The Development of a Regional, High-Resolution OSSE Framework to Study the Impact of Observations from Uncrewed Aircraft Systems on Numerical Weather Prediction

Many sensible weather phenomena are sensitive to boundary layer conditions (e.g., convection initiation, precipitation type, fog, tornadogenesis, snow squalls). Our nationwide observing system, however, has an in-situ gap in the boundary layer, which hinders our ability to accurately forecast these phenomena. Observations of temperature, humidity, and winds from uncrewed aircraft systems (UAS) have recently been suggested as a means of filling this data gap, but questions remain as to how much impact observations from UAS will have on numerical weather prediction (NWP) forecasts and the ideal configuration of a UAS network for weather observations. Such questions should be answered before considerable resources are devoted to creating a nationwide UAS weather network.

In this presentation, we will describe the development of an Observing System Simulation Experiment (OSSE) framework that will be used to study the impact of observations from UAS on a near-operational regional NWP system, the Rapid Refresh Forecast System (RRFS). This OSSE consists of two week-long nature runs, one in the winter and one in the spring, that are created using 1-km Weather Research and Forecasting simulations over the entire contiguous United States. The small grid spacing of the nature run allows for the better representation of mesoscale phenomena, and the relatively long period of integration over a relatively large domain allows for several different phenomena to be examined. As will be shown, the nature run produces realistic mesoscale features, such as convection initiation along a dryline and subsequent upscale development into a mesoscale convective system. Comparisons between simulated and real surface station observations also highlight the realism of the nature runs, though there are some small biases in temperature and humidity. Comparisons between the nature run and a multi-year Multi-Radar, Multi-Sensor (MRMS) climatology show a noticeable high bias in reflectivity and precipitation in the nature runs, which will need to be considered when evaluating the OSSE results. The final portion of the presentation will feature a comparison between real-data observing system experiments (OSEs) using RRFS and OSSEs using only simulated conventional observations. These comparisons between the OSEs and OSSEs will be used to assess whether observation impact and forecast skill in the OSSE is similar to reality. By following this careful approach to constructing the OSSE framework, we should be able to provide more trustworthy recommendations regarding the possible deployment of a UAS network for weather observations.