the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Assessment of Arctic Seasonal Snow Cover Rates of Change
Lawrence Mudryk
Abstract. Arctic snow cover extent (SCE) trends and rates of change reported across recent climate assessments vary due to the time period of available data, the selection of snow products, and methodological considerations. While all reported trends are strongly negative during spring, more uncertainty exists in autumn. Motivated to increase the confidence in SCE trend reported in climate assessments, we quantify the impact of (1) year-over-year increases in time series length over the past two decades, (2) choice of reference period, (3) the application of a statistical methodology to improve inter-dataset agreement, (4) the impact of dataset ensemble size, and (5) product version changes. Results show that the rate of change during May and June has remained consistent over the past decade as time series length has increased, and is largely insensitive to the choice of reference period. Although new product versions have increased spatial resolution, use more advanced reanalysis meteorology to force snow models, and include improved remote sensing retrieval algorithms, these enhancements do not result in any notable changes in the observed rate of Arctic SCE change in any month compared to a baseline set of older products. The most impactful analysis decision involves the scaling of dataset climatologies using the NOAA snow chart climate data record as the baseline. While minor for most months, this adjustment can influence the calculated rate of change for June by a factor of two relative to different climatological baselines.
Chris Derksen and Lawrence Mudryk
Status: closed
-
RC1: 'Comment on tc-2022-197', Anonymous Referee #1, 12 Nov 2022
This paper calculates the rate of change (rate of change: %/decade) of seasonal snow cover in the Arctic region using multiple publicly available snow cover data sets and evaluates the effects of 1) length of data period, 2) reference period, 3) the application of a statistical method to improve agreement between datasets, 4) ensemble size, and 5) differences in product versions on the Arctic SCE trend. The results show that while there is variation in SCE among the data sets, a consistent trend is found in the May and June SCE rates of change in the Arctic region, which is very significant in terms of providing the community with material to judge the reliability of the previously reported Arctic SCE trend and is worthy of publication. However, there are some figures that need to be replaced and there are some errors in the figures (Figure 3 and Figure 5) that may cause confusion to the readers. The manuscript appears to be in the process of being completed and needs to be revised to be clearer referring to the comments (mostly editorial) shown below.
L117: Shouldn’t “section 3.2” be “section 3.1”?
L154: Figure 2. As in Figure “1”, but for June
L167: Figure 3b is for June, NOT October. Please replace them.
L169: Figure 3. As in Figure “1”, but for October
L200: “NOAA-CRD” should be “NOAA-CDR”
L200: “<SD>/SDx is the ratio of the standard deviation from the dataset under consideration (denoted SDx) to the average standard deviation of all datasets (denoted<SDx>)”
This sentence should be “<SD>/SDx is the ratio of the average standard deviation of all datasets (denoted<SD>) to the standard deviation from the dataset under consideration (denoted SDx)”
L227: “Δð ð3” should be “Δð ð2”. In addition, the symbols “P1” through ”P18” are confused with the reference period “P0”-“P3” and should be change.
L234: “3.4” Dataset Adjustments
L240: “andthe” should be “and the”
L263: “Section 3.2-3.4”?
L265: In Figure 5, positive values above zero indicate a stronger SCE rate of change and negative values below zero indicate a weaker SCE rate of change. However, the description in the figure is incorrect (“Weaker SCE change” is written on the positive of the axis).
L266: “peaking at approximately 2%”. This value should be “-2%”
L273: “the rates appear weaker relative to the most recent reference period”. Shouldn’t the “weaker” be “stronger” here?
L274: “3% decade-1”. In Figure 5 I cannot find the impact of “3%” in June. Is this correct value? “2%”?
L276: “reaching 5% decade-1 in June”. The dataset adjustment process seems to weaken SCE rates of change in June. So, “-5%” should be here as also indicated in Figure 5.
Citation: https://doi.org/10.5194/tc-2022-197-RC1 -
AC1: 'Reply to RC1', Chris Derksen, 13 Jan 2023
We are pleased the reviewer sees the values in the analysis for informing the community on the state of Arctic snow cover trends. Apologies for the editorial errors, which we have cleaned up in the revised manuscript. We have responded to all the review comments as outlined in the supplement and made a number of other minor changes to improve the paper, as will be noted in the track changes version of the revised manuscript.
-
AC1: 'Reply to RC1', Chris Derksen, 13 Jan 2023
-
RC2: 'Comment on tc-2022-197', Anonymous Referee #2, 22 Nov 2022
This manuscript looks to understand sources of uncertainty in monthly seasonal snow cover trends, particularly that from selection of dataset, time series length, choice of reference period, and dataset adjustment. A motivating factor is to have snow cover trends be better characterized for climate assessments such as the Arctic Report Card and IPCC reports. The results demonstrate how dataset choice impacts snow cover trends, and I think a study like this is important, particularly to have more robust results in climate assessments. However, some edits are needed before this manuscript is ready to be accepted.
- Line 29: You mention that unlike sea ice, snow disappears in the Arctic every summer. You should clarify that snow permanent snow/ice remains in glaciers. Further, how are glacier areas handled in this study? I did not see that mentioned.
- Line 45: Does the IPCC assessment break down the confidence level by month or season? If so, are all months of changes to Arctic snow cover extent ‘high confidence’ or is there variability throughout the year?
- Line 67: Please add references. Many studies/reviews have given examples of why/how snow is so difficult to measure, model, remotely sense.
- Lines 95-99: It’s unclear to me here which (both?) version of the NOAA-CDR you are using, whether the 180 km or the 24 km one.
- Line 90: I believe the GlobSnow products (not sure about CCI) mask out alpine areas. Is that correct? Did you mask out alpine areas from the other datasets, too? If not, how much land is included in MERRA2, for example, that is excluded from the GlobSnow products? Further, it could be helpful to have an additional figure that maps what land areas are included in the analysis, assuming alpine and/or glacier areas are excluded.
- Line 103: Did you test the sensitivity of the selected 5 mm SWE threshold for determining snow extent? If so, please include a few statements on that. If not, was there a reason you selected 5 mm?
- Table 1: Do any of the Snow Cover Extent products have a SWE/depth limit for when they consider a grid cell to be snow covered? If so, please include a mention and reference to that.
- Figure 3: Any thoughts for why the JASMES product has much more negative October SCE change than the other products?
- Line 233: Dataset Adjustments is missing its subsession number/heading
- Line 274: Figure 5 makes it seem like June’s average impact of reference year is below 3%, maybe only 2%. Am I reading the figure incorrectly?
- Line 274: -1 should be an exponent here
- Line 276: For some datasets, there is an effect in the fall months. Consider saying that there is no net effect on average for all the products.
- Figure 5: The y axis indicates the opposite that is mentioned in the caption. It's labeled that positive values mean a weaker change and negative values mean a stronger change. Can you clarify which one it is?
- Line 364: Could you include a figure or more text to describe what could be missed by only considering monthly trends instead of daily trends?
- Line 384: Are the Brown data products available to download?
Citation: https://doi.org/10.5194/tc-2022-197-RC2 -
AC2: 'Reply on RC2', Chris Derksen, 13 Jan 2023
Thanks for your careful review, and apologies for the editorial errors. We have cleaned these up, and responded to all comments as outlined in the supplement. We have made a number of other minor changes to improve the paper which are noted in the track changes version of the revised manuscript.
Status: closed
-
RC1: 'Comment on tc-2022-197', Anonymous Referee #1, 12 Nov 2022
This paper calculates the rate of change (rate of change: %/decade) of seasonal snow cover in the Arctic region using multiple publicly available snow cover data sets and evaluates the effects of 1) length of data period, 2) reference period, 3) the application of a statistical method to improve agreement between datasets, 4) ensemble size, and 5) differences in product versions on the Arctic SCE trend. The results show that while there is variation in SCE among the data sets, a consistent trend is found in the May and June SCE rates of change in the Arctic region, which is very significant in terms of providing the community with material to judge the reliability of the previously reported Arctic SCE trend and is worthy of publication. However, there are some figures that need to be replaced and there are some errors in the figures (Figure 3 and Figure 5) that may cause confusion to the readers. The manuscript appears to be in the process of being completed and needs to be revised to be clearer referring to the comments (mostly editorial) shown below.
L117: Shouldn’t “section 3.2” be “section 3.1”?
L154: Figure 2. As in Figure “1”, but for June
L167: Figure 3b is for June, NOT October. Please replace them.
L169: Figure 3. As in Figure “1”, but for October
L200: “NOAA-CRD” should be “NOAA-CDR”
L200: “<SD>/SDx is the ratio of the standard deviation from the dataset under consideration (denoted SDx) to the average standard deviation of all datasets (denoted<SDx>)”
This sentence should be “<SD>/SDx is the ratio of the average standard deviation of all datasets (denoted<SD>) to the standard deviation from the dataset under consideration (denoted SDx)”
L227: “Δð ð3” should be “Δð ð2”. In addition, the symbols “P1” through ”P18” are confused with the reference period “P0”-“P3” and should be change.
L234: “3.4” Dataset Adjustments
L240: “andthe” should be “and the”
L263: “Section 3.2-3.4”?
L265: In Figure 5, positive values above zero indicate a stronger SCE rate of change and negative values below zero indicate a weaker SCE rate of change. However, the description in the figure is incorrect (“Weaker SCE change” is written on the positive of the axis).
L266: “peaking at approximately 2%”. This value should be “-2%”
L273: “the rates appear weaker relative to the most recent reference period”. Shouldn’t the “weaker” be “stronger” here?
L274: “3% decade-1”. In Figure 5 I cannot find the impact of “3%” in June. Is this correct value? “2%”?
L276: “reaching 5% decade-1 in June”. The dataset adjustment process seems to weaken SCE rates of change in June. So, “-5%” should be here as also indicated in Figure 5.
Citation: https://doi.org/10.5194/tc-2022-197-RC1 -
AC1: 'Reply to RC1', Chris Derksen, 13 Jan 2023
We are pleased the reviewer sees the values in the analysis for informing the community on the state of Arctic snow cover trends. Apologies for the editorial errors, which we have cleaned up in the revised manuscript. We have responded to all the review comments as outlined in the supplement and made a number of other minor changes to improve the paper, as will be noted in the track changes version of the revised manuscript.
-
AC1: 'Reply to RC1', Chris Derksen, 13 Jan 2023
-
RC2: 'Comment on tc-2022-197', Anonymous Referee #2, 22 Nov 2022
This manuscript looks to understand sources of uncertainty in monthly seasonal snow cover trends, particularly that from selection of dataset, time series length, choice of reference period, and dataset adjustment. A motivating factor is to have snow cover trends be better characterized for climate assessments such as the Arctic Report Card and IPCC reports. The results demonstrate how dataset choice impacts snow cover trends, and I think a study like this is important, particularly to have more robust results in climate assessments. However, some edits are needed before this manuscript is ready to be accepted.
- Line 29: You mention that unlike sea ice, snow disappears in the Arctic every summer. You should clarify that snow permanent snow/ice remains in glaciers. Further, how are glacier areas handled in this study? I did not see that mentioned.
- Line 45: Does the IPCC assessment break down the confidence level by month or season? If so, are all months of changes to Arctic snow cover extent ‘high confidence’ or is there variability throughout the year?
- Line 67: Please add references. Many studies/reviews have given examples of why/how snow is so difficult to measure, model, remotely sense.
- Lines 95-99: It’s unclear to me here which (both?) version of the NOAA-CDR you are using, whether the 180 km or the 24 km one.
- Line 90: I believe the GlobSnow products (not sure about CCI) mask out alpine areas. Is that correct? Did you mask out alpine areas from the other datasets, too? If not, how much land is included in MERRA2, for example, that is excluded from the GlobSnow products? Further, it could be helpful to have an additional figure that maps what land areas are included in the analysis, assuming alpine and/or glacier areas are excluded.
- Line 103: Did you test the sensitivity of the selected 5 mm SWE threshold for determining snow extent? If so, please include a few statements on that. If not, was there a reason you selected 5 mm?
- Table 1: Do any of the Snow Cover Extent products have a SWE/depth limit for when they consider a grid cell to be snow covered? If so, please include a mention and reference to that.
- Figure 3: Any thoughts for why the JASMES product has much more negative October SCE change than the other products?
- Line 233: Dataset Adjustments is missing its subsession number/heading
- Line 274: Figure 5 makes it seem like June’s average impact of reference year is below 3%, maybe only 2%. Am I reading the figure incorrectly?
- Line 274: -1 should be an exponent here
- Line 276: For some datasets, there is an effect in the fall months. Consider saying that there is no net effect on average for all the products.
- Figure 5: The y axis indicates the opposite that is mentioned in the caption. It's labeled that positive values mean a weaker change and negative values mean a stronger change. Can you clarify which one it is?
- Line 364: Could you include a figure or more text to describe what could be missed by only considering monthly trends instead of daily trends?
- Line 384: Are the Brown data products available to download?
Citation: https://doi.org/10.5194/tc-2022-197-RC2 -
AC2: 'Reply on RC2', Chris Derksen, 13 Jan 2023
Thanks for your careful review, and apologies for the editorial errors. We have cleaned these up, and responded to all comments as outlined in the supplement. We have made a number of other minor changes to improve the paper which are noted in the track changes version of the revised manuscript.
Chris Derksen and Lawrence Mudryk
Chris Derksen and Lawrence Mudryk
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 288 | 121 | 16 | 425 | 5 | 4 |
- HTML: 288
- PDF: 121
- XML: 16
- Total: 425
- BibTeX: 5
- EndNote: 4
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1