Dust emission area is the area where Matthews vegetation type (Matthews, 1983) is assigned to desert. Most of Gobi and Takla Makan deserts are indexed as a source region. Dust roll-up scheme proposed by Gillette (1978) is used, i.e., Dust mass loading Fdust is defined as Fdust=C (U-Utr) U2 , where C is a constant, U is surface wind speed and Utr is critical wind speed of dust roll-up (Utr=6.5m/s used). Dust lift-up height (mixed layer) is determined from the vertical profile of potential temperature. RAMS build-in horizontal and vertical advection and turbulent diffusion scheme (user option parameters) are used for dust.
A very severe dust storm was observed in April 14-15, 1998 at Xinjian Province, China, and Satellite image clearly retrieved the onset of this dust front and dust cloud. Daily dust transport simulated by on-line tracer model clearly revealed that eastward travelling strong low-pressure system and cold-frontal line is responsible for this strong dust episode. Model results also clearly indicate that the high/low pressure system chains over the Pacific Ocean efficiently transport the dust cloud from west Pacific region to North America. Results show that dust emitted on April 19 arrived to west coast of United States on days around April 26. Simulated main transport path agreed with NASA/SeaWiFS image data. Numerical results are compared with the intensive lidar observation, ground base SPM (suspended particle matter) observation and high time resolution particle counter data. Time-height cross section analysis shows (i) the dense dust layer exists below 2-3 km (within PBL) and (ii) dust concentration correlated with potential temperature variation.
Model results based on the anthropogenic global SO2 emission data from GEIA are simultaneously tracked by on-line tracer model. Numerical results of SO2 and sulfate are also reported and compared with the ground base observation.