Pinning down blame for complex weather events isn't straightforward. But cutting-edge science is rapidly shrinking the space to argue that the crazy weather we're experiencing isn't due to greenhouse gas emissions.
Around the world, ever more of us seem to be experiencing freak storms, floods and droughts — from catastrophes that devastate whole regions, to local heat waves and floods that leave us thinking: Surely it didn't used to be like this?
Since the early 1990s, scientists have been able to clearly show a rise in the average global temperature due to human-caused greenhouse gas emissions.
With the Earth getting hotter, heat waves become more intense and more frequent. High temperatures evaporate more water, so average global rainfall increases too. With more energy and water vapor in the system, circulation patterns change how weather systems develop, causing severe storms in some regions and drought in others.
Pinning the blame for any given weather event — of even global weather trends — has long been controversial. Yet scientists are getting closer to establishing the link.
A global trend?
In March, the European Academies' Science Advisory Council published a report urging the European Union to up its climate adaptation efforts.
Scientists used data from insurance company Munich Re to show that storms and events such as floods and droughts — and the costs resulting from them — have been on the rise since 1980.
The report's author, Michael Norton, admits the data is not perfect. He points out that more events get reported now, so older figures may be low. Also, the assets at risk will have changed — for example through increased building on floodplains.
But the trend of extreme events happening more frequently as greenhouse gas emissions accumulate in our atmosphere "is pretty obvious."
This becomes clear from comparing their frequency with other natural disasters, such as earthquakes, which remains stable.
"The data from the insurance industry shows that there is significant increase in the number of claims of climate-related events — whereas there is no change in the number of claims to those which are related to geophysical things, which are not related to climate," Norton told DW.
Still, plotting global trends depends on data that may have been gathered differently in different parts of the world, and not necessarily with climate change in mind.
And it gets shakier the further back you go. "Many early disasters would not have been recorded, and details have been lost," Gavin Schmidt a NASA Goddard climate modeler told DW.
More complexity — more uncertainty
And because weather systems are extremely complex, it's not easy to prove a correlation.
"We can lump together all of the heat-wave data from around the world, and we can say with confidence that the number of heat waves and their intensity is increasing over time," Schmidt said.
Rainfall intensity, too, is increasing at the planetary scale.
But the more complex the weather event in question, and the greater number of variables involved, the harder it is to isolate the role of climate change.
Many conditions are involved in the development of mega-storms, from water temperature to air pressure. This makes it complicated to pin down a connection.
Also how people interact with their environment plays a major role in some emanations of extreme weather.
Higher global temperatures can play a role in droughts, but land-use change tends to be a far bigger factor.
And the frequency and intensity of floods tends to be subject to how river systems and flood defenses are managed.
"You can detect a trend from looking at the observations," Friederike Otto of the Oxford University's Oxford Environmental Change Institute told DW, "but you don't know what's causing the trend without doing the attribution."
Advanced computing to attribute individual events
When Hurricane Harvey battered Texas, members of the Trump administration accused the left of "exploiting" the storm to "advance their political agenda."
Hurricane Harvey caused $125 billion in damage (second only to Hurricane Katrina in the US, which cost $161 billion)
But within weeks, a report came drawing from complex, state-of-the-art climate models to show that climate change made Harvey three times as likely to happen.
It was authored by the World Weather Attribution (WWA) project, which calculates the increased likelihood for individual extreme weather events having happened, compared to a human-caused-emissions-free world.
"Looking at the rainfall associated with Hurricane Harvey is something that would not have been possible a few years ago," Otto, one of the report's authors, told DW.
That's partly because scientists have only recently had at their fingertips the computational power for such complex modeling.
Variables such as solar radiation and the amount of aerosols in the atmosphere — as well as greenhouse gases — create the climatic conditions under which many thousands of weather events are possible. It's then down to chance exactly what weather we get (ever heard of the butterfly effect?).
The butterfly effect is called so because it's imagined that a butterfly flapping its wings can cause a typhoon
Putting a figure on the risk of a single event happening compared to the thousands of other possible outcomes is only now possible because scientists can run models innumerable times to see how frequently an event arises.
Another reason it's now possible to better isolate the role played by emissions it that their effect is becoming more significant.
Read more: 2017: The year climate change hit
Extreme weather: more impact, more likely
"The climate is changing, so the signal that you're looking for is also increasing," Schmidt said. "We're seeing continued warming, and so the impacts of that are being felt more clearly throughout the system."
Schmidt says there are limits to what weather attribution can tell us, with many events being impossible to attribute because the models don't have enough fine spatial detail.
Heat waves remain easier to pin down than more complex weather events — and have revealed some startling results.
A WWA study found that southern Europe's unusually hot summer last year was 10 times more likely as a result of climate change, and the intense August heat wave dubbed Lucifer made four times more probable.
"The larger-scale it is, the easier it is to attribute and predict; the smaller-scale, the harder," Schmidt said. "And the more it's related directly to temperature the easier; the more it's related to water, the harder," he continued.
But the science continues getting better — models are becoming more complex and more fine-grained.
As this happens, attribution studies are getting churned out faster.
Claudia Tebaldi, a climate statistician at the United States National Center for Atmospheric Research, says when the field of weather attribution emerged about a decade ago, it took a couple of years to produce a study.
Now, simpler events can me modeled in just a few days — and that's key to informing public debate.
"There are some big events that will be talked about for a long time — like Harvey, because people are still suffering the consequences."
"But for smaller events, it's important to have something to say as (or right after) the event unfolds, because the news cycle moves on," Tebaldi told DW.