SDO/HMI HARP 多模态分层磁极反转线数据集

SDO/HMI HARP 多模态分层磁极反转线数据集是一组从2010年5月至2025年4月期间由美国宇航局太阳动力学观测站的Helioseismic and Magnetic Imager (HMI) Active Region Patches (HARP)中提取的多模态分层磁极反转线 (MPILs) 数据。数据集包含四个不同的层级，每个层级通过使用四种不同的磁场强度阈值来捕获MPIL特征的细微变化。总共提供了6,695个HARP系列，使用Lambert Cylindrical Equal Area (CEA) 投影以12分钟的频率映射。这种分层方法确保每个层级都能捕捉到对极性变化的特定敏感性，使研究人员能够根据一系列科学和操作目标调整他们的分析。在每个阈值层级中，我们提供了与日球物理和空间天气预报相关的六个二进制MPIL掩模，包括MPIL、极性反转区域（RoPI）、正/负极性区域、无符号极性区域和MPIL凸包。此外，还包含了以多元时间序列形式的结构化元数据，允许用户跟踪和分析MPIL属性。数据集整合了极性反转线的时空特征，这对于预测诸如太阳耀斑等事件可能有益。数据集的存储格式为HDF5，包含二进制掩模和数值元数据，为空间天气预报、太阳物理研究和机器学习应用提供了宝贵的资源。

SDO/HMI HARP Multimodal Hierarchical Polarity Inversion Line Dataset is a collection of multimodal hierarchical polarity inversion line (MPIL) data extracted from the Helioseismic and Magnetic Imager (HMI) Active Region Patches (HARP) aboard NASA's Solar Dynamics Observatory (SDO) during the period from May 2010 to April 2025. The dataset comprises four distinct hierarchical levels, each capturing subtle variations in MPIL features using four different magnetic field intensity thresholds. A total of 6,695 HARP series are provided, mapped at a 12-minute cadence using the Lambert Cylindrical Equal Area (CEA) projection. This hierarchical approach ensures that each level exhibits specific sensitivity to polarity changes, enabling researchers to tailor their analyses to a range of scientific and operational objectives. For each threshold level, six binary MPIL masks relevant to heliophysics and space weather forecasting are included, namely MPILs, Region of Polarity Inversion (RoPI), positive/negative polarity regions, unsigned polarity regions, and MPIL convex hulls. Additionally, structured metadata in the form of multivariate time series is provided, allowing users to track and analyze MPIL properties. The dataset integrates the spatiotemporal characteristics of polarity inversion lines, which can facilitate the prediction of events such as solar flares. Stored in HDF5 format, the dataset contains binary masks and numerical metadata, serving as a valuable resource for space weather forecasting, solar physics research, and machine learning applications.