Contrasting drivers of riparian&ndash;stream hydrological dynamics during rainfall events across forest headwater ecoregions

Ledesma, José L. J.; Burkhardt, Paul D.; Bernal, Susana; Musolff, Andreas

doi:10.5194/egusphere-2026-4221

Preprints

https://doi.org/10.5194/egusphere-2026-4221

Preprints

17 Jul 2026

| 17 Jul 2026

Contrasting drivers of riparian–stream hydrological dynamics during rainfall events across forest headwater ecoregions

José L. J. Ledesma, Paul D. Burkhardt, Susana Bernal, and Andreas Musolff

Abstract. Examining how hydrological responses of streams and riparian groundwater interact during rainfall events can help understand runoff generation mechanisms and their links to biogeochemistry. We analysed rainfall–runoff events in four forest headwater catchments spanning semiarid Mediterranean, subhumid Mediterranean, temperate, and boreal ecoregions. Using sub-daily data, we evaluated how hydroclimatic drivers shaped stream and riparian groundwater. We additionally focussed on the slope (slopeLin) and hysteresis (h) of the riparian groundwater–stream relationship, describing coupled responses and connectivity patterns. Rainfall amount and antecedent wetness primarily controlled individual stream and riparian groundwater hydrological response across sites, while the vertical activation of riparian layers, defined by slopeLin, and timing, defined by h, were site-specific. At the Mediterranean sites, deep, low-conductivity layers dominated the lateral connectivity unless sufficient rainfall and wetness allowed activation of shallower flow pathways. At the temperate and boreal sites, shallow, conductive layers were typically active, with temporary disconnection during warm periods at the temperate site and snowmelt-induced activation of upslope sources at the boreal. Clockwise hysteresis, indicating faster riparian than stream responses, prevailed at the temperate and boreal sites. The Mediterranean sites showed weaker or anticlockwise patterns, suggesting activation of other water sources, likely ephemeral tributaries during intense rainfall. Our results show that the activation and hysteresis of riparian–stream connectivity differs across ecoregions, which has implications for solute mobilization. This framework provides a physically grounded perspective for interpreting both rainfall–runoff and concentration–discharge relationships, which can help predict shifts in catchment functioning under changing hydroclimatic regimes.

Received: 14 Jul 2026 – Discussion started: 17 Jul 2026

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

José L. J. Ledesma, Paul D. Burkhardt, Susana Bernal, and Andreas Musolff

Status: final response (author comments only)

RC1:
'Comment on egusphere-2026-4221', Anonymous Referee #1, 28 Aug 2026
General comments
The authors of this manuscript offer an interesting comparison of hydroclimatic factors, runoff response, and the relationships between discharge and groundwater levels across four forested catchments subject to different climate forcings.
The manuscript is well structured and clearly written, though it needs some clarification and improvement. While I recognize the value of comparative studies in advancing our understanding of riparian-stream connectivity, I think that four catchments are insufficient to draw broader conclusions about hydrological processes across different ecoregions or climatic conditions. For instance, there is no evidence that the catchments are representative or that the selected period reflects their respective climate. The selected period is very short (<2 years) for two catchments (Upper Rappbode and Krycklan C2), and the rainfall-runoff events may have occurred in years that were wetter or drier than the average. Furthermore, we cannot know whether the observed differences in runoff response across the four catchments are due more to lithology and soils than to climatic factors. The only direct comparison is possible between the two Mediterranean study areas, which have similar lithology and soil textures and overlapping periods for data analysis. I therefore suggest revising the sentences that imply the results from these four catchments are representative of their respective ecoregions.
Finally, several methodological details should be included to clarify the data analysis (e.g., hysteresis sketches/plots, plots showing the relationship between discharge and groundwater table, and a time series of the main hydrometeorological variables).

Specific comments
Lines 23-26: Here there is too much emphasis on the fact that the different results are due to the different ecoregions, while this research is based on just four study sites that may not be good representatives of their ecoregion. The reference to the concentration-discharge relationships is also not supported by specific data analyses in this manuscript.

Lines 67-75: In this paragraph, please clarify the specific research questions and the specific objectives.

Lines 72-73: It is not clear how it is possible to ‘identify both shared and ecoregion-specific hydrological response’ with data from just four catchments.

Section 2.1: Given that the groundwater table was monitored in the riparian zone, I suggest adding the area of the riparian zones and some key topographic characteristics, such as their width and mean slope.

Line 131: There is a wide range in temporal resolutions spanning from 10 minutes to 1 hour. A 1-hour resolution may be too coarse for short but intense rainfall-runoff events, which may occur at the Mediterranean sites and in Upper Rappbode. Is a 1-hour resolution effective for capturing temporal dynamics that can last for a very few hours?

Section 2.2: For each site, I suggest adding the depth of the well, whether it was installed at the soil-bedrock interface, and the distances from the stream and from the stream gauge.

Lines 177-179: Were these percentage increases compared to other values? Did you perform a sensitivity analysis before choosing these specific values?

Lines 182-187: I recommend adding an equation to explain the composite index, in order to better understand how these values were obtained.

Lines 215-220: A sketch or examples from the study sites would be very helpful to comprehend the slopeLin metric.

Lines 223-238: I recommend adding a sketch or examples of hysteresis loops and of h-index/hAbs values to help understand this analysis.

Line 246: I suggest providing a brief description of the multimodel inference approach, as it may not be very well known. Furthermore, why did you adopt this method instead of multiple linear regressions?

Line 254: Why did you focus only on linear regression models, when non-linear relations and threshold behaviors may be present?

Line 284: There is a missing word between ‘particularly’ and ‘at the semiarid’.

Line 405: Please replace ‘studied ecoregions’ with ‘studied catchments’ as this research is based only on four catchments.

Section 4.3: I do not understand the extended text about the c-Q relationships, which were not the focus of this research. I suggest shortening the text about c-Q relations or providing examples in the manuscript. In addition, I suggest a separate conclusions section that includes only the key findings.

Suggested figures: A time series of the main hydrometeorological variables or some key examples of rainfall-runoff event dynamics in each catchment could be added to the manuscript or included in the supplementary material.
Citation: https://doi.org/10.5194/egusphere-2026-4221-RC1
- AC1: 'Reply on RC1', José L. J. Ledesma, 18 Sep 2026
  
  Thank you for the constructive evaluation of our manuscript. We are glad you found it interesting, structured, and clearly written.
  We have grouped the main points raised in the review into four broader issues. Below, we respond to each of them and outline how we intend to address them in the revised manuscript.
  Issue 1. Representativeness of the study sites and limits of ecoregional inference.
  We agree that four catchments, one within each ecoregional setting, are insufficient to make strong generalizations about hydrological processes across the respective regions, and we will tone down wording that may imply such representativeness. Nevertheless, this study was designed as an event-scale comparison across contrasting environmental settings rather than as a climatological assessment of those regions. In this context, capturing a sufficiently broad range of rainfall-runoff events and associated conditions at each site is particularly relevant, even when the monitoring periods differ in length. The sites also show strong differences in temperature, evapotranspiration and, to a lesser extent, precipitation, providing clearly contrasting hydroclimatic contexts for the event responses. To further support this point, we will present the conditions during the study periods in relation to available long-term climatic conditions at each site.
  We also agree that the effects of climate cannot be isolated from other correlated catchment characteristics, including soils, lithology, topography, and vegetation. However, our intention was not to attribute the observed differences to climate alone, but to compare hydrological behaviour across broader ecoregional settings in which these environmental characteristics naturally covary. We will clarify this point and revise the corresponding wording so that the four catchments are presented as contrasting sites spanning different ecoregional settings rather than as definitive representatives of those regions.
  Issue 2. Additional illustrative plots and time series.
  We agree that additional graphical information would facilitate the interpretation of the analyses. In the revised manuscript, we will therefore include representative examples of the relationships between stream discharge and riparian groundwater table, including contrasting hysteresis patterns, as well as time series illustrating the main hydrometeorological dynamics. These will most likely be included as supporting figures to avoid unnecessarily extending the main manuscript. We will also consider including a schematic figure synthesizing the different hysteresis metrics used, which may help guide the reader through their interpretation.
  Issue 3. Adequacy of temporal resolution for event-scale dynamics.
  We note that the concern regarding a 1-h resolution at the Mediterranean and temperate sites results from a misunderstanding. The 1-h resolution was used only at the boreal site, whereas both Mediterranean sites and the temperate site were analysed at 30-min resolution, as stated in L. 158-161 (“The time resolution used for rainfall-runoff event identification at each site was determined by the coarsest resolution among precipitation, stream discharge, and riparian groundwater table data (Table 1). […]. As a result, a 30-min resolution was used for both Mediterranean sites and the temperate site, while a 1-h resolution was applied at the boreal site”).
  We consider the 30-min resolution adequate for the event-scale dynamics analysed here. Our approach was designed to characterize hydrological responses developing over several hours and throughout complete rainfall-runoff events rather than sub-hourly fluctuations. Indeed, event identification itself relied on sustained discharge changes over 3- and 6-h periods (see current L. 176-187). We will nevertheless make this distinction clearer in the revised manuscript. At the boreal site, the 1-h resolution was imposed by the available groundwater observations and we consider it sufficient for the event-scale metrics examined here, while acknowledging that sub-hourly variability cannot be resolved.
  Issue 4. Justification of the linear model structure in the MMI framework.
  We agree that nonlinear and threshold relationships may occur in event-scale hydrological responses. However, the purpose of the MMI analysis was not to identify the complete functional form of each predictor-response relationship, but to provide a parsimonious and comparable assessment of the first-order direction and relative importance of hydroclimatic controls across response metrics and catchments. We therefore used multiple linear regression models within the MMI framework, which allowed standardized coefficients and partitioned explained variance to be compared consistently among sites.
  We deliberately did not incorporate nonlinear terms, thresholds, or interactions because this would substantially increase the candidate-model space relative to the number of available events and increase the risk of overfitting. In addition, the preceding Spearman correlation analysis does not require linear relationships and provides a complementary assessment of monotonic associations. We will clarify this rationale and explicitly acknowledge that nonlinear or threshold behaviour may contribute to relationships not fully captured by the linear models.
  
  Besides these main points, we will also address all remaining specific and line-level comments in the revised manuscript and provide a detailed point-by-point response at the revision stage.
  Thank you,
  
  The authors
  
  Citation: https://doi.org/10.5194/egusphere-2026-4221-AC1
RC2:
'Comment on egusphere-2026-4221', Anonymous Referee #2, 04 Sep 2026

General comments: This is a very interesting study of runoff-groundwater-streamflow relationships across very different climates. The manuscript was well-written and informative, and I think the study was well-designed.
There are some areas where the manuscript could benefit from clarification or additional information. For example, words like “ecoregion” and “stream discharge” have commonly-used definitions that are inconsistent with their use in the manuscript. More concrete, specific language would greatly improve clarity. I also had concerns about the analytical methods, especially the use of linear models; a justification for why those were chosen would alleviate those concerns.
Line 68: Are these ecoregions or climate regions? Usually ecoregion names have some descriptor of vegetation (e.g., Central European mixed forest) or topography (e.g., southern semi-arid highlands). You should use (and cite) standardized ecoregion names, such as the Digital Map of European Ecological Regions, rather than the climate-based descriptors you are currently using. Here is a link to the DMEER: https://www.eea.europa.eu/en/analysis/maps-and-charts/dmeer-digital-map-of-european-ecological-regions?activeTab=a7caf3b5-7254-4a24-8919-693d4115158b
My other concern with using the term “ecoregion” is that ecology is not a focus of this manuscript. Perhaps Köppen-Geiger climate classifications would be more appropriate. An even better idea would be to simply identify them by the name of the catchments.
Line 116: Which characterization, humid or energy-limited?
Section 2.2: How far were the wells from the streams? If you add this information to your revised manuscript, you’ll also need to specify where it is measured from (i.e., thalweg, center of channel, or some other location).
Line 215: I would be surprised if that relationship is linear. Do you have a citation for this method? Or some justification for using it?
Table 3: Stream discharge measurement should include both volume and time (e.g., m³/s) because it is calculated using volume and velocity. If you are using the word “discharge” to signify something else, such as unit discharge or stage, you should specify that and explain how the variable was calculated. Additionally, for your total discharge, you should specify the time period (e.g., daily, annual).
Section 3.1: You use the phrase “tended to be” a few times in this section. I recommend using more emphatic language instead.
Section 4: There are a lot of variables that drive runoff that are not addressed by your study. You mentioned a few in your introduction (topography, land use, etc.). Of course they can’t all be included, but I would like to see some acknowledgement of those in your discussion. For example, you mentioned drought and a bark beetle infestation that led to vegetation removal at your Upper Rappbode site. How could vegetation cover, or changes in vegetation cover, influence your results?
Lines 444-452: This is another example from my previous comment. Could steep slopes in the catchment play a role here? Your map shows a lot more topographic variability at the Mediterranean sites compared to the temperate and boreal sites. Slope definitely affects runoff, but it wasn’t really addressed in your manuscript other than a quick mention in the site descriptions.
Line 460-461: I agree with this and you should provide some justification here for using linear models.
Section 4.3: This section about solute mobilization feels a bit disjointed from the rest of the manuscript. If you emphasize this more in the introduction, it will tie the whole manuscript together more.

Citation: https://doi.org/10.5194/egusphere-2026-4221-RC2
- AC2: 'Reply on RC2', José L. J. Ledesma, 18 Sep 2026
  
  Thank you for the constructive evaluation of our manuscript. We are glad you found it interesting, well-written, informative, and well-designed.
  We have grouped the main points raised in the review into four broader issues. Below, we respond to each of them and outline how we intend to address them in the revised manuscript.
  Issue 1. Ecoregion terminology.
  We agree that our use of the term “ecoregion” was not sufficiently defined and could be interpreted as referring to a specific standardized classification such as DMEER. This was not our intention. Here, as in Ledesma et al. (2025), we use “ecoregion” in the broader ecosystem-geographical sense of Bailey (2014), i.e., referring to large-scale ecological regions differentiated primarily by macroclimate and associated vegetation and ecosystem characteristics, with additional differences in landform, soils, and other environmental properties. A similarly broad concept is used by the US EPA, which defines ecoregions based on patterns in climate, vegetation, soils, geology, landforms, hydrology, and other ecosystem characteristics (https://www.epa.gov/eco-research/ecoregions).
  We therefore consider the term appropriate for describing the broad contrasting settings represented by our four catchments, but agree that the terminology should be used more rigorously. In the revised manuscript, we will explicitly define our use of “ecoregion”, clarify that semiarid Mediterranean, subhumid Mediterranean, temperate, and boreal are descriptive labels rather than names from a standardized ecoregion classification, and restrict the use of the term where more specific reference to the individual catchments is preferable. We will also avoid wording that could imply strong extrapolation of the results to entire ecoregions.
  Issue 2. Stream discharge terminology.
  We agree that the stream discharge terminology can be further clarified. Stream discharge was normalized by contributing catchment area and expressed as equivalent water depth, a common approach in catchment hydrology that facilitates comparisons among catchments of different size. We will make this explicit in the revised manuscript and ensure that the corresponding time dimension is clearly stated where relevant, as already done for Qavg (mm d−1) and Qmax (mm h−1). Qsum represents the total runoff depth integrated over the complete event, from t0 to tf.
  For dQ, the metric represents the difference between maximum event discharge and discharge at t0 rather than an accumulated discharge over an additional time period. This calculation was performed at the common temporal resolution used for each site, i.e. 30 min at the Mediterranean and temperate sites and 1 h at the boreal site. We will clarify this calculation and ensure that its units are reported consistently in Table 3.
  Issue 3. Unaccounted catchment controls and their implications for interpretation.
  We agree that runoff dynamics can also be influenced by catchment characteristics that were not explicitly included in our event-scale analysis, including topography, soils, geology, and vegetation cover. Our objective was to examine hydroclimatic controls on event responses rather than to provide an exhaustive attribution of all possible runoff controls. We will make this limitation more explicit in the Discussion.
  At Upper Rappbode, vegetation loss associated with drought and bark beetle disturbance did reduce canopy interception and evapotranspiration, thereby increasing effective precipitation and enhancing runoff generation and hydrological responsiveness, as shown by Musolff et al. (2024). We will acknowledge this as an additional control on the observed hydrological behaviour. Likewise, the steeper topography of the Mediterranean sites could contribute to some of the observed response patterns and will be discussed as another site-specific control.
  Issue 4. Justification of the linear model structure in the MMI framework.
  We agree that nonlinear and threshold relationships may occur in event-scale hydrological responses. However, the purpose of the MMI analysis was not to identify the complete functional form of each predictor-response relationship, but to provide a parsimonious and comparable assessment of the first-order direction and relative importance of hydroclimatic controls across response metrics and catchments. We therefore used multiple linear regression models within the MMI framework, which allowed standardized coefficients and partitioned explained variance to be compared consistently among sites.
  We deliberately did not incorporate nonlinear terms, thresholds, or interactions because this would substantially increase the candidate-model space relative to the number of available events and increase the risk of overfitting. In addition, the preceding Spearman correlation analysis does not require linear relationships and provides a complementary assessment of monotonic associations. We will clarify this rationale and explicitly acknowledge that nonlinear or threshold behaviour may contribute to relationships not fully captured by the linear models.
  
  Besides these main points, we will also address all remaining specific and line-level comments in the revised manuscript and provide a detailed point-by-point response at the revision stage.
  Thank you,
  
  The authors
  References
  
  Bailey, R. G.: Ecoregions: The Ecosystem Geography of the Oceans and Continents, 2nd ed., Springer, New York, 180 pp., https://doi.org/10.1007/978-1-4939-0524-9, 2014.
  Ledesma, J. L. J., Musolff, A., Sponseller, R. A., Lupon, A., Peñarroya, X., Jativa, C., and Bernal, S.: The riparian zone controls headwater hydrology and biogeochemistry, doesn't it? Reassessing linkages across European ecoregions, Global Biogeochemical Cycles, 39, e2024GB008250, https://doi.org/10.1029/2024GB008250, 2025.
  Musolff, A., Tarasova, L., Rinke, K., and Ledesma, J. L. J.: Forest dieback alters nutrient pathways in a temperate headwater catchment, Hydrological Processes, 38, e15308, https://doi.org/10.1002/hyp.15308, 2024
  
  Citation: https://doi.org/10.5194/egusphere-2026-4221-AC2

José L. J. Ledesma, Paul D. Burkhardt, Susana Bernal, and Andreas Musolff

Data sets

Event-level hydroclimatic and hydrological data for 209 rainfall–runoff events across four forest headwater catchments José L. J. Ledesma https://www.hydroshare.org/resource/17075f4b1f874f389c69f1f491bdafdb/

José L. J. Ledesma, Paul D. Burkhardt, Susana Bernal, and Andreas Musolff

Viewed

Total article views: 246 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
153	71	22	246	18	19

HTML: 153
PDF: 71
XML: 22
Total: 246
BibTeX: 18
EndNote: 19

Views and downloads (calculated since 17 Jul 2026)

Month	HTML	PDF	XML	Total
Jul 2026	82	40	14	136
Aug 2026	28	14	1	43
Sep 2026	43	17	7	67

Cumulative views and downloads (calculated since 17 Jul 2026)

Month	HTML	PDF	XML	Total
Jul 2026	82	40	14	136
Aug 2026	28	14	1	43
Sep 2026	43	17	7	67

Viewed (geographical distribution)

Total article views: 228 (including HTML, PDF, and XML) Thereof 228 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 18 Sep 2026

Short summary

We compared more than 200 rainfall events in four small forested river basins, from Mediterranean Spain to Germany and Sweden. Rainfall amount and prior wetness controlled how strongly streams and nearby groundwater responded, but the timing and connection between them differed among climates. These differences help explain when water and dissolved substances are mobilised and how headwater environments may respond to a changing climate.


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