Development of a Regional Research Ocean Nowcast-Forecast System Based on HYCOM

PIs: George Halliwell (NOAA/AOML/PhOD)

External PIs: Vassiliki Kourafalou (UM/RSMAS), Ashwanth Srinivasan (UM/RSMAS), and Matthieu LeHenaff (UM/CIMAS)

The overarching goal is to develop a new regional ocean nowcast-forecast system based on the HYbrid-Coordinate Ocean Model (HYCOM). The system has been set up at very-high (0.02°) resolution in a Gulf of Mexico domain and contains four state-of-the-art choices for data assimilation. An initial test of the system using the Sequential Evolutive Extended Kalman (SEEK) filter as the data assimilation method produced encouraging results (Figure 1). The quality of GoM analyses produced by this model and also by other ocean analysis systems are being evaluated at AOML against observations taken in response to the Deepwater Horizon oil spill.

The impact of new data assimilation systems on the quality of ocean analyses and forecasts is a significant focus of the research. This is a research system that will not produce operational real-time ocean analyses for public dissemination. Instead, this system will be used to perform Observing System Experiments (OSEs) to evaluate existing observing systems including those maintained by NOAA/AOML/PhOD and special ocean observations (cruises, drifters, aircraft profiles, etc.) taken in response to the oil spill. Observing System Simulation Experiments (OSSEs) will also be performed using this system to evaluate the impact of new observing systems. Ocean analyses produced by system will be used in retrospective studies of the Deepwater Horizon oil spill. The system is designed to be relocatable and can be set up in other ocean regions to respond to emergencies if needed.

Figure 1. Comparison of SSH from the AVISO altimeter product (left panels) and from an analysis produced by the new 0.02°. Gulf of Mexico ocean forecast system (right panels). At the initial time (8 April 2010, top panels), there are substantial differences between the modeled and observed SSH. As the model assimilates SSH and SST observations over the next several days, the structure of the Loop Current and adjacent cyclones more closely resemble the observed features.