Effects of surface roughness and light-absorbing impurities on glacier surface albedo, August-one ice cap, Qilian Mountains, China

Surface albedo is the main influence on surface melt for Qilian mountain glaciers. Fluctuations in surface albedo are due primarily to variations in micro scale surface roughness (ξ) and light-absorbing impurities (LAIs) in this region. However, combined ξ and LAIs effects over glacier surface albedo are rarely studied and surface roughness rarely considered in the albedo parameterization methods in this 15 region. The present study was conducted in tandem with an intensive photogrammetric survey of glacier surface roughness, LAIs samples and albedo observations along the main flow-line of August-one ice cap during the 2018 melt season. Automatic photogrammetry of surface roughness and automatic observation of glacier surface albedo was conducted at middle of the ice cap in 2018. Detailed analysis indicates a negative power function and positive linear relationship exist between ξ and albedo for snow and ice 20 surface, respectively. ξ could explain 68% of snow surface albedo and 38% of ice surface albedo variation in melt season. Effective LAIs concentration (Cξ) calculated by consider ξ effect over LAIs deposition account for more than 63% of albedo variation at ice surface. Using either ξ or Cξ to estimate ice surface albedo would be a great improvement over some current parameterization methods, such as assuming a constant mean ice surface albedo. A finer resolution of above 50mm and above 100mm is recommended 25 for ice and snow ξ calculations, which explain more albedo variation than coarse resolutions below it. With advances in topographic surveys to improve the resolution, extent and availability of topographic datasets and surface roughness, appropriate parameterizations of albedo based on ξ have exciting potential 1 https://doi.org/10.5194/tc-2020-67 Preprint. Discussion started: 7 April 2020 c © Author(s) 2020. CC BY 4.0 License.

photography(We carried out field observations of snow grain size and corresponding surface roughness based on you suggestion in July of 2020 in August-one ice cap). Especially the manual photogrammetry could give provide direct visual evidence of old snow have rougher surface over fresh snow. The revision of the English grammar are also done to make the lecture easier to read.
Specific comments : 1) The english has to be carefully corrected. Reviewer tried to highlight some errors in the 'technical corrections' section but there are many english mistakes making the reading difficult. Reply: Thanks for your comments. We have revised the English mistakes based on you 'technical corrections' suggestions and polished by language services. r2) Sections have to be read carefully to remove repetitions and to make the conclusion clearer. For example, it is often that in a 'snow section' authors speak about ice albedo measurements. Moreover, reviewer suggests to add small sentences at the end of each section to highlight the result. For example 'over ice cape, the albedo increases while the surface roughness increases and the LAI decreases'. Reply: Thanks for your suggestions. We have revised based on your suggestions, and add small sentences at the end of each section to highlight the result.
3) Explanation and results are insufficient: scatter plots presenting field observations have to be followed by physical analysis, according to what has been shown in the literature. Deeper explanations are needed to better understand the novelty of this work. Reply: We have add physical analysis according literature of similar research or relevant topics. This work not only considers surface layers snow or ice surface characteristics (surface roughness), but also considers LAI, and considers these two factors as important surrogate variables to explain more albedo variation than formerly used parameters. We have cited works such as Jonsell et al. (2003) who mentioned that uppermost surface ice layer, including its density, crystal structures, surface roughness, impurity content and stratification and SSA, are subject to continuous changes and mainly responsible for the observed variability in ice albedo. In this work we abstract the surface layers surface roughness might be an effective surrogate variable to be applied in albedo parameterizations. 4) Figures need to be well called in the text, it is hard to follow the analysis when the text does not refer to the right figure. Moreover, reviewer suggests small changes in the graphs to be clearer (in the 'technical corrections' section). For example, It could help the understanding if authors replot the Figure 3 with different symbols/colors associated to 1) measurements taken over snow cover 2) measurements taken over ice surface. Reply: we have revised accordingly, in the revised manuscript, we differentiate the manual and automatic observations, we also differentiate the snow and ice surface. We also include key meteorological air temperature and precipitation information in the revised manuscript. 5) As said in the 'General comment' section, it is not clear what relationships have been high lighted in this paper. There is a strong relationship between SZA/SSA/LAI/roughness/albedo. The impact of the solar zenith angle and SSA variations needs to be analysed deeper. Reply: Thanks for your excellent suggestions. We have add SSA, LAI, SZA effect in analysis. The impact of the solar zenith angle are mentioned in the observation part. In the August-one ice cap, the ice surface is not sensitive to zenith angle because of high concentration of LAI over surface and melt water. The snow surface is sensitive to zenith angle. For that reason, we calculated daily albedo from the integrated sum of incoming and outgoing shortwave radiation of only 10min albedo data taken when zenith angles were less than 60 o , in an effort to minimize solar zenith angle effects.
6) The Protocols/Measurements sections need more explanations. For example, the accuracy of the instruments should be precise, and references presenting the instruments are missing. Reviewer has some questions about the protocol performed to acquired albedo measurements: How do you estimate the direct/diffuse part of the albedo? Measurements are acquired in clear sky conditions? At which height was located the sensor (ie what is the area actually seen by the sensor) ? What if you measure the aluminium square in addition to the snow surface? Reply: we have detailed the observation in the revised manuscript. We do not differentiate direct and diffused part. The observation include cloudy and clear sky conditions. The sensor is 1m above surface for manual observations, and 1.5m above surface for automatic observations. The surface roughness is measured first. After then, the aluminum square is moved and albedo was measured. 7) There is a problem of unity in equation 2, epsilon is in cm while the h* variable is in meters. Reply: We have revised accordingly. 8) Nothing is new in the discussion part, except the parametrization method. Authors insist on the fact that it is physically-based, but it is not, it is fitted over the area so it is empirical. The sections should be better organized, this is very confusing. Moreover, the new parametrization should be investigated deeper, the associated error needs to be estimated for each type of surfaces (using control points over the snow covered surface, the ice surface). L63: Please put this sentence before, in the section above: 'As the snow melts, insoluble LAIs are retained at the snow surface, so concentrations of LAIs in surface snow increase with snow melt, further reducing snow albedo (Doherty et al., 2013).' Reply: Thanks for your suggestion, we have revised and delete this sentence here accordingly. We add this sentence in the before section. L75: Studies have indicates => Studies indicated Reply: Revised. L80: the distribution LAIs=> the distribution of LAIs Reply: Revised L85: Surface roughness structure => 'Surface roughness features' seems more appropriate. Reply: Revised L88: 'During the past 5 years the August-one ice cap has become darker due to the accumulation of LAIs' => Please add a reference.

Reply: we have revised the physically-based as surrogate variable. The validation of new parameterization method was not present because of lake snow and ice surface LAI observation in 2019.
Reply: Thanks for your suggestion, we have revised as' Successive time-lapse photography of the August-one ice cap has indicate that ice surface become darker due to the accumulation of LAIs in melting season (Qing et al., 2018).' L90: has indicate => has indicated Reply: revised L97: we try to investigate => we investigated Reply: revised L100: at different altitudes and times => and resolution? Or please add the resolution. => This section needs to be rewrite to present clearly the plan of your study. For example: 'first, to study the spatial variation.../second: : : Considering the following structure of your paper: 1rst objective: spatial variation, just the relationship between manual photo/lai/albedo 2nd objective: temporal variation. Reply: Thanks for your valuable suggestion, we have revised. This paragraph is revised as'' L100: Combine with => these measurements were combined with Reply: we have revised it. L108: based on => using only : : : or both : : : Reply: we have revised 'either' as 'only'. L118: It ranges in elevation => the elevation ranges from : : : Reply: Revised. L125. The glacier becomes darker with years? It was said in the introduction. Please clarify. Reply: we have 3 automatic time lapse photography site distributed over the August-one ice cap from terminal to top. The ice surface LAIs (most cryoconites) concentrated over ice surface in summer. For the last 6 years, no accumulation was observed even at top of the ice cap, the concentration of cryoconites has increased because of negative annual mass balance.

Figure 1 Cryoconites covered ice surface (grey part is sun dried cryoconite, brown part is wet cryoconite).
L134: Please specify if you are talking about 3D photogrammetric acquisitions or 2D photogrammetry, and the references need to be adapted (for instance, Manninen = 2D acquisitions with a board, Irvine-Fynn = 3D acquisitions) Reply: we have revised as '3-D photogrammetry' L139: of _1.75 m 2 => it is a very small area, is it representative ? How was the area chosen? Reply: The August-one ice cap is a small ice cap without obvious surface movement. The cap surface shows no relevant morphologies. Structures such as crevasses, cracks, and seracs are not formed. Only water channels formed over ice surface with very small proportion. In this study we selected all ice surface types except water channels. The selected snow or ice surface include smooth snow covered surface, partial snow covered surface, cryoconite holes, weathering crust and smooth ice surface. In the revised manuscript, we have add the selection standard for manual photogrammetry on 12 and 25 July, and 28 August.
L 139: Please clarify: 'by surrounding the target area of snow or ice surface', do you mean by turning around? If yes, what is the space between each step/picture? Reply: Thanks for your question, we have revised as' Seven to fifteen photos were taken at each survey site by surrounding the portable aluminum square. At each cardinal direction of aluminum square, we took 2 or 3 photos to cover the target area. Rough ice surface such as cryoconite holes need take additional photos close to nadir angle.' L143: Please specify the number of sampled areas. Reply: we have revised 'several' as '37' L147: 'Glacier surface albedo was calculated from measurements of up and downward shortwave radiation.' => how were acquired the up and downard shortwave radiation? Please add details about this protocol and sensor.
Reply: we have revised it as' After manual photogrammetry and removal of aluminum square, a Kipp & Zonen CMP11 albedometer consists of two identical pyranometers that measure the incoming global solar and the radiation reflected from the surface below. Albedo is the ratio of the two irradiances. Pyranometer is sensitive to radiation in the wavelength range 0.285 to 2.8μm. The albedometer was mounted on a camera tripod 1m above glacier surface and readings were made in a horizontal plane by means of a spirit bubble.'  L 195: 'from Lettau (1969) developed aerodynamic surface roughness' => 'from Lettau (1969) who developed aerodynamic surface roughness...' Reply: Done L198: adopted => adapted Reply: Done L202: Please add a reference to introduce this metric. Reply:Done L205: This sentence needs to be rewrite: For rough ice surface means more surface area and less concentration of LAIs over ice surface, I don't understand. Reply: we have revised. L207: by consider => by considering Reply: Done L207: this is a new metric? If yes, please change 'defined' by 'introduced' Reply: It is a new metric, we have revised 'defined' as ' introduced' accordingly. L220: 'For that reason, we used only half-hour albedo data taken when zenith angles were less than 60_' => Did you test the accuracy of the sensor considering the zenith angles? Is this a known angle of limitation? Or please add a reference.
Reply: we analyzed the the Kipp & Zonen CNR4 10-min up and downward shortwave radiation data. When the zenith angles are larger than 60 degree, the 10 min data shows great fluctuation especially for snow covered surface. For bare ice, especially when it was smooth and dirty, the zenith angle have very limited effect over albedo. Other studies also suggest measurement under L222: from 11:00 to 17:00 o'clock => what range of zenith angle it is? It strongly depends to the date of your measurements + elevation of the studied site. Please clarify. => It seems that this section should be written in the protocol of albedo measurements, not here in this 'result' section. Reply: we have revised this part and add in the '2.2 data collection' section L225: field investigation => is it based on your automatic photogrammeric measurements? Please clarify. Reply: We have revised 'field investigation' as 'manual observation'. The sentence has revised as' On July 12, manual photogrammetry alone the main flow-line of the August-once ice capindicated that at the ice cap terminals of 4600m' L230: Surface roughness => replace it by 'the associated epsilon measurements decreased ...' Reply: Thanks for your suggestion, we have revised as' The associated ξ measurements decreased decreased from 5.49±1.5cm to 0.5±0.6cm as altitude increased' L236: 'Surface roughness fluctuated between 1.4_0.4cm to 3.3_1.1cm; An increasing trend of surface roughness could be detected as altitude increased (Figure 3e). ' => please, inverse these two sentences to be clearer. Reply: Thanks for your suggestion, we have revised accordingly. L245: 'surface roughness and LAIs decreased as altitude increased.' => not clear because it is only the case in July. In August 3: surface roughness increased as altitude increased and there was some snow cover at the top of the ice cap. Please clarify. Reply: Thanks for your suggestions. We have revised this part based on your suggestions. L247: 'There was a much higher concentration of LAIs on the uncovered ice surface than snow surface' => I don't understand the 'much', the trend is not so strong? It could help the understanding if your replot the Figure 3 with different symbols/colors associated to 1) measurements taken over snow cover 2) measurements taken over ice surface. Reply: Thanks for your suggestions, we have revised Figure 3 based on your suggestions. L248: 'As a consequence, albedo tended to be low on the ice surface and higher on snow-covered surfaces.' => be careful with this sentence, this is also mainly due to SSA values that strongly impact the albedo. There is a strong relationship between SSA/LAI/roughness/albedo. Please rewrite this sentence. Reply: we have revised accordingly. L251: repetition with L240-243 => this section needs to be re-structured in order to remove repetitions and to be clearer on conclusions. Reply: we have rewrite this part and make it more clearer. L260: 'The automatic measurement setup in the middle of the ice cap' => 'The automatic measurement setup was in the middle of the ice cap' Reply: Done. L.264: 'Intermittent snowfall decreased surface roughness and increased albedo' => you should add that it is because there is fresh snow, and fresh snow has high SSA, inducing a high albedo. Reply: we have revised based on your suggestion, Thanks. Figure 4: It could be clearer if you add the intermittent snowfall with vertical lines + air temperatures (to know when it is melting). Reply: Thanks for your excellent suggestion, we have revised Figure 4 based on your suggestion. L265: 'When snow began to melt'=> when is it ? How do you measure that? Do you have air temperature or snow surface temperature? If yes, you could add it on Figure 4 (see comment above). Reply: we have meteorological observations at top of the ice cap including precipitation, surface temperature and air temperature. We have revised Figure 4 and add air temperature accordingly. L276: 'When the snow melted, leaving patchy snow cover, surface roughness increased and albedo decreased quickly to around 0.3_0.05 within two day'. Here again the sentence needs to be rewrite: snow melting = lower SSA = lower albedo, the decrease of albedo is not only due to the increase of surface roughness and LAI. Please clarify.