Tag Archives: climate statistics

Alaska Cooler Since 2000

Yesterday, the Anchorage Daily News ran a feature story in the
Alaska section of the paper titled Study shows Alaska got colder from 2000 to 2010. The article confirmed something I already knew….it’s been cold in the 21st century in Alaska.

I liked the article because it had lots of interesting information about my local region. Unlike so many articles in the local press, no effort was made to draw conclusions about global warming  from what clearly was  a regional study.

Each year I see many articles about the summer ice melt in the Arctic….and each article invariably draws conclusions about climate from what must be short term regional activity.  This article noted something that is easy to forget when thinking climate….regional variation is not unusual and not particularly helpful when evaluating long range worldwide climate trends.

Most of us draw too many conclusions about global climate from our local experience.  The contiguous 48 states of the USA was warm in 2012….regional data…and a short time period.  That experience really doesn’t mean much, when the global view is taken.   A warm USA, a cold Alaska….neither tell us much about global climate.

The article quoted a study that said it was 3.04 degrees F warmer in Barrow and 2.34 degrees  F cooler in for all of Alaska in the first 10 years of the 21st century.  The study used 20 first order meteorological stations.  20 Sites….all of Alaska…. I see a problem.   Alaska is just too big.  20 sites won’t do.   I suspect wild extrapolation of data between test sites (SWAG).

Last month (during a long stretch of cold clear weather) I drove by brother and mother from Providence Hospital to my house in South Anchorage.  The 9 mile trip took about 20 minutes.  My brother (who lives in Hawaii) was fascinated by the changing data displayed by the external thermometer in the car.   The temperature was -1 degree F as we left the Hospital, and quickly got down to -6 F and just as quickly got up to +6F before settling at -2 at my house.  Such swings are normal in Anchorage on clear cold winter nights.

Alaska is 663,267 square miles.   20 sites?  That’s one site for every 33,000 square miles.  South Carolina is a bit over 32,000 square miles.  On a cold winter day it is frequently 40 degrees F cooler in Fairbanks than in Anchorage. They and about 300 miles apart.  Southeast Alaska has weather similar to coastal British Columbia or Washington and the area along the Arctic north of the Brooks Range is similar to the Arctic parts of Siberia.

20 sites…2.34 degrees F…I don’t think so.   Sure it’s been cold…but not 2.34 degrees.  The data cannot possibly be that precise.   The .34 implies a precision that is not supportable by the data.   California is approximately one quarter the size of Alaska.  Imagine predicting temperature for the entire state of California by using data from just 5 sites.

Yep, it’s been cold in Alaska, but not 2.34 degrees F colder.   Alaska has every climate imaginable from a rain forest in Southeast to the frozen tundra of the Arctic.   Any study using only 20 sites must be doing a lot of guessing.   Yes it’s been cold and it’s probably somewhere near 2.4 degrees colder, plus or minus a half a degree or so.

Perhaps I’m overreacting a bit…after all the article did explain that the study used 20 sites, and guessing was clearly in order.   But that is the purpose of this blog….to point out the guessing that is here…there….everywhere in climate science.