Location: Pest Management Research
Project Number: 3032-22000-019-002-A
Project Type: Cooperative Agreement
Start Date: Jul 22, 2024
End Date: Apr 30, 2025
Objective:
The goal of this project is to use Geospatial Artificial Intelligence (GeoAI) to successfully forecast Mormon cricket populations
This project is intended to achieve two objectives:
1) Evaluate the hypothesis that topographically variable sites are centers of endemism for Mormon crickets to recolonize more homogeneous areas of local extinction, and
2) Develop a GeoAI-based, uncertainty-aware method for outbreak forecasting of Mormon crickets.
Successful completion of this one year project will produce a forecasting application constructed as a web-based toolbox, software package, or library for continued use in outbreak forecasting beyond the lifecycle of the project.
Approach:
Mormon crickets at high elevation will develop and hatch over at least an eight-year period and likely much longer, whereas they require two to three years on average at low elevation (R. Srygley, unpublished data). Consequently, soil temperature is a critical factor in determining the persistence of Mormon cricket eggs in egg banks. Because mountains and canyons provide a patchy thermal environment, multiple generations of eggs are likely to be banked in the soil in these regions. Unusual warm events in mountain or canyonlands will rush the development of eggs that have accumulated in egg banks resulting in Mormon cricket outbreaks, banding, and migration from mountain and canyonlands the following spring. With the advances of geographic information science, earth observations and climatic modeling technologies, high resolution spatiotemporal datasets of topographic, environmental, and climatic variables are increasingly available. These spatiotemporal datasets, together with the advances of artificial intelligence (AI) and machine learning algorithms, make it possible to characterize the impact of the spatiotemporal variables on Mormon cricket population dynamics, model the species distribution, and forecast the risk of future outbreaks. In this proposal, we aim to develop uncertainty-aware GeoAI-based machine learning methods to integrate the long-term collections of Mormon cricket data and high resolution of spatiotemporal datasets to model and map the spatiotemporal distribution of Mormon crickets and identify areas where centers of endemism and/or egg banks are likely sources for recruits.
