The Assimilation of Jason-2 Significant Wave Height Data in the North Indian Ocean Using the Ensemble Optimal Interpolation

A parallel module of data assimilation based on the ensemble optimal interpolation (EnOI) scheme has been developed and implanted into the third-generation spectral wave model NOAA WAVEWATCH III version 3.14 (hereafter NWW3). This paper presents an evaluation of preliminary experiments assimilating satellite altimeter significant wave height (SWH) data into the NWW3-EnOI wave assimilation system. Data from Jason-2 are used mainly for assimilation and from Jason-1 for validation. The target computational domain for the experiments is north of 15 degrees S in the Indian Ocean, which is nested into a global implementation of NWW3. A stationary ensemble of model output SWH sampled during a long time integration was examined first, showing that the ensemble error covariances were of significant flow dependence due to the monsoon climate. Experimental results by the NWW3-EnOI were validated against a number of moored buoys situated in the Bay of Bengal and east of the Arabian Sea. It was found that the assimilation of altimeter SWH reduced the root mean square errors in analyzed SWH by approximately 20%-60% at different validation stations. The improvement in model skill can be retained throughout the forecast period. The assimilation is found to have the greatest impact of the assimilation in areas and seasons where and when sea states are dominated by swells.