Toward a general calibration of the Swiss plate geophone system for fractional bedload transport
- 1Swiss Federal Research Institute WSL, Birmensdorf, 8903, Switzerland
- 2Deptartment of Environmental System Sciences, ETH Zürich, Zürich, 8092, Switzerland
- 3Institute of Earth Surface Dynamics (IDYST), University of Lausanne, Lausanne, 1015, Switzerland
- 4Deptartment of Earth and Planetary Science, University of California, Berkeley, 94720, USA
- 1Swiss Federal Research Institute WSL, Birmensdorf, 8903, Switzerland
- 2Deptartment of Environmental System Sciences, ETH Zürich, Zürich, 8092, Switzerland
- 3Institute of Earth Surface Dynamics (IDYST), University of Lausanne, Lausanne, 1015, Switzerland
- 4Deptartment of Earth and Planetary Science, University of California, Berkeley, 94720, USA
Abstract. Substantial uncertainties in bedload transport predictions in steep streams have triggered intensive efforts to develop surrogate monitoring technologies. One such system, the Swiss plate geophone (SPG), has been deployed and calibrated in numerous steep water courses, mainly in the Alps. Calibration relationships linking the signal recorded by the SPG system to the transported bedload can vary substantially between different monitoring stations, likely due to site-specific factors such as the flow velocity and the bed roughness. Furthermore, recent controlled experiments have shown that site-specific calibration relationships can be biased by elastic waves resulting from impacts occurring outside the plate boundaries. Motivated by these findings, here we present a hybrid calibration procedure derived from flume experiments and an extensive dataset of 308 calibration measurements from four different field monitoring stations. Our main goal is to investigate the feasibility of a general, site-independent calibration procedure for inferring fractional bedload transport from the SPG signal. First, we use flume experiments to show that sediment size classes can be distinguished more accurately using a combination of vibrational frequency and amplitude information than by using amplitude information alone. Second, we apply this amplitude-frequency method to field measurements to derive general calibration coefficients for ten different grain-size fractions. The amplitude-frequency method results in more homogeneous signal responses across all sites and significantly improves the accuracy of fractional sediment flux and grain-size estimates. We attribute the remaining site-to-site discrepancies to large differences in flow velocity, and discuss further factors that may influence the accuracy of these bedload estimates.
- Preprint
(2927 KB) -
Supplement
(2183 KB) - BibTeX
- EndNote
Tobias Nicollier et al.
Status: final response (author comments only)
-
RC1: 'Comment on esurf-2022-7', Anonymous Referee #1, 11 Mar 2022
The comment was uploaded in the form of a supplement: https://esurf.copernicus.org/preprints/esurf-2022-7/esurf-2022-7-RC1-supplement.pdf
-
RC2: 'Comment on esurf-2022-7', Dan Cadol, 16 Mar 2022
This paper presents a valuable synthesis of efforts to clean SPG signals and make comparisons across field sites. While complex, I found the ‘apparent’ impact cleaning method to be clear in the end. But key to my understanding was Figure 7. I was a bit lost regarding the thresholds until I came to that valuable figure. Even after cleaning, the bedload flux prediction errors up to 5-fold suggest that there is still work to be done. But this is a major step forward. And it's difficult to know how much of the discrepancy is due to direct bedload measurment uncertainty or the SPG signal features and processing.
My main question/suggestion is also related to Figure 7. Using fixed thresholds is reasonable due to its simplicity. But Fig 7b suggests that thresholds that are functions of both MaxAmp and MaxAmp/f would exclude apparent pulses and retain real pulses more reliably, without the step-like effect of the fixed value thresholds mentioned by the authors. Given the computational and field effort required to obtain the MaxAmp and f_centroid data, this extra bit of analysis seems trivial by comparison. But perhaps the need to automate the signal reduction makes such an approach untenable? Or the exact slope and intercept of the dividing lines in Fig 7b that I am suggesting varies from site to site? A brief exploration of the differences in the MaxAmp vs. MaxAmp/f plots for different sites and flume setups, or just a comment on the differences, might help.
Another possible advantage of sloped thresholds is that you could make them non-overlapping, eliminating the double counting of packets. I wasn’t entirely convinced by the statement that double counting impacts is a non-issue.
In Fig 7c&d, perhaps instead of (or in addition to) just showing the number of packets in the unfiltered and filtered data sets, you could show the fraction remaining after filtering. (ps- The dotted line in Fig 7d disappears for classes 7 & 8. Is this an plotting error, or a meaningful change?)
A few other small comments:
Fig 8b. There are fewer pulses/kg of the two smallest particle classes relative to the third smallest. Is this decline due to saltation? It’s an interesting result, which was masked by the counting of apparent packets in the amplitude-only thresholding method. I would appreciate some thoughts about it in the discussion.
Line 545: Why are there fewer impacts/kg when the particle (or flow) velocity is higher? I would think greater particle velocity would produce more readable impacts, and thus more impacts/kg? I think the text is suggesting that it’s because of more saltation, and thus skipping over the plate. Is this correct? Just a little more clarification of your hypotheses for this feature in the data would be appreciated.
Line 590: It’s good to make clear that uncertainty in the direct measurements used for calibration is very real, and that this may contribute to weaknesses or biases in the predictive power of the modeled estimate. You do mention this, I just think it can be easily forgotten in general, and perhaps merits emphasis.
-
CC1: 'Comment on esurf-2022-7', Roger Kuhnle, 08 Apr 2022
Review of “Toward a general calibration of the Swiss plate geophone system for fractional bedload transport”, by T. Nicollier, G. Antoniazza, L. Ammann, D. Rickenmann, J. W. Kirchner
The manuscript reports on a new method to calibrate the Swiss plate geophone (SPG) which uses a combination of data collected in a laboratory flume and at four different field sites. The SPG has been shown in previous studies to be an excellent indirect method to measure the rate and size of bed load transport in gravel bedded streams and rivers. This study further develops the science of turning impact data into quantitative data of the transport of bedload. The study of bed load transport using impact plates at several field sites has indicated that a general calibration of the SPG has been difficult to develop for a variety of reasons. This study uses a combination of amplitude and frequency to calibrate the SPG in the quest for a general calibration relation to turn impact data into quantitative values of mass and grain size for gravel bed load transport. This manuscript contains much valuable information and should be published, however, suggestions for improvement of the presentation are given below.
Specific comments:
- Lines 246-251: In these lines it is related how all packets were filtered using equation (3) and packets which do not meet this criterion are ignored in further analysis. It is clear to me how and why this was done, however, in Figure 7 there appears to be substantial overlap between real (blue) and apparent (red) peaks measured in the flume experiments. How many real peaks were rejected using this criterion in the flume experiment data? Also, could the authors estimate how many real peaks were rejected from the 4 field data sets considered in the study? I believe text should be added to the manuscript discussing this issue.
- Lines 239-240, and 257-260: In these sentences the lower and upper thresholds for the amplitude-frequency method are described. Is it correct that the lower threshold (V) was based on the minimum grain size of the size fraction and the upper threshold (V Hz) was based on the maximum grain size of the size fraction being considered? The clarity of the text should be improved to make it easier for the reader to interpret the details of how this technique was implemented.
- Figure 9: This Figure is too small and has too much information contained in it. This renders this Figure very difficult to interpret other than for a general impression of the data trends. Consider simplifying this Figure or possibly presenting this information on two Figures.
- Lines 394-400, Table 5: The comparison of the two methods for arriving at quantitative rates and sizes of bed load is interesting. Was the criterion in eq (3) applied to the data before the amplitude histogram (AH) method was implemented? It is clear that the criterion it eq (3) was used as part of the technique for the amplitude-frequency (AF) method. Some text should be added to make it clear as to whether eq (3) was applied in relation to the data before applying the AH method.
Lines 571-598: It is clear that the AF method performed better than the AH method in some cases such as the Erlenbach, however, it is also clear that for the other 3 field sites the AH method yielded results for bed load that were quite close to that obtained from the physical samples. It is also not clear whether the general calibration calculated in this study would give
-
RC3: 'Reply on CC1', Anonymous Referee #3, 27 Apr 2022
Review of “Toward a general calibration of the Swiss plate geophone system for fractional bedload transport”, by T. Nicollier, G. Antoniazza, L. Ammann, D. Rickenmann, J. W. Kirchner
The manuscript reports on a new method to calibrate the Swiss plate geophone (SPG) which uses a combination of data collected in a laboratory flume and at four different field sites. The SPG has been shown in previous studies to be an excellent indirect method to measure the rate and size of bed load transport in gravel bedded streams and rivers. This study further develops the science of turning impact data into quantitative data of the transport of bedload. The study of bed load transport using impact plates at several field sites has indicated that a general calibration of the SPG has been difficult to develop for a variety of reasons. This study uses a combination of amplitude and frequency to calibrate the SPG in the quest for a general calibration relation to turn impact data into quantitative values of mass and grain size for gravel bed load transport. This manuscript contains much valuable information and should be published, however, suggestions for improvement of the presentation are given below.
Specific comments:
- Lines 246-251: In these lines it is related how all packets were filtered using equation (3) and packets which do not meet this criterion are ignored in further analysis. It is clear to me how and why this was done, however, in Figure 7 there appears to be substantial overlap between real (blue) and apparent (red) peaks measured in the flume experiments. How many real peaks were rejected through the use of this criterion in the flume experiment data? Also, could the authors estimate how many real peaks were rejected from the 4 field data sets considered in the study? I believe text should be added to the manuscript discussing this issue.
- Lines 239-240, and 257-260: In these sentences the lower and upper thresholds for the amplitude-frequency method are described. Is it correct that the lower threshold (V) was based on the minimum grain size of the size fraction and the upper threshold (V Hz) was based on the maximum grain size of the size fraction being considered? The clarity of the text could be improved to make it easier for the reader to interpret this fact.
- Figure 9: This Figure is too small and has too much information contained in it. This renders this Figure very difficult to interpret other than for a general impression of the data trends. Consider simplifying this Figure or possibly presenting this information on two Figures.
Tobias Nicollier et al.
Data sets
Field and Flume Claibration Datasets Tobias Nicollier, Gilles Antoniazza, Dieter Rickenmann https://www.envidat.ch/#/metadata/sediment-transport-observations-in-swiss-mountain-streams
Tobias Nicollier et al.
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
525 | 111 | 14 | 650 | 40 | 6 | 7 |
- HTML: 525
- PDF: 111
- XML: 14
- Total: 650
- Supplement: 40
- BibTeX: 6
- EndNote: 7
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1