In the 20th century, when excessive quantities of ozone- depleting chlorinated and brominated hydrocarbons such as CFCs were released into the atmosphere, the ozone layer in the stratosphere - ie at altitudes of 15 to 50 kilometres - thinned out globally.
The Montreal Protocol introduced a ban on these long- lasting substances in 1989.
At the turn of the millennium, the loss of stratospheric ozone seemed to have stopped. Until now, experts have expected that the global ozone layer would completely recover by the middle of the century.
The study, published in the journal Atmospheric Chemistry and Physics, used satellite measurements spanning the last three decades together with advanced statistical methods.
More From This Section
Ozone is formed in the stratosphere, mainly at altitudes above 30 kilometres in the tropics. From there it is distributed around the globe by atmospheric circulation.
The scientists were somewhat surprised that the ozone is thinning out in the lower stratosphere because their models do not show this trend and CFCs continue to decline.
"Thanks to the Montreal Protocol, ozone in the upper stratosphere - ie above 30 kilometres - has increased significantly since 1998, and the stratosphere is also recovering above the polar regions," said William Ball, researcher at ETH Zurich.
Despite these increases, measurements show that the total ozone column in the atmosphere has remained constant, which experts took as a sign that ozone levels in the lower stratosphere must have declined.
The reasons for the continuing decline are still unclear. However, researchers have two possible ex-planations.
On the one hand, climate change is modifying the pattern of atmospheric circulation, moving air from the tropics faster and further in the polar direction, so that less ozone is formed.
Ozone-depleting VSLSs are partly of natural and partly of industrial origin; some are substitutes for CFCs, and although they are less ozone-depleting, they are not neutral either.
"These short-lived substances could be an insufficiently considered factor in the models," said Ball.
"The decline now observed is far less pronounced than before the Montreal Protocol. The impact of the Protocol is undisputed, as evidenced by the trend reversal in the upper stratosphere and at the poles," said Thomas Peter, from ETH Zurich.