Cold Air Outbreaks

Erik T. Smith, Ph.D. Candidate Kent State University Department of Geography

Current Research on Cold Air Outbreaks

Overview

Cold Air Outbreaks (CAOs) are extreme events with potentially large impacts on human health, animals, agriculture, energy industry, etc. With most of the research on CAOs having a regional focus, it's difficult to determine how CAOs have changed through time and how these changes may be related to changes in atmospheric circulation patterns. This study creates a global climatology of CAOs to address the following research questions: 

Research Questions

  1. How have CAOs changed in duration, magnitude, spatial extent, and occurrence in different regions across the globe since 1979?
    1. How well does reanalysis data capture CAOs and how does this compare with CAOs derived from historical surface station data?
    2. How do the two reanalysis datasets (ERA5 and NNR) compare?
  2. Are there seasonal or regional differences in the atmospheric circulation patterns that precede CAOs?
    1. How do circulation patterns influence the magnitude and duration of CAOs?
    2. How do atmospheric teleconnections influence CAOs in different regions?
  3. How well do forecast models predict the location, intensity, and duration of CAOs?

 

Figures

Click on each image to enlarge

NNR

ERA5

Other

Extreme cold event (ECE) Days per Winter Season

ECE Days NNR

Extreme cold event (ECE) Days per Winter Season

ECE ERA

 Regions ERA5

regions era

CAO Days per Winter Season

CAO Days NNR

CAO Days per Winter Season

CAO ERA

Regions NNR

regions nnr

Trend in ECE Days per Winter Season

ece trend nnr

Trend in ECE Days per Winter Season

ece era5 change

Skewness ERA5

skew

Trend in CAO Days per Winter Season

change cao nnr

Trend in CAO Days per Winter Season

trend cao era

Temperature Mid-Winter (oC)

temp

Percent Change in CAO Days per Winter Season

percent NNR

Percent Change in CAO Days per Winter Season

era5 percent

Change in Total CAO Spatial Extent per Hemisphere

spatial extent

Change in Mean CAO Duration 

mean dur nnr

Change in Mean CAO Duration 

mean dur
 

Change in Maximum CAO Duration

max dur nnr

Change in Maximum CAO Duration

max dur
 

Change in Minimum Magnitude per CAO

z nnr

Change in Minimum Magnitude per CAO

min z
 

 

Regional Statistics

Region

Location

1979

CAO Days

2018

CAO

Days

Δ

CAO

Days

Max. Duration

Mean Duration

Min. σ

Begin Date

End date

 

Season Length

1

Eastern U.S.

10

6

-4

-5

-0.4

0.2

-3

-5

-2

2

Canada

14

7

-7

-3

-0.5

0.8

5

0

-6

3

Labrador Sea

15

5

-10

-8

-1.5

0.8

12

20

7

4

Alaska

10

5

-5

-7

-0.6

0.4

40

-42

-82

5

Southeast Asia

7

4

-3

-4

-0.5

-0.1

4

-27

-31

6

Central Russia

8

9

1

2

0.7

-0.1

12

-18

-30

7

Barents-Kara Seas

15

5

-10

-8

-1.1

0.5

16

-11

-27

8

Europe

9

8

-1

-3

0.1

0.0

8

-21

-29

9

N. Africa

9

6

-3

-3

-0.2

0.1

-7

-6

1

10

South Africa

6

2

-4

-6

-0.7

0.1

-3

-10

-7

11

N. Australia

5

4

-1

-2

1.9

-0.7

10

-14

-23

12

South America

9

4

-5

-1

-0.1

0.2

-14

12

26

13

South Pacific

13

18

5

2

0.1

-0.5

4

-18

-22

 

Data

 

NCEP/NCAR (NNR)

ECMWF ERA5

Length of Data

1979 - 2018

1979 - 2018

Temporal Resolution

Mean daily 2-meter temperature

Hourly 2-meter temperature: converted to daily mean temperature

Spatial Resolution

T62 Gaussian grid (94 lat. by 192 long.)

0.25o lat. by 0.25o long. Upscaled to 1o latitude by 1o longitude (181 lat. by 360 long.)

Methods

Cold Air Outbreak Criteria

Magnitude

  •  σ <= -1.96 (2.5th percentile)
  •  Daily mean < 20o C
  • Daily departure from mean > 2o C

Duration

  • Spatial extent must be met for at least 5 consecutive days

Spatial Extent

  • 1,000,000 km2 (1/8 contiguous U.S.)

**An extreme cold event (ECE) is a cold air outbreak (CAO) without the spatial extent criterion.

 

Regionalization

Regions are created by correlating the CAO duration of a central grid point with surrounding grid points where only contiguous grid points with correlations equal to or greater than 0.75 are included in the region. The central grid point is determined by first selecting relative regions, such as the eastern U.S. or Europe, then calculating a local maximum in the total number of CAO days from 1979 – 2018 in each relative region. Relative regions were chosen according to similarity in climatological characteristics, CAO characteristics, and CAO trends during the period of study. This combination of relative and statistically defined regions merges the user’s climatological expertise with data derived inputs to produce a more objectively defined region.

Trends and Significance testing

Trends for the Southern Hemisphere (SH) were calculated over 40 winter seasons (January 1 – December 31) and trends in the Northern Hemisphere (NH) were calculated over 48 winter seasons (July 1 – June 30). Theil-Sen slope estimation was used to estimate the change in the number of CAOs and CAO days, as well as other CAO characteristics, by region and by individual grid point. Because of the limited sample size (39 in NH and 40 in SH), the Theil-Sen slope estimation was calculated from 1000 bootstrapped samples and statistical significance determined from the confidence intervals produced from the bootstrapped samples. A false discovery rate was used to account for the spatiotemporal relationships and better determine the field significance of the data.

 

Poster

 

 

 

 

February 2015 Cold Air Outbreak in Eastern U.S. - Surface Temperatures

 

Research Articles

Smith, E. T., & Sheridan, S. C. (2018). The characteristics of extreme cold events and cold air outbreaks in the eastern United States. International Journal of Climatology38, e807-e820.

Smith, E. T., & Sheridan, S. C. (2019). The influence of atmospheric circulation patterns on cold air outbreaks in the eastern United States. International Journal of Climatology39(4), 2080-2095.

Smith, E. T., & Sheridan, S. C. (2019). The influence of extreme cold events on mortality in the United States. Science of the Total Environment647, 342-351.

 

Previous Research on Cold Air Outbreaks

 

SLP SOM
Figure: Sea-level pressure self-organized map (SOM) from 45oN to 90oN. Graphs represent Oct. - Mar. climatology of SOM pattern to the right.

Abstract

The Characteristics of Cold Air Outbreaks in the Eastern United States and the Influence of Atmospheric Circulation Patterns

Periods of extreme cold impact the mid-latitudes every winter. Depending on the magnitude and duration of the occurrence, extremely cold periods may be deemed cold air outbreaks (CAOs). Atmospheric teleconnections impact the displacement of polar air, but the relationship between the primary teleconnections and the manifestation of CAOs is not fully understood. A systematic CAO index was developed from 20 surface weather stations based on a set of criteria concerning magnitude, duration, and spatial extent. Statistical analyses of the data were used to determine the overall trends in CAOs. Clusters of sea level pressure (SLP), 100mb, and 10mb geopotential height anomalies were mapped utilizing self-organizing maps (SOMs) to understand the surface, tropospheric Polar Vortex (PV), and stratospheric PV patterns preceding CAOs. The Arctic Oscillation (AO), North Atlantic Oscillation (NAO), and Pacific-North American (PNA) teleconnections were used as variables to explain the magnitude and location of mid-latitude Arctic air displacement. Persistently negative SLP anomalies across the Arctic and North Atlantic were evident 1 – 2 weeks prior to the CAOs throughout the winter. The tropospheric and stratospheric PV were found to be persistently weak/weakening prior to mid-winter CAOs and predominantly strong and off-centered prior to early and late season CAOs. Negative phases of the AO and NAO were favored prior to CAOs, while the PNA was found to be less applicable. This method of CAO and synoptic pattern characterization benefits from a continuous pattern representation and provides insight as to how specific teleconnections impact the atmospheric flow in a way that leads to CAOs in the eastern U.S.