Form and function in hillslope hydrology: in situ imaging and characterization of flow-relevant structures

Jackisch, Conrad; Angermann, Lisa; Allroggen, Niklas; Sprenger, Matthias; Blume, Theresa; Tronicke, Jens; Zehe, Erwin

doi:https://doi.org/10.5194/hess-21-3749-2017

Articles | Volume 21, issue 7

https://doi.org/10.5194/hess-21-3749-2017

© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 3.0 License.

https://doi.org/10.5194/hess-21-3749-2017

© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 3.0 License.

Articles | Volume 21, issue 7

Research article

|

21 Jul 2017

Research article |

| 21 Jul 2017

Form and function in hillslope hydrology: in situ imaging and characterization of flow-relevant structures

Conrad Jackisch, Lisa Angermann, Niklas Allroggen, Matthias Sprenger, Theresa Blume, Jens Tronicke, and Erwin Zehe

Download

Final revised paper (published on 21 Jul 2017)
Preprint (discussion started on 17 May 2016)

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'Review of Jackisch - In situ investigation of rapid subsurface flow', Anonymous Referee #1, 15 Jun 2016
- AC1: 'Author's response to review 1 on Jackisch et al. "In situ investigation of rapid subsurface flow: Identification of relevant spatial structures beyond heterogeneity"', Conrad Jackisch, 22 Aug 2016
RC2: 'Interactive comment on “In situ investigation of rapid subsurface flow: Identification of relevant spatial structures beyond heterogeneity” by C. Jackisch et al.', Anonymous Referee #2, 15 Jun 2016
- AC2: 'Author's response to review 2 on Jackisch et al. "In situ investigation of rapid subsurface flow: Identification of relevant spatial structures beyond heterogeneity"', Conrad Jackisch, 22 Aug 2016
RC3: 'Review of Jakisch et al In-situ investigation of rapid subsurface flow ...', Anonymous Referee #3, 15 Jun 2016
- AC3: 'Author's response to review 3 on Jackisch et al. "In situ investigation of rapid subsurface flow: Identification of relevant spatial structures beyond heterogeneity"', Conrad Jackisch, 22 Aug 2016
SC1: 'Comment on „In situ investigation of rapid subsurface flow: Identification of relevant spatial structures beyond heterogeneity“ by Jackisch et al.', Alexander Zimmermann, 22 Jun 2016
- AC4: 'Author's response to comment on Jackisch et al. "In situ investigation of rapid subsurface flow: Identification of relevant spatial structures beyond heterogeneity"', Conrad Jackisch, 22 Aug 2016
RC4: 'Review of Jackisch et al.', Anonymous Referee #4, 22 Jun 2016
- AC5: 'Author's response to review 4 on Jackisch et al. "In situ investigation of rapid subsurface flow: Identification of relevant spatial structures beyond heterogeneity"', Conrad Jackisch, 22 Aug 2016

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

ED: Reconsider after major revisions (02 Sep 2016) by Ross Woods

AR by Conrad Jackisch on behalf of the Authors (25 Nov 2016) Author's response Manuscript

ED: Referee Nomination & Report Request started (28 Nov 2016) by Ross Woods

RR by Christian Stamm (12 Dec 2016)

Suggestions for revision or reasons for rejection

Review „Form and function in hillslope hydrology: In situ identification and characterization of flow-relevant structures” (revised version).

By C. Jackisch et al.

General comments:

Overall, the revised version is much improved. The authors have put a considerable effort in getting a clear structure such that the findings are not only presented side by side but follow a red thread that is obvious to the reader. Due to the improvement, the value of the large amount of presented data and results is now much more evident.

Nevertheless, there are still issue to be resolved to bring the manuscript into a form ready for publication. These aspects are listed below.

Specific comments:

Terminology: Much of the manuscript deals with the distinction between structure(s) and heterogeneity. However, the actual meaning of these notions is not precisely defined (see for example in the abstract L. 11 – 12, p. 24, L. 25 - 26). As I read the manuscript, structure refers to flow paths (for fast transmission of water (and potentially solutes)) that connect through a large fraction of the flow domain while overall heterogeneity is considered to have only local connectivity. To me it seems essential to distinguish explicitly between situations where such flow paths are caused either by specific pore structures that differ from the other parts of the pore space (e.g., root channels, worm burrows) or whether the flow paths emerge as a results of locally heterogeneity in macroscopically homogeneous soils (see for example (Roth 1995)). Only in the first case, there is in principle the possibility to detect flow structures by local measurements (e.g., on soil samples) because one can infer (potential) connectivity based on the type of pore structure that is identified (e.g. a worm burrow).

In the context of this field experiment, this distinction leads to question whether or not observed flow paths at the plot or hillslope scale can be related to the existence of types of pore structures that are different (at a local scale) from the rest of the pore space. If such structures existed this could help to potentially generalize the findings also for other locations.

If I am right with my view, the authors should make the differences between what they call structure and heterogeneity more explicit and be also more specific in what others have already published on this issue. If my interpretation was wrong, there would also be a need for clarification.

Further details:

p. 2, L. 3 – 4: Sentence too generic.

p. 2, L. 5 – 12: Too general, without insight into the topic of this paper. I cannot find the link between form and function in Wittgensteins Tractatus. So why to mention?

p. 2, L. 15: What needs to be revised?

p. 2, L. 16: Flurys experiments were not at the catchment scale but carried out in different soil types.

p. 2, L. 21 – 25: This paragraph is clear and helpful.

p. 4, L. 11: These responses are functions, not form, aren’t they? If not, what do you actually mean by functions as compared to form?

p. 5, L. 31: Why is it not Fig. 2?

p. 7, L. 15: Provide the dates of the experiments.

p. 7, L. 28: Sentence is not clear.

p. 7, L. 29: Why should this time be only depend on the duration of the input signal and be independent of the size of flow domain (imagine the same input duration for a small system where the input leads to steady-state across the entire system as compared to a large one where most of system has not yet been influenced by the input at the time of ending the input)?

p. 8, L. 2: Provide details of the quick sampler.

p. 8, L. 14 – 15: Skip it if you don’t use it for the paper.

p. 10, L. 12: Give the date of the experiment.

p. 10, L. 15: What is facilitated?

p. 10, L. 18: Before or after the natural rainfall? Please provide a schedule where all relevant events are clearly indicated.

p. 10, L. 29: The modeling was not introduced earlier nor is it mentioned later on. I suggest that you briefly compare the prior predictions (based on your prior knowledge of the system) to the actual results. Please also describe the prior assumption about the flow structure you implemented. This provides a very nice opportunity to demonstrate how you gained insight from the experimental work. You should do so even in case that the prior predictions were rather different from the actual outcomes – this would not be interpreted as a weakness (see for example (Holländer, Blume et al. 2009)). Did you consider the natural rainfall for the modeling?

p. 12, L. 27: How was the grid generated?

p. 13, L. 19: Why is this rainfall event only mentioned here?

p. 13, L. 26: According to Fig. 3 in the companion paper the irrigation is indicated after the runoff event. Please clarify.

p. 13, L. 29 – 31: The procedure is not clear. How can two similarity attributes show increasing similarity? Why is a decrease in similarity attributed to the irrigation? This part is essential to this paper but is rather obscure. It might help to have a figure in this manuscript where all relevant events are depicted on a time line and all relevant comparisons for this similarity calculations are demonstrated. Mathematical notations/equations might also facilitate the understanding.

p. 14, L. 5: I do not fully agree with this statement. Upon inspection, it seems to me that the Ks range is clearly the largest in the topsoil with a pronounced range decrease with depth. The data suggest that at the surface there are samples with very high and with very low values (e.g., due to macropores on the one hand, and due compaction on the other hand, respectively). A similar range decline seems to be related to texture.

p. 14, L. 27: Sentence not clear.

p. 16, Fig. 5: The profiles of the T and B locations seem to differ with regard to gravel content and depth of the periglacial material: what are the implications?

p. 16, L. 7: what is the positive bias in the analyses?

p. 17, L. 2: Please mention the recovery explicitly. It is almost hidden in Fig. 7.

p. 19, L. 1: What is inconsistent about the distributions?

p. 20, Fig. 9: I like these figures. The right panel indicates that the response of water content is faster than that of solute transport pointing to the displacement of pre-event water. This aspect should be explicitly discussed in the manuscript.

p. 20, L. 6: Does this remain unclear at that stage or after considering all of the results. Please clarify.

p. 20, 3.3.2: This section is difficult to follow all of the temporal comparisons. A figure with the time line would help.

p. 21, L. 1 – 2: Why should be a difference between connected flow paths and the irregular network of inter-aggregate pores (see (Roth 1995))?

p. 22, Fig. 9: How can one understand this figure without considering the natural rainfall?

p. 22, L. 1: What’s the overall response here?

p. 24, L. 4 – 5, 10: You say the properties were not expected for this soil texture but you provide an explanation, which is not very peculiar (network of inter-aggregate pores). So, is this pore network different from that of soils of similar texture?

p. 24, L. 13 – 14: Sentence not clear.

p. 24, L. 18: Why should all flow paths identified by their soil moisture response be stained? It seems that you confuse water flow and solute transport (see comment above on Fig. 9).

p. 24, L. 20: What are the point-sampling related issues?

p. 24, L. 22: Sentence not clear.

p. 24, L. 25 – 26: See comment above on terminology.

p. 24, L. 28 – 29: This sentence is rather obscure.

p. 25, L. 4 – 6: It would be useful to refer to some literature.

p. 25, L. 10: Which threshold?

p. 25, L. 17: Where can one see these discrepancies?

p. 25, L. 25: This is actually one of the essential questions here: are the flow paths due to structures which differ from other parts of the pore space in their properties. You suggest that they actually do, however such flow path may also emerge from local heterogeneity (see comment above).

p. 27, L. 30 – 31: What are these new ways?

p. 28, L. 4: As mentioned above, I have serious doubts that Wittgensteins work on formal logic really helps in the context of this manuscript.

p. 28, L. 5: This sentence is actually at the heart of the issue and I suggest you bring this aspect up already in the Introduction. Furthermore, there are ramification to what you conclude that should be discussed in more detail. You criticize – based on your findings - the concept of investigating separately relevant soil structures and their hydrological functions/responses. You suggest that irrigation experiments and time-lapse GPR measurements could solve this problem. While this holds true locally, your argument suggests a major problem for (hillslope) hydrology, which is that of generalization and spatial extrapolation. The rational why to investigate structure and functions at least partially separately is that (physical) structures are generally much more stable in time and relatively easy to observe as compared to hydrological functions. Hence, observations of structures allow for estimates about hydrological responses without carrying out for example irrigation experiments at all sites. If you claim that such a separation does not make sense your claim also entails the message that flow experiments are necessary everywhere to make statements about hydrological functions. I suggest that you discuss this aspect in more depth.

p. 29, L. 20: Conclusive with regard to what?

Conclusion:

I think the manuscript has the potential for an interesting contribution to HESS but still needs to address a number of issues as listed above. Some of them actually got only visibly because the current version is much more readable than the original one.

References:
Holländer, H. M., T. Blume, H. Bormann, W. Buytaert, G. B. Chirico, J.-F. Exbrayat, D. Gustafsson, H. Hölzel, P. Kraft, C. Stamm, S. Stoll, G. Blöschl and H. Flühler (2009). "Comparative predictions of discharge from an artificial catchment (Chicken Creek) using sparse data." Hydrology and Earth System Sciences 13: 2069-2094.

Roth, K. (1995). "Steady state flow in unsaturated, two-dimensional, macroscopically homogeneous Miller-similar medium." Water Resources Research 31: 2127 - 2140.

Hide

RR by Anonymous Referee #4 (10 Jan 2017)

Suggestions for revision or reasons for rejection

Review “Form and Function in hillslope hydrology: In situ identification and characterisation of flow relevant structures”
The revised version of the manuscript of Jackisch et al. (2016) investigates the identification of structures relevant for subsurface flow processes in the context of form and function based on soil samples, Ks measurements, TDR, GPR and tracers (KCl and brilliant blue). The revised manuscript is now a very nice study, which would be very helpful for the community to shed light into subsurface flow processes in periglacial cover beds. The idea of form and function is hereby an interesting perspective on the hydrological system and could be extended to many other processes.
The content is relevant for the journal but the modalities still need additional revision, especially in the coordination to the other manuscript of Angermann et al. In the here presented modified version the authors integrated the comments of the reviewers. Structure and story line are getting form. The author’s introduction is plausible and taken the relevant literature into account. They developed the theory of form and function in a very nice way. The materials block is shortened reasonable and most of the unimportant technical details are removed. The benefit of all these measurements to understand subsurface flow patterns is now recognisable.
The description of the results is in the presented form much better except the GPR part.
But the benefit of their experimental design in the view of the presented theory of form and function compared to other experimental studies is still unclear. Their experiments compared to other studies are similar. The same hypotheses are postulated without direct knowledge what is exactly is the reason for the observation.
Form and function is scale dependent. At the hillslope scale or at larger scales it is hard to identify. That is different to the plot scale, where the water-retention curve and the relationship of hydraulic conductivity to pressure head is available from lab experiments because of clear boundary conditions. The authors start to formulate an answer for the next scale, similar to Reggiani (1999, 2000, 2001) on the model level, whit his REW model concept. And at that point the authors give no answer how to solve the problem. Maybe only setup like the experimental hillslopes of Japan or the Biosphere 2 experiments of the Troch group are able to give therefore an answer and should be cited and discussed in the introduction.
The hypotheses are not well formulated. Where is the new direction to understand form and function at the hillslope scale? Here they just add different methodologies and have still many assumptions to be taken into account. The difference between Q2 and Q3 is not strong. From my perspective they can combine them to one. There is still potential for much clearer aims. A figure, which guide through the manuscript would be nice. Which experiments have which aim on which time scale and with which spatial footprint, etc. . And here is the point where the authors should present, what the difference is between the two studies. Currently Jackisch et al. and Angermann et al. have many similarities, whereby the second study has better structure and is clearer formulated compared to the first. The titles sounds too similar. Jackisch is investigating the flow relevant structures and Angermann the subsurface. But are these two topics spreadable, to investigate structures the flow is has to be known and vies versa?
The block of the models (page 10 line 30- page11 line 1) is still not that what I would expect in an experimental study without presenting any simulation results. If they have proven the boundary conditions of the experimental setup with a physical hydrological model, how have they parameterised the entire unknown structures of the hillslope to extract the required irrigation amount and estimated the response time of the system. Why have they not proven their findings of these structures with the model with the observation data in a third step? Here they could have presented a form and function based on observations and virtual experiments (Weiler and McDonnell, 2004) with all its structural uncertainties.
In the materials and method block a short description of the estimation of structural similarity and its mean is missing. A very clear description can be found in the manuscript of Angermann et al. The presented description is hardly understandable.
Most of the plots are still hardly comprehensible and need simplification. Two plots (11 and 12) are used by both manuscripts. Explanation has to be given.
Figure 1: I am wondering how they could be using soil samples and Ks measurements of the wider surrounding area in case of these complex periglacial cover beds and not observations only from the direct surrounding area. The analysis could be completely wrong.
Abstract still needs revision, because they are not able to present form and function.
They have moved many parts to the appendix, but that needs linking to the main manuscript and not only the moving.
Discussion has to be shortened with more linkage to form and function and to model approaches for the larger scale, which could have the potential to investigate the specific problem.
Writing still needs modification and proving.
Abbreviations (Two Way Travel time) have to be defined.

Specific comments:
P3L2: That should be questionable if double peaks are driven by macropore networks. The phenomenon is very special and has not a fast response in the sense of subsurface storm flow. It is slower. But of cause their assumption would be able to investigate the phenomenon. A link to it would need an extra point.
P39-10: That’s not clear; dye tracers are used to understand the interactions between macroporosity and matrix. What do they mean with the requirement of strong assumptions to that topic?
P5L2: change “large number of methods”, they present GPR, dye and salt tracer, soil samples and Ks measurements and irrigation data, that is not large.
P5L13: Add English name to the botanical. Are the conifers important? I just saw deciduous plots.
P7L13: Figure 13 is still hard to understand but important. I would move a simplified version back to the main manuscript. Add in figure 1 station E. It is not clear where it is located.
P7L15: What kind of return period for the specific area are these values?
P8L2: check spelling, is that important? A fast sampling design is obvious for that kind of experiments.
P8L11: The neglecting of lateral flow is common practice in tracer analysis in soil profiles. But by looking in a new direction of form and function should it not be critical discussed? The authors could not just criticise the old studies and then do it exactly the same way.
P8L15: Link to Angermann et al. and remove the isotopes from the appendix. The appendix is not the area, where all not presented material should be moved to.
P9L10: check spelling
P10: Shorten that block
P10L30: Why are the plot sites not proven with the physical model? I am wondering of the intensity of 30 mm in comparison to the plot sites.
P11L8: How have they removed shrubs without disturbing the soil, with chain saw and disc cutter?
P12L1: The presented irrigation is completely different to the plot experiments. Explain the reason and how these experiments still could be comparable. Are the antecedent conditions similar?
P12L26-30: Block has to be rewritten. In that form it is not clear.
P13L25: add duration of the rainfall events and intensities.
Figure 5: Where are these profiles? A link in the figures 2 and 3 would be important to understand the spatial distribution; maybe an arrangement in a catena for the hillslope experiment.
P16L7: What did they mean with a positive bias in the soil data?
P16L11: Where is B7? A signature (line or an arrow to the layer) should be added to the base layer in figure 6.
P17L14: The differences in the figures 6, 15 and 16 are hardly identifiable. The intention of radargrams is unclear.
P17L12-P18L9: Block is hard to follow.
Figure 9: Add abbreviation of the three CHP’s to the caption. Explain the three differences.
P18L4: Where is T7, is it the 7 in figure10 and 3? Then add the T and add a comment to the captions.
Figure 10 needs simplification. The main focus is on 1.5 m depth, figure should highlight that region.
Figure 11: Why have they not selected the same TDR profiles in the temporal development plots and in the distribution functions of the TDR profiles for the apparent velocity? Angermann (figure 8) used the same plot, but here two time steps more are presented, why?
Figure 12 is also in Angermann (figure 9) but there much better explained. What is the reason for the duplication?
P24L5: But that cannot be the aim of an experiment that there is still the need of more data to have representative sample size or to have a perfect conversion. Is not the better solution to understand the process with less sampling by using an intelligent strategy, which they have? Later, in the discussion they even discuss that less effort is possible.
P25L28-31: Here a link to form and function to model approaches and to Reggiani could be an important step.
P26-27: shorten
P28L1-3: I miss models which would be able to take larger scales into account (Weiler and McDonnell, 2004; Lee, Zehe, Sivapalan, 2005ac) and not only the soil core and hillslope scale.

Hide

ED: Publish subject to revisions (further review by Editor and Referees) (31 Jan 2017) by Ross Woods

AR by Conrad Jackisch on behalf of the Authors (18 Mar 2017) Author's response Manuscript

ED: Referee Nomination & Report Request started (23 Mar 2017) by Ross Woods

RR by Christian Stamm (19 Apr 2017)

Suggestions for revision or reasons for rejection

Review „Form and function in hillslope hydrology: In situ identification and characterization of flow-relevant structures from point to hillslope scale” (revised version).

By C. Jackisch et al.

The manuscript has been carefully revised. I only have a few comments to add that may help to clarify some aspects. Basically, the manuscript is ready for publication.

Relevance Wittgensteins Tractatus for this manuscript: At some point, it may be a matter of taste and opinion whether or not to refer to a piece of literature. However, I have to admit that the two sentences the authors refer to (2.033, 4.1241; I could not find a 2.045 in neither the German nor the English version) reading as follows:

2.033 The form is the possibility of the structure.

(German: 2.033 Die Form ist die Möglichkeit der Struktur.
http://tractatus-online.appspot.com/Tractatus/jonathan/D.html)

4.1241 One cannot distinguish forms from one another by saying that one has this property but the other that: for this assumes that there is a sense in asserting either property of either form.

(German: Formen kann man nicht dadurch voneinander unterscheiden, dass man sagt, die eine habe diese, die andere aber jene Eigenschaft; denn dies setzt voraus, dass es einen Sinn habe, beide Eigenschaften von beiden Formen auszusagen.)

do not really provide me any insight that is useful for understanding the actual content of the manuscript nor can I link them to the respective sentence in the manuscript “In
abstract terms the relation of form and function is fundamental for the concept that we can predict a behavior of a system under different forcing by knowing its constructive properties” (p. 2, L. 6 – 8 of the revised manuscript). If there is something in these quotes that make them relevant please explain that content in such terms that average readers like me can also learn something.

p. 7, L. 8: “… to activate all potential flow paths …”: All is probably overstated: If you would prolong the irrigation, more flow paths might get connected.

p. 14, L. 4 – 10, Fig. 4: I think the combination of text and part of the figure is still misleading. The Ksat data in Fig. 4 clearly suggest that the variance decreases with depths. Fig. 13 in the appendix reveals a different results that supports the actual text. I’d suggest to show the data from the appendix in Fig. 4 to be consistent with what is presented to the reader in the main text.

p. 20, Fig. 9: There seems to be some inconsistencies between the explanation of the calculation of the velocity distribution for the Br (p. 9, L. 4 – 5) and Fig. 9: In the text, it is written that the velocity can scale by several orders of magnitude depending how tfix is chosen (p. 9). This considerable uncertainty is not shown in Fig. 9. Accordingly, the interpretation may be strongly influenced by an arbitrary choice of tfix. This should be avoided by displaying the velocity uncertainty.

p. 29, L. 10 – 11: Still, I am not fully convinced with this absolute statement that form and function cannot be approached separately. If we could actually quantify the necessary details of the pore space this separation was possible. However, because small spatial dimensions of the relevant structures (at least perpendicular to the main flow direction) we currently lack methods to actually reveal these structures in undisturbed soils at the plot or hillslope scale. Given this limitation I would agree with the authors that investigating form (of the pore space) in isolation will probably only reveal limited information about flow processes (function) and that methods as described in the manuscript are useful tools for jointly analyzing form and hydrological function of plots and hillslopes.

Hide

RR by Anonymous Referee #4 (28 Apr 2017)

ED: Publish subject to technical corrections (22 May 2017) by Ross Woods

AR by Conrad Jackisch on behalf of the Authors (24 May 2017) Author's response Manuscript

Short summary

Rapid subsurface flow in structured soils facilitates fast vertical and lateral redistribution of event water. We present its in situ exploration through local measurements and irrigation experiments. Special emphasis is given to a coherent combination of hydrological and geophysical methods. The study highlights that form and function operate as conjugated pairs. Dynamic imaging through time-lapse GPR was key to observing both and to identifying hydrologically relevant structures.