STAR: Forecasting populations for conservation: The role of life history and model structure on forecast accuracy

Population forecasts are used to manage threatened and endangered wildlife populations. Natural resource managers use these predictions to anticipate future changes in species abundance, assess extinction risk, and prioritize management interventions. Inaccurate forecasts may lead to erroneous interventions or inefficient uses of limited resources including funding and personnel. Despite the widespread adoption of population forecasts over the last four decades, there have been few efforts to assess the historical performance of these predictions. This project will use the growing number of long-term monitoring datasets collected by research scientists and state and federal agencies to assess the forecast performance of population models. Findings will inform management strategies for threatened populations by identifying the types of data and models that generate accurate forecasts. Outcomes of this project include improved guidance for natural resource managers on effective monitoring strategies for threatened populations, and the development of a framework that can be applied to evaluate other historical ecological forecasts. The research will train the next generation of scientists with modeling and programing skills, handling and development of databases, and the development of AI-ready databases for the scientific community. Population ecologists have been making predictions on the risk of population decline and extinction for almost 40 years. While there has been some past work evaluating forecast ability in stable populations, most assessments of population viability forecasts have been through indirect methods, thus, there is little empirical evidence assessing the long-term accuracy of these forecasts. This project will apply a retrospective approach to assess the reliability of population predictions by developing a publicly available database of historical population viability forecasts linked to updated monitoring data of vertebrates, invertebrates, and plants. This database will be an asset for natural resource managers and scientists studying the properties of ecological forecasts, while the analysis will provide real-world measures of forecast accuracy and precision by comparing predicted trends to updated monitoring data. The project will determine how the life history of target species interacts with statistical survey methods and demographic model details to influence forecast skill. This project provides the first comprehensive evaluation of published population forecasts using monitoring data. 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: 2531659 | Program: 01002627DB NSF RESEARCH & RELATED ACTIVIT | Principal Investigator: Jake Ferguson | Institution: University of Kentucky Research Foundation, LEXINGTON, KY | Award Amount: $400,000 View on NSF Award Search: https://www.nsf.gov/awardsearch/show-award/?AWD_ID=2531659 View on Research.gov: https://www.research.gov/awardapi-service/v1/awards/2531659.html