中文摘要：

极光卵的尺度大小与太阳风-磁层-电离层能量耦合过程紧密相关,准确预测其大小对空间天气研究和预报具有非常重要的意义.本文基于模糊c均值聚类算法,从Polar卫星紫外极光图像中自动提取极光卵边界数据（～1215000个赤道向边界点和-3805000极向边界点）,统计分析其与太阳风等离子体、行星际磁场、地磁指数等之间的相关特性,并构建了以行星际、太阳风为模型参数（模型1）和以行星际、太阳风及地磁指数为模型参数（模型2）的2种极光卵边界多元回归模型.以模型预测的极光卵边界与实际极光卵边界之间的平均绝对误差作为模型评价标准,将本文预测模型与Carbary（2005）模型和Milan（2009）模型进行了对比.结果表明,模型2对极光卵极向、赤道向边界预测的平均绝对误差为1.55和1.66地磁纬度,优于Carbary和Milan模型（Carbary模型极向、赤道向边界的平均绝对误差为2.18和5.47地磁纬度,Milan模型极向、赤道向边界的平均绝对误差为1.71地磁纬度和1.90地磁纬度）.

英文摘要：

The size of the auroral oval is closely related with the solar wind-magnetosphereionosphere coupling process,the accurate prediction of which plays an important role in space weather study and forecast.In this manuscript,a total of 3805000 poleward boundary points and1215000 equatorward boundary points are automatically identified from Polar ultraviolet images by using the fuzzy c-means clustering method.With the massive dataset,we statistically analyze the relationship between the auroral oval boundaries and interplanetary magnetic field（IMF）,solar wind parameters（SWP）,and geomagnetic index.We construct two auroral boundary models by multivariate regression technique,which respectively used the IMF and SWP（model 1）and IMF,SWP and geomagnetic index（model 2）as the model parameters.With the Mean Absolute Deviation（MAD）of the auroral boundaries between the model predictions and the observations being an evaluator,the proposed model is further compared with the models proposed by Carbary（2005）and Milan（2009）.The experimental results show that the MADs of model 2are 1.55 MLAT and1.66 MLAT for poleward boundary and equatorward boundary respectively,which are superior to the boundary models of Carbary（2005）（MADs are 2.18 and 5.47MLAT）and Milan（2009）（MADs are 1.71 and 1.90 MLAT）.