|Effectiveness of distributed temperature measurements for early detection of piping in river embankments |
Authors: Sylvia Bersan et al
I reviewed the original submitted manuscript and I am happy to see that it improved considerably. The authors made a very serious and thorough revision for which I thank them. The paper is now, in my opinion, much better to follow and more coherent. It is based on data from a unique experiment (IJkdijk) and numerical model results. The paper has a geotechnical character, but the analysis and modelling can be of interest to readers of HESS.
There is one aspect that I would suggest to explain more.
I am somewhat worried by the explanation concerning the calibration of the temperature measurements in the field and the possible effect of changing strain during the experiment. The authors reply: “The measurements were performed by a company different form the company that designed and led the test. The calibration procedure was not specified. It was guaranteed that temperature changes were measured with an accuracy of 0.1°C.”. “The DTS system exploited the Raman effect, which is not sensitive to strain”.
Unfortunately, this is not totally true. For example, Hausner et al, 2011 and Hausner and Kobs 2016 elaborate in detail on this.. I advise the authors to study these papers. Standard internal calibration under static conditions may be insufficient for the accuracy (0,1degree) also required by the authors. Also the papers detail on the effect of strain on the (apparent) temperatures. I stand my case, this cannot be neglected by the authors. I reckon the problem in case no (field) calibration data or raw backscatter data are available to the authors. I also see the quality of the data seems nice. However, at least the aspect of (dynamic) strain, accuracy, calibration and thus possibly changes on the (apparent) temperature results should be mentioned and discussed (in a quantitative manner) in the paper.
Hausner, M.B.; Suárez, F.; Glander, K.E.; Giesen, N.v.d.; Selker, J.S.; Tyler, S.W. Calibrating single-ended fiber-optic Raman spectra distributed temperature sensing data. Sensors 2011, 11, 10859–10879.
Mark B. Hausner and Scott Kobs, “Identifying and Correcting Step Losses in Single-Ended Fiber-Optic Distributed Temperature Sensing Data,” Journal of Sensors, vol. 2016, Article ID 7073619, 10 pages, 2016. doi:10.1155/2016/7073619
Minor technical issues:
- Please use hydraulic load in the first 6 sentences of section 3.1 (instead of only load)
- P5L11: specify: “the theory”.
- P7L8: specific discharge [m2/s] is not the same as darcian velocity [m/s].
- P11L3: preferential flow could turn into a pipe with increasing risk, but to my understanding also to drainage and lowering of pore water pressure reducing the risk.
- P11L20: “Classic interpretive” . I am not sure if you should frame it this way. Also your research is based on ‘accidental’ temperature differences which can be detected using DTS. I would refrain from calling this ‘classic’ implying the approach in this paper is totally different, whereas it is mainly applied in a different soil/environment/condition.
- P12L6: Can you already call it a pipe? Or ‘just’ a preferential flow path?
- P12L13: “soil temperature”. I think you mean the local subsurface temperature (soil and the water at that point)
- P14L23: that that
- P15L1-2: “conductivity” here is meant: saturated hydraulic conductivity (not thermal). Please specify
- P16L26: “Two parameters”. Please specify here
- P17L7-8: “A sensor located at downstream side”. Maybe it is worthwhile to add if 1 fibre is enough or that one needs a kind of mesh of fibres to not-miss the temperature anomaly. This come back to the discussion on the best location of DTS measurements in embankments.