Snow processes in mountain forests: interception modeling for coarse-scale applications

Helbig, Nora; Moeser, David; Teich, Michaela; Vincent, Laure; Lejeune, Yves; Sicart, Jean-Emmanuel; Monnet, Jean-Matthieu

doi:https://doi.org/10.5194/hess-24-2545-2020

Articles | Volume 24, issue 5

https://doi.org/10.5194/hess-24-2545-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/hess-24-2545-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 24, issue 5

Research article

|

15 May 2020

Research article |

| 15 May 2020

Snow processes in mountain forests: interception modeling for coarse-scale applications

Nora Helbig, David Moeser, Michaela Teich, Laure Vincent, Yves Lejeune, Jean-Emmanuel Sicart, and Jean-Matthieu Monnet

Download

Final revised paper (published on 15 May 2020)
Preprint (discussion started on 23 Aug 2019)

Interactive discussion

Status: closed

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

- Printer-friendly version

- Supplement

RC1: 'Review of Helbig et al.', Anonymous Referee #1, 23 Sep 2019
- AC1: 'Author comment as response to reviewer comment 1', Nora Helbig, 20 Nov 2019
RC2: 'Intersting, but important questionmarks', Anonymous Referee #2, 13 Oct 2019
- AC2: 'author comment as response to referee comment 2', Nora Helbig, 20 Nov 2019

Peer-review completion

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

ED: Publish subject to revisions (further review by editor and referees) (22 Nov 2019) by Adriaan J. (Ryan) Teuling

Your manuscript on "Snow processes in mountain forests: Interception modeling for coarse-scale applications" has now been reviewed by two referees. They both seem to agree in their assessment that the work is of potential interest and sufficient quality, but that several issues will need to be addressed before the manuscript can be considered for publication in HESS. I have classified these comments as revisions rather than major revisions because new analysis might not be necessary, but given the nature of the comments they might still require substantial effort. In any case the revised manuscript will be returned to the referees for consultation.

When revision your manuscript, please pay particular attention to the following aspects:
- Both referees comment on the use of snow depth of SWE. This is a critical issue. It is important to realize and recognize that snow depth is not a conservative quantity, and that there is a risk for bias when comparing snow depth observations between places with different microclimate conditions. This does not mean that there is no value in making a model based on snow depth observations, but the right arguments should be used to motivate this choice. This is not to reduce "any potential error when converting measured HS values to SWE." as is stated in the manuscript. The real reason (I guess) is that SWE observations are not available because they require more effort in sampling. But that does not make them less relevant. So please be clear in motivating the choices, and acknowledge any potential limitation and the fact that all models will use SWE and not snow depth.
- The choice for the functional relationships that are fitted to be data need to be motivated, and if possible evaluated against alternative models. There should be some objective criteria behind the choice for a particular functional form.
- Please pay particular attention to the writing. If possible, let the manuscript be proof-read by a native speaker/colleague.

I am looking forward to receiving a revised version of the manuscript. Let me know if you have any questions regarding the reviews or the revision.

Best regards

Ryan Teuling

Hide

AR by Nora Helbig on behalf of the Authors (20 Dec 2019) Author's response Manuscript

ED: Referee Nomination & Report Request started (07 Jan 2020) by Adriaan J. (Ryan) Teuling

RR by Anonymous Referee #2 (06 Feb 2020)

Suggestions for revision or reasons for rejection

The data and the model equations are better described now. While this clarified things, I still have some rather fundamental concerns.

• Snow depth vs SWE
I see the reasoning of the authors, but still would argue that for basically all further uses, one would need SWE. The authors themselves state in the introduction “Accurately modelling the spatial distribution of snow water equivalent in forested regions is thus necessary for climate and water resource modelling over a variety of scales.”.
I am still confused about what the snow height actually refers to. If the intercepted height is 10 cm, does this mean that there are 10 cm snow on the trees or that there is so much snow on the trees that this would be 10 cm if distributed on the ground surface?
By using heights instead of SWE the model does not necessarily conserve snow masses, which might provide the model with some (unwanted) flexibility. This issue could at least been looked at by estimating densities backwards (assuming conservation of masses).
Also, when it comes to the conclusions, the use of snow height might cause confusion. What exactly does a statement like “as much as 68 % and on average 43 % of the cumulative snowfall was retained” mean here. Do the % values refer to heights? These might then be quite different from the snow mass (which a not so careful reader of the conclusion might think of).
• Model performance
The performance measures need to be better described, with the information in 3.3 it is not possible to reproduce these. For instance, what range was used for normalization (min-max or some percentiles, the latter would probably be more robust).
I am also wondering how good the model actually is. The performance measures and the figures look nice, but of course, some of the good-looking performance is rather trivial. Figure 4, for instance, looks good because with larger precip events obviously also the interception increases. The study would be much more convincing if the results of the new model were compared to some baseline estimate. I would recommend using some very basic interception model for comparison to better illustrate the added value of the new approach.
• Uncertainties
The authors provide confidence intervals for the different coefficients. This is good, but the more interesting question would be how these translate into uncertainties in the model simulations. For this, all model results should be given with some uncertainty bounds (which could be derived using some MC approach)
• Applicability of the model elsewhere
Validity for a range of conditions: In the previous round of reviews the issue was raised that the validation sites actually are relatively similar and do not span the potential range of conditions. While the author basically agreed with this in their response, the changes in the text do not fully the potential limitations.
• Structure
The authors did not understand the previous comment “Central parts of the methods are described first in the result section.”. I am sorry for the confusion and will try to explain this better. The two fundamental equations pop up in the results and it is not clear where they came from. I understand now based on the authors' response that these equations were derived from the Swiss data. Still, this leaves me wondering: was the form of the equation chosen a priori and then parameter values were estimated based on the data or were a number of functions/expressions evaluated? In the first case, I would expect to see some motivation of the expression in the methods, in the latter case I would like to know which range of expressions has been considered and the decision for one or the other has been taken.
• Language:
Sorry for repeating this example of ambiguous language:
P2L17ff: “In winter as much as 60 % of the cumulative snowfall may be retained in conifer forests”
Would snow in another season not be intercepted?
“.. and as much as 24 % of total annual snowfall may be retained in deciduous forests in the southern Andes”
This reads as if 60% of some total snowfall is intercepted in coniferous forests and 24% are intercepted in deciduous forests in the Andes, i.e. 84% are intercepted in total.
This is a minor detail and one can guess what the authors mean, but in a scientific paper these things should be formulated as clearly as possible. Here, it should be clarified what the % refers to.

As another example: L233ff: “Modeling forest canopy involves several processes such as interception, unloading, melt and drip, and sublimation.”
Modelling forest canopy would involve rather biological processes, what the author mean is something like ‘Modelling the effects of the forst canopy on snow accumulation on the ground ….’

Hide

RR by Anonymous Referee #1 (06 Feb 2020)

ED: Publish subject to revisions (further review by editor and referees) (06 Feb 2020) by Adriaan J. (Ryan) Teuling

AR by Nora Helbig on behalf of the Authors (26 Feb 2020) Author's response Manuscript

ED: Publish subject to minor revisions (further review by editor) (07 Mar 2020) by Adriaan J. (Ryan) Teuling

AR by Nora Helbig on behalf of the Authors (13 Mar 2020) Author's response Manuscript

ED: Publish as is (29 Mar 2020) by Adriaan J. (Ryan) Teuling

AR by Nora Helbig on behalf of the Authors (07 Apr 2020) Manuscript

Short summary

Snow retained in the forest canopy (snow interception) drives spatial variability of the subcanopy snow accumulation. As such, accurately describing snow interception in models is of importance for various applications such as hydrological, weather, and climate predictions. We developed descriptions for the spatial mean and variability of snow interception. An independent evaluation demonstrated that the novel models can be applied in coarse land surface model grid cells.