|This paper has improved from its original version, but still requires substantial revision. The methods are unclear with insufficient detail. Comparisons to other work are provided, but these other approaches are also poorly described so the readers cannot actually understand what the comparisons indicate. Uncertainty is inconsistently presented. Handwaving statements about independence of methods are, in fact, untrue in the case of the ocean. This does not provide the reader confidence in the analysis. |
The paper does not lead to a substantial enhancement in scientific understanding, but reads instead like a technical report on a data product. Because of this, I think the paper would belong better in Earth System Science Data than Biogeosciences.
Page 3, line 24 – The GCP separation of % to land and ocean needs to be qualified with the timeframe to which it refers. This is not a steady state partitioning as suggested by the sentence as written.
Page 4, line 11 - The “full contemporary” exchange would include the large, persistent ocean outgassing in tropics, and uptake at high latitudes See Figure 6.1 of IPCC AR5 ch6, Ciais et al. 2013, these fluxes are +-80 PgC/yr. Similarly “Full contemporary fluxes” for the land would be about +-120PgC/yr as in the original version of this MS. So this net flux is not the “full contemporary exchange”, but only the net contemporary flux. The authors need to explain these differences and find appropriate language that clarifies for the reader.
Section 2.1 - The reader cannot understand to what “ensemble members” are being referred when these have not yet been introduced. The “random draw” approach is not clear. Where do the numbers come from in 10x10x…. ??? 2 is presumably the ocean, but this placement in the equation on line 24 does not correspond to equation 1, enhancing confusion. Figure 1 is not understandable. Table 1 could perhaps help, but there are more than 9 “estimates” there.
This presentation needs much work so that the reader can fully understand the approach.
Section 2.2.4 – State here how many final estimates are produced (2 for marine) so that the reader can understand the error propagation. Do the same for the other sections.
Section 2.5 - Yearly variability in precisely which of the fluxes above? The authors should not assume the reader knows what an “inversion” would provide. Also, the Median was used in RECCAP - what justifies the use of the Mean here?
Page 10, line 30 – Support this statement about model tuning with evidence from the literature.
Page 12, line 18 – Define “bookkeeping” – not all readers know this lingo
Page 12, line 23-27 – This section on comparison to Ciais et al (in revision) is unintelligible because the reader does not know what this other paper does and so what does this comparison actually mean? Later it is stated that Ciais uses an “independent” method, but the authors have not described what this method is sufficiently.
Page 13., line 14. – It is untrue that the RECCAP ocean estimates are independent of those used here. Much of the same pCO2 data is found in the Takahashi flux estimate, a major basis for RECCAP estimates, and that in SOCAT on which these estimates are based. While the Takahashi flux estimate is not directly used here, it is not correct that the ones used are fully independent. The authors should not overstate independence.
Page 13, line 29. This comparison does not “suggest” less IAV in this result than in inversion, it directly shows it. Please be direct in wording.
Figure 2 – Uncertainties must be included on this figure.