Rutgers University is the curator of the NASA / NOAA northern hemisphere snow cover data base that can be accessed via their excellent web site Global Snow Lab. The left margin of the home page allows access to daily, weekly and monthly maps, monthly departure (anomaly map), snow and monthly anomalies (charts) and a data download link. Anomalies are based on a comparison with the 1981 to 2010 mean. All in all it is an impressive resource.
I have not come across this valuable resource before and seldom if ever hear or see the data discussed. Intuitively I feel that snow cover should be a sensitive indicator for climate change and global warming. Where I stay in NE Scotland is on the edge of the northern hemisphere snow belt. Sometimes, when it is cold, we will get a snowy winter, like this year. Other years we get no snow at all. Snow is a sensitive indicator for climate. So what does the data have to say? At first glance remarkably little (see chart below). The mid-winter peaks and late summer troughs have been remarkably stable for a planet rumoured to be melting under the burden of atmospheric CO2 and it is necessary to interrogate the data in some fine detail to tease out the interesting story that the data have to tell.
In summary, for the six months September to February snow area has actually been increasing 1967 to 2014! That has to be a surprise. And for the six months March to August snow area has been decreasing. The trends are generally very gradual and barely significant. But what the data show is that the northern hemisphere is getting snowier winters accompanied by more rapid melt in spring and summer. The latter is not surprising since we know that the lower troposphere is warming (at least we think we know that to be the case).
The annual cycle in N hemisphere snow cover varies between 3 million sq kms (August) and 50 million sq kms (January). Note that the small negative gradient of the regression is biased by the missing summer data for 1968, 1969 and 1971 although beginning the regression in 1972, a small negative gradient remains.
I want to start with a few key observations from the 12 monthly charts shown below.
- January is normally but not always the month with highest snow cover. Occasionally February takes the prize.
- Scrolling down through the months from February the amount of snow cover reduces quickly as it melts reaching a minimum in August. August is always the low point.
- August is an important marker month that will differentiate between a warming and a cooling world. When August snow area begins to increase this would mark the beginning of a new era of snow accumulation.
- From September the snow area expands again. Winter comes early in some parts of the world.
- The data time series for each month are actually remarkably uniform and, as already has been mentioned for the six months September to February, snow area has been increasing 1967 to 2014. And for the six months March to August snow area has been decreasing with time.
To get a handle on what is actually going on it is necessary to look at the maps and in particular the anomaly maps that show where change is taking place (see below).
August is an important month and maps for August 1967 and 2014 are shown below. It is worth looking at 2014 first. There was 2.59 million sq kms of snow and of that, 2.17 was on Greenland. The rest of the northern hemisphere was effectively snow free; presumably the summer snow fields on mountain glaciers are too small to be picked up at this resolution. Compared with August 1967 it can be seen that the Rocky Mountains, the Alps, the Himalayas and Baffin and Ellesmere Islands in Northern Canada had August snow back in 1967 that now melts.
This shows up on the anomaly maps as positive anomalies on the mountains and islands in 1967 and negative anomalies in these same areas in 2014. Looking at the chart for August (above) it can be seen that there is a step down in 1982 and that August snow area has changed little since then. One of the main reasons for the step down in 1982 was the loss of August snow on Baffin Island.
August snow maps
Click on maps and they will open in separate browser window that will ease comparison between pairs of maps.
August snow anomalies
A look at the January chart above shows that January snow cover has been consistently between 45 and 50 million square kms since 1967, though a regression shows a gradual increase with time. However, January 1967 ticks up while January 2014 ticks down and this gives a false picture on the anomaly maps displayed below. I therefore also include anomaly maps for 1968 and 2013 which reverses this random annual bias.
The whole of Russia and the whole of Canada tend to be covered in snow in January throughout the whole period since 1967. The anomalies show up in the areas south of 100% cover. In January 2013, positive anomalies dominated and were spread across the whole apron south of the 100% cover line most notably in the USA, west Europe, east China, Korea and Japan. In January 1968, the Rocky Mountains of the USA, the UK, Korea and Japan show up as negative anomalies. That year there was a strong positive anomaly over S Mongolia.
January snow maps
January anomaly maps
Snow and CO2
The figure below shows how there is no obvious connection between rising CO2 levels and declining snow cover. This of course does not prevent the IPCC from claiming that there is.
Looking in greater detail we see that the snow cycle and CO2 cycles are out of phase. While both are responding to seasonal change there is a greater time lag in the CO2 signal to the orbital / solar stimulus.
Perhaps the most significant aspect about the trends in global snow cover since 1967 is how little it has actually changed. Even though the January maximum shows a positive gradient, it is effectively a flat line. The January anomaly distribution is effectively down to “random weather”.
The August snow minimum has also been invariant since 1982. Virtually all northern hemisphere snow melts apart from Greenland. The August minimum therefore may not be a sensitive indicator since it will only begin to recede again with a shrinkage of the Greenland ice cap. The more rapid spring and summer melt that is most pronounced in the May, June and July data are therefore perhaps the clearest sign of a warming world. The fact that the world is warming is not normally refuted. However, the extent to which it is warming, the causes and the seriousness of the consequences is a still a matter for debate. I see little in this data to warrant panic. Folks who live in areas that used to have August snow on the ground in the Rocky and Himalayan Mountains may notice that it has receded over the last 40 years and be sensitised to climate change in a way that southern California has become sensitised to drought.
It is worth noting that high latitude northern hemisphere will have low to zero greenhouse effect in winter for two reasons. First, the amount of insolation is low (zero for periods N of the Arctic Circle) and second snow cover gives rise to high albedo reflecting what insolation there is straight back to space. This might explain why winter maximum snow cover is unaffected by an enhanced greenhouse effect while the summer melt may be. But equally, there may be other factors that explain the more rapid snow melt from what has become a marginally higher base level. For example, the active Sun in the period 1950 to 2000, or continued natural warming post Little Ice Age.