Dear authors,

Thank you for providing responses to the review. Both reviewers made constructive feedback.
While I notice that the authors responded to the minor comments of R1, I acknowledge that they have not responded to the major points which I copied below:
"My major concern relates to the effects of pollarding. It is still quite unexpected to observe such a small difference between the uncorrected and corrected RWI and LWI chronologies. As mentioned in the discussion, I agree with the authors that this is most likely due to the high replication and the asynchronicity of pollarding events among trees. However, this asynchronocity among events is not shown in the manuscript. I believe it would be valuable to include a figure showing how many trees were affected by pollarding over time."
Also, it is hard to evaluate whether the authors truly addressed the comments of Reviewer 1. For example, the update made to Figure 1 is not shown.

Responses to the comment about Detrending is not well-substantiated, and responses are based on belief is not a strong rationale for why the cut-off was set.
Keeping the chronology as is, because the authors feel comfortable is not a scientific approach. I would advise the authors to be more careful in writing their response statement.

Responses to the comments by Reviewer 2 is more detailed, and I thank the author for being diligent in their responses.

With this said, my decision is "major revisions". Please consider the comments above, very carefully and please indicate clearly in the revisions (with marked changes) where the edits were made. When your revised manuscript is received, I have it reviewed again.

Lastly, please make sure that data used in the manuscript are cited if not original from this work, or had obtained permission to reuse from the original researchers.

Best wishes,
Voary

Dear authors,
Thank you for revising your manuscript. After sending it to reviewers, they found your revision to be satisfactory, pending minor revisions. I agree with the reviewer's evaluation.

Copied below are the comments for your information. Please makes sure to provide link to data if they are available somewhere.

Thank you for your careful attention to the comments.
-----------
The revised manuscript is substantially improved. The authors have addressed the central concern raised in the previous round regarding the effect of pollarding on the climate signal. In particular, the new supplementary figure showing the temporal distribution of pollarding events supports the authors’ interpretation that management effects were largely diluted at the stand level because events were highly asynchronous among trees. Before publication, however, I recommend minor revisions, mostly related to clarity, reproducibility, and consistency.
Main comments
1. The detrending sensitivity analysis should be included in the manuscript or Supplement.
The response to the editor presents a useful comparison of different detrending approaches, including the reviewer-suggested alternative. This is exactly the type of evidence needed to justify the selected method. However, the analysis currently appears only in the response document. Because detrending choice was a major review concern, the comparison should be added to the Supplement, with the acronyms defined clearly, including Fixed-50, ADC/ADCS, PT, and non-PT. The authors should also state briefly in the Methods or Supplement that the correlations with FIC and CRU are not significantly different among detrending choices, preferably reporting the Fisher z-test result or a concise summary.
2. Please clarify the reconstruction workflow and bias correction.
The revised manuscript now includes a cross-dataset calibration-validation framework using FIC for calibration and CRU for validation, followed by final calibration over the full FIC period and quantile mapping. This is a reasonable approach, but the order of operations could be explained more transparently. In particular, the authors should distinguish between:
(1) the raw regression reconstruction used to assess model skill,
(2) the quantile-mapped reconstruction used to improve the distribution and identify extremes, and
(3) the independent validation against CRU.
The current text sometimes gives the impression that these steps overlap. I recommend adding a short workflow paragraph or schematic explanation in Sect. 2.4.3. Please also specify exactly which period was used to train the quantile mapping correction and whether the correction was applied unchanged to the full 1649-2023 reconstruction.
3. The detrending-method sentence needs revision.
The Methods currently state that age-related growth trends were removed using “cubic smoothing splines with a 50% frequency cutoff set to two-thirds of the series length.” This is difficult to parse, I think. Please rewrite this sentence so the detrending choice is unambiguous. For example, state whether each individual series was detrended using a cubic smoothing spline with a 50% frequency-response cutoff at two-thirds of the individual series length, and whether this was applied identically to RW, EW, and LW.
4. Ameliorate the data and code availability statement.
The current statements, “Code will be made available upon request” and “Data will be made available upon request,” are not ideal for reproducibility. Copernicus requires a data availability statement explaining how the underlying research data can be accessed, and recommends depositing data and related metadata in FAIR-aligned repositories with persistent identifiers. Copernicus also encourages software, algorithms, and model code to be deposited in FAIR-aligned repositories and cited with a DOI where possible.
At minimum, the authors should provide repository links or accession information for the tree-ring chronology, reconstructed precipitation series, and scripts needed to reproduce the main analyses and figures. If FIC data cannot be redistributed, the access conditions should be stated clearly.

Minor comments and editorial corrections
The new pollarding asynchrony figure is helpful. In the main text, please specify that the reported percentages are calculated relative to the number of trees available in the chronology in each year.
Please check figure captions for final wording. Figure 5 should clearly distinguish raw reconstruction, bias-corrected reconstruction, instrumental FIC/CRU data, calibration period, validation period, and prediction intervals.
Please clarify whether “extreme” dry and wet years are defined by the ±1.5 SD criterion, the 5th/95th percentiles, SPI categories, or by agreement among all three. The revised manuscript now uses all three approaches, which is good, but the wording should avoid implying that there is a single definition unless one is selected as primary.

Dear authors,

Thank you for considering the comments, and making further corrections. I am happy to accept your contribution for publication in Biogeosciences.

Best,
Voary