Bengaluru: Climate researchers in Germany have produced a comprehensive evidence map using a new machine learning technique on the impact of human-induced or anthropogenic climate change, using data from over 1,00,000 climate impact studies.
The authors used data from the studies and compared them with local data trends in surface temperature and rainfall in a region, which are attributable to humans.
Their findings show that 85 per cent of the world’s population is already living in areas that are experiencing the direct effect of climate change induced by human activity.
The study was published in the journal Nature Climate Change this week.
The authors state that their two-step approach is a new and effective machine-learning focused approach to identifying and assessing climate impact studies exploring anthropogenic emissions.
The approach further provides spatially resolved data (in terms of geography and size of the areas) in a systematic manner.
The researchers have also identified an ‘attribution gap’ where high-income countries, primarily North America and Europe, have twice the amount of evidence attributing changes in climate to human activity when compared to low-income countries.
Trends in global climate impact studies
In their review, the authors found trends among existing studies in terms of data availability. They found that a majority of studies were skewed towards studying the impact of terrestrial ecosystems and freshwater ecosystems. The least studied area was the impact on mountains, snow and ice.
To understand the impact of human-induced climate change on land systems, existing studies have examined how changes in temperature and precipitation annually affect outcomes like crop yields or rate of birth of mammals.
They also compared trends with local ecosystems. They found that 80 per cent of all land on the planet was affected by anthropogenic climate change where 85 per cent of human population resided.
There were also disparities in the quantity of studies over different geographical regions and continents. The authors found that there were more studies coming out of North America, Asia and Europe as compared to South America, Africa and Oceania.
Adjusted for population, the studies from Oceania make up the highest number, while those in Africa and Asia receive the lowest coverage per million inhabitants.
They further found that among climate and impact-related studies from Africa, there was more focus on human and human-managed systems as compared to studies from elsewhere like Europe, potentially indicating that the greater share of research reflected the higher vulnerability of the population in sub-Saharan Africa to climate impacts.