This manuscript brings much wanted information about the Southern Ocean particulate biogenic Ba (Baxs) distribution during winter conditions. Previous studies in the S.O. were all conducted during spring to autumn conditions, showing a seasonal progress of the Baxs signal but lacking information on winter conditions when plankton activity is at a minimum. The mesopelagic Baxs inventory is gauged against integrated PP covering the growth period preceding the sampling, and these data are combined with literature data revealing an interesting correlation.

I wonder why authors, when comparing their data with literature, have considered data from specific expeditions and not from all available data for the S.O. . In particular the Baxs data presented in Dehairs et al. (GBC 1990; INDIGO 3 expedition, 1987) for the same general area as studied by the present authors were not considered. During the INDIGO 3 several stations were occupied along approx. 30°E between 65°S and 57°S. S to N Baxs inventories are similar to values reported in the present ms., confirming indeed that microbial activity in the mesopelagic area is still ongoing during winter period. Further data that have not been included in the Baxs inventroy – PP comparison are those from Dehairs et al. (1997) obtained during Polarstern ANT X/6 expedition along 6°W in early season. If possible these two data sets should be included in the compilation.

Authors do not provide information how integrated PP was obtained for the compilation of literature data. Figure 1 should differentiate the different Baxs data sets.

Authors do not provide any information on sea ice extent relative to position of the southernmost station.

While it makes sense to compare Baxs inventories with PP intensity in the months preceding the sampling, the coinciding Chlorophyll data shown in Fig. S1 still are relatively elevated reaching about 0.5 µg/L at the the PF and in the SAZ, this taking into account that S.O. Chl values of 1 µg/L can be considered bloom values. Please comment.

Not sure Figure 4 adds to the understanding. This figure could be ommitted.

Specific comments

Line 67, page 3: Dehairs et al. 1980 more appropriate as ref. here than dehairs et al. 1997.

Lines 58-59: Surface export is set by the deficit (not excess) of 234Th activity vs. 238U actvity. Specify that 234/238 ratios >1 can occur below the upper 100m, or so, reflecting remineralisation.

Lines 26-28 page 10 and lines 75-76 page 16: Sample numbers n=39 (SPF) and NPF (n=31) pertain to what ? Table S1 shows only data from the present study and not the compilation data set.

Lines 46-47 page 11: Opposite gradients of Baxs and diss. O₂. Please provide more detail.

Lines 65-79 page 16: These sentences are confusing. If there is no significant difference in relative amount of POC remineralized relative to PP (all stations except STZ), then there is no difference in response of Ba relative to PP at the different stations .. ? Only the STZ site behaves differently.

Also, high productivity, low export can be associated with large particles in the surface layer (see Lam & Bishop, 2007). High surface water productivity associated with low export has also been described in Jacquet, Lam, Trull, Dehairs, DSR II, 2011. The possibility that high phyto biomass attracts more grazing and is more depending on recycled production (NH4 based) and thus results in smaller export (more surface water recycling) and possibly lower mesopelagic Ba, is reported also in Dehairs et al., 1992.

Line 76 page 16: “.. are comparable to surface export efficiency obs. In this region …” which region ?

Line 88 page 17: this sentence is unlear. Similar latitudinal trend (of what?) ; higher values NPF (values of what?)

Line 90 page 17: saturated vs undersaturated: specify saturation for dissolved Ba.

Table S2: Add lat. position for each site

Figure S1: top panel indicate AZ, PFZ, SAZ. Why is the STZ station not reproduced here ?

Table S3: Add the Lat-Long range for the basin regions

This manuscript presents a new data set of excess particulate barium (Ba_xs) concentrations in the Southern Ocean during winter conditions. Correlation with integrated PP and data from literature is interesting. My major comment concerns the conclusion that the Ba proxy would have a longer timescale than previously thought. I don’t think that there is a cumulative effect on the Ba_xs signal for the reasons explained below. I suggest that authors revise their discussion (and reformulate abstract & conclusion).

• Chl a data reported in Figure S1 indicate that stations (northern 50°S) experienced production (even of low intensity). This should be compared with Chl a and Ba_xs data from other campaigns (e.g. Cardinal et al., 2005; Blain et al., 2007). The winter period appears to be productive in this sector. This would explain why Ba_xs present similar contents as reported during other seasons.
• Data should be compared to results from the Indigo3, EPO2 and ANTX/6 cruises.
• Line 15-19: Please revise the abstract (and part of the conclusion). I don’t think it’s a question of timescale and cumulative effect. If POC is produced in surface and that remineralization is sustained at mesopelagic depths, Ba_xs will be produced, independently from the season. There is no clues that POC material could accumulate at mesopelagic depths and conducts to latter (weeks to months after the growth season) remineralization and Ba_xs
• Figure 4: not necessary -it does not add to the understanding. It should be (in-depth) compared to contrasts reported in Jacquet et al. (2011; SAZ-SENSE cruise): diatoms vs. flagellate, PP, EP, Fe depletion or enrichment, type of aggregates. The effect of the contrasts on Ba_xs and remineralization during SAZ-SENSE was opposite to these reported during KEOPS (Jacquet et al., 2008) and EIFEX (Jacquet et al., 2008) cruises. This should be compared to the present data set.
• Line 71-82 p11-12 (and Line 90 p17: not clear, please reformulate). Are dissolved Ba and SI available? The SO is globally undersaturated (SI<0.9) or at the equilibrium (0.9<SI<1.1) with respect to barite. Saturation is unusual. Please correct it line 73.
• Also, in productive situations (and deep POC transfer), it is common that Ba_xs at 1000 m depths remains larger than the “180 pM" SO reference.
•  Finally, as recently reported in Jacquet et al. (https://doi.org/10.5194/bg-2020-271; Peacetime cruise) remineralization at mesopelagic depths could be restricted to the upper mesopelagic layer or extend up to 1000 m depending the system functioning during a same season. This leads to major differences in the Ba_xs background reached at 1000 m depths.

Please revise your discussion and conclusion according to these comments.

Review. van Horsten et al. “Early winter 1 barium excess in the Southern Indian Ocean as an annual remineralisation proxy”.

The authors describe particulate Barium, O₂, and potential density profile data from 7 stations in the Southern Ocean along from 59°S to 41°S crossing the Antarctic polar front (51°S) along 30°E south of Africa during GEOTRACES GIPr07 (in early wintertime conditions, June 28-July 13). This is a hard to get and interesting data set. The hypothesis is that since particulate barium should have only a short residence time (days to weeks) in the water column the inventory of particulate Ba would be far lower at times of low productivity that at other times of the year. The authors report particulate Ba concentrations as high as seen in other seasons and infer an active biological carbon pump year-round. The stocks are regressed against annual mean primary production. Comparisons are made with other data sets from the Southern Ocean.  

What I like about the work is the heroic effort to achieve sampling in the wintertime and the excellent primary data arising from the expedition. Also, the goal of finding the correct transfer function relating the inventory of particulate barium in the mesopelagic (an indicator of export) to remotely sensed biomass or primary productivity would be a big plus.

That said, the paper falls short of its goals. The regressions in Figure 3, and manuscript discussion provide no insight. The data south of the polar front are aliased by cloud obscured retrievals of surface chlorophyll and primary productivity (See e.g., Ocean color monthly composites) and fall on a different slope than north of the PFZ. The STZ station is an outlier. The discussion does not sufficiently unify these divergent observations.

Reflecting on other reviewer comments, I am convinced that a more comprehensive analysis of the data (now abundant) from multiple projects need to be considered. I am sure that everyone referenced would have data to share. I echo a need for a fuller hydrographic and dissolved phase framework for data interpretation – the supplemental data are very sparse. There are some issues: (1) methodology: bottle sampling and the missing large particle fraction, and (2) the hypothesis of expected low wintertime concentrations is premised on particle sinking rates that are far too high (50 m d^-1) for the micron sized particles that comprise the bulk of suspended barite (sinking speed ~0.1 m d^-1).

A couple of issues further complicating review is simply the lack of any access to the more complete data sets from the cruise beyond those used in figure 2 or the partially complete data sets used in figure 3. The cruise data should be available as supplemental data and also submitted to the GEOTRACES archives and DOI traceable.

I think the fundamental logic flaw (see comments below) lies on page 12 in the discussion of inferred barite residence times in the mesopelagic. I don’t see a major advance beyond referenced work and don’t support publication of this paper with its present interpretive framework. I encourage the Authors to look again at the results in a larger framework.  

Figure 4 is not needed. It is out of place.  There is room for more figures....

Some detailed comments follow.

P4 line  08, R2=0.83... This seems like a bad validation of the O2 results.

p 4 line 10. The Python language did not exist in 1982. Please fix sentence.

P 5. Lines 18-20. I assume this was an in-line filter, directly connected to the side spigot of the bottle. State what was done. Also, state whether or not the large sinking particle fraction would be sampled.

p 5. line 29. If varying volumes of water were filtered, the blank will not be a constant value. (Ba(filter)-blank(filter))/volume filtered. to get Ba and error should be the s.d/filter blanks / volume filtered. Or is this what you did? I think the calculation was done correctly as error bars vary in size. Please clarify methods.

ALSO please state the assumed particle size fraction that has been sampled.  There is no evidence that bottles adequately sample the sinking particle fraction.

P 6 Line 63. “The data…” Which data?

P9. Fig. 2: O2 scale too compressed to be useful. Authors should provide complete data as supplemental (not just pAl, pBa, Baxs) and submit as soon as possible to GEOTRACES. Include T, S, sigma theta, o2, nutrients, dissolved Ba…

p10... .Lines 31-33. “When taking into account….”. There is something wrong with this sentence.

p 10. Line 34. Very high values can be associated with Rhizosolienia blooms (Bishop, 1988).

P 12. Lines 84 & 85. Line 94-95… “Residence time of barite in mesopelagic days to weeks. & Particle sinking speeds of 50 m d^-1”.   The large particles comprising the flux do sink that fast; however, the subsurface barite is produced by fragmentation of these particles as they sink. The resulting micron sized barites sink at 0.1 m d-1. Thus, the premise of decay to background on the time scale of days to weeks is invalid. Barites in the mesopelagic would have a residence time (by sinking) of hundreds of days – if not years. The sink for these barites is dissolution and reaggregation. As grazing is reduced in the wintertime then dissolution and sinking would dominate.

I’ve not addressed the detailed discussion further as this point and invalidates the key conclusion of the authors.

jim Bishop (UC Berkeley).

p.s. have a look at Bishop 1989 (attached - since it may be hard to find). The mapped representation of barite stocks is virtually the same as sampled here.