Tuesday, 9 January 2018: 1:45 PM
Room 19AB (ACC) (Austin, Texas)
StormGeo provides a 30 day, hourly, temperature and precipitation forecasts for sites throughout Scandinavia. The forecasts are used several times a day by energy traders. The forecasts are used in models of streamflow to predict the amount of hydropower produced the next day, and the next month. The forecasts are also used to look at the demand for power in Scandinavia tomorrow and for the next month. Energy trading is a very competitive business, so continuing improvements to the weather forecast can lead to large monetary gains. StormGeo studied improvements to the forecast for the first three days. A simple bias correction was found to improve forecasts at about 30% of the 40 locations studied. The higher resolution numerical weather prediction model generally had biases closer to zero than the coarser resolution model. Blending model forecasts improved the next day temperature and precipitation accuracy. The hit rate for temperature and precipitation was improved at all sites with model blending and bias correction. This forecast improvement is appreciated by the energy traders. Improvements in forecasting correctly freezing temperatures during precipitation events in winter lead to improvements in the hydropower forecasts. Improvements in temperature forecasts improve the energy demand forecasts.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner