Towards Spatio-temporally Consistent Multi-Site Fire Danger Downscaling with Explainable Deep Learning
Our study analyzes the ability of state-of-the-art CNN and ConvLSTM-based machine learning methods to model the multivariate spatio-temporal structure of the Fire Weather Index (FWI). Authors and corresponding ORCID can be found in the zenodo.json file.
2023_Mirones_FWI_ERL.ipynb is a Jupyter notebook based on the R languaje containing the code necessary to replicate our main results.
environment.yml contains the versions of the python and R libraries employed to reproduce the results of the manuscript. A conda environment with the appropriate versions can be created by typing:
mamba env create -n deep-fwi --file environment.yml