AI for Climate - Forest Fires

For more information see http://www.aivcc.uos.de <http://www.aivcc.uos.de/wordpress/index.php/about/>_.

Motivation

To predict monthly forest fire counts per Modis tile do the following:

For training and validation on ERA5 data run:

$ python preprocess_for_train.py
$ python train_val_regressor.py --mode val

For predicting with CMIP projections run:

$ python preprocess_for_prediction.py
$ python train_val_regressor.py --mode pred
$ python rf_prediction.py
$ python post_process_prediction.py

Name		Name	Last commit message	Last commit date
Latest commit History 50 Commits
datasets		datasets
figures		figures
.gitignore		.gitignore
README.md		README.md
data_visualisation.ipynb		data_visualisation.ipynb
extract_era_id_from_csv.ipynb		extract_era_id_from_csv.ipynb
grouping.PNG		grouping.PNG
model.py		model.py
plot_feature_importances.ipynb		plot_feature_importances.ipynb
plot_fire_count_tile_on_map.ipynb		plot_fire_count_tile_on_map.ipynb
plot_on_map.ipynb		plot_on_map.ipynb
plot_temporal_evolution_fire_count.ipynb		plot_temporal_evolution_fire_count.ipynb
postprocess_prediction.py		postprocess_prediction.py
prepare_cmip6.ipynb		prepare_cmip6.ipynb
preprocess_for_prediction.py		preprocess_for_prediction.py
preprocess_for_train.py		preprocess_for_train.py
requirements.yml		requirements.yml
rf_prediction.py		rf_prediction.py
train_val_regressor.py		train_val_regressor.py