Improved parameterization of dust emission (PM10) fluxes by the gradient method using the Naiman tower data at the Horqin desert in China.

Dust emission/deposition flux has been estimated using the gradient method with the two-level (3 and 15m high) measured PM(10) concentrations and the sonic anemometer measured momentum and kinematic heat fluxes at 8m high from a 20-m monitoring tower located at Naiman (Horqin desert) in the Asian dust source region in China for the winter of November 2007 to March 2008. The time series of measured PM(10) concentration at 3m high is used to identify the dust event and the non-dust event periods. It is found that the dust emission/deposition flux (F(C)) shows a significant diurnal variation with the maximum emission flux of 5.8 kg km(-2)h(-1) at noon and the minimum of -1.6 kg km(-2)h(-1) in the afternoon for the non-dust event cases. Whereas for the dust event cases, the dust emission flux is found to occur when the prevailing winds are westerlies to northerlies with the maximum flux of 1275 kg km(-2)d(-1), while the maximum dust deposition flux of 148 kg km(-2)d(-1) occurs with the prevailing winds of southerlies to easterlies without any diurnal variation. The optimal regression equation between F(C) and the friction velocity (u()) for the dust emission cases is found to be F(C)=9.55 u()(3.13) with the R(2) value of 0.73. However, this regression equation can be improved by taking into account the convective velocity (w()). The resulting optimal regression equation is found to be F(C)=9.3(u()-0.1w())(3.19) for the stable stratification (w()<0) with the R(2) value of 0.77 and F(C)=10.5(u(*)+0.34w(*))(4.11) for the unstable stratification (w(*)>0) with the R(2) value of 0.78, suggesting the importance of the convective velocity on the dust emission flux.