A simple topography-driven and calibration-free runoff generation module

Gao, Hongkai; Birkel, Christian; Hrachowitz, Markus; Tetzlaff, Doerthe; Soulsby, Chris; Savenije, Hubert H. G.

doi:https://doi.org/10.5194/hess-23-787-2019

Articles | Volume 23, issue 2

https://doi.org/10.5194/hess-23-787-2019

© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/hess-23-787-2019

© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 23, issue 2

Research article

| Highlight paper

|

13 Feb 2019

Research article | Highlight paper |

| 13 Feb 2019

A simple topography-driven and calibration-free runoff generation module

Hongkai Gao, Christian Birkel, Markus Hrachowitz, Doerthe Tetzlaff, Chris Soulsby, and Hubert H. G. Savenije

Download

Final revised paper (published on 13 Feb 2019)
Supplement to the final revised paper
Preprint (discussion started on 28 Mar 2018)
Supplement to the preprint

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'Comments on "A simple topography-driven and calibration-free runoff generation module"', Anonymous Referee #1, 15 Apr 2018
- AC1: 'the point by point replies', Hongkai Gao, 07 Jun 2018
RC2: 'This study proposed the HAND-based storage capacity curve (HSC) for runoff generation parameterization in hydrological models', Anonymous Referee #2, 28 May 2018
- AC2: 'the point by point replies', Hongkai Gao, 07 Jun 2018
RC3: 'review', Anonymous Referee #3, 30 May 2018
- AC3: 'the point by point replies', Hongkai Gao, 07 Jun 2018

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

ED: Reconsider after major revisions (further review by editor and referees) (28 Jun 2018) by Fuqiang Tian

AR by Hongkai Gao on behalf of the Authors (02 Aug 2018) Author's response Manuscript

ED: Referee Nomination & Report Request started (21 Aug 2018) by Fuqiang Tian

RR by Anonymous Referee #3 (21 Aug 2018)

Suggestions for revision or reasons for rejection

I would like to thank the authors for their revision, particularly improving discussion section. For some of my comments, however, I do not see a clear response:
1) I’m still not convinced by the use of BB basin to support the conclusion that the new model (HSC) outperforms the TOPMODEL (“We found that the HSC performed better in reproducing the spatio- temporal pattern of the observed saturated areas in the BB compared to TOPMODEL”). The Figures 6 and 8 clearly indicate that both models are significantly overestimating what is considered as ground truth here (observed pattern of saturated area). I do not understand well the argument that the values are not directly comparable. If so, why to compare them? This part is not well linked with the validation of the new approach in its current form.
2) In some places of section 5, it will be interesting to see more specific interpretation of the results. E.g. for statement “Figure 11c interestingly shows that in some catchments, there is almost no threshold…” it will interesting to see some more specific generalisation for which catchments it applies. Or “…where the HSC model performed better are mostly located in the Great Plains, with modest sloping (4.0 degree)…” Does it apply/relate to all catchments with slope 4degree or even less (“HSC outperformed catchments have flat terrain (2.3 degree) with moderate averaged HAND value (26m)”? Some more evaluations (or more specific formulations) allowing some more specific generalisation of results will be interesting here. Where one can expect the new model is better/worser than TOPMODEL/HBV in other regions – outside US?
3) Figure 6 legend is still confusing for me. Would it be possible to make all three columns/methods of maps just with binary colours (is/is not saturated)?
4) Please add the new references to the list.

Hide

RR by Anonymous Referee #1 (04 Sep 2018)

RR by Anonymous Referee #2 (25 Sep 2018)

Suggestions for revision or reasons for rejection

I am appreciated that the authors add more explanations and discussions to improve the manuscript. However, the benefits from the application of the new runoff generation module are still not clear to me. The authors discussed the model uncertainty in the introduction section, while the results have not addressed this issue in their manuscript. The ability of the proposed module to improve the conceptualization of real catchment behaviors in the hydrological model is not convincing. Some of other concerns are listed as follow.
1. The authors defined ‘calibration free’ as one of the major sell points of their work. However, the benefits from the reduced calibration work need to be further interpreted in the results. After the integration of the proposed runoff generation module, the hydrological model still needs calibration. The computation cost caused by the preparation of the HAND curves is even higher than the computation cost reduced by the smaller parameter space. For the practical application of hydrological modeling, running automatic calibration without any pre-calculation is more preferable, unless the authors can provide poofs for that the new runoff generation module is more close to the realism in the catchment. Figures 6-9 compared the simulated and observed saturated areas by various modules. However, I would say the HSC module has not shown higher performance than the Topmodel in Figures 8-9. The Topmodel shows higher performance at low saturated areas (Fig. 9a). Figure 6 has not evaluated the distribution of saturated areas produced by the HBV model, which can be also set up at grid scale. Figure 7 does not make any sense, considering observation is missing.
2. The comparable or better performance produced by the model coupled with the proposed runoff generation module may be caused by either the application of the new runoff generation module or the calibration run with less parameters. The automatic calibration algorithm could produce higher performance when the calibration parameter space is reduced. This need to be further analyzed.
3. The motivation of this sturdy is not clear enough. With the proposed module, more calculation work is needed and produced the same or bit better model performance. The model uncertainty has not been analyzed, and the transferability of the runoff generation module has not been investigated in this study.
In general, limitations in the current work certainly prevented more scientific contributions from this study. The authors should pay more efforts to make their motivation and results more convincing.

Hide

ED: Publish subject to revisions (further review by editor and referees) (13 Oct 2018) by Fuqiang Tian

AR by Anna Wenzel on behalf of the Authors (28 Nov 2018) Author's response Manuscript

ED: Referee Nomination & Report Request started (08 Dec 2018) by Fuqiang Tian

RR by Anonymous Referee #3 (09 Dec 2018)

RR by Anonymous Referee #1 (18 Dec 2018)

ED: Publish subject to minor revisions (review by editor) (10 Jan 2019) by Fuqiang Tian

AR by Hongkai Gao on behalf of the Authors (18 Jan 2019) Author's response Manuscript

ED: Publish as is (24 Jan 2019) by Fuqiang Tian

Short summary

Supported by large-sample ecological observations, a novel, simple and topography-driven runoff generation module (HSC-MCT) was created. The HSC-MCT is calibration-free, and therefore it can be used to predict in ungauged basins, and has great potential to be generalized at the global scale. Also, it allows us to reproduce the variation of saturation areas, which has great potential to be used for broader hydrological, ecological, climatological, and biogeochemical studies.