Examining the Impact of UAS Observations Collected Using Different Sampling Strategies on Ensemble Forecasts of Fog in the Ohio Valley

Dense fog events can be highly impactful, causing major airport delays, aviation accidents, and highway pileups. However, they are often difficult to predict due to a lack of dense spatiotemporal sampling of the nocturnal planetary boundary layer (PBL). Uncrewed Aerial Systems (UAS) represent a potential source of such dense PBL observations; however, the number of UAS needed for a positive forecast impact and the best sampling strategies to pursue with them for the highest impact on a model forecast are unknown. Using a series of Observing System Simulation Experiments (OSSEs) focusing on three impactful fog events in the Ohio Valley, this study seeks to compare the impacts of several UAS sampling strategies and determine how many UAS would be needed to substantially improve fog forecasts for these events. These sampling strategies include a 3d mesonet-like experiment with profiling UAS located at selected surface observing sites and an experiment testing the impact of observations collected by a network of delivery drones. This presentation will cover the selection of the cases for the OSSEs, the development of the high-resolution nature run for each OSSE and the forward operators for creating the synthetic UAS observations, our findings on the optimal localization and error assumptions to use when assimilating UAS observations, and our results from assimilating those observations into a Model for Prediction Across Scales (MPAS) ensemble using the Data Assimilation Research Testbed (DART).

Acknowledgments. This research has been conducted in response to requirements of the Federal Aviation Administration (FAA). The views expressed are those of the authors and do not necessarily represent the official policy or position of the FAA.