This paper describes an interesting and well-conducted study on very small glaciers in the Northern Japanese Alps. Although these tiny snow/ice patches probably have a very limited relevance, their monitoring allows new insights into glaciological processes determining this glacier size class. The evaluation of data specific to these miniature glaciers in comparison with worldwide observations results in valuable conclusions. The paper is mostly clearly written and illustrated and fits well into The Cryosphere. However, when reading the manuscript, a relatively large amount of minor issues and questions came up. At several instances, wrong units and (maybe) wrongly stated results are present that require careful re-reading by the authors. I have two more important conceptual comments that should be addressed during the revision:

Substantive comments:

• Computation of mass balance profiles: The authors present mass balance profiles and elevation gradients for both winter and annual balance. Although they correctly mention and that the comparison of digital elevation models does not deliver local mass balance (allowing the computation of gradients etc.) and discuss emergence velocities of the ice, the conceptual approach remains partly vague and might be questionable for some situations: In fact, only for one of the five investigated glaciers, emergence velocities have been determined in the field. Moreover, the investigated locations only cover a limited elevation range, about at the median glacier elevation. More justification should be provided for the assumption that the measured emergence velocities are directly transferable to the other glaciers, and that the values measured are representative for the entire glacier. Conceptually, for typical alpine glaciers, emergence velocities are small around the median elevation, but increase in magnitude towards the top and the snout. If this also applies to the present situation on glaciers in the Japanese Alps, it might be that the measurements just captured the low emergence velocities at median elevation but the signal across the entire elevation range has been missed and mass balance profiles derived from surface elevation changes are thus biased. I do not consider this possibility as likely as the measured emergence velocities are much smaller than the mass balance rates, and the dynamics on the snow/ice patches is certainly different than on a standard glacier. Nevertheless, the issue needs to be looked at more closely, providing more justification for the assumptions.

• Converting snow/ice volume changes to mass change: The study is based on surface elevation changes that are subsequently converted to a mass change using a density assumption, both for winter and annual balance. In my opinion, there is a high potential for uncertainty that is barely described in the paper so far. The authors rely on some local observations of snow/firn/ice density and consider these values as universal, both regarding all glaciers and all years. This is certainly too much simplified. A variability in densities in the spatio-temporal domain is certain. It is clear that this cannot be measured but an additional uncertainty component that should be estimated based on as much evidence as possible is clearly needed. Furthermore, I am also partly doubtful regarding the chosen numbers and mentioned processes: (1) The density of winter snow corresponds to the observations at an off-glacier snow observatory. The authors mention that snow depth is half that on the glaciers, most likely resulting in smaller densities. Moreover, I would expect the winter snow depth on the glaciers to be particularly high because of a significant portion of wind drifted and avalanche snow. (2) The density of volume change is actually not only composed of the first annual layer’s density, as suggested by the authors, but strongly depends on compaction dynamics in older layers. Although it is argued that transition from snow to ice is occurring during a single year under this climate, more evidence supporting this claim is necessary in my opinion. If incompressible glacier ice (900 kg m-3) is formed during a single year, this should become evident in field observations and images after a year with mass loss, which has not been demonstrated. Figure 2 seems to indicate snow surface almost everywhere. Also, even if compaction of older layers was zero, the strong variability in annual mass balance should lead to differences in the density of volume change between negative (high density of lost material) and positive (lower density of gained material) years – an effect that is not considered at present. I realize that it will be difficult to resolve all these processes without any further observations (that would be difficult if not impossible to acquire). But the problem should be carefully discussed and be incorporated in a better uncertainty estimate of seasonal/annual mass balances. Uncertainties presently are too small in my opinion.

Detailed: comments:

• Line 23: Why no equilibrium line? The ELA was just either above or below the glacier.
• Line 31: only define acronyms if they are also used later in the paper.
• Line 58: “volume” and not “mass” change results from the geodetic method.
• Line 59: Given the multitude of recent studies relying on the geodetic method for glacier change assessment, I would suggest selecting newer and more appropriate references on the methodology.
• Line 64: It would make sense to mention already here that the geodetic method does not provide local mass balance, and thus mass balance profiles, but only elevation change. Given an estimate of spatially distributed (!) emergence velocities AND local density, this elevation change can be converted into a mass balance profile (see major comment above).
• Figure 1: Inset in map is not very clear. It would help to show entire Japan and use colour/shading for the sea.
• Figure 2: mention the year of the images in the caption.
• Table 1: The area stated in the last column is wrong (typo).
• Line 92: At least somewhere in a table, the exact dates of the measurements need to be given. Considering the high mass balance amplitudes, daily ablation/accumulation rates are important, thus also time differences of a few days will be affecting the stated seasonal/annual mass balances.
• Figure 3: Would be easier to read with a larger contour line spacing.
• Line 135: rho is not the ice density but the average density of the lost or gained material, including the effects of the compaction of lower firn/ice layers.
• Line 137: It is not strictly speaking the “stratigraphic system”! This system refers to the absolute maximum (winter) and minimum (autumn) of glacier mass. This date is normally unknown and can rarely exactly be met by monitoring programmes. I am quite sure that the surveys conducted do not correspond to this stratigraphic system (which is not a problem but needs to be stated). Regarding the stratigraphic system, referring to the much newer Cogley et al. 2011 publication would probably be more appropriate.
• Line 154: The choice of using the smallest glacier area over the study period for inferring total mass balance is interesting and should be discussed in more depth. In fact, reference-surface mass balances are then computed that may diverge quite a bit from the conventional mass balance, typically reported in glacier monitoring programmes, related to the actual glacier surface area. Both systems have their pro and cons. The mass gain or loss occurring beyond the minimum perimeter of the glacier could however also be determined and included in the computations. Or is there a reason that glacier extent has not been re-mapped every year?
• Line 159: Why is there only a 10m buffer around the glacier extent? The current glacier extent, or the minimum one? Chances are pretty high that there is remnant some snow close to the glacier in either the first or the second terrain model that would completely distort the correction process. Given that a terrain model can be generated for a bigger perimeter around the glaciers, why did the authors not use all of the stable terrain for this assessment? I would actually EXCLUDE the buffer zone. The text however clearly indicates that the comparison was done WITHIN the buffer zone. Furthermore, in Figure 4 it seems that the buffer is much bigger than the 10m stated. More explanation is needed here.
• Line 161: probably “winter balance error” is meant here.
• Line 169: Please use subscript for w and s after mass balance B (and not just Bw, Bs), always following the terminology proposed in Cogley et al., 2011.
• Line 208: Do you mean “in elevation” instead of “in slope”?
• Line 222: This uncertainty refers to the digital elevation models but does not account for the (important) uncertainty due to unknown density of volume change. This aspect also needs to be considered.
• Table 2: Caption needs to be extended and clarified: What exactly is shown? What are the units? Of course, I would be able to guess correctly but its better to be clear.
• Table 3: The table is too small to read and the headings are partly unclear and not fully explained: “altitude change” => “elevation change”; “correction” => “corrected”. Better give volume change in 10^3 or 10^6 m3.
• Table 4: State dates of surveys in this table as well. Summer balance should be negative and not positive! IMPORTANT: winter and summer balances (in m w.e.) are not adding up to the annual balances stated above (Table 3) and shown in Figure 9 and 10! I cannot track if this impression is due to my wrong understanding of the results shown or an error in the results presented. The difference is partly relevant and far beyond the stated uncertainty bars. This should definitely be looked at in detail during revision.
• Line 247: The analysis of mass balance amplitudes is interesting. However, the unit of all numbers is wrong! Numbers are given in mm w.e., instead of m w.e. as stated in the text and in the figures. I would consistently convert all numbers to m w.e. (i.e. divide them by 1000). Furthermore, the effect of the partly short series (just four years for the studied glaciers) should be discussed. How strongly do the extreme years (2015-2016, high loss; 2016-2017: high gain) affect the result? Are the four years statistically sufficient to draw a final conclusion?
• Line 261: Although I think the measurements likely do not capture the full range of actual emergence velocities, the analysis is well-conducted given the difficulties of direct field observations. Nevertheless, it remains unclear what has actually been done to derive mass balance profiles from the elevation changes. Which values for the emergence velocity have been applied, and how have they been extrapolated over the entire glacier surface, and to other glaciers? More details are needed.
• Line 308: Interesting observation. Any possible explanations?
• Line 313: Well, it is not the mass balance profile that has been measured but elevation change. Actually, typical alpine glaciers would show very similar profiles of elevation change over seasonal and annual periods (the latter only if their mass balance is close to zero)! However, accounting for emergence velocities leads to the reported mass balance profiles that are based on local measurements of mass balance. In that sense, I am still a bit reluctant to accept that emergence velocities are more or less zero throughout the entire elevation range and that mass balance profiles are flat on the investigated Japanese glaciers.
• Line 339: It appears to me that, according to the Abstract of that paper, the stated number is incorrect.
• Appendix: This very long table should rather go into a Supplementary Material and not an Appendix that is actually coming along in the same pdf as the paper. It is information that does not strictly need to be in paper, i.e. can also directly be obtained from the WGMS. Quickly state how the glaciers are ordered in the table. Units for winter and summer balance, as well as amplitude are wrong (mm w.e. instead of m w.e.).

Arie et al: comments by Dr Ian S. Evans, Durham University, for The Cryosphere, on -

Characteristics of mountain glaciers in the northern Japanese Alps.

The authors report very interesting observations on five Japanese glaciers, over several years.  These are clearly exceptional compared with other glaciers, which is attributed to avalanching and perhaps wind drifting, combined with very heavy snowfall.  They are narrow and linear.

All the Tables and Figures are clear and informative.  The writing is concise and the structure is acceptable.  The expression is readable but occasionally imprecise, with a strange use of ‘between’, and there are a few careless errors.

More context would be welcome, both on other possible glaciers among the “ More than 100 perennial snow patches “ in this region, and on comparison with other avalanche-fed glaciers and other glacierets / very small glaciers elsewhere. Can the feeding avalanche tracks be mapped?  Are there cornices in winter, suggestive of wind drifting?  Many aspects of the literature are well covered, but a number of other papers on glacierets or very small glaciers could be considered:

Bosson JB and Lambiel C (2016) Internal structure and current evolution of very small debris-covered glacier systems located in alpine permafrost environments. Frontiers of Earth Science 4(39), 1–17. doi: 10.3389/feart.2016.00039.

Capt M, Bosson JB, Fischer M, Micheletti N and Lambiel C (2016) Decadal evolution of a very small heavily debris-covered Glacier in an Alpine permafrost environment. Journal of Glaciology 62(233), 535–551. doi: 10.1017/jog.2016.56.

E Gachev, K Stoyanov - Present day small perennial firn-ice patches in the mountains of the western Balkan Peninsula. Stud Geomorphol. Carpatho-Balc., 2012 - igipz.pan.pl

EM Gachev - Small glaciers in the Dinaric Mountains after eight years of observation: On the verge of extinction? Acta geographica Slovenica, 2020 - ojs.zrc-sazu.si

E Gachev - Holocene glaciation in the mountains of Bulgaria. Mediterranean Geoscience Reviews, 2020 - Springer

E Gachev, I Mitkov Small glaciers in Pirin (Bulgaria) and Durmitor (Montenegro) as glacio-karstic features. Similarities and differences in their recent behaviour. - Quaternary International, 2019

Gachev E, Stoyanov K and Gikov A (2016) Small glaciers on the Balkan Peninsula: state and changes in the last several years. Quaternary International 415, 33–54. doi: 10.1016/j.quaint.2015.10.042.

Grudd H (1990) Small glaciers as sensitive indicators of climatic fluctuations. Geografiska Annaler Series A Physical Geography 72(1), 119–123. doi: 10.2307/521243.

Leigh JR, Stokes CR, Carr RJ, Evans IS, Andreassen LM, Evans DJA (2019). Identifying and mapping very small (<0.5 km2) mountain glaciers on coarse to high-resolution imagery. Journal of Glaciology 65(254), 873–888. https://doi.org/10.1017/jog.2019.50

Lindh L (1984) Studies on the transitional form between snowpatch and glacier in the Abisko Mountains, Sweedish Lappland. Svensk Geografisk Årsbok, 60, 145–156.

Nojarov P, Gachev E, Grunewald K (2019) Recent behavior and possible future evolution of the glacieret in the cirque Golemiya Kazan in the Pirin Mountains under conditions of climate warming. Journal of Mountain Science 16(1). https://doi.org/10.1007/s11629-018-4957-7

Woo MK and Young KL (2014) Disappearing semi-permanent snow in the High Arctic and its consequences. Journal of Glaciology 60(219), 192–200. doi: 10.3189/2014JoG13J150.

Some of these state that mass balances tend to vary from year to year, positive throughout or negative throughout, rather than between accumulation and ablation areas.

On the ‘Inventory of perennial snow patches…’, the Higuchi et al. GeoJournal paper is only 8 pages.  Might it be worth citing the fuller (81-page) Atlas ? –

Higuchi, K., Iozawa, T.:  Atlas of perennial snow patches in Central Japan. Water Res. Lab., Faculty of Science, Nagoya U., 81 pp., 1971.

Also consider -

Watanabe - Studies of snow accumulation and ablation on perennial snow patches in the mountains of Japan. Progress in Physical Geography, 1988

Minoru Yoshida, Katsuhiro Yamamoto, Kenji Higuchi, Hajime Iida, Tetsuo Ohata and Toshio Nakamura, First discovery of fossil ice of 1000–1700 year B.P. in Japan. Journal of Glaciology 36 (123), 1990 , pp. 258 – 259. DOI: https://doi.org/10.3189/S0022143000009527 [Kuranosuke glacier]

Apart from the detailed corrections, and extending comparisons with avalanche-fed glacierets elsewhere, my main suggestion is to compare the (1.92 to 4.34) ratio of winter mass balance to local weq snowfall, with similar ratios for the other nine glaciers smaller than 0.11 km² in the Appendix (or consider similar ratios for balance amplitude).

DETAILS:

‘likely’ should usually be replaced by ‘probably’.

Line 19  Why not ‘very small avalanche-fed glaciers’ ?  ‘Topographically controlled’ is a broader class, including effects of aspect (shade) and shadow, and shelter from wind drifting snow off a plateau (not in this rugged part of the Japanese Alps !).

69 presumably end-winter: best to give a date for this, here.

73-78 ‘glacial erosion valleys’ are usually termed ‘glacial troughs’.

82 ‘with little debris’

Table 1  Karamatsuzawa cannot be 1.03 km².  That contradicts the Introduction and all the maps.

Also, the average inclination is sometimes close to (altitude range)/ length, but not for Komado and Karamatsuzawa – i.e. it is not overall inclination.  Perhaps define how it was calculated.

95-96 ‘from an altitude range’

155-155 How can you be so precise, unless photos were taken daily?

158  ‘between   …  and …’   ??

Fig.4  Why does the ’10 m’ buffer zone extend much further, especially downglacier in c and d, and upglacier in a ?

Fig.5 caption  Not ‘Each circle represents’ – delete that.  Rather ‘The following numbers of glaciers in each region are included: …’  [ thereafter, repetition of ‘glaciers’ is unnecessary]

179 ‘amounts’

180 ‘accumulation increases with’

186-187  The sentence seems redundant, if you just mention ‘profile and gradient’ earlier.

219 should it be ‘within’ rather than ‘between’?

221 again unsure how ‘between’ is being used.

Fig.7 on left:  ‘slope’

230 & 234:  ‘5 to 13 m’, not 5 to 11 m.  (From Table 4: winter 5.63 to 12.72, summer 7.16 to 11.64.)  Per glacier, summer balances vary up to 1.6 m, so are not exactly constant, just with much smaller variation than for winter.

238 ‘nearly constant’ is an exaggeration.  I would say ‘is much less variable’

243 Add ‘It is reassuring that balances are closely correlated between the five glaciers, over this time period (Figs. 9 and 10).’

Table 3:  seems like ‘correction’ should read  ‘corrected’.  Caption should remind reader what the correction is for.  (The text does not even mention correction, except for lines 111-114 on the DSM: is that relevant, or is the correction for emergence velocity?) 

251-252.  Yes, but it is more logical to say that standard deviations increase linearly with amplitudes.

Fig.11  Not m ! Numbers on y axis are in mm, unless 000 is deleted throughout.

255-262  As emergence velocity has presumably been used above, should this section come earlier?  More explanation of how it was used is needed.

Fig. 11  provides an interesting comparison between regions, but is not the most relevant way of comparing mass balance data with the results in Japan.  Clearly glacier size is important: the larger the glacier, the less important the topographic effects including avalanching and wind drifting of snow.  As the Japanese glaciers are 0.11 km² or smaller, I suggest focussing on comparison with the small number of glaciers in the Appendix which are also of such tiny areas.  I think there are  nine: numbers 20, 21, 27, 51, 56, 74, 139, 151 and 180.  One is in New Zealand, three in the Andes, four in the Alps and one in Apennines (Calderone is not in the Alps.)  As the winter balances in Japan are 1.92 to 4.34 times the average direct snowfall (2.93 m), it would be useful to calculate similar ratios for the other very small glaciers.  I think that would reveal that glaciers e.g. in the Urals have similar ratios, from what has been termed ‘suralimentation’ (‘over-feeding’).  That is , the Japanese glaciers are exceptional (practically ‘outliers’) in absolute terms, but probably not in amplitude relative to direct snowfall.   (More accurate ratios could be calculated from annual snowfall values, rather than e.g. average snowfall at Tateyama Murododaira.)

Table 5:  ‘a^-1’

275 ‘from typical glaciers,’   - although I would rather say ‘from valley glaciers’ because many cirque glaciers also have much snow-drift and avalanche input.

293-2895  Indeed: the dependence of Ural glaciers on wind-drifting (from summit plateaus) and consequent avalanching was noted long ago by Dolgushin:

Dolgushin, L. D., 1961: Main features of the modern glaciation in the Urals. International Association of Hydrological Sciences Publication, 54: 335-347.

295 Not narrow valleys, but cirques.

298-299 delete one of the repeated ‘In addition,’s.

302  delete ‘likely’ ?

313 ‘vertical profiles’

315 ‘become negative throughout’  OR ‘Even if the glacier can become an ablation-area throughout’

325  It might be good to mention (somewhere) that (from the maps in Fig.3 and photos in Fig.2), the glaciers seem to receive avalanche snow throughout – not just at their upstream ends.  Also, the evidence for some wind-drifting effect (despite the lack of plateaus or even rounded summits – the ridges are rugged) comes from the eastward component of aspect of all five glaciers.

336  ‘tend to have annual balances strongly …’

349 Yes, very probably, but by how much ?

349 ‘2°K’

354 delete first ‘the’.

359 ‘very little’  - not ‘almost no’.

362 ‘of a typical’

364 ‘in Japan often had negative winter balance gradients’.   Fig. 13 does not really show negative annual balance gradients, and line 281 states “the gradients vary significantly”.  OK they did not have ELAs: that is because of the lack of positive annual mass balance gradients.

375 Appendix A:  Another 3-orders-of-magnitude error: clearly the amplitude for #1 cannot be 10.7 km.  The 4 columns ABw – SBn must be in mm, not m !

391  ‘fee’ ??

406 ‘Ödenwinkelkees, central Austria,’

429 ‘4(4), 303-311’

439 ‘The Cryosphere’

Incidentally, as a participant in ICG Excursion B7 in 2001 led by N. Matsuoka, with Kotaro Fukui, I was privileged to see the Kuranosuke glacier(et).  Although this was described as a ‘perennial snow patch’, and ‘no active glaciers exist in Japan’, distorted bedding structures in the ice seemed indicative of (possibly former) flow, thus fulfilling the definition of ‘glacier’ (WGI and Cogley et al. 2011: ‘showing evidence of past or present flow’).

-Ian S. Evans