|Review of “The Arctic sea ice extent change connected to Pacific decadal variability” by Yang et al.|
Summary: This paper present as an analysis of changes in western Arctic sea ice over the satellite record and linkages to changes in spring Bering Sea ice extent and modes of variability of Pacific Ocean waters. A notable change in behavior occurred since 2007 resulting in reduced summer ice in the Chukchi Sea and during winter an enhanced interannual variability in the Bering Sea, with higher than normal extent during 2007-2012 and lower during 2013-2018. These changes in ice correlated with changes in Pacific Ocean properties, namely a phase-locking between the Pacific Decadal Oscillation and the North Pacific Gyre Oscillation.
General Comment: This paper is vastly improved over the initial submission. The period has been updated to include the recent shift toward record low Bering Sea ice extents and the analysis is far more detailed and comprehensive. I only noted a few minor details to be to addressed before publication. I applaud the hard work of the authors in addressing my comments and improving the paper.
Intro, 1st sentence: “indicators of global climate change”
Intro, last paragraph, 1st sentence: use “rapid sea ice loss” instead of “fast sea ice loss” to avoid confusion with the “fast sea ice” (fast ice) type.
Sec. 2, first paragraph: Just use “NASA”, spelling out not needed. However, I would suggest spelling out what DMSP, SSMI, and SSMIS stand for. Also remove “team” after “NASA”. What you mean (I think) is that the concentrations are derived using the “NASA Team” algorithm from the passive microwave radiances.
Sec. 2. On the sea ice data, did you use monthly gridded concentration fields and then calculate the monthly extents from those? That is okay, but can introduce a bias because the monthly gridded concentrations conflate temporal and spatial aspects. For example, a grid cell that is has 100% concentration for 5 days in a 30-day month exceeds the 15% threshold and will be considered “ice-covered” for the month and part of the extent. But that’s very different than a cell that has 15% ice each day of the month. Ideally, when calculating monthly total sea ice extent, it is best to calculate individually for each day of the month and then average the daily total extents over the month. This is what is done for the NSIDC Sea Ice Index (https://nsidc.org/data/seaiceice_index/archives), which includes regional analyses for the Bering Sea and the Chukchi Sea (not sure though if those regions are the same as what you use). Of course, for the spatial maps of concentration, you would use the monthly grids. As noted, for extent, there could be a bias. If it’s possible to redo the analysis with the SII numbers – if they fit your spatial region criteria – that would be best. But at the least, I think noting how you calculated the extent – from the monthly average concentration maps.
Sec. 3.1. It sounds like you used Nov. 1978 to Jun. 2019. This gives the maximum length, which is nice. But, with partial years at each end, the annual average trends are going to be skewed some. Probably not much, but for annual average it’s better to include only full years (e.g., 1979-2018). For the seasonal, it is okay as long as any season is full. But you do have for example (Figure 1 and 2) the MAM period, which presumably would be 1979-2019, and the ASO period, which would be 1979-2018. I might suggest sticking with full calendar years for consistency throughout: 1979-2018. Another factor is that the 2019 is at the moment preliminary near-real-time data, which will be replaced with final fields. So, the 2019 values will change. This change is generally small and the NRT product is meant to be as consistent as possible as the final long-term record. Thus, I doubt it will affect the analysis. But if 2019 is used, then it should also be cited: http://nsidc.org/data/nsidc-0081.html.
Sec. 3.1, paragraph starting “In view of…”: “the eastern Siberian Sea” is mentioned. Is this the specific “East Siberian Sea” or the eastern part of some Siberian area? If it is the former, changed “eastern” to “East”.
Sec. 3.1, paragraph starting “Fig. 2 shows…”: “collaborate to the” is awkward. Maybe “result in the”?
Sec. 3.3, 2nd paragraph: “deepening of the Aleutian Low” (add “the”)
Sec. 3.3, paragraph starting “Yeo et al. (2014)…”: remove “until” before “after ~2007 when…”
Sec. 3.4, paragraph starting “A NPGO-related…”: “open water owing to the brine-rejection effect…” – need to add “the”