Down To Earth: A podcast for Geoscientists by Geoscientist Down To Earth: Remote Sensing and Forest Fires
4 snips
Jan 14, 2021 Morgan Crowley, a PhD candidate at McGill University, dives into her innovative research on mapping forest fires using free satellite software. She discusses the impact of climate change on fire seasons in Canada and highlights techniques like the Bayesian updating algorithm for real-time fire monitoring. Crowley also emphasizes the importance of inclusivity in STEM, sharing insights on mentorship for women in geoscience. Her work aims to integrate ecological data for better forest management and improve emergency response to wildfires.
AI Snips
Chapters
Transcript
Episode notes
Bayesian Multi‑Source Fire Mapping
- Morgan Crowley developed a multi-source, per-pixel Bayesian updating algorithm to map fires as they burn.
- The method stacks imagery, weights evidence for each pixel, and runs efficiently in Google Earth Engine versus days-long R runs.
Combine Sensors For Faster Fire Response
- Provide emergency responders the richest possible imagery by combining sensors rather than relying on one source.
- Combining MODIS, VIIRS, Landsat and others highlights each sensor's strengths for faster, actionable burned-area maps.
From Hand‑Drawn Maps To Automated Fire Pipelines
- Morgan moved from qualitative GIS in the US to forest monitoring in Google Earth Engine after seeing limited, hand-drawn and weeks-old fire maps.
- That experience motivated a PhD partnership with the Canadian Forest Service to build multi-source fire pipelines.