Abstract
Space weather driven atmospheric density variations significantly affect low Earth orbit (LEO) satellites, rocket launches, and re-entry events. The Horizon 2020 SWAMI project aims to enhance our understanding of these processes by developing improved neutral atmosphere models and high-cadence geomagnetic indices. This paper provides an overview of the project's progress, including the development of the "MOWA" whole atmosphere model and new $Hp$ indices for enhanced nowcasting and forecasting.
Introduction
Predicting the path of a satellite in Low Earth Orbit (LEO) is surprisingly like predicting the path of a leaf in a storm. While we think of space as a vacuum, the thin upper reaches of our atmosphere—the thermosphere—create enough drag to significantly alter orbits. During solar storms, atmospheric density can spike by up to 800%, turning a routine orbit into a high-stakes guessing game for satellite operators and space debris trackers.
The Space Weather Atmosphere Models and Indices (SWAMI) project is a major European initiative designed to solve this. By blending high-fidelity weather models with faster, semi-empirical space models and introducing high-cadence geomagnetic indices, SWAMI aims to provide the most accurate picture yet of our planet's "whole atmosphere."
The SWAMI Framework
The project stands out by providing an integrated approach to the satellite neutral environment, where main space weather drivers are addressed alongside model improvements. The core objective is the development of the Model of the Whole Atmosphere (MOWA).
Methodology: Blending and Upgrading
The researchers took a multi-pronged approach to unify atmospheric modeling:
- Model Blending (MOWA): They extended the Met Office Unified Model (UM)—typically used for weather and climate—from its 85 km ceiling up to 135 km. They then "blended" it with the Drag Temperature Model (DTM), which covers the 120–1500 km range.
- Physics Upgrades: To make the UM work at high altitudes, they added Non-Local Thermodynamic Equilibrium (NLTE) effects and expanded the radiation scheme to include Far and Extreme Ultraviolet (FUV/EUV) light. Molecular viscosity was also incorporated to ensure stability in low-density regions.
- High-Cadence Indices ($Hp$): They developed the $Hp30, Hp60,$ and $Hp90$ indices (30, 60, and 90-minute cadences). These are "open-ended" versions of the standard 3-hourly $Kp$ index, allowing for more accurate representation of extreme storms without hitting a predefined ceiling.
- Machine Learning Forecasts: The team used Feedforward Neural Networks (FNN) and Gradient Boosting to predict these indices using real-time solar wind data, outperforming traditional persistence models for short-term nowcasting.
Datasets and Long-term Trends
The project utilized a massive database of neutral density observations to re-calibrate the DTM. This includes satellite accelerometer data from missions like CHAMP, GRACE, and GOCE, as well as orbital decay data from Stella and Cryosat2. These datasets cover multiple solar cycles, providing the necessary context for modeling density variations against solar activity.
Key Results and Insights
- Improved Density Accuracy: The new DTM2019 predicts densities approximately 20% lower than previous versions, aligning much better with high-accuracy datasets from TU Delft.
- Stable Whole-Atmosphere Runs: The team successfully ran the UM with a 135 km "lid," proving that first-principles weather models can be pushed into the thermosphere if the right physics (like molecular viscosity) are added.
- Operational Readiness: The project developed the MCM (MOWA Climatological Model), a fast look-up table version of the complex model that satellite operators can run on standard hardware in seconds for re-analysis and predictions of temperature, density, and winds.
Conclusion
The SWAMI project provides a pathway to improved space weather services by addressing both scientific issues and the transition of models into operational services. While the transition zone (120–180 km) remains a complex area for physical blending, the introduction of high-cadence indices and machine learning forecasts marks a significant step forward in satellite safety and orbital prediction.
Reference: Jackson, D. R., et al. (2020). The Space Weather Atmosphere Models and Indices (SWAMI) project: Overview and first results. J. Space Weather Space Clim., 10, 18. https://doi.org/10.1051/swsc/2020019
Summary updated