Delft-FEWS: An operational model integrator

The Delft Flood Early Warning System (Delft-FEWS) is model integration platform used around the world to support numerous operational needs. It grew up as a platform to support flood forecasting. Since its inception, it has been implemented for purposes well beyond flood forecasting. The popularity of the Delft-FEWS platform is derived from the very simple and flexible architecture it implements. Models are linked synchronously through time series, where a time series can be a scalar or a grid. Parameter storage, model connectivity, basin descriptions, operational work flows and predefined display configurations are all stored by FEWS. Any model that runs off a combination of states, parameters and time series can be linked to FEWS thus permitting forecasters to integrate favorite legacy models and newer experimental models side by side. In addition, Delft-FEWS supports uncertainty estimates via ensemble model runs and then calculations of probabilities from those ensembles. Data can be displayed in standard x-y plots, as profiles along a river, within a schematization of a system and in space as grids or basin averages with looping through time. Because there is a broad variety of users, the FEWS application supports a wide range of display functions.
One goal of the FEWS Development project at Deltares is the distribution of knowledge between the groups who conduct the various forms of operational hydrology whether it is flood forecasting or groundwater permitting. The U.S National Weather Service has recently implemented Delft-FEWS into a system they call the Community Hydrologic Prediction System (CHPS). The CHPS implementation offers an informative example of the technology transfer enabled by the FEWS application.

To download the presentation please click here.