GEO轨道相对论电子日积分通量预报统计建模

基于辐射带相对论电子哨声波局地加速理论，将地磁AE指数作为源电子通量和通量各向异性的指标，将地磁Dst指数作为损失机制的指标，利用滑动窗口线性滤波器方法，建立了一个地球静止轨道大于2MeV相对论电子预报模型.利用该模型开展了2000-2009年地球静止轨道相对论电子通量预报试验.研究发现，这10年总预报效率为0.818，2003年的预报效率（0.633）最低，2009年的预报效率（0.856）最高.模型预报效果与持续模型相比有很大提高，略低于利用太阳风参数作为输入的同类预报模型的预报效果.这说明即使在缺少太阳风参数的情况下，该模型利用地磁扰动参数也能取得较好的预报效果.当模型输入参数增加了太阳风速度时，即综合考虑了行星际扰动和磁层扰动对辐射带粒子加速过程的影响，模型逐年的预报效率进一步提升.其中，2005年的预报效率提升了9.5%，这10年的总预报效率增加到0.848，预报值与实测值之间的线性相关系数为0.918，均方根误差为0.422. 

Statistical Model of the Relativistic Electron Fluence Forecast at Geostationary Orbit

Based on the theory of local acceleration by chorus mode wave, AE index is selected as a good indicator for both source electron flux and flux anisotropy, and Dst index is selected as a good indicator for relativistic electron loss. By use of AE and Dst indices as input parameters, a prediction model for relativistic electrons at GEO orbit has been established on the basis of linear filter technology. The results are as follows. The total Prediction Efficiency (PE) of our model for the data from 2000 to 2009 is 0.818. The highest PE is about 0.856, which occurred in 2009. The lowest PE is about 0.663, which occurred in 2003. The prediction accuracy of the model is significantly better than that of the persistence model and is slightly less than the model developed by the same method but taking the solar wind as the input parameters. Furthermore, the model is improved by taking solar wind speed as an additional input parameter. The values of PE from 2000 to 2009 increase every year. The PE of 2005 is increased by 9.5%. The total prediction efficiency from 2000 to 2009 increased to 0.848. The linear correlation coefficient between forecasts and observations is 0.918, and the root mean square error is 0.422.