基于Holt-ARIMA-Lagrange Multiplier的CWC信息熵时序研究

Information entropy time series of CWC based onHolt-ARIMA-Lagrange Multiplier

降水云系的发展过程及其特征分析,是云降水物理学中的一个重要问题.本文选取一次云发展过程中的700 hPa云水含量(Cloud Water Content, CWC)和大气垂直方向上气流速度(Omega, OMG)的1 h值,以信息熵来度量CWC空间分布的混沌程度,辅以OMG的时间变化来判断云的发展,并提出了一种基于多尺度分解、Holt模型、自回归滑动平均模型(Autoregressive Integrated Moving Average Model, ARIMA)和Lagrange Multiplier的组合预测方法.结果表明:1)CWC熵具有非线性和非平稳性;2)在云的不同发展阶段,北方CWC熵序列的均值都小于南方,方差普遍大于南方;3)OMG区域均值与CWC熵的小波低频重构的极值点在时间上有很好的对应关系,相近的极值点在南方中占50%,在北方中占83.3%,表明CWC熵可以在一定程度上反映云系的发展;4)CWC熵序列往往具有多种时间尺度特征,故进行多尺度分解之后再组合建模的Holt-ARIMA-Lagrange Multiplier模型比单一预测方法、单层分解的预测模型更优,准确率提高3%以上.

The development process and characteristic analysis of precipitation cloud system is an important issue in the field of cloud precipitation physics.Here,the 700 hPa Cloud Water Content(CWC)and the 1h value of airflow velocity(omega,OMG)in the vertical direction of the atmosphere are used to measure the chaos degree of CWC distribution via the information entropy and judge the cloud development via OMG time series,hence a combined prediction model is proposed based on hybrid multi-scale decomposition,Holt model,Autoregressive Integrated Moving Average model(ARIMA)and Lagrange Multiplier.The results show that,the CWC entropy has nonlinear and non-stationary characteristics;the clouds over the north have smaller means of the CWC entropy sequence and larger variance compared with those over the south regardless of the cloud development stage;a good temporal corresponding relationship is found between the regional average OMG and the extreme point reconstructed by the wavelet low-frequency of the CWC entropy,and close extreme value points account for 50%in clouds over the south and 83.3%in clouds over the north,showing that CWC entropy can somehow reflect the cloud development;the multiple timescale features of CWC entropy sequences make the multi-scale decomposed Holt-ARIMA-Lagrange Multiplier model more accurate than the single prediction method and single-layer decomposed prediction model,with accuracy improvement of more than 3%.