the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 07 Jan 2026
Characterizing low and high flow spells and their temporal transitions using baseflow estimates
Abstract. Extreme hydrometeorological events such as floods and droughts cause severe socio-economic and environmental impacts. These impacts can be amplified if hazards occur successively before the system can recover. While the drivers of individual extremes are well understood, the spatial variability and timescales of transitions between high and low flow spells remain understudied, especially regarding their implications for operational management. We propose an analytical framework to detect and characterize these spells using daily streamflow data from 643 catchments in France (CAMELS-FR) over the 1970–2021 period. We use a mixed threshold approach combined with baseflow estimation as an indicator for catchment recovery to identify the spells and analyze their frequency, duration, and temporal transitions. The analysis is carried out at catchment scale and at the scale of French operational flood forecasting centers. We find that short duration high flow spells are predominant across France, while long duration high flow spells are concentrated in northern France. Regarding transitions, they are predominantly consecutive occurrences of the same spell type, with consecutive high flow spells being more common. Our analysis reveals that transitions occurring in less than a month from low to high flows show distinct spatial variability, with the shortest transition durations concentrated in the Rhone-Mediterranean and Rhine-Meuse basins. These short term transitions predominantly occur in autumn and early winter. On the other hand, transitions from high to low flows are typically slow, developing over more than 90 days. These findings highlight the importance of enhancing our knowledge on compound events to better adapt flood disaster and drought management to local contexts and their characteristics.
- Preprint
(10424 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-6333', Bailey Anderson, 04 Mar 2026
- AC1: 'Reply on RC1', Guilherme Mendoza Guimarães, 21 Mar 2026
-
RC2: 'Comment on egusphere-2025-6333', Anonymous Referee #2, 10 Mar 2026
Comments on manuscript “egusphere-2025-6333” by Guilherme M. Guimarães.
This manuscript uses an analytical framework to detect and characterize the spatial variability and timescales of transitions between high-flow and low-flow spells. The topic is timely and relevant for the analysis of extreme hydrometeorological events. The methods are original, the work is quantitative, and the overall quality of the manuscript is high. I believe the manuscript is a good fit for HESS, and I recommend it for publication after revisions. I have the following comments for the authors to consider.
Major comments
- The manuscript refers to extreme-event impact databases, but I do not fully understand the role in this study. In hydrology, flood and drought severity is commonly assessed using streamflow time series and deviations from normal conditions. While extreme events can certainly lead to economic losses and other societal impacts, these aspects do not seem to be directly addressed by this paper. I suggest clarifying the relevance of these databases to the present analysis. Otherwise, use less words about the impacts database.
- The authors compare three different baseflow separation methods. However, based on Section 3.1 and Figure 6, the LH and UKIH methods appear to produce quite different results for the description of high-flow and low-flow spells. This raises the question of how robust the results are across methods. Please discuss more explicitly how much confidence can be placed in the findings given these methodological differences ?
- The characteristics of the different spells and their spatial patterns are interesting, and the authors present them thoroughly. However, I wonder whether climate change has affected these spells over the past 30 years. For example, have the duration or transition characteristics of HFS/LFS increased or decreased over time, or are there other temporal trends? Even if a full trend analysis is beyond the scope of this paper, some discussion would be valuable.
- Please explain more explicitly how the two research questions are answered in the manuscript. At present, this connection is not clear ?
- Also I also have the impression that the manuscript is somewhat long, and the discussion of baseflow separation occupies substantial space. Since the main analysis appears to rely on the Lyne-Hollick (LH) method, the authors may wish to consider moving some of the detailed comparison of baseflow separation methods to the Supplementary Material.
Minor comments
- Lines 23–28: Was the 2022 drought more severe than the 2017–2018 drought? I am not sure, but I am curious. Otherwise, it may be better to mention both drought events (2022 and 2017–2018).
- The following paper on compound floods fits well with the content of the second paragraph and may be worth citing: DOI: 10.1126/sciadv.adl4005.
- Line 49: I do not fully agree with the statement that fewer studies investigate this topic using discharge time series than using climate data alone. Please consider rephrasing this statement.
- Lines 48–52: Explainable AI also offers useful approaches for analyzing this type of problem. Please consider adding a sentence from this perspective.
- Lines 59–69: It is not clear why such a detailed explanation of observed impacts and impact databases is needed here. Traditionally, the severity of floods and droughts in hydrology has been assessed based on streamflow. Please consider shortening or rephrasing this paragraph.
- Line 80: This is a good point and makes the paper particularly interesting.
- Figure 3: Could the two subplots be shown using the same regional division? For example, one of the major regional boundaries or the 17 SPC regions.
- Figures 7 and 8: The current colors do not show the spatial patterns very clearly. Please consider using a more divergent color scale.
- Lines 530–542: These findings are not particularly surprising given the climatology of the study area, with high flows in winter and low flows in summer. A more interesting question would be whether the timing of these transitions has changed over the years, for example becoming earlier or later as a result of climate change. However, I understand that this may be beyond the scope of the present manuscript, especially given its current length.
- Line 594: This statement is valid based on the historical data, but how might it change under a changing climate? Please consider adding a brief discussion.
Citation: https://doi.org/10.5194/egusphere-2025-6333-RC2 - AC2: 'Reply on RC2', Guilherme Mendoza Guimarães, 21 Mar 2026
Data sets
Characteristics of low and high flow spells and their temporal transitions in France Guilherme M. Guimarães et al. https://doi.org/10.57745/TNKVAY
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 305 | 223 | 25 | 553 | 12 | 12 |
- HTML: 305
- PDF: 223
- XML: 25
- Total: 553
- BibTeX: 12
- EndNote: 12
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Review of manuscript titled: “Characterizing low and high flow spells and their temporal transitions using baseflow estimates.”
The manuscript develops an analytical framework for drought to flood transitions, which focusses on the conditions which make these events important for decisionmakers. They address two questions: 2. how efficient is the framework at detecting hazards, and 2. How can the framework be used to inform decision-makers? They also assess different event severities lead to transitions, and assess different baseflow separation methods.
The paper is excellent, in my opinion. The method that they propose is, to my knowledge, novel, and their conclusions are interesting and valuable for hydrological science in general and for this emerging area of research in particular. It is well-written and easy to follow, and the figures are clear and useful. My primary suggestions relate to moving the comparison of baseflow separation methods to the supplement, and to tidying up the conclusions. I suggest publication with minor revisions.
Introduction:
I congratulate the authors, because both the abstract and the introduction are very well written, easy to follow, and clearly highlight the value and novelty of this paper.
I would acknowledge the comparison of baseflow methods briefly in the abstract, as this is a substantial component of the analysis.
Methods:
The methods are very clear for the most part. My main comment is that this paper would benefit greatly from a code release which demonstrates the application of the method presented.
The use of severity levels instead of strict threshold levels is a welcome methodological choice, as is the inclusion of transitions between consecutive events of the same type.
In its current form, the paper is quite long. I would consider whether it is worth while to move the comparison of baseflow methods to the supplementary materials. This would shorten your manuscript substantially and allow you to streamline your results.
Briefly comment on how the use of daily mean data may have influenced your results.
L191-193 Do you have any citations which support the idea that these time windows are relevant to risk management?
L207-215 I know that the Hermans-Rasson test addresses this to some extent, but have you considered how possible bimodality in event seasonality might influence this?
Results:
L 313 “as expected” or just, “by definition”
I understand the point of Figure 9, but I find it a bit difficult to follow. Could this information be presented in a more simplistic fashion?
Figure 12 is challenging to interpret due to the large number of parameters and the similar colors. I would suggest moving the donut charts to the supplementary materials to simplify visualization. Or remove the LFS->LFS and HFS->HFS transition panels and then add the information currently in the donut charts as a separate line. Other options might also work to simplify this. In the caption “yellow light colors” should be “light yellow colors”
Figure 13 – The size differences for concentration index are not different enough to interpret easily. I would also suggest simplifying this figure. In almost all of your figures (including this one), the text font is too small. The yellow text is difficult to read. Perhaps change the background color of the circle.
Discussion
L559-561 Why did you choose not to implement a smoothed threshold?
L 569 “our methodology allows”
Conclusions
I like that you have chosen to summarize your conclusions with bullet points, but I think that these should be refocussed to directly address the research aims outlined in the introduction.
Additionally, some of the conclusions, such as discussion of methodologies, should be moved to the discussion. In my opinion, this includes everything after line 672.
Overall, I would suggest reworking this section to be more succinct and focussed on the key contributions made by this paper.