Climate change attribution – calculating the role of climate change in natural disasters
We now know with great certainty that human-caused climate change has made some extreme events more likely or more severe than they would have been in a world with no human influence. This featured as a headline statement in the IPCC Sixth Assessment Report.
How do we know this is the case, and how can we comment on a single event in particular? Through the science of event attribution.
Extreme climate events
Extreme climate events include heatwaves, marine heatwaves, intense storms, bushfires, frosts, sand and dust storms, hail, floods and droughts. Floods can be classified according to the main cause: heavy rainfall (pluvial) including flash floods, river floods (fluvial) and coastal flood. Drought can refer to hydrological drought or agricultural and ecological drought.
Extreme climate events, also known as climate hazards, can lead to disasters – depending on what is exposed to the hazard (e.g., towns, infrastructure) and how vulnerable it is to the hazard. We can decrease the impact of extreme events and lower the chance of them leading to a disaster by reducing their exposure (e.g., mitigation) to the hazard and reducing its vulnerability (e.g., adaptation).
Due to the natural variability in our climate, extremes come along from time to time – but by definition, they’re rare. Some large-scale drivers affect the odds of extreme events, for example in eastern Australia hot events are more likely during El Niño, floods are more likely during La Niña. The elephant in the room is climate change, altering the climate and changing the odds of extremes.
Was climate change to blame? Unpacking event attribution
For any event, it is important to disentangle the network of causes and influences behind the event, and carefully analyse if and how human influence affected each of them; then assess the overall effect. To do this, we employ careful analysis and modelling of the climate of the real world compared to a simulated world without human-caused climate change. The analysis can be conditional on the large-scale drivers of the event, for example by examining the chances of a record heat event during an El Niño year.
If human-caused climate change was a factor, then we can’t simply give a black-and-white statement like “climate change caused the event”. Instead, the effect must be carefully explained and put in context, in one of two ways:
- An event of the magnitude experienced was more (or less) likely than it would have been without climate change – carefully stating the geographic domain and length of time of the extreme event in question. For example, the Australia-wide summer temperatures in 2012-2013 were a record at the time and breaking this national record for the summer season was estimated to be 2.5 to 5 times more likely due to climate change.
- An event was more (or less) severe than it would have been in a world without climate change. For example, the record dry and hot October in Tasmania in 2015 was around 5-10% drier and around half a degree warmer than it would have been without climate change.
Some studies find that climate change didn’t change the odds of an event to a distinguishable degree beyond natural variability. This is found primarily around rainfall events and drought where natural variability is very high, and a human influence or ‘signal’ doesn’t ‘emerge’ above this variability.
However, many prominent studies have found a clear and attributable influence of human-caused climate change in increasing the odds of extreme events. Studies (that can be explored via this database) include heat extremes for Australia as well as in specific cities and regions; marine heatwaves in the Tasman Sea, northern Australia and on the Great Barrier Reef; and extreme fire weather events.
Learning from past events
Some examples to illustrate the range of influences and the attribution of particular climate extreme events:
- A large fire event is a function of the condition of the landscape and vegetation, the fire weather, ignitions. Climate change may affect all of these, with strong evidence that it has made dangerous fire weather more likely in many places, including during the Black Summer.
- A large river flood is a function of the prior state of the system, including soil moisture (known as ‘antecedent’ conditions), as well as the rainfall event itself. Drivers such as El Niño affect the odds of large floods. Climate change may affect the rainfall event through influencing antecedent conditions, the weather system that delivers the rain (‘dynamic’ effects) and providing a warmer atmosphere (‘thermodynamic’ effects). A human signal has not been identified distinct from natural variability in three very wet seasons and wet events assessed in Australia so far in 2011, 2012 and 2016.
- Severe spring frost events that damage crops when they are flowering are mostly ‘radiation’ frosts in Australia and are a function of the large-scale meteorological driver of a cold outbreak, as well as the local scale processes producing clear skies and a lack of wind. There is some evidence that climate change increased the cold outbreak and clear sky components of the risk of frost events. This offset the increase in average temperature – but the net effect is less clear than other types of extremes.
Explore the event attribution examples described here and more examples on Carbon Brief’s interactive map explorer.