2016/08/17

Edmonton's warmest years (or NOAA anomalies, NOAA problems)

I've seen this chart show up in various places online:
  • NOAA National Centers for Environmental information, Climate at a Glance: Global Time Series, published July 2016, retrieved on August 13, 2016 from http://www.ncdc.noaa.gov/cag/

It's done by the National Oceanic and Atmospheric Association (NOAA) who are out of the US, and it shows how the average temperature for each year compares to the average for the 20th century.

That got me wondering what a similar chart for Edmonton would look like, and here it is:


The Edmonton data is all-over-the-place compared to the NOAA chart, but there are a few reasons for that:
  • The scales are quite different, with the NOAA chart measuring a maximum anomaly of 0.9°C from the 20th century average, where Edmonton has values all the way up to 3°C and down to -4°C.
  • The NOAA chart is averaging temperatures for the entire world - land and ocean. Obviously the ocean is great at moderating temperatures, but even with Edmonton data we've seen that looking at averages removes the noisiness and really flattens things out.
Here is the Edmonton chart again, this time with the NOAA data overlayed on the same scale:


And here we can see that Edmonton's variation is much higher than what NOAA records, although again that is to be expected since NOAA is looking at the entire planet.

One thing that stands out in the Edmonton data is the number of warm years in the 1980s. This seems a bit surprising, because I'm not sure that anyone who lived here would remember them that way.

To doublecheck things, here is the Blatchford data as well as the data from the Edmonton International Airport:


This chart only goes back to 1961, because that is when the data for the International Airport starts.

We shouldn't expect the two airports to be a perfect match here, because the math behind them is a little bit different: Blatchford is compared against the average for the entire 20th century, where for EIA it's just 1961-2000. But even with that the data does line up pretty well. For the most part the cool years were cool at both airports, and the warm years were warm, and the warmest years were definitely in the 1980s.

This doesn't necessarily mean that those winters were particularly warm though. This is looking at the average temperature for an entire year, so it could be a warm winter, or a hot summer, or warm nighttime temperatures, or some combination of all of that.

For an explanation of where these numbers actually come from, here is an example of the calculation using the 2015 data for Blatchford:


The top section of this chart shows the 20th century average temperature for each day of the year, as well as the actual temperature that was recorded in 2015. The bottom section shows the difference between those for each day. Those differences for each day are then averaged for the year.

For 2015 the result is 2.5°C above the 20th century average. And while we can see that 2015 was consistently above the average throughout the year, it got a lot of its "points" during a warmspell in late January with several weeks that were 10+°C above the 20th century average.

So with all of that out of the way then, here are Edmonton's warmest and coldest years:


The 1980s are the big winner here, taking 1st, 2nd and 4th spots. The 2000s also had 3 years in the top 10, in the 6th, 8th and 9th spots. So far, the 2010s have only had one year in the top 10, with 2015 coming in at #3 (2016 isn't included, but I'll talk about it a bit later). And the rest of the top 10 is filled out with 1992 in 10th, 1976 in 5th, and the 7th spot is way, way back in 1889.

I've mentioned before that I'm a little suspicious of the very old data, and 1889 at #7 seems a little weird, especially considering that the #1 coldest year was just two years earlier in 1887. It makes me wonder if maybe someone moved the thermometer too close to the woodstove that year?

For the coldest years, 5 of them were in the 1880s, and the rest are scattered up through the 1950s. In terms of recent cold years, 1996 is the only really cold one coming in at #11.

(a few of these years - notably 1880, 1887, 1992 - are missing significant chunks of data. I didn't want to just leave them out entirely though, so in cases where data is missing I have substituted in the 20th century average for that day. For the purposes of this calculation that keeps things neutral, and doesn't unfairly reward or penalize any year. For 1880 and 1887 this likely makes them appear warmer than they actually were, and they still rank as #1 and #2 coldest).

For one more comparison, here is an animation with the warmest and coldest years for Edmonton and Calgary:


The Calgary data starts a few years late, so it actually misses many of Edmonton's coldest years. But beyond that there is quite a bit of overlap in the top and bottom 10.

1981, 1987, 1988, 2006 and 2015 are in the top 10 for both cities, although with slightly different rankings.

For the coldest years, 1887, 1927, 1950 and 1951 are in both lists. And 1996 is close, in Calgary's #6 spot and Edmonton's #11.

I'd mentioned being suspicious of 1889 in 7th spot for Edmonton, and it was a warm year in Calgary too - although not that warm. In Edmonton it was almost 2°C above the 20th century average, compared to 1°C above in Calgary. That still makes it one of the warmest years in Calgary for 40 years, but it doesn't come close to the top 10.

Finally, going back to NOAA:


Here is the Edmonton data, with the yearly numbers pushed to the background, and the 5-year average added on top to remove some of the noise. The Edmonton data isn't a perfect fit - the large 1980s hump being a notable difference - but for the last 40 years we can see that both NOAA and Edmonton have been consistently above the 20th century average.

I haven't included 2016 in any of these comparisons today, because the way the math works it's not really fair to compare a year that's only half over against all of the others. So far though, 2016 has definitely been warm. In a few weeks I'll do a day-to-day comparison to see exactly where 2016 ranks.

No comments:

Post a Comment