DICast® Dynamic Integrated foreCast

  • Overview
  • History

The Dynamic Integrated foreCast (DICast®) system is tasked with ingesting meteorological data (observations, models, statistical data, climate data, etc.) and producing meteorological forecasts at user defined forecast sites and forecast lead times. In order to achieve this goal, DICast® generates independent forecasts from each of the data sources using a variety of forecasting techniques. A single consensus forecast from the set of individual forecasts is generated at each user-defined forecast site based on a processing method that takes into account the recent skill of each forecast module.

DICast® is a licensed technology of University Corporation for Atmospheric Research.

The DICast® system was first developed at NCAR in the Fall of 1998 with the goal of generating completely automated, timely, accurate forecasts out to ten days at thousands of international locations. Potential applications of this system include the transportation systems, precision agriculture, and general public-oriented forecasts. See Applications for other future applications of this technology.

The DICast® system ingests data from multiple sources and applies automated forecasting techniques to each data source. Each of these forecast modules produces an "independent" forecast. The forecast skill is then improved using a fuzzy logic scheme to combine the individual forecasts.

Input data used thus far includes the National Weather Service (NWS) Global Forecast System (GFS) and Eta models, the NWS MAV and MEX MOS, European Forecast Center's ECMWF, climatology and observational data. Supplemental mesoscale models including MM5 and WRF have also been utilized for special applications. A suite of forecasting algorithms is also applied to each as appropriate. These include Dynamic Model Output Statistics (MOS) 'smart' interpolation schemes, and persistence.

The system is designed to generate forecasts of several standard meteorological parameters at a set of user configured locations for each forecast lead time. The individual forecasts are combined using a weighted sum. The weights used in the combination are adjusted daily to reflect the recent performance of the forecast modules.