一种NRLMSISE-00模型太阳辐射校正方法

根据空间天气的状态，调整大气模型的相关输入参数能够减小模型的计算误差.通过对比CHAMP卫星在轨大气密度探测数据与NRLMSISE-00模式的计算结果发现，通过调整F_10.7的输入，使轨道大气密度积分的模式计算结果与探测结果之间的误差达到最小，此时的F_10.7被称为理想F_10.7输入（F*）.进一步的分析发现，F*与太阳紫外辐射MgII指数存在很好的相关性，因此可以选择其他的太阳紫外辐射代理参数取代F_10.7，从而减小模型计算误差.本文采用神经网络技术，建立新的太阳紫外辐射代理参量F_euv与MgII，F_10.7等的对应模型，能够根据当日参数值计算F_euv.研究结果表明，新的代理参数能够有效减小NRLMSISE-00的计算误差. 

Optimizing the NRLMSISE-00 Model by a New Solar EUV Proxy

Solar EUV radiation is one of the sources for atmospheric density variation in thermosphere. NRLMSISE-00 model is widely used to simulate the density variation in the satellite orbit. However, this model always exhibits large density errors compared with satellite observations．In this work, the model output is compared with CHAMP satellite observation on the same position, and an ideal value of proxy for solar EUV (F*) is chosen through changing the input F_10.7. Further investigation shows that F* is closely related to another solar EUV index MgII. By using of neuron network technique, the correspondence relationship between the new proxy of solar EUV (F_euv) and F_10.7 and MgII. The target parameter is F*. The new proxy substitutes F_10.7, and is input into NRLMSISE-00 model. The results show the difference of thermosphere density between those detected by CHAMP and GRACE-A satellite and that calculated by the NRLMSISE-00 model is dropped significantly.