2021/07/19

Heatwaves & Heat Islands: Part 2

We are back to talk about the heat island effect again today, and how heatwaves may or may not exaggerate the temperature differences between rural and urban stations.

In Part 1 we looked at the individual daily temperatures during 3 of Edmonton's most notable heatwaves - 2021, 1961, and 1924. On extremely hot days:

  • The High temperatures for all of the stations around Edmonton are generally pretty close to one another.
  • The Lows temperatures for the Blatchford station in Edmonton are generally several degrees warmer than at the outlying stations, but that occurred in all 3 of the heatwaves going back to 1924.

Today we want to see if there are any trends to be seen, and we will be focusing on a few Edmonton-area Environment Canada weather stations.

  • Blatchford is in Edmonton, about 2km north of downtown with data from 1880-today.
  • Campsie is 100km northwest of Edmonton with data from 1913-today. 
  • Calmar is 40km southwest of Edmonton with data from 1916-2007.
  • The International is 30km south of Edmonton with data from 1961-today. 
  • Stony Plain is 40km west of Edmonton with data from 1966-today. 

Campsie is the most truly rural station of the bunch:

 

Summer Temperature History

To start with, here is the history of 10-year average temperatures for these stations for the "summer" months: from June 1st through August 31st.

Blatchford is the longest line in dark blue, and for the Highs temperatures it is mostly in the middle of the other stations. Stony Plain in Pink is the most recent station to start, and in the beginning its summer Highs were well below the other stations, although since about 2005 it has closed most of the gap.

For the Lows Blatchford is well above all of the other stations, with Stony Plain in the second spot. The Lows for Calmar, Campsie and the International are all clustered together.

In terms of trends the average summer Highs for most of the stations look pretty flat, with the exception of Stony Plain which has jumped up closer to the other stations. For the average summer Lows all of the stations have warmed, except for the International which has dropped 0.2°C since the 1960s. For the other stations Blatchford is up 3.9°C over 140 years, Campsie is up 2.5°C over 100 years, Calmar was up 2.2°C over 91 years, and Stony Plain is up 0.7°C over 55 years.

So in general the average summer Highs around Edmonton aren't all that much different than they were a century ago, but the summer Lows are up by 2.5~4°C. We talked about this more in Why Heatwaves Aren't the Whole Story. (also, today we are just looking at summer temperatures, but the temperature changes during the winter are much larger).

The word of the day is "Relative" because what we really want to know is how the temperatures at these stations have changed compared to one other. If the heat island effect disproportionately affects big cities, then theoretically Blatchford's temperatures should have warmed faster than the outlying areas.


Summer Temperature Comparison
This chart uses the temperatures at Blatchford as a reference, and it shows the average difference each summer (June 1 - August 31) for each of these stations. 

For the average summer Highs Calmar was actually a little warmer than Blatchford after around 1930. And as we saw earlier Stony Plain's Highs started 1°C or more colder than Edmonton, caught up to the city for 2010~2015, and then just recently have been 0.8°C cooler again. For Campsie and the airport the differences have bounced around, but similar to Stony Plain for the past few years they've been about 0.5°C cooler than the city. 

For the average summer Lows Stony Plain has been closest to Edmonton, at around 1°C cooler than the city since the 1960s. For the other stations the average Lows are now around 3~3.5°C cooler than the city, and that gap has grown by about 1°C since records started.

Is that ~1°C change the Heat Island Effect? It could be, although for Campsie and Calmar it looks like most of the change happened from 1915-1960, and things were pretty flat after that. In 1960 Edmonton's population was 269,314, and the city had a much smaller footprint than it did today.

For the airport it does look like the gap widened around 2000, but that's probably because the airport is weird. Here's the same chart, except that it is flipped to use the airport as the reference station:

Using the airport as a reference we are only looking at 1960-today. And around 2000 Blatchford did get warmer than the airport, but so did Campsie and Stony Plain. Stony Plain has urbanized and grown since 2000, but Campsie hasn't, so this is probably more about how notoriously cold the airport is, rather than the Heat Island Effect.

Summing up so far:

  • Blatchford's summer Highs are pretty close to the neighbouring stations, and have been since 1915.
  • Blatchford's summer Lows are 2~3.5°C warmer than the neighbouring stations, and have been since at least the 1960s. Going further back to 1915 Blatchford was still warmer, but by slightly less at maybe 1.5~2.5°C.

So now we know what the history of typical summer temperatures looks like. What about the heatwaves?


30°C Days Temperature Comparison
This is the same chart again, except this time instead of looking at the whole summer we are focussed on days where Blatchford's High hit 30°C. 

On thing to keep in mind with this chart is that we're dealing with a very small sample size. Edmonton averages about 4 30°C days each year, but many years only get 1 or 2, and some get none. And so each bar in this chart is the average temperature difference from only a handful of days. That's why for the Highs Campsie has 1 giant bar in 2011 for a single day when Blatchford hit 31.5°C, and Campsie was down at 20.5°C. Also, in this chart years which had lots of 30°C days will be an average of all of those days, and might look less extreme than single days. But we wanted to talk about what the Heat Island Effect looks like on Edmonton's hottest days, and these are the hottest days.

For the Highs the differences are pretty small, with Blatchford typically 0.5~1°C warmer than the other stations on these very hot days. And as for the question of how that has changed over time: it really hasn't, has it? Stony Plain has actually closed the gap with Edmonton by about 0.5°C, and the other stations are all around where they were 50 or 100 years ago.

For the Lows the differences are larger than the Highs again. Stony Plain is pretty close to Blatchford with overnight Lows about 1°C cooler on these hot days, and that gap has closed from 2°C 50 years ago. For the other stations the gap is in the range of 4~5.5°C, which is up from 2~3.5°C a century ago. Again though, for Calmar and Campsie most of the change happened from 1915-1960. For the International the gap has increased from 3.7°C to 4.6°C since 1960.


32°C Days Temperature Comparison
We can play this game one more time with 32°C days, which happen on average about every other year. 32°C days are uncommon enough that most of the bars in this chart represent the temperature differences on a single day (or maybe two) when Blatchford hit 32°C.

For the Highs, in recent years when Blatchford has hit 32°C:
  • Campsie has generally been about 1.5°C, but sometimes it's or 3.5°C cooler, and a few times it has been warmer. Going back through history most of these bars are in the 0/1.5/3.5°C coller range (with 1932 as a standout at 7.7°C)
  • The Airport has been 1.5~3°C cooler, which hasn't changed since 1960.
  • Stony Plain has been 1~2°C, while in the 1970s & 1980s the difference was 2~3°C.
  • Calmar doesn't have recent data, but from 1915-2007 it was in the 1~2°C range without much change.
For the Lows when Blatchford has hit 32°C:
  • Campsie Lows are about 5~7°C cooler, which has been pretty typical since about 1960. Before that the difference was around 4~5°C.
  • The Airport's Lows are also 5~7°C cooler, compared to 3~5°C in the 1960s & 1970s.
  • Stony Plain's Lows range from 1°C warmer to 2.5°C cooler than Blatchford's. That is warmer than in the 1960s & 1970s when the difference was 4~5°C.
  • And Calmar was in the range of 3°C for 1915 through 1960, and then 4°C for 1960 through 2007.

15°C Lows Temperature Comparison
One notable thing about the 2021 heatwave was the large number of warm overnight Lows which we had, so here is a final version of the chart for nights when Blatchford stayed about 15°C. Many of these dates will overlap with what we saw in the 32°C chart, but it's worth focussing in on those hot nights in particular. (once again though, this is a very small sample size)

And what do we see?

On those hot night the Highs at the outlying stations were generally ~1°C cooler than Blatchford, although there doesn't seem to be much of a change from 1915 to 1960 to 2021.

And for the Lows Blatchford was 2~5°C warmer than the other stations. Most of the stations look pretty flat, although with Campsie it again looks like the gap got bigger from 1940~1960.

You can also set the dashboard to show the really hot Lows like 17°C, or 19°C, but prior to 1960 there were almost none of those and so the history is shorter.


Summary
So what was the point of all of this?

This image shows the daily temperatures from the great heatwave of 2021. In the past on this blog we have repeatedly discussed how overstated the impact of the Heat Island Effect is. And yet...when I saw these temperatures my first thought was "Wow, that's some big Heat Island Effect."

But when we take a closer look, was this really the Heat Island Effect? The High temperatures were actually quite close across the region. And for the Lows Blatchford was definitely much warmer than neighbouring stations, but that has been the case since at least 1915.

So today leaves us with some unanswered questions. Blatchford is warmer than its surroundings, but that was also mostly the case a century ago, so what is going on? All that we can do is speculate, but a few ideas come to mind: 
  • The downtown area might have a microclimate which makes it warmer than the surroundings. That might be why there is 8,000 years of history of Indigenous use of areas like Rossdale, rather than at a place like the airport? (disclaimer: I do not know anything about the Indigenous history of locations like Stony Plain or Campsie)
  • Blatchford is still a big empty field with a few old runways, next to a freeway and a railyard. Maybe it doesn't actually feel the Heat Island Effect the way an downtown does.
  • Maybe the Heat Island Effect's biggest impact is early on, when a city grows from 0-400,000 people. As a city spreads (or sprawls) outwards the Heat Island might affect a larger area, but it doesn't get more concentrated.
  • Maybe the Heat Island Effect is really, really big, and even a location like Campsie 100km away still feels the majority of it.
That's all speculation though, and speculation is terrible. For all I know a Wizard did it. 

But I think there are 2 big takeaways from today:
  • When we see a 5°C temperature difference between Blatchford and the airport, that does not mean that the airport is the "natural" temperature, and that the big city has warmed itself up by 5°C. The vast majority of that temperature difference also existed 50 or even 100 years ago. And we now know that is also the case during extreme temperatures like heatwaves.

  • For the Calmar & Campsie stations with long histories it does look like the city's summer Low  temperatures got relatively warmer by about 1°C from 1915~1960, but then after that things flattened out.

1 comment:

  1. Your posts always make me think. This time I am thinking about how much of the heat island effect is observed versus actual. Fruit walls and wind-chill perceptions would be my context for comparison. Protecting marginally hardy fruit trees from wind has an observable impact on tree survival and production. This hyper-local observation does not hold true across the heat island effect, only locally where shelter is maintained, but is still scalable throughout the city (increased survival rates for trees planted leeward of structures favorably altering prevailing winds). Perhaps the air mass is so massive and dynamic that it is not measurably altered at all by our puny local structures but we still experience these local alterations all the same and then we report them in our multitudes?

    ReplyDelete