Impact of Assimilating the Special Radiosonde Observations on COAMPS Arctic Forecasts During the Year of Polar Prediction

Abstract

We quantify the impact of assimilating the extra radiosonde observations launched in August 2018 during the Year of Polar Prediction (YOPP) on numerical weather prediction forecasts over the Arctic using the U.S. Navy's Coupled Ocean/Atmosphere Mesoscale Prediction System (COAMPS^®). The Naval Research Laboratory Atmospheric Variational Data Assimilation System (NAVDAS) is used to assimilate the observations and provide the initial conditions for the COAMPS forecasts. Four data assimilation experiments are conducted as part of an observing system experiment (OSE) to investigate the impact of assimilating the extra radiosonde observations on COAMPS short-term forecasts (up to 48 h). The four experiments explore the impacts of assimilating the additional radiosonde observations in NAVDAS in cold-start or warm-start configuration. Based on the results from the four-month-long data assimilation tests, we find that assimilating the extra YOPP radiosonde observations improves the overall accuracy of the COAMPS short-term atmospheric forecasts over the Arctic for both cold-start and warm-start initializations. The warm-start configuration makes the most significant improvement to the COAMPS Arctic forecasts during the test period. The domain-averaged mean RMSE of the 48 h forecast from the warm-start assimilation is about 17% less than that from the cold-start initialization.