Relatively Extreme: Heat Waves in NL

At the risk of stating the obvious: It’s been hot in Newfoundland this summer. This july we broke more than a few temperature-related records, gaining the attention of both local and national media. The phrase we kept hearing last month: ‘heat wave’. But whether that term fits depends who you ask.

Heat waves are one of the most confusing weather events meteorologists track, mostly because there seem to be as many definitions for heat waves as there are meteorologists. Criteria vary dramatically between organizations, countries, and even within countries (the U.S. alone has several definitions), but all require

  • temperatures to rise above a set level
  • for a minimum number of consecutive days

By Environment Canada’s definition (highs at or above 32oC for at least 3 days), heat waves are extremely rare in Newfoundland. For example, St. John’s hasn’t officially hit that 32oC mark once, let alone for three days. But this is a geographically limited definition, suitable only for warmer parts of the country. It’s useful for identifying events that pose serious health hazards, but will miss many locally significant events. And certainly this July felt like a heat wave: a long stretch of unusually high temperatures (unusual for us, at least).   That character of ‘unusual’ has to count for something, right? Fortunately, the World Meteorological Organization (WMO) agrees, and offers a more flexible definition based on local climatology. To the WMO, a heat wave occurs when

  • daily maximum temperatures exceed local climatology by at least 5oC
  • for a minimum of 5 consecutive days

Individual heat waves are separated by at least one day that does not meet the temperature requirement. By this definition, heat waves can occur anywhere, but will always be an extreme event – locally rare, with the potential to stress the surrounding environment. We can consider these long ‘hot spells’, a broader family of events that meet the temperature criteria of a heat wave, but without any consideration of duration.

Applying the WMO’s definitions to St. John’s can help us put this July in context and estimate how unusual it was. First, we need to select a threshold; on average, the daily summer high in St. John’s is 18.7oC. By WMO standards, hot spells (and heat waves) should then exceed 23.7oC; let’s round this to 24oC. Using this definition, a typical St. John’s summer will experience 8.5 hot spells. Heat waves are much less common; on average, only one occurs every 2.7 years.

So what about this year? To date in 2014, St. John’s has had six hot spells and (drumroll)… four heat waves. All began in July, and the last ended on August 5th. In climatological terms, that’s more than we’d typically expect in a decade

So is this climate change?
Whenever we see something like our record breaking sequence of July heat waves, we need to reconsider our definition of ‘normal’ climatology. How well does this unusual series of rare events fit with our past climate? With a little statistics (click here if you really want to see the details), we can address a few relevant questions:

Q: How unusual is it to have 4 heat waves?
A: Unusual, but not impossible. We’d expect four or more heat waves roughly once a century. Other years have seen multiple heat waves; for example, three occurred in 1984, and two occurred in 1998, 2003, 2004, and 2008.

HW_count

Q: Were these events usually long?
A: Nope. This year we’ve had two 5 day and two 6 day heat waves. St. John’s has had much longer events, including one in 2003 that lasted 13 days.

HW_Max

Q: Were these events unusually warm?
A: Again, no. The warmest event had a peak temperature of 29.5oC; we’d expect one heat wave in forty to hit these temperatures.  As for breaking temperature records: this is still two degrees below St. John’s all time high of 31.5. (Note: we did, however, break humidex records, which accounts for the impacts of humidity on perceived heat.  But that’s a whole other mess to sort through…)

HW_Duration

So: is this July a sign of climate change? Alone, no – within the context of our past climate, this July falls firmly into the category of ‘novel, but not unexpected’. But there are indications that these are becoming more common; of the 27 heat waves that have occurred between 1950-2014, 41% happened on or after 1998.   Years with multiple heat waves are apparently more common (it’s happened 4 times since 1998, but only once from 1950-1997). Altogether, this suggests that the past climate record may no longer reflect our present. While it’s still too early to say what this means for our future, it tells us we should keep an eye on our thermometers in coming years.

Heat Wave Stats: The sausage gets made

To assess probability, I used a point process model for hot spells (see here). This blends several distributions to model:

  • the frequency of hot spells (a Poisson fit to annual counts)
  • the duration of individual spells (a geometric distribution)
  • the maximum hot spell temperature (a generalized extreme value distribution)

There are a few assumptions that are violated here – specifically, hot spells are assumed to be independent and identically distributed, which is not necessarily true due to autocorrelation in temperature. Similarly, a geometric distribution assumes all trials are independent, which again is violated by autocorrelation (the chance of exceeding the threshold increases with the previous day’s temperature; our t isn’t fixed!). The first issue is partly addressed by ‘declustering’ the data; while individual days are definitely not independent, individual hot spells separated by sufficient time will be. Here, we consider a one day separation between spells enough to ensure independence – really, we should probably use something longer (2-3 days), but it doesn’t drastically change the results. As for the concerns around the geometric distribution: they aren’t addressed, but we’ll file it under ‘good enough’.

To fit the distributions, hot spells were identified in bias-corrected station data for St. John’s Airport (get it here); these were defined as periods with daily maximum temperatures above 24oC. This threshold choice matches the WMO definition of a heat wave well, but is also justified by threshold tests commonly used in generalized Pareto (GP) distributions (specifically, GP parameters are stable for thresholds above 23-23.7oC).

Once hot spells were identified, the starting date, duration, and peak temperature were recorded.   Using data from 1950-2011, the following distribution parameters represented the best fit to our hot spell model:

  • Frequency (Poisson Model): annual counts of hot spells were used to fit the distribution; results gave lambda = 8.53 (the mean number per year). With an annual count variance of 11.37, the model is somewhat overdispersed but remains a reasonable fit to the data.
  • Duration (Geometric): A geometric distribution is useful when estimating the length of unusual events – or number of trials before a success occurs. Here, the question is how many days typically pass before a heat wave ends. It uses a single parameter, theta, which here gives the probability a day below the threshold will occurring. Fitting to the hot spell durations gives theta = 0.3633.
  • Maximum Temperature (Generalized Pareto Distribution): Fitting the generalized Pareto distribution to hot spell temperature maxima (recorded as the highest temperature recorded during each spell) gives a scale parameter of 2.407 and a shape parameter of 0.026. Based on 529 individual hot spells, the results are relative robust (low parameter uncertainty).

In order to estimate how unusual our record breaking string of four heat waves is, I created a huge number of fake ‘years’, in which i) the number of hot spells was determined by a random sample of the fitted Poisson distribution ii) and the duration of these spells was given by a random sample of the fitted geometric distribution. In 106 ‘years’, the best-fit parameters given above give a 1.3% chance that 4 or more heat waves will occur. Taking into account uncertainty in the Poisson and geometric parameters, the 99% confidence interval on this is [0.4%, 3.15%].

Crushed Hopes: The Climatology of the Victoria Day Long Weekend on the Avalon

Image
A typical early summer scene in St. John’s.

In much of Canada, the Victoria Day weekend marks an unofficial start to the summer season. Across the country, cottages and cabins are reopened, camping gear is dusted off, and patio umbrellas begin to bloom. We enthusiastically shed our winter wardrobe and embrace summer’s promise for outdoor fun… just as long as the weather cooperates. A little good weather can go a long way to making this weekend memorable, while bad weather can quickly sour our late spring exuberance and leave us doubting that summer will ever arrive. But this is Canada, and cold wet weather is common in late May; in weather-battered Newfoundland, it’s the norm. So perhaps our best bet for maximizing our enjoyment of the weekend is to arm ourselves with realistic weather expectations – that is, a firm understanding of the May 24th weekend climate, based on observations at some of the island’s airports (St. John’s, Gander, and Stephenville).

Let’s focus on St. John’s, since the majority of our population lives on the Avalon Peninsula. I’m also a ‘bad news first’ kind of person, and the Victoria Day climate in St. John’s certainly qualifies. While everybody has his or her own impression of what constitutes ‘good’ weather, it’s doubtful the typical St. John’s Victoria Day forecast fits anybody’s definition. Cool temperatures, frequent fog, and a very high chance of rain conspire to dampen any optimism that summer is just around the corner. Based on the last 30 years (1984-2013), a forecast for the three-day weekend would read:

“An 83% chance of rain over the weekend, with 17mm expected from midnight Friday through midnight Monday. Daily highs near 11oC, cooling to 2oC overnight. Morning fog expected, mixed with rain and drizzle. There is a slight chance conditions will clear up briefly, before clouds and fog return.”

Let’s break it down in greater detail. We’ll focus on daytime hours alone (6AM to 9PM), since we’re happy to sleep through late night bad weather. On average over the holiday weekend, St. John’s can expect only seven hours with sunshine (clear to mainly clear skies), nine daytime hours with precipitation, and a full thirteen (!) with fog. Roughly, 1 in 4 Victoria Day weekends remain foggy and/or cloudy throughout all daylight hours, and 5 in 6 see daytime rain at some point over the weekend. Put all of this together, and the chances of a glorious weekend full of summer’s promise are…well, not great.

Maybe we should lower our expectations, and hope for just one good day. After all, one in three’s not bad, right? Looking through the climate record, we can check how often we see at least one day with a solid stretch of good weather during the daytime. We’ll set our bar relatively low, and expect i) fewer than 5 hours with precipitation, ii) at least 5 hours with clear-to-mainly-clear skies, and at least an 8 hour stretch with temperatures above a modest 5oC. Unfortunately, results aren’t encouraging: only 2 of every 5 years will have any days meeting these criteria. Only four years in the past thirty featured two good weather days, and there is one solitary year with good weather throughout. All told, seventeen of the ninety Victoria weekend days in the past 30 years meet our (admittedly subjective) criteria.

Enough of the bad news; let’s talk coping strategies. Avalon residents could simply resign themselves to poor weather, and stoically head to the cabin or campsite regardless, armed with rain gear and a few packs of cards. Or we could dig a little deeper into the climatology and see if there’s a better time or place to usher in a start to summer. For example, we could consider trading in Victoria Day for America’s Memorial Day weekend, which arrives a week later. This makes a noticeable difference in St. John’s, with the number of good weather days rising to 24 out of 90, with at least one good weather day appearing in 19 of the last 30 years. That’s a full 23% improvement in our chances, just for shrugging off a minor monarchic tradition.

Better yet, we could consider leaving the Avalon in search of better weather. Conditions improve a little in Gander (20/90 days had good weather), and a lot in Stephenville (38/90 days), suggesting this might be a great weekend to check out the island’s interior or West Coast. This second option looks particularly appealing, with 12 of the last 30 years featuring two or more good weather days, and only 8 years with none. For Newfoundland weather, those are tremendous odds. Gander and Stephenville also see an improvement from a shift from Victoria to Memorial Day, respectively gaining an additional 6 and 4 good weather days.

May24
The U.S. wants you to have a better weekend than the U.K. 

So if you live on the Avalon, don’t get your hopes up for good weather this Victoria Day. Unless you’re willing to drive west, in which case maybe we can carpool? I’ll bring the snacks, and a Petition to Recognize Memorial Day.