Title: "Integration of Machine Learning and Remote Sensing Data for Evaluating Climate Extremes in the Agricultural Production System of Prince Edward Island"

Abstract: 

Climate change threatens global agriculture by altering weather patterns and intensifying extreme events, posing challenges for Prince Edward Island (PEI), Canada, where rainfed potato farming is crucial. Potatoes, highly sensitive to climatic variability, require precise monitoring to enhance resilience. This study utilizes high-resolution satellite remote sensing to assess climate impacts on PEI potato crops across four fields (2021–2022). Multispectral data from Landsat 8, PlanetScope, and Sentinel-2A estimated evapotranspiration (ETc), soil moisture, drought conditions, and yield. Sentinel-2A showed strong correlations between NDVI and crop coefficient (Kc), with ETc peaking at 4.0 mm/day (2021) and 3.7 mm/day (2022). Soil moisture indices derived from Landsat 8 and MODIS correlated well with field data, confirming reduced moisture under high temperatures. Long-term drought analysis (2012–2022) identified 2020 as the most severe drought year, with significant seasonal droughts in June 2021 and 2022. Machine learning models, particularly gradient tree boosting (GTB), effectively predicted yield (R² = 0.71–0.78, RMSE = 2.82–5.96 t/ha). This study underscores the value of remote sensing and ML in optimizing water use, managing drought, and improving yield prediction for sustainable agriculture in PEI.

Friday March 21, 2025, 9:00 am, AVC 278N

Everyone is welcome.