In situ observed relationships between snow and ice surface skin temperatures and 2 m air temperatures in the Arctic \\
Pia Nielsen-Englyst, Jacob L. Høyer, Kristine S. Madsen, Gorm Dybkjær, Rasmus Tonboe, Gorm Dybkjær, Emy Alerskans \\
\section*{Summary}
The authors have done a good job in improving the manuscript. It is much more focussed and more clear what the intention is.\\
However, although I do acknowledge the importance of understanding the difference between the skin and 2m temperature for the interpretation of satellite products, the majority of the insights presented in this manuscript are still not new. What this manuscript does is bringing these insights together and add an quantitative analyses of the impact of clouds. In that respect it does have merit. But I still miss a clear recommendation. This is not really a statistical analyses. It shows processes and features, but how to apply this knowledge, no hint is given. Based on this paper my conclusion is that you should not interpret the satellite Tskin in terms of T2m because there is a strongly varying surface temperature inversion. This conclusion is not made, nor is a method given of how to apply this so you actually can interpret Tskin in terms of T2m. The added value of this manuscript is therefore still limited. \\
Furthermore, there are other points that need to be addressed before publication can be considered.\\ \\
The way the surface energy balance is described in Section 3 is still incorrect/unclear. Since you discuss the skin temperature, I assume that you describe the surface energy balance of a skin layer, which is an infinitesimal thin layer without heat capacity. In that case there is an instantaneous balance between the different fluxes, and for each surface that does not change phase M is per definition 0. The warming or cooling of the medium below the surface affects the surface temperature through G. That also includes the effect of latent heat release when refreezing occurs. This affects the temperature of the medium and with that the temperature gradient close to the surface and thus the conductive heat flux at the surface. M can only be non zero when a phase change is possible. In that case the medium is usually almost isothermal and cannot raise in temperature any further being limited to the melting temperature. When the medium is isothermal G is basically 0. The excess heat results in M being positive, indicating melt occurs. M thus cannot be negative.\\
This is the usual way the surface energy balance is described in this field and in models. This makes it also more transparant to understand since there is a clear difference between M and the ground heat flux G. In your description you seem to mix them up, and that is only possible when the surface layer you describe has a finite thickness and a heat capacity. You have to rephrase this, preferably using the skin layer formulation since that is what you are studying. Furthermore, in your description it is sometimes unclear whether you are looking at the surface (the interface between atmosphere and snow/ice/land) or whether you are looking at the medium below the surface.\\
Furthermore, although the English is not bad, there are numerous statements that are not clear, not specific enough or raise questions, in addition to numerous small mistakes or typos. I have tried to mark them below, but since I am no native speaker as well, I recommend that this manuscript is checked for English language as well before publication.
\noindent
Below follows a list of more specific points.
\section*{Specific comments}
P1 L15: Rephrase: 'makes the surface colder than' into 'cools the surface with respect to'\\
P1 L16: Remove: 'often'\\
P1 L16-17: Remove: 'and the two.... certain conditions'. Is not specific enough, you specify when in the next sentence.\\
P1 L18: I guess that with 'best agreement' you mean with difference of less than 0.5 $^{\circ}$C?\\
P1 L18: How often is T2m $<$ Tskin? Can you give an estimation in \% of time for example?\\
P1 L19: Rephrase 'when it is cold' , this is a subjective statement, negative radiative balance with a non melting surface, or temperatures well below 0.\\
P1 L22: Add 'mean overcast T2m-Tskin difference'\\
P1 L24: Is cloud limited the same as clear sky effect (L25)?\\
P1 L25: Replace 'assessed' by 'tested' or studied. (in short space 3 times assessed)\\
P1 L25: Replace 'The clear sky effect has been assessed' by 'To this end we test three different...'\\
P1 L28: Why are the smallest biasses found during summer?\\
\noindent
P2 L6-7: Would be nice if you could mention more than one mechanism responsible for the amplification.\\
P2 L7: Insert 'surface' before 'mass balance', or remove 'atmospheric' before 'warming'. This way you include oceanic warming as well and its impact on the tidewater glaciers.\\
P2 L8: Add that the other part is the result of increased calving rates.\\
P2 L14-15: Add 'increase in projected' between 'the' and 'surface air temperature'\\
P2 L15 remains a vague statement: Add at least one reason how it may contribute. Intensification of weather systems? Change of flow patterns?\\
P2 L20: Replace 'not available everywhere' by 'rare' and add 'available' just before 'time series'\\
P2 L21: with 'ice regions', do you mean land ice and sea ice regions combined?\\
P2 L23: Replace 'this means that' with 'consequently' \\
P2 L24-25: remove 'due to .... system'. Too much repetition.\\
P2 L28: Add 'clear sky' before 'surface temperature' and 'all sky' before '2 m air temperature'.\\
P2 L30: Add 'and the role of clouds on this relationship' just before 'as we do here'.\\
P2 L32: Formulate more explicite: replace 'an imbalance between the radiative fluxes' with 'a negative net radiative balance'. And replace 'especially' by 'this mostly occurs'\\
P3 L1: remove: 'the inversion continues al the way to the surface' This is a strange statement since you already explained that the inversion is forced at the surface. \\
P3 L2-3: Rephrase: The surface-drive.... snow/air interface. the temperature inversion does not cause the temperature difference but is the temperature difference and the skin temperature is actually the temperature of an infinitesimal thin layer without heat capacity, in this case the snow surface at interface with the air. \\
P3 L5: the dominating factors? where does this refer to? dominating factors in what?\\
P3 L13: First mention of 'differences' refers to differences between what?\\
P3 L14: Explain for what the T2m in the coupled model was corrected for by using Tskin?\\
P3 L16-19: Nicely formulated. \\
P3 L23: Replace: 'an effort has also been made' by 'we'.\\
\noindent
P4 L2: I guess you refer to calendar years, not mass balance years? And how do you handle the data which do not cover full years? PROMICE is full years data but ARM, ICEARC, SHEBA, TARA are not.\\
P4 L4: remove Tskin, or reformulate: Tskin observations are derived from the long wave radiation. You do not have both parameters as observables.\\
P4 L29: Same question as before: an albedo of 0.3 is already low. In case of an albedo of 0.3, is the surface still fully snow covered? Or is this partly snow covered? My guess is that this might also represent cases with partly snow cover. \\
P4 L22: Related to the ARM sites, how long is the period of observations per year that you use? Biased to winter spring? Or otherwise?\\
\noindent
P5 L6: In this case you only have winter/spring data. How do you handle the averaging in this case? Period averages per calendar year? How does this affect the results?\\
P5 L19: Where does ASFG stand for?\\
P5 L20: Rephrase: 'Five different levels... anemometer.' The mast contains five different levels varying in height..., on which temperature/humidity probes and a sonic anemometers are mounted. \\
P5 L23-25: Rephrase: 'Three surface....2007' e.g. 'Three different methods to measure surface temperature were deployed: a General.....radiometer, for which data is available over the period April to September2007'.\\
P5 L25: How do you handle the differences in available periods when you average?\\
P5 L27-28: Rephrase: first state which measurements you use most of the time, then what you use instead in case these are unreliable. e.g. 'which is based on Epply observations, and in cases where epply was known to be wrong.....'\\
\noindent
P6 L4-5: Rephrase: Remove reference to the hovercraft here, provide information on the length of the observational period and distance, and how they measured air and surface temperature from the hovercraft. I guess the weather station was installed on the hovercraft? \\
P6 L7: Rephrase, obvious that temperature is measured with a temperature sensor.\\
P6 L9: Reprase: 'Build to withstand ...sea ice' Do you mean an ice reenforced ship or an ice breaker?\\
P6 L11: Remove 'a' befor 'part'.\\
P6 L14: Replace 'had' with 'deployed'.\\
P6 L16: How do you handle the different length observational periods in the averages?\\
P6 L22-26: Rephrase: I am not sure I understand what you mean here. I guess the skin temperature you derive using these different sensors will differ, because of the different spectral range they measure in, not covering the whole spectrum in which the surface emits. And you have to correct for that. But what do you mean by 'sky temperature which is reflected'? Why is the sky temperature of importance? You are looking at the surface. My suggestion is to start with a sentence about the different emissivities and spectral ranges resulting in different Tskin. Then state that reflection at the surface of the radiation emitted by the sky affect the observations. And finally explain that the combination of the generally lower temperatures of the sky combined with the lower emissivities of the sky compared to the surface make this a small effect and that you will neglect it.\\
P6 L26-27: Do I understand correctly that they looked for emissivities that resulted in the best correspondence of both types of observations?\\
P6 L22 - P7 L4: I find this part a bit chaotic. Please try to focus this more, with less repetition of statements.\\
\noindent
P7 L1: You present the difference between the two sensors. But which one represents the total surface temperature best? Or is that the third broad band sensor? And how do you translate from a smaller band to a borad band? Or is the point that they provide already reasonable values of Tskin? How do you know what the 'truth' is?\\
P7 L5 Replace 'with' by 'of'. Add behind DMI_Q, showing a good correlation (value) and a small bias(value) when comparing these methods.\\
P7 L5-6: Remove: 'There is a.....comparison'. This sentence is unclear.\\
P7 L7: Replace 'if' by 'is'.\\
P7 L11: You changed the name of the subsection, but not the text of the section itself. Add Long-wave equivalent befor cloud fraction.\\
P7 L15: It is still not clear to me what you apply on all stations, and what is station dependent. As far as I understand from your description you apply this equation to all sites and you also apply LWD_cloudy = sigma T^4 to all sites. Thus no station dependent relations for the upper or lower limit of the LWd T2m relation. Kuipers Munneke et al., Int J. Climatol., 2011. describe how you can derive for each station a polynomial function that follows the upper and lower limit of LWd as a function of T2m. This way the functions better represent the observations. Then you determine for every single observation pair LWd T2m the CCF, as you describe. It must be clear in you description what you apply the same way to all stations, and what is station dependent. The way described by Kuipers Munneke is best. But you can also discuss how sensitive the results are for this choice.\\
P7 L23: To what does characteristics refer to? Characteristics of what?\\
P7 L24: remove 'net' in front of 'surface energy balance' (also later in the section)\\
P7 L24-25: rephrase 'between the atmosphere..... ocean'. e.g. at the interface of the atmosphere with the snow, ice, land or ocean surface.\\
P7 L25: Add 'as' behind 'written'.\\
P7 L26: Replace 'upwelling' by 'reflected'\\
P7 L28: Add 'defined' between 'are' and 'positive'.\\
P7 L29 - P8 L2: See comment above about the definition of the surface energy balance.\\
\noindent
P8 L4-5: This sentence is not clear. What does 'short wave radiation input refer to? The incoming short wave radiation or the net short wave radiation? It is most fair to compare the net short wave with the net long wave flux. Also note that Maykut 1986 describe sea ice conditions, not land ice, nor seasonally snow covered regions. Especially the importance of clouds can differ considerably. \\
P8 L7 (twice): Replace 'non-radiative' by 'turbulent heat' since you do not intent the ground heat flux here. \\
P8 L7: Add before 'On average,' The latter is related to the fact that ....'\\
P8 L8: Replace '. However, because ' by 'and since', and replace 'balance' by flux\\
P8 L10-11: Rephrase this sentence, e.g. 'Note that the surface energy fluxes are strongly related to the surface winds as the turbulent mixing is a function of wind speed.' The link with the previous bit of the paragraph is evident this way.\\
P8 L11-12: Rephrase, this is too simple and not clear. Are you only discussing winter conditions here? In general in the arctic during winter SWD is negligible, irrespective of cloudy or clear sky conditions. I guess you mean: Under clear sky conditions, when SWD is negligible, LWU... (remove the reference to winter, it is SWd that is the objective factor, not time of year) Furthermore, add that this results in a negative radiative balance cooling the surface and driving a positive sensible heat flux. \\
P8 L 13-15: Rephrase, you first mention stable stratification, and then you explain that the surface temperature is lower than the air temperature, Better the other way around. Furthermore, you repeat the reference to stable stratification in L16.\\
P8 L20-21, I prefere consistency within this paper than with other papers. Since katabatic or inversion winds are the same, use only one term in this manuscript and perhaps mention that for a certain region, the other term is often used.\\
P8 L22: remove 'both'.\\
P8 L26: Replace 'thus the' by 'reduces the'\\
P8 L30-31: Instead of presenting this as length of (polar) night and day, more objective to present this in terms of available incoming shortwave radiation. \\
P8 L31-32: Remove 'The temporal...scales.' This sentence more or less repeats the previous sentence. In both sentences you mention variability on different time scales without mentioning what time scales.\\
\noindent
P9 L9: I am very surprised that the coldest month is December. Usually the coldest month is February/March, and this also seems the case for most of your sites when looking at Figure 5a. Please check if this is correct, and if the difference with February/March is statistical significant. If not, you cannot make this statement. You also state 'not shown', but you do show this in figure 5. Furthermore, in this sentence and the following, do you refer to KAN_U or all stations?\\
P9 L13-14: Isn't the larger variability also due to the larger Pole equator temperature difference in winter and spring resulting in relative high temperatures in the advected air?\\
P9 L19: Replace 'exception is' by 'exception are the'\\
P9 L23: Replace 'appear to be' by 'are'\\
P9 L29: Replace 'very likely' by 'mostly'\\
P9 L33: Remove 'tend to'\\
P9 L32-P10 L2: Formulate more direct: This is explained by the LAB sites having surface melt ....' and remove the phrase 'ceiling of variability' but only keep the part about the upper limit of the melting point.\\
\noindent
P10 L3: How can figure 5 present daily variability? You present monthly averages in this figure. I suggest to replace 'daily' by 'monthly' or 'seasonal'.\\
P10 L4-5: Formulate more directly: 'Figure 6 illustrates...'\\
P10 L6: Remove 'without gaps'. Not continuous already indicates there are gaps.\\
P10 L11: Where does 'large differences' refer to? differences between individual stations or between ACC and LAB?\\
P10 L12-13: Where can I see that the Rnet is negative at night? And is this the case for all sites, also the most northern ones?\\
P10 L14-15: It is not incorrect, but the length of the melt season is temperature dependent, and thus for sites at lower altitudes and thus higher temperatures, the melt season is longer.\\
P10 L16: Replace 'from' by 'of'\\
P10 L24: Replace 'often katabatic winds' by 'are often of katabatic origin'\\
P10 L29: Remove 'binned'. and add that these are examples for two sites to illustrate the relation.\\
P10 L29-31: Remove 'The middel .... each bin' this is caption information.\\
P10 L32: Remove 'binned'.\\
\noindent
P11 L12: Remove 'resulting in and inversion wind'. \\
P11 L13: Replace 'inversion' by 'katabatic'. I prefer consistent terminology in this paper over consistency with other papers. You can explain this early in the manuscript. Now it gives the impression that these are two different things and they are not. \\
P11 L14: Remove 'it seems that'. The nature is the same/comparable between Greenland and Antarctica.\\
P11 L17: Replace 'from' by 'of'\\
P11 L19: Replace 'effects' by 'effect'. (also following sentences)\\
P11 L22: Refer to 'next section' instead of number.\\
P11 L27-30: Where can I see the seasonal variations?\\
P11: In reference to figure 9, is there for the Greenland ice sheet a pattern in the cloud free and cloud covered frequencies? Lower vs higher ablation zone? Or North vs South or East vs West? \\
P11 L32: Rephrase sentence. The figure presents cloud cover, not LWd, as this sentence now suggests. \\
P11 L32: Rephrase, 'average slope' of what doe you calculate? \\
\noindent
P12 L1-2: present slopes in a small table. And at least describe the results. Does your 95\% confidence interval mean that the slope is significantly different from 0? I think you need to use a different test for that. (same question for the slopes presented later in the manuscript)\\
P12 L15-16: What do you present in column 'all months' in case the is no annual average? Perhaps is makes more sense not to present 'all months' in case no full year is present.\\
P12 L18: Make this section 4.4 instead of 4.3.1.\\
P12 L19: Replace 'that can only be observed during' by 'which are only available under'\\
P12 L20: Replace 'in' by 'during'\\
P12 L21: What do you mean by 'within 1-3 days'? Do you mean 'averages of observations typically over a 1-3 day period'?\\
P12 L21-23: Rephrase: first say that satellite averages may thus differ, and than how they differ given the results presented in the previous section. Now you repeat the fact that under clear sky conditions the surface is colder twice in one sentence. Also not necessary to repeat when the satellite can and cannot measure, that is already stated the sentence before. For example: However, these satellite averages will differ from the all sky average temperature, since the Tskin is typically lower under clear sky conditions compared to cloudy conditions. This difference is referred to as the clear sky bias. \\
P12 L24: Add 'from satellites' behind 'observations'.\\
P12 L24: Replace 'therefore' by 'thus'.\\
P12 L25: Replace 'on' by 'off' and add 'on Tskin estimations from satellites' behind 'bias'.\\
P12 L25: Remove: 'by using.... windows'.\\
P12 L28: Add: 'by using.... windows'.\\
P12 L28: Replace: 'the cloud' by 'a cloud'\\
P12 L30: Replace: 'has been' by 'is'\\
P12 L30-31: Move the part between bracket to between 'observations' and 'with all sky', and move 'for all sites' to the start of the sentence.\\
\noindent
P13 L3-6: Rephrase, this sentence is not clear. I don't think I understand what you mean. Furthermore, you only describe for a specific situation what happens. It does not explain why for so many stations the differences increase for increasing periods. You also do not explain why for some stations the clear sky values are actually higher.\\
P13 L9: This acn be much better explained than only the range in Tskin. The smaller range is due to the less negative, or even positive radiative balance in summer. That also explains the smaller Tskin range for ACC, and for LAB the Tskin is at melting point and cannot be raised any further. \\
P13 L11: But you can make these figures for the part of the year that you do have observations!! \\
P13 L13: Add 'the period' after 'except for'. What is the spread in these figures? Can you present those as an transparant orange band around your observations?\\
P13 L13: Replace 'observed' by 'presented' and replace 'an effect' by 'the result'\\
P13 L15: Replace both 'is' by 'are', and add 'for' before 'Jan-Mar.'\\
P13 L16-17: Add 'not shown'\\
P13 L19-20: Rephrase: section 4.3.1 does not present a close relationship between Tskin in absolute sense and CCF. Phrase this as Tskin being affected by clouds.\\
P13 L18: Relationship of what with skin temperature? and remove 'surface'\\
P13 L19-27: Why do you present the relation of Tskin with the inversion strength? Since you wish to use Tskin as a proxy for T2m why not present that relation?\\
P13 L21: Replace 'corroborated' by 'shown'\\
P13 L21: It is difficult to judge the different stations from this plot, and the decrease in inversion strength is not very clear from this figure.\\
P13 L24-25: Present slopes in a small table. And at least discuss what they show.\\
P13 L25-27: I prefer to have an integrated results and discussion section, but if you do have a discussion section, this sentence should be in the discussion section and not in the results.\\
P13 L29-31: Rephrase, be more specific on when the coldest month occurs. Furthermore, remove the references, I don't see why it is necessary to use these references to confirm that it is colder in winter than summer.\\
\noindent
P13 L29 - P14 L28: This part only repeats what was already stated in the result section. I am missing a discussion about the bins in occurrence of cloud cover, the slope in the change in inversion strength with cloud cover is also not discussed, their values and the variability in them.\\
\noindent
P14 L3: Explain what categories do you refer to.\\
P14 L5: When indicating what conditions you refer to with winter, refer to the limited amount of incoming solar radiation instead of low sun and polar night.\\
P14 L6-7: Around noon and early afternoon in spring and summer? Move 'around noon and early afternoon' to before 'in spring and summer'.\\
P14 L7: Replace 'where' by 'when'.\\
P14 L8: Twice you refer to 'differences ' , but between what? Different sites in the Arctic?\\
P14 L9: What do you mean by 'closest coupling'? Rephrase: During summer and local noon the atmosphere is closest to neutral, the inversion is about 0, and then satellite observed Tskin will have the best agreement with T2m.\\
P14 L12: Replace 'downwards' by 'towards the surface'.\\
P14 L12: Replace 'also' by 'especially' or remove.\\
P14 L12: Replace 'inversion occurs' by 'inversion occur'.\\
P14 L13: Rephrase: with increasing wind speed the inversion strength decreases.\\
P14 L14: Remove 'and not at calm winds'\\
P14 L18: Remove sentence ' This is in agreement... ice sheet.' This is already stated in line 15.\\
P14 L20: Categories of what?\\
P14 L25: Relationship between what?\\
P14 L29: Replace 'at' by 'and'\\
\noindent
P15 L1: What do you mean by screen? Remove the sentence 'we therefore ... PROMICE data. In addition,'\\
P15 L2: Remove 'actual' and replace 'on' by 'in'.\\
P15 L2: Add what these uncertainty are. What is the uncertainty in your estimated temperature inversion, what is the uncertainty introduced by not taking height changes into account, what is the uncertainty resulting when you do take it into account?\\
P15 L3: Remove: 'and again... sites' Not necessary to repeat this statement\\
P15 L4-5: Reformulate, this sentence is not clear. What do you wish to asses? The uncertainty introduced by using sensors with different spectral specifications? Why do you refer to clear sky here? Did you mean: 'To asses the impact of different spectral characteristics of the used radiometers on the observed clear sky temperature inversion, ...' \\
P15 L7: 'Slope' of what? And here you see a sign change, you have to say something about that. Likely, the trend is not significantly different from 0 in both cases. But you have to show this.\\
P15 L9: 'Trend' in what? \\
P15 L12: Start new sentence after 'interval', the sentence becomes incomprehensible.\\
P15 L14-16: Reformulate: You have to be more specific on what affects the bias and how that relates to different time windows. This is the only thing new compared to what you already present in the results.\\
P15 L10-16: Again repeat of Results and not much more information. I am missing a proper analyses and explanation of the the differences resulting from the averaging period and between the different stations (see comment about location sites on Greenland in the results part)\\
P15 L17: You do not show anywhere that IR satellite retrievals usually show a cold bias. Reformulate to introduce this properly. \\
P15 L17-21: As I understand, using satellites you derive Tskin, and based on the impact of clouds, that generally results in a cold bias of Tskin compared to all sky Tskin. Is it correct that this is fully explained by this? \\
Rephrase 'another part' because the fact that the satellite Tskin is compared to observations of T2m is another issue. Basically you cannot compare Tskin to T2m because of the varying temperature inversion. Reformulate this last part as well, to make a clear distinction between an issue with how you measure and an issue arising from how you interpret the observations.\\
P15 L23: Replace 'deployments' by 'site' or 'stations'.\\
P15 L23-25: Reformulate sentence into two shorter sentences to improve readability.\\
P15 L29: Reformulate, it is a bit strange that Tskin itself influences Tskin.\\
P15 L30: Remove 'tight', replace 'controls' by 'control'.\\
P15 L32: Start new sentence at 'and the generation' to improve readability.\\
\noindent
P16 L1: Reformulate: with the presented results it is not possible to interpret the satellite derived Tskin in terms of T2m. You first need to make of statistical model that includes all these effects.\\
P16 L2-3: Reformulate: In case you use NWP analyses for the correction of the satellite retrieval, the product is not an independent estimate of the temperature.\\
P16 L3: remove reference to start of satellite era. In terms of climate change assessment it is still a reasonably short period that satellite products are available.\\
\nointent
Figure 1: Elevation is in m above sea level.\\
Figure 3: Add correlation coefficient, bias and Root Mean Square Difference. It appears that the broad band is higher than the narrow band. Furthermore, add grid lines and upper axis and axis on the right side. Make the figure of similar design as figure 4. All figures should have the same general design. Check them all!! \\
Figure 4: Reformulate the caption: Add the abbreviations Tskin and T2m, add what regions this stations represents, and also rephrase the part about the orange line. The orange line is the temperature difference T2m - Tskin. Furthermore, add what the standard deviations represent which is the variability in the monthly mean.\\
Figure 5: Add grid lines, similar to figure 4. And add that Tskin is the surface temperature and T2m the air temperature at 2m. Also be consisten with you axis description, Tskin written out or not, T2m written out or not, T2m - Tskin written as temperature difference or not. Check for all figures!!\\
Figure 6: Add the abbreviations Tskin and T2m in the caption. Also be consistent in the use of T2m-Tskin as axes description or Temperature difference (see figure4)
Figure 7: Add that these are monthly averages, similar to figure 5.\\
Figure 8: I still do not see the purpose to present the bins in the middle and lower panels. You do not refer to it in the manuscript and to the lower plots not at all. Also present the standard deviations in the upper plot as a band around the signal itself, as in figure 4. Add what regions these stations represent.\\
Figure 11: Remove 'Similar to Figure 7.a but with...' and rephrase into an independent caption. The dotted lines indicate the maximum number of sunlight hours.\\
Figure 13: Reformulate: for each month into seasonal cycle or monthly averages. Add what does the grey band signify. What is the variability in the orange line? You can present this as a transparant orange band around the line.\\
Figure 14: Add that these are monthly averages. Why not show a scatter plot of Tskin vs T2m with the different stations as different colors?\\
Table 2: Reformulate caption: 'under different circumstances' should be something like 'different seasons'. What does 'all months' refer to? An annual average? And what does SICE .|. DMI_Q mean, describe in caption. Perhaps you should not present all months in case it does not cover a full year.\\ |
- Printer-friendly version
- Supplement
