This repository contains all the code for processing the data as well as code for the DL and RF models used in the paper Otero and Horton, 2022: https://eartharxiv.org/repository/view/3447/
A total of 6 deep learning architectures are tested to predict both, precipitation amounts and extreme precipitation events (>95th and >99th). A baseline model is also include to benchmark the DL performance.
- DNN_models_comparison.ipynb: Contains the DeepFactory class with all the models architectures.
- Random_Forest.ipynb: Runs RF regressor and classifier point-wise.
In addition, a layer-wise-relevance propagation (LRP) is applied to assess the importance of the predictors.
The data used for this study can be download from https://cds.climate.copernicus.eu/cdsapp\#!/dataset/reanalysis-era5-pressure-levels.