1

Are Exceptionally Cold Vermont Winters Returning?

Dr. Jay Shafer

July 1, 2015

Lyndon State College

Jason.Shafer@lyndonstate.edu

3

Outline

• What the cold forecast well?• How cold was it?• Arctic air characteristics• Regional climate controls• Climate trends• Activity – statistical method to predict seasonal weather

4

Dynamical Model Forecasts and Verification

Verification: 2013-14Forecast

http://iri.columbia.edu/our-expertise/climate/forecasts/seasonal-climate-forecasts/

http://www.esrl.noaa.gov/psd/data/usclimdivs/

5

Dynamical Model Forecasts and Verification

http://iri.columbia.edu/our-expertise/climate/forecasts/seasonal-climate-forecasts/

http://www.esrl.noaa.gov/psd/data/usclimdivs/

ForecastVerification: 2014-15

6

Seasonal Snowfall 2014-15

Above average snowfallin fact, winters are getting wetter, and the climate is cold enoughto sustain snow, so wintershave gotten snowier

7

Heating Degree Days

• HDD = 65 deg F - (daily avg temperature)• For example, high = 40, low = 20, daily avg temperature = 30• 65-30 = 35 HDDs • HDDs correlate well with energy use

8

Energy Use vs HDDs

150000 160000 170000 180000 190000 200000 210000 220000 230000 240000 2500006000

6250

6500

6750

7000

7250

7500

7750

8000

8250

8500

7585

6668

7205

8123

Lyndon State College Fuel Oil Consumption (Gallons)

Cum

ulati

ve S

t. Ja

y H

DD

s (O

ct 1

- M

ay 3

1)

2013-14

9

Heating Degree Day Trends

1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 20106000

6500

7000

7500

8000

8500

9000

f(x) = − 5.47765376823307 x + 18245.077668221

St. Johnsbury, VT: Cumulative Heating Degree Days Time Series

Oct

1 -

May

31

Cum

ulati

ve H

DD

s

Averaging 5 fewer HDDS a season, so about 100 fewer HDDSnext 20 years, a few percent less energy usage, on average.

Over the last century, heating demand has declined5-10 % due to winter warming.

10

Why were these last two winters so cold?

11

North American Circulation Pattern

12

Arctic Air Mass Example – January 23, 2104

Arctic High Pressure

13

Where did the arctic air originate?Arctic air masses typically have a long residencetime over high latitude continental regions withsnow cover associated with high pressure systems

14

Sampling arctic air at Lyndon State

15

16

Arctic Air Mass Characteristics• Dry (typically sunny)• Very cold (less than -30 deg C)• Stable (difficult to get

precipitation/clouds to form)• Isothermal low-level air mass

(coldest air mass at the surface)

17

Very cold, stable air

18

19

Seasonal Controls of Winter Weather

Variability

20

El Niño/La Niña (Nino3.4) vs. Burlington Winter Temperatures

-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

El NiñoWarmN=6

El NiñoColdN=4

La NiñaWarmN=7

La NiñaColdN=3

21

El Niño/La Niña (ENSO) Conclusions

• ENSO does not explain the variability • Other factors are at play, complex interactions of tropics and high

latitudes • Other areas of the US have significant winter ENSO relationships, but

not the Northeast US• ENSO has little to no effect on winter conditions in the Northeast US

22

Snow Cover – Land Surface Feedbacks

23

http://climate.rutgers.edu/snowcover/

Snowy Octobers last two years

Theory: snowier fall conditions produce an increased risk for cold air mass developmentand eventual movement into middle latitudes –

In other words, if there’s more snow in the fall,then somewhere in the Northern Hemisphere,there is bound to be an enhanced risk of arctic airmasses moving southward away from the arctic.

This is an example of a positive feedback loop, annegative correlation.

24

Winters following high October Eurasian Snow Cover

Arctic air pathway is open more often

25

Winters following low October Eurasian Snow Cover

Cold air is shy and remains furthernorth – Alaska and northern Canada

26

Fall Sea-Surface Temperature Anomalies Preceding Cold Winters

North and Central Atlantic Oceanis average to cold.

Oceans play a significant rolein forcing the atmosphereover longer time periods.

27

Fall Sea-Surface Temperature Anomalies Preceding Cold Winters

North and Central Atlantic Oceanis warm to average.

Strong dipole of SSTs in north and central Pacific

28

Arctic Oscillation

29

30

Vermont Climatic Changes

31

Heating Degree Day Trend – Temperature Trend

1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 20106000

6500

7000

7500

8000

8500

9000

f(x) = − 5.47765376823307 x + 18245.077668221

St. Johnsbury, VT: Cumulative Heating Degree Days Time Series

Oct

1 -

May

31

Cum

ulati

ve H

DD

s

Averaging 5 fewer HDDS a season, so about 100 fewer HDDSnext 20 years, a few percent less energy usage, on average.

Over the last century, heating demand has declined5-10 % due to winter warming.

32http://journals.ametsoc.org/doi/full/10.1175/2011WCAS1096.1?prevSearch=[Contrib%3A+alan+betts]&searchHistoryKey=

Betts 2011

Winters are getting shorter and the growing season is increasing.

33

Betts 2011Winters are getting shorter through other proxies.

34

Seasonal Forecasting

35

Statistical Prediction Methods

• Relate numerical index values of oceanic temperature patterns (or whatever variable you choose) to the following winter’s temperature and precipitation• For example, you could look at the ENSO state, which represents El

Niño/La Niña, the largest oceanic oscillation on monthly to yearly time scales• We will relate the Arctic Oscillation to show how this is related to

winter temperatures

36

Statistical Methods: Fall Snow Cover

• Siberian snowfall during October has an effect on winter temperature patterns over the Northern Hemisphere.• There is well documented literature on the topic physically connecting

the two – as mentioned earlier• Enhanced fall snow cover enhances the Hemispheric cold air reservoir and

creates a greater potential for winter cold in the mid and high latitudes• Complex interaction involving stratosphere and troposphere, but it has been

physically described – still need “weather” events to move cold air south

37

Climate Prediction Center Forecasts

July, August, September Outlook

Precipitation forecasting is much more difficultthan temperature forecasting.

38http://iri.columbia.edu/our-expertise/climate/forecasts/seasonal-climate-forecasts/

http://www.cpc.ncep.noaa.gov/products/predictions/long_range/seasonal.php?lead=3

Fall (Sept – Nov) Outlook

39

Winter (Dec– Feb) Outlook

40

Conclusions

• Are cold winters returning? • No, not for the long haul• However, natural variability will continue to produce cold spells and

occasional prolonged cold weather like the last two winters

• Climate models struggle with seasonal forecasting and processes as snow cover-land surface feedbacks• Vermont winters are getting shorter (especially with their late arrival),

but they can have intense stretches as they have in the past• The next ten winters will probably be like the last ten winters

41

Activity

• Correlate Arctic Oscillation state with season cumulative HDDs• Hypothesis: Arctic Oscillation phase during winter has an effect on

Vermont seasonal temperatures• Excel sheet is available at: https://

drive.google.com/file/d/0B3NtxLJnOImFRUxHQU9qMnNJazg/view?usp=sharing

42