| Popis: |
Our current capability of space weather prediction in the Earth's radiation belts is limited to only an hour in advance using solar wind monitoring at the Lagrangian L1 point. To mitigate the impacts of space weather on telecommunication satellites, advancing the lead time of the prediction is a critical task. We develop a prototype pipeline called "Helio1D" to forecast ambient solar wind conditions (speed, density, temperature, tangential magnetic field) at L1 with a lead time of 4 days. This pipeline predicts Corotating Interaction Regions (CIRs) in which their compressed stream interfaces and high-speed streams can increase high-energy fluxes in the radiation belts. The Helio1D pipeline connects the Multi-VP model, which provides real-time solar wind emergence at 0.14 AU, and the 1D MHD model. Using the long-term data from Multi-VP, we benchmark the Helio1D pipeline for solar wind speed against the observation data in 2004 - 2013 and 2017 - 2018. We developed a framework based on the Fast Dynamic Time Warping technique that allows us to continuously compare time-series outputs containing CIRs to observations to measure the pipeline's performance. In particular, we use this framework to calibrate and improve the pipeline's performance for operational forecasting. Since the 1D MHD model is computationally inexpensive, we provide daily ensemble forecasting of 21 members, including several targets around the Earth to account for the uncertainties. This pipeline can be used to feed real-time, daily solar wind forecasting to predict the dynamics of the inner magnetosphere and the radiation belts. In this presentation, we will share the lessons from this research-to-operation project and discuss ways to effectively implement operational space weather pipelines. |
