Published June 14, 2023 | Version v1.2
Software Open

Mesogeos: A multi-purpose dataset for data-driven wildfire modeling in the Mediterranean

  • 1. National Observatory of Athens
  • 2. University of Valencia

Description

Mesogeos is a multi-purpose dataset facilitating the development of deep learning models for wildfire-related applications in the Mediterranean.

Mesogeos can be used to extract ML-ready datasets tailored to several wildfire applications like wildfire danger forecasting, final burned area prediction, burned area size prediction and others.

For more: https://arxiv.org/abs/2306.05144 

More information and a link for downloading the dataset can be found here: https://github.com/Orion-AI-Lab/mesogeos

A website is maintained here: https://orion-ai-lab.github.io/mesogeos/

Mesogeos contains variables related to the ignition and spread of wildfires, as well as for several fire drivers (vegetation, weather, human-related) for the years 2006 to 2022 at a daily 1km x 1km grid.

Specifically it contains the following variables:

Datacube Metadata:

  • Temporal Extent: (2006-04-01, 2022-09-29)
  • Spatial Extent: (-10.72, 30.07, 36.74, 47.7), i.e. the wider Mediterranean region.
  • Coordinate Reference System: EPSG:4326

 

Files

Orion-AI-Lab/mesogeos-v1.2.zip

Files (12.0 MB)

Name Size Download all
md5:c70480424e632b4b815b02f3f2788b08
12.0 MB Preview Download

Additional details

Related works

Funding

European Commission
TREEADS - A Holistic Fire Management Ecosystem for Prevention, Detection and Restoration of Environmental Disasters 101036926
European Commission
DeepCube - EXPLAINABLE AI PIPELINES FOR BIG COPERNICUS DATA 101004188