Hartmann et al. present work on infiltration experiments across a moraine chronosequence in the Swiss alps, spanning almost 14,000 years. They performed infiltration experiments on four plots of different ages, with each plot being subdivided into three subplots where different precipitation intensities were applied.

The work is a heavily revised version of a previously submitted manuscript. I was one of the original reviewers back then and suggested a rejection. This submission was deemed different enough to be considered a new submission, and I would agree with this assessment. The way the authors reworked the current manuscript makes it much more enticing and sets it apart further from Hartmann 2020a and 2020b (in my opinion).

In the introduction, the authors describe the need for identifying different flow patterns for potential integration into landscape evolution models. It would be interesting to revisit this idea in the discussion. What does the work suggest such an integration could look like, and more importantly, what would the practical differences be between the different flow types for such a model (especially considering that some of the differences between the plots appear minor, even though they were found to be significant)?

I am also wondering if the amount of rocks has an effect on the flow type. If a large fraction of the profile is taken up by rocks, percolating water will be restricted to the space between the rocks. The authors do include flow types that take into account rocks in the profile. From what I understand, though, this applies mostly to homogeneous flow that happens around the rocks. What would happen if a larger rock led to an effective partitioning of an otherwise homogeneous wetting front? (That is, if the soil below the rock remained dry)

Further, it appears that some profiles exhibited a significant portion of rocks in the upper soil layers. Overland flow was not measured, but it could be beneficial to talk a little more about the potential impact of less water infiltrating at these sites.

In the discussion, the authors mention very briefly that the edges of the plots were not analyzed. I might have missed this earlier, but does this only apply to the outer edges of the 1x1.5 m plots or also to the borders between plots 1 and 2 and plots 2 and 3? If so, how big of a buffer was included? I could imagine that interactions around the inner boundaries could have an impact, too.

I think the revisions are a little more than just minor, but I am confident that the authors can address them.

General comments:

Page 4, Lines 9-11

What is the reasoning for having two plots of ~the same age?

7, 24+

Can you describe what the practical differences are between the flow types?

7,31

Do these indices depend on the effective width of the profile? If there is a flow restriction, for example from a rock, wouldn’t that lead to a “compression” of the water flux through the narrower width? Is it possible that water from one experiment gets drawn into another subplot through matric forces?

9,13

Given the low n, there is a chance that the trend is random, even if it’s statistically significant, no?

Fig 4

When there were rocks at or immediately below the surface, what happened to the water that couldn’t access that space? Did it run off?

12, 16-18

Which appears to be the case for most profiles…?

13, 6-7

Is it really…?

Fig 6 and 14, 6-7

Curious to read why sometimes SAD is greater for lower intensities and sometimes greater at higher intensities.

18, flow type classification

A table with the percentages would be good here so that the reader doesn’t have to piece together everything. Fig 10 is nice, but it is a little difficult to compare the length of the bar sections after “Matrix flow between rocks”. Maybe something for an appendix.

20, Fig 11

I’m wondering if the percentage of rocks in a profile affects this as well. This harkens back to my earlier comment in which way rocks affect all these flow characteristics.

21, 5-8

I was thinking about this the entire time while reading the manuscript. Can you estimate infiltrated volume or surface runoff? I would imagine the rocks close at the surface play a huge role here and not just the soil properties.

23, 1-3

This makes sense to me. It could be a combination of both larger diameter and longer roots.

23, 10-16

I like that you bring this up. My initial interpretation would have been that the different external factors of the sites (which also affect landscape evolution) are more important than age.

23, 30-31

Didn’t you argue on the previous page that hydrophobicity could affect the infiltration patterns…?

25, 29-30

This is an important point that needs to be included in the methods (unless I missed it somehow). Does this only refer to the outer edges or to the inner boundaries as well? How much was excluded?

Review by John Nimmo.

This paper provides an extensive and valuable set of field observations of the subsurface flow patterns generated by three different irrigation intensities over four members of a soil chronosequence. As in previous works using similar methods, this study offers quantitative analysis of unsaturated flow features that otherwise would be evaluated subjectively and without quantification.

The main value is in providing evidence to elucidate how factors including soil age, input intensity, vegetative cover, and others influence the depth and homogeneity of the distribution of the infiltrated water. In particular, a major issue is the distinction between preferential and homogeneous flow patterns, understanding of which has tremendous importance to water supply and water quality matters, as well as to agriculture and ecosystem health. The linkage to physical phenomena is primarily through classification into six categories based on a modified version of the scheme of Weiler and Flühler (2004). The paper provides useful documentation of soil developmental processes over 13500 years. Together with an earlier study of Hartmann et al. (2020a), it provides evidence of the differences resulting from calcareous-vs.-siliceous parent materials.

The data analysis is extremely thorough. A great variety of statistical methods are employed, perhaps more than necessary. I see little or no value in the P_dye analysis because the constraint of monotonicity is a serious shortcoming that could distort the interpretation of how water behaves in the profile.

Presentation of multifactor comparisons of many individual experiments is unavoidably complex, and is done here (figures 3-10) through an organization that requires the reader’s time and effort to understand and evaluate, but it does show the results in a way that the effects of soil age, irrigation intensity, and spatial variability can be directly seen.

The main problem I find in the manuscript is confusion and inconsistency concerning the classification of types of subsurface flow. Much of this relates to the term “finger flow”, for which I don’t find a clear definition in this paper, and which seems to be used in different ways.

Some background from my own understanding: Three main categories of preferential flow are commonly used—funneled flow, which is directed into particular downward paths as a result of heterogeneities of the medium that provide faster flowpaths through the more conductive material; fingered flow, which is initiated at flow instabilities in the wetting front and sustained in downward preferential paths by the greatly enhanced hydraulic conductivity of the newly wetted material; and macropore flow, which proceeds through elongated continuous pores over significant distances within the medium.

I see these categories to be represented in the scheme of Weiler and Flühler (2004) (hereafter referred to as WF2004), which is designed specifically for use in interpreting dye-tracer results. Macropore flow needs matrix interaction to be visible, as acknowledged in the first three categories of WF2004. I see the term “matrix heterogeneous flow” as a synonym for funneled flow, and it is quite adequate in that usage. Instability-initiated fingered flow would be difficult or impossible to distinguish from matrix heterogeneous flow when the only evidence is from pictures of dye-tracer distribution. Thus it is appropriate to group both of these flow modes together as in the fourth WF2004 category, “Heterogeneous matrix flow and fingering”. Absence of preferential flow is reasonably called homogeneous matrix flow in the fifth category. In the present study, the use of the WF2004 classification scheme is a suitable approach for evaluating dye-tracer patterns in terms of preferential flow. It is extended reasonably with the added sixth category to accommodate effects of large stones in the soil. The other modifications adopted here are poorly explained, and appear to deviate significantly from some widely understood general features of preferential flow, and from the evidence available from this study as I understand it. Below, I explain these issues further in relation to finger flow and macropore flow.

1. Finger flow

Instability-initiated fingers are possible, though my expectation in such heterogeneous soil is that these are likely to be rapidly channeled into funneled flowpaths. Based on the images and other available information in the present study, I doubt that it is possible to discern whether instability-initiated fingering is an active process. In 18:24 (location noted as page:line) the term “finger flow” seems to mean any preferential flow that is identified by finger-like patterns of dye tracer, not limited to the downward-moving fingers of wetness generated at a wetting-front instability. The finger-like patterns in the dye could result from other modes of preferential flow. If what is meant is just that the patterns have a finger-like shape, without regard to specific process, “finger flow” would be better replaced by the general term “preferential flow”. This issue occurs also in 1:14, 21:4, 22:3-6, 22:16, 23:21-22, 24:5-8, and 26:4-13. On the other hand, the specific mode of instability-initiated finger flow is the subject of 23:4-6 and 24:11—25:2. It also is strongly related to the effects of hydrophobicity in 22:6 – 23:9. These passages need clarification and consistency. Overall, finger flow must be explicitly defined and the term used consistently. If the paper actually does claim that instability-generated finger flow is detected in these experiments, there needs to be justification for how this can be determined.  

2. Macropore flow

There needs to be more discussion of the possible effects of macropores. The soils are likely rich in narrow macropores that result from growing and decaying roots (apparent in the images of both young and old soils), and other bioactive processes. If such macropores convey significant water that then has some degree of interaction with soil matrix material, they could create flow pattern features of the types observed. The statements in 18:15-22 are hard to understand and accept, where it is implied that finger flow can be distinguished from macropore flow, and stated that no macropore flow was found. If there are reasons to justify ruling out active macropore flow, they need to be carefully explained.

I cannot make sense of the statements in 7:30-33, which seem to imply that finger flow can be distinguished from macropore flow, but then contradict that in saying that no such differentiation is made. Then there is confusion in the statement that narrow finger flowpaths could somehow be misclassified as macropore flow with high (but not low or intermediate) interaction.

Section 3.2 (18:12-30) needs to be rewritten for consistency with other clarifications. The category “Macropore flow with high interaction/ Finger flow” is mentioned here and in Figure 10, but it is not mentioned in the definition of the categories on page 7 and is not in the scheme of WF2004.

Overall:

This paper is dense with useful information and provides insights into the development of preferential flow paths during landscape evolution and several other important facets of unsaturated flow in calcareous soils. It needs revision for consistency and adherence to evidence and general understanding of the different types of preferential flow paths. Because the basic experimental work and presentation of data are sound, I have classed these revisions as minor, though I see them as extremely important.

Minor comments:

6:16-18. Rewrite for clarity. Use of “below” in line 16 suggests that the excavation is downward to produce horizontal planes, but “vertical profiles” in 17 suggests otherwise. Does “below” mean “downslope of”? The operation suggests that a trench was first excavated off to the side of the plot to provide access for vertical profiling. More details on this would be helpful.

7:17. What is meant by “amount”? The number of flow paths?

7:28-29. Clarify—maybe make two sentences. Start with a clear description of  the problem caused by rocks. Then the solution devised.

7:31-33 Why “misclassified”? What is unreasonable about “macropore flow with high interaction”?

7:34. Proportion in relation to what? PFF needs to be defined more clearly.

12:8-10. Split sentence into two, for clarity.

14:1. Replace “Whereas” with “In contrast,” or similar expression.

17:4-5. It seems at best to be a very subtle effect for the middle portion of the profiles to be less significantly different. Maybe not worth mentioning.

21:30-31. Delete “influence the water transport and”.

22:2-3. Word missing from sentence?

22:4-5. The point is not about the water transport in general but the pattern of the water transport that is affected. Insert “pattern” or some similar expression.

23:22. Word wrong or missing.

23:29. “Matric potential”, not “Matrix potential”.

23:32. The paragraph starting here, and also the next one, are all about the two older soils. This should be made clear to the reader in the first sentence at line 32. Consider rearranging discussion from this point through 25:13 in order to proceed in the logical order of young to old.

25:24-27. Confusing. Which of the plots were less affected by the direct application of water? Why is there consideration of the boundaries in this?