Data Sheet 1_Explainable machine learning of the MCP dark count observed by Earth-orbiting space telescope.pdf
收藏NIAID Data Ecosystem2026-05-10 收录
下载链接:
https://figshare.com/articles/dataset/Data_Sheet_1_Explainable_machine_learning_of_the_MCP_dark_count_observed_by_Earth-orbiting_space_telescope_pdf/31962750
下载链接
链接失效反馈官方服务:
资源简介:
Radiation in low Earth orbit (LEO) poses a critical risk to satellite electronics and optics, and clarifying the causes of sudden variations in Earth’s inner radiation belt is therefore essential. Using the dark count rates on the micro-channel plate (MCP) detector of the Hisaki space telescope collected over the period 2013–2018, we identified several enhancement events with amplitudes of a factor of 2–5. A two-stage regression analysis with model training on 2013–2016 and detailed event analysis in 2017–2018, combining satellite orbital parameters with Geostationary Operational Environmental Satellite (GOES) measurements (X-rays, magnetic fields, protons, and electrons) and the Symmetric-H component (SYM-H) index, detected events in August 2018 and September 2017, as well as precursor variations occurring about 2 days earlier. While such events are expected to be attributed to coronal mass ejections, SHapley Additive exPlanations (SHAP) analysis revealed an unexpected contribution from non-delayed X-ray variations, suggesting that solar flares may directly affect detectors within minutes. This finding indicates that non-delayed X-ray variations can act as a distinct, event-specific driver of transient dark count enhancements, highlighting the event-to-event variability of radiation-induced detector responses in LEO.
创建时间:
2026-04-08



