Thursday, 26 January 2012: 1:30 PM
Development and Calibration of a Rigorous Fraternal Twin Ocean OSSE System in the Gulf of Mexico
Room 340 and 341 (New Orleans Convention Center )
The use of rigorous OSSEs for observing system design and the evaluation of data assimilation systems (DAS) is less advanced for the ocean compared to the atmosphere. A prototype fraternal twin ocean OSSE system has been developed that uses two substantially different configurations of the HYbrid Coordinate Ocean Model (HYCOM) that mimic the performance of different model types. Development, evaluation, and calibration of this system is being conducted in the Gulf of Mexico. For the fraternal twin approach to be viable, errors between the two model configurations that are used for the nature run and the DAS must have similar magnitudes and properties as the errors between present-day, state-of-the-art ocean models and the true ocean. Also, the model configuration chosen as the nature run must be realistic, i.e. possess a model climatology, synoptic eddies, currents, etc. that agree with observations to within pre-specified limits. Multi-year non-assimilative simulations of Gulf of Mexico synoptic-scale variability have been performed using both model configurations. Comparison of the output from these two runs to each other and to ocean observations from satellites, cruise profiles, and surface drifters demonstrates that these basic criteria are met. The initial application of the OSSE system is to determine the impact of targeted quasi-synoptic aircraft observations of the upper ocean on the accuracy of ocean analyses and forecasts in the Gulf of Mexico. To evaluate and calibrate the system, aircraft observations taken in response to the Deepwater Horizon oil spill from a NOAA P-3 hurricane research aircraft during nine flight days between 8 May and 9 July 2010 are used. The evaluation is performed by comparing results from an OSSE that quantifies the impact of synthetic P-3 observations to an OSE using identical procedures that quantify the impact of the actual observations. If conclusions are similar, the system is viable; else, a suitable calibration must be derived. Results of this evaluation will be presented.
Supplementary URL: