The dimension and environmental impact of the dust storm phenomenon has been realized only recently. There are few previous works on the climatic control of dust storm frequency, especially in the field of time series analysis. Actually, there is a poor correspondence of climatic parameter mean values and dust storm data. Although generally negative, rainfall — dust-storm correlations do not reveal the physical causes of dust storm generation, as is shown by lagged cross-correlation and spectral analysis. Temperature data may indicate seasonal variation of dust storms by extreme values, but they are no reliable defining factor owing to high persistence over shorter time series. The same is found for mean wind speeds, but negative interrelations with atmospheric pressure point to the importance of cyclogenesis and convective cells in dust storm generation in Asia.
Three different seasonality patterns are described: (i) a single dust storm maximum in spring typical of summer rain areas; (ii) a single storm maximum in summer in areas with bimodal (winter and monsoonal) rainfall; (iii) an extended spring and summer dust storm maximum in areas with unimodal winter rains. Additional data from northern Africa fit into this pattern.
Spectral energy of dust storm, rainfall, and temperature series for northern India and northern China (1976–1986) is mainly concentrated in the wavelength of seasonal variations, but some supra-seasonal signals indicate fluctuations between 3·6 and 5·5 years with various phase dislocations. In China, this interannual variability is supposed to be linked with dynamic shifts in circumpolar vortex dynamics.