There are so many studies of climate change and its consequences around the world that human experts are reaching their limits when it comes to sifting through the results. Researchers have now trained AI to identify, evaluate, and summarize scientific publications about climate change and its consequences. With the help of machine learning, based on more than 100,000 studies, they show that the effects of man-made climate change can already be demonstrated on 80 percent of the Earth’s surface. So 85 percent of the world’s population is affected.
Climate change is causing rising temperatures, uneven distribution of precipitation, increased frequency of extreme weather events, melting glaciers, and rising global sea levels. Countless studies deal with different impacts on the environment and people, partly at the local and regional level, and partly at the global level. The Intergovernmental Panel on Climate Change evaluates many of these studies for its reports. But given the vast amount of scientific evidence, it is almost impossible for human experts to get an overview.
Effects on the environment and society
A team led by Max Callahan of the Mercator Institute for Research on the Global Commons and Climate Change in Berlin evaluated more than 100,000 studies with the help of machine learning and created a map of the world where relevant evidence of location-specific impacts of climate change is. registered. “We conclude that the effects of man-made climate change can be found in 80 percent of the Earth’s land area, where 85 percent of the population lives,” the researchers wrote.
“These effects appear in a number of different systems and at different scales, covering a wide range of areas of research, from glaciology to agricultural engineering and from marine biology to migration and conflict research.” It was only able to include a limited number of studies on this topic. On the other hand, Callaghan and colleagues’ approach has allowed a comprehensive view of the studies available to date.
AI helps with evaluation
To do this, the researchers used an artificial intelligence capable of evaluating natural language. “We have developed this software further so that it can identify, localize and categorize studies of observed climate impacts – on a scale beyond what is manually possible,” explain Callaghan and colleagues. Using this technique, they found more than 102,000 publications relevant to understanding the effects of climate change.
Artificial intelligence scored, among other things, the region relevant results for related studies. The researchers compared this to more data on trends in temperature and precipitation, which can be traced back to human activities with a high degree of probability. In this way, they created a map of the world showing areas where the effects of man-made climate change have already been documented. The result: “The majority of the world’s population lives in regions where trends in temperature and/or precipitation can be at least partially attributable to human influence,” according to the authors.
Data gaps in poor countries
The researchers also found that related publications were unevenly distributed across different regions. “Our results reveal a significant gap,” they wrote. “Strong evidence of the potential effects of man-made climate change is twice as common in high-income countries as in low-income countries.” They found the strongest evidence for Western Europe, North America, and South and East Asia. In contrast, the evidence was lower for Africa and South America in particular. “This imbalance suggests that the lack of evidence in individual studies is due to the fact that these sites are studied less intensively, rather than because there is no effect in these areas,” the researchers said.
According to the authors, their approach is well suited to summarize previous scientific evidence and clarify potential research gaps. “In general, it should be noted that this type of automated assessment process cannot replace rigorous assessment by experts,” they wrote. “However, the method can identify a large number of studies that could indicate a man-made impact on the impacts of climate change.”
Cowell: Max Callahan (Mercator Institute for Research on Global Commons and Climate Change, Berlin) et al., Natural Climate Change, doi: 10.1038/s41558-021-01168-6
“Alcohol buff. Troublemaker. Introvert. Student. Social media lover. Web ninja. Bacon fan. Reader.”