1.    Title

On the time scale of meteorological, soil moisture, and snow drought indices to assess streamflow drought over catchments with different hydrological regime: a case study using a hundred Chilean catchments

2.    Research question

From abstract and introduction

• i) what temporal scales of (the frequently used and easy to calculate) SPI and SPEI meteorological indices can be used as proxies for streamflow drought in catchments with different hydrological regimes?
• ii) considering that the soil acts as a natural reservoir to maintain streamflow during periods of reduced P, can a soil moisture drought index be used to assess streamflow droughts instead of the SPI and SPEI meteorological indices?
• iii) considering that snowmelt is an important moisture contribution to streamflow and surface water availability during spring and summer months in catchments with a pronounced snow influence, can a SWE (Snow Water Equivalent) index be used to assess streamflow droughts instead of the SPI and SPEI in those catchments?

3.    Goals

From introduction and research question

• To define temporal scales for SPI and SPEI for using them as proxies for streamflow droughts in different hydrological regimes.
• To evaluate the suitability of a soil moisture drought index for using it instead of SPI and/or SPEI meteorological indices to assess streamflow drought.
• To evaluate the suitability of SWE index to assess the streamflow droughts in snow-influenced catchments.

4.    Findings

From abstract

• i) there is no single meteorological, soil moisture, or snow drought index and temporal scale that could be used to characterise all streamflow droughts across Chile, and
• ii) the greater the snow influence in a catchment, the larger the temporal scale of the drought index to be used as proxy of streamflow drought.

5.    How do you rate this paper in absolute terms? Poor to fair, good, very good to excellent.

Good to very good.

6.    Recommendation.

My decision is between “accepted after minor revisions” to “accept without changes”.

7.    Confidential comments to the editor

The paper has been well written: the results are clear (however, some improvements are still possible), well presented, and explained. The research questions are pertinent because it is necessary to improve the understanding of drought propagation in the hydrological cycle.

The proposed minor changes in the document (in “comments to the author”) do not affect the paper quality. Also, I made some specific questions that must be easily answered by the authors.

That is why my decision between “accepting” and “accepting after minor changes”.

8.    Comments to the author

I had a hard time following the explanation of the results of the Spearman correlation analysis. What variables were compared in this analysis? I believe it is easier to present these results in a table, preserving the text explanation. The text alone makes it tremendously difficult to understand what is meant.

9.    Questions

How do you feel about your assumption that the SSI-1 is the most representative of streamflow deficit conditions for the several catchments? Is it possible that catchments with higher storage (e.g., snow-influenced basins) should characterize their hydrological droughts with SSS for longer durations? Why yes, or why not, do you believe that SSI-1 is suitable for all your catchments?

You say that SSI-1 integrates catchment-scale hydrological processes. I think that it would be feasible for Chilean catchments ¿Is it (the SSI-1) suitable for larger catchments than those catchments in Chile? ¿Do you think that SSI-1 integrates the catchment-scale processes in larger basins?

Is the SWE suitable for all snow basins in Chile??? If the answer is no, ¿is that the reason why the spatial scales 1, 3, and 6 months were selected?

Is it possible to use other linear or non-linear analyses to establish the relationship between the indices and SSI-1? Correlation analysis gives a first look at the relationships, but it is a linear version of the analyses. It may be that non-linear relationships that are possibly hidden in the treated data cannot be seen. Here a question arises: is the Event Coincidence Analysis (ECA) strong enough as an alternative method of assessing the relationship between droughts and their triggers? ¿Why?

In the case of the ECA, ¿were the same thresholds selected for the several indices? That is, if moderate hydrological drought events were selected in the SSI-1 (SSI-1<-1.0), it was assumed that they were triggered by moderate meteorological/soil/snow events (e.g. SPI<-1.0, SPEI<-1.0, ...). ¿Do you consider merging the severities of hydrological and meteorological/soil/snow droughts? (e.g. that some severe droughts SSI-1<-1.5 are triggered by moderate meteorological droughts: -SPI<-1.0, or SPEI<-1.0-).

Pag. 13, you say:

While the analysis was done for diverse lags, the cross-correlation values and precursor coincidence rates, decreased gradually from zero lag (lag = zero months) to a lag of 12 months.

¿Is true that sentence for all the analyzed catchments? If the answer is "yes", ¿what do you think is the process that drive the flow in the catchments? The response is especially interesting in the nival and nivo-pluvial catchments, where the runoff should be delayed some weeks with respect to snow precipitation.

This study offers a holistic methodology for comprehending streamflow droughts across diverse hydrological contexts through the utilization of numerous hydrological and meteorological indices across Chile. In general, the manuscript is well-written.  Nonetheless, substantial revision is necessary to rectify numerous methodological flaws and improve the manuscript's readability. My remarks are as follows:

Major comments:

- Concerning the application of a one-month timescale to represent hydrological drought, I have several reservations. One potential drawback is that the temporal scale of one month may be exceptionally susceptible to short-term fluctuations in streamflow, which may result in an overestimation or underestimation of drought conditions. Furthermore, the index may be impacted by short-term anomalies caused by particular weather phenomena, which may introduce extraneous data and potentially obscure longer-term patterns of hydrological drought. In addition, the responses of hydrological processes to precipitation events, such as groundwater recharge and base flow contribution to streams, may be delayed. These delayed effects might not be captured on a one-month scale, resulting in an inadequate representation of streamflow conditions. Critically, a one-month scale may not sufficiently capture the seasonal dynamics of streamflow in catchments with substantial snowmelt contributions, particularly if the SSI is computed for months in which snowmelt does not predominate as a hydrological process.

- Also, catchments characterized by intricate hydrogeology may exhibit non-linear correlations among precipitation, soil moisture, and streamflow. As such,  a one-month SSI may not provide an accurate representation of real-world drought conditions.

- The sensitivity of the coincidence analysis between the SSI-1 and meteorological, soil moisture, and snow drought indices to the thresholds used to define each index remains uncertain.

- The selection of time scales and indices is significantly influenced by the accessibility and reliability of data, especially for soil moisture and snowpack, and streamflow data. Data may be scarce or unreliable in some catchments, thereby compromising the precision of drought assessments.

- In the same context, there is substantial variation in the extent to which snowmelt contributes to streamflow across catchments and years. In order to evaluate the effects of streamflow drought on a snow water equivalent (SWE) index, comprehensive calibration and validation procedures are necessary.

- The intricate correlation between streamflow and soil moisture is susceptible to the impact of vegetation and soil properties. This intricacy may not be comprehensively encapsulated by a single drought index.

- I suppose the selected catchments are irregular in terms of total area and dominant hydrological and meteorological conditions. This may imply that hydrological regimes within a given catchment may vary considerably, which makes it difficult for single index or specific time scale to represent an entire catchment.

- The justification for the selection of these particular 100 catchments should be clarified better.  It is necessary to include a table or figure that compares the various hydroclimatological and land use characteristics of the chosen catchments.

- The results are not adequately discussed in relation to the existing literature concerning Chile and beyond. It is my belief that introducing the "Discussion" and "Results" sections independently is crucial.

Specific comments:

- The title is considerably lengthy. It should be revisited to be more concise and informative.

- In the abstract, it is suggested to emphasize the importance of the research outcomes and the possible ramifications they may have on the domains of hydrology or drought monitoring. It can also be enhanced by presenting quantitative results.

- Prevent "lump" referencing; that is, do not refer excessively to illustrate a point or provide a description in the paper (e.g., L22-24).

- The in-text citation format must be meticulously revised in accordance with the journal's specifications.

- L37: Please include agricultural and socioeconomic droughts.

- L60: "climate factors" <> "climate conditions".

- In the text, SWE and SWEI are utilized interchangeably.

- The explanation is absent for Equation 3.

- Equation's numbering needs to be revised.

Overall, the presented research is a solid and valuable addition to drought science, posing relevant questions for drought management and monitoring. The introduction is well-structured, offering a comprehensive review of drought indices. The methods and results are clearly articulated. I have only a few general and minor comments to offer for further enhancement.

General comments:

• While the science questions are well-framed and articulated, the takeaways in the abstract seem somewhat general. The authors could enhance these by distilling the results into more compelling and informative points. Focusing on 'what/why' questions rather than providing simple 'yes/no' answers.
• The method section notes that the soil moisture and snow drought index are calculated using a non-parametric approach, in contrast to the parametric method used for other drought indices. I suggest the authors discuss the implications of using different methods. It would be insightful to conduct analysis demonstrating the uncertainties in drought assessment that arise from using various approaches and distribution functions.
• In the Results section, before the correlation analysis, adding a section that gives an overview of the drought characteristics in the region would provide useful context. This overview should include aspects such as the seasonality of drought, as well as the spatial distribution, frequency, and severity of droughts based on different indices.
• This study concentrates on the CAMELs headwater catchments. I recommend the authors provide discussions from regional perspectives, e.g., upscaling the catchment-scale insights to a wider regional context.

Minor comments:

1. Please define some technical terms that may not be commonly understood by all readers. For example, clarifying the meaning of 'nival hydrological regime'.
2. Line 285: given that the correlation between SWEI and SSI are generally low (75th percentile < 0.4), the following statement lacks solid grounding: “snow accumulated during the cold season influences the Q observed in the next hot season.” And later in Line 395: “The SWEI showed the highest values over nival catchments, followed by nivo-pluvial and pluvio-nival catchments, respectively”. While true in a comparative sense, the performance of SWEI in nival catchments is still lower than that of SPI. In general, with the results presented, it is not a strong argument that snow drought is critical for Q drought even for catchments with seasonal snowpack. Given the presented results, it is not a convincing argument to suggest that snow drought is a critical factor for discharge drought, even in catchments with seasonal snowpack. A more detailed examination of historical snow droughts in this region could offer a better perspective. Investigating whether severe snow droughts are a rare occurrence in this region and, consequently, whether they are not a reliable indicator for discharge droughts, may be able to provide a more robust conclusion.