Adelaide is the hot place to be right now. We’re in the middle of an extreme, enduring heatwave, and the city’s residents are suffering. Indeed, we’ve had rolling blackouts as the power system fails to meet peaking loads, and more people are suspected to have died from heat stress over the last week than were killed in the infamous Ash Wednesday bushfires of 1983.
Now, make no mistake, the month of January in Adelaide is renowned for its hot, dry weather. For instance, back in 1908 the city felt the brunt of a run of 6 days above 40C, and in 1939 the temperature soared to its historical single-day record (still unbeaten) of 46.1C. So given this context, an obvious question is whether the current heatwave is anything remarkable, and can a climate change signal be detected in this event? The standard scientific answer I would usually give is something like this: ‘Extreme weather has occurred in the past, and it is not possible to definitively attribute any one unusual event to climate change. That said, a higher frequency of intense heatwaves like this is consistent with the expectations of a rising global temperature‘ (see figure showing ‘more record hot weather’). But in the case of Adelaide’s 2009 heatwave, a bit of deeper investigation does indeed suggest that a climate change link is very likely.
Starting on January 26, we’ve had daily maxima of 36.6C, 43.2C, 45.7C [3rd hottest day ever recorded], 43.4C, 43.1C, 41.1C, 40.6C and 38.8C. This also included the hottest night ever recorded in South Australia, when around midnight on 29th Jan, it dropped to a minimum of 33.9C. For the last 5 days the temperature has not dropped below 25.9C at night. The run of 6 days above 40C equals the record from 101 years ago. The current 5-day forecast is for 38C [update: actual = 36.3C], 38C [33.0C], 37C [35.6C], 40C [43.9C] and 34C [41.5C]; if this holds, we’ll have had a string of 12 days above 35C, or perhaps 13 if Saturday nudges up a degree or so over the forecast. The heatwave is hitting more than Adelaide by the way — Melbourne got its 2nd hottest day on record and Tasmania its hottest ever. The town of Kyancutta on the Eyre Peninsula sizzled at 48.2C.
Historical records from the Bureau of Meteorology show that there have been 6 previous ‘heatwave events’ (here defined as >35C) that lasted 8 days, many more of 7 days, more still of 6, and so on. This is useful information for analysis, because it turns out that the return time of any given string of hot days is logarithmically related to it’s length (see below for how I know this). From Bureau records, we can infer that if the current heatwave does last for 12 days, such an extreme outlier should only occur, by chance (i.e. if the climate is not trending), once every 400 years or so. If it goes for 13 days, then that’s roughly a 1 in 1000 year event — such is the nature of a logarithmic relationship! Statistically speaking, there is always a danger in extrapolating beyond the bounds of your data, but in the case of rarely (or never) observed events, there is little other empirical recourse (a mechanistic simulation such as a general circulation model would also give useful inference on this matter).
So why might I be as bold as to suggest a climate change link with this current heatwave? Well, the hottest night ever recorded is certainly notable, and is consistent with the expectations of a greenhouse gas (GHG) induced warming. You see the basic hypothesis goes that if solar forcing is causing the surface warming, we’d expect there to be relatively more hot days (when the sun is shining) compared to hot nights (when the sun is hitting the other side of the planet and the warmth is maintained by the insulating effect of GHGs and high altitude clouds). If it’s a build-up of GHG that is driving the warming, you’d conversely expect relatively more hot nights, because the ‘atmospheric blanket’ has thickened. Another way to think about this commonsense prediction is to consider deserts, which have few high clouds and so tend to get very cold at night. Or in the extreme, to look to the Moon, which has no GHG at all and so has daily temperatures soaring above 100C and night times plummeting to below -100C! But let’s not read too much into one extremely hot night in Adelaide — it could still just be unusual weather, after all. Better in this case to reflect on large meteorological data sets which quite clearly confirm the GHG prediction, showing that relatively more warming has occurred at night compared to the day time. Another nail in the solar warming idea (but that’s another story).
Okay, so if it’s not the hottest night on record that seals the deal in my mind, what is it? Well, it’s another recent heatwave actually — one that occurred only 10 months ago in Adelaide. The record-smashing March 2008 heatwave persisted for a whopping 15 straight days and surpassed the previous longest stretch of above 35C degree temperatures recorded in any Australian capital city (formerly held by Perth in 1988). Poor old Kyancutta took a hot hit then too, recording 13 days straight of over 40C temperatures. The hottest night ever in March in Adelaide, of 30.2C, was also recorded during this event. Overall, last year’s heatwave was not as consistently hot as this year’s event (see chart), but it will still likely hold the record for duration. A frequency analysis on this monster implies that it was a 1 in 3,000 year event in a stationary climate! Read the media release by SARDI to show how Dr Warwick Grace calculated its return time. (Put another way, that’s about as likely as tossing a coin 12 times and getting all heads [or tails] — try it, I dare you…)
Now global warming is quite clearly expected to both increase the frequency of heatwaves (i.e., greater number of events per unit time) and cause those heatwaves that do occur to be hotter and to last longer (on average). This is a fairly simple expectation one derives from a change in the average temperature, as the top figure in this post indicates. The Intergovernmental Panel on Climate Change Fourth Assessment Report (2007) has this to say about it:
“Since 1950, the number of heat waves has increased and widespread increases have occurred in the numbers of warm nights… In several regions of the world, indications of changes in various types of extreme climate events have been found. The extremes are commonly considered to be the values exceeded 1, 5 and 10% of the time (at one extreme) or 90, 95 and 99% of the time (at the other extreme). The warm nights or hot days are those exceeding the 90th percentile of temperature, while cold nights or days are those falling below the 10th percentile… In the last 50 years for the land areas sampled, there has been a significant decrease in the annual occurrence of cold nights and a significant increase in the annual occurrence of warm nights. Decreases in the occurrence of cold days and increases in hot days, while widespread, are generally less marked. The distributions of minimum and maximum temperatures have not only shifted to higher values, consistent with overall warming, but the cold extremes have warmed more than the warm extremes over the last 50 years. More warm extremes imply an increased frequency of heat waves.“
So, in Adelaide we have two freakishly rare extreme events happening with a 10 month period. How likely is that? Well, if the events are totally independent, we’d expect the joint likelihood of two such heatwaves (of 0.25% probability per year [the 2009 event] and 0.033% per year [2008 event], respectively), occurring within the same 12 month period, to happen about once every 1,200,000 years. Is that unlikely enough for you? But if there is ‘autocorrelation’ (dependencies between the two events due to a linked cause — such as climate change), this calculated probability is not valid.
What exactly do I mean by this? Well, the heatwave that struck Europe is 2003 provides a good way to illustrate my final point, thanks to a neat analysis published in Nature in 2004. The authors of this study estimate that warming to date has at least doubled the probability of such an extreme heatwave occurring. Moreover, under ongoing heating, climate models suggest that by 2040, this extraordinarily hot summer (in historical terms) will be just a run-of-the-mill average summer. By 2060, it will be among the coolest of summers the future residents of Europe will thereafter ever experience.
The figure below really says it all. Such is the nature of coping with a ‘gradually warming’ climate…