the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 10 Aug 2026
Large droughts are not large small droughts: hydro-meteorological drought generation processes vary with streamflow drought size in the Alpine region
Abstract. Streamflow droughts can be caused by a range of processes which can show strong heterogeneity from basin to basin. In contrast to local droughts, spatially extensive droughts seem to have more spatially homogeneous driving processes. However, the underlying mechanisms that cause this difference in the generation processes of small vs. large droughts are not well understood. This study presents an observation-based analysis that bridges this scale gap between large and small streamflow droughts. We assess how drought generation processes vary with drought size by classifying droughts into different types using streamflow drought observations and different hydro-meteorological variables. We find that: (1) most large drought events are characterized by one or two drought types even in hydro-climatically complex regions such as the Alps; (2) each drought type has its own spatial footprint and only specific drought types can lead to large droughts; and (3) in recent years, large droughts have increasingly been characterized by a mix of drought types. Our results highlight that the most common drought type in a catchment might not always lead to large droughts, which can be important for water management.
Competing interests: Manuela Brunner is an editor with HESS. The authors declare no further competing interests.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(6815 KB) - Metadata XML
-
Supplement
(6027 KB) - BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2026-3487', Anonymous Referee #1, 10 Sep 2026
-
RC2: 'Comment on egusphere-2026-3487', Anonymous Referee #2, 24 Oct 2026
In the manuscript, “Large droughts are not large small droughts: hydro-meteorological drought generation processes vary with streamflow drought size in the Alpine region”, the authors assess the spatial connection between small and large hydrological droughts in the Alpine region. They conclude that the combination of small droughts does not always generate a large one, and they also identify the specific cases in which this propagation takes place and which type of hydrological droughts are more prone to trigger this connection. I think the manuscript is well written and perfectly fits in HESS. However, I have some concerns that need to be addressed before the acceptance of the manuscript for publication.
General comments:
- I would suggest being clearer on the definition of what constitutes a large and a small drought from the beginning of the manuscript. I understand that large and small droughts are related to their spatial extent (you provide a mathematical definition in the methods section), but it would be nice for readers to have a clear idea from the beginning. In line with this comment, I think the title is a bit confusing because "large" is used to describe both extent (location perspective) and frequency (event perspective).
- The authors use ERA5-Land as the dataset for meteorological and snow-related variables. Regarding snow variables, it is well known that snow persistence is overestimated (Monteire et al., 2023). Some evaluations attribute this difference to a cold bias in mid-latitudes and high-altitude regions (Cao et al., 2020; Pelosi et al., 2020). Could this overestimation be conditioning the snow-dependent drought definition?
- Regarding the hydrological drought definition, the authors use a threshold limit computed as the 20th percentile of the streamflow time series. I have some concerns about it. Could you please better justify your choice based on the points below?
- Is the 20th percentile threshold a correct approach to capture the hydroclimatic variability of this climate?
- I understand you are using a unique threshold for the whole period; however, in part of the analysis, you split your sample into two periods. How does the 20th percentile change in these two periods? Is this change conditioning your results?
- Why are you using this methodology instead of the standardisation index approach?
- Why are the periods 1970-1993 and 1997-2017 selected? Is the year 1993 a break point in the time series?
- In relation to the drought definition based on the event perspective, the criteria used might be highly dependent on your results. Using daily time series and a concomitance threshold of 50% may not capture the natural propagation delay between snowmelt and streamflow in the catchment, which can bias results toward the precipitation deficit and mask the snow-streamflow nexus.
Specific comments:
- The study claims to assess droughts in the Alpine regions; however, the Italian Alps and part of the French Alps are not included in the analysis. I suggest stating clearly that the analysis is mainly conducted on the northern Alps and that the outcomes apply to these regions.
- In the last paragraph of section 2.2. Drought definition and classification, I would explain the classification scheme of Brunner et al. (2022) in more detail to improve understanding of the drought types and their relationships.
- Figure 1. I suggest clearly stating which panels refer to a single catchment (A and B) and which refer to all catchments analysed (C).
- I suggest moving all figures in the supplementary material to the main text. I don't think there is a figure limit in HESS.
References
Monteiro, D. & Morin, S. Multi-decadal analysis of past winter temperature, precipitation and snow cover data in the European Alps from reanalyses, climate models and observational datasets. Cryosphere 17, 3617–3660 (2023).
Pelosi, A., Terribile, F., D’Urso, G. & Chirico, G. B. Comparison of ERA5-Land and UERRA MESCAN-SURFEX reanalysis data with spatially interpolated weather observations for the regional assessment of reference evapotranspiration. Water 12, (2020).
Cao, B., Gruber, S., Zheng, D. & Li, X. The ERA5-Land soil temperature bias in permafrost regions. Cryosphere 14, 2581–2595 (2020).
Â
Citation: https://doi.org/10.5194/egusphere-2026-3487-RC2
Data sets
Large-sample hydro-meteorological dataset for the Alps Christopher Schlemper, Jonas Götte, and Manuela Brunner https://doi.org/10.4211/hs.f1c12fa6c5be4c61a5ec617fa62e13d6
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 221 | 142 | 44 | 407 | 104 | 43 | 52 |
- HTML: 221
- PDF: 142
- XML: 44
- Total: 407
- Supplement: 104
- BibTeX: 43
- EndNote: 52
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This manuscript presents a valuable, observation-based framework for understanding how hydro-meteorological drought generation processes differ between localized and spatially extensive streamflow droughts in the Alpine region. While the adaptation of a qualitative drought typology into an automated classification scheme is a strong conceptual step, the study's broader conclusions are weakened by some methodological limitations. Specifically, the requirement for absolute temporal synchrony to define large drought events mathematically penalizes natural river routing and upstream-downstream lag times, likely skewing the identified spatial footprints of certain drought types. Also, the reliance on a static historical baseline without addressing climate non-stationarity, and with the omission of intermediate land surface states—such as root zone soil moisture, which is critical for accurately representing plant-accessible water—limits the classification scheme's physical depth. I see addressing these constraints as necessary before these findings can be robustly integrated into a regional Drought Early warning system. The following are more specific comments:
- The manuscript defines streamflow droughts using a 20th-percentile threshold applied to a 30-day smoothed time series calculated over the 1970–2017 period. While the authors cite precedent for this approach, they need to explicitly explain why the 20th percentile is the appropriate physical baseline for this specific hydroclimate. Furthermore, applying a static baseline across 48 years of rapid Alpine warming assumes a stationary climate, which could significantly skew the decadal trend analysis presented in Figure 6.
- The automated classification categorizes droughts based strictly on preceding rainfall or snowmelt deficits. This completely bypasses terrestrial moisture buffering. Evaluating root zone soil moisture is critical in this context, as it represents the plant-accessible water stored in the root zone and serves as a highly accurate stock-based indicator for analyzing water availability. Omitting this intermediate storage metric forces complex, temperature-driven events into an overly broad rainfall deficit category.
-Â A large drought is defined as an event where at least 50% of the catchments are simultaneously under drought. Because the dataset explicitly over-represents low-elevation catchments, and rainfall deficit droughts are most common in these lower northern regions, the 50% metric mathematically favors rainfall deficits. The authors should quantify this spatial bias against the true topographic distribution of the Alps.
- The authors frame the study relevance around recent severe European droughts, specifically noting the 2018 and 2022 events in the introduction. However, the dataset ends in 2017, omitting the most significant compound droughts of the modern era and weakening the claims about recent trend shifts.
- Under the event perspective, the authors define a large drought using a daily time series to identify a time interval when at least 50% of catchments in the domain are under drought simultaneously. This strict requirement for catchments to be in drought at the same time mathematically favors spatially homogeneous meteorological anomalies, like rainfall deficits, which can affect large regions simultaneously. On the other hand, requiring absolute simultaneous occurrence penalizes drought types that propagate via upstream-downstream connections in the river network. As the authors note, regional and elevation-dependent temperature differences and sub-surface processes can delay the transmission of the drought signal to the river, causing relative time shifts between catchments and asynchrony in drought onset. Because of these time shifts, a snowmelt deficit drought might affect a large number of catchments over its total duration but fail to affect 50% of them simultaneously on a single day. Consequently, the conclusion that snowmelt deficit droughts generally remain small in extent may simply be a methodological artifact of the event perspective penalizing natural asynchrony, rather than an accurate reflection of their total spatial footprint.Â