New approach for estimating the ocean observation impacts on coupled assimilation using EFSO

Ensemble Forecast Sensitivity to Observation (EFSO) is a powerful technique that can efficiently identify the beneficial/detrimental impacts of every observation in ensemble-based data assimilation (DA) systems. EFSO has been widely used in atmospheric DA in recent years, but not yet applied to any ocean or coupled DA systems. This study extends EFSO to oceanic DA system by introducing a novel error norm that includes the forecast error in ocean salinity. This new approach was implemented to the CFSv2-LETKF for quantifying the individual impact of ocean observations. In the data-denial experiment, the CFS forecast errors were significantly reduced by removing the detrimental ocean observations detected by EFSO, showing the validation of impact estimation and the great potential of EFSO to be applied as a data selection criterion.