the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 07 May 2025
Similarities between sea ice area variations and satellite-derived terrestrial biosphere and cryosphere parameters across the Arctic
Abstract. Satellite time series availability for the Arctic Ocean and adjacent land areas allows for cross-comparisons for cryosphere vs. vegetation parameters. Previous studies focused on correlation analyses between vegetation indices (derivatives of the normalized difference vegetation index (NDVI)) of tundra regions and sea ice extent for selected months. We have refined these analyses through consideration of distinct sea ice basins and all months, extension to south of the treeline, and included cryosphere essential climate variables such as snow water equivalent (SWE; March as proxy for annual maximum) and mean annual ground temperature (MAGT) in permafrost areas. The focus was on 2000–2019 considering data availability. As a first step, we derived trends. Changes across all the different parameters could be specifically determined for Eastern Siberia. Linkages between de-trended sea ice area (SIA) and NDVI across tundra regions was confirmed, where lower sea ice extent correlates with higher NDVI. The regional extension beyond the treeline revealed linkages for Northern European Russia and partially correlations of sea ice variations with land parameters over northern Scandinavia. Differences compared to previous studies ending in 2008 were found for the Kara Sea region and adjacent land area, indicating recent changes. In case of ground temperatures, high significant correlations were found for more distant sea ice basins than for NDVI, where the adjacent sea ice basins were more relevant. Negative and positive significant correlations can be found for March SWE depending on SIA month and region. Also, other months than September (sea ice extent minimum) were found to have high correlations vs. land-based variables, with distinct differences across sea ice basins. The fraction of data points with significant correlations north of 60° N is higher for SWE and MAGT than for the NDVI derivatives. Fractions for SWE are higher for Eurasia than Northern America. Autumn (incl. October and November) and mid-winter (incl. February, March) were most relevant for both investigated cryosphere-related parameters permafrost temperature and March snow water equivalent. Although similarities could be found between TI-NDVI and MaxNDVI, a higher proportion of significant correlations was observed for TI-NDVI. The datasets provide a baseline for future studies on common drivers of essential climate parameters and causative effects across the Arctic.
- Preprint
(38122 KB) - Metadata XML
- BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on egusphere-2025-1358', Anonymous Referee #1, 04 Jun 2025
The manuscript by Bartsch et al. describes how pan-Arctic datasets of mean annual ground temperature at 2m depth (MAGT), snow water equivalent (SWE) and NDVI (as a proxy for plant growth) correlate with sea ice area (SIA). Sea ice loss is one of the main causes of the amplified warming of the Arctic, and together with changes in atmospheric humidity this influences MAGT, SWE and plant growth. Such links have been shown previously from observations, remote sensing and models (see e.g. Bhatt et al., 2010, 2014, 2017; Buchwal et al., 2020; Macias-Fauria et al., 2012, 2017; Parmentier et al., 2015; Rehder et al., 2020; Screen et al., 2012; Yu et al., 2021). This study aims to differentiate itself from this previous work by using satellite data where possible, and by focusing more on regional correlations rather than those made across the whole Arctic.
While I appreciate the attempt by the authors to look further into this topic, I feel that the manuscript in its current form is a missed opportunity to learn something truly novel. In particular, I had hoped that this study would go beyond mere correlations by identifying causal links, and by showing more detail. More detailed regional analyses between sea ice and the terrestrial environment have been done for example by Parmentier et al. (2015) who performed a pan-Arctic pixel-wise correlation between local sea ice conditions and temperature and modeled methane emissions, and who argued a causal link in autumn but not in spring. Rehder et al. (2020) used causal-effect networks to identify temporal links to the land near the Laptev Sea, and showed that spring-time correlations in sea ice and atmospheric variables were both related to large scale atmospheric circulation, not to each other, although sea ice loss had a weak effect on the near coastal environment in summer. Regional links between NDVI and sea ice have also been shown before (see e.g. Yu et al. 2021 and the paper by one of the co-authors of this study, Macias-Fauria et al. 2017). In addition, see also chapter 10 of the 2017 AMAP report (the authors incorrectly state on line 45 that this report did not include vegetation trends). Btw, reverse links have also been argued, where terrestrial vegetation growth lowers surface albedo, affecting climate and subsequently sea ice loss (Zhang et al., 2020).
While many of these previous studies relied on models or reanalysis datasets, this study aims to use remote-sensing datasets as much as possible. However, the authors use the TTOP model to determine soil temperature at 2 m depth. While this model uses land surface temperature (LST) from MODIS as an input, it also uses reanalysis data when MODIS LST is unavailable. Moreover, it models the soil temperature depending on for example land cover and surface wetness. While the TTOP model is probably the best estimate we have for permafrost extent at the moment, it is still a (hybrid) model. If the authors wanted to compare to satellite data only, rather than reanalyses or models, it would have made more sense to compare to MODIS LST directly. Moreover, 2 m depth is rather deep in the Arctic, where the active layer is typically shallower than 1 m. Any warming signal would be strongly attenuated and lagged at 2 m depth, which makes it difficult to make instantaneous correlations.
The current study also shows correlations at short and long distances, but it is not clear whether these correlations have a common distant cause or whether they represent an internal dynamic in the Arctic. Are they due to large scale atmospheric circulation affecting both sea ice and the terrestrial variables? Or are they due to local feedbacks dominated by sea ice decline? Unfortunately, the answer to this question is left in the middle by the authors, who present the dataset as a baseline for further analyses of drivers and dependencies. The paper would have been much stronger if it included a proper discussion on the underlying causes for the apparent correlations.
While I generally appreciate the effort by the authors, they could have done a more detailed spatial analysis given the high spatial detail of the source data, and they should have provided better causal insights rather than showing correlations with little context. Unless the authors go beyond their basic presentation of the data, and given the fact that they urge others to take their dataset further, I feel like this manuscript in its current form would work better as a data paper (e.g. in ESSD) rather than as a research article.
Some further comments:
- Line 45: Vegetation is extensively discussed in the 2017 SWIPA report (see chapters 8 and 10).
- Line 81: wouldn’t frequent cloud cover be a problem for MaxNDVI as well? Easy to miss the peak season in frequent cloudy parts of the Arctic, adding uncertainty to interannual variability in peak NDVI values.
- Line 126-128: did this model result match the observations well?
- Line 150: which parameters? Reference?
- Line 157: what’s the pixel size?
- Line: 164-165: This dataset appears to contain only the trend over the entire time series, not the original high temporal data used for the correlations, and there are no details on how the underlying dataset was processed. Is there a reference describing this?
- Line 177-178: This assumption appears reasonable, but did you test whether it’s true?
- Figure 1: could you give the area north of 60°N a different color from the one south of 60°N? Makes it easier to see the domain instead of just highlighting a latitudinal band.
- Table 4: what’s the resolution of the source data for sea ice area?
- Line 201: entire northern hemisphere or only north of 60°N?
- Line 255: I can imagine that correlations to more distant basins may show up but here we need better argumentation for why these correlations appear because at these large distances it may just as well be large scale atmospheric circulation affecting both (e.g. teleconnections related to Rossby wave propagation).
- Line 269: a positive correlation may not be that surprising. For example, the Fram Strait is an area of sea ice export, which can actually be enhanced in warm summers because the strong sea ice melt makes the ice thinner and more mobile, subsequently leading to more export through this area. As such, a warmer Arctic leads to more sea ice in the Fram Strait, explaining positive correlations. These kinds of internal dynamics need to be considered when interpreting correlations to sea ice.
- Line 275: do you mean figure A7 instead of A2?
- Line 279-280: This is unclear. Relevant for TI-NDVI in what way?
- Line 336: Please specify why solar absorption trends are increasing in this region. Is it less cloud cover or changes in surface albedo from e.g. earlier snow melt and/or shrubification?
- Line 376-377: This sounds interesting, but what would be the reason for this temporal lag?
- Line 379-384: I’m not sure I’m following this. Why would warmer summers and increased absorption of radiation lead to regional cooling in the autumn?
- Line 407: Unclear. Which “following ones”?
- Line 433-435: I’m not sure how this agrees with Sasgen et al. (2024) since they explicitly state that they did not look at the influence of sea ice.
- Line 449: why define the abbreviation FT for “freeze-thaw” if you only use it one more time?
- Figure A2: please replace “source” in the caption with the actual reference.
- Table A2: what does the “2000?” mean in the table caption?References
Bhatt, U. S., Walker, D. A., Raynolds, M. K., Comiso, J. C., Epstein, H. E., Jia, G., Gens, R., Pinzon, J. E., Tucker, C. J., Tweedie, C. E., and Webber, P. J.: Circumpolar Arctic Tundra Vegetation Change Is Linked to Sea Ice Decline, Earth Interact., 14, 1–20, https://doi.org/10.1175/2010EI315.1, 2010.
Bhatt, U. S., Walker, D. A., Walsh, J. E., Carmack, E. C., Frey, K. E., Meier, W. N., Moore, S. E., Parmentier, F.-J. W., Post, E., Romanovsky, V. E., and Simpson, W. R.: Implications of Arctic Sea Ice Decline for the Earth System, Annu. Rev. Environ. Resour., 39, 57–89, https://doi.org/10.1146/annurev-environ-122012-094357, 2014.
Bhatt, U. S., Walker, D. A., Raynolds, M. K., Bieniek, P. A., Epstein, H. E., Comiso, J. C., Pinzon, J. E., Tucker, C. J., Steele, M., Ermold, W., and Zhang, J.: Changing seasonality of panarctic tundra vegetation in relationship to climatic variables, Environ. Res. Lett., 12, 055003, https://doi.org/10.1088/1748-9326/aa6b0b, 2017.
Buchwal, A., Sullivan, P. F., Macias-Fauria, M., Post, E., Myers-Smith, I. H., Stroeve, J. C., Blok, D., Tape, K. D., Forbes, B. C., Ropars, P., Lévesque, E., Elberling, B., Angers-Blondin, S., Boyle, J. S., Boudreau, S., Boulanger-Lapointe, N., Gamm, C., Hallinger, M., Rachlewicz, G., Young, A., Zetterberg, P., and Welker, J. M.: Divergence of Arctic shrub growth associated with sea ice decline, Proc. Natl. Acad. Sci., 117, 33334–33344, https://doi.org/10.1073/pnas.2013311117, 2020.
Macias-Fauria, M., Forbes, B. C., Zetterberg, P., and Kumpula, T.: Eurasian Arctic greening reveals teleconnections and the potential for structurally novel ecosystems, Nat. Clim. Change, 2, 613–618, https://doi.org/10.1038/nclimate1558, 2012.
Macias-Fauria, M., Karlsen, S. R., and Forbes, B. C.: Disentangling the coupling between sea ice and tundra productivity in Svalbard, Sci. Rep., 7, 8586, https://doi.org/10.1038/s41598-017-06218-8, 2017.
Parmentier, F.-J. W., Zhang, W., Mi, Y., Zhu, X., Huissteden, J., Hayes, D. J., Zhuang, Q., Christensen, T. R., and McGuire, A. D.: Rising methane emissions from northern wetlands associated with sea ice decline, Geophys. Res. Lett., 42, 7214–7222, https://doi.org/10.1002/2015GL065013, 2015.
Rehder, Z., Niederdrenk, A. L., Kaleschke, L., and Kutzbach, L.: Analyzing links between simulated Laptev Sea sea ice and atmospheric conditions over adjoining landmasses using causal-effect networks, The Cryosphere, 14, 4201–4215, https://doi.org/10.5194/tc-14-4201-2020, 2020.
Screen, J. A., Simmonds, I., Deser, C., and Tomas, R.: The Atmospheric Response to Three Decades of Observed Arctic Sea Ice Loss, J. Clim., 26, 1230–1248, https://doi.org/10.1175/JCLI-D-12-00063.1, 2012.
Yu, L., Leng, G., and Python, A.: Varying response of vegetation to sea ice dynamics over the Arctic, Sci. Total Environ., 799, 149378, https://doi.org/10.1016/j.scitotenv.2021.149378, 2021.
Zhang, W., Döscher, R., Koenigk, T., Miller, P. A., Jansson, C., Samuelsson, P., Wu, M., and Smith, B.: The Interplay of Recent Vegetation and Sea Ice Dynamics—Results From a Regional Earth System Model Over the Arctic, Geophys. Res. Lett., 47, e2019GL085982, https://doi.org/10.1029/2019GL085982, 2020.Citation: https://doi.org/10.5194/egusphere-2025-1358-RC1 -
AC1: 'Reply on RC1', Annett Bartsch, 27 Jun 2025
Many thanks for the in depth review and comments! Please find our comments and responses to your questions below.
Reviewer comment: The manuscript by Bartsch et al. describes how pan-Arctic datasets of mean annual ground temperature at 2m depth (MAGT), snow water equivalent (SWE) and NDVI (as a proxy for plant growth) correlate with sea ice area (SIA). Sea ice loss is one of the main causes of the amplified warming of the Arctic, and together with changes in atmospheric humidity this influences MAGT, SWE and plant growth. Such links have been shown previously from observations, remote sensing and models (see e.g. Bhatt et al., 2010, 2014, 2017; Buchwal et al., 2020; Macias-Fauria et al., 2012, 2017; Parmentier et al., 2015; Rehder et al., 2020; Screen et al., 2012; Yu et al., 2021). This study aims to differentiate itself from this previous work by using satellite data where possible, and by focusing more on regional correlations rather than those made across the whole Arctic.
- Comment: Note, further key novel points are the consideration of vegetation and cryosphere at the same time, previous studies have focused on one parameter only. A comparison through a consistent setup for vegetation and land cryosphere parameters was so far not made. Also the temporary extension, evaluating all months in combination with consideration of sea ice basins goes beyond past studies. There has been an example for NDVI versus monthly sea ice (Yu et al. 2021), but without considering different sea ice basins.
Reviewer comment: While I appreciate the attempt by the authors to look further into this topic, I feel that the manuscript in its current form is a missed opportunity to learn something truly novel. In particular, I had hoped that this study would go beyond mere correlations by identifying causal links, and by showing more detail. More detailed regional analyses between sea ice and the terrestrial environment have been done for example by Parmentier et al. (2015) who performed a pan-Arctic pixel-wise correlation between local sea ice conditions and temperature and modeled methane emissions, and who argued a causal link in autumn but not in spring. Rehder et al. (2020) used causal-effect networks to identify temporal links to the land near the Laptev Sea, and showed that spring-time correlations in sea ice and atmospheric variables were both related to large scale atmospheric circulation, not to each other, although sea ice loss had a weak effect on the near coastal environment in summer.
- Reply: Thanks for pointing out the two studies. Parmentier et al. (2015) show that linkages might exist even further south than for the extent we have chosen plus the inclusion of sea ice across the Hudson bay and Canadian Archipelago (unfortunately this was not covered in the sea ice basin data set that we used), what should be addressed in the discussion in addition to comparison of the findings and discussion of potential causal links of Parmentier et al. (2015) as well as Rehder et al. (2020).
Reviewer comment: Regional links between NDVI and sea ice have also been shown before (see e.g. Yu et al. 2021 and the paper by one of the co-authors of this study, Macias-Fauria et al. 2017). In addition, see also chapter 10 of the 2017 AMAP report (the authors incorrectly state on line 45 that this report did not include vegetation trends). Btw, reverse links have also been argued, where terrestrial vegetation growth lowers surface albedo, affecting climate and subsequently sea ice loss (Zhang et al., 2020).
- Reply:
- Please note that we acknowledge the extensive previous work on NDVI in the paper, including the mentioned studies Yu et al. 2012 and Macias-Fauria et al. 2017 (see section 2.1 and Table 3). In addition, an in depth analyses has been, however, so far missing for the cryosphere parameters.
- Thanks for pointing out the misplacement of the phrasing on vegetation and AMAP. This sentence relates to Comiso and Hall (2014) mentioned before, so needs to be moved before the AMAP mentioning. Note, that we correctly list in Table A1 that the AMAP study considered vegetation.
- Thanks for pointing out Zhang et al., 2020, which should be considered in the discussion.
Reviewer comment: While many of these previous studies relied on models or reanalysis datasets, this study aims to use remote-sensing datasets as much as possible. However, the authors use the TTOP model to determine soil temperature at 2 m depth. While this model uses land surface temperature (LST) from MODIS as an input, it also uses reanalysis data when MODIS LST is unavailable. Moreover, it models the soil temperature depending on for example land cover and surface wetness. While the TTOP model is probably the best estimate we have for permafrost extent at the moment, it is still a (hybrid) model. If the authors wanted to compare to satellite data only, rather than reanalyses or models, it would have made more sense to compare to MODIS LST directly. Moreover, 2 m depth is rather deep in the Arctic, where the active layer is typically shallower than 1 m. Any warming signal would be strongly attenuated and lagged at 2 m depth, which makes it difficult to make instantaneous correlations.
- Reply:
- Indeed, the signal at 2m depth is attenuated. The reason to anyway use it comes from (1) that 2m depth is commonly used to represent permafrost presence (e.g. Obu et al. 2021, and product documentation) and (2) that initial observations in Bartsch et al. (2023) pointed to potentially high correlations with sea ice. And our results in the new manuscript have confirmed these and demonstrate a significantly higher linkage with sea ice than all other parameters, despite of the attenuation. But we agree that the issue of attenuation should be discussed.
- Note, that the datasets used come from ESA Permafrost CCI which were created using a transient model version of CryoGRID in order to obtain time series. A TTOP model was used in the past for the creation of the ESA DUE GlobPermafrost dataset representing equilibrium conditions 2000-2016, what cannot be used for our trend analyses.
Reviewer comment: The current study also shows correlations at short and long distances, but it is not clear whether these correlations have a common distant cause or whether they represent an internal dynamic in the Arctic. Are they due to large scale atmospheric circulation affecting both sea ice and the terrestrial variables? Or are they due to local feedbacks dominated by sea ice decline? Unfortunately, the answer to this question is left in the middle by the authors, who present the dataset as a baseline for further analyses of drivers and dependencies. The paper would have been much stronger if it included a proper discussion on the underlying causes for the apparent correlations.
- Reply: Literature suggests that the underlying dynamics differ between the analysed parameters. For example, as mentioned in the discussion, Sasgen et al. (2025) suggest large scale atmospheric circulation playing a role for ice sheets and permafrost. The linkage between sea ice in the proximity and SWE as well as NDVI has been pointed out regionally before. We agree that the discussion should be restructured to cover these aspects better and in more detail, also considering the above suggested references.
Reviewer comment: While I generally appreciate the effort by the authors, they could have done a more detailed spatial analysis given the high spatial detail of the source data, and they should have provided better causal insights rather than showing correlations with little context. Unless the authors go beyond their basic presentation of the data, and given the fact that they urge others to take their dataset further, I feel like this manuscript in its current form would work better as a data paper (e.g. in ESSD) rather than as a research article.
- Reply:
- Please note that with the present setup new results are shown for all parameters, and specifically for the cryosphere this type of correlation analyses was not done to date.
- Could you clarify what is meant with ‘more detailed spatial analysis given the high spatial detail of the source data’?
Further reviewer comments:
- Line 45: Vegetation is extensively discussed in the 2017 SWIPA report (see chapters 8 and 10).
- Reply: As mentioned above, the placement of the sentence is a mistake. This sentence refers to Comiso & Hall. We state the consideration of vegetation in AMAP/SWIPA in Table A1.
- Line 81: wouldn’t frequent cloud cover be a problem for MaxNDVI as well? Easy to miss the peak season in frequent cloudy parts of the Arctic, adding uncertainty to interannual variability in peak NDVI values.
- Reply: Yes, this is why we have used TI-NDVI which is bi-monthly (16 days periods) instead of growing season summed NDVI (GSSNDVI) which is constructed on a daily basis (Park et al. 2016). And it is indeed an issue to be mentioned also in the discussion.
- Line 126-128: did this model result match the observations well? - Reply: A decreasing amount of snow over North America was also found through satellite observations by Pulliaianen et al. (2020)
- Line 150: which parameters? Reference? - Suggestion for extended phrasing: … cryosphere as well as vegetation parameters 2000 onwards, due to specifically the availability of MODIS and denser coverage by Landsat (see e.g. Tables 2 and 3)
- Line 157: what’s the pixel size? - 25km, but we have been using for sea ice basins aggregated values.
- Line: 164-165: This dataset appears to contain only the trend over the entire time series, not the original high temporal data used for the correlations, and there are no details on how the underlying dataset was processed. Is there a reference describing this? - reply: We suggest to add the following text: A circumpolar and circumboreal time series of satellite data of AVHRR- GIMMS and MODIS in the period 1982-2019, respectively, were processed. The following parameters related to the growing season as the yearly maximum NDVI (MaxNDVI), and time integrated NDVI (TI-NDVI) were computed and analyzed from these data series. Maximum NDVI (MaxNDVI) is the annual maximum NDVI value observed during the period of peak phytomass during growing season, typically in late July and early August for the Arctic (Frost et al. 2025), and somewhat earlier for the Boreal region below 60° N. The following datasets were processed:
- Advanced Very High Resolution Radiometer (AVHRR), 1982–2019, in the form of the Global Inventory Modeling and Mapping Studies 3 g V1.2 dataset (GIMMS-3g+) with a spatial resolution of 1/12° (~8 km) (Pinzon and Tucker, 2014; Pinzon et al., 2023).
- Moderate Resolution Imaging Spectroradiometer (MODIS), 2000–2019, using 16-day NDVI product from the Terra (MOD13A1, v. 6.1) with a spatial resolution of 500 m. Pixels with a Summary Quality Assurance value ≥3 (indicating cloudy and compromised observations) were removed.
- Time-integrated NDVI (TI-NDVI) is the sum of maximum NDVI values within set compositing periods during May–September, calculated for datasets with daily temporal resolution. TI-NDVI includes phenological variations throughout the growing season; therefore, it better represents gross primary production (Tucker and Sellers, 1986), and is better correlated with climate variables than MaxNDVI (Bhatt et al., 2010; Bhatt et al., 2021). NDVImax and TI-NDVI may exhibit different correlations with climatic parameters like temperature (Yan et al. 2022). No filtering processes or other amendments have been done to the datasets and we have pretty much followed the procedure in Frost et al. (2025).
- Line 177-178: This assumption appears reasonable, but did you test whether it’s true?
- Reply: Considering the whole Northern hemisphere (as well as Eurasia and North America separately), the average March SWE was found to be the closest monthly average to peak SWE by Pulliainen et al., (2020) [see extended data Table 1)]. We suggest to added the following text::
- Average SWE in March was selected for this study and interpreted to represent maximum SWE conditions, following Pulliainen et al., (2020).
- Figure 1: could you give the area north of 60°N a different color from the one south of 60°N? Makes it easier to see the domain instead of just highlighting a latitudinal band.- Reply: We agree that this would be useful
- Table 4: what’s the resolution of the source data for sea ice area? - Reply: 25km
- Line 201: entire northern hemisphere or only north of 60°N? - Reply: Thanks for pointing out, it should be north of 60°N
- Line 255: I can imagine that correlations to more distant basins may show up but here we need better argumentation for why these correlations appear because at these large distances it may just as well be large scale atmospheric circulation affecting both (e.g. teleconnections related to Rossby wave propagation).
- Reply: Note that we come back to that in the discussion, e.g. see discussion of Sasgen et al. (2024)
- Line 269: a positive correlation may not be that surprising. For example, the Fram Strait is an area of sea ice export, which can actually be enhanced in warm summers because the strong sea ice melt makes the ice thinner and more mobile, subsequently leading to more export through this area. As such, a warmer Arctic leads to more sea ice in the Fram Strait, explaining positive correlations. These kinds of internal dynamics need to be considered when interpreting correlations to sea ice.
- Reply: Thanks for pointing out, we agree that this should be covered in the discussion.
- Line 275: do you mean figure A7 instead of A2?
- Reply: Yes, thanks for spotting!
- Line 279-280: This is unclear. Relevant for TI-NDVI in what way?
- Reply: for the adjacent land area
- Line 336: Please specify why solar absorption trends are increasing in this region. Is it less cloud cover or changes in surface albedo from e.g. earlier snow melt and/or shrubification? - Reply: This observation comes from Letterly et al. 2018 (based on reanalyses), who discuss earlier snow melt timing as a general reason for reduction of albedo in some areas across the Arctic.
- Line 376-377: This sounds interesting, but what would be the reason for this temporal lag? - Reply: There is a mistake here. ‘following’ should be replaced with ‘preceding’. The reasoning is provided at the end of the paragraph.
- Line 379-384: I’m not sure I’m following this. Why would warmer summers and increased absorption of radiation lead to regional cooling in the autumn? - Reply: The observation is that there are summers with high NDVI followed by earlier sea ice formation in the autumn. The processes listed are rather speculation, thus we suggest to remove the part ‘(such as …. 4.1)’ and change ‘turn into’ to ‘lead to’ and add a comment that more investigations are needed for clarification.
- Line 407: Unclear. Which “following ones”? - Reply: After June
- Line 433-435: I’m not sure how this agrees with Sasgen et al. (2024) since they explicitly state that they did not look at the influence of sea ice. - Reply: We refer here to long distance linkages in general, not specifically sea ice. This could be change for example to ‘Such long distance linkages across the Arctic were also found ice sheets and permafrost …’
- Line 449: why define the abbreviation FT for “freeze-thaw” if you only use it one more time? - Reply: Thanks for spotting
- Figure A2: please replace “source” in the caption with the actual reference. - Reply: Thanks for spotting
- Table A2: what does the “2000?” mean in the table caption? - Reply: Thanks for spotting. This column should be removed, this was from an old version of the table.
References in replies:
Bartsch, A., Strozzi, T., and Nitze, I.: Permafrost Monitoring from Space, Surveys in Geophysics, https://doi.org/10.1007/s10712-023-09770-3, 2023
Frost, G.V. , Bhatt Uma S. , Macander M.V., Logan, B.T. , Walker D.A., Raynolds, M.K., Magnússon, R. I , Bartsch, A., Bjerke , J.W., Epstein, H.E. , Forbes, B.C. , Goetz, S.J , Hoy, E.E. , Karlsen S.R. , Kumpula, T. , Lantz, T.C. , Lara, M. J. , López-Blanco, E., Montesano, P.M., Neigh, C.S.R. , Nitze, I., Orndahl, Kathleen M. , Park, T. , Phoenix, G.K. , Rocha, A.V. , Rogers, B. M., Schaepman-Strub, G., Tømmervik,H., Verdonen, M., Veremeeva, A., Virkkala, A-M., & Wiagl, C.F. 2025. The changing face of the Arctic: four decades of greening and implications for tundra ecosystems Frontiers in Environmental Research, 13 , 2025 DO10.3389/fenvs.2025.1525574
Letterly, A., Key, J., and Liu, Y.: Arctic climate: changes in sea ice extent outweigh changes in snow cover, The Cryosphere, 12, 3373–3382, https://doi.org/10.5194/tc-12-3373-2018, 2018.
Obu, J., Westermann, S., Barboux, C., Bartsch, A., Delaloye, R., Grosse, G., Heim, B., Hugelius, G., Irrgang, A., Kääb, A. M., Kroisleitner, C., Matthes, H., Nitze, I., Pellet, C., Seifert, F. M., Strozzi, T., Wegmüller, U., Wieczorek, M., and Wiesmann,A.: ESA Permafrost Climate Change Initiative (Permafrost_cci): Permafrost extent for the Northern Hemisphere, v3.0, CEDA, https://doi.org/10.5285/6e2091cb0c8b4106921b63cd5357c97c, 2021b.
Park, T., Ganguly, S., Tømmervik, H., Euskirchen, E.S., Høgda, K.A., Karlsen, S.R., Brovkin, V., Nemani , R.R., Myneni, R.B. 2016 Environ. Res. Lett. 11 084001
Pinzon, J. E., Tucker, C. J., Bhatt, U. S., Frost, G. V., and Macander, M. J. (2023).Global vegetation greenness (NDVI) from AVHRR GIMMS-3g+. NASA’s Open Data Portal, 1981–2022. doi:10.3334/ORNLDAAC/2187
Sasgen, I., Steinhoefel, G., Kasprzyk, C., Matthes, H., Westermann, S., Boike, J., and Grosse, G.: Atmosphere circulation patterns synchronize pan-Arctic glacier melt and permafrost thaw, Communications Earth & Environment, 5, 375, https://doi.org/10.1038/s43247-024- 01548-8, 2024.
Tucker, C. J., and Sellers, P. J. (1986). Satellite remote sensing of primary production. Int. J. Remote Sens. 16, 1395–1416. doi:10.1080/01431168608948944
Yan, J., Zhang, G., Ling, H., & Han, F. 2022. Comparison of time-integrated NDVI and annual maximum NDVI for assessing grassland dynamics, Ecological Indicators, 136, 2022, 108611https://doi.org/10.1016/j.ecolind.2022.108611.
References in reviewer comments:
Bhatt, U. S., Walker, D. A., Raynolds, M. K., Comiso, J. C., Epstein, H. E., Jia, G., Gens, R., Pinzon, J. E., Tucker, C. J., Tweedie, C. E., and Webber, P. J.: Circumpolar Arctic Tundra Vegetation Change Is Linked to Sea Ice Decline, Earth Interact., 14, 1–20, https://doi.org/10.1175/2010EI315.1, 2010.
Bhatt, U. S., Walker, D. A., Walsh, J. E., Carmack, E. C., Frey, K. E., Meier, W. N., Moore, S. E., Parmentier, F.-J. W., Post, E., Romanovsky, V. E., and Simpson, W. R.: Implications of Arctic Sea Ice Decline for the Earth System, Annu. Rev. Environ. Resour., 39, 57–89, https://doi.org/10.1146/annurev-environ-122012-094357, 2014.
Bhatt, U. S., Walker, D. A., Raynolds, M. K., Bieniek, P. A., Epstein, H. E., Comiso, J. C., Pinzon, J. E., Tucker, C. J., Steele, M., Ermold, W., and Zhang, J.: Changing seasonality of panarctic tundra vegetation in relationship to climatic variables, Environ. Res. Lett., 12, 055003, https://doi.org/10.1088/1748-9326/aa6b0b, 2017.
Buchwal, A., Sullivan, P. F., Macias-Fauria, M., Post, E., Myers-Smith, I. H., Stroeve, J. C., Blok, D., Tape, K. D., Forbes, B. C., Ropars, P., Lévesque, E., Elberling, B., Angers-Blondin, S., Boyle, J. S., Boudreau, S., Boulanger-Lapointe, N., Gamm, C., Hallinger, M., Rachlewicz, G., Young, A., Zetterberg, P., and Welker, J. M.: Divergence of Arctic shrub growth associated with sea ice decline, Proc. Natl. Acad. Sci., 117, 33334–33344, https://doi.org/10.1073/pnas.2013311117, 2020.
Macias-Fauria, M., Forbes, B. C., Zetterberg, P., and Kumpula, T.: Eurasian Arctic greening reveals teleconnections and the potential for structurally novel ecosystems, Nat. Clim. Change, 2, 613–618, https://doi.org/10.1038/nclimate1558, 2012.
Macias-Fauria, M., Karlsen, S. R., and Forbes, B. C.: Disentangling the coupling between sea ice and tundra productivity in Svalbard, Sci. Rep., 7, 8586, https://doi.org/10.1038/s41598-017-06218-8, 2017.
Parmentier, F.-J. W., Zhang, W., Mi, Y., Zhu, X., Huissteden, J., Hayes, D. J., Zhuang, Q., Christensen, T. R., and McGuire, A. D.: Rising methane emissions from northern wetlands associated with sea ice decline, Geophys. Res. Lett., 42, 7214–7222, https://doi.org/10.1002/2015GL065013, 2015.
Rehder, Z., Niederdrenk, A. L., Kaleschke, L., and Kutzbach, L.: Analyzing links between simulated Laptev Sea sea ice and atmospheric conditions over adjoining landmasses using causal-effect networks, The Cryosphere, 14, 4201–4215, https://doi.org/10.5194/tc-14-4201-2020, 2020.
Screen, J. A., Simmonds, I., Deser, C., and Tomas, R.: The Atmospheric Response to Three Decades of Observed Arctic Sea Ice Loss, J. Clim., 26, 1230–1248, https://doi.org/10.1175/JCLI-D-12-00063.1, 2012.
Yu, L., Leng, G., and Python, A.: Varying response of vegetation to sea ice dynamics over the Arctic, Sci. Total Environ., 799, 149378, https://doi.org/10.1016/j.scitotenv.2021.149378, 2021.
Zhang, W., Döscher, R., Koenigk, T., Miller, P. A., Jansson, C., Samuelsson, P., Wu, M., and Smith, B.: The Interplay of Recent Vegetation and Sea Ice Dynamics—Results From a Regional Earth System Model Over the Arctic, Geophys. Res. Lett., 47, e2019GL085982, https://doi.org/10.1029/2019GL085982, 2020.Citation: https://doi.org/10.5194/egusphere-2025-1358-AC1
-
AC1: 'Reply on RC1', Annett Bartsch, 27 Jun 2025
-
RC2: 'Comment on egusphere-2025-1358', Anonymous Referee #2, 31 Jul 2025
The manuscript "Similarities between sea ice variations and satellite-derived terrestrial biosphere and cryosphere parameters across the Arctic" presents an analysis of primarily satellite-derived variables to examine the correlation between sea ice area in different basins across the Arctic, and NDVI, MAGT (mean annual ground temperature) and SWE. The objectives of the work were to extend earlier analyses to a pan-Arctic perspective and identify correlations between not only adjacent basins but also distant basins for all months of the year. A dataset has been produced (Arctic Sea Ice and Land Parameter Correlations - ASILaC), which appears to be the main product of the study and the authors state that this dataset will provide a baseline for future studies on common drivers of essential climate parameters and causative effects across the Arctic.
My main impression of the manuscript is that the work is essentially describing the results of a dataset that has been produced but it doesn't really go any further to examine or discuss the processes or mechanisms driving the correlations that have been found - for example why there can be simultaneously negative and positive correlations between sea ice in adjacent and distant basins and the terrestrial variables and what this means. A lot of the discussion comes across as a repeat description of the results, but with some comparison to previous studies, but it doesn't really shed new insight or provide suggestions for how the dataset could be utilised in future work to actually delve deeper into the "why" behind the results presented. I think this fine if it is the main objective of the work (as has been stated to a certain degree), and I think the title of the manuscript is appropriate, but as already mentioned by Reviewer 1, I feel that at the moment the manuscript, as it is, would work better as a data paper in a journal such as ESSD. I recommend that the main output of the analyses - the ASILaC dataset - be mentioned as one of the objectives in 1.2.4 as this appears to be the baseline that could be used in future studies.
Minor comments
Line 32 - missing references in "MORE"?
Line 44 - remove "thus" before "therefore"
Line 88 - first time CAVM is introduced. It is referenced but I think the full description of the abbreviation should be provided, as it was still not clear in Fig.1 what CAVM was referring to.
Line 194. I did not fully understand why the trend analysis was only applied to the number of frozen days between 1 March to 31 July. Is this a period when the number of frozen days is likely to be most variable over long time series?
Line 202 - The SWE data are provided at monthly time resolution but the authors have chosen to only use the March SWE as a proxy for the max SWE. Is there some reason for not analysing correlations between monthly SWE and monthly SIA? The authors discuss that high negative correlations of March SWE with SIA may have implications for wildlife and reindeer herding (Line 423-424) but presumably the variations of SWE throughout the whole winter period (and not just March) would also be just as important from this perspective?
Line 237. change "are" to "area"
Line 250. "statistiscal" to "statistical"
Fig.4, 5, 6, 7. Perhaps add into the caption that the numbers eg. Beau-1, Beau-2, Beau-3, .... are referring to the correlations for the different months of the year. I appreciate this might be quite obvious to most readers but I didn't get it immediately.
Citation: https://doi.org/10.5194/egusphere-2025-1358-RC2 -
AC2: 'Reply on RC2', Annett Bartsch, 04 Aug 2025
Reviewer comment: My main impression of the manuscript is that the work is essentially describing the results of a dataset that has been produced but it doesn't really go any further to examine or discuss the processes or mechanisms driving the correlations that have been found - for example why there can be simultaneously negative and positive correlations between sea ice in adjacent and distant basins and the terrestrial variables and what this means. A lot of the discussion comes across as a repeat description of the results, but with some comparison to previous studies, but it doesn't really shed new insight or provide suggestions for how the dataset could be utilised in future work to actually delve deeper into the "why" behind the results presented.
- Reply: we agree that the discussion needs to be restructured and extended with reference to more relevant studies (see also reviewer #1 comments and response)
Reviewer comment: I think this fine if it is the main objective of the work (as has been stated to a certain degree), and I think the title of the manuscript is appropriate, but as already mentioned by Reviewer 1, I feel that at the moment the manuscript, as it is, would work better as a data paper in a journal such as ESSD. I recommend that the main output of the analyses - the ASILaC dataset - be mentioned as one of the objectives in 1.2.4 as this appears to be the baseline that could be used in future studies.
- Reply: We agree that the creation of the dataset would be useful to add as an additional objective.
- Note, beyond the creation of the dataset, key novel points are the consideration of vegetation and cryosphere at the same time, previous studies have focused on one parameter only. A comparison through a consistent setup for vegetation and land cryosphere parameters was so far not made. This resulted in the identification of so far undocumented potential linkages, which led us to the decision to submit to The Cryosphere instead of ESSD.
Reviewer comment: Line 194. I did not fully understand why the trend analysis was only applied to the number of frozen days between 1 March to 31 July. Is this a period when the number of frozen days is likely to be most variable over long time series?
- Reply: This period characterizes spring thaw on the northern hemisphere (e.g. Mortin et al. 2012). A change in number of frozen days in this period represents a change in spring timing.
Mortin, J., T. M. Schrøder, A. Walløe Hansen, B. Holt, and K. C. McDonald (2012), Mapping of seasonal freeze-thaw transitions across the pan-Arctic land and sea ice domains with satellite radar, J. Geophys. Res., 117, C08004, doi:10.1029/2012JC008001.
Reviewer comment: Line 202 - The SWE data are provided at monthly time resolution but the authors have chosen to only use the March SWE as a proxy for the max SWE. Is there some reason for not analysing correlations between monthly SWE and monthly SIA? The authors discuss that high negative correlations of March SWE with SIA may have implications for wildlife and reindeer herding (Line 423-424) but presumably the variations of SWE throughout the whole winter period (and not just March) would also be just as important from this perspective?
- Reply: We have chosen March SWE for consistency with the annual representation of the other land parameters (mean annual ground temperature, TI-NDVI and Max-NDVI are annual measures). But we agree, it would be useful to also look into monthly SWE in a follow on study.
Citation: https://doi.org/10.5194/egusphere-2025-1358-AC2
-
AC2: 'Reply on RC2', Annett Bartsch, 04 Aug 2025
Status: closed
-
RC1: 'Comment on egusphere-2025-1358', Anonymous Referee #1, 04 Jun 2025
The manuscript by Bartsch et al. describes how pan-Arctic datasets of mean annual ground temperature at 2m depth (MAGT), snow water equivalent (SWE) and NDVI (as a proxy for plant growth) correlate with sea ice area (SIA). Sea ice loss is one of the main causes of the amplified warming of the Arctic, and together with changes in atmospheric humidity this influences MAGT, SWE and plant growth. Such links have been shown previously from observations, remote sensing and models (see e.g. Bhatt et al., 2010, 2014, 2017; Buchwal et al., 2020; Macias-Fauria et al., 2012, 2017; Parmentier et al., 2015; Rehder et al., 2020; Screen et al., 2012; Yu et al., 2021). This study aims to differentiate itself from this previous work by using satellite data where possible, and by focusing more on regional correlations rather than those made across the whole Arctic.
While I appreciate the attempt by the authors to look further into this topic, I feel that the manuscript in its current form is a missed opportunity to learn something truly novel. In particular, I had hoped that this study would go beyond mere correlations by identifying causal links, and by showing more detail. More detailed regional analyses between sea ice and the terrestrial environment have been done for example by Parmentier et al. (2015) who performed a pan-Arctic pixel-wise correlation between local sea ice conditions and temperature and modeled methane emissions, and who argued a causal link in autumn but not in spring. Rehder et al. (2020) used causal-effect networks to identify temporal links to the land near the Laptev Sea, and showed that spring-time correlations in sea ice and atmospheric variables were both related to large scale atmospheric circulation, not to each other, although sea ice loss had a weak effect on the near coastal environment in summer. Regional links between NDVI and sea ice have also been shown before (see e.g. Yu et al. 2021 and the paper by one of the co-authors of this study, Macias-Fauria et al. 2017). In addition, see also chapter 10 of the 2017 AMAP report (the authors incorrectly state on line 45 that this report did not include vegetation trends). Btw, reverse links have also been argued, where terrestrial vegetation growth lowers surface albedo, affecting climate and subsequently sea ice loss (Zhang et al., 2020).
While many of these previous studies relied on models or reanalysis datasets, this study aims to use remote-sensing datasets as much as possible. However, the authors use the TTOP model to determine soil temperature at 2 m depth. While this model uses land surface temperature (LST) from MODIS as an input, it also uses reanalysis data when MODIS LST is unavailable. Moreover, it models the soil temperature depending on for example land cover and surface wetness. While the TTOP model is probably the best estimate we have for permafrost extent at the moment, it is still a (hybrid) model. If the authors wanted to compare to satellite data only, rather than reanalyses or models, it would have made more sense to compare to MODIS LST directly. Moreover, 2 m depth is rather deep in the Arctic, where the active layer is typically shallower than 1 m. Any warming signal would be strongly attenuated and lagged at 2 m depth, which makes it difficult to make instantaneous correlations.
The current study also shows correlations at short and long distances, but it is not clear whether these correlations have a common distant cause or whether they represent an internal dynamic in the Arctic. Are they due to large scale atmospheric circulation affecting both sea ice and the terrestrial variables? Or are they due to local feedbacks dominated by sea ice decline? Unfortunately, the answer to this question is left in the middle by the authors, who present the dataset as a baseline for further analyses of drivers and dependencies. The paper would have been much stronger if it included a proper discussion on the underlying causes for the apparent correlations.
While I generally appreciate the effort by the authors, they could have done a more detailed spatial analysis given the high spatial detail of the source data, and they should have provided better causal insights rather than showing correlations with little context. Unless the authors go beyond their basic presentation of the data, and given the fact that they urge others to take their dataset further, I feel like this manuscript in its current form would work better as a data paper (e.g. in ESSD) rather than as a research article.
Some further comments:
- Line 45: Vegetation is extensively discussed in the 2017 SWIPA report (see chapters 8 and 10).
- Line 81: wouldn’t frequent cloud cover be a problem for MaxNDVI as well? Easy to miss the peak season in frequent cloudy parts of the Arctic, adding uncertainty to interannual variability in peak NDVI values.
- Line 126-128: did this model result match the observations well?
- Line 150: which parameters? Reference?
- Line 157: what’s the pixel size?
- Line: 164-165: This dataset appears to contain only the trend over the entire time series, not the original high temporal data used for the correlations, and there are no details on how the underlying dataset was processed. Is there a reference describing this?
- Line 177-178: This assumption appears reasonable, but did you test whether it’s true?
- Figure 1: could you give the area north of 60°N a different color from the one south of 60°N? Makes it easier to see the domain instead of just highlighting a latitudinal band.
- Table 4: what’s the resolution of the source data for sea ice area?
- Line 201: entire northern hemisphere or only north of 60°N?
- Line 255: I can imagine that correlations to more distant basins may show up but here we need better argumentation for why these correlations appear because at these large distances it may just as well be large scale atmospheric circulation affecting both (e.g. teleconnections related to Rossby wave propagation).
- Line 269: a positive correlation may not be that surprising. For example, the Fram Strait is an area of sea ice export, which can actually be enhanced in warm summers because the strong sea ice melt makes the ice thinner and more mobile, subsequently leading to more export through this area. As such, a warmer Arctic leads to more sea ice in the Fram Strait, explaining positive correlations. These kinds of internal dynamics need to be considered when interpreting correlations to sea ice.
- Line 275: do you mean figure A7 instead of A2?
- Line 279-280: This is unclear. Relevant for TI-NDVI in what way?
- Line 336: Please specify why solar absorption trends are increasing in this region. Is it less cloud cover or changes in surface albedo from e.g. earlier snow melt and/or shrubification?
- Line 376-377: This sounds interesting, but what would be the reason for this temporal lag?
- Line 379-384: I’m not sure I’m following this. Why would warmer summers and increased absorption of radiation lead to regional cooling in the autumn?
- Line 407: Unclear. Which “following ones”?
- Line 433-435: I’m not sure how this agrees with Sasgen et al. (2024) since they explicitly state that they did not look at the influence of sea ice.
- Line 449: why define the abbreviation FT for “freeze-thaw” if you only use it one more time?
- Figure A2: please replace “source” in the caption with the actual reference.
- Table A2: what does the “2000?” mean in the table caption?References
Bhatt, U. S., Walker, D. A., Raynolds, M. K., Comiso, J. C., Epstein, H. E., Jia, G., Gens, R., Pinzon, J. E., Tucker, C. J., Tweedie, C. E., and Webber, P. J.: Circumpolar Arctic Tundra Vegetation Change Is Linked to Sea Ice Decline, Earth Interact., 14, 1–20, https://doi.org/10.1175/2010EI315.1, 2010.
Bhatt, U. S., Walker, D. A., Walsh, J. E., Carmack, E. C., Frey, K. E., Meier, W. N., Moore, S. E., Parmentier, F.-J. W., Post, E., Romanovsky, V. E., and Simpson, W. R.: Implications of Arctic Sea Ice Decline for the Earth System, Annu. Rev. Environ. Resour., 39, 57–89, https://doi.org/10.1146/annurev-environ-122012-094357, 2014.
Bhatt, U. S., Walker, D. A., Raynolds, M. K., Bieniek, P. A., Epstein, H. E., Comiso, J. C., Pinzon, J. E., Tucker, C. J., Steele, M., Ermold, W., and Zhang, J.: Changing seasonality of panarctic tundra vegetation in relationship to climatic variables, Environ. Res. Lett., 12, 055003, https://doi.org/10.1088/1748-9326/aa6b0b, 2017.
Buchwal, A., Sullivan, P. F., Macias-Fauria, M., Post, E., Myers-Smith, I. H., Stroeve, J. C., Blok, D., Tape, K. D., Forbes, B. C., Ropars, P., Lévesque, E., Elberling, B., Angers-Blondin, S., Boyle, J. S., Boudreau, S., Boulanger-Lapointe, N., Gamm, C., Hallinger, M., Rachlewicz, G., Young, A., Zetterberg, P., and Welker, J. M.: Divergence of Arctic shrub growth associated with sea ice decline, Proc. Natl. Acad. Sci., 117, 33334–33344, https://doi.org/10.1073/pnas.2013311117, 2020.
Macias-Fauria, M., Forbes, B. C., Zetterberg, P., and Kumpula, T.: Eurasian Arctic greening reveals teleconnections and the potential for structurally novel ecosystems, Nat. Clim. Change, 2, 613–618, https://doi.org/10.1038/nclimate1558, 2012.
Macias-Fauria, M., Karlsen, S. R., and Forbes, B. C.: Disentangling the coupling between sea ice and tundra productivity in Svalbard, Sci. Rep., 7, 8586, https://doi.org/10.1038/s41598-017-06218-8, 2017.
Parmentier, F.-J. W., Zhang, W., Mi, Y., Zhu, X., Huissteden, J., Hayes, D. J., Zhuang, Q., Christensen, T. R., and McGuire, A. D.: Rising methane emissions from northern wetlands associated with sea ice decline, Geophys. Res. Lett., 42, 7214–7222, https://doi.org/10.1002/2015GL065013, 2015.
Rehder, Z., Niederdrenk, A. L., Kaleschke, L., and Kutzbach, L.: Analyzing links between simulated Laptev Sea sea ice and atmospheric conditions over adjoining landmasses using causal-effect networks, The Cryosphere, 14, 4201–4215, https://doi.org/10.5194/tc-14-4201-2020, 2020.
Screen, J. A., Simmonds, I., Deser, C., and Tomas, R.: The Atmospheric Response to Three Decades of Observed Arctic Sea Ice Loss, J. Clim., 26, 1230–1248, https://doi.org/10.1175/JCLI-D-12-00063.1, 2012.
Yu, L., Leng, G., and Python, A.: Varying response of vegetation to sea ice dynamics over the Arctic, Sci. Total Environ., 799, 149378, https://doi.org/10.1016/j.scitotenv.2021.149378, 2021.
Zhang, W., Döscher, R., Koenigk, T., Miller, P. A., Jansson, C., Samuelsson, P., Wu, M., and Smith, B.: The Interplay of Recent Vegetation and Sea Ice Dynamics—Results From a Regional Earth System Model Over the Arctic, Geophys. Res. Lett., 47, e2019GL085982, https://doi.org/10.1029/2019GL085982, 2020.Citation: https://doi.org/10.5194/egusphere-2025-1358-RC1 -
AC1: 'Reply on RC1', Annett Bartsch, 27 Jun 2025
Many thanks for the in depth review and comments! Please find our comments and responses to your questions below.
Reviewer comment: The manuscript by Bartsch et al. describes how pan-Arctic datasets of mean annual ground temperature at 2m depth (MAGT), snow water equivalent (SWE) and NDVI (as a proxy for plant growth) correlate with sea ice area (SIA). Sea ice loss is one of the main causes of the amplified warming of the Arctic, and together with changes in atmospheric humidity this influences MAGT, SWE and plant growth. Such links have been shown previously from observations, remote sensing and models (see e.g. Bhatt et al., 2010, 2014, 2017; Buchwal et al., 2020; Macias-Fauria et al., 2012, 2017; Parmentier et al., 2015; Rehder et al., 2020; Screen et al., 2012; Yu et al., 2021). This study aims to differentiate itself from this previous work by using satellite data where possible, and by focusing more on regional correlations rather than those made across the whole Arctic.
- Comment: Note, further key novel points are the consideration of vegetation and cryosphere at the same time, previous studies have focused on one parameter only. A comparison through a consistent setup for vegetation and land cryosphere parameters was so far not made. Also the temporary extension, evaluating all months in combination with consideration of sea ice basins goes beyond past studies. There has been an example for NDVI versus monthly sea ice (Yu et al. 2021), but without considering different sea ice basins.
Reviewer comment: While I appreciate the attempt by the authors to look further into this topic, I feel that the manuscript in its current form is a missed opportunity to learn something truly novel. In particular, I had hoped that this study would go beyond mere correlations by identifying causal links, and by showing more detail. More detailed regional analyses between sea ice and the terrestrial environment have been done for example by Parmentier et al. (2015) who performed a pan-Arctic pixel-wise correlation between local sea ice conditions and temperature and modeled methane emissions, and who argued a causal link in autumn but not in spring. Rehder et al. (2020) used causal-effect networks to identify temporal links to the land near the Laptev Sea, and showed that spring-time correlations in sea ice and atmospheric variables were both related to large scale atmospheric circulation, not to each other, although sea ice loss had a weak effect on the near coastal environment in summer.
- Reply: Thanks for pointing out the two studies. Parmentier et al. (2015) show that linkages might exist even further south than for the extent we have chosen plus the inclusion of sea ice across the Hudson bay and Canadian Archipelago (unfortunately this was not covered in the sea ice basin data set that we used), what should be addressed in the discussion in addition to comparison of the findings and discussion of potential causal links of Parmentier et al. (2015) as well as Rehder et al. (2020).
Reviewer comment: Regional links between NDVI and sea ice have also been shown before (see e.g. Yu et al. 2021 and the paper by one of the co-authors of this study, Macias-Fauria et al. 2017). In addition, see also chapter 10 of the 2017 AMAP report (the authors incorrectly state on line 45 that this report did not include vegetation trends). Btw, reverse links have also been argued, where terrestrial vegetation growth lowers surface albedo, affecting climate and subsequently sea ice loss (Zhang et al., 2020).
- Reply:
- Please note that we acknowledge the extensive previous work on NDVI in the paper, including the mentioned studies Yu et al. 2012 and Macias-Fauria et al. 2017 (see section 2.1 and Table 3). In addition, an in depth analyses has been, however, so far missing for the cryosphere parameters.
- Thanks for pointing out the misplacement of the phrasing on vegetation and AMAP. This sentence relates to Comiso and Hall (2014) mentioned before, so needs to be moved before the AMAP mentioning. Note, that we correctly list in Table A1 that the AMAP study considered vegetation.
- Thanks for pointing out Zhang et al., 2020, which should be considered in the discussion.
Reviewer comment: While many of these previous studies relied on models or reanalysis datasets, this study aims to use remote-sensing datasets as much as possible. However, the authors use the TTOP model to determine soil temperature at 2 m depth. While this model uses land surface temperature (LST) from MODIS as an input, it also uses reanalysis data when MODIS LST is unavailable. Moreover, it models the soil temperature depending on for example land cover and surface wetness. While the TTOP model is probably the best estimate we have for permafrost extent at the moment, it is still a (hybrid) model. If the authors wanted to compare to satellite data only, rather than reanalyses or models, it would have made more sense to compare to MODIS LST directly. Moreover, 2 m depth is rather deep in the Arctic, where the active layer is typically shallower than 1 m. Any warming signal would be strongly attenuated and lagged at 2 m depth, which makes it difficult to make instantaneous correlations.
- Reply:
- Indeed, the signal at 2m depth is attenuated. The reason to anyway use it comes from (1) that 2m depth is commonly used to represent permafrost presence (e.g. Obu et al. 2021, and product documentation) and (2) that initial observations in Bartsch et al. (2023) pointed to potentially high correlations with sea ice. And our results in the new manuscript have confirmed these and demonstrate a significantly higher linkage with sea ice than all other parameters, despite of the attenuation. But we agree that the issue of attenuation should be discussed.
- Note, that the datasets used come from ESA Permafrost CCI which were created using a transient model version of CryoGRID in order to obtain time series. A TTOP model was used in the past for the creation of the ESA DUE GlobPermafrost dataset representing equilibrium conditions 2000-2016, what cannot be used for our trend analyses.
Reviewer comment: The current study also shows correlations at short and long distances, but it is not clear whether these correlations have a common distant cause or whether they represent an internal dynamic in the Arctic. Are they due to large scale atmospheric circulation affecting both sea ice and the terrestrial variables? Or are they due to local feedbacks dominated by sea ice decline? Unfortunately, the answer to this question is left in the middle by the authors, who present the dataset as a baseline for further analyses of drivers and dependencies. The paper would have been much stronger if it included a proper discussion on the underlying causes for the apparent correlations.
- Reply: Literature suggests that the underlying dynamics differ between the analysed parameters. For example, as mentioned in the discussion, Sasgen et al. (2025) suggest large scale atmospheric circulation playing a role for ice sheets and permafrost. The linkage between sea ice in the proximity and SWE as well as NDVI has been pointed out regionally before. We agree that the discussion should be restructured to cover these aspects better and in more detail, also considering the above suggested references.
Reviewer comment: While I generally appreciate the effort by the authors, they could have done a more detailed spatial analysis given the high spatial detail of the source data, and they should have provided better causal insights rather than showing correlations with little context. Unless the authors go beyond their basic presentation of the data, and given the fact that they urge others to take their dataset further, I feel like this manuscript in its current form would work better as a data paper (e.g. in ESSD) rather than as a research article.
- Reply:
- Please note that with the present setup new results are shown for all parameters, and specifically for the cryosphere this type of correlation analyses was not done to date.
- Could you clarify what is meant with ‘more detailed spatial analysis given the high spatial detail of the source data’?
Further reviewer comments:
- Line 45: Vegetation is extensively discussed in the 2017 SWIPA report (see chapters 8 and 10).
- Reply: As mentioned above, the placement of the sentence is a mistake. This sentence refers to Comiso & Hall. We state the consideration of vegetation in AMAP/SWIPA in Table A1.
- Line 81: wouldn’t frequent cloud cover be a problem for MaxNDVI as well? Easy to miss the peak season in frequent cloudy parts of the Arctic, adding uncertainty to interannual variability in peak NDVI values.
- Reply: Yes, this is why we have used TI-NDVI which is bi-monthly (16 days periods) instead of growing season summed NDVI (GSSNDVI) which is constructed on a daily basis (Park et al. 2016). And it is indeed an issue to be mentioned also in the discussion.
- Line 126-128: did this model result match the observations well? - Reply: A decreasing amount of snow over North America was also found through satellite observations by Pulliaianen et al. (2020)
- Line 150: which parameters? Reference? - Suggestion for extended phrasing: … cryosphere as well as vegetation parameters 2000 onwards, due to specifically the availability of MODIS and denser coverage by Landsat (see e.g. Tables 2 and 3)
- Line 157: what’s the pixel size? - 25km, but we have been using for sea ice basins aggregated values.
- Line: 164-165: This dataset appears to contain only the trend over the entire time series, not the original high temporal data used for the correlations, and there are no details on how the underlying dataset was processed. Is there a reference describing this? - reply: We suggest to add the following text: A circumpolar and circumboreal time series of satellite data of AVHRR- GIMMS and MODIS in the period 1982-2019, respectively, were processed. The following parameters related to the growing season as the yearly maximum NDVI (MaxNDVI), and time integrated NDVI (TI-NDVI) were computed and analyzed from these data series. Maximum NDVI (MaxNDVI) is the annual maximum NDVI value observed during the period of peak phytomass during growing season, typically in late July and early August for the Arctic (Frost et al. 2025), and somewhat earlier for the Boreal region below 60° N. The following datasets were processed:
- Advanced Very High Resolution Radiometer (AVHRR), 1982–2019, in the form of the Global Inventory Modeling and Mapping Studies 3 g V1.2 dataset (GIMMS-3g+) with a spatial resolution of 1/12° (~8 km) (Pinzon and Tucker, 2014; Pinzon et al., 2023).
- Moderate Resolution Imaging Spectroradiometer (MODIS), 2000–2019, using 16-day NDVI product from the Terra (MOD13A1, v. 6.1) with a spatial resolution of 500 m. Pixels with a Summary Quality Assurance value ≥3 (indicating cloudy and compromised observations) were removed.
- Time-integrated NDVI (TI-NDVI) is the sum of maximum NDVI values within set compositing periods during May–September, calculated for datasets with daily temporal resolution. TI-NDVI includes phenological variations throughout the growing season; therefore, it better represents gross primary production (Tucker and Sellers, 1986), and is better correlated with climate variables than MaxNDVI (Bhatt et al., 2010; Bhatt et al., 2021). NDVImax and TI-NDVI may exhibit different correlations with climatic parameters like temperature (Yan et al. 2022). No filtering processes or other amendments have been done to the datasets and we have pretty much followed the procedure in Frost et al. (2025).
- Line 177-178: This assumption appears reasonable, but did you test whether it’s true?
- Reply: Considering the whole Northern hemisphere (as well as Eurasia and North America separately), the average March SWE was found to be the closest monthly average to peak SWE by Pulliainen et al., (2020) [see extended data Table 1)]. We suggest to added the following text::
- Average SWE in March was selected for this study and interpreted to represent maximum SWE conditions, following Pulliainen et al., (2020).
- Figure 1: could you give the area north of 60°N a different color from the one south of 60°N? Makes it easier to see the domain instead of just highlighting a latitudinal band.- Reply: We agree that this would be useful
- Table 4: what’s the resolution of the source data for sea ice area? - Reply: 25km
- Line 201: entire northern hemisphere or only north of 60°N? - Reply: Thanks for pointing out, it should be north of 60°N
- Line 255: I can imagine that correlations to more distant basins may show up but here we need better argumentation for why these correlations appear because at these large distances it may just as well be large scale atmospheric circulation affecting both (e.g. teleconnections related to Rossby wave propagation).
- Reply: Note that we come back to that in the discussion, e.g. see discussion of Sasgen et al. (2024)
- Line 269: a positive correlation may not be that surprising. For example, the Fram Strait is an area of sea ice export, which can actually be enhanced in warm summers because the strong sea ice melt makes the ice thinner and more mobile, subsequently leading to more export through this area. As such, a warmer Arctic leads to more sea ice in the Fram Strait, explaining positive correlations. These kinds of internal dynamics need to be considered when interpreting correlations to sea ice.
- Reply: Thanks for pointing out, we agree that this should be covered in the discussion.
- Line 275: do you mean figure A7 instead of A2?
- Reply: Yes, thanks for spotting!
- Line 279-280: This is unclear. Relevant for TI-NDVI in what way?
- Reply: for the adjacent land area
- Line 336: Please specify why solar absorption trends are increasing in this region. Is it less cloud cover or changes in surface albedo from e.g. earlier snow melt and/or shrubification? - Reply: This observation comes from Letterly et al. 2018 (based on reanalyses), who discuss earlier snow melt timing as a general reason for reduction of albedo in some areas across the Arctic.
- Line 376-377: This sounds interesting, but what would be the reason for this temporal lag? - Reply: There is a mistake here. ‘following’ should be replaced with ‘preceding’. The reasoning is provided at the end of the paragraph.
- Line 379-384: I’m not sure I’m following this. Why would warmer summers and increased absorption of radiation lead to regional cooling in the autumn? - Reply: The observation is that there are summers with high NDVI followed by earlier sea ice formation in the autumn. The processes listed are rather speculation, thus we suggest to remove the part ‘(such as …. 4.1)’ and change ‘turn into’ to ‘lead to’ and add a comment that more investigations are needed for clarification.
- Line 407: Unclear. Which “following ones”? - Reply: After June
- Line 433-435: I’m not sure how this agrees with Sasgen et al. (2024) since they explicitly state that they did not look at the influence of sea ice. - Reply: We refer here to long distance linkages in general, not specifically sea ice. This could be change for example to ‘Such long distance linkages across the Arctic were also found ice sheets and permafrost …’
- Line 449: why define the abbreviation FT for “freeze-thaw” if you only use it one more time? - Reply: Thanks for spotting
- Figure A2: please replace “source” in the caption with the actual reference. - Reply: Thanks for spotting
- Table A2: what does the “2000?” mean in the table caption? - Reply: Thanks for spotting. This column should be removed, this was from an old version of the table.
References in replies:
Bartsch, A., Strozzi, T., and Nitze, I.: Permafrost Monitoring from Space, Surveys in Geophysics, https://doi.org/10.1007/s10712-023-09770-3, 2023
Frost, G.V. , Bhatt Uma S. , Macander M.V., Logan, B.T. , Walker D.A., Raynolds, M.K., Magnússon, R. I , Bartsch, A., Bjerke , J.W., Epstein, H.E. , Forbes, B.C. , Goetz, S.J , Hoy, E.E. , Karlsen S.R. , Kumpula, T. , Lantz, T.C. , Lara, M. J. , López-Blanco, E., Montesano, P.M., Neigh, C.S.R. , Nitze, I., Orndahl, Kathleen M. , Park, T. , Phoenix, G.K. , Rocha, A.V. , Rogers, B. M., Schaepman-Strub, G., Tømmervik,H., Verdonen, M., Veremeeva, A., Virkkala, A-M., & Wiagl, C.F. 2025. The changing face of the Arctic: four decades of greening and implications for tundra ecosystems Frontiers in Environmental Research, 13 , 2025 DO10.3389/fenvs.2025.1525574
Letterly, A., Key, J., and Liu, Y.: Arctic climate: changes in sea ice extent outweigh changes in snow cover, The Cryosphere, 12, 3373–3382, https://doi.org/10.5194/tc-12-3373-2018, 2018.
Obu, J., Westermann, S., Barboux, C., Bartsch, A., Delaloye, R., Grosse, G., Heim, B., Hugelius, G., Irrgang, A., Kääb, A. M., Kroisleitner, C., Matthes, H., Nitze, I., Pellet, C., Seifert, F. M., Strozzi, T., Wegmüller, U., Wieczorek, M., and Wiesmann,A.: ESA Permafrost Climate Change Initiative (Permafrost_cci): Permafrost extent for the Northern Hemisphere, v3.0, CEDA, https://doi.org/10.5285/6e2091cb0c8b4106921b63cd5357c97c, 2021b.
Park, T., Ganguly, S., Tømmervik, H., Euskirchen, E.S., Høgda, K.A., Karlsen, S.R., Brovkin, V., Nemani , R.R., Myneni, R.B. 2016 Environ. Res. Lett. 11 084001
Pinzon, J. E., Tucker, C. J., Bhatt, U. S., Frost, G. V., and Macander, M. J. (2023).Global vegetation greenness (NDVI) from AVHRR GIMMS-3g+. NASA’s Open Data Portal, 1981–2022. doi:10.3334/ORNLDAAC/2187
Sasgen, I., Steinhoefel, G., Kasprzyk, C., Matthes, H., Westermann, S., Boike, J., and Grosse, G.: Atmosphere circulation patterns synchronize pan-Arctic glacier melt and permafrost thaw, Communications Earth & Environment, 5, 375, https://doi.org/10.1038/s43247-024- 01548-8, 2024.
Tucker, C. J., and Sellers, P. J. (1986). Satellite remote sensing of primary production. Int. J. Remote Sens. 16, 1395–1416. doi:10.1080/01431168608948944
Yan, J., Zhang, G., Ling, H., & Han, F. 2022. Comparison of time-integrated NDVI and annual maximum NDVI for assessing grassland dynamics, Ecological Indicators, 136, 2022, 108611https://doi.org/10.1016/j.ecolind.2022.108611.
References in reviewer comments:
Bhatt, U. S., Walker, D. A., Raynolds, M. K., Comiso, J. C., Epstein, H. E., Jia, G., Gens, R., Pinzon, J. E., Tucker, C. J., Tweedie, C. E., and Webber, P. J.: Circumpolar Arctic Tundra Vegetation Change Is Linked to Sea Ice Decline, Earth Interact., 14, 1–20, https://doi.org/10.1175/2010EI315.1, 2010.
Bhatt, U. S., Walker, D. A., Walsh, J. E., Carmack, E. C., Frey, K. E., Meier, W. N., Moore, S. E., Parmentier, F.-J. W., Post, E., Romanovsky, V. E., and Simpson, W. R.: Implications of Arctic Sea Ice Decline for the Earth System, Annu. Rev. Environ. Resour., 39, 57–89, https://doi.org/10.1146/annurev-environ-122012-094357, 2014.
Bhatt, U. S., Walker, D. A., Raynolds, M. K., Bieniek, P. A., Epstein, H. E., Comiso, J. C., Pinzon, J. E., Tucker, C. J., Steele, M., Ermold, W., and Zhang, J.: Changing seasonality of panarctic tundra vegetation in relationship to climatic variables, Environ. Res. Lett., 12, 055003, https://doi.org/10.1088/1748-9326/aa6b0b, 2017.
Buchwal, A., Sullivan, P. F., Macias-Fauria, M., Post, E., Myers-Smith, I. H., Stroeve, J. C., Blok, D., Tape, K. D., Forbes, B. C., Ropars, P., Lévesque, E., Elberling, B., Angers-Blondin, S., Boyle, J. S., Boudreau, S., Boulanger-Lapointe, N., Gamm, C., Hallinger, M., Rachlewicz, G., Young, A., Zetterberg, P., and Welker, J. M.: Divergence of Arctic shrub growth associated with sea ice decline, Proc. Natl. Acad. Sci., 117, 33334–33344, https://doi.org/10.1073/pnas.2013311117, 2020.
Macias-Fauria, M., Forbes, B. C., Zetterberg, P., and Kumpula, T.: Eurasian Arctic greening reveals teleconnections and the potential for structurally novel ecosystems, Nat. Clim. Change, 2, 613–618, https://doi.org/10.1038/nclimate1558, 2012.
Macias-Fauria, M., Karlsen, S. R., and Forbes, B. C.: Disentangling the coupling between sea ice and tundra productivity in Svalbard, Sci. Rep., 7, 8586, https://doi.org/10.1038/s41598-017-06218-8, 2017.
Parmentier, F.-J. W., Zhang, W., Mi, Y., Zhu, X., Huissteden, J., Hayes, D. J., Zhuang, Q., Christensen, T. R., and McGuire, A. D.: Rising methane emissions from northern wetlands associated with sea ice decline, Geophys. Res. Lett., 42, 7214–7222, https://doi.org/10.1002/2015GL065013, 2015.
Rehder, Z., Niederdrenk, A. L., Kaleschke, L., and Kutzbach, L.: Analyzing links between simulated Laptev Sea sea ice and atmospheric conditions over adjoining landmasses using causal-effect networks, The Cryosphere, 14, 4201–4215, https://doi.org/10.5194/tc-14-4201-2020, 2020.
Screen, J. A., Simmonds, I., Deser, C., and Tomas, R.: The Atmospheric Response to Three Decades of Observed Arctic Sea Ice Loss, J. Clim., 26, 1230–1248, https://doi.org/10.1175/JCLI-D-12-00063.1, 2012.
Yu, L., Leng, G., and Python, A.: Varying response of vegetation to sea ice dynamics over the Arctic, Sci. Total Environ., 799, 149378, https://doi.org/10.1016/j.scitotenv.2021.149378, 2021.
Zhang, W., Döscher, R., Koenigk, T., Miller, P. A., Jansson, C., Samuelsson, P., Wu, M., and Smith, B.: The Interplay of Recent Vegetation and Sea Ice Dynamics—Results From a Regional Earth System Model Over the Arctic, Geophys. Res. Lett., 47, e2019GL085982, https://doi.org/10.1029/2019GL085982, 2020.Citation: https://doi.org/10.5194/egusphere-2025-1358-AC1
-
AC1: 'Reply on RC1', Annett Bartsch, 27 Jun 2025
-
RC2: 'Comment on egusphere-2025-1358', Anonymous Referee #2, 31 Jul 2025
The manuscript "Similarities between sea ice variations and satellite-derived terrestrial biosphere and cryosphere parameters across the Arctic" presents an analysis of primarily satellite-derived variables to examine the correlation between sea ice area in different basins across the Arctic, and NDVI, MAGT (mean annual ground temperature) and SWE. The objectives of the work were to extend earlier analyses to a pan-Arctic perspective and identify correlations between not only adjacent basins but also distant basins for all months of the year. A dataset has been produced (Arctic Sea Ice and Land Parameter Correlations - ASILaC), which appears to be the main product of the study and the authors state that this dataset will provide a baseline for future studies on common drivers of essential climate parameters and causative effects across the Arctic.
My main impression of the manuscript is that the work is essentially describing the results of a dataset that has been produced but it doesn't really go any further to examine or discuss the processes or mechanisms driving the correlations that have been found - for example why there can be simultaneously negative and positive correlations between sea ice in adjacent and distant basins and the terrestrial variables and what this means. A lot of the discussion comes across as a repeat description of the results, but with some comparison to previous studies, but it doesn't really shed new insight or provide suggestions for how the dataset could be utilised in future work to actually delve deeper into the "why" behind the results presented. I think this fine if it is the main objective of the work (as has been stated to a certain degree), and I think the title of the manuscript is appropriate, but as already mentioned by Reviewer 1, I feel that at the moment the manuscript, as it is, would work better as a data paper in a journal such as ESSD. I recommend that the main output of the analyses - the ASILaC dataset - be mentioned as one of the objectives in 1.2.4 as this appears to be the baseline that could be used in future studies.
Minor comments
Line 32 - missing references in "MORE"?
Line 44 - remove "thus" before "therefore"
Line 88 - first time CAVM is introduced. It is referenced but I think the full description of the abbreviation should be provided, as it was still not clear in Fig.1 what CAVM was referring to.
Line 194. I did not fully understand why the trend analysis was only applied to the number of frozen days between 1 March to 31 July. Is this a period when the number of frozen days is likely to be most variable over long time series?
Line 202 - The SWE data are provided at monthly time resolution but the authors have chosen to only use the March SWE as a proxy for the max SWE. Is there some reason for not analysing correlations between monthly SWE and monthly SIA? The authors discuss that high negative correlations of March SWE with SIA may have implications for wildlife and reindeer herding (Line 423-424) but presumably the variations of SWE throughout the whole winter period (and not just March) would also be just as important from this perspective?
Line 237. change "are" to "area"
Line 250. "statistiscal" to "statistical"
Fig.4, 5, 6, 7. Perhaps add into the caption that the numbers eg. Beau-1, Beau-2, Beau-3, .... are referring to the correlations for the different months of the year. I appreciate this might be quite obvious to most readers but I didn't get it immediately.
Citation: https://doi.org/10.5194/egusphere-2025-1358-RC2 -
AC2: 'Reply on RC2', Annett Bartsch, 04 Aug 2025
Reviewer comment: My main impression of the manuscript is that the work is essentially describing the results of a dataset that has been produced but it doesn't really go any further to examine or discuss the processes or mechanisms driving the correlations that have been found - for example why there can be simultaneously negative and positive correlations between sea ice in adjacent and distant basins and the terrestrial variables and what this means. A lot of the discussion comes across as a repeat description of the results, but with some comparison to previous studies, but it doesn't really shed new insight or provide suggestions for how the dataset could be utilised in future work to actually delve deeper into the "why" behind the results presented.
- Reply: we agree that the discussion needs to be restructured and extended with reference to more relevant studies (see also reviewer #1 comments and response)
Reviewer comment: I think this fine if it is the main objective of the work (as has been stated to a certain degree), and I think the title of the manuscript is appropriate, but as already mentioned by Reviewer 1, I feel that at the moment the manuscript, as it is, would work better as a data paper in a journal such as ESSD. I recommend that the main output of the analyses - the ASILaC dataset - be mentioned as one of the objectives in 1.2.4 as this appears to be the baseline that could be used in future studies.
- Reply: We agree that the creation of the dataset would be useful to add as an additional objective.
- Note, beyond the creation of the dataset, key novel points are the consideration of vegetation and cryosphere at the same time, previous studies have focused on one parameter only. A comparison through a consistent setup for vegetation and land cryosphere parameters was so far not made. This resulted in the identification of so far undocumented potential linkages, which led us to the decision to submit to The Cryosphere instead of ESSD.
Reviewer comment: Line 194. I did not fully understand why the trend analysis was only applied to the number of frozen days between 1 March to 31 July. Is this a period when the number of frozen days is likely to be most variable over long time series?
- Reply: This period characterizes spring thaw on the northern hemisphere (e.g. Mortin et al. 2012). A change in number of frozen days in this period represents a change in spring timing.
Mortin, J., T. M. Schrøder, A. Walløe Hansen, B. Holt, and K. C. McDonald (2012), Mapping of seasonal freeze-thaw transitions across the pan-Arctic land and sea ice domains with satellite radar, J. Geophys. Res., 117, C08004, doi:10.1029/2012JC008001.
Reviewer comment: Line 202 - The SWE data are provided at monthly time resolution but the authors have chosen to only use the March SWE as a proxy for the max SWE. Is there some reason for not analysing correlations between monthly SWE and monthly SIA? The authors discuss that high negative correlations of March SWE with SIA may have implications for wildlife and reindeer herding (Line 423-424) but presumably the variations of SWE throughout the whole winter period (and not just March) would also be just as important from this perspective?
- Reply: We have chosen March SWE for consistency with the annual representation of the other land parameters (mean annual ground temperature, TI-NDVI and Max-NDVI are annual measures). But we agree, it would be useful to also look into monthly SWE in a follow on study.
Citation: https://doi.org/10.5194/egusphere-2025-1358-AC2
-
AC2: 'Reply on RC2', Annett Bartsch, 04 Aug 2025
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 451 | 72 | 15 | 538 | 14 | 34 |
- HTML: 451
- PDF: 72
- XML: 15
- Total: 538
- BibTeX: 14
- EndNote: 34
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1