As egg-laying reptiles, sea turtles depend on the temperature of the soil to determine the sex of their hatchlings. However, with climate change expected to raise temperatures globally, the sex ratios of their offspring are at risk of becoming imbalanced, threatening their populations. To analyze the potential impacts of climate change on sea turtle hatchlings, Fuentes and Porter (2013) used the Niche Mapper™ microclimate model to project soil temperatures at three green turtle nesting sites in the northern Great Barrier Reef under various emissions scenarios. They compared their results to existing correlative models, as well as to observational data of current climate conditions. The authors found that because the microclimate model required more data inputs than the correlative models, it produced more detailed predictions. However, both ultimately reached the same overall prediction that under an extreme emissions scenario, temperature increases are likely to skew sex ratios to favor female hatchlings as early as 2030. By 2070, turtle populations may become exclusively female or at worst, will no longer be able to incubate.—Katie Huang
Fuentes, M.M.P.B., Porter, W.P., 2013. Using a microclimate model to evaluate impacts of climate change on sea turtles. Ecological Modeling 251, 150–157.
Fuentes and Porter used the Niche Mapper™ microclimate model to estimate soil temperatures at Bramble Cay, Raine Island, and Sandbank 7, three green turtle nesting grounds in the northern Great Barrier Reef. Temperatures were projected under current climate conditions as well as under conservative and extreme emissions scenarios of climate change for 2030 and 2070. The two major data inputs required for the model were climate maximum and minimum data and physical properties of the soil. To calculate the climate inputs, the authors took hourly measurements of unshaded nesting grounds from November 2007 to March 2008. Variables such as daily air temperature, relative humidity, and wind speed were obtained from various data sources for the three sites. Soil properties were taken from samples of nesting beach dunes from November 2007. The authors calculated their reflexivity properties and ran a sensitivity analysis in order to determine their potential effect on soil temperature. Other soil properties such as thermal conductivity ranges and specific heat values were taken from other data sources. Fuentes and Porter also compacted the existing model to allow users to input conditions such as variations in soil water level or snow deposition that would affect hatchlings. In order to simulate projected increases in temperature, they increased the minimum and maximum air temperatures in the microclimate submodel by projected seasonal increments. To apply the temperature data to sea turtle hatchling sex ratios, they assumed sand temperatures at 29.3°C produced a 1:1 sex ratio, while temperatures below 27.8°C produced all males and temperatures above 30.8°C produced all females. They then assumed that the proportion of females increased linearly between 27.8 and 30.8°C. In order to analyze the accuracy of the microclimate model, the authors compared their results with those of correlative models comparing air and sea surface temperatures with soil temperature. They then ran the microclimate and correlative models for the current climate scenario and compared those results with observational data.
The temperatures constructed with Niche Mapper™ were not significantly different from observational data. The model was able to explain over 58% of variation in sand temperature at each of the nesting grounds, making it an appropriate model for future climate change scenarios if applied with caution. Based on soil temperature projections at 50 cm, Bramble Cay was the nesting site most susceptible to warming, while Sandbank 7 was the least. Under an extreme emissions scenario, Bramble Cay would produce mainly female hatchlings by 2030, while the other two sites would still produce male and female hatchlings. By 2070, Bramble Cay may reach temperatures near the upper thermal threshold for incubation, while Sandbank 7 would produce mainly female offspring. The 2070 projections for Raine Island differed by model, with the microclimate model predicting only female offspring and the correlative model suggesting that a small proportion of male offspring would still be produced. However, these results do not account for other influential factors such as nesting behavior, embryonic development, and potential capacity to adapt. Since sea turtles lay eggs at various depths and nest at different beach locations, they may learn to adapt to temperature changes in the future. Also, the models only predicted temperatures for one depth, leading to an incomplete prediction of future incubating environments. The authors suggest that temperatures at various depths need to be modeled in order to create more accurate predictions of sea turtle hatchling sex ratios.
Both the microclimate and correlative models were able to explain the variability in sand temperature at the three nesting grounds. However, the microclimate model is more useful in determining the suitability of nesting environments because it creates hourly projections at multiple specified depths. While the correlative model produced strong correlations between the mean monthly observed and modeled temperatures, it did not clearly illustrate the effects of daily fluctuations. The differences can be attributed to the amount of data inputted into each model: the microclimate dataset is more extensive and thus produces more detailed results. Furthermore, the microclimate model is capable of projecting soil temperatures at multiple depths simultaneously, while separate correlative models are needed for each desired depth. With further use, the microclimate model may be refined to help implement more effective short-term management strategies by creating site-specific approaches. However, both models are capable of producing similar range projections, with differences varying from 0.04 to 1.32°C. Despite the variation, the models ultimately reached the same projective implications.