Read-only mirror of https://github.com/MeteoSwiss/c4dl-lightningdl — MeteoSwiss. Issues & pull requests at the source.
Find a file
Repository files (latest commit first)
Filename Latest commit message Latest commit date
2022-11-22 17:01:09 +01:00
c4dllightning Fix URL in setup.py 2022-10-03 14:25:03 +02:00
data Initial commit 2022-01-20 15:05:04 +01:00
figures Initial commit 2022-01-20 15:05:04 +01:00
models Initial commit 2022-01-20 15:05:04 +01:00
results Updates before release 2022-03-10 17:09:55 +01:00
scripts Deleted unnecessary lines from training.py 2022-11-22 17:01:09 +01:00
.gitignore Initial commit 2022-01-20 15:05:04 +01:00
LICENSE Add license 2022-03-14 16:36:21 +01:00
README.md Update paper status in README.md 2022-10-03 14:44:54 +02:00
setup.py Fix URL in setup.py 2022-10-03 14:25:03 +02:00

This repository contains the machine learning code used in the paper: Multi-source seamless lightning nowcasting with recurrent deep learning, accepted to Artificial Intelligence for the Earth Systems. Preprint available: https://arxiv.org/abs/2203.10114

Installation

You need NumPy, Scipy, Matplotlib, Tensorflow, Numba, Dask and NetCDF4 for Python.

Clone the repository, then, in the main directory, run

$ python setup.py develop

(if you plan to modify the code) or

$ python setup.py install

if you just want to run it.

Downloading data

The dataset can be found at the following Zenodo repository: https://doi.org/10.5281/zenodo.6325370

Download the NetCDF file. You can place it in the data directory but elsewhere on the disk works too.

Running

Go to the scripts directory and start an interactive shell. There, you can find training.py that contains the script you need for training and plots_lightning.py that produces the plots from the paper.