Ensemble CME Modeling Constrained by Heliospheric Imager Observations
Ensemble CME Modeling Constrained by Heliospheric Imager Observations
复制标题
DOI：
10.1029/2020av000214
复制
发表时间：
2020-09
期刊：
影响因子：
8.4
通讯作者：
L. Barnard;M. J. Owens;C. Scott;C. A. D. Koning
中科院分区：
文献类型：
--
作者：
L. Barnard;M. J. Owens;C. Scott;C. A. D. Koning
预测日冕物质抛射（CME）的到来是空间天气预报的一个关键目标。在实际空间气象预报中，太阳风数值模型被用于这一任务，集合技术作为改进这些预报的一种手段正在得到越来越多的探索。目前，这些预测不受现场和遥感观测的限制，例如美国国家航空航天局（NASA）STEREO航天器上的日光层成像仪（HI）记录太阳风和CME的白色光图像。我们报告的案例研究的四个日冕物质抛射，并显示如何HI观测可以用来提高技能，减少这些事件的集合后报的不确定性。使用计算效率高的太阳风模型，我们产生了200个成员的集合后报，在最佳估计的均匀分布内扰动建模的CME参数。通过比较模型化的日冕物质抛射侧翼的轨迹与HI观测，我们计算出每个系综成员的权重。加权CME到达时间的集合分布，提高了技能，减少了每个事件的后报不确定性。对于这四个事件，相对于未加权的集合，加权的集合显示出到达时间误差的平均降低20.1 ± 4.1%，以及到达时间不确定性的平均降低15.0 ± 7.2%。如果能够获得真实的实时HI观测数据，这种技术可以应用于业务空间天气预报。因此，由于NASA和欧洲航天局目前正在计划下一次空间天气监测任务，我们的概念验证研究提供了一些证据，证明了在这些任务中包括人工智能的潜在价值。
Predicting the arrival of coronal mass ejections (CMEs) is one key objective of space weather forecasting. In operational space weather forecasting, solar wind numerical models are used for this task and ensemble techniques are being increasingly explored as a means to improve these forecasts. Currently, these forecasts are not constrained by the available in situ and remote sensing observations, such as those from the heliospheric imagers (HIs) on the National Aeronautics and Space Administration's (NASA's) STEREO spacecraft, which record white‐light images of solar wind and CMEs. We report case studies of four CMEs and show how HI observations can be used to improve the skill and reduce the uncertainty of ensemble hindcasts of these events. Using a computationally efficient solar wind model, we produce 200‐member ensemble hindcasts, perturbing the modeled CME parameters within uniform distributions about the best estimates. By comparing the trajectory of the modeled CME flanks with HI observations, we compute a weight for each ensemble member. Weighting the ensemble distribution of CME arrival times improves the skill and reduces the hindcast uncertainty of each event. For these four events, the weighted ensembles show a mean reduction in arrival time error of 20.1 ± 4.1%, and a mean reduction in arrival time uncertainty of 15.0 ± 7.2%, relative to the unweighted ensembles. This technique could be applied in operational space weather forecasting, if real‐time HI observations were available. Therefore, as NASA and the European Space Agency are currently planning the next space weather monitoring missions, our proof‐of‐concept study provides some evidence of the potential value of including HIs on these missions.