Imprints of Increases in Evapotranspiration on Decreases in Streamflow during dry Periods, a large-sample Analysis in Germany&nbsp;

Bruno, Giulia; Strohmenger, Laurent; Duethmann, Doris

doi:https://doi.org/10.5194/egusphere-2024-2678

Preprints

https://doi.org/10.5194/egusphere-2024-2678

Preprints

04 Sep 2024

| 04 Sep 2024

Imprints of Increases in Evapotranspiration on Decreases in Streamflow during dry Periods, a large-sample Analysis in Germany

Giulia Bruno, Laurent Strohmenger, and Doris Duethmann

Abstract. Decreases in streamflow (Q) during dry periods have the potential to negatively affect river ecosystems and human societies, and understanding their causes is crucial to anticipate them. The contribution of increases in catchment actual evapotranspiration (E) to decreases in Q during dry periods remains poorly quantified. To address this gap, we performed a data-based analysis for 363 small (< 1000 km²) catchments without substantial water management influences in Germany over 1970–2019. We quantified trends in the magnitude of summer low flows, i.e., the minimum 7-day Q during summer months (7dQ_{min, JJA}). We attributed these trends to their main potential predictors (namely, long-term variations in E, summer precipitation, P, as well as spring and winter P as proxies for storage). Furthermore, we assessed potential changes in the annual P-Q relationship during a multi-year drought in the early 1990s, and investigated whether these changes were related with trends and anomalies in E and P. Summer low flows decreased significantly in 31 % of the catchments, with a median trend of -3.7 % decade^-1 across all catchments. Increases in E were a relevant driver of these decreases particularly in relatively drier Eastern catchments (contribution to long-term dynamics of 7dQ_{min, JJA} of 35 % based on multiple linear regression, and correlation coefficient between trends in 7dQ_{min, JJA} and in E of -0.74). Changes in the P-Q relationship occurred in 26 % of the catchments that experienced a multi-year drought between 1989 and 1993, with lower Q than expected from the relationship before the drought. These changes occurred in catchments with concurrent strong increases in E (median trend of 6.1 % decade^-1). Our findings point to the importance of increases in E, especially in dry catchments, when assessing potential future decreases in Q during dry periods for water management and climate adaptation strategies.

Received: 28 Aug 2024 – Discussion started: 04 Sep 2024

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.

Download & links

Preprint (PDF, 1369 KB)

Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
Preprint (1369 KB)

Supplement (846 KB)

Download & links

The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.

Journal article(s) based on this preprint

17 Sep 2025

Imprints of increases in evapotranspiration on decreases in streamflow during dry periods, a large-sample analysis in Germany

Giulia Bruno, Laurent Strohmenger, and Doris Duethmann

Hydrol. Earth Syst. Sci., 29, 4473–4489, https://doi.org/10.5194/hess-29-4473-2025,https://doi.org/10.5194/hess-29-4473-2025, 2025

Short summary

Giulia Bruno, Laurent Strohmenger, and Doris Duethmann

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2024-2678', Anonymous Referee #1, 26 Sep 2024
This paper analysed the summer low flows by using the datasets from 363 small catchments. The relationships among evapotranspiration (E), precipitation (P) and streamflow (Q), as well as the storage (S) were quantified. Results showed that summer low flows decreased significantly, and increased E played the main driver in the eastern catchments. In addition, the P-Q relationship changed in 26% of catchments between 1989 and 1993. Generally, the structure of the paper is clear and well-organized, however, there are some concerns.
Specific comments:
The coefficients of a and b in equation (2) should be vary with wind direction and elevation of gauges. Does authors calculate them for each gauges? If so, please provide the analysis which are the main uncertainty for precipitation datasets.

It is quite difficult to understand the equation (4), where the dynamics of storage was approximated by P_mam and P_djf. Please provide more explanations. In addition, since baseflow is 0.66 in this area, which means the soil moisture and groundwater both plays important roles in runoff variation. But it seems they are not taken into account in the analysis.

For multiple linear regression in predicting summer low flows, the authors showed the R2 in four clusters which exhibited good performance in Table S1. However, there is spatial variation among different gauges so the coefficients vary at each gauge, does the regression in the gauge scale follow the same trend with cluster?

Could authors discuss why the r= -0.49 in P_djf for the easter cluster in lines 258?

For 363 catchments, there is only 15 catchment with negative Cp-q rel values which distributed sparsely in Fig.8(a). I am curious about the possibility caused by data process uncertainty.

For Fig.5 and Table S1, it showed that P_jja played a major contribution to Q changes in Pre-Alphine and South-Central cluster, while E played much more contribution in Eastern and Northern area. Does it relate to elevation changes?
Citation: https://doi.org/10.5194/egusphere-2024-2678-RC1
- AC1: 'Reply on RC1', Giulia Bruno, 10 Dec 2024
  
  Please refer to the attached pdf.
  
  Citation: https://doi.org/10.5194/egusphere-2024-2678-AC1
RC2:
'Comment on egusphere-2024-2678', Anonymous Referee #2, 02 Oct 2024
Bruno et al. analyzed the decreasing trend of low flow in 363 small catchments in Germany. They also attributed the decrease to the increase of ET. They further unraveled that the change of P-Q relationship during drought produces lower flow which is generally due to the increased ET. They conducted this work based on observations of P and Q, empirical expression of subsurface storage, and ET derived based on water balance and statistical analysis. I think studying the decrease of low flow and trying to find the major drivers is very important in the climate change background. The data and analysis are generally reliable, the structure and the writing are good. However, there the following concerns which need clarification from authors for further review.
I was more or less confused by the overall idea of the authors. If you wanted to check if the decrease of the flow is caused by increased ET, and you also have calculated the water balance, why not do a straightforward analysis of the overall change of P, ET, Q, and S. Then it is easy to get if the decrease of flow is mainly driven by ET. Then you can do the analysis in your manuscript as a follow-up. Otherwise, I feel the conclusions are even not that convincing as, for example, the decrease of low flow might be just because of the shift of the timing of streamflow.

You mentioned you used Kirchner’s approach to calculate S_dyn which needs that S is the main control of Q generation. I am wondering if the 363 catchments you used meet this requirement and where is your analysis for this?

So, how do you quantify the uncertainties in E you derived from water balance as I am not sure the uncertainties in S_dyn.

Also, I have to say, for the catchments with areas ranging from 50-150km², the lateral groundwater flow is significant which has been discussed in Ying Fan’s paper ‘Are catchments leaky?’ and also quantified in our research (not published yet). Therefore, Equation 3 might be problematic.

Line 161, P_DIF and P_MAM are used as proxies of storage processes. How and why they e used as proxies?

All conclusions occur in less than 30% of the catchments, so how do you think about the generality of the study?

I am wondering why ET is increasing? The land cover change or the temperature increasing? Authors listed a lot of attributes of the catchments in table 1 but are limited used in the analysis.

Have you cited the paper talking about the similar thing? Tran et al. (2023), Frontiers in Water.

I suggest returning this manuscript to authors for major revision.
Citation: https://doi.org/10.5194/egusphere-2024-2678-RC2
- AC2: 'Reply on RC2', Giulia Bruno, 10 Dec 2024
  
  Please refer to the attached pdf.
  
  Citation: https://doi.org/10.5194/egusphere-2024-2678-AC2

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2024-2678', Anonymous Referee #1, 26 Sep 2024
This paper analysed the summer low flows by using the datasets from 363 small catchments. The relationships among evapotranspiration (E), precipitation (P) and streamflow (Q), as well as the storage (S) were quantified. Results showed that summer low flows decreased significantly, and increased E played the main driver in the eastern catchments. In addition, the P-Q relationship changed in 26% of catchments between 1989 and 1993. Generally, the structure of the paper is clear and well-organized, however, there are some concerns.
Specific comments:
The coefficients of a and b in equation (2) should be vary with wind direction and elevation of gauges. Does authors calculate them for each gauges? If so, please provide the analysis which are the main uncertainty for precipitation datasets.

It is quite difficult to understand the equation (4), where the dynamics of storage was approximated by P_mam and P_djf. Please provide more explanations. In addition, since baseflow is 0.66 in this area, which means the soil moisture and groundwater both plays important roles in runoff variation. But it seems they are not taken into account in the analysis.

For multiple linear regression in predicting summer low flows, the authors showed the R2 in four clusters which exhibited good performance in Table S1. However, there is spatial variation among different gauges so the coefficients vary at each gauge, does the regression in the gauge scale follow the same trend with cluster?

Could authors discuss why the r= -0.49 in P_djf for the easter cluster in lines 258?

For 363 catchments, there is only 15 catchment with negative Cp-q rel values which distributed sparsely in Fig.8(a). I am curious about the possibility caused by data process uncertainty.

For Fig.5 and Table S1, it showed that P_jja played a major contribution to Q changes in Pre-Alphine and South-Central cluster, while E played much more contribution in Eastern and Northern area. Does it relate to elevation changes?
Citation: https://doi.org/10.5194/egusphere-2024-2678-RC1
- AC1: 'Reply on RC1', Giulia Bruno, 10 Dec 2024
  
  Please refer to the attached pdf.
  
  Citation: https://doi.org/10.5194/egusphere-2024-2678-AC1
RC2:
'Comment on egusphere-2024-2678', Anonymous Referee #2, 02 Oct 2024
Bruno et al. analyzed the decreasing trend of low flow in 363 small catchments in Germany. They also attributed the decrease to the increase of ET. They further unraveled that the change of P-Q relationship during drought produces lower flow which is generally due to the increased ET. They conducted this work based on observations of P and Q, empirical expression of subsurface storage, and ET derived based on water balance and statistical analysis. I think studying the decrease of low flow and trying to find the major drivers is very important in the climate change background. The data and analysis are generally reliable, the structure and the writing are good. However, there the following concerns which need clarification from authors for further review.
I was more or less confused by the overall idea of the authors. If you wanted to check if the decrease of the flow is caused by increased ET, and you also have calculated the water balance, why not do a straightforward analysis of the overall change of P, ET, Q, and S. Then it is easy to get if the decrease of flow is mainly driven by ET. Then you can do the analysis in your manuscript as a follow-up. Otherwise, I feel the conclusions are even not that convincing as, for example, the decrease of low flow might be just because of the shift of the timing of streamflow.

You mentioned you used Kirchner’s approach to calculate S_dyn which needs that S is the main control of Q generation. I am wondering if the 363 catchments you used meet this requirement and where is your analysis for this?

So, how do you quantify the uncertainties in E you derived from water balance as I am not sure the uncertainties in S_dyn.

Also, I have to say, for the catchments with areas ranging from 50-150km², the lateral groundwater flow is significant which has been discussed in Ying Fan’s paper ‘Are catchments leaky?’ and also quantified in our research (not published yet). Therefore, Equation 3 might be problematic.

Line 161, P_DIF and P_MAM are used as proxies of storage processes. How and why they e used as proxies?

All conclusions occur in less than 30% of the catchments, so how do you think about the generality of the study?

I am wondering why ET is increasing? The land cover change or the temperature increasing? Authors listed a lot of attributes of the catchments in table 1 but are limited used in the analysis.

Have you cited the paper talking about the similar thing? Tran et al. (2023), Frontiers in Water.

I suggest returning this manuscript to authors for major revision.
Citation: https://doi.org/10.5194/egusphere-2024-2678-RC2
- AC2: 'Reply on RC2', Giulia Bruno, 10 Dec 2024
  
  Please refer to the attached pdf.
  
  Citation: https://doi.org/10.5194/egusphere-2024-2678-AC2

Peer review completion

AR: Author's response | RR: Referee report | ED: Editor decision | EF: Editorial file upload

ED: Reconsider after major revisions (further review by editor and referees) (24 Dec 2024) by Khalid Hassaballah

AR by Giulia Bruno on behalf of the Authors (03 Feb 2025) Author's response Author's tracked changes Manuscript

ED: Publish subject to revisions (further review by editor and referees) (02 Apr 2025) by Khalid Hassaballah

Technical Comments and Clarification for Improvement:

1- Clarification on PDJF and Summer Low Flows (Referee #1, Question 4):
In the author’s response to Referee #1 regarding Question 4, further clarification is needed on how increases in PDJF (precipitation during December–February) could have indirectly contributed to decreases in summer low flows in the Eastern cluster. Specifically, how can increased precipitation potentially result in lower streamflow during summer? This seems counterintuitive and should be better explained or corrected.

2- Trend Analysis in Abstract and Main Text (L20–21):
The statement: "Summer low flows decreased (increased) significantly in 31% (2%) of the catchments, with a median trend of -3.7% per decade across all catchments.” needs further examination. If 67% of the catchments showed no significant trend, what is the rationale for computing the median trend across all catchments, including those without significant trends? Additionally, it would be more informative to report the median trend separately for the 31% of catchments showing decreasing trends and the 2% with increasing trends. This would help to clarify the magnitude and direction of significant changes.

3- Ambiguity in Trend Percentages (L225–227):
The sentence: “Trends in 7dQmin, JJA were significantly negative in 31% of the catchments and significantly positive in 2% of them (negative in 77% and positive in 23%, Fig. 4b), with a median (interquartile range, IQR) of -3.7 (-7.5/-0.6)% per decade across all catchments (Fig. 4c).” is confusing. It is unclear how the 31% and 2% (which refer to statistically significant trends) relate to the 77% and 23% figures (which seem to include non-significant trends). Please clarify the distinction between statistically significant trends and overall direction of trends, and how these percentages are used in analysis and interpretation.

4- Over-Reliance on Previous Studies:
While referencing previous literature is important for contextualization, several of the author’s responses rely heavily on prior studies without providing sufficient evidence from the current analysis. It is recommended to strengthen the manuscript by presenting clearer links between your results and your arguments, supported by direct evidence from your own data.

5- Language and Grammar Corrections:
Please revise the manuscript for language accuracy. For example, in Section 4.4:
“Due to the generally coarse resolution of satellite-derived E estimates, we computed E from observed P and Q, and estimates of the Sdyn of the catchments from Q data as a first order approximation of S.”
This sentence could be rewritten for clarity and correctness. A suggestion:
“Given the generally coarse resolution of satellite-derived E estimates, we derived E using observed P, Q, and estimates catchment dynamic storage (Sdyn) from Q data, as a first-order approximation of S.”

Hide

AR by Giulia Bruno on behalf of the Authors (26 May 2025) Author's response Author's tracked changes Manuscript

ED: Publish subject to technical corrections (09 Jun 2025) by Khalid Hassaballah

Technical Comments and Clarification for Improvement:
1- Clarification on PDJF and Summer Low Flows (Referee #1, Question 4):
Authors response: “We argued that increases in winter precipitation (PDJF) might have indirectly contributed to decreases in summer low flows in the eastern cluster via feedbacks with evapotranspiration (E). Specifically, increases in PDJF might have supported increases in storage (S) recharge in winter and thus in S levels at the beginning of the growing season. In turn, increases in S levels at the beginning of the growing season potentially contributed to increased vegetation growth and therefore E, by further contributing to reduced summer low flows.”

Further Comment by the Editor:
The conceptual linkage between increased precipitation, higher soil moisture or groundwater storage, enhanced vegetation, increased evapotranspiration is well-documented in ecohydrology and hydrological literature. However, in your response, the term storage (S) is ambiguous. Please clarify whether you are referring to soil moisture storage or groundwater storage.
While I agree that increased vegetation growth can lead to higher evapotranspiration and a subsequent reduction in soil moisture, this mechanism does not directly apply to groundwater storage. On the contrary, an increase in winter precipitation that enhances groundwater recharge would typically support higher summer low flows, not lower. This appears to contradict the mechanism you propose.
Additionally, I encourage you to address the directionality of this relationship. If a decrease in winter precipitation (PDJF) will also lead to a reduction in summer low flows, how do you reconcile that with the claim that an increase in PDJF also results in reduced summer low flows? If both increasing and decreasing PDJF will lead to the same outcome (reduced summer low flows), a clearer justification or an alternative explanation is needed to resolve this apparent inconsistency.

Hide

ED: Publish subject to technical corrections (13 Jun 2025) by Giuliano Di Baldassarre (Executive editor)

AR by Giulia Bruno on behalf of the Authors (11 Jul 2025) Author's response Manuscript

Journal article(s) based on this preprint

17 Sep 2025

Imprints of increases in evapotranspiration on decreases in streamflow during dry periods, a large-sample analysis in Germany

Giulia Bruno, Laurent Strohmenger, and Doris Duethmann

Hydrol. Earth Syst. Sci., 29, 4473–4489, https://doi.org/10.5194/hess-29-4473-2025,https://doi.org/10.5194/hess-29-4473-2025, 2025

Short summary

Giulia Bruno, Laurent Strohmenger, and Doris Duethmann

Supplement

https://doi.org/10.5194/egusphere-2024-2678-supplement

Giulia Bruno, Laurent Strohmenger, and Doris Duethmann

Viewed

Total article views: 1,051 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
554	104	393	1,051	52	31	41

HTML: 554
PDF: 104
XML: 393
Total: 1,051
Supplement: 52
BibTeX: 31
EndNote: 41

Views and downloads (calculated since 04 Sep 2024)

Month	HTML	PDF	XML	Total
Sep 2024	80	21	46	147
Oct 2024	32	4	57	93
Nov 2024	25	3	95	123
Dec 2024	35	16	103	154
Jan 2025	13	8	82	103
Feb 2025	20	1	2	23
Mar 2025	12	11	0	23
Apr 2025	16	4	0	20
May 2025	12	6	1	19
Jun 2025	16	10	2	28
Jul 2025	12	10	1	23
Aug 2025	114	8	3	125
Sep 2025	167	2	1	170

Cumulative views and downloads (calculated since 04 Sep 2024)

Month	HTML	PDF	XML	Total
Sep 2024	80	21	46	147
Oct 2024	32	4	57	93
Nov 2024	25	3	95	123
Dec 2024	35	16	103	154
Jan 2025	13	8	82	103
Feb 2025	20	1	2	23
Mar 2025	12	11	0	23
Apr 2025	16	4	0	20
May 2025	12	6	1	19
Jun 2025	16	10	2	28
Jul 2025	12	10	1	23
Aug 2025	114	8	3	125
Sep 2025	167	2	1	170

Viewed (geographical distribution)

Total article views: 1,079 (including HTML, PDF, and XML) Thereof 1,079 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 17 Sep 2025

Download

The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.

Preprint (1369 KB)
Metadata XML

Short summary

Decreases in streamflow during dry periods threaten ecosystems and society, and increases in evapotranspiration may contribute to them. From data for small catchments in Germany, summer low flows decreased over 1970–2019 and increases in evapotranspiration relevantly contributed. Stronger-than-expected decreases in streamflow during the 1989–1993 drought occurred in catchments with increases in evapotranspiration. Increases in evapotranspiration need full consideration for streamflow prediction.


Total:	0
HTML:	0
PDF:	0
XML:	0