openEO

GFM Temporal Aggregation

GFM Temporal Aggregation

Description

This UDP performs temporal aggregation of Copernicus Emergency Management Service Global Flood Monitoring (GFM) products for a user-selected area, time range, set of GFM layers, and temporal statistic. It is intended for rapid flood-event summaries such as maximum observed flood extent, flood frequency, cumulative flooded observations, or counts of valid observations over an analysis period.

The workflow is based on the openEO Platform Global Flood Monitoring tutorial and loads the GFM collection from the EODC STAC API:

https://stac.eodc.eu/api/v1/collections/GFM

Supported GFM Layers

The bands parameter accepts one or more of the following GFM item assets:

LayerDescription
ensemble_flood_extentEnsemble flood extent derived from the individual flood algorithms. This is the default layer and the main layer for mapping observed flood occurrence.
ensemble_water_extentEnsemble water extent product.
ensemble_likelihoodEnsemble likelihood layer expressing confidence in the detected flood or water signal.
reference_water_maskPermanent and seasonal reference water bodies, useful for distinguishing known water from flood events.
exclusion_maskAreas excluded from flood mapping or interpretation.
advisory_flagsAdvisory quality or interpretation flags.
dlr_flood_extentFlood extent output from the DLR algorithm.
list_flood_extentFlood extent output from the LIST algorithm.
tuw_flood_extentFlood extent output from the TU Wien algorithm.
dlr_likelihoodLikelihood layer from the DLR algorithm.
list_likelihoodLikelihood layer from the LIST algorithm.
tuw_likelihoodLikelihood layer from the TU Wien algorithm.

When multiple bands are requested, each band is aggregated independently and preserved as a separate output band.

Temporal Statistics

The statistic parameter selects the reducer applied over the full temporal extent:

StatisticTypical use
maxMaximum value over the period. For binary flood-extent layers this indicates whether a pixel was ever mapped as flooded during the period. This is the default.
minMinimum value over the period. Useful for identifying pixels that consistently satisfy a condition, depending on the selected layer encoding.
meanMean value over the period. For binary flood-extent layers this can be interpreted as flood frequency across available observations.
medianMedian value over the period, useful as a robust central tendency for likelihood or flag-like layers.
sumSum over the period. For binary flood-extent layers this represents the number of flooded observations.
countNumber of valid observations contributing to the temporal aggregation.

The result has no temporal dimension: each selected GFM layer is reduced to one value per pixel for the requested analysis period.

References

Contact

  • APEx

    prime investigator