Freeze/Thaw Cycles: Part 1

Today we are going to take a look at the history of Edmonton's freeze/thaw cycles.

This post was written way back in mid-January under the assumption that we would have a typically melty February. As we will see today, during February we on-average get about 10 days above freezing.  Instead though, February 2019 was meltless and was the coldest February in generations. But now it seems that our long, long coldsnap is finally over, so let's talk about melting.

The question of how exactly to count freeze/thaw cycles is actually a little bit complicated. If you are interested in the specifics and assumptions behind the numbers today those are outlined in a separate post Freeze/Thaw Cycles: Appendix. The short version is:

• we will be using Daily High & Low temperatures to count freeze/thaws.
• the numbers today won't necessarily be 100% accurate, but for year-to-year comparisons they should at least be consistent.

With that house-keeping out of the way...

Freeze/Thaw Cycles Each Year

Here we have the number of freeze/thaw cycles each year going back to 1880. The 5-Year average appears as a dotted red line to show how things have changed over time.

Right now we average about 91 freeze/thaw cycles per year. In recent years that has ranged from around 120 for a year like 2004 down to around 70 for 20014, but most years are in the range of 80-100.

That average of about 90 freeze/thaw cycles per year has been relatively consistent for the last 90 years, going back to 1930. And before that the average was quite a bit higher, with around 110~115 freeze/thaw cycles per year.

So for the last 90 years the number of freeze/thaw cycles has not changed much, and before that we used to get more? That can't be right, can it? These numbers might seem counter-intuitive, but we will want to keep this in mind today: a freeze/thaw cycle needs a thaw and a freeze.

Hopefully by the end of today we will have a better idea of what is going on here.


Freeze/Thaw Cycles Each Month

Here the freeze/thaw cycles have been broken out by month, and the chart cycles through the average for each decade going back to the 1880s. The faint grey lines in the background show the ranges of all of the decades, and the 2010s are dotted-in in red for an easy comparison. Both the 1880s and 2010s only include 9 years (1881-1889 and 2010-2018), so they aren't quite full decades.

The months which have the most freeze/thaw cycles are March, April, October & November. That is probably not surprising, because those are the main spring/autumn months. And over the decades those months have all averaged from 12-18 cycles each month.

What might be a bit surprising is that we also have quite a few freezes/thaws in the middle of winter, with December, January & February all averaging 6-11 cycles each month. That is about half the number of cycles that the big months get, and it is not a new thing. In the 1880s January averaged 6 freeze/thaws, but in the 1890s it was 9. In more recent history the 1980s were at 10, the 1990s were down at 7, and the 2000s and 2010s were both up at 11. Right now we have been up at the top of the historic range for January, but that still only works out to a few extra freeze/thaw cycles per month.

Focusing on the red line for the 2010s:

• January & February are near the top of the historic range.
• March & December are in the middle of the range.
• April, May, September, October & November are all at or near the bottom of the range.

In the first chart today we saw that our freeze/thaw cycles have actually decreased over time. Here we can see that in January & February the numbers have gone up a bit, but for the rest of the months they have dropped.


Freeze/Thaw Cycles Each Week

In this chart we have broken things down even further to the number of freeze/thaw cycles for each week of the year. (the numbers for week 4 of each month have been been adjusted so that they are comparable to all of the other 7-day weeks)

Looking at the historic range (the grey cloud in the background) the number of freeze/thaw cycles trends upwards each week from January through the end of March, and then with the beginning of April it starts to drop off rapidly to the end of May. In the autumn things are reversed, with a fairly quick increase starting in September and peaking at Halloween, and then a slightly slower dropoff towards New Year's.

When we look at the line for the 2010s we again see:

• January & February weeks are near the top of the range, which means maybe 1 extra freeze/thaw cycle per week on average. 
• Early-spring (March & early-April) are middle of the pack.
• Late-spring (Late-April & May) are near the very bottom.
• Autumn & early-winter (September, October & November) are near the very bottom.
• December is middle of the pack.

So why are our recent numbers so low in late-spring, autumn, and early-winter? Earlier we had said that a freeze/thaw cycle needs a thaw and a freeze, and now we are going to look at why that is important.




Here we have the same data which we were just looking at - the average number of freeze/thaw cycles per decade - but we have also added the number of Highs Above Freezing in red, and the Lows Below Freezing in blue.

As this chart cycles through the decades we see that:

• During the winter from November-March the number of freeze/thaw cycles follows the red line for Highs Above Freezing very closely. In the winter most nights are going to be cold, and so the only way to have a freeze/thaw cycle is to have a warm day.
• For April-October the freeze/thaw cycles follow the blue line for Lows Below Freezing. In the spring/summer/autumn most of the days are warm, and so cold nights are the trigger for a freeze/thaw.

Depending where we are during the year the freeze/thaw cycles have different drivers. In the winter they are driven by warm days, and in the spring & autumn the cold nights are important.

Here we are just focusing on the 2010s:

• We have gained freeze/thaw cycles in January-March because of more warm days.
• We have lost freeze/thaw cycles in April-May and September-November because of fewer cold nights.
• We have also lost freeze/thaw cycles in late-November and December because of fewer warm days.

Why have January & February warmed up while November & December haven't? Who knows. But it was something which we also saw in more detail in The Months through the Years.


Freeze/Thaw Cycles Each Year, Again

And so here we are back where we started, with the number of freeze/thaw cycles each year. But this time we have also added the number of Highs above freezing in red, and the Lows below freezing in blue.

• The number of Highs above freezing has stayed very consistent, with an average of about 281 per year since the 1890s.
• The number of Lows below freezing has dropped a bit, from 192 a century ago to 174 today.
• We saw earlier that for the past 90 years our average number of freeze/thaw cycles each year has stayed fairly constant at around 90. Before that it was about 20 higher at 112.

When we think of freeze/thaw cycles we probably focus on the thawing part of it, but that has stayed about the same over the years. What has changed is that we have lost some freezing nights, and with that we have also lost some freeze/thaw cycles.

It might seem strange that our number of Highs above Freezing has not changed much over the last century, but it is something that we have seen before, most recently in The Months Through the Years:

The background of this chart is the distribution of our High temperatures since 1996. 

The 1996-2018 average High for each day is the white line which runs through the middle of the chart. The red line which appears shows the average Highs from 1880-1889. Our modern Highs are warmer in January, February, and late November, but for most of the year things have not changed very much. 

In comparison here are the 1880-1889 Lows:

For the Lows temperatures our modern averages are shifted up 3-5°C compared to where they were a century ago.

With that temperature shift our "season" of Lows-below-freezing has gotten shorter by about 2 weeks in April and October, which is why we have lost on-average about 20 Lows below freezing. But for the Highs-above-freezing the season has stayed consistent as mid-March to early-November.

1920s & 1930s

One final question from our yearly chart is why the number of freeze/thaw cycles took a sharp step down around 1930. This chart shows the 1910s through 1940s so that we can take a closer look at what was going on there.

There is a lot going on in this chart, but here are some things to watch for:

• The 1910s & 1920s both had a lot of Lows Below Freezing (blue line) in April and May, and a lot of Highs Above Freezing (red line) in October. So cool spring nights and warm autumn days meant a lot of freeze/thaw cycles.
• In the 1930s the number of cool spring nights and warm autumn days both fell. Those might seem like they are opposites, but they both cause the number of freeze/thaw cycles to drop.
• Then in the 1940s the number of cool spring nights stayed fairly low. It did have more warm autumn days than the 1930s, but also less warm days in November & December.

We don't know why these changes happened from decade-to-decade, but this explains why the chart shows such a noticeable change.


The 2010's

Today we have mostly focused on how things have changed over the different decades, but in this final chart we are comparing the individual years from 2010-2018 (and Jan-Feb 2019).

For 2018 the notable months were May which had very few freeze/thaw cycles, and then September, October & November had lots. And for 2019 so far January was average, while February was one of only 15 months since 1880 with no days above freezing.

And here are some of the extremes:

• January 2010 had 7 freeze/thaw cycles, while in 2012 there were 17.
• February 2019 just had 0, while 2014 had 3, and 2016 had 19.
• December 2010 had 0, while 2011 had 19.

Over the last century the number of freeze/thaw cycles has been very consistent, on average. But on a year-to-year and month-to-month basis there is a lot of variation.

Summary

So do we get more freeze/thaw cycles than we used to?

• For January & February we might get about 1 more per week than the historical average, although things have not changed very much since the 2000s.


• For the rest of the year we have lost freeze/thaw cycles in the late-spring and early-fall, and also in December for some reason.

January & February are probably the months when we are most likely to notice freeze/thaw cycles, because warm mid-winter days are a bit unexpected, and also because there is a lot of snow to melt. So what probably happens is that certain weeks or months stick in our minds - it is easy to remember a particularly melty January or February - and then we forget all about the frozen December which came before it.