Return Level Analysis of the Hanumante River Using Structured Expert Judgment: A Reconstruction of Historical Water Levels

Round 1

Reviewer 1 Report

Minor

1. The information of experts of Table 1. and Table A.2-4 don't match each other.
2. Can you check that the water level 2.4 m has real return periods of 2-years (in Figure 7.)  based on the results of the experts' survey?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The paper presents a method to reconstruct historical water levels at a cross section of Hanumante River in Nepal for the period 1990-2019, through the objective evaluation of assessments from a panel of selected ‘experts’ (chosen among citizens, water specialists, students). The authors apply a so called Structured Expert Judgment (SEJ) to systematically analyze and weight the opinions of the various experts and evaluate the related uncertainties.

The need to reconstruct missing data is essential because the evaluation of risk levels, anomalies, design parameters must be conducted on sufficiently long time series (at least several decades). This task is challenging, and sometimes almost impossible, in ungauged basins where data are scarce or not available at all. In these cases, especially in certain areas of the world, the engagement of population might represent a very valuable support. The method presented in this paper follows this approach, and although it may lead to highly inaccurate results (as the authors themselves say), it still can represent a useful contribution to improve the knowledge of flood risk of the area.

I think the topic is interesting and the paper is overall well written. However, in my opinion, especially from the hydrological point of view, there are conspicuous inaccuracies and lack of depth that should be properly addressed before considering the paper suitable for publication.

First of all, a return analysis for flood risk evaluation should be done on river discharge rather than water level. Water levels depend on cross section geometry, and it is not clear if the geometry of HM04 has remained reasonably constant through the years (actually the authors say at row 111 that ‘average river width decreased from six in 1964 to two meters in recent times’ , but it is not clear if this refers specifically to location HM04).So a historical analysis based only on water levels may not be meaningful to assess the probability of occurrence of a given event.  Changes in river cross section and corresponding specific stage-discharge relationships to convert water levels into discharge values should be evaluated. I understand this task is probably impossible to achieve for the specific site, so the authors should at least provide some evidence that the cross section of HM04 has not changed significantly in the period 1990-2019, or should limit their presentation to the reconstruction of historical water levels without assigning return periods.

Also, I think that the authors should validate, at least roughly, their result with precipitation data.They state at rows 533-538 their intent to do so, but ‘without any success of acquiring relevant data’. Actually historical records of precipitation at Kathmandu Airport (at least monthly, but probably also daily) are available; for example monthly data are included in the World Weather Records (WWR) and can be downloaded from the web. Furthermore, there are several global gridded datasets derived from remote sensing products, or merged satellite-gauge products, or climatological models. These data may not be optimal for small scale studies due to their coarse spatial resolution , but they still represent a valuable source of climatological information.Precipitation data could be shown for comparison or also used to apply simple rainfall-runoff models.

Broadly speaking, in my opinion the authors should be more convincing when explaining what is the added value of SEJ, and why they consider their method more adequate to retrieve missing water levels and flood risk evaluation with respect to other basically low-cost method (for example search for available historical photographic documentation, application of simple rainfall-runoff models, etc.). I think that some sentences about SEJ being ‘the best available option at hand’ (Row 506), or that ‘a comparable risk exists with measured data’ (row 451) sound naive, without a proper explanation and discussion.

Other specific remarks and/or questions:

I confess I find quite unrealistic that experts could remember monthly water levels of 20 years ago, although I guess there might be cultural aspects -of which I am unaware- that may help in this task (for example, the habit of keeping written records for personal use, relationship with social events, etc.). Since the data finally used are the annual maxima, why haven’t the authors chosen to formulate the questions in such terms? (i.e. ‘which was the maximum water level in 1990?’ etc). This would have reduced the number of questions and maybe also the uncertainty of the estimation; is there any reason why it is more advisable to ask questions separately for each month?

Row 133: Is it conventional to inform the experts about the target/non target questions or shouldn’t they be ‘blind’? Also, regarding the calibration questions, obviously they all refer to recent years (since no older data are available): do the authors think this might lead to a biased evaluation of the experts’ scores?

Rows 277, 284: ‘GEVfit’ should be 'gevfit' and ‘GEVinv’ should be ‘gevinv’: Matlab is case-sensitive

Figure 5: Confidence interval for 1992 is very narrow. Which is the explanation for this?

Row 540, ‘further research could be done on the expected damages due to floods or about the possibilities of an early-warning-system’: Given the philosophy of the research presnetd in this paper, community based early warning systems (CBEWS), for which a wide literature exists, should be mentioned.

Citation (13) is incomplete.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

I think the authors have addressed most of my comments, clarifying several concepts in the text.

Some aspects could still be improved, additional analysis would be beneficial for the study and hopefully will be tackled in future research, but I think the revised version of the manuscript is suitable for publication in Water.