We thank Referee #1 for having read our manuscript and for giving us his/her comments. Below, we report each comment to reply to them in as orderly a manner as possible. The comments are reported in a normal text while the reply is highlighted by italic text.

The paper may be interesting, but it is not ready for publication. For example, Section 4, introducing the results, is almost impossible to read. There is inconsistency in terminology, the presentation of tables lacks care, there are awkward expressions in English, and many results are not shown. As such, I cannot evaluate the paper at this stage.

Reply: We thank the referee#1 for having read our manuscript. We are working to improve the manuscript following the supplied comments.

In this section:

There is confusion between the terms "series" and "chronologies".

Reply: We will better clarify when we refer to individual series and when to site chronologies.

I do not understand why there are four EWBI chronologies and four LWBI chronologies coming from three sites.

Reply: The four EWBI and the four LWBI derive from the different percentage of EW and LW pixels considered in the analysis for each site. Since for this species there is a lack of standard procedures, various percentages of pixels within the frame (i.e., 25, 50, 75, and 100 %) were considered both for EW and LW as stated at P5/L116-118. However, we recognize that in the present form, this concept could be not clear, thus we will work to make it clearer or consider removing these methodological results, even if we think that they are useful to try to start to delineate standard procedure also for this species but can be considered too specific details for the present work.

Line 189: What does "remaining" mean?

Reply: we will remove the unnecessary and confusing “remaining”.

The table is difficult to read because the second row in column 2 is for EPS*-related values, while in column 4, EPS*-related values are presented in square brackets. There is no consistency.

Reply: It will be modified as also suggested by referee#2 (i.e., *EPS) to make it clearer and consistent. The whole table will be reshaped to improve the readability.

In the same table, there are additional lines labeled "PCA." I do not know why.

Reply: Thank you for highlighting this inconsistency. We will edit the table for clarification: the inconsistent rows will be edited in accordance with what is reported in the text, and we will modify the column name from “Valley” to “Chronology code”.

The paragraph at line 194: To which figure/table do the results refer?

Reply: We will consider adding these results to the supplementary material in the amended version of the manuscript or consider removing these methodological results, even if we think that they are useful to try to start to delineate standard procedure also for this species.

Line 194: What does "from the difference" mean?

Reply: To obtain the DBI series and then the DBI chronologies, differences between LWBI and EWBI series were performed. We will try to clarify the concept.

Line 200: Did you mean to say, "The DBI site chronologies showed similar decadal variabilities"? The term "trend" might be confusing.

Reply: We thank you for the suggestion. We will modify the text accordingly.

Line 201: What does "in quantity but not in time" mean?

Reply: The entire sentence will be removed; it is an unnecessary description of the chronology that can be easily inferred by Table 1 and the Figure 2.

Line 202: Replace "duration" by "chronology length" which is more commonly used.

Reply: The entire sentence will be removed; it is an unnecessary description of the chronology that can be easily inferred by Table 1 and the Figure 2.

"(not shown)" appearing twice in the section is excessive. Such results should be shown.

Reply: We will consider adding these results to the supplementary material in the amended version of the manuscript or consider removing these methodological results, even if we think that they are useful to try to start to delineate standard procedure also for this species.

We thank Referee #2 for having read our manuscript and for giving us his/her comments. Below, we report each comment to reply to them in as orderly a manner as possible. The comments are reported in a normal text while the reply is highlighted by italic text.

--------------------------------

Cerrato et al. explores the value of tree-ring blue intensity – a surrogate parameter for wood density - from European Larch in the Southern Rhaetian Alps as a proxy of past temperature. The use of blue intensity is still in an explorative state and therefore more tree species and sites are needed to bring this field forward. The idea of the current study is therefore compelling and should be of interest to the dendrochronological community and beyond. With this said, I find the state of the current manuscript to be far from being ready for publication. I am sorry to be negative, but parts of the manuscript are at present untidy, there are sometimes important details missing from the text and the meaning of some sentences and paragraphs is not comprehensible.

Reply: We thank the referee#2 for having read our manuscript and for the encouraging words provided. We are working to improve the manuscript following the supplied comments.

For example, the authors write that four EWBI and LWBI chronologies, respectively, are produced in the current work, but they do not give any explanation how this is done given that there are three sites that have been sampled.

Reply: The four EWBI and the four LWBI derive from the different percentage of EW and LW pixels considered in the analysis for each site. Since for this species there is a lack of standard procedures, various percentages of pixels within the frame (i.e., 25, 50, 75, and 100 %) were considered both for EW and LW as stated at P5/L116-118. However, we recognize that in the present form, this concept could be not clear, thus we will work to make it clearer or consider removing these methodological results, even if we think that they are useful to try to start to delineate standard procedure also for this species.

Also, how is it possible to get 16 deltaBI chronologies from four EWBI/LWBI datasets? The authors need to carefully check the manuscript for these inconsistencies or unclarities and correct or clarify before a proper review of the manuscript can be performed. 

Reply: The 16 DBI chronologies derived from the permuted difference between EWBI and LWBI series from which the chronologies were built. We will consider removing these methodological results, even if in our opinion they might be usefulto start to delineate standard procedure also for this species. 

Also, the manuscript requires a careful language check.

Reply: We relied on the service offered by American Journal Experts (certificate validation n. E502-D45C-A2F2-D167-022P). We will report your complaints to American Journal Experts asking for a new language check with different editors.

When it comes to the methodological aspects of the study, I wonder why the authors have decided not to wash the samples in either ethanol or acetone as the standard praxis advocates. Resin and other impurities may pose serious implications for both high- and low-frequency variability in the resulting BI time-series, potentially obscuring the climate-related signal. This issue should be properly addressed in the manuscript and the methodology clearly motivated if the authors should choose to proceed with unwashed samples.

Reply: As stated by the referee in the review opening, the BI is still in an explorative state, and thus in literature sample processing is approached with different modalities. To our knowledge there is no tested standard procedure for Larix decidua Mill., but we followed Wilson et al. 2014, 2019, 2021) who processed samples for BI analysis without washing them. The same approach was followed by other authors where samples were only polished (e.g., Dannenberg and Wise, 2016; Dolgova, 2016; Arbellay et al., 2018; Heeter et al., 2020). On the other hand other papers refer to samples treated with solvents (e.g., Campbell et al., 2007; Björklund et al., 2014, 2015, 2020; Babst et al., 2016; Brookhouse and Graham, 2016; Buckley et al., 2018; Buras et al., 2018; Akhmetzyanov et al., 2020; Cao et al., 2020; Blake et al., 2020; Davi et al., 2021). We discussed this point in the materials and methods section, but we can certainly better underline this point in a new version of the manuscript. In our opinion, one of the strengths and greatest originality of the paper we have presented lies in being the second on BI of Larix decidua and the first approaching the use of BI data from this species as climate proxy. To our knowledge there is no tested standard procedure to follow. However, our results indicate that a climate signal, at least in the high-frequency domain, is retained with high levels of significance albeit the samples were not treated with solvents. We agree with the referee about the necessity to investigate the effect of the resin and other impurities removal, as well as about the influence of the CooRecorder frame dimensions and percentage of retained pixels on the results. In our opinion, determining the effects that impurities in the cores can induce in measurement results deserve dedicated work with properly designed experiments. This is a very interesting point for further investigation, we will focus on in a dedicated paper.

I have listed some specific points below that require either clarification or rephrasing.

P1/ L13: Note that there are other proxies that have similar or higher temporal resolution than tree-rings.

Reply: We will account for this in the amended version of the manuscript.

P1/L18: “The results showed that BI data from European larch share greater variance with June–August mean temperatures than total ring width measurements.” Would be informative to get some numbers here.

Reply: We will add it in the amended version of the manuscript. A proposed modification will be: “The results in the high-frequency domain showed that BI data from European larch explain up to 37 ± 8 % of the June–August mean temperatures variance in the study area, which is 78 % higher than that explained by total ring width measurements for the same period.”

P1/L18: unclear what is meant by “temporal and spatial quite-stationarity” and “regression indices”??

Reply: We will rephrase the sentences. A proposed modification will be: “The BI correlation values with the average June-August temperature are quite stable in time and space. In fact, the correlation values obtained considering a 50-year moving window range between 0.40 and 0.71 (mean value 0.57), resulting significant for the western and central Mediterranean areas for all the time windows considered.”

P2/L45: “Among the climatic proxies that can be used (Trachsel et al., 2012), dendrochronology represents an excellent tool for reconstructing the climatic variations that occurred in the past.” Dendrochronology is not a proxy but the science that is based on chronological dating of tree-rings. Please rephrase the sentence. Also, change “tree-ring-based dendroclimatology” in the next sentence. Dendrochronology is always based on tree-rings and it is thus redundant to state this.

Reply: We agree with referee#2, dendrochronology will be replaced by “dendrochronological data” while the successive phrase will be rephrased to avoid pleonasms.

P2/L59: “innovative methods such as maximum wood density (MXD)” Both TRW and MXD are conventional proxies that have been used in climate reconstructions for decades. I would therefore be careful saying that MXD is innovative or novel.

Reply: We agree with the referee#2 and removed the word “innovative”, making the sentence confusing.

P3/L85: change to Mount Vióz, Mount Adamello and Mount Presanella

Reply: It will be done on the amended version of the manuscript.

P3/L88: I cannot see that the coordinates of the three sampling sites have been provided in the manuscript.

Reply: We will add coordinates in the Study area section.

P4/L95/figure caption: “Numbers within square bracket indicate sample size” clarify if this refers to the number or cores or the number of trees?

Reply: They refer to the total number of sampled trees. It will be clarified and modified according to other suggestions in the amended version.

P4/L99: “The precipitation distribution reaches a minimum in winter and a maximum in summer, at 140.8 mm and 288.1 mm, … “ clarify which month/months the precipitation totals are provided for, and also over which period.

Reply: It will be clarified in the amended version as: “In the 1961–1990 period the gridded precipitation dataset return a distribution with a minimum in winter (December– February) and a maximum in summer (June – August), showing cumulative values of 172.2 mm and 292.4 mm, respectively, whereas the mean annual value is 1017.1 mm (Crespi et al., 2018; Carturan et al., 2012; Brunetti et al., 2006)”. Difference in values compared to the original version will occur after the upgrade and reanalysis of the precipitation dataset.

P5/L108; “… to highlight the ring boundaries. They were then scanned at 3200 dpi …” Split sentence for better readability.

Reply: It will be done in the amended version.

P5/L117: Mention briefly why these Coorecorder settings were adopted when measuring BI? Were they the ones that gave the highest Rbar values, or were they simply arbitrarily selected? 

Reply: The selected values were a compromise between both the sample and average tree-ring widths and the measurements necessities. A detailed study about the influence of the frame dimension on the Rbar will be considered for the future. The paragraph will be modified as follows: “To measure the minimum value of latewood BI (LWBI) and the maximum value of earlywood BI (EWBI), we applied a frame width of 100 pixels (equal to 0.8 mm at 3200 dpi Vs. a maximum sample diameter of 5.12 mm), providing a good guarantee to select the least disturbed portion and avoid any disturbance of the core. Depths of 50 and 200 pixels were used to measure LWBI and EWBI. The adopted frame size values were considered a good compromise between the average size of the woody structure and the measurement requirements. "

P5/L122: Why 16 DBI datasats? What is the difference between these datasets? Please clarify. Also, it is a common praxis to wash the samples in alcohol or ethanol to remove the discoloration caused by resin and other impurities. A better explanation of why this step was omitted is required.

Reply: The 16 DBI chronologies derived from the differences between the four LWBI and the four EWBI series (considering the 25, 50, 75 and 100 % of the pixels within the measuring frame applied to the cores as described at P5/L116-118) from which the chronologies were built, as described in Materials and Methods (from each LWBI series, all EWBI series were subtracted, obtaining in this manner 16 DBI series for each core). However, we will consider removing these methodological results, even if in our opinion they could be useful to establish a standard procedure also for this species. Regarding the resin and impurities removal, as stated by the referee in the review opening, the BI is still in an explorative state, and thus in literature sample processing is approached with different modalities. To our knowledge there is no tested standard procedure for Larix decidua Mill., but we followed Wilson et al. 2014, 2019, 2021) who processed samples for BI analysis without washing them. The same approach was followed by other authors where samples were only polished (e.g., Dannenberg and Wise, 2016; Dolgova, 2016; Arbellay et al., 2018; Heeter et al., 2020). On the other hand other papers refer to samples treated with solvents (e.g., Campbell et al., 2007; Björklund et al., 2014, 2015, 2020; Babst et al., 2016; Brookhouse and Graham, 2016; Buckley et al., 2018; Buras et al., 2018; Akhmetzyanov et al., 2020; Cao et al., 2020; Blake et al., 2020; Davi et al., 2021). We discussed this point in the materials and methods section, but we can certainly better underline this point in a new version of the manuscript. In our opinion, one of the strengths and greatest originality of the paper we have presented lies in being the second on BI of Larix decidua and the first approaching the use of BI data from this species as climate proxy.

P5/L126: “BI sample series belonging to the same individual were averaged to create the individual BI series.”

Reply: we will substitute “sample series” with “core series” for clarity. The sentence will be as follows: “BI cores series belonging to the same tree were averaged to create the individual BI series.”.

P5/L127: “Some individual BI series showed an age trend; thus, they were standardized using a modified negative exponential curve. If the modified negative exponential curve failed to fit the trend of the individual series, they were standardized using a negative or a horizontal line. “This sentence gives the impression that only some of the series were treated for age trends. If so, a better explanation is needed to why the standardization was not adopted universally, when wood density is known to have an age trend and is therefore commonly standardized before climate reconstruction.

Reply: All the series were treated for age trends. The series that exhibited an exponential negative trend were standardised using an exponential negative curve. Following the procedure introduced in ARSTAN software (Cook 1985, Cook and Holmes 1986; Cook and Holmes 1999), if the negative exponential curve does not fit with the trend shown by the considered tree-ring series, the alternative standardisation method to attenuate the age trend is using a negative slope or a horizontal line. Differently, if the question refers to why some trees show a negative exponential trend in BI values that can be fit by a negative exponential equation whereas other do not, probably it depend i) by the environmental history of the tree (even if trees that resulted macroscopically damaged or live in a dynamically geomorphological context were excluded from the sampling) or ii) by the reaching (or not) of the pith during the sampling (some individuals show hearthwood damaged by fungus or bacterial activity and thus only the most external portion was available/suitable for the analysis), but these are normal issues encountered and accounted also in TRW series.

Cook, E. R., 1985. A Time Series Approach to Tree-Ring Standardization. Ph.D. dissertation, University of Arizona, Tucson. 

Cook, E. R., and Holmes, R. L., 1986: User’s manual for program ARSTAN. In Holmes, R. L., and Adams, R. K. (eds.), Tree-Ring Chronologies of Western North America: California, Eastern Oregon, and Northern Great Basin. Tucson: Laboratory of Tree- Ring Research, University of Arizona, 50–56.

Cook E.R., Holmes R.L., 1999. Users Manual for Program ARSTAN., Tucson, Arizona.

P5/L131: “The mid-low-frequency domain was obtained using the same filter as a low-pass filter.” Unclear which filter the authors refer to.

Reply: The sentence will be rephrased as follows: “The mid-low-frequency domain was obtained by means of a low-pass Gaussian filter with a window length of 30 years and a sigma of 5 years.”.

P5/L134: “… were considered for the site chronologies.” (suggestion)

Reply: Thank you for the suggestion, we will follow it .

P6/L136: “representativeness of each chronology compared to an infinite hypothetical population“ (suggestion)

Reply: Thank you for the suggestion, we will follow it .

P6 “3.2 the paragraph grouping in this section needs some correction (e.g., see line L158/159).

Reply: We thank the referee to have highlighted this typo. The divided paragraphs were merged into one coherent paragraph describing the gridded dataset creation. 

L151: “… and interpolating the longest and homogenized meteorological series available for the Alpine region “ this section would really benefit from more info. Which are the met series the authors are referring to? Where are they located, how far from the study sites and at which elevation? How representative are these data for the studied sites? What about the accuracy in the early part of the record – especially in regard to precipitation? In the introduction the authors mention that the reliability decreases prior to 1875. It is then appropriate to use these records as calibration targets?

Reply: All the requested details are reported in the cited papers (i.e., Brunetti 2006, 2012, Brunetti 2014 and Crespi 2018). The methodology permits to reconstruct the climatologies and the anomalies with a high degree of confidence. It considers a minimum of 5 meteorological stations (only considering those within 150 km from the sampling stands) in 1800 to grow to ~50 in 1900 and ~200 in 2000. A diagram of the number and position of the used  meteorological stations will be added to the supplementary material as necessary.

P6/L166: “high-resolution spatially continuous interpolated gridded dataset” information about the spatial and temporal resolution should be provided, as well as its temporal coverage. Also, which parameters were used from the CRU dataset?

Reply: All the requested details are reported in the cited papers (i.e., Brunetti 2006; 2012; Brunetti 2014 and Crespi 2018) but, following the suggestion, we will insert some information from them here. Details about the used CRU-TS datasets will be added to the amended version of the manuscript. A proposed sentence structure for this last point will be: “In addition to the site-specific series reconstructed as described above, the Climatic Research Unit (CRU) time series (TS) version 4.07 of a high-resolution spatially continuous interpolated gridded mean temperature and precipitation datasets (Harris et al., 2020) was accessed through the CEDA archive (https://data.ceda.ac.uk/badc/cru/data/cru_ts/cru_ts_4.07, last accessed 10 October 2023) to assess the spatial coherence of the dendroclimatic signal.”.

P7/L87: Please provide more details around the 24 BI chronologies that were obtained. It is not clear how and why 16 deltaBI chronologies were constructed. Also, as I understand three sites were sampled (fig. 1), but four EWBI and LWBI chronologies obtained? How is this possible?

Reply: The four EWBI and the four LWBI derive from the different percentage of EW and LW pixels considered in the analysis for each site. Since for this species there is a lack of standard procedures for BI, various percentages of pixels within the frame (i.e., 25, 50, 75, and 100 %) were considered both for EW and LW as stated at P5/L116-118. The 16 DBI chronologies derived from the differences between the four LWBI and the four EWBI series; from each LWBI series, all EWBI series were subtracted, obtaining in this manner 16 DBI series for each core from which the chronologies were built, as described in Materials and Methods. However, we recognize that in the present form, this concept could be not clear, since also referee#1 highlight the same misunderstanding, thus we will work to make it clearer or consider removing these methodological results, even if we think that they are useful to try to start to delineate standard procedure also for this species but can be considered too specific details for the present work.

P//L187: “… the remaining chronologies …” not clear which remaining chronologies the authors refer to. Please clarify or rephrase the sentence.

Reply: we will remove the unnecessary and confusing “remaining”.

P7/L189: “…whereas the mean interseries correlation increased by 0.05 at most” unclear, needs clarification.

Reply: Thank you for the suggestion, the sentence will be rephrased as follows in the amended version of the manuscript: “Considering the EPS, the chronologies spanned between 1502 and 2015 CE for a total of 514 years, whereas the mean interseries correlation show a generally slightly increase.”.

P7/table 1: 

unclear what “valley” in the table head is referring to.

Reply:Thank you for highlighting unclear column name, we will edit the table for better explaining , modifying “Valley” in “Chronology code”. 

[EPS* 𝑟𝑟̅] and [EPS* 𝜌𝜌̅] may be misinterpreted as the EPS statistics multiplied by the interseries correlation coefficients (also, in the table caption it is referred to as *EPS and not EPS*).

Reply: Thank you for the suggestion, it will be modified as suggested (i.e., *EPS) to make it clearer and consistent.

Table captions are placed above and not below the table.

Reply:Thank you for noting it, it will be moved as requested.

“correlation with other chronologies” which parameter and which period? Also, are the correlations performed on high-pass filtered data or using raw chronologies?

Reply: Thank you for the suggestion, we will improve the caption following your requests. 

Why provide information on the number of trees that have not been used in this study?

Reply: We agree, unnecessary numbers will be removed.

P8/L202: “Moreover, the maximum sample depth of each chronology was comparable in quantity but not in time, with the duration reduced by approximately 100 years between ANBO and PALP and another 100 years between PALP and BARC” ?? Meaning unclear.

Reply: Thank you for the suggestion, the sentence will be removed; it is an unnecessary description of the chronology that can be easily inferred by Table 1 and the Figure 2.

P9/L214: “PC1 (ANBO+BARC+PALP)” unclear if we are still talking about deltaBI here?

Reply: No, we are not referring to the DBI sensu strictu, but we are referring to the first principal component resulting from the PCA applied to the three selected DBI chronologies (i.e., 25 % of the darkest pixel of LWBI and 100 % of the pixel of EWBI) of the study area. This part was dealt at P8/L203-204 we indicate: “Due to the coherence shown by DBI chronologies, a PCA was performed considering all the chronologies to better preserve the common variability, likely related to climate (Seftigen et al., 2020).” and at lines 206-209: “The PCA confirmed that the considered chronologies shared a large portion of the original variance. In fact, the first principal component (PC1) explained more than 80 % of the variance alone, and all the chronologies were strongly positively correlated with it. Thus, PC1 (ANBO+BARC+PALP) was selected to represent the areal chronology (Fig. S2 in the Supplementary Material for further details)”. We will modify the two paragraphs to better explain the concept.

P9/L220: it should be explained how the correlation confidence interval was obtained.

Reply: Thank you for your suggestion. Details about the method used to calculate the confidence interval will be added in the amended version of the manuscript. A proposed sentence will be: “Correlation indices were calculated in the R-project statistical environment by means of the treeclim (Zang and Biondi, 2015) package. The bootstrapping procedure described in DENDROCLIM2002 (Biondi and Waikul, 2004) was applied to calculate the correlation indices and their 95 % confidence interval by means of the percentile confidence interval method (Zang and Biondi, 2015; Dixon, 2001).”.

Dixon P.M., 2001. Bootstrap Resampling. In: Encyclopedia of Environmetrics., Wiley. https://doi.org/10.1002/9780470057339.vab028

P9/Sect. 4.1 What about the other BI parameters (EWBI and LWBI) and their climate signal imprint? Also, how is the BI signal compared to the TRW signal? If this is an exploratory study of BI from Larix (as stated in the introduction) then why do the authors limit the climate response analysis to just deltaBI?

Reply: As stated in the result section “The analysis of the BI chronologies and their correlation with meteorological datasets revealed that the influence of the considered percentage of the LWBI darkest pixels, as well as that of the EWBI lightest pixels, slightly affected the results. However, in general, DBI values obtained from a small percentage of the darkest LWBI pixels (i.e., 25 %) and from a large percentage of the EWBI pixels (i.e., 100 %) yielded better results considering the values of their correlation with meteorological data”. Moreover, as stated in the Material and Method section: “to correct the heartwood/sapwood discolouration that characterises European larch, 16 Delta BI (DBI) datasets were calculated and analysed as differences between the EWBI and LWBI datasets (Björklund et al., 2015, 2014).”. These are the main reasons why EWBI and LWBI were not considered separately even if they were tested during the study. We will work in the amended version of the manuscript to make this concept much clearer, considering also to add some information about this topic in the supplementary materials if necessary.

About the comparison with TRW signal we will better underline the results coming from previous study on the area (Coppola et al., 2012, 2013; Cerrato et al., 2018).

P9/L241: what is meant by “raw spatial correlation”?

Reply: we thank the referee#2 to have highlighted this typo. “raw” will be deleted in the amended version of the manuscript. The new version will be: “Since then, a strong increase in correlation was observed on the area (Fig. S3 in the Supplementary Material).”

P12/249 “what is meant by “static correlation”?

Reply: It is a term defined in Zang and Biondi (2015) to indicate the correlation performed on the entire considered period in contrast to ‘moving’ used on a moving window. We agree with the refree#2 that this term is not commonly used, thus we will remove it in the amended version of the manuscript.

Zang C., Biondi F., 2015. treeclim: an R package for the numerical calibration of proxy-climate relationships. Ecography, 38 (4), 431–436. https://doi.org/10.1111/ecog.01335

P12/figure 4 caption: it should be indicated over which season the temperature has been aggregated.

Reply: Thank you for your suggestion. It will be added to the caption in the amended version of the manuscript.

P12/L266: “row data and low-frequency domain”?

Reply: We corrected the typo.

P15/L314: the authors argue that the deltaBI transformation of the data may be sufficient to mitigate the heartwood-sapwood discoloration issue. Is this just speculations, or has a comprehensive comparison between the deltaBI and LWBI actually been made to draw these conclusions?

Reply: It is not a speculation as it was tested in the first stage of the analysis, but not reported to focus the manuscript on the most significant results. However, the significant attenuation of the difference between heartwood and sapwood can be appreciated in the figure reported in the attached file. We can add it in the Supplementary material.

I recommend the authors to work on streamlining the discussion and chiseling out the main message. The section is in its present form very long and often repeats what has already been stated in the results (for example L370 – spatial correlation analysis).

Reply: We thank the referee for the comment, we will follow it shortening and focussing the discussion on the most important message.

----------------

Figure 1 (attached): Solid black thick lines identify mean raw BI series of latewood (LW), earlywood (EW) and DBI (delta). Solid grey thin lines identify the raw individual series. Red area identifies the portion of the series where the transition from heartwood to sapwood occurs. To be noted the significant attenuation of the sapwood influence on the DBI values compared to both EW and LW.

Dear Editor,

We thank you for your reply and for the opportunity to submit an amended version of the manuscript for a new round of revision. We have addressed the referees' comments point by point, and we will submit a revised version of the draft as soon as possible.

Best regards,

Riccardo Cerrato on behalf of all the authors