Evaluation and Application of a Rigorously Validated Ocean OSSE System to Improve Ocean Model Initialization for Coupled Hurricane Prediction

Abstract:

A new fraternal-twin OSSE system has been configured over an extended North Atlantic Ocean domain to evaluate ocean observing strategies with respect to improving coupled hurricane forecasts. It is the first ocean OSSE system to employ all rigorous validation and calibration methods developed for atmospheric OSSE systems. It is based on two different configurations of the Hybrid Coordinate Ocean Model (HYCOM). Model configurations used for the unconstrained Nature Run (NR) and data-assimilative Forecast Model (FM), including lower horizontal and vertical resolution for the FM, produce the substantially different physics and truncation errors required of a credible OSSE system. Evaluation of the NR demonstrates sufficient realism in terms of mean climatology and of statistical properties of eddy and current variability to represent the “true” ocean. System evaluation is performed by comparing OSSE’s to Observing System Experiments (OSEs) that are identical except for assimilating actual observations. Based on RMS error increases, the OSSE system tends to overestimate impacts by <10% on average and thus requires only small calibration. The system is then applied to evaluate ocean observing strategies during the 2014 hurricane season. Example results are presented that evaluate different strategies for conducting airborne ocean profile surveys prior to Hurricane Edouard. The impact of the actual pre-storm survey performed prior to Edouard is evaluated against alternate survey strategies. Results demonstrate the importance of surveying as large a region as possible with ocean profiles separated by < 100 km. Modest additional advantages are achieved by profiling both temperature and salinity with CTDs compared to temperature profiles alone from XBTs. The HYCOM-HWRF coupled hurricane prediction model can now be initialized with analyses produced by the OSSE system. The impact of airborne survey design on actual hurricane forecasts will also be discussed.