Image: Temperature outlook for the United States issued on January 1, 2025 by the Climate Prediction Center.
The article
That Arctic blast can feel brutally cold, but how much colder than ‘normal’ is it really?, derived from this post, appeared in The Conversation on January 6, 2025.
The article
Don’t let these temperatures fool you: The extreme winters of the past are gone | Opinion, derived from this post, appeared in the Detroit Free Press on January 8, 2025.
What is Normal?
I write this on January 1. 2025 in advance of a cold snap about to occur in the eastern and southeastern parts of the United States (U.S).
There is high certainty that there will be a cold air outbreak in a few days.
There is also a forecast that the month of January will be colder than normal. This monthly forecast is less certain.
Such a cold air outbreak is not an unexpected or extraordinary event. This is true even in a warming climate.
I have been writing about cold air outbreaks in the context of climate change for years. Here is a topical collection entitled,
What I want to write about here is – what is normal?
I am motivated by two things.
First, in the coming days we will see figures in the news that compare the current temperatures to “normal.”
Second, normal is changing.
These two things combine in a way that it can easily be construed that present-day cold events are larger and colder than they are in an objective sense.
There are a couple of definitions of normal at work here.
- First, if we are going to present data as a difference from normal, then we need to define normal.
- Second, we often think of normal as what we expect; what we are used to.
Both of these are changing.
One of the most used standards for defining a science-based normal is a 30 year average of temperature and precipitation. There are protocols that at the end of every 10 years that these normals are updated. The two most recent thirty year spans are 1981 – 2010 and 1991 – 2020.
We know that Earth is getting warmer, so we expect the more recent 30-year span to be warmer than the earlier one. If we report a 2024 cold snap as differences from normal, then that cold snap will show larger deviations from normal for the more recent period. Even though it is the same temperature, it will appear colder.
Since 30 years is a significant part of a human life, for people less than 40 or so, this fits nicely into what they expect or are used to.
On the other hand, if you are interested in an absolute sense of how much Earth has warmed, the protocol of updating the averaging baseline every decade is not so good. It makes the warming appear smaller than it is; it complicates the communication of science-based knowledge.
To get a measure of how much it has really warmed, it is better to select a 30-year averaging span that occurred prior to the rapid temperature increase of recent decades.
The span I prefer is 1951 – 1980. Warming signals begin to emerge quite strongly after 1980. The 1951 – 1980 span is also the one used by NASA as a baseline.
Just to say, I have had my students tell me that the climate normal of my youth is no longer relevant. And, in many cases, I agree with that. It is essential that people “normalize” their experiences; it is part of their well being. On the other hand, it is quite important that we know how much it has warmed in a more absolute sense.
With that, I want to present a few plots which I made from the service provided by NASA’s Goddard Institute of Space Studies, more often known as GISS.
There are many interesting features in these plots, but we will focus on the eastern and southeastern parts of the United States. (North America, the U.S. and Canada, are located in the top, left quadrant of these figures.)
All of these plots are for January.
This first image is the difference between two 30-year averages. The 1951 – 1980 average is subtracted from the 1991 – 2020 average. Most of the world shows warming, and in the U.S. and Canada, North America, there is a warm spot of more than 4.0°C in north-central Canada. From this figure we can conclude two things. First, if we were to take any single day or month and compare it to a 1991 – 2020 average and then a 1951 – 1980 average, in the east half of the U.S., we would expect differences up to 2°C. Second there is a spatial pattern of larger differences, for example, there have been larger increases around the Great Lakes than Florida.
Now I am going to pick a few years just to show the visual effect of comparing to different averaging baselines. The first is 1977, when I was living in Tallahassee, Florida. It was a very cold winter. It even snowed and people raked the snow under trees to take pictures.
Here is January, 1977 compared to a 1991 – 2020 average. (Note that the color scale changes between these images.)
and here is January, 1977 compared to a 1951 – 1980 average. (Note that the color scale changes between these images.)
Focusing on the eastern U.S. and Canada: when January, 1977 is compared to the more recent 1991 – 2020 average it appears much colder than “normal” and the cold snap has a larger spatial extent. Again, note the color scales are different in the two plots by about 2°C.
Next I pick 1994, because it is a more recent year that is remembered as very cold. I was living in the Washington, D.C. region at that time.
First, January 1994 compared to 1991 – 2020: (Note that the color scale changes between these images.)
Next, January 1994 compared to 1951 – 1980: (Note that the color scale changes between these images.)
The message that follows from doing these comparisons is that because there is such a strong warming trend on the planet, and for our interests in the eastern U.S., if we were to compare a relatively cold, current-day temperature to a recent baseline average, it is easily reported as a cold event that is larger and colder than it is in an absolute sense.
As an example of this point, consider January of 2018, which had a cold air outbreak that was widely reported.
First, January 2018 compared to 1991 – 2020: (Note that the color scale changes between these images.)
Next, January 2018 compared to 1951 – 1980: (Note that the color scale changes between these images.)
For 2018, the comparison to the 1991 – 2020 average shows the entire eastern half of North America as below average, whereas compared to the 1951 – 1980 average, 2018 is above average north of the Great Lakes. We see here an example of the averaging period changing whether we are seeing warming or cooling.
The differences in western Canada are particularly stark as 2018 is, in fact, more than 4°C warmer when compared to the older base period.
What will occur in the East and Southeast in the next week is objectively and dangerously cold. It will feel very cold as we are normalized to warmer temperatures than we were 30 and 60 years ago. However, as an objective measure of warming, these temperatures are not lower than they were 30 and 60 years ago. We are not seeing a paradoxical winter is getting colder as the climate gets warmer.
From a physics-based perspective, the cold pool in the Arctic approaches the same low temperatures today that it approached 60 years ago. It takes longer after the onset of polar night for that to occur. The cold pool is smaller and of shorter duration. Because of changes to weather patterns, cold air might show up in peculiar places. When the Arctic cold pool is displaced to the eastern half of the U.S., it still gets cold. But, for not as long and not over as big an area as it used to. The extreme cold of the past is getting more rare and that will continue to be the trend in the coming decades.
Something Extraordinary about the Great Lakes
This is a forecast for early January 2024.
In pink are areas below 0°F.
What is extraordinary is the area above 0°F centered in Michigan, Ontario, and the Great Lakes.
This image is from the Washington Post (An Arctic Blast is Coming) and was made by Ben Noll.
This “warm spot” is indicative of the very high water temperatures of the Great Lakes, shown in this figure. In many places the lake water temperatures are still above 40°F. The lake ice cover is only 1.3%. This open warm water sets up the region with enormous potential for lake effect snow. This figure is from the Great Lakes Environmental Research Laboratory.