Abstract
Dust emission estimated with a 4D-Var data assimilation system using ground-based lidar network data was compared with vegetation growth data based on visual observations in the Gobi desert in Mongolia in the spring of 2007. The dust emission flux estimated with the data assimilation system was less than that estimated without data assimilation in the dust event of May 21-30 and was the opposite in the event of March 25-April 3. The threshold surface friction velocity estimated from the results of the data assimilation was less than 0.3 m s-1 in the dust event of March 25-April 3 and was ∼0.36 m s-1 in the event of May 21-30. The difference between the two events was qualitatively explained by the vegetation growth data. The accumulated precipitation during the period was ∼2 mm. The results show that vegetation growth with slight precipitation in the Gobi desert may significantly reduce dust emission.