BetaThis is a new service. Your feedback will help us improve it. Tell us what you think

Flood Early Warning

Flood Early Warning estimates where flooding is likely in the next six hours and triggers warnings to affected households and responders.

Keywords

  • Tags: Forecasting
  • Tags: Resilience
  • Tags: Environment
  • Tags: Public safety
  • Stage: Production
  • Type: Hydrological model with learned residual correction
  • Sector: Planning and environment
  • Language: English

How does our product work?

A physical catchment model produces the primary forecast. A learned component corrects that forecast using the recent record of where the physical model has run high or low for each catchment.

The decision to issue a public warning is always taken by a duty hydrologist, who sees both the corrected and uncorrected forecasts.

Overview

Warning lead time is the single largest determinant of whether people move belongings, vehicles and themselves in time. Extending it by even two hours changes outcomes materially.

The learned correction is used because the physical model has known, stable biases in particular catchments that are easier to measure than to model from first principles.

Owner and responsibility

Environment Agency

Catriona Munn

flood-warning@environment.example.gov

See more information

More detailed information on the system

Here you can get acquainted with the information used by the system, the operating logic, and its governance in the areas that interest you.

System description

The learned component can only adjust the physical forecast within a bounded range. It cannot suppress a warning that the physical model calls for, and a duty hydrologist can override either.

This is a deliberate asymmetry: the cost of a missed flood warning is not comparable to the cost of a false one.

References

Annual flood warning performance report, including lead time and false alarm rate by catchment.

Was this information useful?

Would you like to give feedback? Your feedback will help us develop our algorithms further.

If you wish for the system owner to get back to you personally, please contact them directly using the email address above. For more general comments, use the box below.