We can no longer rely on historical data to predict extreme weather

A satellite image of Hurricane Harvey on August 24

A satellite image of Hurricane Harvey on August 24 (NASA / NOAA /)

Floods and other dangerous weather events are becoming more intense and frequent as our climate warms. Historically, we have always been able to predict these ends by looking at how often they have happened in the past. But a new study published on Wednesday in 2004 Science sciences reveals how many of these forecasts are missing. It is only a decade since, according to the findings, that the climate has shifted so dramatically that the frequency of extreme events in the past is no longer a reliable predictor.

These forecasts help us design flood maps and design infrastructures to withstand even severe events. But if our forecasts are wrong, it means we can no longer design new homes, roads and bridges based on the storms of the past. Raising limbs – such as tropical cyclones, heat waves and heavy thunderstorms – will force us to change our designs and structures that can withstand these changes.

It is difficult to understand the effect of human overheating on extreme events. The atmosphere is chaotic by nature and extreme accuracy is by definition rare, giving scientists little information to understand. Noah Diffenbaugh, a field scientist at Stanford University, and a team of climate scientists have described a history of extreme hot, humid and dry weather conditions from 1961 to 2005 in a climate forecast. This framework incorporates both historical event-based forecasts and climate models, which incorporate forecasts of future temperature increases into their estimates.

Over the next decade, however, people continued to burn fossil fuels, and the record-setting weather conditions were whipping up areas around the world. Seven of the 10 hottest years on record between 2006 and 2017 and huge storms, like Hurricane Harvey in 2017, have caused more destruction than ever.

Given the extra heat, Diffenbaugh wanted to check how well historical data could predict recent extreme events. He used data from 1961 to 2005 to develop probabilities of hot, humid and dry extremes in the northern hemisphere between 2006 and 2017. Separately, Diffenbaugh also used climate models to compare real-world record-breaking waves, thunderstorms and storms. past frequency, as well as projections based on climate models.

Diffenbaugh's forecasts based on historical data have not been good for this decade. The results underestimate extreme events, especially hot and humid. Compared to historical data projection, real extreme hot days have increased by at least 50% in Europe and East Asia. And observed extremes were also 50% more frequent in the United States and Europe than historically predicted. "I was very surprised," says Diffenbaugh. "I had a sinking that the framework my team had developed in recent years had some flaws."

But that wasn't necessarily the problem. The context is good when we foresee extreme events that occurred in the last part of the 20th century. But in the last decade, the extra growth we have created is so significant that extreme weather patterns are different from those of the past. Meanwhile, the Diffenbaugh climate models tested were able to accurately predict the frequency of events that set the record between 2006 and 2017. "Climate models for the near future include what really happened," says Diffenbaugh. "Even if they were future forecasts at that time."

"Noah Diffenbaugh's paper is innovative and combines both models and observations to prove that the probabilities of extremities change rapidly due to global warming," says Erich Fischer, a clinical scientist at ETH Zurich who did not participate in the study. "The document has implications for risk management."

Diffenbaugh's findings have a huge impact on designing new climate change infrastructures and updating existing structures. It seems that now we cannot calculate the probability of a 500-year flood event based solely on past floods, so we have tended to plan the risks in the past. These findings suggest that we need to use a combination of historical information and climate models to map the future in the context of climate change.

This is because climate models are good at predicting changes in large areas, it is not going to tell a planner how often a particular river that passes through their city can flood. These local events are just too difficult to predict at the moment. Some states are already trying to cope. In California, the 2016 Climate Safe Bill was adopted to develop a design process for new roads, bridges and other structures so that it can cope with a changing climate. Diffenbaugh is part of the working group created by the bill. He says we need to use the results of climate models along with local historical data to better prepare ourselves for the future. Fischer adds: "The climate and, consequently, the probability of an extreme happening today are different than 10 or 20 years ago." To prepare, we must use past observations and future forecasts to literally keep our cities above water.