摘要
基于获得的海水浴场逐日两次观测资料及同期NCEP数值模拟结果(提取各种相关变量),通过求取两者之间的相互关系,并采用最优子集方法确定了对各预报要素具有重要意义的影响因子。针对海水浴场的气温、降水及能见度等要素的预报,应用支持向量机方法建立了相应的预测模型,最终获得不同海水浴场在不同预报时效、不同预报要素的数值产品释用结果。经过对比分析,各预报要素释用后的结果较释用前在预报准确率方面有较大的提高。
Based on the correlation between the observed data and the numerical simulation results from NCEP,the main influencing factors for different forecasting objects, for instance, temperature, precipitation, visibility and so on, were obtained. Furthermore, the SVM provided an effective and feasible way to construct the corresponding forecast models. Comparing with the direct outputs of the model, the interpretation and utilization of numerical simulations showed great improvement.
出处
《海洋预报》
2015年第2期14-23,共10页
Marine Forecasts
基金
国家海洋局海洋公益性行业科研专项(201005036)
