资源环境科技发展态势分析平台

Large uncertainties exist in real-time predictions of the 2015 El Nio event, which have systematic intensity biases that are strongly model-dependent. It is critically important to characterize those model biases so they can be reduced appropriately. In this study, the conditional nonlinear optimal perturbation (CNOP)-based approach was applied to an intermediate coupled model (ICM) equipped with a four-dimensional variational data assimilation technique. The CNOP-based approach was used to quantify prediction errors that can be attributed to initial conditions (ICs) and model parameters (MPs). Two key MPs were considered in the ICM: one represents the intensity of the thermocline effect, and the other represents the relative coupling intensity between the ocean and atmosphere. Two experiments were performed to illustrate the effects of error corrections, one with a standard simulation and another with an optimized simulation in which errors in the ICs and MPs derived from the CNOP-based approach were optimally corrected. The results indicate that simulations of the 2015 El Nio event can be effectively improved by using CNOP-derived error correcting. In particular, the El Nio intensity in late 2015 was adequately captured when simulations were started from early 2015. Quantitatively, the Nio3.4 SST index simulated in Dec. 2015 increased to 2.8 A degrees C in the optimized simulation, compared with only 1.5 A degrees C in the standard simulation. The feasibility and effectiveness of using the CNOP-based technique to improve ENSO simulations are demonstrated in the context of the 2015 El Nio event. The limitations and further applications are also discussed.