Geostatistical upscaling of rain gauge data to support uncertainty analysis of lumped urban hydrological models

Muthusamy, Manoranjan; Schellart, Alma; Tait, Simon; Heuvelink, Gerard B. M.

doi:https://doi.org/10.5194/hess-21-1077-2017

Articles | Volume 21, issue 2

https://doi.org/10.5194/hess-21-1077-2017

© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 3.0 License.

Special issue:

Rainfall and urban hydrology

https://doi.org/10.5194/hess-21-1077-2017

© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 3.0 License.

Articles | Volume 21, issue 2

Research article

|

20 Feb 2017

Research article |

| 20 Feb 2017

Geostatistical upscaling of rain gauge data to support uncertainty analysis of lumped urban hydrological models

Manoranjan Muthusamy, Alma Schellart, Simon Tait, and Gerard B. M. Heuvelink

Download

Final revised paper (published on 20 Feb 2017)
Supplement to the final revised paper
Preprint (discussion started on 20 Jun 2016)

Interactive discussion

Status: closed

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

- Printer-friendly version

- Supplement

RC1: 'Review', Anonymous Referee #1, 05 Jul 2016
- AC1: 'Reply to Interactive Comment of anonymous referee #1', Manoranjan Muthusamy, 27 Sep 2016
RC2: 'Review I', Anonymous Referee #2, 15 Jul 2016
- AC2: 'Reply to Interactive Comment of anonymous referee #2', Manoranjan Muthusamy, 27 Sep 2016
SC1: 'Comment to “ Geostatistical upscaling of rain gauge data to support uncertainty analysis of lumped urban hydrological models”', Hannes Müller, 12 Aug 2016
- AC3: 'Reply to Interactive Comment of Hannes Müller', Manoranjan Muthusamy, 27 Sep 2016

Peer-review completion

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

ED: Reconsider after major revisions (11 Oct 2016) by Peter Molnar

Dear Authors,

I was alerted by HESS to the fact that my summary assessment of your paper has not reached you and you are waiting for next steps. I apologise for that, in my books this was done but I have apparently not correctly executed this task.

As you have seen both reviewers submitted detailed and partly critical but in my opinion constructively critical reviews of your paper, and you have indicated in the online open discussion how you will change the manuscript. Both reviewers asked for major revisions and will be reviewing again your revised manuscript.

I would like to invite you to submit your revised manuscript paying close attention to the main issues raised by both reviewers. In my opinon these were:

(1) Focus on novelty in your work (and remove trivial statements). In your respons to Referee 1 you have provided a good summary of what is new in the paper. Please make sure this reflects in a similar systematic list in the revised manuscript. Also both Referees pointed to the need for clearer explanations. This doesn't need to make the paper longer.

(2) What is the uncertainty in the results given the short dataset. Indeed you explain that a longer period was used to build the semivariograms, you may show the results of those tests in Supplementary Materials for instance. There are also other new results/figures that you include in the response to the Refferee which may find itself in Supplementary Materials and/or the revised manuscript.

(3) Influence of methdology on results, e.g. the assumption of independence and classes for pooling, etc. Please in your revision explain well the rationale behind the methodological steps you chose and how they may be affecting the results. Both referees raise some very important questions which many readers will have and they should be dealt with.

I am looking forward to your revision.

Peter Molnar

Hide

AR by Manoranjan Muthusamy on behalf of the Authors (04 Dec 2016) Author's response Manuscript

ED: Referee Nomination & Report Request started (17 Dec 2016) by Peter Molnar

RR by Anonymous Referee #2 (20 Dec 2016)

Suggestions for revision or reasons for rejection

Dear authors,
I have enjoyed reading this version of the paper, the text is much clearer and the method and the discussion related to the use of it are well explained and interesting.
I have some minor comments/suggestion for you to consider prior for the publication of the paper:

Introduction section (and also in the conclusion)
1. “Despite recent advances in radar technologies rain gauge measurements are still considered to be the most accurate way of measuring rainfall, especially at short temporal averaging intervals (< 1 hour), which are of most interest in urban hydrology studies (Seo and Krajewski, 2010; Villarini et al., 2008). “ The references are at the wrong location in the sentence. Should be:
"Despite recent advances in radar technologies rain gauge measurements are still considered to be the most accurate way of measuring rainfall, especially at short temporal averaging intervals (< 1 hour, e.g. Seo and Krajewski, 2010; Villarini et al., 2008), which are of most interest in urban hydrology studies (REF)".
Where REF could be the paper by Ochoa-Rodriguez et al. (2015) for example. Seo and Krajewsky and Villarini did not discuss urban applications in their papers.

Section 2.2
2. “Figure 3 shows time series of daily rainfall averaged over the network for 2012 and 2103.” – should be 2013.

Section 3.1
3. “For a catchment of this size (400 × 200 m2) it is very unlikely to have a response time of more than 30 min.” – I would claim that even 30 min is too long for a response time for an urban catchment in this scale. Probably rainfall data of <10 min are required, especially for rainfall intensity larger than 10 mm/h. See for example: Ochoa-Rodriguez et al. (2015), which you cite, and:
Schilling, W. (1991). "Rainfall data for urban hydrology: what do we need?" Atmospheric Research, 27(1-3), 5-21.
Zhu, H.-j., and Schilling, W. (1996). "Simulation errors due to insufficient temporal rainfall resolution -- Annual combined sewer overflow." Atmospheric Research, 42(1-4), 19.

Section 3.5
4. “is known as block kriging” – please add a reference to a paper or a textbook explaining what block kriging is.

Section 3.6
5. back-transformation and linear extrapolation. Let assume that at a given time t the highest rain intensity was 10 mm/h at a given gauge and that the highest rain intensity ever observed was 20 mm/h (somewhere in the domain). Can it be that the highest stochastically simulated value at time t will exceed 20 mm/h? In other words – using your method what will be the highest rainfall intensity that can be simulated? Is it limited by the observed data? Is there any other physical limitation?

Section 4.1
6. I strongly encourage the authors to have a look at the paper by Peleg at el. (2013) which is very relevant to your discussion. They are discussing in details the changes of rainfall spatial correlation with aggregation time and the large errors of weather radar for the <5 min time scales (relevant to comment 1 above).
Peleg, N., Ben-Asher, M., and Morin, E. (2013): Radar subpixel-scale rainfall variability and uncertainty: lessons learned from observations of a dense rain-gauge network, Hydrol. Earth Syst. Sci., 17, 2195-2208.

General
7. It would be nice to have some examples of the stochastic fields that you are generating. For example, showing one time step with the gauges marked (and labels indicating the observed rainfall intensity) and the simulated rainfall intensity around the field.

Hide

RR by Anonymous Referee #1 (04 Jan 2017)

Suggestions for revision or reasons for rejection

General comments
The intention of the paper has already been stated in the previous reviews of its first version. As it has not been changed, the focus of this review is on the improvements based on the former critics.
The structure of the paper has not been changed too much, but with the changes the paper focuses more on the methodology which supports the understanding. Furthermore, the single steps of the methodology are explained in more detail such that they could be reconstructed easily.
A major critic was the small sample size, which of course has not changed, but their impact is comprehensively analyzed. The results indicate that the influence seems to be negligible for the target of the paper.
The differences between tipping bucket errors as well as spatial uncertainty is also differentiated, explained and their influences on the variograms are illustrated.
Overall, the paper has been highly improved within the review process and could be published after the following comments have been considered.

Specific comments
Remark: [Page: Line]
[8: 5]: “spatial aggregation” This term simply means a spatial average, does it?

Technical corrections
Some comments on grammar, spelling or style. Although I did not focus on these aspects, I could find several errors and strange/complicated sentences. Here are a few examples:
[12: 28]: “The reason …” Long and complicated sentence
[14: 29]: “As a result …” A comma after “that” would highly help to understand the structure of the sentence
[15: 2] “…they reduced…” should be present tense
[15: 6] “… averaging interval are…” should be plural
[15:26] “… catchments, in similar climates…“ I can’t see a reason for the comma here
[15: 27] ”This estimate of uncertainty…” three times the same construction in two sentences
Consequently, I would recommend proofreading from a native speaker.

Hide

ED: Publish subject to technical corrections (13 Jan 2017) by Peter Molnar

AR by Manoranjan Muthusamy on behalf of the Authors (27 Jan 2017) Author's response Manuscript

Short summary

In this study we develop a method to estimate the spatially averaged rainfall intensity together with associated level of uncertainty using geostatistical upscaling. Rainfall data collected from a cluster of eight paired rain gauges in a small urban catchment are used in this study. Results show that the prediction uncertainty comes mainly from two sources: spatial variability of rainfall and measurement error. Results from this study can be used for uncertainty analyses of hydrologic modelling.