Review comments on the manuscript egusphere-2024-2831, entitled: Estimation of the 3-D geoelectric field at the Earth's surface using Spherical Elementary Current Systems

 by Liisa Juusola et al.

The authors tried to derive the geoelectric field at the Earth's surface from magnetic field variations measured in the vicinity. The various components of the E-field are estimated with the help of the SECS approach and by using the 3-D induction model PGIEM2G. This approach is applied to IMAGE Magnetometer Network area. Convincing results are obtained in this way, which compare quite favorably with GIC measurements in gas pipelines. The computational design of the framework is suitable for running it in near-real time for estimating space weather hazards, resulting from GICs in the Fenno-Scandian region.

In spite of these generally positive ratings, the study would gain, when improvements were made in a number of cases.

Open issues

1. One thing, the authors obviously have not taken into account is the effect of prompt penetrating electric fields on the geoelectric field. As shown by Brändlein et al. (2012) doi:10.1029/2012JA018008, the ionospheric Hall current, driven by the prompt penetration field, causes ground-based magnetic signatures, but it does not cause geoelectric fields on the ground. At mid-latitudes significant effects of this process can be observed. I am not aware that anyone has studied this effect at auroral latitudes. This point should be discussed.
2. Table 3: Larges ground E-fields are predicted at the end of 7 Sep. 2017 for a location close to the transformer of Namsos. It should be checked if measurements of ground currents are available at that station. In case there are, they should be compared with the predictions. This would make the study much more convincing and relevant for application.
3. In the Introduction it is mentioned that a second layer is introduces below the Earth's surface. From the following sections it is not clear what this extra layer physically represents. How does it account for lateral conductivity variations?
4. Another statement is that the radial component of E_DF is not required to be zero. What is the effect of that assumption? What does it physically imply? These two latter assumptions are pointed out as important assets of the presented approach. Therefore, they should be better explained to the readers.

The authors describe a new set of techniques to model the geoelectric field using curl free as well as the divergence free geomagnetic field. They work through a series of simplification of Maxwell's equations to derive the relationships and point out interesting insights into the induced geoelectric field properties. The model does require a good representation of the ground conductivity which can be a limitation for many other locations. Overall this is an excellent contribution to the research area and will be interesting to apply in locations outside the Scandinavian region.

Minor comments:

Abstract: I would not have a citation embedded in the abstract ((Kruglyakov & Kuvshinov, 2018)

Line 9: coefficients

Line 17: with orders of magnitude

Line 24: , a solid understanding

Line 25: scarce

Line 28: A couple of more linking sentences would be useful. E.g. To achieve an intercomparison of results we ... "do things ..."

Line 52: surface

Line 150: You make an excellent point about the induced fields tending to cancel each other out.

Line 188: geoelectric

Line 300: It is not entirely clear at this point that the SMAP model with PGIEM2G is a prerequisite for the modelling to work to compute CF from DF. Can you clarify that here?

Line 347: good data are available

Figure 3 caption: Last sentence says Bx, By, Bz but that is -B_theta, B_phi, -B_r rather than r, theta, phi as written.

Figure 5: Conductivity is in a diverging blue-white-red color scale - could you change it to a linear one (i.e. no white in the middle). This applies to other figures or plots with linear increasing rather than positive/negative variations

Figure 10: similar comment about linearly increasing colors. Also there doesn't seem to be any red in the plots.

Figure 18: the label on the colorbars are not legible