Existing models are out of date and bear little relation to the reality on the ground, according to the international team of researchers behind the studies.

“When dust emission models were developed, there were few continuously varying global data sets available and simplifying assumptions were made for their implementation,” they write in the JGR Atmospheres paper. Those simplifications included assuming that Earth’s surface has no vegetation, that the majority of dust was emitted by North Africa and the Middle East, and that there was an infinite amount of dry, loose sediment on the surface.

In reality, the researchers found, the truth is anything but. When they calculated annual calibrated dust emissions from 2001 to 2020 using daily satellite observations of dust emission sources every 500 meters across Earth, the team found much less dust is emitted on the surface than the dominant model assumes.

The new analysis in Science of the Total Environment also discovered that Earth’s main dust sources shift over the course of the year between deserts in East Asia, the Middle East and North Africa, and Australian and North American shrub lands — variations hidden by the current model.

Atmospheric dust affects both the climate and human health, even in areas far from its source. Overall, the researchers write in the JGR Atmospheres study, the old model differs from the satellite observations by up to two orders of magnitude.

“Current models have only been telling a fraction of the story,” Adrian Chappell, a professor of climate change impacts at Cardiff University’s School of Earth and Environmental Sciences and both papers’ lead author, said in a news release.

Using the old model runs the risk of delaying scientific advances and improving climate change projections, the researchers warn. They propose that the field adopt the new, more sensitive model to strengthen continued research.