【作者】 张容焱； 宋德众； 李永尧； 吴滨；

【Author】 Zhang Rongyan Song Dezhong Li Yongyao Wu Bing The Climate Center of Fujian, 350001, Fuzhou

【机构】 福建省气候中心；

【摘要】 本文利用最优子集、神经网络、EOF-CCA和均生函数4种常用统计方法,对福建春季、前汛期(雨季)、后汛期(台风季)和秋冬季进行年度、季度旱涝的短期气候预测试验。结果表明,最最优子集、神经网络模型具有较好的预报能力,EOF-CCA模型有一定的预报能力,均生函数模型相对较差。更多还原

【Abstract】 Using four statistics method in common -optimization subset recursive, nerve network, EOF-CCA and mean create function-make a year and quarter experiments research of short-term climate prediction short-term weather estimate of flood/drought about spring, monsoon, typhoon season and autumn- winter season in Fujian. The result indicate that optimization subset recursive and nerve network has the good forecast ability, the model of EOF- CCA has stated forecast ability , but mean create function model has worse forecast ability correspondingly.更多还原