EGMS vs GNSS in Groningen: does satellite radar see the same subsidence?
Question: does EGMS show the same subsidence rate as GNSS stations on the ground? EGMS (European Ground Motion Service) measures ground motion with the Sentinel-1 radar satellites.
Data:
- EGMS L2b Calibrated, release 2020–2024: how each point moves towards or away from the satellite (the line of sight), for each satellite track separately;
- GNSS stations from the Nevada Geodetic Laboratory: the position of each antenna every day, above the Groningen gas field.
Method: for each station we take the EGMS points within 100 m. We convert the GNSS motion to the direction of the satellite. Then we compare the speed of motion in mm per year. We set the pass/fail limits before the calculations.
1. How much data we have
- 35 GNSS stations: 27 subsiding and 8 stable
- 125 station–track pairs with complete data (radius 100 m)
2. Is the subsidence rate the same?
How to read the figure: each point is one station seen from one satellite track. The horizontal axis shows GNSS and the vertical axis shows EGMS, both in mm per year. The diagonal line means perfect agreement. The grey band (±1.5 mm/yr) is the allowed difference. Filled points are subsiding stations, empty points are stable stations.
Almost all points are inside the grey band. Points from ascending tracks (blue) are a little above the line, and points from descending tracks (orange) are a little below it. This is a small difference in east–west motion between GNSS and EGMS. It is within the limit, and we did not study it further.
3. Results for the criteria set before the calculations
| Criterion | Result | Limit | Passed |
|---|---|---|---|
| K1: Is the subsidence rate the same? | 99% of pairs with |dv| <= 1.5 mm/yr | >= 80% | ✔ yes |
| K2: Are the time series the same? | median std of the difference 4.7 mm | <= 5.0 mm | ✔ yes |
| K3: Does EGMS show no false motion at stable stations? | max |dv| 1.62 mm/yr | <= 1.5 mm/yr | ✘ no |
Result by the rule set in advance: PARTIAL: the subsidence rate agrees, but the time series or the check at stable stations is outside the limit.
How to read the table: K1 and K2 pass, so the subsidence rate and the changes over time agree. K3 fails, but only just, at one stable station. So the formal result is partial agreement.
4. One station over time
Station AME1
How to read the figure: each panel is one satellite track (motion towards the satellite, LOS) or the vertical motion (U). The line is EGMS. The dots are daily GNSS positions converted to the same direction. Both go down at the same speed.
Conclusion
- EGMS shows the subsidence above the Groningen gas field well. The GNSS stations on the ground show the same.
- The formal result is partial agreement, because the check at stable stations (K3) failed, but only just.
- Limitation: EGMS is calibrated with a model based on GNSS stations, so this comparison is not fully independent. This is why the second post compares EGMS with levelling, which does not use GNSS.
Sources: EGMS (Copernicus Land Monitoring Service); GNSS: Nevada Geodetic Laboratory (Blewitt et al. 2016, 2018); method and limits: Koudogbo et al. (2025) EGMS Validation Report, Fabris & Nazari (2026).
The full notebook and all the code are on GitHub: Showcase_1_GNSS_Groningen.ipynb in the egms-check repository.