A pre-calibration approach to select optimum inputs for hydrological models in data-scarce regions

Tarawneh, Esraa; Bridge, Jonathan; Macdonald, Neil

doi:https://doi.org/10.5194/hess-20-4391-2016

Articles | Volume 20, issue 10

https://doi.org/10.5194/hess-20-4391-2016

© Author(s) 2016. This work is distributed under
the Creative Commons Attribution 3.0 License.

https://doi.org/10.5194/hess-20-4391-2016

© Author(s) 2016. This work is distributed under
the Creative Commons Attribution 3.0 License.

Articles | Volume 20, issue 10

Research article

|

28 Oct 2016

Research article |

| 28 Oct 2016

A pre-calibration approach to select optimum inputs for hydrological models in data-scarce regions

Esraa Tarawneh, Jonathan Bridge, and Neil Macdonald

Download

Final revised paper (published on 28 Oct 2016)
Supplement to the final revised paper
Preprint (discussion started on 01 Jun 2016)

Interactive discussion

Status: closed

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

- Printer-friendly version

- Supplement

RC1: 'Interactive comments on ‘A pre-calibration approach to select optimum inputs for hydrological models in data-scarce regions’ By Tarawneh et al.', Anonymous Referee #1, 27 Jun 2016
RC2: 'RC', Anonymous Referee #2, 07 Jul 2016
- AC2: 'Authors' response to Anonymous Referee #2', Esraa Tarawneh, 22 Jul 2016
AC1: 'Authors' response to Anonymous Referee #1', Esraa Tarawneh, 22 Jul 2016
AC3: 'The Wala catchment system', Esraa Tarawneh, 02 Aug 2016

Peer-review completion

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

ED: Reconsider after major revisions (11 Aug 2016) by Elena Toth

The paper presents a comparison of the rainfall-runoff modelling results obtained with different input and watershed characterization, applied to a data-scarce region, well representative of many parts of the worlds, and therefore may certainly be of interest for a wide hydrologists’ audience.
The two referees have a positive opinion of the paper and I warmly invite the Authors to submit the revised version they are already working on, taking into account all the comments made by the Referees.
I will ask the opinion of Ref 2 on the revised version, while I believe that the Authors have properly addressed the points raised by Ref 1, but the use of monthly instead of daily data (see comments by Ref 1 and the Authors' reply) should be better clarified in the text (it is not explained in the revised version) and consider adding in the paper the results got at daily scale.
On the comparison of rainfall input: the authors themselves acknowledge in the discussion of the results, the use of a rainfall generator is questionable for filling time-series (and even more so in this case, where the two stations are far from the catchment and their data are not even included in the daily local set).
But since the performances are very poor also when excluding the only station where the weather generator is massively used, there are certainly problems in the ability of the remaining 3 raingages (where there almost no missing data, so the weather generator is not used) to represent the precipitation field over the catchment, probably due also to the lack of gauges in the eastern part of the study area; the authors should consider further comments on such issue (some comments are actually already included in the revised version).

Hide

AR by Esraa Tarawneh on behalf of the Authors (30 Aug 2016) Author's response Manuscript

ED: Referee Nomination & Report Request started (01 Sep 2016) by Elena Toth

RR by Anonymous Referee #2 (28 Sep 2016)

Suggestions for revision or reasons for rejection

Review of the “A pre-calibration approach to select optimum inputs for hydrological models in data-scarce regions”

Regarding the use of monthly data for model evaluation, rather than daily:
It is standard practice that large scale watersheds are evaluated based on monthly rather than daily data. This is mainly due to: (1) larger errors associated with the data at daily time step that are usually compromised when aggregating to monthly; (2) assumptions and simplifications in large scale models that are not designed to represent very detail small temporal and spatial scale processes.
As I mentioned previously (comment#14, below please find it), the Wala basin is located in semi-arid part of the world; and the short-term events (intense daily or sub-daily rainfall and floods) are major concern. While the model is evaluated at monthly, the daily (and sub-daily) events may be under estimated. The authors have not addressed or discussed this issue in their article. I strongly recommend to report this in the paper at least in the discussion part. I understand the main goal of the paper is to address the importance of data monitoring, etc., but even in that case the author can argue this issue to support their goal and objectives.
“my previous comment #14: Page 10, L12-16: (i) When looking at the top left graph in Fig. 9, the high performing statistics are usually representing low flows, while high flows are almost completely underestimated (or not predicted at all). How the authors will consider both low flows (important for drought) and high flows (major events and important for flooding and water saving) explain in their model evaluation and scenario selection? (ii) As indicated in several places in the text, the Wala basin is located in semi-arid area that is prone to intense (sub-daily) rainfall events. My concern is that how model evaluation at the monthly scale will ensure representation of locally important short-term events? (iii) As mentioned previously, the data are compared at downstream station that might be affected by operation of a dam.”

Page 5, Line 12-13: Repeated statement. Please remove one.
Page 9, Line 29-30: As already mentioned in my previous review, you cannot run the SWAT model on monthly basis. It operates on daily time step, and it just prints the outputs on monthly or yearly formats depending on the user’s need. Please rewrite the statement (all SWAT users are quite familiar with the model operation, may be no need to mention how you ran the model but refer to post processing of the data only).

Referee Report: PDF

Hide

ED: Publish subject to minor revisions (Editor review) (30 Sep 2016) by Elena Toth

AR by Esraa Tarawneh on behalf of the Authors (05 Oct 2016) Author's response Manuscript

ED: Publish as is (07 Oct 2016) by Elena Toth

Short summary

The study presents a rapid, robust and reproducible approach to select optimum inputs for hydrological models. We show significant variation in pre-calibrated model performance due to the choice of different combinations of land-use, soil and weather datasets, particularly in data-scarce regions most in need of computationally-efficient approaches to optimise reliability of models and support decision-making. Our case study is directly relevant to planners and decision-makers in Wala, Jordan.