1. General comments.
The authors have improved the manuscript and addressed main concerns of the two reviewers. The formulations were chosen to be less generalized and more specific to the investigated avalanches and the friction calculation is now more reasonable and understandable. The differences of energy sources between friction and entrainment is now supported by an additional calculation of entrainment.
However, other points are not considered, which I will list below. There is still a problem that presented data is not able to support a conclusion. The main problem is the insufficient description of the used methods.
I disagree with the assumption that the IRT camera is a useful tool for qualitative comparisons for temperatures after an avalanche stopped. The presented verification and existing literature does not allow this statement. This requires more careful formulations.
2. Specific comments
2.1 Conclusions are not supported by the data
The main conclusion is that friction was the main energy source, which was dependent on the drop height. This is solid after discussing Figure 9 and Table 2. The authors should only be more precise that it is in fact dependent on the effective drop height, which is dependent on the growth index (Eq 3).
However, the conclusion that entrainment varied between the avalanches does not hold, because of insufficient analysis and presentation of the data. The authors presented warming due to entrainment of 0.08 and 0.3 degrees. The calculation is very questionable. Firstly, it is not clear to me how the temperature difference between mass released and mass entrained is calculated, which was used in Eq. 4. I can only assume it is the difference of T_Prelease and T_Ptrack, averaged over the solid lines shown in Figure 9, with excluding the grey areas. Besides this ambiguity it is concerning that that these plotted results do not correspond to the calculated values in Table 2, to my opinion. I see a positive difference between the blue and the orange temperature profile in average for both cases, maybe even a larger one for avalanche #2 (left). In Table 2 this impression is not represented. A smaller (and negative!) change was mentioned for avalanche #1. This is directly affecting the conclusion “varied between avalanches”. The fact that Table 2 states that T_Ptrack is 0.1 degree cooler in average compared to T_Prelease is very concerning. A warming due to entrainment cannot be explained with Eq. 4, as the authors stated, it would result in a cooling. The authors need to be clearer that they do not make two mistakes here, firstly wrong averaging, secondly, a sign error.
In the discussion section the authors stated that deposits from powder cloud has consistently lower temperatures than the warm dense core (line 615ff). It is not clear to me if the author refer to the general impression of IRT images or to Figure 7, where the authors presented lower temperatures for two transects out of 4, which show lower temperatures for a thin deposition area. This is one example of how the reader does not know exactly how the authors argue and leave the reader unsure (see section 2.2 below for other examples). For both interpretations I have concerns. For the IRT images my opinion is that also qualitative relations may not be possible and therefore the authors may not be able to state this conclusion (see comments to IRT images in section 2.3).
For the lower temperatures in the thin deposits compared to the dense core in Figure 7 (transect L1 and L2), I am unsure by the presented material if these transects are indeed covering deposits in the middle of the avalanche, where the authors assume deposits of the dense core. Comparing Figure 6b, showing the transects, and Figure 1b, showing visual photo of the avalanche, I would assume especially for L1 that this transect covered an erosion zone in the middle of the avalanche (while at the side L1 is indeed covering thin deposition). The higher temperatures in the middle of the avalanche could be those of the exposed warmer bed surface of the avalanche (warmer temperatures are deeper in the snowpack). Another observation makes me thinking that rather erosion areas were analysed in L1 and L2. Figure 1b shows that also at L3 and L4 there are areas of thin deposition at the sides of the transects. Why aren’t there differences in temperatures for those transects? I assume, because L3 and L4 covered deposition, while L1 and L2 covered erosion in the middle of the avalanche? Please ensure the reader that not an erosion zone was analysed in L1 and L2 in Figure 7. This makes the conclusion in lines 615ff problematic that deposits from the powder cloud has consistently lower temperatures than the warm dense core, as well as the statement in lines 600ff that the IRT can be used to “…could be the differentiation of flow regimes in the deposition area”.
2.2 Description of methods
It is still unclear to me (although requested by reviewer #2) how the released and entrained mass was determined, both for avalanche #3, for which the TLS scans are completely available, and for avalanche #2 with incomplete TLS scans. This is quite important since it is affecting the conclusion “entrainment energy varied between avalanches” (Eq. 4). Please add a description how the area of release vs. entrainment was determined. This area was maybe then used to calculate mass with depth and density from profiles and TLS data? The reader is forced to guess here. Please also add how the “averaged density” of the release and entrainment was determined. I assume with profiles and it is reasonable to assume which profiles, but I think this information could be added for more clarity
For avalanche #2 the scan before the avalanche is missing partly. The authors stated that the erosion was quite homogeneous, that’s why they could extrapolate to the missing parts. This would also include the assumption that the surface before the avalanche was quite homogeneous, since the scan before the avalanche is missing. Please discuss that there were for example no larger drifts present, or other rough areas. How was the extrapolation done afterwards? With the perimeter of the avalanche determined with GPS measurements? Please be a bit more precise here.
Both reviewer wanted more information on the BTS measurements, which was not added as the comments suggested. BTS measurements were interpolated for Figure 8. I know now that they took 3-5 min, but I am unsure if it is for one location with several measurements in depth, or indeed for one single temperature measurement. 3-5 min for the latter is certainly long enough. If the authors would have done 10 measurements in 3 min at different depth, I would be worried if the not mentioned Time Constant for the thermocouples was met. Please be clearer here and also provide information of the vertical resolution.
Please provide information on the horizontal resolution of the regular spaced pits which were used to determine the depth of the deposits.
Figure 8: It is not clear to me how the solid line showing the surface of the avalanche deposits, as well as the surface of the ground (dotted) was determined. With a combination of GPS measurements of the transect and TLS measurements?
The caption and the text for Figure 4 is confusing. I am unsure what was measured at the surface and what at the corresponding layer in the profile. The authors state in the text that the surface temperature was measured “along the erosion layer” in the avalanche path at P_track (line 192f), which I would interpret as the temperature of the bed surface of the avalanche. However, the grey colour in Figure 4, as well as the figure caption, that T_P_track was related to the surface of undisturbed snow next to the avalanche. Please be clearer here.
Also, why was not the corresponding layer temperature (not surface temperature) compared with the IRT of the bed surface at time 0 min similar to T_P_release? And why did the authors not include avalanche #3 in the verification? If there are no surface temperatures available, the authors have data from the profiles and the corresponding layers at the release and along the track for time 0 min.
In Figure 4 the IRT data is presented as continuous lines. What was the temporal resolution of taking pictures? I assume no video was made during one hour.
Please mention in the methods section for what the summer DEM data is used. Right now the summer DEM is in section 2.5 TLS, while in fact it is only used for IRT, to georeferenced the images and calculate the mass of the avalanches. Please also mention the source of the summer DEM.
In Figure 6 the IRT pixels of the transects were related to real distance using the width of the avalanche. I suggest to have pixel number on the x-axis and rather indicate for the width of the avalanche in the Figure, since there is obviously no clear relationship between pixel size and real distance.
There are some rounding issues using Eq. 2 and 3. I calculate 1.39 instead of 1.5 (line 435), and 0.96 instead of 0.9 (line 454).
2.3 Qualitative interpretation of IRT images
To my opinion the authors should be more careful with the qualitative interpretation of the IRT images. In the introduction the authors concluded based on their literature review that IRT can be seen as a useful qualitative tool for snow applications (line 86ff). But the before cited study of Schirmer and Jamieson showed the opposite. After only some seconds (!) they found relative warm areas in an exposed snow profile which were quite probably not warmer in the unexposed snowpack, with increasing relative differences in the next few minutes. This means that relative differences are not possible in such a scenario of exposing snow to a new environment. I would assume that an avalanche with warm particles in motion exposed to cold air may act in a very similar way as exposing a snow profile. In both cases the area of interest is certainly not in a thermal equilibrium with the surrounding atmosphere. This is also the main difference to cirrus clouds, which the authors refer to as a similar example, for which IRT is successfully used (line 596ff). I would suggest to change this formulation here in line 86ff, since I would say this cannot be concluded with cited literature.
In Figure 7 thin deposits have slight differences in temperatures (<1 degree) compared to deposits from the dense core (see also my comment above that I am unsure if there were really deposits covered by transects L1 and L2 in the middle of the avalanche). This is a qualitative usage, and I have a strong concern with this qualitative usage. Visible by eye in Figure 1b is that there was certainly a difference in surface roughness between the two areas. If this was the main reason why the IRT images showed differences, then I do not see a benefit to a visual picture. There needs to be some more investigations that there is more value in an IRT image compared to a visual image for distinguishing deposition areas. The IRT images were taken less than one minute after the avalanche stopped (line 167). The cooling speeds shown in Figure 4 are on a much longer time scale (several minutes to up to one hour). This is why I do not like the authors’ reply. I am interested in a much shorter time scale to trust Figure 7, which is the time between the deposition and the time of the image used for Figure 7. The authors correctly argued that the coarse resolution may not result in similar problems as in Schirmer and Jamieson (2014). The authors argued that over a footprint of 1 m there may be a more isotropic signal character. However, this is only a hypothesis, which can be easily tested with the existing data the authors have. One can argue using their argument, that there is a qualitative difference based on the small scale roughness how much more isotropic the signal character is. This is why there should be more done with existing data. As I understand, the authors made IRT videos of the moving avalanche, eventually stopped the videos and took pictures. So there is data available leading to the data shown in Figure 7 which can be analysed on different cooling speed based on roughness differences. No laboratory experiments are needed for this investigation. |
- Printer-friendly version
- Supplement
