Project discussion notes
Team theme, tentative area of interest, or question:
Wildfires, post-disturbance reorganization, zombie forests
Day 1: March 12, 2024 - CU Boulder
Selected scientific question:
What variables are best at predicting whether an area affected by wildfire in the Southern Rockies will recover as a forest or transition into shrub or grassland?
Day 2: March 13, 2024 - CU Boulder
- Established data sources: CBI, LCMAP, CU Boulder Disturbance Stack, CHELSA
- Loaded/downloaded data, and established the data processing pipeline and end product: a tabular dataset with each pixel relating to a disturbance event and the corresponding features, with the ground truth label being 1 = recovered to forest and 0 = recovered to shrub/grassland (after a 10 year period)
- preparing analysis using XGBoost and SHAP analysis for feature importance ranking
Day 3: March 14, 2024 - Virtual due to snow :(
Last update:
2024-05-09