基于遗传算法优化的ENSO指数的动力预报模型反演

Retrieval of the non-linear dynamic forecast model of El Nino/La Nina index based on the genetic algorithm optimization

基于NCEP/NCAR提供的1958-1995年全球月平均海温距平场再分析资料,采用动力系统反演思想和遗传算法途径,进行了El Nino/La Nina指数的动力预报模型的参数优化和模型反演,从上述海温资料中重构了Nino3海温距平指数的非线性动力模型.模型预报试验结果表明,遗传算法具有的全局搜索和并行计算优势能够客观、有效地反演海温指数的动力预报模型,对Nino3海温指数和El Nino/La Nina事件进行较为客观准确的预测,为El Nino/La Nina预测提供有益的研究参考.

Based on the 1958-1995 global month-average SSTA field reanalysis data from NCEP/ NCAR, using dynamical system retrieval idea and GA Genetic Algorithm approach, the parameters of E1 Nino/La Nina exponential dynamical forecast model are optimized and the nonlinear dynamical model of Nino3 SSTA index is reconstructed from above SST data. The analytical result and forecast test show that GA has the advantage of global optimum search and parallel calculation, which enables objective and effective retrieval of SST index dynamical forecast model. So it can effectively forecast the Nino3 SST index and E1 Nino/La Nina event and afford useful reference to E1 Nino/La Nina forecast.