Can we attribute a specific drought to climate change?

Discussing the role of climate change in contemporary weather extremes has led to flippancy, defensiveness (Malcolm Turnbull could barely broach the subject of rainfall variability without getting warned by the Nationals), or, for less scientifically minded former prime ministers, comparisons to animal sacrifice.

But while Australia largely ignores the impact of climate change, recently embarrassing itself at the Pacific Islands Forum, the science behind attribution has become increasingly sophisticated over the past decade and, with increasingly stark data, relatively clear cut.

Studies as far back as the early 2000s found links between climate change and the likelihood of certain weather events, notably the 2003 European heatwave, but leader of the Earth System and Climate Change (ESCC) Hub, David Karoly, says there was still a “science opinion on this ten years ago, [that] scientists couldn’t identify a clear human contribution to extreme events [and] couldn’t separate it from natural variability”.

“And so some people were, and still are, of the opinion that for all extreme events, scientists are unable to identify a human-caused contribution,” Karoly said. “That is no longer correct.”

Both the ESCC Hub and Climate Council have published recent collections of local attributions studies, while the American Meteorological Society offers a global compilation in their 2016 report “Explaining Extreme Events from a Climate Perspective”. The report ran parallel to their annual State of the Climate research and outlines both clear-cut cases (i.e. jumps in liability for coral bleaching events) as well as more complex areas of research (frost risk over south-west Western Australia).

While denialism and disinterest clearly play a strong part in a political discourse, typified by Australia’s agriculture minister declaring he doesn’t “give a rats” if the drought is “man-made or not”, there’s also more understandable confusion around attribution.

Significantly, Karoly says there’s the difference between “amplitude” and “frequency” of extreme events. This means that while disasters like bushfires might not be that much more intense compared to historic records, they are more frequent; for cyclones, it’s the reverse.

And there are other complicating factors: the exponentiality of the impact of just a degree or two increase in global temperature; the range of different methodologies used in attribution; and even simple linguistic challenges.

Science, it turns out, is hard.

“[But] if you ask an appropriate question: not ‘was this event due to climate change or was it due to natural variability?’, but ‘what is the relative contribution to the frequency of this event, or the amplitude of this event?’, scientists can make relatively clear, high confidence answers.”

Events with the largest long-term trends, such as heatwaves, sea levels, bushfires and coral bleaching, are often easier to attribute to global warming. The process starts with observed data, such as “the substantial increase in the frequency of the number of days above 35 celsius, or the number of days above 40 celsius, in Melbourne, over the last 50 years.”

(Click to enlarge)

Because observations don’t inherently specify a cause, attribution researchers then run climate model simulations, which can involve comparing counterfactual experiments to real-world shifts.

Effectively, this involves “parallel universes … for a world in which there has been no human-caused climate change: no increase in greenhouse gasses, no warming of the ocean temperatures”. The better the long-term data, the easier it becomes to find a “climate change signal”.

On the flip-side, attribution becomes trickier when observed data involves more complicating factors (i.e. the effect of wetter atmosphere, shifting circulations and heatwave-related evaporation events on droughts) and/or is not as well-represented in global climate models (hailstorms, tropical cyclones, etc).

“So it’s much more complicated for rainfall than it is for temperature changes,” Karoly said. “With rainfall, it varies geographically, and varies time of year. We’re actually in many parts of Australia getting more summertime rainfall but declines in southern parts of Australia wintertime rainfall.”

(Click to enlarge)

This has not stopped groups such as the The World Weather Attribution Project attempting near real-time attribution data for all weather extremes. In terms of the future, there are lesser-known knock-on impacts to consider, such as airborne dust, and work on better modelling “for bushfires, for tropical cyclones, for droughts … and often that means going to higher spatial resolution”.

But responsibility for the actual next stage of attribution research, adaptation and preventative strategies, lies not just with subsequent research agencies but a political and media class willing to listen and, as flippant as it might sound, even defer to the experts.