the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 08 Sep 2025
Experimental investigation of temperature-induced variations in the diffuse attenuation coefficient of young sea ice: depth and wavelength dependence
Abstract. The light attenuation properties of sea ice play a critical role in regulating the radiative energy budget and primary production in polar regions. Previous studies have shown that the diffuse attenuation coefficient of downwelling irradiance (Kd) of young, thin sea ice is sensitive to changes in ice temperature (Tice). However, the depth- and wavelength-dependent behavior of Kd under changing Tice conditions remains poorly understood. To address this gap, a comprehensive investigation was conducted using synchronized measurements of Kd and Tice from a cold laboratory experiment, a freezer experiment, and field observations in Liaodong Bay. The underlying mechanisms were further explored by analyzing the evolutions of sea ice microstructures and inherent optical properties, supported by Mie scattering theory and freezing equilibrium principles. Results revealed a distinct depth dependence: a negative correlation between Kd and Tice was found in 95.96 % of surface-layer measurements, while a positive correlation occurred in 38.38 % of bottom-layer measurements. The overall correlation between Kd and Tice was weaker in bottom layers compared to surface layers. This contrasting behavior is attributed to depth-dependent changes in the relative volume fractions of air and brine, which in turn affect scattering properties. No consistent trend in the Kd–Tice relationship was found across wavelengths; however, increased variability was occasionally observed in the 443–555 nm range. These findings enhance our understanding of the thermal sensitivity of light attenuation in sea ice and provide valuable insights for improving radiative transfer parameterizations in climate models and marine ecosystem simulations.
- Preprint
(2442 KB) - Metadata XML
- BibTeX
- EndNote
Status: closed
- RC1: 'Comment on egusphere-2025-2373', Anonymous Referee #1, 06 Oct 2025
-
RC2: 'Comment on egusphere-2025-2373', Anonymous Referee #2, 18 Oct 2025
Review of manuscript entitled “Experimental investigation of temperature-induced variations in the diffuse attenuation coefficient of young sea ice: depth and wavelength dependence” by Yu Zhang et al.
This manuscript describes a set of measurements designed to estimate the diffuse attenuation coefficient of young sea ice as a function of ice temperature. Two measurement realizations took place under laboratory conditions, the third took place under field conditions.
This manuscript addresses a question of interest to TC readership, but I have reservations about the methodology and related analysis.
I completely agree that the diffuse attenuation coefficient is a metric that is informative for characterizing the role of sea ice in a surface energy budget. But I wonder whether the domains studied in these experiments were optically thick enough to observe the diffusion domain? In the cold lab experiment, I question whether optical sensors at 3.5, 9.5, and 16.8 cm are far enough in optical-depth space from top or bottom (what’s the total ice thickness?) boundaries to observe a radiation field that is not attenuated by losses at the top or bottom surfaces. Same question applies to the freezer experiment with sensors at 10.9, 14, 17, 21.8 cm depth. Empirical measurement of diffuse attenuation coefficients require the assumption that the domain is horizontally infinite and that top/bottom boundaries are not a significant light sink. In fact, comment on lines 224 – 225 (Kd(412) showed multiple instances of negative value) suggests a finite-domain problem. In any half space (finite depth) domain, different behavior between an extinction coefficient near the top surface and an extinction coefficient near the bottom surface should be expected. Given this, I am not at all surprised that in line 267, the authors state that “surface layers typically showed negative correlations between Kd (λ) and Tice , while positive correlations were more often observed in the bottom ice layers.” It is possible that these relationships have everything to do with the finite ice thickness and failure to record the transport of light in an optically thick domain, and little to do with the physical properties of the ice being measured. And, in fact, the total ice thickness is not reported for any of the experiments-- which makes this extra challenging.
Did the ice grow all the way to bottom of vessel in the lab experiment and freezer experiment? If so, was that ice isothermal? Was the bottom boundary condition a dark surface or a bright surface? It is very difficult to assess the results of this experiment without this information.
Also, there is no information about the incident light sources—was illumination direct? Or diffuse? Or a mixture? How was stray light controlled?
Additionally, on line 202, the term “correlation coefficient (R )” is referenced. But Fig 5 legend states “slopes of linear regressions”. It is not possible to assess these results, if I don’t know what is plotted. I would suggest that, in this context, assessment of correlation coefficient has little physical meaning that can be related to the ice, other than whether it is positive or negative.
213: “For example, during the first warming phase (Fig. 5c)” – how would I know from the figure that the data shown in this plot are from a warming phase? The 6 phases (c – g) need to be denoted on the timeseries (fig 5a)—I can’t even tell which are warming, which are cooling. Also, it appears the data presented in Figure 5 were taken as the temperature changed—so this isn’t an equilibrium situation? But some intervals were constant temperature? Language says “warming” or “cooling” phases, suggesting there was no thermal equilibrium. The temperature curve on Fig 5a shows some of these intervals had small high frequency temperature cycling, but others had large cycling.
I don’t believe “Data available upon request” is an acceptable state for Data Availability with TC.
Technical/ minor comments
33: distribution has declined?
38, 40: disconnect: describing albedo reduction specifically related to reduced concentration/extent and then state that this albedo reduction allows more solar radiation to enter the ice, elevating ice temperature. This isn’t causal.
59: define ‘limited effect’. I think these references show significant effect!
64: more pronounced? I don’t know that those references actually say that?
100: how was the ice grown—only looking at figure 1 do I deduce the authors are talking about lab-grown sea ice. There is nothing in description here that specifies the ice growth—initial salinity? Temperature forcing? Insulation?
112: access holes exceeding 1 m diameter? An access hole that large could create a notable discontinuity in the horizontal homogeneity of the ice domain. How were the irradiance detectors positioned within the hole? Looking sideways? Or does the hole freeze around a probe tree? Fig 2d shows ‘frozen observation point’—but optically it looks quite different.
123: I don’t know what ‘stratified techniques’ means?
134: so samples warmed to just above 0 C prior to physical property characterization? Were the samples thermally stable?
142 – 152: was a_CDOM considered?
Fig 3 contamination from top/bottom boundaries. Side boundaries too. Expect bottom Kd to always be high since no light gets backscattered from below, so deepest Ed is always small.
Fig 5a: assume black markers represent ice temperature? Is the ice in the cold lab experiment frozen all the way to bottom of vessel, implying that there is no temperature gradient in this ice? Why some intervals have steady temperature (1 Dec – 14 Dec) and other intervals show short-interval temperature cycling (23-29 Dec, 16 – 20 Dec, 22 – 26 Dec)? Fig 5 b – g relate to which different temperature change intervals (not stated in figure caption, nor in text)?
199: “The results show that once sea ice formed (indicated by a stable Tice < -2 ℃)” All of the Tice are < -2 C.
202: “correlation coefficient (R ), but Fig 5 legend says “slopes of linear regressions”. Completely impossible to assess this if I don’t know what is plotted.
213: “For example, during the first warming phase (Fig. 5c)” – how would I know from the figure that the data shown in this plot are from a warming phase? The 6 phases (c – g) need to be denoted on the timeseries (fig 5a)—I can’t even tell which are warming, which are cooling. Also, it appears the data presented in Figure 5 were taken as the temperature changed—so this isn’t an equilibrium situation? But some intervals were constant temperature? Language says “warming” or “cooling” phases, suggesting there was no thermal equilibrium. The temperature curve on Fig 5a shows some of these intervals had small high frequency temperature cycling, but others had large cycling.
229 – 230: “Overall, Kd (555) exhibited a decreasing trend with increasing Tice and an increasing trend with decreasing Tice”. Trend and value are two different things.
244: “anomalous measurements… hard to account for” What does this mean?
271: were these counts of air and brine inclusions normalized to a standard volume?
288: absorption of solar radiation—so, this comment is applicable only to the natural ice sample? Otherwise, it’s not clear which experiment this discussion relates to.
289: “The increase in Vair contributes to the rise in Kd , thereby explaining the observed negative correlation between Kd and Tice”. The logic of this sentence is difficult to follow—increased Vair results from increased Tice, increasing Tice and increasing Kd would be a positive correlation.
307: “The newly formed air bubbles may subsequently migrate upward.” Are the authors referring to temperature-gradient migration? If so, it’s important to identify the temperature gradient in the ice.
Fig 9: The relationships presented in this figure appear to have very low correlation, so am having difficult time understanding how they back up and illustrate the discussion in lines 300 – 319.
Fig 10: What depths do Sample ID number correspond to? That needs to be stated in the caption.
389: “the sea ice remained frozen throughout our measurements and did not undergo a complete freeze–thaw cycle”. What does this mean? Sea ice (with temperature above about -35C) is always at its salinity determined melt/freeze point!
408: “transport”? I don’t see the relevance of transport.
Citation: https://doi.org/10.5194/egusphere-2025-2373-RC2
Status: closed
-
RC1: 'Comment on egusphere-2025-2373', Anonymous Referee #1, 06 Oct 2025
This study uses a freezer experiment and field observations in Liaodong Bay to investigate the depth- and wavelength-dependent behavior of Kdunder changing Tice. For now, the description and figures are not clear enough. There were many missing details, and the information in the figure titles was limited. This made it difficult for readers to understand. Please see my comments below.
Major comment
The authors list three experiments in the Results section. The results were clearly different, but the authors did not synthesize them. Firstly, the ice layer structure in the three experiments was different. This made the following comparisons questionable. Secondly, the spectral variation of Kd in the three experiments differs. Only Experiment 1 shows that Kd is wavelength-dependent. Thirdly, the depth dependence of Kd in Experiment 3 is different. These clear differences were not analyzed in the manuscript, so the depth- and wavelength-dependent behavior of Kd observed in this study is questionable.
It seems that only the ice sample from Experiment 3 was analyzed further in the Discussion Section. I didn’t see any reference to the other two experiments. Furthermore, most of the analysis in the Discussion section was not sufficiently robust. The corresponding comments can be found below.
Other comments
1 Introduction
The introduction did not explain why it is necessary to understand the wavelength and depth dependencies in the variations of Kd(λ) with Tice. For example, what can the difference in Kd between the surface and bottom layers of ice result in? Similar questions arise for Kd at different wavelengths.
2 Date
This part lacks some important information. Firstly, the thickness of the three observed ice was missing. This is more important for Experiment 3. Furthermore, the ice layer structure in three experiment was totally different. For example, the first layer of Experiment 1 was from 3.5-9.5 cm,whereas in Experiment 2 it was 11-14 cm. Is it reasonable to make a direct comparison? How the layer structure affects the following results? Secondly, it seems that the ice temperature in the laboratory experiments was different. However, I did not see the relevant description. Thirdly, the errors of each method should be shown. Some of the ice physical properties in Fig. 4 were very similar. The error of the method directly affects whether the result is credible.
2.3.3 In the calculation of absorption coefficient, I didn’t see the description of ap and acdom. Were both parameters observed?
2.3.4 How Equation 4 was obtained is unclear. Is it empirical? Furthermore, the ice microphotographs were obtained in Section 2.3.2. Thus, the distribution of inclusion numbers can be observed.
The optical size of the brine tube is not equivalent to its spherical radius. The latter value is larger than the former.
3 Result
It is not easy for readers to understand kd12 kd23…It will be better to name them as the surface layer or bottom layer. Furthermore, why do you also calculate kd13? This information has been shown in kd12 and kd23.
Each observation lasts for a long time. Which day and ice temperature is shown in Fig. 3?
Fig 4 How much is the error of temperature sensor? The difference in ice temperature between the first layer and last layer was no more than 0.5.
Fig 5 This figure is difficult to read. What are these six temperatures? Why did you choose these six temperatures?
4 Discussion
This section uses ice samples for analysis. However, it is not described where these samples were from. Were six samples from one experiment or three experiments? Why were these samples chosen? I suspect these samples were taken during observations of Experiment 3. This is because only this experiment has ice physical properties (Fig. 4). This is the main shortcoming of the manuscript. The results from the three experiments were not the same.
L274 Why does this part use a completely different layer structure? This means that this section is not relevant to Section 3.
L276 In general, the uncertainty of gravimetric volume method for Va was 100–200%, while that of Vb was 20–30%. The present Va of the three ice layers was between 8% and 5%. Considering the uncertainty of the density measurement method, this difference is so small. Therefore, the Va value is not reliable.
Figure 8a is also not credible. As described in Section 2.3.2, the ice microstructure was obtained using thin ice sections. Therefore, the different thicknesses of the thin ice samples directly result in different inclusion numbers.
I don’t understand what Fig. 9 is trying to show. Clearly, there was no significant correlation between Va and Vb in your ice sample.
L322 “the scattering coefficient of different ice layers were calculated using Mie theory”. Mie theory is used to calculate the scattering efficiency, which was set to 2 for the sake of simplicity in Section 2.3.4. So, the following calculation is not relevant to Mie theory.
L323 Again, this section uses a different layer structure to those in L274 and Section 3.
The whole section 4.2 is confusing. The Kd is the combined effect of ice scattering and absorption. It is unreasonable to compare the Kd and the scattering coefficient directly.
L330 What is “freezing equilibrium model and the equivalent sphere theory”
L331 “In the surface layer, the observed negative correlation between Kd and Tice implies that both σ and g increase with Tice”. How do you get the g of ice?
Fig 10 Which experiment does it come from? The results shown in Fig. 10 appear to contradict those shown in Figs. 8b and 8c. Sample 2 has small Va and Vb, but a large scattering coefficient. Sample 3 has larger Va and Vb than Sample 2, but a smaller scattering coefficient.
L351 “The observed variations in Kd with Tice exhibited wavelength dependence especially in the shortwave range for cold laboratory experiment.” Yes, but in the other two experiments, there was no wavelength dependence. So, it is still doubtful.
Citation: https://doi.org/10.5194/egusphere-2025-2373-RC1 -
RC2: 'Comment on egusphere-2025-2373', Anonymous Referee #2, 18 Oct 2025
Review of manuscript entitled “Experimental investigation of temperature-induced variations in the diffuse attenuation coefficient of young sea ice: depth and wavelength dependence” by Yu Zhang et al.
This manuscript describes a set of measurements designed to estimate the diffuse attenuation coefficient of young sea ice as a function of ice temperature. Two measurement realizations took place under laboratory conditions, the third took place under field conditions.
This manuscript addresses a question of interest to TC readership, but I have reservations about the methodology and related analysis.
I completely agree that the diffuse attenuation coefficient is a metric that is informative for characterizing the role of sea ice in a surface energy budget. But I wonder whether the domains studied in these experiments were optically thick enough to observe the diffusion domain? In the cold lab experiment, I question whether optical sensors at 3.5, 9.5, and 16.8 cm are far enough in optical-depth space from top or bottom (what’s the total ice thickness?) boundaries to observe a radiation field that is not attenuated by losses at the top or bottom surfaces. Same question applies to the freezer experiment with sensors at 10.9, 14, 17, 21.8 cm depth. Empirical measurement of diffuse attenuation coefficients require the assumption that the domain is horizontally infinite and that top/bottom boundaries are not a significant light sink. In fact, comment on lines 224 – 225 (Kd(412) showed multiple instances of negative value) suggests a finite-domain problem. In any half space (finite depth) domain, different behavior between an extinction coefficient near the top surface and an extinction coefficient near the bottom surface should be expected. Given this, I am not at all surprised that in line 267, the authors state that “surface layers typically showed negative correlations between Kd (λ) and Tice , while positive correlations were more often observed in the bottom ice layers.” It is possible that these relationships have everything to do with the finite ice thickness and failure to record the transport of light in an optically thick domain, and little to do with the physical properties of the ice being measured. And, in fact, the total ice thickness is not reported for any of the experiments-- which makes this extra challenging.
Did the ice grow all the way to bottom of vessel in the lab experiment and freezer experiment? If so, was that ice isothermal? Was the bottom boundary condition a dark surface or a bright surface? It is very difficult to assess the results of this experiment without this information.
Also, there is no information about the incident light sources—was illumination direct? Or diffuse? Or a mixture? How was stray light controlled?
Additionally, on line 202, the term “correlation coefficient (R )” is referenced. But Fig 5 legend states “slopes of linear regressions”. It is not possible to assess these results, if I don’t know what is plotted. I would suggest that, in this context, assessment of correlation coefficient has little physical meaning that can be related to the ice, other than whether it is positive or negative.
213: “For example, during the first warming phase (Fig. 5c)” – how would I know from the figure that the data shown in this plot are from a warming phase? The 6 phases (c – g) need to be denoted on the timeseries (fig 5a)—I can’t even tell which are warming, which are cooling. Also, it appears the data presented in Figure 5 were taken as the temperature changed—so this isn’t an equilibrium situation? But some intervals were constant temperature? Language says “warming” or “cooling” phases, suggesting there was no thermal equilibrium. The temperature curve on Fig 5a shows some of these intervals had small high frequency temperature cycling, but others had large cycling.
I don’t believe “Data available upon request” is an acceptable state for Data Availability with TC.
Technical/ minor comments
33: distribution has declined?
38, 40: disconnect: describing albedo reduction specifically related to reduced concentration/extent and then state that this albedo reduction allows more solar radiation to enter the ice, elevating ice temperature. This isn’t causal.
59: define ‘limited effect’. I think these references show significant effect!
64: more pronounced? I don’t know that those references actually say that?
100: how was the ice grown—only looking at figure 1 do I deduce the authors are talking about lab-grown sea ice. There is nothing in description here that specifies the ice growth—initial salinity? Temperature forcing? Insulation?
112: access holes exceeding 1 m diameter? An access hole that large could create a notable discontinuity in the horizontal homogeneity of the ice domain. How were the irradiance detectors positioned within the hole? Looking sideways? Or does the hole freeze around a probe tree? Fig 2d shows ‘frozen observation point’—but optically it looks quite different.
123: I don’t know what ‘stratified techniques’ means?
134: so samples warmed to just above 0 C prior to physical property characterization? Were the samples thermally stable?
142 – 152: was a_CDOM considered?
Fig 3 contamination from top/bottom boundaries. Side boundaries too. Expect bottom Kd to always be high since no light gets backscattered from below, so deepest Ed is always small.
Fig 5a: assume black markers represent ice temperature? Is the ice in the cold lab experiment frozen all the way to bottom of vessel, implying that there is no temperature gradient in this ice? Why some intervals have steady temperature (1 Dec – 14 Dec) and other intervals show short-interval temperature cycling (23-29 Dec, 16 – 20 Dec, 22 – 26 Dec)? Fig 5 b – g relate to which different temperature change intervals (not stated in figure caption, nor in text)?
199: “The results show that once sea ice formed (indicated by a stable Tice < -2 ℃)” All of the Tice are < -2 C.
202: “correlation coefficient (R ), but Fig 5 legend says “slopes of linear regressions”. Completely impossible to assess this if I don’t know what is plotted.
213: “For example, during the first warming phase (Fig. 5c)” – how would I know from the figure that the data shown in this plot are from a warming phase? The 6 phases (c – g) need to be denoted on the timeseries (fig 5a)—I can’t even tell which are warming, which are cooling. Also, it appears the data presented in Figure 5 were taken as the temperature changed—so this isn’t an equilibrium situation? But some intervals were constant temperature? Language says “warming” or “cooling” phases, suggesting there was no thermal equilibrium. The temperature curve on Fig 5a shows some of these intervals had small high frequency temperature cycling, but others had large cycling.
229 – 230: “Overall, Kd (555) exhibited a decreasing trend with increasing Tice and an increasing trend with decreasing Tice”. Trend and value are two different things.
244: “anomalous measurements… hard to account for” What does this mean?
271: were these counts of air and brine inclusions normalized to a standard volume?
288: absorption of solar radiation—so, this comment is applicable only to the natural ice sample? Otherwise, it’s not clear which experiment this discussion relates to.
289: “The increase in Vair contributes to the rise in Kd , thereby explaining the observed negative correlation between Kd and Tice”. The logic of this sentence is difficult to follow—increased Vair results from increased Tice, increasing Tice and increasing Kd would be a positive correlation.
307: “The newly formed air bubbles may subsequently migrate upward.” Are the authors referring to temperature-gradient migration? If so, it’s important to identify the temperature gradient in the ice.
Fig 9: The relationships presented in this figure appear to have very low correlation, so am having difficult time understanding how they back up and illustrate the discussion in lines 300 – 319.
Fig 10: What depths do Sample ID number correspond to? That needs to be stated in the caption.
389: “the sea ice remained frozen throughout our measurements and did not undergo a complete freeze–thaw cycle”. What does this mean? Sea ice (with temperature above about -35C) is always at its salinity determined melt/freeze point!
408: “transport”? I don’t see the relevance of transport.
Citation: https://doi.org/10.5194/egusphere-2025-2373-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 1,405 | 35 | 12 | 1,452 | 23 | 26 |
- HTML: 1,405
- PDF: 35
- XML: 12
- Total: 1,452
- BibTeX: 23
- EndNote: 26
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This study uses a freezer experiment and field observations in Liaodong Bay to investigate the depth- and wavelength-dependent behavior of Kdunder changing Tice. For now, the description and figures are not clear enough. There were many missing details, and the information in the figure titles was limited. This made it difficult for readers to understand. Please see my comments below.
Major comment
The authors list three experiments in the Results section. The results were clearly different, but the authors did not synthesize them. Firstly, the ice layer structure in the three experiments was different. This made the following comparisons questionable. Secondly, the spectral variation of Kd in the three experiments differs. Only Experiment 1 shows that Kd is wavelength-dependent. Thirdly, the depth dependence of Kd in Experiment 3 is different. These clear differences were not analyzed in the manuscript, so the depth- and wavelength-dependent behavior of Kd observed in this study is questionable.
It seems that only the ice sample from Experiment 3 was analyzed further in the Discussion Section. I didn’t see any reference to the other two experiments. Furthermore, most of the analysis in the Discussion section was not sufficiently robust. The corresponding comments can be found below.
Other comments
1 Introduction
The introduction did not explain why it is necessary to understand the wavelength and depth dependencies in the variations of Kd(λ) with Tice. For example, what can the difference in Kd between the surface and bottom layers of ice result in? Similar questions arise for Kd at different wavelengths.
2 Date
This part lacks some important information. Firstly, the thickness of the three observed ice was missing. This is more important for Experiment 3. Furthermore, the ice layer structure in three experiment was totally different. For example, the first layer of Experiment 1 was from 3.5-9.5 cm,whereas in Experiment 2 it was 11-14 cm. Is it reasonable to make a direct comparison? How the layer structure affects the following results? Secondly, it seems that the ice temperature in the laboratory experiments was different. However, I did not see the relevant description. Thirdly, the errors of each method should be shown. Some of the ice physical properties in Fig. 4 were very similar. The error of the method directly affects whether the result is credible.
2.3.3 In the calculation of absorption coefficient, I didn’t see the description of ap and acdom. Were both parameters observed?
2.3.4 How Equation 4 was obtained is unclear. Is it empirical? Furthermore, the ice microphotographs were obtained in Section 2.3.2. Thus, the distribution of inclusion numbers can be observed.
The optical size of the brine tube is not equivalent to its spherical radius. The latter value is larger than the former.
3 Result
It is not easy for readers to understand kd12 kd23…It will be better to name them as the surface layer or bottom layer. Furthermore, why do you also calculate kd13? This information has been shown in kd12 and kd23.
Each observation lasts for a long time. Which day and ice temperature is shown in Fig. 3?
Fig 4 How much is the error of temperature sensor? The difference in ice temperature between the first layer and last layer was no more than 0.5.
Fig 5 This figure is difficult to read. What are these six temperatures? Why did you choose these six temperatures?
4 Discussion
This section uses ice samples for analysis. However, it is not described where these samples were from. Were six samples from one experiment or three experiments? Why were these samples chosen? I suspect these samples were taken during observations of Experiment 3. This is because only this experiment has ice physical properties (Fig. 4). This is the main shortcoming of the manuscript. The results from the three experiments were not the same.
L274 Why does this part use a completely different layer structure? This means that this section is not relevant to Section 3.
L276 In general, the uncertainty of gravimetric volume method for Va was 100–200%, while that of Vb was 20–30%. The present Va of the three ice layers was between 8% and 5%. Considering the uncertainty of the density measurement method, this difference is so small. Therefore, the Va value is not reliable.
Figure 8a is also not credible. As described in Section 2.3.2, the ice microstructure was obtained using thin ice sections. Therefore, the different thicknesses of the thin ice samples directly result in different inclusion numbers.
I don’t understand what Fig. 9 is trying to show. Clearly, there was no significant correlation between Va and Vb in your ice sample.
L322 “the scattering coefficient of different ice layers were calculated using Mie theory”. Mie theory is used to calculate the scattering efficiency, which was set to 2 for the sake of simplicity in Section 2.3.4. So, the following calculation is not relevant to Mie theory.
L323 Again, this section uses a different layer structure to those in L274 and Section 3.
The whole section 4.2 is confusing. The Kd is the combined effect of ice scattering and absorption. It is unreasonable to compare the Kd and the scattering coefficient directly.
L330 What is “freezing equilibrium model and the equivalent sphere theory”
L331 “In the surface layer, the observed negative correlation between Kd and Tice implies that both σ and g increase with Tice”. How do you get the g of ice?
Fig 10 Which experiment does it come from? The results shown in Fig. 10 appear to contradict those shown in Figs. 8b and 8c. Sample 2 has small Va and Vb, but a large scattering coefficient. Sample 3 has larger Va and Vb than Sample 2, but a smaller scattering coefficient.
L351 “The observed variations in Kd with Tice exhibited wavelength dependence especially in the shortwave range for cold laboratory experiment.” Yes, but in the other two experiments, there was no wavelength dependence. So, it is still doubtful.