Review of “Bedload transport fluctuations, flow conditions and disequilibrium ratio at the Swiss Erlenbach stream: results from 27 years of highresolution temporal measurements” by Dieter Rickenmann

CONTRIBUTIONS AND AUDIENCE

In this paper, the author uses 27 years of Swiss plate geophone recordings in the Erlenbach to analyze the dynamics of bedload transport in mountain streams. The paper considers an original protocol that compares measured transport with capacity transport. The results show that transport intensity and fluctuations are strongly correlated with pre-flood conditions, and more specifically with sediment availability in the river bed.

I found the article well-written with an exhaustive review of the literature, but also very difficult to read, mainly due to the excessive number of parameters and abbreviations (e.g. 15 flow parameters Q_x with different x-indices). But I found the methodology interesting and think that the proposed framework may be useful for future work.

I have no reservations about publication with minor revisions.

COMMENTS

Line 158: If I've understood correctly, the signal emitted by the >10mm fraction is used as a proxy to estimate the transport rate associated with the 4mm to 10mm fraction, by calibration? Perhaps this could be written more explicitly.

Line 174: does “zero values” means periods with no impulses in the time series?

Line 215: it is not trivial to apply a bedload equation in such a torrent context. Could you specify which part of the upstream reach and which grain size distribution were considered for the calculations? (in particular I suppose grain size varies a great deal through time?)

Line 253: is there any reason why you use Qbtot and not Qbred in Eq2?

Line 257: what are Qb,s and Qb,e?

Line 286: correct "in the in the"

Line 286-9: it means that larger Q are associated with smaller Qbm? 

Line 283: the analysis presented below does not really explain why the wider spread in period B. Or maybe I missed an information?

Line 317: How did you delimitate these periods? Visually?

Line 346: why 30 min?

Line 354:  is “log values of EdM (ACF_LEdM_m30)” correct?

Fig7: I guess it is an important result, but it is still difficult for me to appreciate the physical meaning of the relation between Qb and the autocorrelation coefficient…

Line 361: how did you define (measure) the critical discharge?

Line 263: it may be usefull to recall here the definition “discharge threshold at the end of an event”. Same for Qs.

Fig9 and 10: why do we observe a positive correlation of Qe with QbT and a negative correlation with QbM? I missed something…

Fig10: what about the correlation between Qs and LQbt?

Lines 383-386: brackets are not necessary

Line 483: because more in-channel material immediately available for transport?

Line 443: Phase 1 has been mentioned in different places for explaining the data in period B. I think it could be interesting to recall the exact definition of phase 1 used in this paper and also to indicate the t*/tc* range concerned?

Line 481: this sentence is not clear

Line 489-490: Your work conclude that the bed plays important role in what was measured. We can regret that you don’t propose a paragraph dedicated to your direct bed observations (if any?). and do you have indications on the size of the transported materials for both periods?

This work presents an analysis of an extensive set of geophone and piezoelectric sensor data that are used to infer bed sediment transport in an Alpine stream. The primary focus of the paper the collection and processing of these particular data, and correlations in the results.

The paper is closely related to the work of Rickenmann (2020), which presented an analysis of the same data, but based on events rather than per-minute observations. Although it is reassuring to see that the change in methodology does not significantly change the results, in my view a shortcoming of the paper is the similarity to this previous work.

My main comment is therefore to suggest that a more direct comparison of the strengths and weaknesses of the two approaches is made, in order to demonstrate that the method presented here is indeed a substantial advance. In particular, the scientific benefit would be demonstrated by showing that the new methodology has the ability to test specific scientific hypotheses that could be tested with the previous approach.

I concur with anonymous referee #1's list of comments, and suggest a few more minor clarifications:

Abstract: a certain amount of jargon is used here (disequilibrium ratio, lag time, critical discharge, coefficient of variation,  clockwise/anticlockwise transport behaviour) much of which is likely to be unclear to people who have not already read the paper.

line 78: define coefficient of variation

line 219: "xx% of the particles are finer" presumably refers to particle mass, rather than particle number?

line 262: Clarify exactly what the 'kernel smoothing' does (presumably a type of low-pass filter?) and what the unit of bandwidth is? (I'd usually understand bandwidth to be measured in Hz).