Thunder probability nowcast

 

The Finnish Meteorological Institute (FMI) has developed the Smartmet Nowcast (SNWC) system to provide accurate and timely nowcast information to end-users in the Scandinavian forecast domain. SNWC combines observation-based data with the NWP nowcast model and integrates it with the 10-day forecast, ensuring rapid and automated weather forecast production. 

SNWC was integrated into FMI's operational forecast production pipeline in 2021, initially focusing on key parameters like temperature, humidity, wind speed, precipitation, and total cloud cover. Different methodologies were used to handle the diverse characteristics of these parameters. Over time, the system has evolved to include machine learning techniques, incorporating new parameters like thunder probability and wind gusts into the production process since autumn 2023. 

The thunder probability nowcast follows a methodology similar to pySTEPS commonly used for radar-based nowcasts. This involves generating a present state based on observed lightning and creating a nowcast based on motion vector fields. Before blending the thunder nowcast with the operative forecast, a four-hour nowcast with 15-minute timesteps is created, part of the high-resolution nowcast family (HRNWC) alongside total cloud cover and precipitation accumulation. This thunder probability nowcast information is then integrated temporally with NWP data to account for dynamic evolution. 

Even though the thunder probability nowcast is created with quite simple methods, it has improved accuracy and quality of thunder nowcast greatly. This is mainly due to the systems ability to incorporate the observed lighting information with NWP forecast in high updating frequency of 15 minutes. The high time frequency is some way covering a lack of convection, but generating convection producing system might evolve thunder probability nowcast even further.