See attached file.

In this article, the authors present a methodology to automatically compute Energy Magnitudes (Me). They apply the methodology to the GEOFON catalogue (2011-2023). The authors present several quality checks and statistical analyses of the dataset analyzed. They further compare their Me estimates with those made available by IRIS. The codes used both for off-line and real-time computations are made openly available.

The article if clear and well-written, and represents an important contribution by adding an additional magnitude estimation to the reference Geofon catalog. I suggest only minor revisions below:

Abstract: Maybe add a brief sentence explaining the added value of Me estimates?

Line 24: “… the low frequency end…”, in practice we often measure Mw from the low frequency end of spectra, but really it represents the static (f = 0 Hz) component. Maybe just add that?

L26: Correct to “… fraction of the total energy being radiated…”; “energy” is currently missing.

L 29: “parameter”; singular, not plural.

L 36: I’m not familiar with the methodologies to compute Me in detail, so I was a bit surprised to read that you compute Me from P waves, in opposition to S waves, which carry most of the energy. I guess it’s related to the SNR. For the more unaware readers, maybe add a brief explanation on why you compute Me from P waves?

L 44: Distance range: 20˚ to 98˚? I guess 98˚ is related to the P-wave shadow zone. Why disregard near source recordings? Add a brief explanation, again for the sake of the more unaware readers.

L 47: Suggestion: add “each” before “single station”

Eq 1: Please double check this equation – is it dimensionally correct? Maybe I’m missing something…

L 53: Clarify what is “a wide range of plausible focal mechanisms”

L 55: How much is “just before” the P wave arrival?

L 57: Correct to “a single event-station pair”

L 66: Can you provide a rationale for starting in 2011?

L 86: Wouldn’t you want to take into account static station corrections, once you’ve analyzed a large enough dataset? It seems like that would provide more robust Me estimates. It’s a common correction when computing ML.

L 87: 246 networks: Do you have a smart way to cite the DOIs of all those networks in your work??

L 90: It was surprising to me that you find entire networks outside the 5-95 percentiles. Wouldn’t it be enough to exclude stations outside the 5-95 percentiles? It’s not very clear to me why you need to exclude entire networks.

L 96: “Anomaly score”: maybe give a brief explanation on the grounds on which this method flags anomalous amplitudes?

Datasets D0, D1, …: it’s not clear to me if you apply cumulatively or independently the quality criteria D1 -> D3. Please clarify.

Figure 3: Can you overlay the dataset D6, in front of the black dots and behind the lines? In case the figure doesn’t become illegible, it would be nice to see how much we lose from D3 to D6.

L 127: Change to “with the intra-event, equal to…”

L 127: phi = sqrt(…) , I believe a square (^2) is missing in the last parameter, phi_S.

L 137: East African Rift: active, but cratonic…

L 149: ”Similar to our approach” instead of “Like us”

Fig 6e: Do you really need to use a log x scale? A big part of the plot is empty…

L 166: “on the analysis”

L 174: I believe it should be BB instead of HF, right before “, i.e.,”

Figure 8: Very nice! It shows a lot of inter-station variability…

L 180: Is it really 50 deg? All other styles of faulting have 60 deg.

L 180: Just write out that OF means “other faulting styles”

L 196: “, where different faulting…”

L 197: It was a surprise to read here that the method does not consider radiation pattern. When in line 52 you write that G(f) is computed for a range of plausible focal mechanisms, I though you took into account the focal mechanism, therefore the radiation pattern. Please clarify in the text. It seems like you should take into account the focal mechanism/radiation pattern, to get better Me estimates…

Fig 11: Maybe easier to read in a 2 rows x 4 columns plot? Top-BB, Bottom-HF.

Finally, it would be really nice to have a “Conclusions” section, where you summarize the main take-away lessons from your new Me catalog. Why is this catalog useful? What new things are we learning from it?

Great work! Thank you for this contribution.