the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 13 Mar 2026
Role of ionospheric and ground contributions in driving GIC: Northern Europe at the peak of the May 2024 superstorm
Abstract. We examine the geoelectric field and geomagnetically induced current (GIC) in Northern Europe during the May 2024 superstorm using a recently developed method: The divergence-free part of the geoelectric field (EDF), associated with rapid magnetic field variations, is estimated from ground-based magnetic field observations using spherical elementary current systems. The curl-free part of the geoelectric field (ECF), associated with charge accumulation, is estimated from EDF using coefficients that depend on ground conductivity and linearly relate ECF to EDF in the time domain. We apply the method to both regional 10 s International Monitor for Auroral Geomagnetic Effects (IMAGE) magnetometer observations and global 1 min SuperMAG observations by adopting a global triangular grid adapted to local magnetometer density. We compare the resulting GIC in the Finnish benchmark power grid and conclude that, in the absence of ground conductivity information and higher cadence data, EDF estimated from 1 min magnetic field observations alone can provide a reasonable proxy for GIC activity in a power grid compared to 10 s geoelectric field. However, polarization of the geoelectric field due to lateral variations in ground conductivity can produce intense GIC at substations connected to transmission lines traversing regions of enhanced geoelectric field. GIC at such substations may be poorly described by EDF alone.
- Preprint
(16912 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2026-1110', Anonymous Referee #1, 20 May 2026
-
RC2: 'Comment on egusphere-2026-1110', Anonymous Referee #2, 21 May 2026
The authors present an interesting study expanding on previous efforts to describe the geoelectric field using divergence-free (DF) and curl-free (CF) Helmoltz decomposition. This provides important intuition about the role of the source and telluric currents and could be used more widely in other geoelectric and GIC studies.
General comments:- It should be made more clear (especially on Lines 79-80 when you lay out the outline of the paper) that you first undergo a validation effort for your modelling using a geomagnetic storm from 1998 before moving on to the specific results for the May 2024 storm. It is confusing for the reader to see the title about May 2024 in Northern Europe, but then Section 2.1. 2.2, and 2.3 are all about June 1998 storm.
- Recommend to either remove Section 2.4, Section 3.4 and Figure 12, or mention that Central Europe will be investigated in the abstract and title. As it is written, these sections seem out of place in the context of a paper which primarily focuses on Fennoscandia.
Line 9: I strongly disagree with this assessment that 1 min E_DF can be used as a “reasonable proxy” for GIC. Your results show large variations in peak GIC depending on the choice of method which, without GIC data validation, makes it very difficult to assess which E field model is preferable. For example, the 60 s DF underestimates the 10 s DF+CF |GIC| by a factor of 5 at Alajarvi. Similar discrepancies are seen at other sites in Table 1 as well as in Figures 7-10. From a hazards perspective, a factor of 5 in GIC is a significant difference in risk to transformers and other relay equipment. If the 10 s CF+DF can be validated against GIC measurements as being the more accurate method, then it highlights the need for higher cadence magnetic datasets, and the necessity of incorporating ground conductivity into GIC modelling. (Note that there are also cases which see extreme reductions in GIC when using 10 s CF+DF compared to 60 s DF, so it is not a clear one-to-one correspondence where one method always sees higher risk; rather it is a complex story that can only be properly addressed using GIC data for validation.).
Line 54-55: The way you have written this implies that the ground conductivity does not impact the divergence-free field. However, I think it is important to clarify that both DF and CF electric fields depend on the ground conductivity implicitly. The ground magnetic field measurements are a superposition of the primary ionospheric source magnetic field and the secondary induced magnetic field. Thus, the total ground magnetic field used in Equation 1 already implicitly contains information about the ground conductivity via the secondary induced field.
Section 2.4: This seems out of place in the context of this paper.
Line 142: I think it is a typo and E_DF should be E_CF.
Line 164: Specify whether the 1-D model was constructed by you via some inversion method, or cite the relevant authors that produced it.
Line 223: Don’t use the word “data” to refer to modelled electric field and be clear about the distinction between GIC versus E. It should read: “The majority of the peak GIC values calculated from 10 s |E| model and all peak GIC values calculated form 10 s |E_DF| model occurred…”
Line 225: Was there no available GIC data during the May 2024 storm? I am surprised if Finland does not have any active GIC monitoring.
Line 250: The geoelectric field is not dominated by longer periods. It tends to have higher frequency content compared to the geomagnetic data. This can be explained by standard MT equation E = ZB where Z is a high-pass filter. There are many examples where lower sampling rate of 1 min geomagnetic data will miss spikes in geoelectric field driven by higher frequency dB/dt. See Trichtchenko (2021) for discussion.
Line 275: Is there a geological reason why the total E field is amplified in southern Finland compared to the DF field?
Line 293-297: I would remove this paragraph. It feels extremely speculative to guess why the Norway station tripped given that it is an entirely different network topology. Keep in mind that the TE mode electric field parallel to the coast would be decreased by the inductive ocean effect. Transmission lines parallel to the coast would thus see a "reverse coast effect” with a reduced electric field and reduced GIC. See Cordell and Unsworth (2025) for discussion of this.
Line 298-399: Everything in this section thsu far has shown that the 1 min data and DF field might be significantly underestimating the GIC at many locations. You also never make a direct comparison between 10 s CF+DF and 60 s CF+DF. That comparison would more definitively show the impact of 10 s versus 60 s data for the total field. But I largely disagree with your conclusion that 1 min data are sufficient.
Figure 11: Is there a reason why you expand the region of interest to a much larger area of the North Atlantic and eastern Greenland? Previous figures have focused on Fennoscandia where magnetometer coverage is dense. If you expand to include such a large area then there is extremely sparse magnetometer coverage over most of the map area shown in Figure 11 which might make SECS much less reliable. The text in Section 3.3 seems to focus mostly on interpretation over Fennoscandia, so why not just show the area you discuss in the text and the area that has good coverage?
Line 373: This sentence about the importance of E_CF in determining GIC amplitudes should be highlighted more prominently in the abstract. In order to do a proper GIC hazard analysis, you cannot ignore galvanic effects.
Line 374-375: There is a tendency to assume that the “external” and “internal” mathematical representations from SECS are a perfect mapping of ionospheric and telluric currents. In reality, the external and internal components may not be accurate representations of what is actually happening physically.
Line 432: Disagree that E_DF can provide reasonable proxy for GIC hazard when it has 5 factor variance from other (potentially more accurate) methods which incorporate E_CF.
Line 434: 1 min data is not sufficient to capture rapid variations in the electric field of interest to power networks via harmonics and transients. The most notable effects seen are often associated with sudden, rapid variations <1 min.
Line 436: Conclusion number 3 appears to contradict conclusion number 1.
Line 441-444: In order to assess the long-term risk, you need the full geoelectric field. Using the E_DF which potentially underestimates the risk by a factor of 5 is not sufficient for hazard analysis.
References
Cordell, D. R., & Unsworth, M. J. (2025). The influence of the geoelectric coast effect on geomagnetically induced currents. IEEE Transactions on Power Delivery, 1–11. IEEE Transactions on Power Delivery. https://doi.org/10.1109/TPWRD.2025.3544488
Trichtchenko, L. (2021). Frequency Considerations in GIC Applications. Space Weather, 19(8), e2020SW002694. https://doi.org/10.1029/2020SW002694
Citation: https://doi.org/10.5194/egusphere-2026-1110-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 697 | 2,096 | 61 | 2,854 | 54 | 67 |
- HTML: 697
- PDF: 2,096
- XML: 61
- Total: 2,854
- BibTeX: 54
- EndNote: 67
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The authors present results from large scale geoelectric field modelling during the May 2024 storm in Northern Europe, using the approach after Juusola (2025b) which also works in the absence of ground conductivity models. The ground electric field response (E) is divided into a divergence-free (DF) and curl-free (CF) part, where the DF part represents the large scale geoelectric field and is derived by magnetic field observations only. CF is computed from a ground conductivity model (S-map) and can locally have much larger amplitude and show strong polarization especially valid for GIC in power grids. The authors set out to see if EDF provides a good proxy for GIC, using various data sets (IMAGE, SUPER-MAG, BEAR) for the May 2024 geomagnetic storm. They then present several hypothetical scenarios using the Finnish benchmark power grid to assess potential GIC levels. Then the authors turn to a broader scale looking at model outputs for ionospheric equivalent currents and electric field models for northern Europe for two time-steps during the May 2024 storm 22:29 when the electrojet was exceptionally far south and 22:36 at peak GIC time. A next chapter sees the inclusion of the modelling results for one location in central Europe (observatory NCK) where a 1-D conductivity assumption has been shown to be sufficient. Then SUPER-MAG data are used to model global ionospheric current density, of which the northern hemisphere is shown in several figures for several time-steps.
I find that there is a lot of very interesting content, but that the overall structure of the manuscript is confusing with little introduction and explanation why certain aspects are pulled together. The line of thought is sometimes lost when observational results (what really occurred during the May 2024 storm) is alternated with very speculative aspects (moving the Finnish power grid to the Norwegian coast, and the northwards shift by 1° that is not very well explained). Some of these excursions in the argumentation are very unrealistic and could be omitted to shorten the text and make everything more concise and comprehensible. Maybe data availability to validate the models is a problem, but this isn’t discussed. Why was BEAR data used for one site if much newer MT data from across Scandinavia should be available? The example from the Hungarian observatory is also very far-fetched. Continuous electric field observations (including the May 2024 storm) are for example available from GFZ Potsdam for a site in north Germany. The quantification of model performance is also very broad-brush – can the authors give better estimates for the comparison of e.g., GIC derived from EDF or E total?
I would encourage a shortening and restructuring of the manuscript, leaving out the most extreme hypothetical examples and strengthening the introduction and overall flow and argumentation of the text. Some of the figures need improvement and figure captions should be made more comprehensive.
General remarks: