| As a systematic model that can reasonably simulate the process of climate change,the coupling climate model has certain physical characteristics of the atmosphere and oceans.However,the parameter value setting in the coupling mode is not perfect,and in order to constrain the model deviation and quantify the uncertainty of the prediction,the mode parameters need to be optimized.The current coupling mode multi-parameter optimization method pays attention to the sensitivity of parameters to a single module to improve the predictive capability of the module,while ignoring the interaction between the coupling mode multi-module and the parameter,which limits the ability of the multi-parameter optimization method to improve the mode forecast.Based on the research of common data assimilation and parameter optimization methods,this paper chooses 5-variable conceptual coupled mode system(5VCCM)as the background mode to construct a "twin" experimental framework,and in-depth study of data assimilation based on Kalman Filter(KF)And parameter optimization methods.Compared with the parameter optimization method with Ensemble Kalman Filter(EnKF)and the Ensemble Adjustment Kalman Filter(EAKF),the Data Assimilation with Enhanced Parameter Correction(DAEPC)can further improve the accuracy of model prediction.According to the phenomenon of using single-coupled-component observation information to adjust different parameters to make the RMSE of the module state different in the experiment,and in view of the limitations of the existing multi-parameter estimation methods that only use the singlecomponent observation information to modify the parameters to improve the singlecomponent prediction ability,this article based on the DAEPC method,a coupling mode multi-parameter optimization method based on multi-coupled-component is proposed.First,in the 5VCCM model system,the atmospheric and ocean observation data are used to establish the sensitivity information of the parameters to each component.Based on this,the parameters of collaborative optimization are selected,and the observation information of the multi-coupled-component is used to synchronously adjust multiple parameters in the coupling mode,so as to achieve the purpose of improving the accuracy of multiple components prediction at the same time.The experimental results show that whether in the perfect mode or the deviation mode,the sensitivity information is used to select the appropriate number of parameters for collaborative estimation.Compared with the model prediction with only the assimilation process,this method can reduce the RMSE of the atmospheric and ocean components at the same time and improve the analysis and forecasting capabilities of components in the coupled model.In order to verify the application potential of the proposed parameter optimization method in the actual forecasting system.It is verified in meridional overturning circulation box model(MOCBM)with more complex physical dynamics.In the perfect mode experiment,this method can improve the prediction accuracy of the atmospheric and ocean components at the same time;in the deviation mode experiment,although the prediction ability of multiple coupling components cannot be improved at the same time,it has an improving effect on the prediction of the atmospheric single component.In an intermediate coupled circulation model(ICCM)with actual physical meaning and state-space distribution characteristics,compared to component predictions with only assimilation processes,this method reduces the RMSE of atmospheric and sea surface temperature predictions at the same time.In terms of spatial distribution,the prediction error of atmospheric temperature in high latitude areas and sea surface temperature in central ocean areas is reduced.Therefore,the parameter optimization methods can also improve the analysis and forecasting capabilities of multi-coupled-component in more complex actual forecasting systems. |
