【摘要】 水文预报中一般是直接采用实测资料建立滑动自回归（ＡＲＭＡ）模型，现采用实测资料减去季节均值后的中心化变量作为模型中的变量，既可消除季节因素的影响，也可利用季节均值为预报服务。再辅以迭代最小二乘法使参数处于线估计状态。ＡＲＭＡ模型在雅砻江泸宁站至小得石站河段的应用效果很好。更多还原

【Abstract】 A multiple input,single output ARMA model is employed to tackle a flood forecasting problem in the lower reaches of the Yalongjiang River.The model variables are normalized by substracting seasonal mean values from the measured values which are identified from the data record.The parameters of ARMA model are estimated by the least square method and are allowed to vary by using the rescursive least squares technique.The results so far obtained for this river reaches are promising.更多还原