NASA is preparing to use the Artemis II crewed lunar flyby as a live proving ground for a suite of forecasting models designed to predict solar particle storms up to 24 hours before they strike. The mission, which will send four astronauts beyond low-Earth orbit for the first time since the Apollo era, depends on accurate space-weather predictions to protect both crew and spacecraft electronics. Six distinct models, each covering different warning windows, are being integrated into a single alert system that will be tested under real deep-space conditions.
What is verified so far
Two NASA centers are leading the forecasting effort. At Goddard Space Flight Center, the Moon to Mars Space Weather Analysis Office coordinates solar monitoring and model development, providing a central hub for experts who continually watch the Sun for signs of trouble and advise mission planners on radiation risk. That work is closely tied to the operational responsibilities of the Space Radiation Analysis Group at Johnson Space Center, which manages the warning system that will serve the crew in real time and translate scientific forecasts into concrete guidance for flight controllers.
The warning system tying these models together is the Integrated Solar Energetic Proton Alert/Warning System, or ISEP, which NASA’s radiation group describes as its primary space-weather decision tool for Artemis missions. ISEP bundles six models (MAG4, SEPSTER, UMASEP, REleASE, SEPMOD, and STAT) into a single framework. Each covers a different piece of the prediction chain, from early magnetic precursors on the Sun to the arrival of high-energy particles at the spacecraft. By design, the system is meant to avoid dependence on any single model and instead provide a layered picture of risk.
MAG4, for instance, uses magnetic free-energy proxies derived from solar magnetograms to generate Poisson-based probabilities for flares above M-class and X-class thresholds, fast coronal mass ejections, and solar proton events. According to model documentation, it focuses on the likelihood of these events within the next 24 hours, giving Artemis II its longest formal lead time for a potentially hazardous burst of radiation. That daylong window is crucial for launch decisions and for scheduling activities that are especially sensitive to elevated particle flux.
Not every model in the suite works on that timescale. UMASEP v2.0 takes a different approach, analyzing real-time solar and particle data to flag when protons above 10 MeV are likely to arrive, with a forecast window of roughly seven hours. REleASE operates on an even shorter fuse, providing just 30 to 90 minutes of advance warning, before particle intensities are expected to rise. Other components, such as SEPMOD and STAT, help estimate how an event will evolve and how confidence in a forecast should change as new measurements come in. The staggered coverage matters because a single model cannot reliably track a solar event from its magnetic precursors all the way through particle arrival at the spacecraft. By layering predictions across different timescales, ISEP aims to give flight controllers a progressively sharpening picture of incoming radiation.
External support comes from NOAA’s Space Weather Prediction Center, which has committed to provide formal decision-support operations for Artemis II and to coordinate closely with NASA’s radiation experts at Johnson. NOAA has identified a “significant solar radiation storm” as the primary space-weather concern for the mission, particularly during the translunar and return phases when the crew will be far from Earth’s protective magnetic field. To rehearse those workflows, the prediction center ran a multi-day testbed simulation built around scenario-based solar events during a notional in-progress Artemis II flight, treating the mission as an operational laboratory for both forecasting and communication procedures.
NASA has also created a dedicated center of excellence for probabilistic solar energetic particle forecasting and so-called “all-clear” determinations. The CLEAR initiative brings together specialists from Goddard, the Community Coordinated Modeling Center, NASA’s radiation analysis group, NOAA, and several academic and industry partners. Its goal is to refine methods for quantifying the probability that conditions will remain quiet during specific windows, such as planned extravehicular activities, so that mission planners can weigh radiation risk alongside other operational constraints instead of relying on binary go/no-go judgments.
On the hardware side, the Orion capsule carries the Hybrid Electronic Radiation Assessor, or HERA, an onboard detector that functions as both a measurement instrument and a real-time warning system. NASA’s radiation protection concept for Orion calls for crew sheltering during solar particle events, with HERA triggering alerts when dose rates climb above predetermined thresholds. The instrument already flew on the uncrewed Artemis I mission, and peer-reviewed data published in Nature validated its deep-space radiation measurement capability. That flight produced the first Orion-specific radiation dataset, giving modelers empirical benchmarks for what the sensors, electronics, and crew will encounter when astronauts are aboard.
Launch policy ties these elements together. NASA’s published weather criteria for Artemis II include a constraint barring liftoff during severe or extreme solar activity, citing the risk that elevated energetic particles pose to both electronics and communications. Because those criteria explicitly reference solar radiation conditions, the performance of forecasting models such as MAG4 and UMASEP will have a direct say in whether the rocket leaves the pad at all, potentially delaying launch if predictions indicate a significant chance of a major storm within the first day of flight.
What remains uncertain
The biggest open question is how well these models will perform under operational pressure. Public documents describing the NOAA testbed exercise confirm that teams rehearsed scenario-based responses and information flows, but they do not report detailed accuracy rates, false-alarm frequencies, or statistical skill scores for the individual models or for the ISEP system as a whole. Without those metrics, it is difficult for outside observers to judge whether the 24-hour MAG4 forecast or the seven-hour UMASEP window will deliver the reliability needed when a real solar event threatens the crew.
Another gap involves how HERA’s real-time onboard data will feed back into ground-based models during the mission. Artemis I showed that HERA can measure radiation in deep space, but no publicly available operational report has described whether those measurements have been used to recalibrate or validate the six models that make up ISEP. It remains unclear, for example, whether dose rates seen inside Orion during quiet periods have been systematically compared with model-based expectations, or how quickly updated in-flight measurements will be incorporated into forecasts during Artemis II.
The integration of CLEAR’s probabilistic methods with ISEP’s more deterministic alerts is also not fully described in available sources. CLEAR is designed to quantify the likelihood of both hazardous events and safe “all-clear” intervals, yet current public descriptions do not spell out how those probabilities will be presented to flight controllers alongside ISEP’s warnings, or how disagreements between different models will be resolved when time-sensitive decisions (such as whether to proceed with a planned maneuver) must be made.
Finally, Artemis II itself will occur during a relatively active phase of the solar cycle, but the range of possible conditions during the mission window remains broad. Forecasting models can estimate probabilities, not certainties, and even a well-predicted storm could arrive slightly earlier or later than expected, complicating efforts to align launch, translunar injection, and high-risk operations with quieter periods. Until the mission flies and the models are tested against real events with a crew at stake, the true operational value of NASA’s multi-layered space-weather strategy will remain, in important ways, unproven.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
