The Insurance Council of Australia increased its estimate for wildfire damage claims to over $700 million, expected to increase.
Treasurer Josh Frydenberg met with the CEO’s of insurers Insurance Australia Group, Suncorp Group and QBE Insurance Group, and local heads of Germany’s Allianz and Switzerland’s Zurich Insurance Group, plus heads of Commonwealth Bank of Australia and Westpac Banking Group.
Risk & Cause…
In this blog, we will look at the causal factors behind Australia’s potent wildfire season 2019-20, including the role of long-term climate trends. This will, in fact, lead us to uncover ways in which ocean temperature patterns influence temperature and rainfall across adjacent lands. And make for some of the most predictable elements of the global climate system…
Australia, December, 2019
In the final month of 2019, Australian wildfires were rarely far from the headlines.
Many parts of the country have seen a high incidence of damaging fires, with parts of New South Wales enduring unprecedented wildfire coverage.
There, approx. 4 million acres were burned between the 1st Jul to 31st Dec, with some 9,000 structures destroyed.
Victoria has also been hit hard, with a state of disaster declared for the worst-hit areas.
Almost the entire periphery of Australia, where populations are generally highest, has seen at least one wildfire since late Oct.
Nationally, there have been 20 directly attributed fatalities to date.
Explanations for the high wildfire activity
Explanations for the high wildfire activity tend to derive from a serious deficit in 2019’s winter-spring rainfall. Also some exceptionally hot spells of weather mid-spring through the first month of summer.
‘We will look at causal factors, including the role of climate change.’
Figure 1: Maps showing, for Sep-Dec, (left) mean maximum air temperature (°C; at 2 m) and (right) total rainfall (mm), both in comparison to the 1981-2010 average.
In Sep-Dec 2019, almost the entirety of the country saw far less than the typical rainfall. In most areas, less than half the average fell. This is a drought on a phenomenal scale on a par with the most severe in modern history anywhere on Earth.
‘This is a drought on a phenomenal scale on a par with the most severe in modern history anywhere on Earth.’
Resulting parched soils have heated faster under strong sunshine. This has contributed to some intense heatwaves. Most notably, 17th-18th Dec broke Australia’s country-average max temperature record twice. The figures of 40.9°C and 41.9°C, respectively, are hot enough that dry vegetation may spontaneously ignite.
For this blog piece, I’m taking the explanation further, to look at what’s behind the low rainfall and high temperatures.
‘For this blog piece, I’m taking the explanation further, to look at what’s behind the low rainfall and high temperatures.’
There are two key driving forces behind Australia’s weather patterns to be considered.
The first is the Indian Ocean Dipole (IOD), – a pattern of sea surface temperatures in the Indian Ocean. Once or twice a decade, the western half becomes markedly warmer than the eastern half. This makes for a positive IOD. There’s also a counterpart, a negative IOD, with a warmer eastern half.
Whichever half is cooler tends to see less rainy weather.
Sep-Dec 2019 has seen a strong positive – or warmer west – IOD. This has promoted drier weather than usual across the southern half of Australia especially. Rain-bearing cyclones in the Southern Ocean have been kept further south (Figure 2), away from the country.
‘Rain-bearing cyclones in the Southern Ocean have been kept further south (Figure 2), away from the country.’
Less rain also meant less cloud cover to hold back daytime temperatures, which have frequently soared.
Figure 2: (Left) map showing Sep-Dec mean 500 hPa geopotential height anomaly across Australia and surrounds and (right) scatter plots against the IOD index of (top) total Sep-Dec rainfall (mm) and (bottom) number of days within Sep-Dec reaching at least 35°C.
…and Atmospheric Oscillations
Large-scale difference in atmospheric pressure have been the second key driving force.
Specifically, differences in mean sea level air pressure (MSLP) between Tahiti (South Pacific) and Darwin, Australia. This is used by the Australian Bureau of Meteorology as a basis for an atmospheric index called the Southern Oscillation Index (SOI).
Unlike the IOD, the SOI hasn’t been up to much lately; values have been weakly negative (varying near -2.0 to -0.5 hPa. Very low and high readings would be near -4.0 and +4.0 hPa, respectively).
Although this suggests sea level has not had great involvement in creating wildfire conducive conditions, it does mean that it’s doing nothing to help mitigate hot, dry conditions.
Research shows (Figure 3) that a positive SOI (higher mean sea level difference) encourages wetter conditions in the southern half of Australia. (Especially (see appendix for whole-Australia plots) with cooler average temperatures – in this case, less days reaching 35°C or more during Sep-Dec). Even over a less concentrated period of time, such as weakly, a positive SOI appears to make a significant impact on rainfall.
Figure 3: Scatter plots against the SOI of (left) total Sep-Dec rainfall (mm) and (bottom) number of days within Sep-Dec reaching at least 35°C.
Considering Climate Change… Decade Comparisons
It’s difficult to attribute specific extreme events to climate change, but we can look at how the probability of such extreme events is being affected.
First, let’s look at rainfall amounts during the typically cooler, wetter half of the year in southern Australia (Apr-Oct). Here, the picture is far from clear (Figure 4).
Figure 4: Plot of decadal mean total rainfall (mm) between 01/04 and 31/10, for the 1950s through 2010s.
There was a much wetter run of years in the 1980s and 90s, compared to which 2000-2019 were relatively dry. However, look back at 1950-69 and reanalysis data indicates an even drier couple of decades.
On the other hand, data from the Australian Bureau of Meteorology (BOM), presented here, suggests the 1950s and 60s were much wetter. Which is the most accurate? That’s not an easy question to answer.
Reanalysis is a process of estimation used to provide full global coverage even where weather stations don’t exist. This is prone to error, so could be underestimating the 1950-69 rainfall. Alternatively, there might have been some very dry areas without weather stations. This would have caused the all-stations average to overestimate.
Regardless, we can at least see that rainfall hasn’t been especially high this past decade. In fact, the 2010s mean is strongly skewed upward (approx + 11 mm) by a very wet 2016 (Figure 5).
Figure 5: Plot of yearly total rainfall (mm) between 01/04 and 31/10, for 1950-2019.
Troublesome Temperatures and Vegetation Growth
Now let’s look at the frequency of very hot days (reaching at least 35°C) during peak drying-out months (Sep-Dec).
The higher the temperature, the lower humidity tends to be, increasing likelihood of tinder-dry vegetation status. In this way, very hot weather increases fire risk – even after wet winters. In fact, a wet winter may result in a larger amount of vegetation susceptible to drying out, further raising the risk.
Figure 6: Plot of decadal mean number of days within 01/09 to 31/12 reaching at least 35°C, for the 1950s through 2010s.
Clearly, very hot days have been considerably more numerous during the 21st Century so far, compared to the 1950s through 90s. An increase of two days is substantial when talking about such a large area across many years at a time.
‘Clearly, very hot days have been considerably more numerous during the 21st Century so far, compared to the 1950s through 90s. An increase of two days is substantial when talking about such a large area across many years at a time.’
Bringing It All Together
Considering all of the above, I believe it’s fair to conclude that climate change has increased the likelihood of very hot days leading to heightened wildfire risk, especially during strongly positive IOD events, or periods with a persistently negative SOI.
‘Considering all of the above, I believe it’s fair to conclude that climate change has increased the likelihood of very hot days leading to heightened wildfire risk, especially during strongly positive IOD events, or periods with a persistently negative SOI.’
It’s likely that when the IOD or SOI are supporting high rainfall and cooler temperatures, low risk wildfire seasons will still occur.
When they aren’t, however, wildfire seasons may reach unprecedented severity.
Oceans are Handy Forecasters
Large-scale ocean temperature patterns like the IOD also exist in the Atlantic and Pacific basins. They too have some significant impacts on temperature and rainfall patterns across adjacent lands.
‘It’s likely that when the IOD or SOI are supporting high rainfall and cooler temperatures, low risk wildfire seasons will still occur.
When they aren’t, however, wildfire seasons may reach unprecedented severity.’
Oceans tend to change temperature much more gradually than the atmosphere. So, these large-scale patterns are of great help for long-range forecasting. They make for some of the most predictable elements of the global climate system.
Increased wildfire risk can be foreseen weeks, even months in advance, with the right tools to hand…
Head Meteorologist at MetSwift