Jedrzej Mosiezny, GMV; Lorenzo Porcelli, GMV; Jie Chen, GMV; Valerio Levi, GMV; Alejandro Pastor, GMV; Sean Bruinsma, CNES; Dennis Thesing, German Aerospace Center (DLR); Andreas Bachlechner, German Aerospace Center (DLR); Thurid Mannel, German Aerospace Center (DLR)
https://doi.org/10.64861/ZSTN9644
Keywords: space weather, thermospheric density, atmospheric drag, geomagnetic storms, LEO cataloguing, orbit propagation, orbit determination, uncertainty realism, EU SST, NRLMSIS, DTM, JB2008, ballistic coefficient, space situational awareness
Abstract:
Low Earth Orbit (LEO) has entered an era of sustained growth driven by large satellite constellations and an expanding mix of commercial and institutional missions. In this regime, the quality of a space object catalogue is generally limited not only by sensor coverage, but also by the ability of orbit propagation (OP) and orbit determination (OD) algorithms to model atmospheric drag under rapidly changing thermospheric conditions. Space-weather events, including coronal mass ejections (CMEs), solar flares, high-speed solar wind streams (HSS), and corotating interaction regions (CIRs), trigger geomagnetic storms which in turn induce atmospheric heating . These events result in abrupt density changes and long recovery tails, leading to degraded prediction accuracy and overly optimistic covariance estimates precisely when conjunction assessment and re-entry monitoring require robust uncertainty characterisation.
This paper presents the EU SST R&D activity “Catalogue impact of space weather”, which targets operationally actionable handling of space weather-driven drag uncertainty in LEO. The study comprises: 1) a survey and classification of relevant space weather drivers and thermospheric density models suitable for routine cataloguing; 2) quantitative OP and OD performance assessments over representative disturbed and quiet periods; and 3) the definition and trade-off analysis of mitigation strategies that can be embedded in an operational EU SST cataloguing production chain. The objective is to move beyond qualitative “storm versus quiet” assessments and deliver actionable thresholds, guidelines, and trade‑offs suitable for an operational EU SST environment._x000D_
First, we present a storm-focused literature review of thermospheric density modelling under operational constraints, including input availability, licensing, and computational cost. Empirical density models remain the practical baseline for large-scale cataloguing yet exhibit distinct storm time‑behaviour that motivates careful model selection and configuration. The review establishes an event taxonomy and identifies the indices and proxies used by operational models (e.g., solar flux‑proxies and geomagnetic indices), providing a consistent basis for scenario design and for linking space weather severity to OD/OP performance outcomes.
Second, we introduce a two-step analysis pipeline that mirrors operational cataloguing practice to assess the impact of space weather. In the OP focused step, a state is fitted prior to a selected space weather event and propagated through the disturbed interval (up to one week, consistent with typical LEO operational horizons). The baseline scenario uses observed space weather drivers during the fit window and historically forecast (“future in the past”) drivers during the prediction window, alongside a quiet time reference scenario to isolate disturbance driven effects. A sensitivity analysis covers: 1) dynamical model fidelity, 2) fit window length, 3) disturbance placement inside or outside the fit window, 4) observed versus forecasted space weather inputs, and 5) the thermospheric density model family (NRLMSIS class models, and other operationally viable empirical alternatives such as DTM and JB2008). Performance is assessed against high accuracy reference ephemerides provided by public sources and collaborating operators, using orbit fit residuals, component wise and normed position differences on the prediction horizon, and the estimated ballistic coefficient response as a proxy for density related modelling error. In the OD focused step, the analysis is repeated with real EU SST observations. Altitude-dependent fitting intervals are used, and estimation behaviour is explicitly characterised through post-fit residuals by measurement type, rejected-observations fractions, state differences with respect to reference orbits, and the evolution and realism of the estimated covariance at the OD epoch and along the prediction interval.
To avoid confounding attribution of errors to space-weather effects, the pipeline includes automated manoeuvre detection. A complementary signal‑processing approach, smoothing along‑track timing residuals over an orbital period window and detecting peaks in higher‑order derivatives, has demonstrated the capability to recover manoeuvres present in operational plans, and is implemented for automatic outlier identification.
The study analyses 13 geomagnetic storms of different intensity levels (G1-G5) primarily associated with CMEs and approximately 40 active and inactive LEO objects over the period May 2022 to September 2025. Storm cases are benchmarked against six quiet periods with nominal solar activity. The interpretation of scenarios benefits from a user-facing dashboard enabling interactive exploration of performance metrics across scenarios. The resulting insights are discussed with a focus on space weather effects on the OP and OD performance in an EU SST operational frame.
Finally, we define and assess mitigation strategies with strong emphasis on operational viability, considering required inputs, computational cost, and integration complexity. Recommended actions include adaptive covariance scaling and realism enforcement techniques, model configuration switching during disturbed conditions, and catalogue maintenance heuristics that prioritise tracking resources when uncertainty inflation is warranted. The work provides a reproducible methodology and early evidence that “storm-proofing” a LEO catalogue requires both physics-aware modelling choices and data-driven operational heuristics.
Date of Conference: September 15-18, 2026
Track: Atmospherics/Space Weather