A Preliminary Analysis of the Impacts of Assimilating UAS Data in NSSL’s Experimental Warn-on-Forecast System

From 24–25 March 2023, a severe weather outbreak occurred across the lower Mississippi River Valley. An isolated supercell produced the violent, long-track EF4 Rolling Fork, MS tornado, which was the deadliest in the area in more than 50 years. Despite the increasing frequency of severe weather events, the ArkLaMiss region lies in a weak radar coverage zone with few stationary observing platforms. High-resolution temporal and spatial observations of the atmospheric boundary layer (ABL) are necessary for creating more accurate forecasts, such as those produced by NSSL’s Warn-on-Forecast System (WoFS). Therefore, this study analyzes the impacts of assimilating Uncrewed Aerial System (UAS) observations in the analysis and prediction of the Rolling Fork, MS tornadic supercell. This severe weather event was sampled using various instrumentation during the PERiLS 2023 field campaign, including three UAS CopterSonde instruments located at three separate, stationary profiling sites near the Louisiana/Mississippi/Arkansas border. The CopterSondes collected ascending data every 15–30 minutes, with cadence depending on storm motion and other environmental factors. Simultaneously, the WoFS launched real-time forecasts covering a 900 by 900-km domain centered over the moderate risk area issued by the SPC’s Day 1 Convective Outlook. During this mission in particular, the UAS profiling sites were colocated with the center of the WoFS domain. For this study, two experiments are conducted using the WoFS data assimilation and prediction system. The control experiment is the baseline WoFS, while the UAS experiment assimilates the temperature, dewpoint temperature, u-wind, and v-wind components from the CopterSonde UAS observations in addition to all other observations in the baseline WoFS. 0–3-h forecasts are initialized every hour starting from 1700 out to 0000 UTC the next day. This talk will discuss the preliminary results from both experiments, with a focus on forecast trends and biases associated with the assimilation of the UAS observations.