data_code.zip

Limitations such as finite ensembles and imperfect physical parameterizations introduce sampling and model errors, reducingbackground ensemble spread and underestimated error covariances, further leading to filter divergence. This can cause parameters to converge to inappropriate values, resulting in failed parameter estimation. To mitigate this issue, state-of-the-art inflation methods are commonly employed in ensemble filtering data assimilation to increase prior variances and alleviate filter divergence. In this study, we conduct an in-depth investigation of a novel adaptive covariance inflation algorithm (t-X) within the framework of an observation system simulation experiment (OSSE) based on anintermediate coupled model (ICM) and the Ensemble Adjustment KF(EAKF), aiming to develop a joint approach for optimizing both model parameters and initial fields simultaneously. The influence of parameter optimization on model deviation correction and climate estimation is investigated through a series of assimilation experiments. This study introduces a novel inflation algorithm t-X for parameter estimation and validates its feasibility based on theICM used for El Nino and Southern Oscillation (ENSO) simulation and prediction. Results indicate that the t-X algorithm performs well in parameter estimation, synthetic state analysis, and ENSO real-time prediction. Optimizing parameters concurrently with initial fields further enhances the simulation and prediction capabilities of the model.