Towards observation-based gridded runoff estimates for Europe

Gudmundsson, L.; Seneviratne, S. I.

doi:https://doi.org/10.5194/hess-19-2859-2015

Articles | Volume 19, issue 6

https://doi.org/10.5194/hess-19-2859-2015

© Author(s) 2015. This work is distributed under
the Creative Commons Attribution 3.0 License.

https://doi.org/10.5194/hess-19-2859-2015

© Author(s) 2015. This work is distributed under
the Creative Commons Attribution 3.0 License.

Articles | Volume 19, issue 6

Research article

| Highlight paper

|

22 Jun 2015

Research article | Highlight paper |

| 22 Jun 2015

Towards observation-based gridded runoff estimates for Europe

L. Gudmundsson and S. I. Seneviratne

Download

Final revised paper (published on 22 Jun 2015)
Preprint (discussion started on 18 Nov 2014)

Interactive discussion

Status: closed

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

- Printer-friendly version

- Supplement

RC C5829: 'Towards observation based gridded runoff estimates for Europe', Balazs Fekete, 26 Dec 2014
- AC C6375: 'Response to Reviewer \#1', Lukas Gudmundsson, 05 Feb 2015
RC C5975: 'Review of ms hessd-2014-452 by L. Gudmundsson and S. I. Seneviratne', Anonymous Referee #2, 08 Jan 2015
- AC C6384: 'Response to Reviewer \#2', Lukas Gudmundsson, 05 Feb 2015
AC C6373: 'General Response to the Reviewers Comments', Lukas Gudmundsson, 05 Feb 2015

Peer-review completion

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

ED: Reconsider after major revisions (21 Feb 2015) by Vazken Andréassian

AR by Lukas Gudmundsson on behalf of the Authors (17 Mar 2015)

ED: Referee Nomination & Report Request started (22 Mar 2015) by Vazken Andréassian

RR by Anonymous Referee #3 (23 Mar 2015)

Suggestions for revision or reasons for rejection

Review of the paper entitled
Towards observation based gridded runoff estimates for Europe
Submitted to : HESS
By Lukas Gudmundsson and Sonia Seneviratne

I was not involved in the first round of revision, and I thus take this revised paper as a new one.

SYNTHESIS
1. In simple words, I would tell that this paper presents a method to upscale and then interpolate the observed streamflow from small catchments. A validation is proposed based on (i) cross validation, (ii) large catchments which were not part of the initial catchment dataset, (iii) independent evaporation measurements. A comparison is provided based on a combination of LSM data.

WHAT I LIKED IN THIS PAPER
2. This paper is scientifically sound (I appreciated the three steps validation step in particular). It is rather well written (although sometimes I feel that the authors use unneeded complex words for things which are after all very simple).

WHAT I DID NOT LIKE SO MUCH IN THIS PAPER
3. I always have problem with machine learning techniques which I like to classify as “fancy statistical methods” because they are black-box (sometimes only grey…) sometimes difficult to follow… and always difficult to INTERPRET PHYSICALLY. My preferred approach would have been to use a simple regression (Q=a*P^b*E^c… eventually with monthly varying parameters, because these parameters are easy to interpret) and THEN use the machine learning technique to interpolate the residual. It would probably have given exactly the same efficiency… but it would have been much easier to interpret. I have no objections to have a black-box approach to map residuals (because by definition, the residual is impossible to understand), I always consider it a pity to mix what is understandable with what is not.

RECOMMANDATIONS
4. You must do something about the negative RFM values in Fig. 11… you can’t just write that you know they exist!
5. Fig 11 : there is certainly a lot of information to gather from the differences in Fig 11. May I suggest mapping the differences in the Turc-Budyko non dimensional graph (i.e. x=P/E0 and y=Q/P, each catchment is represented by a point and the colour of the point would present the difference between the two models), this could help explain physically the differences.

CONCLUSION
Overall, I liked this paper and I recommend its publication. Here are my suggestions to improve its readability:
6. You should try to synthesize your approach using simpler words in the introduction (for example, what you do is a “spatial interpolation” but I did not find the word in your paper);
7. Line 122 : you should not include “LSM runoff” in the “validation data paragraph… it is just a comparison (just change : 2.2 Comparison data and then 2.3 Validation data
8. Line 324 : do something for those negative values.

Hide

RR by Balazs Fekete (07 Apr 2015)

ED: Reconsider after major revisions (07 Apr 2015) by Vazken Andréassian

AR by Lukas Gudmundsson on behalf of the Authors (05 May 2015) Author's response Manuscript

ED: Publish as is (20 May 2015) by Vazken Andréassian

AR by Lukas Gudmundsson on behalf of the Authors (21 May 2015)

Short summary

Water storages and fluxes on land are key variables in the Earth system. To provide context for local investigations and to understand phenomena that emerge at large spatial scales, information on continental freshwater dynamics is needed. This paper presents a methodology to estimate continental-scale runoff on a 0.5° spatial grid, which combines the advantages of in situ observations with the power of machine learning regression. The resulting runoff estimates compare well with observations.