Page History

The following This is a description of the latest operational release of GloFAS v3.23. For an overview of other GloFAS releases, please see: GloFAS versioning system

Summary

GloFAS version 3.2 was implemented as a minor change to provide access to the Global Flood Monitoring (GFM) beta version through its web interface (www.globalfloods.eu). 

The  Global Flood Monitoring (GFM) is designed to provide a continuous global, systematic monitoring of flood events, with significantly enhanced timeliness of flood maps for emergency response and improved effectiveness of Rapid Mapping activation requests (through better identification of the area of interest). The  beta version contains flood products that are based on an ensemble approach integrating three robust, cutting edge algorithms developed independently by three scientific teams. The data processing architecture underlying the different scientific algorithms is based on the data cube concept, whereby SAR images are geocoded, gridded, and stored as 'analysis-ready data' (ARD) in an existing spatio-temporal SAR data cube. Users have access to consensus flood maps (Observed Flood Extent layer) where a pixel is marked as flooded when at least two algorithms classify it as water.

3 is associated with a change in ECMWF IT infrastructure and the necessary modifications to a number of software. Here is a summary of the main changes to the GloFAS web service:

• Introduction of near real-time flood mapping-related products from the Global Flood Monitoring (GFM) beta version including:
  • S-1 Observed Flood Extent (data format: raster (GeoTIFF) and vector (shapefile)). The layer i dentifies the pixels covered by floodwater. Pixels that are usually underwater (such as lakes and reservoirs) are identified based on the monthly Sentinel-1 Reference Water Mask are not part of the Sentinel-1 Observed Flood Extent.
  • S-1 Observed Water Extent (data format: raster (GeoTIFF) and vector (shapefile)). The layer i dentifies the pixels classified as open and calm water using Sentinel-1 SAR backscatter intensity and is derived using the ensemble flood mapping algorithm.
  • Exclusion Mask (data format: raster (GeoTIFF)). The layer i ndicates the pixel locations where the SAR data could not deliver the necessary information for a robust flood delineation. It combines static effects leading to no-sensitivity in flood mapping, water-look-alikes, strong typography, and radar shadows.
  • Uncertainty Values (data format: raster (GeoTIFF)). The layer is g enerated along with the binary map product as a simplified appraisal of trust in the ensemble flood extent detection approach.

• Introduction of reference products from the Global Flood Monitoring (GFM) beta version including:
  • S-1 Reference Water Mask (data format: raster (GeoTIFF) and vector (shapefile)). The layer i dentifies the pixels classified as open and calm water, both permanent and seasonal, using Sentinel-1 SAR backscatter intensity using an ensemble flood mapping algorithm.
  • Affected Population layer (data format: raster (GeoTIFF)). The layer is extracted from the Global Human Settlement (GHS) layer and from the GHS-POP dataset. Represents the estimate of the number of people affected by the flood.
  • Affected Land Cover layer (data format: raster (GeoTIFF)). The layer provides information for quick assessment of affected land cover or land use types, e.g., how much agricultural/cropland area is affected by the flood extent.

• Introduction of new data access functionalities, including: 
  • Function to define Area of Interest.
  • Visualization of the available products for user-defined AOIs.
  • GFM products download option for user-defined AOIs (per layer or for all product layers).
  • Notifications configuration via the app and/or Twitter.

  Key specifications of the products

Input satellite data

Satellite sensor

S-1 - Synthetic Aperture Radar (SAR)

Acquisition mode

Interferometric Wide Swath (IW)

Data product

Level-1 - Ground Range Detected (GRD)

Spatial sampling

20 x 22 m

Global revisit frequency

Europe: ~1-3 days. Rest of the world: ~3-14 days, depending on location

SAR polarization scheme

VV + VH

General product specifics

Geographic coverage

Global

Pixel size

20 m

Timeliness

<8 hours after S-1 image acquisition.

Thematic accuracy

Target threshold of >70-80%, based on the Critical Success Index

1. New version of LISFLOOD

LISFLOOD version 3.2.0 is used in this GloFAS release. It includes important developments, such as:

• enhanced code capability to handle inputs;
• new netCDF readers based on xarray that allow for more flexibility and efficiency, dramatically decreasing the time required to load large inputs thus leading to improved computational performances;
• extensive use of unit tests to verify the functioning of the model and enhance collaborative development of the code;
• improvements of the management of latitude and longitude grids to allow maximum precision of the reference system;
• porting of the code to python3.

The modelling of the hydrological processes has not changed compared to the version of the model used in the previous GloFAS release.

For more information, please visit the dedicated GitHub space: https://github.com/ec-jrc/lisflood-code/releases

2. Enhancement of the web reporting points layer

• Introducing over 400 new river stations (grey symbols, for monitoring the hydrological forecast) in the reporting point layer
  • 279 in Africa
    • 45 in Mozambique, 29 in Cote d'Ivoire, 23 in Togo, 22 in Guinea, 19 in Cameroon, 16 in Chad, 14 in Namibia, 11 in Benin and Nigeria, 10 in Sudan and Zambia, 9 in Republic of Congo, 7 in Malawi and Rwanda, 6 in Senegal and Tanzania, 5 in Burkina Faso and Mali, 4 in Gabon and South Africa, 3 in Congo and Guinea Bissau, 2 in Algeria, Morocco and Zimbabwe and finally 1 new station in Botswana, Central African Republic, Ethiopia and Lesotho
  • 73 in Central America
    • 64 In Guatemala and 9 in Honduras
  • 65 in South America
    • all 65 in Argentina
  • 28 in Asia
    • 23 in China and 5 in Bangladesh

Technical details