The Impacts of Assimilating Various Densities of Uncrewed Aircraft System Observations on Regional NWP Forecasts in an OSSE
Journal Article
Overview
abstract
Abstract; Uncrewed aircraft systems (UASs) have emerged as an option for increasing the number of routine observations within the in situ observational gap in the lower troposphere. Before deploying a nationwide network of UASs, however, it is necessary to determine what impact UAS observations will have on weather forecast model accuracy and assess the relative benefits of various UAS networks. Our goal is to help address this knowledge gap by examining the impact of assimilating varying densities of UAS observations on Rapid Refresh Forecast System (RRFS) forecasts. To do this, an observing system simulation experiment (OSSE) is used that consists of two week-long nature runs over the contiguous United States. Five different networks in which UASs execute hourly vertical profiles up to 2 km AGL are examined, with the spacing between UAS sites varying between 300 and 35 km. Results show positive impacts from assimilating UASs, with observations from the 35-km UAS network reducing 6-h root-mean-square errors by over 15% in the lower atmosphere. It is also shown that the benefit per UAS in the bulk verification statistics decreases as more UASs are added to the network. Examining a low cloud ceiling case shows that UASs can improve cloud forecasts when there are minimal clouds at the analysis time owing to a better representation of above ground moisture, though the UAS impact was minimal when using the coarsest UAS network. Altogether, these results suggest that UASs can improve RRFS forecasts, and benefits can be obtained from less than a hundred UASs.; ; Significance Statement; There are relatively few weather observations in the lower atmosphere, which is problematic because many types of high-impact weather, such as thunderstorms, are sensitive to conditions in this layer. Observations from drones are one possible way to better observe the lower atmosphere, though the impact of these observations on weather forecast models is unclear. Using a simulated atmosphere as a surrogate for reality, we demonstrate that weather-observing drones can improve forecast models, with 6-h forecast errors being reduced by more than 15% when using over 6000 drones. Notable benefits can also be obtained with less than 100 drones spaced across the contiguous United States, which suggests that even a small fleet of weather-observing drones can improve forecasts.;
