TEL:+86 13943095588

ziņas

mājās > Uzņēmums > JAUNUMI > ziņas > Current Status of Low-Earth Orbit Satellite Reentry Prediction Technology: A Mature System with Accuracy Bottlenecks to Break

Current Status of Low-Earth Orbit Satellite Reentry Prediction Technology: A Mature System with Accuracy Bottlenecks to Break

Mar 25 2026

After decades of technological iteration and accumulation, a relatively complete technical system has been formed for LEO satellite reentry prediction, with two core technical approaches established: object-oriented modeling and spacecraft-oriented modeling. Major international space agencies have developed dedicated prediction toolchains to provide technical support for satellite reentry risk management, yet significant accuracy bottlenecks remain.

Current Status of Low-Earth Orbit Satellite Reentry Prediction Technology: A Mature System with Accuracy Bottlenecks to Break

I. Existing Prediction Method System

NASA’s Object Reentry Survival Analysis Tool (ORSAT) is a typical representative of object-oriented modeling. This tool decomposes a satellite into a set of discrete components and simulates the independent trajectories of each fragment after the parent vehicle breaks up. Developed in 1995, it has been updated to Version 7.1, which features a full rewrite of the thermal demise model, adopts a forward-time/centered-space numerical scheme, and incorporates a pyrolysis model for fiber-reinforced plastic (FRP) materials, greatly improving the accuracy of component breakup prediction.

A diversified landscape of mainstream global prediction tools has taken shape. In addition to NASA’s ORSAT, ESA’s SCARAB and DRAMA, CNES’s PAMPERO and DEBRISK, and JAXA’s ORSAT-J together constitute the current mainstream toolset. Their core functions focus on predicting spacecraft breakup during reentry, calculating casualty areas for surviving debris, and thereby quantifying reentry risks to the human population on Earth.

Technically, existing prediction methods rely on three core elements:

First, ballistic coefficient estimation, which serves as the prediction foundation. Typical statistical methods determine the ballistic coefficient by processing Two-Line Element (TLE) data, but its accuracy heavily depends on the quality and density of TLE data. Studies have shown that TLE data and the SGP4 model can only serve as initial conditions for numerical integration and cannot meet the requirements of precise reentry prediction.

Second, atmospheric density models. The widely used NRLMSISE-00 model has notable deficiencies: it systematically underestimates thermospheric hydrogen density by approximately 100% during low solar activity, leading to large short-term density prediction errors.

Third, propagation algorithms. Traditional numerical integration methods incur high computational costs and are poorly suited for large-scale constellation prediction. The high-efficiency semi-analytical propagator developed by Leonid Space achieves a speedup of over 3500 times compared with the Orekit reference implementation and 4.5 times relative to DRAMA, enabling Monte Carlo analysis for large satellite ensembles.

II. Actual Level of Prediction Accuracy

Validation results released by Leonid Space in 2025, based on 934 real reentry events spanning 1961–2024 and covering six solar cycles, provide a comprehensive assessment of current prediction accuracy. The study classifies prediction performance into three scenarios:

Under perfect knowledge conditions (using actual space weather records and drag characteristics derived from the satellite’s final year of flight data), the median error in one-year deorbit time prediction is 6.0 days (1.6%).
Under historical conditions (estimating drag parameters using only the preceding 30 days of tracking data, with space weather taken from actual records), the median error is 18.6 days (5.1%).
Under fully predictive conditions (using only contemporaneously available solar cycle forecasts as input), the median error reaches as high as 45.5 days (12.4%).

These data demonstrate that the accuracy of space weather prediction directly determines overall prediction performance: the error under fully predictive conditions is 7.5 times that under perfect knowledge conditions. The unpredictability of solar and geomagnetic activity has become the greatest bottleneck for current technology.

For short-term prediction, performance is relatively favorable. A reentry prediction case in May 2025 showed that the predicted time 9 days before reentry was 06:37 UTC on May 10 with an uncertainty of ±3.28 hours, while the actual reentry time was 06:16 UTC, falling within the predicted window—indicating that prediction accuracy improves significantly as reentry approaches.

Nevertheless, long-term prediction remains severely challenging. Under normal solar activity, orbit prediction errors within 24 hours can be controlled within 100 meters, whereas during periods of intense solar activity, deviations between the actual and predicted satellite positions may reach 1 kilometer, with errors accumulating and amplifying over longer prediction horizons.

In terms of institutional capability comparison, NASA’s ORSAT excels in handling complex spacecraft configurations, ESA’s DRAMA performs prominently in large-scale debris environment analysis, and the Leonid method achieves approximately a fourfold improvement in prediction accuracy for well-characterized satellites under the “historical conditions” test compared with ESA’s DRAMA & DISCOS toolchain.

III. Major Technical Limitations

Current LEO satellite reentry prediction technology is constrained by four key limitations:

Fundamental deficiencies in atmospheric density models: Existing empirical models lack a rigorous physical foundation and cannot accurately capture complex thermospheric dynamic processes. Errors reach approximately 10–15% during quiet space environment periods and exceed 100% during disturbed periods, reflecting insufficient understanding of the Sun–Earth system coupling mechanism.
Uncertainties in satellite characteristic parameters: Critical parameters such as spacecraft ballistic coefficient, mass distribution, and surface properties are difficult to obtain precisely. Most on-orbit targets lack complete detailed data, and ballistic coefficients can only be inferred from on-orbit behavior. Assuming a constant ballistic coefficient introduces errors comparable to those from density uncertainty.
Bottlenecks in space weather prediction: The high variability of solar activity makes thermospheric conditions difficult to measure accurately. Current models yield prediction errors of 30–50% for thermospheric density surges caused by magnetic storms, with forecast lead times of less than 12 hours.
Constrained computational resources: Traditional high-precision numerical simulation methods are computationally expensive and cannot support real-time prediction for large-scale constellations. Computational load increases exponentially during Monte Carlo uncertainty analysis.