General comments:

This is a well written and interesting study on the origin of numerous observations of blue/green suns in 1831. There will certainly always be some uncertainty, but in my opinion this is a convincing study and according to the evidence presented in the manuscript it seems indeed highly plausible that the 1831 eruption was the one of Ferdinandea. The paper is well suited for Climate of the Past and I recommend publishing the article after considering the following (mainly minor) comments.

Apart from the specific comments below it would be good if the paper would also briefly address the following points:

• Were there reports of wildfires in Europe in August 1931?
• How does the time lag between the eruption and the occurrence of colored suns differ between Ferdinandea and other eruptions, e.g. Krakatoa?

Specific comments:

Lines 77 – 80: Perhaps you can here already refer to the appendices. When reading these sentences I was asking myself: How many newspapers and journals were searched and which ones? This information is provided later, but it would also fit here.

Line 122: “These observations have accordingly not been included in the present analysis.”

Probably because Fig. 1 is right below this sentence I was asking myself, whether the point in China refers to the observations of this Mandarin compendium (which is not the case, as I learned later).

Figure 5: what is the meaning of the color of the circles. This is not explained as far as I can tell. In particular: What is the black circle in the "no blue sun" area?

Fig. 6: I think this Figure is appropriate for this manuscript, but I would add a disclaimer that this is only a very crude depiction of the plume extent and motion.

Line 194: “yields a transport rate of about 0.97 deg hour^-1”

This refers to longitude, right? I suggest mentioning this explicitly to avoid confusion.

Section 3.5: I would add that zonal winds in the stratosphere at mid-latitudes depend on the seasons and are easterly (westward) in summer and westerly (eastward) in winter. In August the winds would be easterlies, i.e. westward winds, which supports your hypothesis. The zonal wind reversal in the middle atmosphere typically occurs in September. This information can be found in a standard text book on atmospheric dynamics.

Equation (1): what are the limitations of this equation? What assumptions is it based on? There must be limits to the parameter ranges in which the equation is applicable, e.g. if Delta M ist very small, tau may be negative, which does not make sense.

Line 242: “These three observational phases”

It’s not entirely clear which three phases you mean here.

Line 258: “and using the corresponding magnitude range (either 8 < \Delta M < 12 or 12.5 < \Delta M < 16.5)”

Which range is used in which case?

Figure 8 and related explanations in the text: I don’t fully understand, how the optical depths and their error bars are estimated. You use equation 1 and obtain information on the solar zenith (or elevation angles), but how is \Delta M determined for each case? This seems highly arbitrary and should introduce significant uncertainties.

Please also explain, how the error bars in Fig. 8 were determined.

Line 289: “extinction co-efficient Q”

Q needs to be dimension-less for equation (2) to yield the correct units. This already implies that Q is the “extinction efficiency”, not an “extinction co-efficient”. Coefficients typically have units of 1/length, X-sections of length^2 and the Mie efficiencies (scattering or extinction) are dimension-less.

Line 340: “The eruption has been assigned a VEI of 3, which is associated with a total volume of erupted tephra of the order of 0.1 km^3.”

This amount of erupted tephra is certainly probably associated with large uncertainties, right? I suggest mentioning this.

Line 364: “This hypothesis”

Which hypothesis do you mean here? That Ferdinandea was the source of the aerosol leading to the blue sun observations? Or your hypothesis on the additional release of sulfur?

Line 421: „This hypothesis“

Again, which hypothesis?

Line 441 – 443: As mentioned above, the zonal winds in the stratosphere in July/August will most likely have been easterly/westward, in good agreement with your hypothesis.

Citations (occurs many times in the paper): “et al” -> “et al.”

General Comments

This paper presents the hypothesis that an important climate forcing event in the 19th century and seen in the records of Greenland ice cores, was due to the eruption of an undersea volcano (Ferdinandea) off the coast of Sicily that occurred in 1831. 

This is an innovative, interesting and scholarly research paper and it should be published. It is innovative in its use of "eye-witness" observations of certain atmospheric optical phenomena associated with volcanic aerosols to estimate size, composition, transport and even the amount of S in a little known volcanic eruption.  It is interesting because it reveals new information on processes affecting the injection of aerosols into the stratosphere as well as perhaps solving the mystery of a sulphate signal in the Greenland ice core record thought to be from an 1831 tropical eruption.  It is scholarly (there are > 120 references) as it involves collection and research of historical documents and records in several different languages, as well as covering topics in volcanology, atmospheric optical physics and dynamics and even some radiative transfer.

Specific Comments

The hypothesis presented relies on three key points:

1. That a small (VEI 3) volcanic eruption could inject aerosols into the stratosphere.
2. That the event injected sufficient SO2 to register a signal in the Greenland ice core.
3. That the aerosols were transported westwards in agreement with the (indirect) observations presented.

I cannot comment on point 2 as I am not sufficiently expert to assist the discussion, but I can provide some thoughts on points 1 and 3.

Point 1. The idea that an undersea, relatively minor eruption could produce stratospheric aerosols seems surprising.  However, as the authors note, the eruption of Krakatau in December 2018 did just that.  More recently, the 12 August 2021 eruption of the undersea volcano Fukutoku-Okanoba, off Japan generated a column that reached up to 18 km (in the stratosphere).  These eruptions were monitored by modern satellite instruments and provide unambiguous evidence of stratospheric injection.  The 1963 eruption of Surtsey was reported to have generated a column up >9000 m which is stratospheric at the high latitude of Surtsey.  Thus there is evidence that near sea-surface volcanic eruptions can generate tall columns, even if the amount of solid-rock material is not large.  The mechanism for this is thought to be due to a combination of tropospheric convective instability and the interaction of hot material with sea water, generating additional  convective available potential energy (CAPE) which drives the buoyancy of the column.  I think the authors should perhaps describe this in more detail as they must convince readers that a small VEI eruption can generate a stratospheric aerosol – crucial to their hypothesis.

Point 3.  The authors estimate the zonal transport of the aerosol to be ~20 m/s westwards.  This is quite fast.  There are no detailed vertical wind profiles available for 1831 but there are modern climatologies and also good knowledge of the atmospheric circulation based on solid meteorological foundations.  Eruptions from Mt Etna (a volcano close to Empedocles/Ferdinandea) generate volcanic plumes that predominantly move eastwards and sometimes northwards and southwards, but rarely westwards. This is because these eruptions are mostly tropospheric where the winds in summer are from the west.  Modern climatologies of the vertical zonal winds suggest that up to the tropopause at 40 N during the NH summer winds are from the west.  Above the tropopause the winds gradually shift to the east in accordance with thermal winds caused by the vertical gradient of temperature. As the sign of the gradient changes from negative to positive the winds change from eastwards to westwards.  But, according to modern zonal wind climatologies for midlatitude NH summers, the magnitude of the westward winds is much less than 20 m/s.  At 70 hPa (~18 km) it is generally <10 m/s and does not reach ~20 m/s until 10 hPa (~30 km).  It is possible that in August 1831 the stratospheric winds were anomalously high, but that would be somewhat speculative and convenient.  I wonder then whether the authors should re-evaluate the speed of transport or perhaps put an error range on their estimate to allow for this inconsistency.

None of these points are serious enough for me to suggest a revision is needed and I am happy to recommend publication subject to technical corrections.  Indeed I found the paper so interesting that I wondered whether the authors had exhausted all observations, such as reports from ships logs (perhaps too scant?), were pumice rafts observed? or could the possibility that the aerosols reached the upper stratosphere be sustained (where there are stronger winds), in which case one might expect observations of PSCs during the NH winter.  I believe the earliest documented evidence for PSCs dates back only to the 1870s. 

Minor comments 

These are really quite minor.

1. I don't think it is necessary to have a full-stop after the longitude/latitude directions (e.g. N. should be N)
2. Extinction coefficient and extinction efficiency factor are both acceptable but just use one. I think extinction efficiency is most commonly used.
3. I suppose that sometimes observations were hampered by cloudiness. One cannot expect a particularly complete set of observations of the Sun so I think some inconsistencies could be explained by lack of visibility.
4. Were there any lunar observations? (The Moon was 43% illuminated and 23 days old on 14 August 1831).
5. I noticed in Table 1 A22 that there was mention of a "black dot" observed on the Sun. Could this be a sunspot? If so,it adds weight to the drastic diminution of light from the Sun, as observing sunspots requires considerable rejection of sunlight.  There are > 200 observations of sunspots made in 1831 (see: https://onlinelibrary.wiley.com/doi/abs/10.1002/asna.201111601) some with notes.  These records are held at the archives of the Royal Astronomical Society in London, so it might be possible for one of the authors to view them and investigate whether anything unusual was noted on solar observations during August 1831.

This is an interesting paper that comes up with a new hypothesis concerning the now-unknown (previously attributed to Babuyan Claro) 1831 eruption. The volcanic signature seen in ice cores is attributed to an eruption in the vicinity of Sicily, which according to its assigned VEI should not have had a big impact on the stratosphere. However, associated cross-tropopause exchange could have injected sulfur from this plume into the stratosphere. The authors present a detailed study of sightings of coloured sun, which beautifully align. They are consistent with a plume that travelled around the globe form East to West in the lower stratosphere. The paper is very interesting and thought-provoking. I think this is the sort of papers that play an important role in the discussions in the community. Combining historical documentary sources and scholarly expertise with scientific reasoning is interesting. This is fascinating and should be published with some minor revisions. However, at some instances this reasoning should be strengthened, as is outlined below.

• The time (and space) between the eruption and the first observations is very small. Sightings of coloured suns were made immediately (days) after the eruption and near the location of the eruption. This sound plausible, but is this to be expected? We read that a radius of ca. 0.5 um is required for the volcanic aerosols to have this effect. I am not a microphysicist, but in the case of tropical eruptions, sulfate aerosols, forming from the gas phase, need some time to grow to that size. This could take 2-3 months. During that time the cloud would long have circled the globe. Perhaps this is different in this case, but I would appreciate a discussion of the aerosol formation and growth process.
• Is there a role of ash or is ash too large anyway for that?
• How likely is it that “tropospheric” plumes at these latitudes reach the stratosphere due to trop-strat exchange? There are examples from other locations, but is there also evidence from this region, e.g., from Etna eruptions?
• 20 m/s is superfast. Perhaps you can provide some context for this?
• You mention 20CRv3, but then do not show it. I agree that for this region and time period, this data set might not show much, but a plot would still be nice (the data go back to 1806 on NCAR’s website).