Where to start?

The following 5 Jupyter notebooks that allow to familiarise with the tasks of:

  1. Visualize the Flood Event (3arcmin or 1 arcmin from CaMa-Flood, give a flood_case name and define an area)
    • /home/pad/Notebooks/Inundation-Forecast.ipynb
  2. Extract Sentinel-1 EO data (20m from GFM. 375m from VIIRS, use the same flood_case name and area )
    • /home/pad/Notebooks/Extract_GFM_Inundation.ipynb
      • Conda environment yaml neeeded for GFM : /home/mocb/gfm_env.yaml
    • /home/pad/Notebooks/Extract_VIIRS_Inundation.ipynb
  3. Benchmark the CMF Model with GFM and with VIIRS Observations (use the same flood_case name)
    • /home/pad/Notebooks/Bench_CMF_GFM_Inundation.ipynb
    • /home/pad/Notebooks/Bench_CMF_VIIRS_Inundation.ipynb
    • /home/pad/Notebooks/Bench_CMF_MODIS_Inundation.ipynb

What next?

To benchmark automatically a series of floods there are python scripts and a dedicated EO data extractions:

Promising new developments from the Atlantis project C4E-2026 to support Satellite flood extent data extraction. Here how to install Atlantis.

These steps worked well for me testing a VIIRS data extraction on the HPC or on the MacOs

#Instruction to install and run Atlantis on the HPC
cd $PERM
git clone -b main git@github.com:ECMWFCode4Earth/atlantis.git
cd atlantis
module load uv
git lfs install
git lfs pull
git pull
make setup
make demo

After the installation and demo one can extract any flood-case on VIIRS. You can look at any event in the years: 2012, 2013, 2014, 2015, 2016, 2017, 2018, 2019, 2020, 2023, 2024, 2025, 2026. Here an example for the Panama 2016 flood:

uv run atlantis fetch --event Panama_2016 --source viirs --bbox "-70 -40 -50 -20" \
--start-date 2006-01-08 --end-date 2006-01-08 --plot --harmonise --output ./data/Panama_2016

Ex. output in png (also tif files are produced):

run as it is from the /perm/${USER}/atlantis directory and you will have the output under .data/Panama_2016.

The name of the event and the directory you change as you see fit. The Bounding Box has "Lon_min Lat_min Lon_max Lat_max" .

Atlantis works also on the MacOs. I had to install the uv (with curl) and lfs (with conda) packages first, but then it worked well there too. Below the instructions for MacOs.

#Instruction to install and run Atlantis on a Mac
curl -LsSf https://astral.sh/uv/install.sh | sh
source $HOME/.local/bin/env
conda install git-lfs

git clone -b main git@github.com:ECMWFCode4Earth/atlantis.git
cd atlantis
git lfs install
git lfs pull
git pull
make setup
make demo

Feedback welcome! pad@ecmwf.int

  • No labels