A rapid-update wind analysis and nowcasting system based on mesoscale model, radar, and surface data

Predicting sudden and significant changes in wind power generation is a challenging problem. Such changes can be forced by large-scale weather features as well as local-scale weather events. The large-scale weather systems can be predicted by numerical weather prediction (NWP) models with reasonable accuracy. However, the current NWP models have limited ability in forecasting ramp-events caused by the sudden and significant wind changes corresponding to mesoscale features, such as passages of cold front or thunderstorm outflow.

Given the rapid evolving and small-scale nature of many ramp events, an accurate and rapid-updated analysis that combines high-resolution data can provide a good alternative to NWP forecasts for use in ramp-event wind energy prediction. Over several decades, NCAR has developed data assimilation, nowcasting, and forecasting systems targeted at the improvement of very short-term forecasts of high-impact weather events. In addition to WRF-based data assimilation systems, NCAR developed the unique high-resolution data assimilation system VDRAS for convective-scale nowcasting applications. VDRAS is designed for the assimilation of high-resolution observations from radar and surface networks using a four-dimensional variational (4D-Var) scheme of a cloud-resolving model. One unique feature of VDRAS is its rapid update cycle, typically 6-15 minutes. VDRAS has been implemented in real-time operational environments since 1998. The frequently updated analysis from VDRAS through the assimilation of high-density observations provides the current atmospheric state, which is crucial for nowcasting wind and initializing high-resolution numerical models. Although VDRAS has been widely used for obtaining frequent analyses, its capability on wind nowcasting is yet to be proven. In this presentation, we will describe and discuss preliminary results of VDRAS wind nowcasting and its impact on energy nowcasting from our recent case studies.