Snow Cover changes on regional scale

[Tealight]

I finished calculating regional snow extent data and will post my analysis here. The main snow cover thread doesn't quite fit for this detailed long term analysis. At the moment all data is still in one long list, but after formatting we can graph things like snow extent for region x in month y. I attached a map showing all regions and an example for Greenlands snow extent.

Eventually regional graphs should also get daily updates on my main snow cover webpage
https://sites.google.com/site/cryospherecomputing/snow-cover

Data Download (csv & formatted ExcelSheet)
https://github.com/NicoSun/CryosphereComputing/tree/master/ScienceData/Snow_Cover_Regional

[Tor Bejnar]

I note the two lowest August Greenland snow extent years matche the two lowest August Arctic sea ice extent years.

[Tealight]

The data is now available as a formatted Excel sheet through the Link in the first post. I try to post 2-3 regions every day until we covered all of them.

Today we continue the northern regions with northern Canada and Alaska. They feature a pattern I've seen in several northern regions. Early summer snow cover was lowest in the 2010-2012 region and increases moderately for earlier years and slightly later years.

[Tealight]

Today we continue with Canada. This post is about the Canadian Rockies and Central Canada. The next one is exclusive to eastern Canada. Check the region map in the first post for the exact boundaries.

For the Rockies May, June and October seem to have a consistent downward tend in snow cover over the last 20 years. The other months are very close to either the maximum extent or to the minimum extent. Only November has a very slightly positive trend in snow cover.

In Spring/Autum snow cover in Central Canada is highly variable from year to year. However October and November show a very clear upwards tend. So the data is confirming the theory that the higher atmospheric moisture from an increasingly ice free Arctic causes more snow cover in the autumn.

[Tealight]

Eastern Canada doesn't have a downwards trend during Spring. Instead it's positive like in the autumn. The theory here is that eastern Canada receives so much more snowfall that it offsets any increased snow melt during spring. See the Northern Hemisphere snow cover thread for discussion, maps & graphs (around post 200).

snow cover trend per month
April: slight increase
May: clear increase
June: neutral
July: slight increase
Aug: None
Sep: clear increase (although at very low level)
Oct: clear increase
Nov: clear increase

[Tealight]

Today we continue with the Arctic region with Siberia. This post is about the Eastern and Central Siberia. The next one is exclusive to western Siberia. Check the region map in the first post for the exact boundaries.

Like the Arctic in Northern America early summer snow cover was lowest in the 2010-2012 region and increases for earlier years and later years. For eastern Siberia this is most pronounced in June and for Central Siberia in May.

Autumn Snow Cover shows no significant trend for eastern Siberia, possibly due to almost complete coverage by the end of October. Central Siberia experiences a slight increase in October and November snow cover.

[Tealight]

West Siberia varies more noticably from year, but has again an early summer low point around 2010-2012. October shows the typical long-term autumn increase in snow cover which we have observed in other regions. November only has a very small increase due to the large year to year variability.

[Tealight]

Today we complete the Arctic region with Eastern Europe and Scandinavia.

Both regions are heavily influenced by the North Atlantic and snow extent varies considerably from year to year. Neither follows the typical spring decrease and autumn increase trend of other Arctic regions. Overall Eastern Europe experiences a modest decline in snow cover over most month, but as 2017 shows new record high snow cover (for the last 20 year period) is also possible.

The only months with a trend in Scandinavia are February and October showing a slight increase and June with a slight decrease.

[Tealight]

We finish the Arctic regions with the mean extent over the whole year to see when each region starts to melt and at what rate. For example Alaska melt starts early at the end of April, but takes until the end of June. In contrast Eastern Siberia doesn't start until mid May, but also finishes at the end of June.

The 2000s and 2010s mean snow extents of Alaska and Eastern Canada show the magnitude of low spring snow cover or high autumn snow cover compared to the whole year.

[Archimid]

This is so great. I'm downloading some of this data to play with it. Thank you Tealight. /bow

[Shared Humanity]

I agree. This is great stuff Tealight.

[Tealight]

This is so great. I'm downloading some of this data to play with it. Thank you Tealight. /bow

I agree. This is great stuff Tealight.

Thanks, I have been quite busy lately but will continue with the mid-latitudes. Feel free to post your own findings.

The US Pacific varies considerably from year to year like Europe, but also shows a similar trend to the northern regions just a few months offset. The long-term spring snow cover decrease is in April instead of May/June and autumn snow cover increase is delayed into January.

[Tealight]

The US rockies show significantly less variability and clearer trends. It's a very clear long-term decrease in spring and a modest increase during autumn and winter.

[Tealight]

Surprisingly (at least for me) the US Mid-West shows the same a long-term increase in spring snow cover as the US North-East and Eastern Canada. Maybe it's a feature of the increasingly ice free Great lakes to release more moisture at the end of winter. In autumn the regions show the typical long-term increase in snow cover as well.

[Tealight]

Now finishing North America with the US South East and US South West. In both regions snow is pretty much unknown except for a few days in winter, but these are becoming more common or effect a wider area. January and February both show a long term increase in snow cover.

[Tealight]

When looking at the regional USA mean snow extent over the whole year the Pacific Coast and southern regions are pretty much irrelevant and Alaska has a far shorter snow free period than the continental states. Even though the US Mid-West and US Rockies are at the same latitude the Mid-West melts faster than the snow on the mountains. No surprises here due the high altitude regions being colder. In autumn however both regions gain snow cover at the same rate indicating that temperature isn't a factor in autumn. (at least for these regions)

[Tealight]

Western and Southern Europe behave similar to southern USA. Snow only falls in mid to late winter and shows a long-term increase. However the year to year variation is still the dominating factor.

[Tealight]

When averaging through the huge year to year variations, central Europe and to some extent southern Europe as well go against the typical long-term increase in autumn snow cover. Only late winter shows an increase.

Central Europe months with trends:
Nov: decrease
Dec: decrease
Jan increase

Southern Europe months with trends:
Nov: decrease
Dec: decrease
Jan increase
Feb increase
Mar increase
Apr: decrease

[Tealight]

Overall Western Europe and Southern Europe have no meaningful snow cover. On the Central European graph we can see a typical weather phenomen in late winter bringing cold and snowy weather to the continent and peak snow cover for the region.

[bbr2314]

Question for you, Tealight. Would you be able to reproduce the US graphs with two curves for each 10-year period? (i.e., 1999-2008, and 2009-2018?). Or 11-years, whatever the split is. Curious to see how the curves shift when we compare longer averages vs. individual years and I would think that would be most informative re: ongoing trends.

[Tealight]

Question for you, Tealight. Would you be able to reproduce the US graphs with two curves for each 10-year period? (i.e., 1999-2008, and 2009-2018?). Or 11-years, whatever the split is. Curious to see how the curves shift when we compare longer averages vs. individual years and I would think that would be most informative re: ongoing trends.

Sure I can show 10 year averages, but you have to be careful in not selcting a very specific range which happended to be very high or low. We only have 21 years of data so I chose the middle period (2004-2013) as a control period.

2004-2013 compared to 1999-2008 and 2009-2018
Rockies: pretty consistently between the other two averages
Mid-West: mostly between other two except early December significantly higher
North-east: same as the other two averages
South: similar noise to other two averages

[bbr2314]

Question for you, Tealight. Would you be able to reproduce the US graphs with two curves for each 10-year period? (i.e., 1999-2008, and 2009-2018?). Or 11-years, whatever the split is. Curious to see how the curves shift when we compare longer averages vs. individual years and I would think that would be most informative re: ongoing trends.

Sure I can show 10 year averages, but you have to be careful in not selcting a very specific range which happended to be very high or low. We only have 21 years of data so I chose the middle period (2004-2013) as a control period.

2004-2013 compared to 1999-2008 and 2009-2018
Rockies: pretty consistently between the other two averages
Mid-West: mostly between other two except early December significantly higher
North-east: same as the other two averages
South: similar noise to other two averages

Thanks for that, very helpful! Can you do Eastern Canada as well?

It is interesting that it looks like the Northeast follows the Midwest's pattern, but less exaggerated.

The Rockies also provide a very notable trend of decreasing spring, summer, and autumnal snow  -- perhaps the most strikingly consistent of any region.

[Tealight]

Thanks for that, very helpful! Can you do Eastern Canada as well?

It is interesting that it looks like the Northeast follows the Midwest's pattern, but less exaggerated.

The Rockies also provide a very notable trend of decreasing spring, summer, and autumnal snow  -- perhaps the most strikingly consistent of any region.

I already posted 10 year Eastern Canada averages, just not the same periods as the USA. Anyway here are all four Canadian regions with 1999-2008, 2004-2013 and 2009-2018 periods.

[Shared Humanity]

Noticeable positive, early season anomalies in Northern, Central and Eastern Canada.

[bbr2314]

Noticeable positive, early season anomalies in Northern, Central and Eastern Canada.

The Rockies decrease / shift early season is probably the next most noticeable trend. It almost seems like cryospheric balance is gradually shifting closer to Greenland where it can be more easily maintained thanks to proximity to 21,000,000 KM^2 of ice.

[Tealight]

It's time to finish the project with the last four mid-latitude regions in Asia. This post is for the low elevation Central Asia and East Asia.

Central Asia months with trends:
Oct: increase
Nov: increase
Dec: increase
Jan decrease
Feb: increase
Apr: decrease

East Asia months with trends:
Oct-Mar: increase

Both regions show the typical autumn-winter increase in snow extent. In spring central asia shows decrease and east asia shows no trends.

[Tealight]

The final regions are Tibet and the mountain ranges north of it. (I named the region Central Asian Mountains)

Tibet months with trends:
March to May: increase
June to February: decrease

Central Asia Mountain months with trends:
October - February: increase
May: decrease

Tibet goes completly against the general trend with increases in spring snow cover and decreases for most of the year including autumn and winter. A more detailed look follows tomorrow with daily values of 10 year averages. The mountains north to it follow the general trends with increased autumn-winter snow cover and decreased spring snow cover.

[interstitial]

Tealight did you compute R² for those trends? Some of them don't look like they have very strong correlation.

[Tealight]

Tealight did you compute R² for those trends? Some of them don't look like they have very strong correlation.

I just added a linear trendline in excel to get any quick impression from the noisy data. Some months have R² almost 0 and some as high as 0.6.  We are analysing a highly weather dependent variable so I don't expect high R² values to begin with. That's why climate scientists use 10 to 30 year averages.

[Tealight]

As promised here is the detailed 10 year mean snow extent, though for Tibet it's still quite noisy. The only smooth period is mid to late summer showing a decline in snow cover. The Tibet region includes the glaciated Himalayas, so snow extent never reaches zero.