This is an interesting paper on the impact of the Southeat Madagascar Bloom on the  oceanic CO2 sink, a problem that has not been previously addressed (to my knowledge). The results are convincing but critically depend on the work of Chau et al. (2021) which is in review. As my expertise is more on the impact of physical processes on the bloom biology I am unable to to make a judgment  on the work of Chau et al. (2021) regarding the reconstruction of surface pCO2 and air-sea CO2 fluxes on which the results in this paper so much depend. Assuming that Chau et al. (2021) is correct / passes the review process and is published, I think that application of that approach to the Madgascar Bloom provides some new insights into its role as a CO2 sink. The  observational aspects of the paper clearly  show the strong interannual variability associated  with  the bloom, but the interpretation of these in terms of CO2 hinges on Chau et al. (2021). In light of this I am happy to recommend publication once the Chau et al. (2021) has passed peer review and been accepted for publication, after correction of a few minor issues (listed below).

Minor comments:

lines 275-277 - given that wind speed is spatially (as well as temporally) variable I am not sure how taking account of spatial variability would affect the calculations described here. Does it matter?

Figure 3 mark positions of eddies

Figure S2 add lat & long on axes, also Figure S6 bottom

Figure S3 mark positions of eddies

This study makes use of in-situ data from the OISO Project, among others, to understand the impact of the South-East Madagascar Bloom on the oceanic carbon sink. This is a first attempt to investigate how this large sporadic austral summer bloom can potentially contribute to the oceanic carbon sink. The physical drivers of this bloom have long been investigated with various possible mechanisms, although it seems to be coming down to a combination of a few factors which can contribute to the initiation of the bloom. On this note, the question of its impact on the oceanic carbon sink were previously raised but the lack of in-situ data hinders any research.

This very interesting study is well-structured and well-written. And with the right data, it provides new insights on the biogeochemical signature of this bloom. It clearly shows the difference between fCO₂ during a bloom and non-bloom (or low bloom) year, and that this difference is due to biological processes during the boom. And that it acts as a CO₂ sink (between -1.7 and -2.7 TgC/month).

Having said that, I am slightly less impressed with the reconstructed fCO₂ and air-sea CO₂ fluxes from Chau et al. (2021), which is still in review. I would expect that the Chau et al. (2021) is acepted before the current manuscript. I am also a bit puzzled because from Figures 11, S10 and S11, it seems that the impacts of the previous bloom years (1999, 2000, 2004, 2006, 2008, 2012-14) on the reconstructed fCO₂ and air-sea CO₂ flux are almost non-existent, whereas a significant drawdown is found for the 2020 one. I think that this variability from the CMEMS-LSCE-FFNNN is interesting, and deserve to be included and possibly explained in the text.

However, these do not take away the importance of in-situ data in this data-limited region and I am sure that the few comments can be easily addressed by the authors and that might help to improve this already good paper. Thus, I recommend this paper for publication, once the comments have been addressed, and the Chau et al. (2021) paper accepted. 

Minor Comment:

On lines 318-319, the authors mentioned the presence of a clear signal of the SEMC retroflection. A recent paper by Ramanantsoa et al. (2021) discussed the early-retroflection, retroflection and no retroflection of the SEMC, and the impacts of the early-retroflection on the SEMB. I recommend including this citation is the discussion.