The authors present an analytical calculation for the shell currents in a thin-layer model for the ionosphere. This is in principle an interesting topic, relevant to this journal, and the paper is generally well written. However, there is one major premise that the authors should better justify: the use of a tanh function for B_Φ. The main results of the paper depend strongly on this shape of B_Φ. For example the conclusion that the currents are maximum in the middle of the layer is to be expected from this assumption by arguments from symmetry.

It is true, as the authors mention, that this step-up function is widely used in various models. However, it is usually employed in the case of homogenous media. The thermosphere/ionosphere system, even when only considered between 100 and 400 km, is certainly not homogenous. There will be significant vertical gradients in neutral and plasma densities, temperatures, composition, etc. Thus, it is not evident why a highly symmetric function such as tanh is appropriate here.

The manuscript titled "Determination of Vertical Profiles of Shell Currents in the Ionosphere" by Romashets and Vandas extends their work published as Romashets and Vandas (2020) and proposes an alternative approach for the modeling of shell currents in the ionosphere. This paper is well-written and the mathematical equations are properly derived. As the topic of the manuscript is within the scope of Annales Geophysicae, I recommend it to be accepted with a minor revision.

Minor comments:

1. My suggestion is that the authors can include an appendix of the symbols that they have used for the equations and include a brief description of them for the ease of the readers.

2. My second comment stems from the fact that the authors have not presented any comparison of their results from other manuscripts. Has the shell currents in the ionosphere not been derived from any other methods in literature or the approach that the authors have presented is the only one so far? If any such approach has been mentioned previously in literature, then the authors should compare their results and present the benefits of their approach. For example, is their approach leading to better accuracy or is it easier to implement or is it better in terms of comparison with observational data? I found this aspect to be missing in the discussions so far and this part should be worked upon as it will give a broader perspective to the readers.

Apart from these comments, I found the manuscript to be free from typos and grammatical errors and hence I recommend it to be accepted after these revisions.