春季平均气温的多层递阶长期预报

本文以D.G.Lainiotis在估计理论方面的工作为理论依据（文献[1]），运用文献[2][4][5]中给出的方法，扼要介绍了一种高精度自适应预报－－多层递阶预报在春季平均气温长期预报上的应用。它克服了通常方法中因固定参数模型所造成较大误差的缺陷，在预报中充分考虑到系统的时变特性，将预报问题分害成对系统的时变参数预报及对系统的状态（输出）预报两部分。文章利用了黑龙江省某地区1961－1989年二十九年的春季平均气温的历史资料，对1990－2010年二十一个年份的春季平均气温进行了长期预报。这一测试结果表明这一方法比其它方法更准确。

A Prediction Method-Multi-Level Prediction and Its Application in the Meteorological Observation

This paper relys on the theoritical work on estimation devoloped by D. G. Lainiotis the method given by referenced[1] is used. The paper also simply intreduces a high accurate adaptive prediction method - multi-level prediction and its application in the meteorological observationl.New method overcomes the shortcoming of having large error which always appears in the normal method where the model with fixed parameters is used.In the procedure of predication tone-varying property parameters and the othe is the prdiction of the states of the system 29 yerars(1961-1989)data is used in thos paper,and the estimation of 21 years meteorological observation is made.The test resrlt shows that this method is more accurate than noramal methods.