Some issues are so large that they are particularly difficult to see. The perfect example of such a problem is climate change.
The fundamentals are straightforward: the climate is warming as a result of the usage of fossil fuels. However, since the nitty-gritty is so broad and intricate, our knowledge of it is constantly changing. Humans are evolving at such a breakneck pace that it's nearly difficult to stay up.
"We estimate that the number of studies relevant to observed climate impacts published per year has increased by more than two orders of magnitude since the first assessment report (AR) of the Intergovernmental Panel on Climate Change (IPCC) in 1990," scientists write in a new paper led by first author and quantitative data researcher Max Callaghan of the Mercator Research Institute on Global Commons and Climate Change (MCC) in Germany.
"Manual expert evaluations are already being pushed to their limits by the exponential rise of peer-reviewed scientific papers on climate change."
Comprehensive Study
Of course, this struggle is a problem in and of itself because how can humans comprehend the problem of climate change if the magnitude of the problem exceeds our ability to analyze, measure objectively, and understand it?
Even traditional meta-analysis analyses carried out by human experts are restricted to "dozens to hundreds of papers."
Related Article: Climate Tipping Points Inevitably Leads to Dire Environmental Consequences
Using AI
One answer to the 'huge literature' problem is to use artificial intelligence (AI) rather than humans to filter through the almost infinite and ever-expanding mountain of published climate science.
Callaghan and co. used a deep-learning language analysis AI technology called BERT to discover and categorize over 100,000 scientific articles describing the consequences of climate change in their latest research - yep, another one to add to the list.
While the researchers recognize that automated studies like this are no substitute for human specialists' thorough judgments, their techniques can accomplish things that humans can't.
Sifting Through Massive Amounts of Data
In this case, that required sifting through massive amounts of data, finding a wide range of climatic impacts, mapping them out throughout the globe, and interpreting them in light of anthropogenic contributions to historical temperature and precipitation patterns.
We must be cautious, though, because machine-learning studies like these - especially at such a large scale - might contain false positives and other types of uncertainty, according to the researchers.
"While traditional evaluations can provide reasonably exact but incomplete images of the data," the researchers write, "our machine-learning-assisted technique provides a broad preliminary but a quantifiably uncertain map."
However, the AI study had already produced some alarming numbers before then.
According to the study, 80 percent of the world's surface area (excluding Antarctica) currently has temperature and/or precipitation patterns that can be ascribed at least in part to human influence on the climate - and these climatic impacts already affect 85 percent of the world's people.
Climate Change Severity
Of course, we didn't need any artificial superbrain to inform us that climate change was a major issue, but where climate impacts can and can't be readily detected - depending on where research has been spatially concentrated - is instructive.
High levels of evidence of impacts on human and environmental systems were co-located with attributed temperature or precipitation changes for almost half of the world's land (48%) - housing three-quarters (74%) of the worldwide population.
In other words, there's much overlap between effects on the natural world and study into human-caused contributions to climate change in regions like Western Europe, North America, and South and East Asia.
However, the correlations aren't as strong in other locations - perhaps because, unfortunately, there aren't enough climate scientists looking into those specific regions yet.
"The lack of evidence in individual studies is because these locations are less intensively studied, rather than because there are no impacts in these areas," the researchers argue, adding that this "attribution gap" is due to both geographic (inhospitable or sparsely populated areas) and economic considerations (low-income countries are significantly less studied).
"In the end, we hope that our global, live, automated, and multi-scale database will assist in kickstarting a slew of climate effect assessments on specific themes or geographic regions," the team concluded.
"If science develops by standing on the shoulders of giants, giants' shoulders are becoming harder to reach in an era of ever-expanding scientific literature. Our computer-assisted evidence mapping method will help you get ahead."
Also Read: Three Scientists Awarded 2021 Nobel Prize for Finding Patterns in World's Chaotic Climate
For more environmental news, don't forget to follow Nature World News!