UBIMET’s high-precision weather forecasts, powered by its proprietary core model, the UBIMET Consolidated Model (UCM), are a key component of the TWINVEST project. The forecasts are directly integrated into the platform for user access and serve as input for both the energy forecasting model and maintenance scheduling processes.
This system combines various numerical weather models such as the ECMWF model, the GFS model, the ICON model and UBIMET’s proprietary RACE model. In the UCM framework, these models are first harmonised to be available in a standard temporal and spatial resolution. Then they are combined by using a dynamic, machine-learning-based weighing algorithm that takes the different strengths of the models into consideration. After a final post-processing step—using statistical calibration and downscaling methods—the UCM generates the forecast. For the domain of Europe, forecasts are available with a temporal resolution of 15 minutes and a spatial resolution of 100×100 meters for all standard parameters and a forecast horizon of 14 days.
Additionally, the short-range forecasts utilise live meteorological data streams such as radar, satellite, and lightning data to describe current weather conditions accurately and provide high-precision forecasts for the upcoming hours. This is particularly important for the forecasting of severe weather conditions that might cause damage to unprotected infrastructure.
Finally, besides the highly relevant wind parameters (speed, direction, and gust), parameters such as temperature and precipitation are provided as well to deliver a complete description of the weather conditions at the Demo Sites.
© visuals: UBIMET