limits of offspring feminization in long-lived organisms like turtles. ie. Males are generally surplus to population growth rate requirements. Arguments around inbreeding, genetic integrity and diversity are completely valid, but most turtle species have experienced severe bottlenecks in their evolutionary history and given that most turtles live (and breed) for well over 50 years, these impacts may be minimal.
Fig. 1. Possible relationship between population growth rate and increasing feminization of offspring.
'Population Viability With Increasing Feminization"
So with the worst case scenario that turtles have no capacity to respond to warming climates, what might increasing feminization mean for them? Where is the tipping point where feminization becomes detrimental to population growth?
I created some population models whereby six populations were connected through limited dispersal (1% per year to each population). The turtle life history patterns were loosely based on painted turtle (Chrysemys picta) parameters, which I have used previously (Spencer and Janzen 2010). Males mature at 6yo and females 8yo. For the basic scenarios, females can produce up to 15 eggs per clutch and and 80% of the population produce a second clutch each year. Annual mortality decreases with age with eggs/hatchling mortality rates set at 75%pa (nest predation) and adult mortality set at 5%pa. A standard deviation of 5-10% was imposed for each iteration. Initial population sizes were set at 1000 and carrying capacities were set at 5000 per population. Maximum lifespan was set at 75yo.
The scenarios we compared are shown in Figure 2. I used Vortex 10 PVA software to create each scenario, which had 100 iterations over 200 years. A second round of scenarios was also created where a catastrophe was imposed on each population. Survival rates for each age/stage group declined for one year by 80%. These catastrophes occurred randomly 4 times over 200y.
Fig. 2. Scenarios created to test for the effects of increasing feminization of offspring that may occur under climate change scenarios. I created six populations linked through populations dispersal. Se ratios of offspring produced each year in each population was manipulated such that population specific sex ratios could be specific. Vortex 10 was used to simulated each scenario to compare parameters such as population growth rate (stochastic- stoch r) and extinction probabilities (PE). red populations represent scenarios where only females are being produced; Blue populations represent scenarios where only males are being produced and; mixed blue/red represent where both males and females are being produced- dominant color indicates dominant sex per clutch.
"Let's Look At The Obvious First"
Alright we have to do it. If we go from 50/50 sex ratios to 100% females (extreme feminization), will populations go extinct? Well yes of course they will. At least some sperm is required to fertilize eggs, although as Jeff Goldblum has told us "Life, uh, finds a way".
Fig. 3. Population sizes (meta) of the six populations where offspring sex ratios were set at 50/50 (blue) and 100% female (red). Over 100 iterations, extinction probabilities were 0% and 100% respectively.
"Males or Females? What's the Better Sex (for populations)?"
I ran two scenarios to simulate global cooling and global warming. In each of these scenarios, all populations produced 100% offspring of either sex, except for one population that produced 10% offspring of the opposite sex (scenario 2 in Fig. 2). Clearly, females win hands down in terms of population stability. Male dominated offspring scenarios lead to a 50% probability of extinction, which suggests that cooling climates may be more of an issue for species with this form of TSD (Type 1A)- a possible reason why dinosaurs may have died out. It would certainly be informative to evaluate whether periods of cooling led to increased turtle extinctions or population crashes in species with TSD compared to species with GSD.
Fig. 4. Population sizes (meta) of the six populations where offspring sex ratios were set at 100% female except for one population where 90% of offspring are female (blue) and 100% male in all populations except for one population where 90% of offspring are male (red). 100 iterations, extinction probabilities were 0% and 50% respectively.
For turtles with TSD type 1A, the risks of extinction are greater in periods of cooling and probably limit distributions at higher latitudes and altitudes, but in periods of warming, population growth rates are likely to become enhanced with increased female offspring production (Fig. 5). But where is the tipping point. I ran the third scenario in Fig. 2, where the levels of male production fluctuated from 0-50% in all populations but also 2-10% in only one population ie. males could only get to other populations through normal dispersal rates.
The implications are significant. From a landscape perspective, individual populations becoming entirely feminized may not lead to any changes in population growth rates if any population within the connected landscape, or metapopulation, is producing a small number of males each year (5-10% males on average per nest). Hence extrapolating population declines or extinctions from nest thermal profiles from limited numbers of nesting beaches is not likely to be valid, no matter how long they have been monitored for. The longevity of turtles provides them resilience and the reproductive potential of adult females ensures population growth rates are maintained (or even enhanced) despite limited numbers of males moving around the network.
"More Females, More Resilience"
More female turtles in populations does provide greater population resilience compared to the GSD-50/50 offspring sex ration scenario (Fig. 6). Populations can respond quicker after a catastrophic event and maintain positive population growth rates (stoch r= +5% compared to -0.02%).
Fig. 6. Simulating catastrophes whereby all survival parameters decrease by 80% randomly for one year four times over 200 years. Population sizes (meta) of the six populations where offspring sex ratios were set at 100% female except for one population where 90% of offspring are female (blue)- no catastrophe; 100% female except for one population where 90% of offspring are female (green)- catastrophe; and 50% female in all populations (red). 100 iterations, extinction probabilities were 0% for all scenarios.
The scenarios above do not make any assumptions that turtles will adjust nesting behaviors in response to changing nesting thermal profiles. They also make no assumptions that extra dispersal will occur as population numbers increase or part of behavioral shifts includes moving to new populations in response to changing nesting thermal profiles. But they are informative to demonstrate that turtles can handle total feminization of most populations as long as there are a small number of males being produced somewhere across the landscape. Populations also demonstrated increased resilience, although a caveat here is that allele frequency is approximately 20-40 times greater in populations with 50/50 offspring sex ratio scenarios. Given the longevity of turtles and considering most populations have regularly gone through bottlenecks in their evolutionary history, "quantity" over "quality" probably helps ensure turtles avoid extinction when populations crash. Hence an army of available female turtles ensures that reproductive output is maintained. Besides, evolutionary theory predicts that dioecious species should produce a balanced primary sex ratio maintained by frequency-dependent selection, hence imbalances in population sex ratios may drive females to disperse throughout metapopulations to source male producing nesting habitat that will not be available at local nesting areas under warming climate scenarios.
I have not really touched on the role of selection on maternal nest site choice. There is significant high quality research that has spanned 30-40 years looking at the role of TSD in driving maternal nest site behavior, as well as, the evolution of pivotal temperatures. Yet any response of these four sex-determining traits (maternal nesting behavior, pivotal temperatures, nesting phenology, and nest depth) to changing climates will also mitigate catastrophic impacts of nest feminization and mortality. This is not downplaying the impacts of climate change on turtles. Turtles face a number of challenges to avoid extinction and factors like climate change and human development will significantly impact their habitats into the future. With limited resourcing and funding, as well as, government apathy or will, we need to prioritise species and populations. Part of the prioritization process includes ranking and mitigating threats. The question is whether feminization of nesting grounds is a threat to a species? From a population perspective, these models are worst case scenarios and suggest that feminzation is not a threat to turtles with TSD. Mass nest mortality on the other-hand, will impact populations, but again, the full impacts must be evaluated from a landscape of meta-population perspective.
This blog has not been through any form of peer review and maybe I have made it all too simplistic and I welcome your comments below. I also welcome any collaboration to turn this into a peer-reviewed publication (if it is worthwhile). Contact me at r.spencer@uws.edu.au
Fig. 5. (a) Population sizes (meta) of the six populations where offspring sex ratios were set at 100% female in all populations (green); 50% female in all populations (red); 100% in all populations except for one population where 90% of offspring are female (black); 100% in all populations except for one population where 95% of offspring are female (purple); 100% in all populations except for one population where 98% of offspring are female (blue).100 iterations, extinction probabilities were 0% for all simulations except for the 100% female scenario (100% PE) and the 98% female one population scenario (10% PE). (b) Similar scenarios- 50% female in all populations (blue); 100% in all populations except for one population where 90% of offspring are female (green); 100% in all populations except for one population where 98% of offspring are female (red) -- but nest predation was set at higher rates of 85% (SD 15%).
"More Females, More Resilience"
More female turtles in populations does provide greater population resilience compared to the GSD-50/50 offspring sex ration scenario (Fig. 6). Populations can respond quicker after a catastrophic event and maintain positive population growth rates (stoch r= +5% compared to -0.02%).
Fig. 6. Simulating catastrophes whereby all survival parameters decrease by 80% randomly for one year four times over 200 years. Population sizes (meta) of the six populations where offspring sex ratios were set at 100% female except for one population where 90% of offspring are female (blue)- no catastrophe; 100% female except for one population where 90% of offspring are female (green)- catastrophe; and 50% female in all populations (red). 100 iterations, extinction probabilities were 0% for all scenarios.
"What does it all mean?"
The effects of feminization at an individual nest level are significant but those effects are dampened when looked at the level of (meta) population. Population dynamics and growth rates are affected by smaller changes in other life history parameters. For example, the tipping point for a population appears >98% feminization of nests in a meta-population, however, if you lower nest predation rates by 10% or adult mortality by far less, the effects of feminization at a population level are overcome (Fig. 7).
Are turtles with TSD better equipped to handle warming ground temperatures? Some nesting grounds are likely to become "death traps", but other nesting grounds that were too cool for incubation will become available- classic source-sink scenarios within metapopulations. The main question is whether there is enough connectivity in systems to allow turtles to move around and this will need to be assessed on species by species basis. Agriculture may limit their ability to disperse and potentially increase their risks of mortality. It may also homogenize nesting habitats, which in turn will homogenize nest thermal profiles at a landscape level. Thus increasing the risks of complete broad-scale feminization of offspring in these areas.
The effects of feminization at an individual nest level are significant but those effects are dampened when looked at the level of (meta) population. Population dynamics and growth rates are affected by smaller changes in other life history parameters. For example, the tipping point for a population appears >98% feminization of nests in a meta-population, however, if you lower nest predation rates by 10% or adult mortality by far less, the effects of feminization at a population level are overcome (Fig. 7).
Fig. 5. Population sizes (meta) of the six populations where offspring sex ratios were set at 100% female in 5 populations and 98% in one population with normal (75% blue) and low (65% red) nest predation rates. 100 iterations, extinction probabilities were 30% and 0% respectively.
Are turtles with TSD better equipped to handle warming ground temperatures? Some nesting grounds are likely to become "death traps", but other nesting grounds that were too cool for incubation will become available- classic source-sink scenarios within metapopulations. The main question is whether there is enough connectivity in systems to allow turtles to move around and this will need to be assessed on species by species basis. Agriculture may limit their ability to disperse and potentially increase their risks of mortality. It may also homogenize nesting habitats, which in turn will homogenize nest thermal profiles at a landscape level. Thus increasing the risks of complete broad-scale feminization of offspring in these areas.
The scenarios above do not make any assumptions that turtles will adjust nesting behaviors in response to changing nesting thermal profiles. They also make no assumptions that extra dispersal will occur as population numbers increase or part of behavioral shifts includes moving to new populations in response to changing nesting thermal profiles. But they are informative to demonstrate that turtles can handle total feminization of most populations as long as there are a small number of males being produced somewhere across the landscape. Populations also demonstrated increased resilience, although a caveat here is that allele frequency is approximately 20-40 times greater in populations with 50/50 offspring sex ratio scenarios. Given the longevity of turtles and considering most populations have regularly gone through bottlenecks in their evolutionary history, "quantity" over "quality" probably helps ensure turtles avoid extinction when populations crash. Hence an army of available female turtles ensures that reproductive output is maintained. Besides, evolutionary theory predicts that dioecious species should produce a balanced primary sex ratio maintained by frequency-dependent selection, hence imbalances in population sex ratios may drive females to disperse throughout metapopulations to source male producing nesting habitat that will not be available at local nesting areas under warming climate scenarios.
I have not really touched on the role of selection on maternal nest site choice. There is significant high quality research that has spanned 30-40 years looking at the role of TSD in driving maternal nest site behavior, as well as, the evolution of pivotal temperatures. Yet any response of these four sex-determining traits (maternal nesting behavior, pivotal temperatures, nesting phenology, and nest depth) to changing climates will also mitigate catastrophic impacts of nest feminization and mortality. This is not downplaying the impacts of climate change on turtles. Turtles face a number of challenges to avoid extinction and factors like climate change and human development will significantly impact their habitats into the future. With limited resourcing and funding, as well as, government apathy or will, we need to prioritise species and populations. Part of the prioritization process includes ranking and mitigating threats. The question is whether feminization of nesting grounds is a threat to a species? From a population perspective, these models are worst case scenarios and suggest that feminzation is not a threat to turtles with TSD. Mass nest mortality on the other-hand, will impact populations, but again, the full impacts must be evaluated from a landscape of meta-population perspective.
This blog has not been through any form of peer review and maybe I have made it all too simplistic and I welcome your comments below. I also welcome any collaboration to turn this into a peer-reviewed publication (if it is worthwhile). Contact me at r.spencer@uws.edu.au
Comments
Post a Comment