If you don't like the weather, wait 5 minutes

Today we're going to look at Edmonton's seemingly-erratic weather.

To get a sense of what I'm talking about, here are our High temperatures for the last few months:

November started out really cold, and it ended really warm. December started really warm and ended really cold. January went cold-warm-cold-warm-cold. And February has mostly been pretty cold. Is this sort of variation normal? Can't we ever have nice, "average" temperatures?

This is a theme we've tackled a few times before in:
  • How Normal is Average - during the summer about 80% of the days are within ±5°C of the average temperature, but in the winter that range doubles to ±10°C.
  • Does Edmonton Get Chinooks - we don't necessarily get chinooks like Calgary does, but we do usually see an an echo of whatever they get.
  • 'Tis the season of -20°C - even in the middle of winter we get a lot of warm days to balance off the cold ones.
What we're going to focus on today is how often we swing from warm days to cold days, and then back again.


This is the chart we'll be working with, but it needs a bit of explanation.

This is a chart that we use all the time around here, and it shows daily High temperatures throughout the year. We're going to stick with High temperatures today because they are more variable than the Low temperatures. And so if it feels like our temperatures are behaving erratically then we'll see that more easily with the Highs.

The red and dark blue areas in the chart are the Highs for 2017, and they show whether each day was above or below the average for that time of year. The orange and lighter blue areas in the background show the temperature distribution that Edmonton has seen since 1996, with the warmest, second warmest, coldest and second coldest. The 25th and 75th percentiles are shown as a grey band.

Those 25th and 75th percentiles will be what we're focusing on today. I didn't want to look at just any ol' temperature swing, but wanted to specifically look at when we swing from being warmer than normal to colder than normal, or visa versa. The 25th and 75th percentiles work as handy cutoffs for that.

So with all of that said then: the yellow lines in this chart show every time that the temperature swung from cold (below the 25th percentile) to warm (above the 75th percentile) or the opposite. In 2017 that happened 38 times.

There's a formula controlling all this, and the lines might not exactly match what someone would eye-ball off of the chart, although they should be close. When the temperature spends lots of time hanging around the average that doesn't count as a swing, and if the temperature stays warm or cold for a long time then this chart tries to pick a consistent start or end point. Normally I don't like to invent a bunch of rules for these charts, but we're going to be applying this to 136 years of data, and so it's important to have a standard methodology rather than doing it all by hand.

2017 - Take 2

In 2017 the High temperatures swung between the 25th and 75th percentiles a total of 38 times. But a lot of those happened during the summer when the actual change in temperature was relatively small.

This version of the chart takes all of those swings, and it filters out any where the change in temperature was less than 15°C. That is just an arbitrary cutoff, but a 15°C swing in High temperatures is going to feel like a substantial change no matter where we are during the year - whether it's going from 20°C down to 5°C in June, or from -10°C up to 5°C in February.

Applying the 15°C cutoff filters out the small swings during the summer months, and we are left with a total of 18 swings of 15°C or more for 2017. That includes some very large swings, like in February when things warmed up by 36°C over the course of a week, and in December when temperatures dropped by 31°C over about 2 weeks.

It's important to keep those timeframes in mind, because here we're not necessarily talking about the rapid temperature changes of chinooks (some of Edmonton's largest day-over-day increases are here). Today is more about situations like "It was really warm last week, so why is it absolutely freezing this week?"


As a comparison, here we have 2016. The way that we're measuring things 2016 had a total of 32 swings, compared to 38 in 2017. But when we limit that to the changes of 15°C or more both years had 18. The largest changes for 2016 were a 23°C drop over a week at the end of April, followed by a 26°C warmup in early May.

The beginning of 2016 was the end of a large El Niño, and we can see that for February through late-April the temperatures were mostly well above the average, with only one large temperature swing during those three months.


Here we have 2015. It had 38 swings in total, compared to 32 for 2016 and 38 for 2017. For 15°C swings it had 19, or one more than 2016 and 2017. The largest swings here were a 25°C drop over 4 days in January, and then a 26°C drop over 11 days in February.

The end of 2015 was part of the same El Niño that started 2016, and here again we can see that for October and November the temperatures were mostly above average. But then things did become erratic again briefly during December.

2010 to 2014

And finally, this chart cycles through the 15°C swings for 2010-2014. Here's the summary:
  • 2010: 13 swings of 15°C or more. The largest were a 28°C warmup over 12 days in May, and a 27°C drop over 9 days in November
  • 2011: 15 swings, including a 30°C warmup over 7 days in January, and a 31°C drop over 9 days in November
  • 2012: 14 swings, including a 30°C 5-day drop in January, followed by a 31°C 5-day warmup.
  • 2013: 19 swings, including a 26°C 6-day drop in January and a 31°C 6-day warmup in May
  • 2014: 20 swings, including a 29°C 11-day warmup in January and a 32°C 9-day warmup in December.

Those are just a few examples, and all the years look pretty similar. With these charts it's tough to compare year-to-year though, or to get a sense of what's going on. That's what we're going to try to do next.


This chart takes the same approach to counting temperature swings, and it applies it to all of Environment Canada's data going back to 1881. In the earlier charts today we'd looked at all the temperature swings, and at the swings that were 15°C or more. This chart also adds counts for the 25°C and 30°C swings.

When we look at all of the temperature swings, we see that the 38 in 2017 was not unusual. Years like 2012 and 2014 were quite a bit higher with 46. And since the 1880s the 5-Year average has stayed fairly constant, ranging from the low-30's to the mid-40's.

When we move to the 15°C swings the 18 in 2017 were again very typical. Recently a year like 2010 only had 13 of them, but 2003 had 25. And again, the average has been around 18 since the 1880s.

2017 starts to look a bit unusual when we move to the 25°C and 30°C swings. 2017 had 7 swings of 25°C or more, including 2 of 30°C or more. Recently 2009 and 2011 both had about that many swings. To find a year which had significantly more we need to go all the way back to 1990, with 9 swings of 25°C or more, including 5 of 30°C or more. Going all the way back to the 1880s it looks like the number of large temperature swings has dropped a bit, although that's probably because we don't see High temperatures of -30°C or -40°C anymore.

So 2017 was not exceptionally erratic, but in terms of really large temperature swings it was a little bit on the high-side of what we would typically see.


Here we have 1990, which is the only recent~ish year which had more large swings than 2017 did. 1990 had 41 swings total, 25 of 15°C or more, 9 of 25°C or more, and 5 of 30°C or more.

It looks a lot like all of the other charts from today, but there are a few really extreme months here. From mid-January through late-February there were a series of 30°C swings, and then in November and December the temperatures bounced around constantly.


Today we've looked at how often Edmonton's temperature swings from above-average to below-average, or visa versa. This isn't just warm days or cold days, but is how often our weather switches between noticeably warm and noticeably cold.

This discussion was prompted by the question: "Was 2017's weather more erratic than normal?" That's a big question, and so to try to answer it I settled on the approach that we're using here - counting temperature swings between the 25th and 75th percentiles. That choice was arbitrary and it doesn't hold any deep meaning, but it did give us a consistent methodology to apply.

On average we saw that these temperature swings happen about 35 times per year, or about every week-and-a-half. That number varies from year-to-year, and during the summer the temperature swings are smaller than they are during the winter. But overall the numbers have stayed reasonably consistent over the last century, although the very largest swings have decreased a bit in the last few years.

The title today "If you don't like the weather, wait 5 minutes" is a quotation that's originally attributed to Mark Twain. He would have been referring to the weather in New England, but it's also a popular sentiment in Alberta. And it turns out that it's not entirely wrong, but rather than every 5 minutes it's more accurately about every 10 days.


February Rain

Today we're continuing our series on rain-in-months-that-you-wouldn't-expect-to-have-rain, with February.

Previously we've looked at:
That means that during the winter of 2017-2018 we've had rain in November, January and now February, but we didn't get any in December so I don't have a writeup for it yet.

Here's the recent history of how often the International Airport has recorded rain during February:

And it's actually pretty sparse, with only 15 rainy days since 1995 that recorded any rain. 

The big standout here is Valentine's Day, 2015 which had a whopping 7.4mm.

In terms of long-term trends:

And again, things are pretty quiet. 

In the 56 years since 1961, there have been 20 that recorded rain during February. Although 7 (now 8 including 2018) of those have been since 2007, so it's possible that the frequency is increasing.

One thing that surprises me is that February is less rainy than January, even though it's usually a few degrees warmer.

January versus February

January has had rain in 33 of the last 56 years, compared to February's 20.

Averages don't have a lot of meaning when the rain is so infrequent, but January's average rainfall is 1.65mm compared to 0.65mm for February. And for a bit more context, March also doesn't get much rain averaging 1.1mm, but then April gets 15mm, and our rainiest month July gets 93mm.

February is 2 or 3 days shorter than January, so there is that to consider. But February is also typically the least snowy of the main winter months, and apparently that applies to rain as well.

Update for February 13, 2018:

Here is the first chart, updated for the rain that fell on February 13, 2018.

The International recorded 0.4mm of rain, and Blatchford recorded 0.7mm of generic precipitation (Blatchford doesn't distinguish between rain and snow, which is why this chart use data for the International).

0.4mm isn't nothing, but it is less than the 1.2mm that fell in 2017 on February 18, and it's also obviously much less than the 7.4mm in 2015.


Birkebeiner Weather

This weekend is the Birkebeiner cross-country ski festival, and the weather looks like it should be pretty good. Let's take a look back at the history of its weather.

Temperature History

These temperatures are for Blatchford, so they aren't necessarily representative of Elk Island where the event is held, but this should at least give an idea of the warm years and the cold years.

1985 was the first time that the Birkebeiner was held in Edmonton, and it was also the coldest with 3 Lows below -25°C and 3 Highs below -20°C. 2008 was another cold one, with 3 Lows below -25°C, and the event was cancelled that year when the windchills were measured down to -46°C.

The warmest Birkebeiner was 2011 with 3 Highs above 5°C, and 3 Lows just below freezing. Recently there had been a string of warm years though, with 2011, 2012, 2013, 2015, 2016 and 2017 all have days above freezing.

The average High temperature has been -4°C, and the average Low -14°C.


This chart shows the measured snowdepth for each year. These numbers are from the International Airport, which again won't be a perfect match for Elk Island,

Going back through the Birkebeiner's History, it looks like the event was cancelled due to lack of snow in 1987, 2001, 2006, and 2016. (and also in 2008 because of extreme windchills)

In 2018 we're at 34cm of snow, which is the 5th highest in the Birkbeiner's history.

Fresh Snow (or sometimes Rain)

This chart shows how much precipitation - either snow or rain - fell during each Birkbeiner weekend. The snowiest year was 1988 with 18cm falling during the Saturday, and the rainiest was 2015, with 7.4mm falling during the Saturday.


January Review / February Preview

A chilly end to January makes it a little tough to remember some of the warm days that we had earlier in the month.

High Temperatures

Here are the high temperatures for January, and we started off warm, got cold, got warm again, and then got cold again.

We didn't set any records this month - either all time or recent - but on the 18th we did hit 8.6°C which was very close to the all-time record of 9.0°C from 2009. We looked at that heatwave in detail in Record Watch - January 17, 18 & 19.

I like to tell people that January always gets some warm days, and in the last 20 years the average has been 11 days with Highs above 0°C. This year we were just a bit below that with 10, and that included 4 Highs above 5°C.

We also had 3 Highs below -20°C, and the last time that we had that many cold Highs was in 2012.

Overall though, the average High for January 2018 was -4.3°C, and that's just a touch warmer than the 20-year average of -4.8°C.

Low Temperatures

Here we have the Low temperatures, with the familiar rollercoaster. We set a recent-coldest (since 1996) Low on New Year's Day, at -27.6°C, and then a recent-warmest Low on January 19th at -1.9°C.

The average Low for the month was -14.3°C. which was just a touch below the 20-Year average Low of -13.7°C.

The deepfreezes felt like they lasted forever, but honestly we were right about where we would expect to be. We had 9 Lows below -20°C including 3 below -25°C, and that's about as typical as it gets for January.

Warm and Cold Months

This chart could use some explanation, and for that it's probably easiest to go back to How Warm is 2017 - the Months. Generally though, it's showing how the average temperature for each month compares to the average for the 20th century, and also to the average for the last 5 years.

January was about 4°C warmer than the 20th century average for January's. But recent January's have all been pretty warm - the 5-Year average is 5.3°C above the 20th century. So January 2018 was warmer than history, but a little bit cooler than the last few years.

As this chart cycles through the years January 2001 really stands out, at 10.8°C above the 20th century average. If we look back to the temperature tables for earlier, we can see that that January had 25 Highs above 0°C, and the coldest Low was only -15°C,

Monthly Snow

January was pretty snowless, right up until the storm we got on the 25th, 26th and 27th. Those 3 days totalled 11.8cm, and that brought our total for the month to 15.1cm. That's a bit below the average of 21cm, but a bit above the 12cm that we got in January last year.

Cumulative Snow

For the winter so far, January brings our cumulative snow total to 60.5cm. That's just about where we were at this time last year, and below the average of 80cm.


Here we have the snowdepths measured at the International Airport going back to 1995. We'd been pretty low so far this winter, although the storm at the end of the month bumped us up to about 18cm.

It's been a few years since we've had a really snow winter - 2010-2011, 2012-2013 and 2013-2014 all reached the 44cm-52cm range. We looked at that in more detail in Snowdepth - Part 2 - The Big Melt.

And here we can see how the snowdepths at the International and Blatchford compare to the average. The recent snowstorm pushed both stations up to around the average. Earlier in the month though, the International had been at 8cm while Blatchford was at 2cm.

In the past I've been suspicious of Blatchford's snowdepth measurements. We see an example of that here, because Blatchford jumped from 1cm of snow on the ground all the way up to 24cm, even though we only got about 12cm of snow during the storm. In comparison the International jumped from 8cm to 18cm, which is more reasonable. For these snow measurements I've always used the International as the main reference, and I will continue to do that.

This chart also shows the snowdepth last year, and it saw significant melting in both January and February.


And finally, here we have the recent history of February temperatures.

February is a lot like January, with usually 10-15 Highs above 0, and 5 Lows below -20°C. It's also on-average the least-snowy of the main winter months.


Edmonton's Largest Snowfalls

Last year we looked at how often Edmonton gets giant snowfalls, and from recent history we saw that we get:

  • 10cm (4"): usually about 2-5 times per year
  • 20cm (8"): maybe every-other-year
  • 30cm (1'): none in the last decade, 3 in the last 30 years.
I was never completely happy with some of the charts in that original post, and so today we're going to take a second look.

Snowfall Events of 5cm or More

This chart shows every multi-day snowfall event with a total of 5cm or more, going back to 1880. Blatchford's snow records are shown in orange and they stop in 2006, and the International Airport's are in blue and they run from 1960 to today. 

One important distinction here is that this chart doesn't show one-day snowfall totals, but instead it takes consecutive days with snow and combines them into snowfall "events." The reason we're using this approach is because snow will often start in the evening and continue on through the morning. Environment Canada's numbers would split that across the two days, but it makes more sense to group them together. 

And so today we are going to define a Snowfall Event as: the total snow on consecutive days where each day received at least an inch (2.5cm) of snow.

There's also a caveat here for the "X Events" counter at the top of the chart, because it double-counts on days that Blatchford and the International both recorded large Events on the same day. I don't think that's a big deal since we're not trying to be particularly scientific, and we'll compare Blatchford and the International later on.

From this chart it's tough to discern any trends, and for that it's probably easiest to go back to the charts from last year:

This chart from last year shows the number of 10cm+ snowfalls each winter, and it hasn't changed much over time. There were a few years around 1900 with a lot of big snowfalls, but other than that 1-4 per year has been pretty typical.

In the post last year I think that we covered the trends (or lack thereof) pretty well. Today we're more interested in the big Events:

25cm or More

This version of the chart only shows the Events with 25cm of snow or more.

Edmonton's largest snowfall Events were: 
  1. 47.5cm over 3 days, from April 18-20, 1955
  2. 46.5cm over 5 days, from December 27-31, 1893
  3. 43.5cm over 3 days, from April 1-3, 1948
  4. 41.9cm over 3 days, from February 29-March 2, 1900
  5. 40.6cm on 1 day, October 25, 1885, and 40.6cm over 2 days on April 6&7 1991.
Most of those are multi-day totals, with the only single-day snowfall being the 40.6cm on October 25, 1885. A few other big days were November 15, 1942 with 39.9cm; May 2, 1886 with 38.1cm; and October 16, 1991 with 28.6cm.

The most recent 25+cm snowfall Event was 25.9cm over 6 days from January 12-17, 2011. And since 2000 there were only 2 other big Events: 35.2cm over 5 days from May 2-6, 2003; and 28.1 over 2 days on April 14&15, 2002.


Here we're looking at all snowfall Events of more than 5cm, since 1995. Because there is a bit more space the labels on this chart are for anything greater than 20cm.

There are 4 Events here where both Blatchford and the International recorded more than 20cm of snow: November 10 1996, April 15 2002, January 20 2003 and May 6 2003. Those dates have both the orange text and the blue text.

There are 4 other 20cm Events that happened after 2007, and so only the International has any data for those.

And then finally there were 4 Events where Blatchford recorded more than 20cm but the International didn't (although it was close twice), and 1 Event where International broke 20cm but Blatchford didn't.

Blatchford vs The International

This chart shows the 46 years where both Blatchford and the International were recording snow data, from 1960 to early 2007. It cycles between both stations, Blatchford only and the International only.

We can see that there's a lot of overlap between the two. Across these years Blatchford recorded 351 of these large events, with an average of 10.9cm. At the International there were 361 events, with an average of 10.2cm.

But we'll end off with one more graph from last year:

Today we've talked about large snowfalls - in the range of 5cm, 10cm, 20cm and beyond. And this chart shows how often they happen. At both Blatchford and the International almost half of the time when it shows we get less than 1cm, and we only get more than 5cm about 16% of the time.


Record Watch - January 17, 18 & 19

We've just had a couple of pretty warm winter days, and today we're going to put them into context.

January 17

On Wednesday we hit a High of 6.7°C. That' was the 6th warmest January 17th recorded in Edmonton, down a bit from the 7.5°C we had last year in 2017. The record is 8.3°C set in 2009.

The Low of -5.2°C was well below the 0°C record for warmest-Low set in 1976.

January 18

Our High temperature this year on January 18 was 8.6°, which was a perfect match for last year which was also 8.6°C. So 2017 and 2018 both tied for 3rd place, behind 8.9°C in 1900 and 9.0°C in 2009.

The Low this year was -5.2°C, which is in about 10th warmest spot. In comparison, 2017's Low was -1.5°C. The record warmest-Low was 3.3°C in 1900 again, and the coldest was -43.9°C in 1886.

January 19

January 19th's High of 5.2°C put it in 13th place, well below the record of 10°C set in 1944.

The Low of -1.9°C was the 6th warmest, with 1891 holding the record at 2.2°C. The record Low for January 19th is -49.4°C set in 1886, which is the coldest temperature ever recorded in Edmonton. 3 days all share the honour of having records at -49.4°C: January 19 1886, January 21 1886, and February 3 1893.

Top 50 for January 15-21

And here we have the 50-warmest days recorded in Edmonton for the week of January 15-21st.

The lowest record for this week is 7.2°C for January 16, which was set way back in 1884. The highest is 10.7°C for January 20th set in 1981, followed closely by 10°C for January 19th set in 1944. The records for the other 4 days are all between 8.3°C and 9.1°C.

The grey shading breaks things out by decade, and so far the 2010's have 10 days in the Top-50 with entries in 2013, 2014, 2017 and 2018. The 1940s have 7 days, the 1890s have 6 days, and then the 2000s have 5. The cutoff for the Top-50 is 6.1°C, so our 5.2°C on the 19th this year doesn't make it in.


2017-2018 Coldsnaps So Far

Today is the end of a short coldsnap, so we're going to check-in and see how the winter of 2017-2018 compares so far.

This is a followup to Coldsnaps & Deepfreezes from last year. And if you're interested in Edmonton's historic deepfreezes, please see It was the winter of '69.

Lows Below -20°C

This chart shows all of the days with Low temperatures below -20°C, going back to the unusually cold winter of 2010-2011. The number are how many continuous days were below -20°C.

This most recent coldsnap felt like it lasted forever, but it was really only 5 days. Looking back through recent years 5-6 days is a pretty typical length.

Between Christmas and New Year's this year we had 9 days with Lows below -20°C, and the last time we had a coldsnap that long was back in Dec/Jan 2010-2011.

So far this winter we've had 14 days below -20°C, and the average is about 23. We've already passed low winters like 2015-2016 with 6 and 2011-2012 with 10, but still have a long way to go before we catch 2010-2011 which had 48.

Lows Below -25°C

In our two coldsnaps so far this winter we've had 9 Lows below -25°C, which isn't a huge number, but which is more than most of the years here. The only recent winters with more were 2010-2011 with 16 and 2013-2014 with 17.

Highs Below -20°C

This year our two coldsnaps have also included two 3-day stretches of Highs below -20°C, and we hadn't had one of those since January 2012. And with 6 Highs below -20°C so far we're already above all these other years, except for 2010-2011.

Normally Highs below -20°C happen in December or January, but both 2011 and 2014 had 2 each in early March.

Highs Above 0°C

For a change of pace, here are our warm days so far, with Highs above 0°C. The numbers here are a little messy to read, but that's a good thing because it means that there are lots of them.

The 15 straight Highs above freezing that we just had in December was the longest mid-winter streak of all the years here. And with 32 days above freezing so far we're about halfway to the average.

Highs Above 5°C

And finally, here we have the days with Highs above 5°C. We'll hopefully see quite a few more of these as we get closer to spring, but with 13 so far we've almost caught 2012-2013. The 7-day streak in December was again on the long side of things for the middle of winter.