Review of “Back to pristine levels: a meta-analysis of suspended sediment transport in large German river channels”

This paper presents new analysis on suspended concentration changes in many German rivers during the 1990-2010 period. The authors show that there is a significant decrease in surface suspended concentrations by an average of -0.92 mg/L/yr related to a decrease of sediment supply. Interestingly, the suspended sediment load is (in 2010) very close to the reconstructed Early Holocene sediment load before human impact on the landscape. Although it is difficult to single out the cause of this observed sediment decline, the authors argue that the most likely explanation is a change of agricultural practices, i.e. the expansion of conservation agriculture and increasing building of retention basins. Overall, this is a very interesting and well-written paper, with robust statistical analysis and arguments. This manuscript is worth being published in ESurf providing some minor revisions (see below):

Specific comments:

• Lines 34-47: this paragraph is a bit too long. I suggest reducing it by 50%
• Line 60: “In 2009/10 about 38% of arable land were under soil conservation” should be “In 2009/10 about 38% of arable land was under soil conservation”
• Line 65 “periode” should be “period”
• Lines 100-101: what is the pore size the coffee filters that were used?
• Line 109: is there a rationale for choosing 150 sample/year as the cutoff number of samples for considering annual SSC representative? Or is it a value chosen arbitrarily? (Which is fine, but maybe specify it in the main text).
• Lines 130-132 and 170-171: what are the reason(s) for not using rating curves to correct for missing measurements? A sensitivity test (e.g. in the supplementary materials) for a few rivers comparing mean SSC with and without correction for missing measurements using rating curve approach and discharge may be useful here to demonstrate that the data gaps do not affect the calculated mean SSC and observed trends. Also, are the 150 data points evenly distributed throughout seasons for all rivers?
• Line 141 “the assumption of LSR is violated” It would be clearer if you stated what this assumption was first in case people are not familiar with linear least squares regression
• Line 184 “Warwick (2015)” should this be “Warrick (2015)” as before? If not and it is a separate reference, it is missing from the reference list
• Line 214 “artificial areas” I see this is how CORINE names this group of classes (which contains Urban fabric, industrial units, mining and construction sites, artificial vegetated areas) but the name itself is a bit vague. Maybe at line 216 you could specify the kind of land cover class this refers to.
• Lines 247-248: in figure 4, is the “average SSC” the average during the 1990-2010 period?
• Lines 303-305: this is important. Many large river studies have shown that SSC increases (sometimes by a factor 2 to 5) with depth. This increase SSC with depth often related to increase of suspended silt and sand content above the riverbed. Is there any existing study on the Rhine or other German rivers reporting the extent of possible suspended sediment concentration change with depth? Although probably unlikely, a change in the proportion of fine/coarse sediment supply could lead to a decrease of surface SSC (fine sediments) but increase of SSC at depth and therefore an increase of depth average SSC?
• Lines 322 and 325: technical correction, replace “Fig. 4” by “Fig. 5”
• Line 349 “decreases sediment loads had no major impact” should be “decreases in sediment loads had no major impact” or “decreasing sediment loads had no major impact”
• Line 390 “SSCs changes between 1970 and 2020 are unraveled by the compilation of the residual SSCs which were calculated using the daily SSCs and the long term average SSC for each station.” This is not really clear. Are these the residuals from your Sen slope regression?
• Figure 11: The dashed lines show the start and end of your study period. It might be clearer if you labelled them.

Review of Hoffmann et al “Back to pristine levels: a meta-analysis of suspended sediment transport in large German river channels”

In this manuscript Hoffmann and colleagues analyse the suspended sediment concentration (SSC) and load of 9 large river catchments in Germany. Over a 20-year period (1990 – 2010) they find that nearly all (49/62) of the measurement stations recorded a continual decline in mean annual SSC. By the end of the measurement period the SSC seems to be closing in on a natural base level previously observed before human activity in the Rhine valley. The authors then analyse multiple possible driving processes for this trend. This analysis is thorough but they are not able to propose a single likely process driving this trend and so instead suggest a combination of smaller factors, such as soil erosion controls and small-scale flood defences may contribute to the observed trend.

I enjoyed reading this manuscript and found it very thought provoking. I thought the analysis on the whole is thorough and well done, but I feel some more comparison between the basins/stations could offer some insight into the proposed controlling factors. For example, catchments with higher arable land use percentages could be more impacted by any change in soil conservation practices while catchments which have seen flooding in the recent past are more likely to be targeted for flood defences such as rainwater retention basins potentially resulting in a greater decline. Comparative analysis between the catchments may offer some more support for the closing arguments which is currently lacking. I believe these changes, and some other line by line comments I make below, would be minor in nature.

Line by line comments:

Line 44: Waterways may need a definition

Line 63: Rainfall erosivity needs a definition here

Line 63: Periode should be period

Line 65: The original area could be included to add some further contextualisation about the impact of this change.

Line 72: It is not immediately clear what is meant by Work-daily. Perhaps a short definition is required.

Line 144: It is not clear why the Mann-Kendall test is being used or what it tests from this description.

Line 145: I also found the description of the Sen’s slope unclear. What is the magnitude of the trend?

Line 145: Calculate should be calculated

Line 171: “identify changes is suspended…” should be identify changes in suspended…

Line 172: Units of these metrics would help to explain them.

Line 173: I am not sure what is meant by “reactivity of river catchments”

Line 180: When is the log-linear regression analysis used?

Line 184: Missing word: “Surface runoff generating rain fall events”.

Line 202: rainfall erosivity still needs a definition.

Line 231: Why is there no change seen at these 6 stations? Could there be any information derived from these stations? (and again on lines 255 and 260 for the seasonal data).

Line 270: Do these analysis techniques identify the same stations as not changing?  

Line 349: “Decreases in suspended loads”

Line 366: “into the Black Sea”

Line 378: TOC needs a definition

Line 419: The reference to Figure 7 should be to figure 9.

Line 423: Is the rate of the decline between 1990 and 2010 worth discussing? It is 10x times faster than what is potentially seen in the stratigraphy.

Line 433: It is not clear what “an increase in rainfall erosivity between April and November by 2.1 % per year or 42 % from 1990 to 2010” means. Is there an increase in erosivity each year between April and November?

Line 438: The acronym USLE is not defined

Line 455: It is not clear where this data can be seen in figure 10.

Figure 3: The points for stations with no significant difference are not always easily visible