Changes in Seasonal Snow

During the past century, snowfall has been well monitored in populated regions of Europe, Russia, and North America. Much of this information is archived as snow depth or fresh snowfall totals, from which snow-water equivalence (SWE) must be estimated. Beginning in the 1960s, visible-wavelength satellite imagery has provided detailed mapping of snow-covered area in the Northern Hemisphere (see chapter 1).

Because snow extends to low elevations and latitudes, snow cover is very sensitive to warming or cooling. This is clear in the satellite record, which shows major declines in snow-covered area in spring and summer in the Northern Hemisphere. Spring melt shifted earlier by more than 2 weeks since the early 1970s, and the time of peak snow-covered area has shifted from February to January. The largest snowpack decline has occurred in midlatitude regions, roughly corresponding with the March-April 0°C to 5°C isotherms in the band 40° N to 60° N. The rate of change of Northern Hemisphere snow cover in these months is -0.62 x 106 km2 per decade

(1967-2010). Summer (June through August) snowpack declined at -0.79 x 106 km2 per decade from 1972 to 2010. This is equivalent to snow cover reductions of 7.5% and 52% in the spring and summer, respectively, relative to the mean values over the period of record. Winter snowpack is more robust, as temperatures remain below freezing from November to February in most areas of the Northern Hemisphere that experience seasonal snowfall. Annual average snow cover declined at a rate of -0.34 x 106 km2 per decade from 1972 to 2010: a net loss of 5.2% for this period.

Observed reductions in snow cover are not spatially uniform. Snowfall has had no statistically significant changes in some regions, and others have seen increases in snow cover. Southern Canada (below 60° N) experienced an average of 78 snow days per year (days with solid precipitation) for the period 1955-2004, but station data indicate opposite trends in western and eastern parts of the country. Snow events in western Canada declined by 2.9 days per decade during this period, whereas central and eastern Canada experienced an increase of 1.7 snow days per decade: respective changes of -14.5 and +8.5 days over the 50-year period. The frequency of snow days does not provide direct information on total snowfall, as SWE may change with little or no shift in snow-covered area or snow frequency. This does not seem to be the general case in midlatitudes, however, as snow cover, frequency, and depth are declining in conjunction, whereas winter rainfall events are becoming more frequent.

There are limited direct observations of snowpack trends at high elevations. Terrain variability in alpine environments prohibits straightforward satellite assessments of snowpack. Snow transect and snow-pillow data in the western United States and in the European Alps indicate a general decline in snowpack at all elevations in the latter half of the 20th century, with local exceptions. Declining glaciers and earlier peak runoff in stream flow from mountain headwater regions are consistent with diminishing snow accumulations at high altitudes, but these observations could also be a consequence of warmer temperatures in spring and summer.

The record from the Southern Hemisphere is also sparse. Snowpack measurements dating from the early 1960s in the mountainous regions of southeastern Australia indicate similar trends to northern midlatitudes: no major changes in maximum winter snowpack, but significant declines in snow cover and SWE in late winter and spring. The signal from the mountain snowpack in New Zealand and the Andes is not as clear. Interan-nual variability is high in these regions, governed by South Pacific synoptic variability (e.g., ENSO cycles), potentially masking decadal-scale trends.

0 0

Post a comment