Bittig et al developed a new method to construct a pCO2 climatology in the Baltic Sea. The method is novel, clever and provides a clear and significant improvement towards what currently exists and the results are useful for multiple applications stated in the MS itself. I have a couple of comments below that I would like the authors to consider, which I believe would strengthen the MS. Overall though, I do recommend publication of the MS.

The MS itself currently has some shortcommings that I would like the authors to address:

1. Model comparison: If I follow the methodology correctly, the basis of the method is formed by model pCO2 and its variability, which is then (in my own simple words) corrected by observations. As a reader, I would be curious to see a comparison between the actual model output and the pCO2 climatologies obtained. Are they very different? 
2. My previous point would also to some sort explain how dependent the method is on the initial model state. If the final pCO2 established through the method (steps 2 and 3 in their Figure 1) is not far away from the original model pCO2, one could conclude (a) that there is not sufficient observations, (b) the model already does a perfect job, so why performing steps 2-3 in Figure 1 or (c) the model is very sensitive to the baseline model pCO2 (step1 in Figure 1).
3. The same is true for trends. Are model trends significantly different from the trends obtained in step 4 of Figure 1?
4. The authors in-depth discuss methodological differences with other climatologies (e.g. Becker et al 2021, Parard et al 2016) but there is no actual comparison. As a reader I would be very interested on how this new climatology compares with the state of the art in terms of output
5. In a method-heavy publication, it is difficult to not overload the reader in the main text, while still maintaining the required detail for reproducibility. I believe the authors have done a great job offering enough detail in the main text so that the reader can follow, while providing all necessary detail in an appendix. I would have like though a few more lines in the main text on how the error calculation is done.
6. Evaluation: Lines 177-178 “We therefore stepped back from a quantitative attempt but provide just a qualitative picture (Fig. 5).” I dont understand this statement. From a readers perspective you dont gain more from Figure 5. And - despite all shortcomings in a quantitative assessment - this is a new method that may be used in high profile applications (UN stocktake, CDR activities, extreme event studies) so it behooves the authors well to be as quantitative as possible about their approach and potential errors/uncertainties. I believe the authors should do as many evaluations/comparisons as they can, highlighting obviously the caveats attached to them.
7. A smaller remark on line 16: Maybe state what year or decade the anthropocene started - I had to look this up as I was not aware of it (and it makes a difference whether the ocean absorbed 25% or more)

Review of « A regional pCO2 climatology of the Baltic Sea from in situ pCO2 observations and a model-based extrapolation approach» by Bittig et al. submitted for discussion in ESSD

This manuscript reports a new pCO2 climatology for the Baltic sea which is based on a new extrapolation method of pCO2 observations. The presented dataset (actually only available through a temporary link) is composed of a spatially gridded climatology of the surface water pCO2 of the Baltic sea basin at a resolution of three nautical miles based on data collected between 2003 and 2021. A mean average linear trend of pCO2 over the considered period at the same spatial resolution is also presented.

Theses estimates are based on the conjunct use of two datasets:

- Surface pCO2 measurement values from the SOCAT version 2022

- pCO2 estimates from a model (ERGOM version 1.2 ) tuned for the Baltic sea.

The interpolation method presented in this manuscript is original. It is based on a EOF decomposition of the model dataset to obtain spatial patterns of pCO2 variability which are then constrained through an optimisation process with the observational values. The strength of the method relies on an ensemble approach which allows an uncertainty estimate for each grid cell.

My general opinion is that this is a very interesting manuscript. It reports an original and robust method to extrapolate data. The manuscript is well written and is supported by proper illustrations. It certainly deserves publication. I have only minor concerns which I hope can improve this overall good manuscript. I particularly appreciated the fact that most of the details have been wisely presented in the appendix allowing the reader to have an easy to read main manuscript.

General comments :

I regret that the final data product (Climatology and long term trend gridded product) ois not clearly presented. I would appreciate a small section describing the dataset (Format, Size, units, etc).

The section 4.2 of the discussion (« Comparison with other pCO2 mapping approaches is not necessary ») is interesting to read but it does not give some discussion elements to compare the presented dataset with other climatologies. It is a general discussion on mapping approaches which could be added in the introduction.

Section 4.3 of the discussion (« Baltic Sea pCO2 climatology ») is discussing the reasons that have led the authors to produce a climatology rather than a monthly gridded product over the entire period. I do not disagree with these arguments but I believe that they could be presented earlier in the manuscript (Section 2 for example).

Section 4.4 of the discussion (« Long-term pCO2 trend ») could be simplified by adding a table which would compare the trend in this study to other studies in the Baltic sea.

Specific comments :

L26 : What is meant by « smart extrapolation approaches» ? I would suggest just mentioning “extrapolation approaches”.

L 150-151 : « We can not observe a tendency of the mapping approach to give extreme values or outliers in absence of observations (compare Fig. 4a). » May be I wrongly understand this sentence but I am not sure to understand how this figure is showing this. This need to be clarified.