Overview:

This manuscript presents new d2Hwax data from soils, river and marine sediments along the Chilean coast. The results show that there is a constant apparent fractionation in humid regions, whereas in arid regions evapotranspiration contributes to the d2Hwax signal. The d2Hwax of C29/C31 is shown to also be related to the aridity gradient, and potentially can reflect vegetation type changes. d2Hwax of marine sediments reflect the terrestrial d2Hwax input.

Review:

This manuscript is very interesting, provides novel data and important global insights and is structured and written very well. I congratulate the authors for a well-presented paper. The MS presents novel and systematic data, combined with a wide array of global databases (climate, vegetation etc.) and an updated d2Hwax database. The manuscript is particularly interesting in its assessment of the evaporation effect on d2Hwax in arid regions, and provides a global perspective on this process. The modeling and model parametrization are explained very well and lay out the method for utilizing this method in other places. 

I recommend publishing the paper pending some minor and textual comments.

Minor:

L116. Please explain how you calculate the uncertainty of the d2H values (e.g., average of duplicates? long-term error? error of the A6? etc..).

L297. This correlation is for soils and lakes combined?

L350. Why not present the annual average from each site compared to the OIPC data? Or, maybe just average the growing season months? In addition, it would be useful if you could provide the average residuals and standard deviation of the residuals (what is the difference between measured and OIPC data in permil).

L498 – 519. The statistical test shows that the marine, river and soils d2H overlap and are not statistically different from one another. However, Fig. 7 shows that marine sediments are, on average, heavier from rivers and soils and don’t really overlap at the 1 sigma level. Is this of importance? The Peru current flows northward, so ocean mixing would cause the opposite effect. Maybe higher contribution from coastal sediments (that should be heavier than the rivers based on Fig. 1c)?

Textual comments:

 L33. Add the abbreviation d2H_wax (instead of line 37)

L65. Notation d13C X2

L112. Notation H3+

L181. The wording here is not so clear (what is the purpose of this test? Testing the similarity of two populations?). Can you please rephrase.

L198. Reference format

L199. missing “back to isotopic ratios”

L273. Table 1 - IGSN not defined

Table 2. df, is this the same as the number of samples used for the regression? If so, I think number of samples is a more straightforward definition of this.

L475. Should be ‘explained’ not ‘exposed’

L547. Maybe ‘also’ instead of ‘more strongly’

The manuscript contains a dataset of δ²H_wax values from different terrestrial, riverine and marine sediment samples across environments with different aridity indexes in Chile. These δ²H_wax values are accompanied by δ²H values of precipitation, and a large set of different environmental characteristics, with the aim of identifying how well sediment derived δ²H_wax values track δ²H values of precipitation, and with that provide new insights into the validity of using δ²H_wax values for paleoclimatic reconstructions. The results show that on a global scale, the obtained δ²H_wax values follow δ²H values of precipitation quite well. However, within the dataset itself, across the aridity gradient, other environmental drivers also appear to become important in shaping δ²H_wax values. The latter even seems to differ for δ²H values of the two studied leaf wax n-alkane C-chain lengths, possibly related to changes in vegetation types. Lastly, the δ²H_wax values found in marine sediment samples reflect the δ²H_wax values from the terrestrial and riverine sediments.

I enjoyed reading the manuscript, which is written well with clear explanations of objectives and implications. The data presented here provides interesting new insight into how well δ²H_wax values track δ²H values of precipitation on a global scale, but also considers in more detail deviation of δ²H_wax from the expected δ²H_pre pattern by different drivers, like changes in evapotranspiration and vegetation type, along the aridity gradient. I think this is a valuable new approach to gain more insights into the drivers of δ²H_wax values. Although I am not an expert in the modelling approach and therefore cannot judge its accuracy very well, the explanation of the model was clear enough that I could follow what was being done. I only have a few minor comments that may help further strengthen the manuscript, and I think this paper is suitable for publication after these minor points have been addressed.

Minor comments:

1. Although changes in δ²H_wax values as an effect of differences in vegetation type are discussed in section 4.2, it is not addressed in the manuscript introduction. In L47-L59 I believe it might be valuable to already introduce the possible effects of species variation on δ²H_wax along the aridity gradient where changes in plant community composition may occur. Additionally, for the discussion section 4.2, it could be considered that even within taxonomically/physiologically constrained groups like herbaceous plants or eudicots, species differences in ε_wax/pre can still be very large (Chikaraishi et al., 2004, Phytochem.; Gao et al., 2014, PLoS ONE; He et al., 2020, Geochim. Cosmochim. Acta; Baan et al., 2023, Geochim. Cosmochim. Acta). This complicates the interpretation of the effect of the broad term ‘vegetation type’ changes on δ²H_wax values and requires more detailed knowledge on the integration of n-alkanes and their δ²H values from different plant species into the sediment (as you state in L540).
2. Even though in Fig. 2A & B all of the datapoints from the Chilean dataset would be considered to fall within error among the datapoints in the global dataset (i.e. roughly falls on the expected line in a global δ²H_pre gradient), it seems that once the Chilean dataset is isolated, there is no longer a strong relationship between δ²H_wax and δ²H_pre Can the authors comment on the relation between δ²H_wax and δ²H_pre within the Chilean dataset? From Fig. 2A & B it looks like this relationship is not very strong, and if this is the case, could this be an effect of uncertainty in δ²H_pre values, or an effect of additional environmental/biological control on δ²H_wax values? As a result of this, what magnitude of error could be introduced when reconstructing δ²H_pre from δ²H_wax for a given site that may be subject to additional environmentally/biologically induced variation in δ²H_wax values over time? I suppose the latter is difficult to answer quantitatively, but perhaps the authors can comment on this. Overall, I find the different comparisons made in Fig. 2 very interesting, but it might be valuable to clear up the interpretation and implications of the results on different spatial scales.
3. L527-L533: The results presented suggest that changes in the hydrological and vegetation characteristics of a given study site over time (i.e. irrespective of its current aridity state) can introduce some error in the reconstructed δ²H_pre from sedimentary δ²H_wax values, which is somewhat in contrast to the statement made in L527. As such, the continuation of this paragraph seems to be slightly opposing the initial statement, as hydrological changes may not be reflected in δ²H_pre Perhaps this paragraph could be slightly revised to provide a better overview of the nuances required for paleoclimate reconstructions from δ²H_wax values.

Phrasing/textual comments:

L30: Italicize ‘n’ in ‘n-alkanes’. This is not consistently done throughout the manuscript.

L52: although ‘less negative’ is not incorrect, I find that this can be a somewhat confusing term. More straightforward referencing between different δ and ε values could simply be ‘higher’ or ‘lower’ than (in this specific case ‘higher’). Also goes for further on in the manuscript (e.g. L286 and L287).

L65: superscript of ‘13’ and subscript of ‘wax’ should be fixed.

L93: Was the internal standard also used as a recovery standard to account for losses during sample processing? This is not mentioned later in the paragraph regarding n-alkane quantification (L107).

L198: Reference format: should not be in separate brackets?

L445: ‘They’ refers to herbaceous plants? ‘Deeper rooting depths’ relative to what (other vegetation types or with aridity, I presume the latter, but it is not entirely clear from this sentence)?

L493-494: ‘…, as they integrate over larger regions.’ seems a bit confusing at the end of the sentence since ‘… as they average both vegetation and climatic variability to a greater extent …’ is already mentioned before (i.e. last part of the sentence is redundant I think). Perhaps change to something like: ‘… as they integrate both vegetation and climatic variability over larger regions than soil samples.’

Table 4: This table seems somewhat redundant, as the p values are already shown in Fig. 7. The table itself could perhaps be moved to a supplemental info document or somehow processed into the text, if manuscript length would be in issue.