So the record-breaking production at the mountain tops has implications for the entire carbon cycle by which the Earth's crust pushes up to form mountains, crumbles, washes with rivers and rainwater to the sea, and eventually settles to the bottom to form new rock.
"This work takes the trend between soil production rates and chemical weathering rates and extends it to much higher values than had ever been previously observed," Mr Larsen said.
Mr Larsen described the study site in the Southern Alps as "an extremely rugged mountain range", with rainfall of 10 metres a year and slopes of about 35 degrees.
To collect samples, he and co-author Andre Eger, then a graduate student at Lincoln University, were dropped from a helicopter onto remote mountain tops above the tree line.
They then hiked down to an appropriate test site, collected samples and took them back up to their base camp.
The pair stayed at each of the mountaintop sites for about three days.
"I've worked in a lot of places - this was the most challenging fieldwork I've done," he said.
Researchers later brought soil samples back to the University of Washington and measured the amount of Beryllium-10, an isotope that forms only at the Earth's surface by exposure to cosmic rays.
Those measurements showed soil production rates on the ridge tops ranging from 0.1 to 2.5 mm per year, and decrease exponentially with increasing soil thickness.
The peak rate was more than twice the proposed speed limit for soil production, in which geologists wondered if in places where soil was lost very quickly, the soil production couldn't keep up.
In earlier work, Mr Larsen had noticed vegetation on very steep slopes and so proposed the project to measure soil production rates at some of the steepest, wettest locations on the planet.
The new results show that soil production and weathering rates continue to increase as the landscape gets steeper and erodes faster, and suggest that other very steep locations such as the Himalayas and the mountains in Taiwan may also have very fast soil formation.
"A couple of millimetres a year sounds pretty slow to anybody but a geologist," said co-author David Montgomery, a University of Washington professor of Earth and space sciences.
"Isaac measured two millimetres of soil production a year, so it would take just a dozen years to make an inch of soil.
"That's shockingly fast for a geologist, because the conventional wisdom is it takes centuries."
The researchers believe plant roots may be responsible.
The mountain landscape was covered with low, dense vegetation, and the roots of those plants reached into cracks in the rocks, helping break them apart and expose them to rainwater and chemical weathering.
"This opens up new questions about how soil production might happen in other locations, climates and environments."