CAREER: Investigating the ecological drivers and consequences of reproductive mode in thermal plasticity in plethodontid salamanders

Safeguarding the resilience of America’s vast natural resources depends on scientific knowledge to help streamline and direct monitoring efforts. Forecast models are critical for prioritizing resources in a fast-paced, changing world, but knowing what data is needed to accurately parameterize these models remains a challenge. For example, wild animal populations are made up of both males and females, and it is generally assumed that both are equally sensitive to environmental changes. However, new research is revealing that this may be an oversimplification. Knowing when and where such vulnerabilities are most likely to arise will be key to increasing the efficiency and efficacy of monitoring operations. To validate findings, the project will leverage machine learning to analyze model outputs, while also applying cutting-edge biotechnology to track the small-bodied amphibians in their natural environment. These efforts will contribute to the generation of large, publicly available ecological monitoring datasets, ideal for AI model training. The project will also advance the education and training of the nation’s future STEM workforce with a new, hands-on research course at Virginia Tech, with datasets also implemented as modules in classes and summer data camps for undergraduates and high schoolers. In each case, students will be trained in the use of Artificial Intelligence to automate, debug, and translate code, enhancing data literacy and teaching foundational concepts in computer programming through engagement in relevant, real-world challenges. The goal of this project is to advance knowledge of the ecological conditions in which differences in thermal plasticity are most likely to arise and when they will be important for population adaptation. Plethodontid salamanders are well suited to these investigations as they exhibit natural variation in trait-specific life histories associated with abiotic conditions, and thermal acclimation capacity has been shown to vary by trait and reproductive condition. This work will provide robust empirical tests of the ecological factors underlying trait-specific thermal plasticity by characterizing and comparing variation in thermal acclimation capacity (Aim 1) and thermoregulatory behavior (Aim 2) across natural elevational gradients. Empirically derived estimates, together with machine learning algorithms, will then be used to construct and analyze mathematical models exploring how variation in individual energetics and population recruitment emerge from realistic patterns of trait-specific plasticity (Aim 3). Finally, empirical patterns and model predictions will be validated in marked, wild populations of salamanders using biotechnology-enabled remote monitoring, which will generate large, publicly-available datasets spanning multiple dimensions of ecological data (Aim 4). Both mechanisms will serve to train the next generation of scientists in use of AI to enhance programming capabilities, data literacy, and quantitative reasoning. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria. NSF Award ID: 2543042 | Program: 01003031DB NSF RESEARCH & RELATED ACTIVIT,01002627DB NSF RESEARCH & RELATED ACTIVIT | Principal Investigator: Jeanette Moss | Institution: Virginia Polytechnic Institute and State University, BLACKSBURG, VA | Award Amount: $1,018,844 View on NSF Award Search: https://www.nsf.gov/awardsearch/show-award/?AWD_ID=2543042 View on Research.gov: https://www.research.gov/awardapi-service/v1/awards/2543042.html