Simple and approximate estimations of future precipitation return values

Benestad, Rasmus E.; Parding, Kajsa M.; Mezghani, Abdelkader; Dyrrdal, Anita V.

doi:https://doi.org/10.5194/nhess-17-993-2017

Articles | Volume 17, issue 7

https://doi.org/10.5194/nhess-17-993-2017

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

Special issue:

Risk and uncertainty estimation in natural hazards

https://doi.org/10.5194/nhess-17-993-2017

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

Articles | Volume 17, issue 7

Research article

|

03 Jul 2017

Research article |

| 03 Jul 2017

Simple and approximate estimations of future precipitation return values

Rasmus E. Benestad, Kajsa M. Parding, Abdelkader Mezghani, and Anita V. Dyrrdal

Download

Final revised paper (published on 03 Jul 2017)
Supplement to the final revised paper
Preprint (discussion started on 26 Jul 2016)
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: 'Review nhess-2016-229', Anonymous Referee #1, 13 Sep 2016
- AC1: 'Response to RC1: 'Review nhess-2016-229'', Rasmus Benestad, 17 Oct 2016
RC2: 'Interesting approach, but further discussion advisable', Reik Donner, 24 Sep 2016
- AC2: 'REsponse to RC2: 'Interesting approach, but further discussion advisable', Rasmus Benestad, 17 Oct 2016

Peer-review completion

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

ED: Reconsider after major revisions (further review by Editor and Referees) (04 Nov 2016) by Thorsten Wagener

AR by Rasmus Benestad on behalf of the Authors (16 Dec 2016) Author's response Manuscript

ED: Referee Nomination & Report Request started (30 Jan 2017) by Thorsten Wagener

RR by Reik Donner (26 Feb 2017)

Suggestions for revision or reasons for rejection

The authors have clearly improved the presentation of their results as compared to the original discussion paper. I still believe that their approach is interesting and possibly relevant and as such deserves publication in NHESS. However, prior to final acceptance, I feel that a couple of points need to be finally clarified.

In the response letter (reply to referee 1), the authors agree that “it is a bit misleading to call it [the obtained type of return-level estimate] an upper bound” and state that they use the term potential sensitivity instead. However, the terms “upper bound” and “upper limit” still occur at various places, particularly the abstract (l.9), method description (ll.90,92), conclusions (ll.285,290) and supplementary material (l.79). While I understand that it might be unavoidable to use such terms in some cases, it is still not fully clear from the reasoning presented in the manuscript why this “potential sensitivity” (of precipitation to temperature changes) is actually an upper bound for the expected precipitation changes and might not be further magnified by other (temperature-unrelated) effects. Clarifying this in the manuscript (with probably just a few more words) still appears necessary.

The order or figures in the Supplementary Material is not fully clear to me. If this order shall represent the order of appearance in the main manuscript (as I suppose) the correct order should be SM2, SM3, SM1, SM4, etc. Moreover, if I am not fully mistaken, the R scripts used for obtaining the presented results had been provided in the Supplementary Material of the discussion paper, but not in that of the final paper; thus, the corresponding statement and referencing in ll.100-101 should be corrected.

As has also been criticized in the discussion paper, the variable mu is used in different ways, i.e., for the annual wet-day mean precipitation (l.142) and the observed monthly mean (l.150), which is quite confusing. I suppose that the statistical exponential model of the wet-day mean precipitation as considered by the authors has firstly been developed on an annual basis, but is later being used separately for each calendar month. If this is correct, it would be helpful if the authors could state this explicitly within Section 2.2.

In ll.144-145, the authors state that they assume that “factors other than temperature that are affecting wet-day precipitation are stochastic and stationary”. However, the validity of the stationarity assumption seems to be tested only for the wet-day frequency (Section 1 of the Supporting Material). A few words on this aspect at the mentioned position in the text would be helpful.

Several references appear incomplete, especially Benestad et al. 2012a (pages missing, also in SM), Berg et al. 2013 (volume and pages missing, typo in journal name), Takayabu et al. (volume and pages missing) and Benestad 2008 (only SM – what is this reference?).

In general, the validity of the exponential approximation of the PDF of the wet-day precipitation might have been shown in the existing literature, but the quality of this approximation should still be discussed briefly in the manuscript or at least in the SM. Specifically, I understand the authors’ argument that they consider only moderate extremes (1 in 20 years, 95% quantiles, cf. p.2 of the response letter to the original reviews) – but then, they should state this explicitly in the text and emphasize that the approach is likely to perform less well if even rarer extremes are considered (say, 50 or 100-year return levels).

In connection with Fig. SM9, it would be good if the authors could highlight the stations with statistically significant trends (e.g., by a black circumference of the filled circles). In the caption of Fig. SM8, it is mentioned that there are regions with significant trends; these are, however, hard to assess in Fig. SM9. It should also be briefly stated how significance should be assessed here (i.e. if the statistical independence assumption of a classical t-test would hold or if there are sufficiently strong serial correlations in the historical records that would call for more complex testing approaches like block bootstrapping).

In relation with Fig. SM10, please add a very short explicit statement on the selection criteria for the chosen stations. Was it just time series length and data quality? This is just to clarify that they might not be any selection bias.

Technical comments:
- Line 24: MunichRe is a reinsurance, not an insurance company.
- Line 41: “demands have limited”
- Data and Methods: Since the introductory paragraphs of this section set the stage for the details presented in the following subsections, use of past tense is rather unusual here. Please consider using present tense here instead.
- Line 122: remove “in the SM”
- Line 255: Table 4 should be Table 2
- Line 281: better start the conclusions in present perfect tense, i.e., “we have proposed…”
- SM, line 1: “additional analyses that address some…”
- SM, line 14: “influence”
- SM, line 20: “not too sensitive”
- SM, line 26: “increase over southern Norway”
- SM, line 28: “typically by the order of 0.1 mm/day…”
- SM, line 41: “consistent with a near-constant…”
- SM, line 50: “suggest the highest”
- SM, line 57: it is not clear what “its density” refers to in this sentence
- SM, line 59: “wet-day mean precipitation than temperature…”
- SM, line 90: “with predominantly orographic”
- SM, line 102: “South America”
- SM, line 125: “is often comparable…” – if this is really often the case, one or two corresponding references would be reasonable
- Fig. SM1, caption: please add a statement that the results refer to a single selected station only
- Fig. SM2, caption: please use \ln to suppress italics in the presentation of the logarithm functions
- Fig. SM4, caption: “the relative change in comparison to…”
- Fig. SM5, caption: “between the seasonal cycles in…”
- Fig. SM6, caption: “long-term linear trends”
- Fig. SM10, caption: “The inset shows…”

Hide

RR by Anonymous Referee #3 (24 Mar 2017)

ED: Publish subject to minor revisions (further review by Editor) (24 Mar 2017) by Thorsten Wagener

AR by Rasmus Benestad on behalf of the Authors (05 Apr 2017) Author's response Manuscript

ED: Publish as is (28 May 2017) by Thorsten Wagener

AR by Rasmus Benestad on behalf of the Authors (29 May 2017) Manuscript

Download

Article (739 KB)
Full-text XML

Short summary

We propose a strategy for quantifying the maximum effect a temperature change has on heavy precipitation amounts, making use of the limited available sources of information: laws of physics, seasonal variations, mathematical estimation of probability, and s large number of climate model results. An upper bound is estimated rather than the most likely value.