the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 12 Nov 2025
Influence of Hydro-climatic Factors on Streamflow Patterns in Chilean Catchments
Abstract. Chile's extreme latitudinal gradient (18°–56° S) and diverse climatic regimes make it a natural laboratory for understanding hydro-climatic controls on streamflow evolution under climate change. This study quantifies maximum streamflow trends and their relationships with precipitation, temperature, and soil moisture across 38 catchments representing nine distinct Chilean climate zones over 2000–2021. We integrated field streamflow observations, gridded climate datasets (CR2MET), and satellite soil moisture data (SMAP/SMOS), applying Mann-Kendall trend analysis, Theil-Sen slope estimation, and Spearman correlation tests Results reveal significant declining streamflow trends in 67 % of winter-rainfall catchments (median slope: -1.12 mm/year, p<0.1), while summer-rainfall and tundra regions show the most substantial temperature increases (+0.06 °C/year). Mediterranean and temperate climates exhibit the strongest precipitation-streamflow correlations (ρ = 0.20–0.79), with soil moisture acting as an intermediate control mechanism. Our findings indicate that accumulated precipitation is the dominant driver in 53 % of analysed catchments, with soil moisture modulate the precipitation-runoff relationship, particularly in water-limited environments. These results provide critical insights for water resource management in Chile's diverse climatic regions and contribute to understanding hydro-climatic linkages in transitional climate zones globally.
- Preprint
(13195 KB) - Metadata XML
-
Supplement
(8245 KB) - BibTeX
- EndNote
Status: open (until 08 Mar 2026)
- RC1: 'Comment on egusphere-2025-5318', Anonymous Referee #1, 24 Feb 2026 reply
-
RC2: 'Comment on egusphere-2025-5318', Anonymous Referee #2, 25 Feb 2026
reply
The authors aim to 1) quantify annual maximum flows trends across 38 catchments in Chile (Q1); (2) identify dominant hydro-climatic controls on streamflow evolution by analysing relationships among streamflow, precipitation, temperature, and soil moisture variables (Q2); and (3) assess regional variations in climate-streamflow relationships to determine how different climate zones respond to hydro-climatic drivers (Q3). To this end, the authors use seven indices focused on annual maxima of precipitation, remotely sensed soil moisture, and streamflow, and compute the correlations between these indices for each basin separately over the period 2000-2021. While understanding the main drivers of high flows is relevant, I am afraid that the proposed methodology and analysis do not help answer the question, and the results do not support the conclusions. Moreover, the novelty of the study is unclear, and the objectives are only partially achieved. The overall comments are presented first, followed by the details.
General comments
- Basin selection and Köppen-Geiger classification
It is unclear how the authors selected the catchments and why those, rather than other basins, were chosen or discarded. Similarly, the criteria for assuming that a basin with multiple Köppen-Geiger classifications is representative of that group are not specified. Is a basin with a fraction of its area classified as a particular climate group, representative of that group? The authors offer no argument to support that idea. Additionally, several streamflow gauges appear to measure water flowing in canals rather than in riverbeds (e.g., Canal Vilama en Vilama), and human interventions (e.g., dams, diversions) are poorly presented in the document. How do the authors address the presence of dams in the Copiapó River or the Elqui River basins? The presence of dams would directly affect the results, particularly high and low flows, as the authors highlight throughout the document (e.g., L400, 503, 605).
- Runoff generation
The methodology proposed by the authors focuses on a series of annual maxima at the annual temporal scale for indices based on precipitation, soil moisture, and streamflow. However, more in-depth analysis is required before assuming that the (separately computed) annual maxima values are correlated. If the dates of the different annual maxima (e.g., cumulative 7-day precipitation and hourly streamflow) are not the same, why should the values be correlated? Additionally, some of the basins are strictly snow-driven, while others have a mixed hydrological regime. In such basins (e.g., snow-driven Río Elqui or Río Grande catchments), most precipitation falls in winter, while high flows occur in spring/summer due to snowmelt. How can the authors expect a correlation between the computed indices in such basins? The methodology does not address the seasonality of precipitation and streamflow and does not account for snow at all, which is very relevant for several of the basins the authors included in their analysis. Further, the combination of precipitation and antecedent soil moisture conditions can greatly impact peak flows during a storm. By treating these variables as “independent” (i.e., assuming that the maximum annual values of their suggested indices are comparable), the authors neglect relevant runoff-generation processes.
- Statements not supported by the results
There are many statements in the document that point to data or information that were never presented in the results (e.g., elevation, snow, presence of dams, etc.), for example (L514-515): “This pattern suggests that warming amplifies evapotranspiration and reduces snow contributions, ultimately limiting water availability.” Moreover, some of the statements are vague or too general (e.g., “Correlations confirm that precipitation […] is the primary driver of streamflow and soil moisture”, L570) and can be inferred without conducting the study. This weakens the discussion and conclusions, and jeopardizes the paper's novelty. Please see the details below.
- Overall document readability
In my opinion, the introduction of the document requires more work. It summarizes several previous studies, but it is unclear in terms of highlighting the novelty of the study or its relevance. Further, most of the figures correspond to maps, while the analysis is conducted at the catchment scale, without spatial heterogeneity. Thus, most of the figures could be combined for simplicity. Moreover, the authors refer to “North” (e.g., L400), “central-southern regions” (e.g., L41 or 336), or “Patagonia” (e.g., L635) without specifying what particular areas they are referring to. Additionally, the authors used several names to identify different climate groups throughout the manuscript (e.g., “Winter-rainfall regions (mediterranean, temperate, semi-arid)”, L499 vs. “ET(w)–BSk(w)”, L410). This makes the manuscript very difficult to read.
- Length of period analyzed
The authors used a 20-year period for their analysis. Simultaneously, the authors state that the second half of that period is influenced by a megadrought (L43-44: “Analysis of 106 Chilean catchments during the recent megadrought (2010-2020)”, or L86: “The last decade has been marked by an unprecedented megadrought”), which greatly influences the presence/absence of trends. Using only 20 years can be considered particularly short for trend analysis. Instead of focusing just on hourly streamflow, the authors could also explore daily streamflow, which covers longer periods of time, to make their result more robust.
- Research questions
The authors posed three research questions that partially address Q1 by computing temporal trends – using the Man-Kendall test and the Sein-Theil slope– over a 20-year period (2001-2021). Q2 (“identify dominant hydro-climatic controls on streamflow evolution …”), although relevant, is not addressed and is poorly supported by the results, since the authors do not analyze high flow events, but just focus on annual (independent) maxima indices. Such methodology impedes the analysis of runoff generation and of how precipitation, antecedent soil moisture conditions, isotherm 0°c, etc., interact to generate runoff. This point is heightened by the lack of 1) analysis of the diversity of processes driving high flows in rainfall-driven, snow-driven, or mixed hydrological regimes (all present in the study basins), and 2) spatial heterogeneity within each catchment. Regarding Q3 (“assess regional variations in climate-streamflow relationships…”), the authors poorly discuss why some of the basins within the same area show mismatching trends (e.g., Figs. 2b, 2h, 3h, 3i, 4h, 5b, 5e, 5i), weakening the overall conclusions. Moreover, the authors do not show the spatial distribution of Köppen-Geiger climate groups for Chile and do not indicate which fraction of these areas is covered by the selected basins. Including such analysis would put the representativeness of the selected catchments into context and, subsequently, the robustness of the conclusions.
Specific comments
- L21-29: The authors use the present tense to highlight the current effects of climate change on water resources. I suggest using more up-to-date references to support their statements. Here are some suggestions that could be useful: (i) https://doi.org/10.1139/er-2021-0109, (ii) https://doi.org/10.1016/j.scitotenv.2022.159854, and (iii) https://doi.org/10.1016/B978-0-323-99714-0.00008-X.
- If I understand correctly, the authors aim to address relationships between streamflow, climate, and soil moisture. Although answering these questions could be useful in a climate change assessment, these topics do not necessarily involve climate change (as the first paragraph states; L21-29). In my opinion, the manuscript’s introduction would greatly benefit from reducing its focus on climate change.
- L116-117: It is unclear to me what the criteria are to select (or discard) a Köppen-Geigger group. What do the authors mean by “A climate zone was included in a study area if it covered at least 10% of the zone's surface [What zones’ surface?] or 15% of a specific catchment” (How were the catchments previously selected to apply such criteria?).
- In my opinion, Figure 1 would greatly improve if the authors added the regions of Chile mentioned in the text, as well as a spatial distribution of the Köppen-Geiger classification. Also, the text in Fig. 1c, 1g, and 1k is difficult to read. Also, panels have different font sizes for numbering. If the coordinates for each panel are needed, I strongly suggest increasing the font size. Please also increase the font size of the distance scales.
- The selection of some catchments is peculiar. Why would the authors choose streamflow stations measuring streamflow in canals (e.g., Canal Lauca en Sifón n°1, Canal Vilama en Vilama, Canal Tilomonte antes represa, Canal Mal Paso después de bocatoma)? Also, if the authors aim to study streamflow trends – particularly if they use hourly streamflow records (L182) –, how would they isolate the effects of dams in these basins: Río Copiapó en la ciudad de Copiapó, río Copiapó en La Puerta, and río Elqui en La Serena (e.g., L324, L504)?
- What is the hydrological regime of the different basins? I think this would influence the climate attributes required to characterise high streamflow events.
- When gap-filling streamflow, how were the donor catchments selected? Is the pool of donor catchments the same as in Table 1? Also, how did the authors evaluate the performance of each “infilled record”(L253-254)? Did they conduct a cross-validation? How many of the annual maximums correspond to gap-filled values?
- In my opinion, without clarifying the hydrological regime of the basins (e.g., rainfall-driven, snowfall-driven, mixed, etc.), it is very difficult to determine whether the Spearman Correlation between the annual maximum variables makes sense or whether the climate descriptors are sufficient to characterise the catchments. For example, in L286, the authors state: “Conversely, in snow-dominated or arid regions, temperature and snowmelt dynamics often exert stronger control on streamflow variability (Viviroli et al., 2011; Zhai & Tao, 2017).” Which of the climate attributes suggested by the authors would represent the effect of snow (e.g., L331 or L345)?
- Also, in several basins (particularly in northern and central continental Chile), the wet season occurs in winter, while maximum streamflow occurs in spring and summer due to snowmelt. In such basins, why would the maximum rates of daily precipitation be correlated with the annual maximum hourly streamflow values (e.g., L365-366)?
- Before comparing the annual correlation between the annual maximum time series (L278-280), I wonder whether the variables correspond to the same events each year. For example, does Qmaxannual occur on the same day as Ppmaxannual or PP7maxannual? If that is not the case (for different basin-year combinations), what processes would explain such behaviour? This could be key to understanding and explaining trends in annual maximum hourly streamflow and to verifying whether the selected climate attributes are representative of the processes generating runoff.
- The patterns found in Figs. 2b, 2c, and 2h are interesting. Is there an explanation for the mismatches in the trends of annual maximum hourly streamflow in these three regions?
- L339: When the authors mention “Annual maximum precipitation”, are they referring to PPmax or PP7max?
- L371: How do the authors conclude that trends in Tmaxannual (Fig. 5) are also trends in “both regional warming and the elevation-driven upward shift of the 0 °C isotherm” (L371)? Or do they refer explicitly to the days associated with the annual maximum temperature? Also, how did the authors identify the summer events (L376)? Did they analyze more than just the annual maxima (e.g., seasonal maxima)?
- L396-397: I wonder how the authors lead to this conclusion when analysing the trend of annual maximum soil moisture: “These results suggest that in drier summer-rainfall regions, irregular rainfall events may counterbalance drying tendencies when soil moisture is measured at a three-hour resolution”. Did they quantify other events than just the annual maxima?
- Table 2 shows results for “Q trend”, “PP trend”, and “T trend”. However, such a acronyms haven’t been defined so far. Do they correspond to the annual maximum values?
- The authors use the Köppen-Geiger classification and a climate classification (e.g., L499, 501). In my opinion, the manuscript would benefit from using a single clustering approach for simplicity.
- Figure 12 displays acronyms that haven’t been previously defined (e.g., Q_T, PP_T).
- How do the authors reach this statement (L524-525): “The streamflow trends reveal a clear and consistent decrease across Mediterranean, temperate, and winter-rainfall climates, indicating a strong sensitivity to winter precipitation, as also observed in Europe (Blöschl et al., 2019).”?
- L534-535: The authors state: “These findings confirm the critical role of winter precipitation in sustaining streamflow, contrasting with the weaker or non-significant changes found in tundra and summer-rainfall climates.” What “streamflow” are the authors referring to? Floods? Low flows? In summer/winter?
- What results support this statement (L539-542)? “Global warming intensifies these effects, especially in drier catchments. While Vicuña et al. (2013) proposed that extreme heat might increase the contributing runoff area, the results here indicate that enhanced evapotranspiration instead reduces streamflow availability.” Also, I think the authors are misinterpreting Vicuña et al. (2013), since Vicuña et al. analyzed temperature on rainy days, while the authors don’t.
- What results support this statement: “In contrast, summer-rainfall climates display mostly stable or positive soil moisture trends, especially in semi-arid basins where convective summer storms sustain short-term replenishment despite high temperatures.”?
- In my opinion, this statement is too vague (L571-572): “Correlations confirm that precipitation – especially multi-day accumulation – is the primary driver of streamflow and soil moisture, reinforcing findings from Europe (Blöschl et al., 2019).” Such a statement can be assumed without conducting this study. Also, the authors haven't explored the temporal mismatch between the annual maximum hourly streamflow and the maximum cumulative precipitation.
- L591-593: In this case, I think correlation is not causation, and more in-depth analysis is required (e.g., event-to-event) to demonstrate such causation. Also, the small number of basins analyzed does not support such a generic statement.
- L630-635: What results support this statement?
- L635-636: None of the results support this conclusion.
Technical suggestions
- L24-27: The statement reads very similarly to the previous one in L21-22. Maybe rephrase to avoid repeating the idea?
- L30-31: I wonder if the authors could expand the list of citations for such a statement.
- L32: “… in the Atacama” region? Desert?
- L33-34: I think it would improve the manuscript if the authors could explain what they mean by “however, recent studies have yielded diverse and sometimes contrasting results.” What in particular are they referring to?
- L49: I think the paragraph works well without: “For many years, global studies primarily focused on streamflow, temperature, and precipitation.”
- L64: This sentence (“Other studies emphasise the dominance of climatic drivers over oceanic signals in shaping hydrological regimes.”) reads strangely. The authors have primarily mentioned climate-related drivers in the previous two paragraphs (L49-57 and L59-64).
- L84: I think the sentence would benefit if the authors clarified the spatial extent of the mega-drought.
- L86-89: I think this idea is repeated. It can also be found in L44-45.
- I wonder if the authors would consider reordering the order of the paragraphs. For example, paragraph 1 (L21-29) describes general climate change impacts; it then moves on to continental Chile (paragraphs 2 and 3, L30-47), then to worldwide impacts (L49-79), and finally returns to the effects on Chile.
- L108: “68°W”
- L120: What do the authors mean by “ET(w)”? I think such a term has not been previously defined. Similar comment for other acronyms (e.g., BSk(s), BSk(w), etc.).
- In Table 1, I believe there is a typo. Shouldn’t it be "Río Copiapó“ instead of “Río Copiaó”?
- I think the information provided in L216-217 is the same as the data provided in L196-198.
- L306: I wonder if “sensitive” is more accurate than “susceptible in this context. Also, in L307, the authors state: “with higher elevation modulating the magnitude of these declines.” Is this related to orographic precipitation enhancement, precipitation partition into rain and snow, or something else?
- When explaining the results in Fig. 2, the authors use the panel titles rather than the numbering (a, b, c, …). I think the text would benefit from using the numbering instead of the climatic zones. For example, in L312, the authors state “in … Cfb(i)-Cfb, …”, but this region corresponds to the last panel in Figure 2, so interpreting the figure and the text is difficult. This comment also applies to Fig. 3.
- How did the authors find this conclusion (L318): “likely reflecting oceanic influence and differences in catchment size.”?
- 2, 3, 4, 5, and 6 would benefit from mentioning the units of the Theil-Sen slope.
- L443-444: I think the sentence would benefit if the authors explained what they mean by “northern zones”.
- In the title of Table 2, I wonder whether “uncoloured” could be changed to “in black”.
- 9 would benefit from increasing the font size of the axis text. Also, are all correlation statistically significant? Not showing which correlations are statistically significant makes it difficult to draw conclusions.
- I wonder if Figures 9-11 can be combined into a single figure, with a message that’s easier to see. The way it is currently presented looks repetitive, and the message is unclear. This level of detail could be in the Supplements (note that very similar plots are also shown there).
- Code availability: I strongly suggest that the scripts be made freely available when the paper is published. For transparency, the study should be reproducible.
Citation: https://doi.org/10.5194/egusphere-2025-5318-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 283 | 185 | 23 | 491 | 69 | 220 | 115 |
- HTML: 283
- PDF: 185
- XML: 23
- Total: 491
- Supplement: 69
- BibTeX: 220
- EndNote: 115
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The manuscript analyses streamflow trends in 38 Chilean catchments over the 2001–2021 period and examines their correlations with precipitation, air temperature, and soil moisture. Mann-Kendall and Theil-Sen trend analyses are applied to in situ streamflow records, a local gridded precipitation dataset (CR2MET), and the GLDAS-Noah reanalysis soil moisture product (rather than the SMAP/SMOS satellite datasets mentioned in the abstract).
However, the manuscript is poorly structured and difficult to follow. Several statements are not supported by quantitative results derived from the analysis. Moreover, the study lacks a clearly articulated scientific question or testable hypothesis. Although three specific objectives are stated, two of them are not fully addressed in the Results and Discussion sections. The datasets and methodological approaches employed are standard and do not present methodological innovation. In addition, the maps shown in the Results section do not depict spatially continuous hydrological variables and would be more appropriately represented using coloured symbols. Some of the conclusions are speculative and not sufficiently supported by the evidence provided.
In its current form, the manuscript does not constitute a substantial contribution to the hydrological literature. The major and minor issues identified are detailed below.
Major comments:
MC1. The motivation of the manuscript is insufficiently developed. The Introduction reviews several studies from Chile and other regions documenting increases in air temperature, decreases in precipitation and streamflow, and the growing attention given to soil moisture as a potential explanatory variable of streamflow trends. However, it does not provide a clear conceptual framework describing the relationship between hydroclimatic drivers and streamflow patterns, as suggested by the title. Furthermore, the Introduction does not clearly state whether the study aims to identify spatial, temporal, or spatio-temporal streamflow patterns. It also lacks a strong justification for the three stated objectives: (i) quantifying trends in annual maximum streamflows; (ii) identifying dominant hydroclimatic controls on streamflow; and (iii) assessing regional variations in climate–streamflow relationships. The importance and novelty of addressing these objectives are not sufficiently articulated.
MC2. Methodology is not appropriate to address Specific Objective 2. Specific Objective 2 aims to “identify dominant hydro-climatic controls on streamflow evolution by analysing relationships among streamflow, precipitation, temperature, and soil moisture variables”. While the identification of dominant hydro-climatic controls on streamflow evolution is indeed a relevant and interesting research question, particularly at the event scale, the methodological approach adopted in the manuscript is not adequate to support such inference. The study evaluates trends in annual maximum streamflow, precipitation, air temperature, and soil moisture, and subsequently computes the Spearman rank correlation coefficient among these variables. However, correlation does not imply causation, and the literature provides numerous examples of spurious correlations. High values of the Spearman coefficient alone do not constitute evidence of “dominant hydro-climatic controls” on streamflow evolution. To robustly identify dominant controls, a more rigorous experimental or modelling framework would be required, explicitly designed to disentangle causal relationships and quantify relative contributions. Such an approach is not implemented in the present manuscript.
MC3. Methodology is not appropriate to fully address Specific Objective 3. Specific Objective 3 seeks to “assess regional variations in climate–streamflow relationships to determine how different climate zones respond to hydro-climatic drivers”. However, as discussed in MC2, the methodology is limited to analysing trends in annual maximum streamflow, precipitation, air temperature, and soil moisture, followed by the computation of Spearman rank correlation coefficients among these variables. This approach allows the identification of regional differences in trends and in correlation coefficients among annual maximum variables. Nevertheless, it does not provide a robust framework for assessing variations in “climate-streamflow relationships”. The manuscript does not clearly define how this relationship is conceptualised or evaluated, particularly with respect to the temporal scale adopted for the analysis. Moreover, the study does not explicitly define the “hydro-climatic drivers” assumed to control streamflow responses at the catchment scale. The analysis implicitly assumes that Köppen–Geiger climate classes, annual maximum precipitation, and air temperature are the primary climatic controls on annual maximum streamflow. However, it does not account for the potential modulating effects of catchment characteristics (e.g., soil properties, land cover, geology) or large-scale climatic teleconnections (e.g., ENSO, PDO, SAM). A more rigorous experimental design is required to clearly define and quantify the climate-streamflow relationship, specify the relevant temporal scale, and isolate the influence of different hydro-climatic drivers. Such an approach is not implemented in the current manuscript.
MC4. Definition and naming of climate zones. The Köppen–Geiger climate classification adopted in this study is based on an outdated, locally developed dataset (Sarricolea et al., 2017), rather than the state-of-the-art global classification provided by Beck et al. (2023). The use of a non-standard and older classification limits the comparability of the results with recent international studies published in leading hydrological journals. In addition, the manuscript does not assign descriptive names to the climate classes listed in the first column of Table 1. This omission significantly hinders the readability of the Results and Discussion sections. For example, line 524 refers to “Mediterranean, temperate, and winter-rainfall climates”, while line 545 mentions “winter-rainfall climates -particularly Mediterranean, temperate and semi-arid regions”. In neither case is it clear which of the nine climate classes defined in Table 1 is being referenced. To improve clarity and accessibility, I strongly recommend providing a concise textual description for each climate class (e.g., “cold semi-arid/tundra” instead of “Bsk(s)-ET(s)-Bsk(s)(i)”). This would greatly facilitate interpretation and ensure consistency throughout the manuscript.
Beck, Hylke E., Tim R. McVicar, Noemi Vergopolan, Alexis Berg, Nicholas J. Lutsko, Ambroise Dufour, Zhenzhong Zeng, Xin Jiang, Albert IJM Van Dijk, and Diego G. Miralles. High-resolution (1 km) Köppen-Geiger maps for 1901–2099 based on constrained CMIP6 projections. Scientific Data 10, 724 (2023).https://doi.org/10.1038/s41597-023-02549-6.
MC5. Data sources. The selection and treatment of the data sources are not adequately justified. First, the “Data sources” section (L175) states that “The study focuses on identifying trends and correlations in extreme hydrological events”. However, the analysis is limited to annual maxima of selected variables, without a formal definition of extreme events (e.g., criteria for event start and end, minimum inter-arrival time, threshold selection). Annual maxima alone do not constitute an event-based extreme analysis. Second, the rationale for selecting the 38 streamflow stations is not provided. Some catchments are subject to significant anthropogenic regulation (e.g., the Lautaro reservoir), which, as acknowledged by the authors (L324–325), masks natural streamflow variability. If natural variability is altered, it is unclear how the study can robustly identify “dominant hydro-climatic controls on streamflow” (SO2). The manuscript does not explain why stations with near-natural regimes and consistent high-frequency (e.g., hourly) records across Chile were not prioritised. Moreover, although L241–242 mention data gaps due to natural and operational factors, there is no information on the maximum allowable gap length, whether hourly data were consistently available throughout 2000-2021, or how temporal consistency was ensured. The use of logarithmic relationships (Equations 11 and 12) to fill gaps in annual maximum streamflows is not supported by references or validation analyses demonstrating their reliability. Third, the choice of the GLDAS-Noah v2.1 soil moisture product is not justified. No comparison is made with other widely used gridded datasets (e.g., ERA5, ERA5-Land, GLEAM, SMAP, ESA-CCI SM), nor is a literature review provided to support the suitability of GLDAS-Noah for Chile or South America. In addition, the manuscript does not specify which soil layer(s) from GLDAS-Noah were used. Fourth, the study period (approximately 20 years) is short for robust trend detection in annual maxima (the core of this manuscript). The potential limitations associated with such a short record length are not discussed. This is a critical issue for a study focused on trend analysis, and it requires either a detailed justification or the use of datasets with longer temporal coverage. Finally, L196–197 indicate that catchment-scale averages of gridded datasets were computed using area-weighted approaches, but the software and specific procedures used are not described, which limits the reproducibility and transparency of the analysis.
MC6. Variable selection and definition. The selection of variables such as PP7daily,i , SM3daily,i, SM7daily,i , etc., appears arbitrary and lacks clear physical or methodological justification. For example, it is not evident why the annual maximum of a 3-hour soil moisture estimate (SMmaxannual) should be expected to correlate meaningfully with the annual maximum streamflow. The manuscript does not provide a conceptual or process-based explanation supporting these choices. If the objectives are to “identify dominant hydro-climatic controls on streamflow evolution” (SO1) and to “assess regional variations in climate-streamflow relationships” (SO2), a more appropriate approach would be to analyse the specific precipitation events that trigger the annual maximum streamflow at each catchment. Such an event-based framework would allow a clearer linkage between forcing and response. Conversely, if annual maxima of precipitation, streamflow, and soil moisture are treated as independent variables, without explicitly accounting for causal relationships, then it would be preferable to employ widely recognised extreme precipitation indices (e.g., Rx1day, Rx5day, R95pTOT, PRCPTOT). Using established indices would enhance comparability with the international literature and facilitate positioning the findings within the broader context of extreme hydroclimatic research.
MC7. Unnecessary figures. The maps presented in the Results section (Figures 2–8) do not display spatially continuous hydrological variables. Instead, they represent discrete catchment-based results. As such, the current map format is unnecessarily repetitive and could be consolidated into one or two multi-panel figures covering the entire study area (Chile). Each catchment could be represented by a single symbol, coloured (e.g., brown for decreasing trends, blue for increasing trends) and left blank when no statistically significant trend is detected. This would improve clarity, reduce redundancy, and facilitate comparison across variables. Even in their current form, some potentially interesting spatial patterns are visible; for example, a single catchment showing a decreasing trend among predominantly non-significant trends (Figure 2b), or clusters of catchments with decreasing trends surrounded by others without significant trends (Figure 2h). However, these spatial configurations are not analysed or discussed in the manuscript. A more focused and interpretative discussion of such patterns would substantially strengthen the Results section.
MC8. Results and Discussion sections are difficult to follow. In addition to the issues raised in MC4 regarding the ambiguous use of climate zone terminology, two further aspects substantially limit the clarity of the Results and Discussion sections. First, the naming of variables is often ambiguous, preventing a clear understanding of what is being analysed. For example, L89 refers to “Annual maximum soil moisture trends (Fig. 6)” without specifying whether this corresponds to SMmaxannual, SM3maxannual, or SM7maxannual. Similar ambiguity appears in references to annual maximum precipitation (e.g., L339, Figure 3). Precise and consistent variable naming is essential for clarity, particularly when multiple related indices are introduced. Second, several interpretative statements are not clearly linked to supporting figures or tables. For instance, the statement “In summer-rainfall climates, however, the signal is weaker or even positive” (L435-436) does not specify which of the nine climate classes listed in Table 1 are being referenced, nor what the signal is being compared against. Likewise, “The hydro-climatic patterns identified in the Mediterranean and temperate zones of Chile...” (L454) lacks a clear correspondence with the climate classes defined earlier. Statements such as “Overall, the results confirm the central role of precipitation...” (L469) and “Streamflow exhibits strong positive correlations with accumulated precipitation” (L475-476) are not explicitly tied to a specific figure or table that demonstrates these findings. To improve readability and scientific rigour, all interpretative claims should explicitly reference the corresponding figures or tables, and terminology should be used consistently and unambiguously throughout the manuscript.
MC9. Unsupported claims in the Results and Discussion sections. The Results and Discussion sections contain several statements that are not supported by analyses presented in the manuscript. In multiple cases, interpretations extend beyond the scope of the data and methods described. Examples include:
“They are further modulated by elevation” (L344). The role of elevation was not analysed or quantified in the study.
“Results suggest that in drier summer-rainfall regions, irregular rainfall events may counterbalance drying tendencies when soil moisture is measured at a three-hour resolution” (L396-397). Rainfall events were neither formally defined nor analysed.
“In temperate rainy climates (Cfb(i)–Cfb), negative soil moisture trends were also present, albeit weaker than in arid winter-rainfall regions” (L402-403). The relative magnitude (intensity) of trends was not evaluated in a way that supports this comparison.
“With the most substantial decreases occurring in the north, indicating progressive soil drying driven by reduced rainfall and snowmelt” (L414-415). Snowmelt processes were not analysed, and no evidence was presented demonstrating that reduced rainfall or snowmelt drives the reported soil moisture trends.
“Soil moisture emerges as an intermediate variable, strongly associated with precipitation, and in Mediterranean and temperate climates, it modulates runoff responses” (L471-472). The manuscript does not analyse how soil moisture modulates runoff responses to precipitation.
“The results here indicate that enhanced evapotranspiration instead reduces streamflow availability” (L540-541). Evapotranspiration was not analysed.
“However, in purely tundra climates, soil moisture shows weak or even opposite signals because snow and frozen ground dominate the hydrological response…” (L551-552). Snow and frozen ground processes were not evaluated.
“The spatial variability of soil moisture trends reflects the interplay between precipitation seasonality …” (L554-555). Spatial variability of soil moisture trends was not explicitly analysed beyond catchment-level summaries.
“However, this is partially mitigated by consistency with longer records (Boisier et al., 2018)” (L610–611). No quantitative comparison with longer records is presented to substantiate this claim.
Overall, several interpretations introduce processes (elevation effects, event dynamics, snowmelt, evapotranspiration, frozen soils) that were not analysed within the methodological framework of the study. All interpretative statements should be directly supported by the presented analyses, or clearly framed as hypotheses or speculative interpretations requiring further investigation.
Minor comments
Minor comments will only be made in an eventual new review round of this manuscript.