Simulated Operational Testing of the Prototype Implementation of the SOFIE Model: The 2025 Space Weather Prediction Testbed Exercise

During the May 2025 SWPT exercise, the physics-based SOFIE model demonstrated its operational viability by successfully simulating solar energetic particle events in significantly less than real time, achieving a 4-day forecast in 5 hours through optimized grid configurations and collaborative feedback.

The Gist

The Space Weather "Crystal Ball" That Actually Works

Imagine the Sun as a massive, temperamental lighthouse. Sometimes, it doesn't just shine; it sneezes. These "sneezes" are called Coronal Mass Ejections (CMEs), and they blast clouds of magnetic fields and high-energy particles (Solar Energetic Particles, or SEPs) into space.

If these sneezes hit Earth, they can fry satellites and give astronauts dangerous radiation burns. For decades, predicting exactly when and how hard these sneezes will hit has been like trying to forecast a hurricane while blindfolded. We have models, but they are often too slow to be useful in an emergency, or they are too simple to be accurate.

This paper is about a new, super-smart "weather station" called SOFIE (Solar wind with Field lines and Energetic particles) that finally passed a major stress test.

The Big Test: "The Artemis Drill"

In May 2025, a group of space experts (from NASA, NOAA, and universities) gathered for a massive drill called the Space Weather Prediction Testbed (SWPT). Think of this as a fire drill for space. They pretended that the Sun was sneezing right now, and they had to predict the danger to astronauts on a mission to the Moon (like the upcoming Artemis II mission).

They used two famous historical "sneezes" (from 2017 and 2001) as their practice targets. The goal? To see if SOFIE could predict the radiation storm faster than real-time.

How SOFIE Works: The "Digital Twin"

Most weather models are like a simplified map. SOFIE is different; it's a digital twin of the entire solar system's atmosphere.

1. The Background: First, it builds a 3D model of the solar wind (the constant breeze blowing from the Sun).
2. The Eruption: When a CME is spotted, it injects a "virtual firehose" of plasma into the model.
3. The Race: It then simulates how that firehose pushes particles, how they speed up, and how they travel along magnetic "highways" toward Earth.

The problem? Doing this math is usually like trying to count every grain of sand on a beach while running a marathon. It takes too long. By the time the model finished the calculation, the real storm had already passed.

The "Aha!" Moment: Speeding Up the Simulation

During the drill, the team realized they were trying to be too perfect. They were using a super-high-resolution grid (like a 4K camera) for the entire solar system, even for parts of space where no one was looking.

The Analogy: Imagine you are filming a car chase. You don't need a 4K camera focused on the trees in the background or the clouds in the sky. You only need that high detail on the car and the road it's driving on.

The team made a brilliant tweak:

• Coarse Grid: They blurred the background (the empty space far from the Sun and Earth) to save processing power.
• Fine Grid: They kept the high-definition focus only on the path of the CME and the route to Earth.

The Result: This was like switching from a slow, heavy truck to a sleek race car.

• Before: The model took days to run.
• During the Drill: Using this new "smart focus" method, SOFIE predicted a 4-day storm in just 5 hours.

The Verdict: Fast and Accurate

The paper shows that SOFIE didn't just run fast; it ran well.

• The Early Warning: It took a few hours to get the first "heads up" (because it needed time to see the CME speed up), but once it started, it predicted the storm's peak and how long it would last with impressive accuracy.
• The Trade-off: They tested three different "focus" settings. The most detailed setting was the most accurate but slower. The "smart focus" setting was slightly less detailed but fast enough to give astronauts time to get to shelter.

Why This Matters for the Future

For humans to travel to Mars, we need to know when to hide. We can't wait for a storm to hit before we know it's coming.

This paper proves that physics-based models (which understand the actual laws of nature) can be fast enough to be used in real-time operations. SOFIE is no longer just a research toy; it's becoming a practical tool that could one day be the "Space Weather Channel" for astronauts, telling them exactly when to duck for cover.

In short: The team built a super-fast, super-smart simulator that can predict solar sneezes before they hit, giving us the precious time we need to keep our space explorers safe.

Technical Summary

1. Problem Statement

Solar Energetic Particles (SEPs) pose severe radiation hazards to satellites, ground-based technologies, and, critically, human spaceflight missions beyond Earth's magnetosphere (e.g., Artemis missions). While physics-based models offer a more robust physical understanding of SEP acceleration and transport compared to empirical models, they have historically been considered too computationally expensive for operational, real-time forecasting. The core challenge addressed in this paper is determining whether a complex, physics-based SEP model can meet the latency and accuracy requirements of operational space weather centers (specifically NOAA/SWPC) to support future human exploration.

2. Methodology

The study utilized the SOFIE (SOlar wind with FIeld lines and Energetic particles) model, an integrated physics-based suite built within the Space Weather Modeling Framework (SWMF). The model was tested under simulated operational conditions during the 2025 Space Weather Prediction Testbed (SWPT) exercise at NOAA's Space Weather Prediction Center (SWPC).

• Model Components:

  • AWSoM-R: Simulates the ambient 3D solar wind background using magnetohydrodynamic (MHD) equations.
  • EEGGL: Generates and propagates Coronal Mass Ejections (CMEs) using a Gibson-Low magnetic flux rope configuration.
  • M-FLAMPA: Solves the Parker transport equation for energetic particles, simulating shock-driven acceleration and transport along magnetic field lines (neglecting cross-field diffusion for efficiency).
  • Note: The test particle solver (PARMISAN) was excluded for this operational test to prioritize speed.

• Experimental Design:

  • Events: Two historical, well-observed SEP events were simulated: the 10 September 2017 event and the 4 November 2001 event.
  • Inputs: The model ingested real-time (simulated) data including synoptic magnetograms (GONG/MDI), EUV/X-ray flare data, coronagraph images (LASCO) for CME speed, and active region locations.
  • Operational Constraints: The team simulated the workflow of a forecast center, including data latency (e.g., waiting ~1 hour for CME speed estimates) and limited computational time windows (4–6 hours per event).
  • Grid Optimization: A key methodological step involved testing different grid resolutions.
    • Setup 1: Default high-resolution grid.
    • Setup 2: Coarsened grid in the solar corona (SC) by a factor of 2 (used on-site during the exercise to improve speed).
    • Setup 3: Hybrid approach (coarse background, fine resolution for HCS and Earth-directed cones).

3. Key Contributions

• First Operational Testbed: This represents the first time the SOFIE model was run on-site at NOAA/SWPC under simulated operational conditions with direct feedback from forecasters and radiation safety operators (SRAG).
• Performance Validation: It demonstrates that a physics-based model can transition from "research-only" to "operational-ready" by achieving significant speedups without catastrophic loss of accuracy.
• Grid Resolution Trade-off Analysis: The study provides a quantitative analysis of the trade-off between grid resolution and computational cost, establishing that coarsening the background solar corona grid significantly reduces runtime while maintaining acceptable predictive skill for decision-making.
• Workflow Integration: It outlines a practical workflow where background solar wind is pre-computed, and CME/SEP simulations are triggered upon event detection, allowing for rapid "catch-up" with real-time.

4. Key Results

• Computational Speed:

  • Using 1,000 CPU cores, SOFIE completed a 4-day SEP simulation in approximately 5 hours (Setup 2) for the 4 November 2001 event.
  • For the 10 September 2017 event, a 4-day simulation was completed in ~13 hours.
  • Crucially, the model ran significantly faster than real time after the initial input delay, providing multi-day forecasts well before the event decayed.

• Predictive Accuracy:

  • SEP Fluxes: The model successfully reproduced the onset, peak intensity, and decay phases of SEP events for both >10 MeV and >100 MeV proton channels.
  • Threshold Crossing: While the earliest forecast (threshold crossing) lagged observations by ~1 hour (due to input data latency and initial computational catch-up), the model provided accurate forecasts for the subsequent 4-day evolution.
  • CME Morphology: Synthetic white-light coronagraph images matched LASCO observations, correctly capturing CME propagation direction and speed.
  • Statistical Metrics: For the 4 November 2001 event, Setup 1 (high res) achieved a Spearman correlation >0.9 with GOES data. Setup 2 (coarse res) maintained >92% of data points within an order of magnitude of observations.

• Latency Analysis:

  • The primary delay in the earliest warning (onset) was driven by the ~1-hour latency in obtaining CME kinematic parameters and the time required for the simulation to "catch up" to real-time.
  • Once the simulation caught up (e.g., within 1.19 hours of real time for Setup 2), it provided a leading time of >90 hours for the 4-day profile.

5. Significance

• Operational Viability: The study proves that physics-based SEP models are no longer just theoretical tools but can meet the strict latency requirements of operational space weather centers.
• Human Spaceflight Support: The ability to predict SEP fluxes significantly faster than real time is critical for the Artemis program and future Mars missions, allowing mission controllers to make timely decisions regarding sheltering astronauts from radiation storms.
• Strategic Implementation: The findings suggest a dual-model strategy for operations:
  1. Run a coarse-grid model (Setup 2) immediately upon event detection to provide rapid, order-of-magnitude estimates within the first few hours.
  2. Run a high-resolution model (Setup 1) subsequently to refine the forecast and improve accuracy for the later decay and ESP (Energetic Storm Particle) phases.
• Future Roadmap: The paper identifies specific areas for future improvement, including better estimation of seed particle populations, dynamic adjustment of mean free path parameters, and reducing input data latency through new instruments like the CCOR-1 coronagraph.

In conclusion, the 2025 SWPT exercise marked a milestone in space weather forecasting, demonstrating that the SOFIE model is robust, fast enough for operations, and capable of providing critical radiation safety data for human space exploration.