the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 12 Jan 2026
Groundwater Hysteresis Increasingly Decouples Flowing Network Length from Streamflow as Snow Shifts to Rain
Abstract. Flowing stream networks expand and contract in response to dynamic groundwater levels. Field studies generally associate greater flowing network length (L) with higher streamflow (Q), but this neglects potential hysteresis caused by nonequilibrium groundwater flow after rain and snowmelt. Using a new version of the Distributed Hydrology Soil Vegetation Model (DHSVM), we predict that groundwater hysteresis may decouple L from Q across large (> 100 %) variations in Q. In a 27 km2 snowy volcanic watershed, seasonal anomalies in measured stream ionic concentration indicate an outsized contribution from longer subsurface flowpaths during recession, supporting our L-Q hysteresis hypothesis and refining our model calibration. The model can reproduce observed stream network elasticity (from field surveys), and the predicted network length anomaly mirrors seasonal anomalies in measured stream ionic concentration (r = −0.92), suggesting that the model can capture the seasonal reconfiguration of groundwater flowpaths. A warmer climate is expected to cause a partial transition from snow to rain resulting in flashier streamflow, but our simulations predict that seasonal groundwater hysteresis would dampen storm-scale stream network elasticity, thereby significantly increasing L-Q hysteresis on daily to monthly timescales (p < 0.01). Conceptual models of stream networks should consider the potential effects of groundwater hysteresis in headwaters catchments, especially in a changing environment. More broadly, our investigation highlights how spatially distributed process-based hydrological modeling can sometimes reveal emergent hydrological behaviors that are not apparent from sparse field data.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Hydrology and Earth System Sciences.
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
(2997 KB) - Metadata XML
-
Supplement
(2390 KB) - BibTeX
- EndNote
Status: final response (author comments only)
- CC1: 'Comment on egusphere-2025-6294', Giacomo Medici, 16 Jan 2026
-
RC1: 'Comment on egusphere-2025-6294', Anonymous Referee #1, 25 Feb 2026
This study from Boardman et al. describes reconfiguration and application of a spatially distributed, physically based hydrologic model to simulate flowing network length and streamflow in a headwater catchment. The model shows expansion and contraction of the flowing stream network relative to discharge, and the authors compare the model outputs to a common field-based description of L-Q using a power-law relationship. An important contribution of the work is that the model reveals nonmonotonicity in the simulated L-Q relationship, which the authors state is due to the hysteretic nature of the flowing stream length and flow in the watershed driven by groundwater.
General comments:
- Both the science and the presentation of the manuscript are very high quality. I have several relatively minor comments that in my opinion the authors should address; however I suggest that the paper be published in HESS upon completion of these minor comments.
- In my opinion, this paper represents an important step forward for the headwater modeling community.
Specific comments:
- L28: This definition of headwater streams doesn’t align with what current literature suggests (Cf. Golden et al., 2025; https://doi.org/10.1038/s44221-024-00351-1). I recommend revision or inclusion of a citation where other folks have adopted this definition that you all use.
- L176: I really appreciate the written description of the model formulation. However, it would also be useful for future studies if equations were presented here that describe how the processes were parameterized. Please consider inclusion of these.
- L196: It seems odd to jump from S1 and S2 to S10.
- L260: Can reaches be less than 30-m in the formulation? If so, where is that possible and how are those lengths determined.
- L286: Didn’t you state that this version of DHSVM has a new instream routing algorithm? Wouldn’t prior sensitivity analyses of DHSVM therefore not be applicable here?
- L315: Just for clarification: this is comparing Qsim and Qobs at the outlet of the catchment, right? No other catchments upstream have flow data (at least that are used to calibrate the model)?
- L324: It’d be helpful to know what the benchmarks are that you’re considering here.
- L326: It might be helpful to let the reader know that you’ll talk about power law anomaly in the next section. It took me awhile to understand what L_A and EC_A were referring to.
- L330: It sounds like L_A and EC_A weren’t incorporated into model calibration, but rather were included after calibration to constrain the realizations. Was it not possible to incorporate these directly into the calibration algorithm as well?
- L393: It feels like it might be worthwhile to include some information (maybe a short paragraph) on the climate scenarios and how you ran these through DHSVM.
- Fig 1: When showing the stream network, it would be helpful to also identify where the dry reaches are. E.g., with grey lines or something similar.
- L485: Why not use a Mann Kendall test here and include the p-value? That seems like it would provide more statistical merit to the analysis than the sign of the Sen's slope.
- L491: This is a super interesting finding. One point that could be worthy to include in your discussion: It does seem like there’s quite a bit of scatter in the (simulated) L-Q plots shown in the studies from both Gao and Mahoney, which also demonstrate nonmonotonicity in L-Q… (https://doi.org/10.1016/j.jhydrol.2021.126522; https://doi.org/10.1016/j.jhydrol.2023.129422). Are these models also incorporating a degree of hysteresis (albeit perhaps not quite as explicitly as is being done here)?
- Fig 4. Same as fig. 1 – it’d be helpful to show the dry streams on these maps.
- L726: could just be worthwhile to note that EC isn’t always conservative and therefore may not universally be a good tracer.
Citation: https://doi.org/10.5194/egusphere-2025-6294-RC1 -
RC2: 'Comment on egusphere-2025-6294', Anonymous Referee #2, 27 Feb 2026
This manuscript presents a substantial and carefully executed investigation of stream-network dynamics and groundwater–surface water interactions under changing climatic conditions. The study combines distributed hydrological modelling with water-quality observations and analyses conducted across multiple temporal scales, representing a considerable modelling and analytical effort. The manuscript is relatively long, although generally well written, and addresses a timely and relevant topic, particularly in the context of ongoing transitions from snow- to rain-dominated hydrological regimes.
Overall, the analyses are thorough and the main findings appear physically plausible and scientifically meaningful. The study provides an extensive exploration of network dynamics and their climatic controls, and the effort invested in model implementation and evaluation is evident throughout the manuscript. The comments below are primarily intended to further clarify aspects of process interpretation and model representation and to help strengthen the Ms.
Major comments
A central contribution of the manuscript is the interpretation of the observed decoupling between flowing network length and discharge in terms of groundwater hysteresis. The modelling results clearly demonstrate the emergence of hysteretic L–Q behaviour. At the same time, the mechanisms described in the Results suggest that this behaviour may arise from differences in how distinct storage components translate into network activation versus discharge generation. For instance, portions of the drainage network may become active without contributing proportionally to outlet discharge, while rapid runoff components may increase discharge without producing equivalent network expansion. This suggests that groundwater hysteresis may operate in combination with activation and connectivity dynamics associated with faster flow pathways. This interpretation is somewhat less explicit in the title, abstract, and introduction, where groundwater hysteresis appears as the primary explanatory mechanism. Clarifying this causal chain would help strengthen the physical interpretation of the results and more clearly distinguish groundwater memory effects from activation–connectivity processes governing network expansion.
A second aspect concerns the use of power-law scaling as a reference framework for interpreting L–Q behaviour. The simulated relationships shown (e.g., Fig. 3) appear to display a tendency toward flattening at higher inputs, suggesting behaviour that may depart from ideal scale-free growth. Because the double-logarithmic representation can partially mask such curvature, presenting the relationship also in natural (non-logarithmic) space could help clarify the interpretation. A more cautious framing of the power law as one possible benchmark rather than an expected scaling relationship may therefore be beneficial. Exploring alternative saturating functional descriptions could further support interpretation of the simulated L–Q dynamics (see also Durighetto et al., 2025). Importantly, the main conclusions of the manuscript do not appear to rely critically on the assumption of power-law scaling.
The study represents an impressive modelling effort combining distributed simulations with hydrochemical information and multi-scale analyses. The emergence of L–Q hysteresis within the model appears closely related to the representation of surface–subsurface interactions and to the coupling between surface routing and subsurface storage components. While these processes are clearly important, their implementation can be difficult to visualize from the present description. Providing a schematic representation and/or a more explicit description of exchange mechanisms would help readers better understand how hysteresis emerges within the modelling framework. It may also be useful to clarify which aspects of the inferred mechanisms are directly supported by observations and which primarily emerge from internal model dynamics.
Finally, the manuscript addresses an important question linking stream-network dynamics to hydroclimatic regime shifts. Given the breadth of analyses presented, some streamlining of the manuscript and a clearer separation between model results, mechanistic interpretation, and broader climatic implications could further improve readability and help highlight the novel contribution related to snow-to-rain transitions.
In line with the above arguments, the Discussion (which is quite extensive and properly puts this paper in the context of the literatyure in general) could benefit from a somewhat clearer comparison with previous studies addressing specifically L–Q hysteresis. Several observational and process-based investigations have shown that hysteretic behaviour may emerge from differing activation and connectivity dynamics across events and catchments (e.g., Jensen et al., 2019; Bujak-Ozga et al., 2023; Zanetti et al., 2024, Hydrological Processes). Positioning the mechanism proposed here relative to these interpretations may further reinforce the contribution and clarify its novelty.
Minor / inline comments
Line 28: The manuscript refers to headwater streams; however, the study catchment (~30 km²) may not strictly fall within a typical headwater definition. Clarifying terminology or briefly explaining its relevance for the selected basin would improve consistency. Mentioning the broader global relevance of non-perennial streams could also help contextualize the study.
Line 62: The terms non-monotonic and hysteretic appear at times closely associated. As these concepts describe different behaviours, clarifying terminology and maintaining consistent usage throughout the manuscript would improve interpretability.
Line 71: The statement that groundwater level significantly controls flowing network length may depend on dominant runoff-generation mechanisms. A brief clarification of the conditions under which groundwater exerts primary control on L would be helpful.
Line 74 (see also Line 744): Functional L–Q scaling, hysteresis in the L–Q relationship, and groundwater–discharge hysteresis are not necessarily fully equivalent processes. Clarifying these distinctions throughout the manuscript would improve conceptual clarity.
Lines 118–119: Zanetti et al. (2024, Hydrological Processes) discuss hydrological processes that potentially generate hysteresis within the water-(im)balance framework and may provide useful context here.
Line 120: I think a hierarchical organization within the drainage network does not necessarily imply a unique relationship between flowing length and discharge, but only a one-to-one relation between total flowing length and spatial configuration of the active network.
Line 134: If the selected basin is intended as representative, the research question could potentially be framed more broadly in terms of mountain catchments.
Lines 178–210: The surface–subsurface coupling description would benefit from a schematic representation or clearer formulation (in the main paper) to facilitate visualization of model processes.
Section 2.1.5: I'm surprised by the fact that evaporation from channels might be important.. anyways this point better fits the discussion.
Model parametrization: Unless I am missing something, additional information regarding assumptions on riverbed composition, hyporheic exchange processes, interaction depth, hydraulic conductivity, and parameterization of channel–subsurface exchanges would improve transparency/reproducibility.
Line 359: The use of a power-law formulation is arbitrary: Fig. 3 suggests that alternative functional descriptions could better capture the simulated dynamics.
Line 415: Further clarification of the mechanisms responsible for simulated network disconnections would be informative.
Line 465: The statement that observed scaling reproduces a power-law relationship is not fully evident from the results; representation of L and Q in natural space may help clarify this point.
Line 467: Clarifying whether a maximum active network length is implicitly constrained by the modelling domain (see lines 259–264) would aid interpretation.
Figure 4: This figure is particularly informative and could benefit from expanded discussion and captioning. Clearer panel identification may also improve readability.
Section 3.3: Maybe i'm missing something here but clarifying the exact quantitative role/value of conductivity data within the analysis would strengthen interpretation.
Lines 745–748: The discussion might more explicitly acknowledge here mechanistic formulations that address hysteresis and non-stationary network dynamics
Citation: https://doi.org/10.5194/egusphere-2025-6294-RC2
Data sets
Data and Code for Stream Network Hysteresis Study Using DHSVM in Sagehen Creek Basin Elijah Boardman https://doi.org/10.5281/zenodo.17958145
Video supplement
Supplemental Video - Groundwater and Stream Network Animation Elijah Boardman https://mountainhydrology.com/sagehen_groundwater_stream_animation/
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 353 | 217 | 18 | 588 | 54 | 11 | 19 |
- HTML: 353
- PDF: 217
- XML: 18
- Total: 588
- Supplement: 54
- BibTeX: 11
- EndNote: 19
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
General comments
- Very good research on a general topic that spans between several fields of geosciences (hydrogeology, siliciclastic sedimentology and hydrology as a minimum). However, some minor issues need to be fixed.
- Snow hydrology in a mountain region is a growing topic due to the climate change.
Specific comments
Line 35. “other fluvial processes”. You need to specify all of them (erosion, sedimentation, landslide triggering or whatever you retain relevant).
Line 39. “Longer time scales”. Please, specify the order of magnitude for these scales.
Lines 40-41. “and hydrogeological properties(transmissivity)”. You should back-up this statement with references. See research on aquifer transmissivity and groundwater that converges:
- Dietrich, C. R., and Garry Neil Newsam 1989. A stability analysis of the geostatistical approach to aquifer transmissivity identification. Stochastic Hydrology and Hydraulics 3, 293-316.
- Agbotui, P.Y., Firouzbehi, F. and Medici, G., 2025. Review of effective porosity in sandstone aquifers: insights for representation of contaminant transport. Sustainability, 17(14), 6469.
Lines 134. Ok the research questions are clear. Have you been so explicit for the description of the general goal?
Line 140. You need inserting basic equations when describing routing modelling. The four below are not sufficient in the manuscript.
Line 360. Double information with same equation on line 49.
Figures and tables
Figure 1. I prefer the conceptual models below, and the graphs above.
Figure 2. The second option would be splitting the figure in two parts.
Figure 4. Months on horizontal axes difficult to read. Please, enlarge them.
Figure 4. Splitting the figure in two parts to fix the issue?
Figure 7. I suggest the conceptual models below, and the graphs above also for this figure.
Figure 9. Do you need an approximate spatial scale?