Variations in Satellite Derived Snow and Ice Coverage in the Northern Hemisphere

One of the advantages of using the new MEaSURES Cryospheric dataset is determining the satellite derived extent of snow on land and sea ice on ocean surfaces from the same format for the Northern Hemisphere. The MEaSURES dataset is a combination of several satellite platforms and frequencies. Since the MEaSURES cryospheric dataset is gridded to the EASE-II grid, the extent locations of the snow and sea ice coverage can be compared between two normally independent observations. Therefore, a total areal extent of land snow and sea ice totals can be calculated between 1979 and 2012. The area calculations of snow and ice extent are done by accumulating each grid box reported weekly prior to 1999, and daily since 1999 for the Northern Hemisphere. The extent for the sea ice coverage is than added to the snow extent to give a total extent coverage. The cryospheric total extent coverage, we will call, cryototal for this paper, can then be determined. It should be noted that fresh water lakes as well as the Greenland Ice Sheet are not included in the cryototal area. The yearly cryototal is predominately driven by the maximum snow cover extent in the winter period and the minimum sea ice cover extent in the summer period, though there are variations in the area throughout the annual cycle.

The maximum cryototal extents and dates of occurrence are determined by taking the maximum daily/weekly extents between 1 January and 1 June. The annual maximum cryototal extent is approximately 62 million km2, usually occurring by late January through the mid-February for the study period. The snow coverage maximum extent is around 47 million km2 and the sea ice coverage extent would be just under 16 million km2. The cryototal, snow, and sea ice extent maximum areas each year do not vary much throughout the study period and linear regression analyses found no trends in the extent. However, when the date of the maximum area is collected, the dates of the maximum extent have been changing over the study period. With all three parameters, the date of the maximum area has been later in the year throughout the study period. For the maximum cryototal area, earlier years were more likely to have their maximum extent occur in mid-January, while later years had their maximum extent occur closer to the middle to end of February. It should be remembered that the earlier years, pre-1999, the extents are only calculated on a weekly basis which could be effecting the dates, however, the differences should only be fewer than 7 days, which is not the case. The maximum snow and sea ice extents are similar, with little change over time of the total amount of land or sea ice covered. However, linear regression trends for dates of the maximum extent indicate later occurrences in the winter; in January for snow extents, and late February to middle March for sea ice extents.

                The annual minimum cryototal extent for the study period has been declining, especially over the last 10 years of the study. The cryototal extent is around 10 million km2 in 1980 and falls to under 6 million km2 by 2012, with a study mean extent of just under 9 million km2. The decline in the cryototal area is basically a result of the decline in sea ice area extent, which falls from 7 million km2 to under 3 million km2 by the end of the period. The minimum extent of snow cover does not change over the study period. The date of the annual minimum cryototal extent is generally around the end of August to the beginning of September with no trend in the date. The date of the minimum snow extent is the middle of August and also has no trend in the date. Meanwhile, the date of the minimum sea ice area does have a trend, changing from the end of August to the beginning of September for the study period. 

                The melt season length, here calculated as the time from the maximum extent to the minimum extent for each parameter.  For the cryototal melt season length, the mean would be 213 days, ranging from 189 to 238 days. There is a very slight linear regression trend from longer cryototal melt seasons in the earlier years compared to the later years. The snow melt season length is very similar to the cryototal length with the mean length being 204 days with a range of 90 days over the study period. The trend is almost identical to the cryototal melt season trend. The sea ice melt season length is approximately 187 days, however, there is no trend indicated in the length of the melt season. This seems counter intuitive since the date of minimum extent is later in the year, indicating a longer melt season, though in reality the maximum sea ice extent date is also later in the year so the length of season stays the same.

                The freeze-up season is indicating similar features in the snow coverage and cryototal area, where the length of the accumulation of snow coverage starts to increase after the end of July until it maximum area during the winter. The length of time for the accumulation of snow to take place has been increasing over the period which follows the decrease in time for the melting. For sea ice extent, the length of time between the minimum area and the maximum area for the following winter is roughly the same length during the period, which corresponds well to the lack of a melt season change.

                In summary, the solar radiation and geographic conditions have a large role on the changes in maximum extents that occurs each year. Minimum extents especially for sea ice coverage are changing compared to the snow cover.  The changes in the timing of the maximum and minimum extents indicates an interesting finding which needs to be further investigated. Furthermore, how each parameter changes because of the other also needs to be investigated.