Scale Separation of Spatial Rainfall Heterogeneity Based on Three Methods

XIAO Bei,FANG Jiabei,YANG Xiuqun,REN Xuejuan
DOI: https://doi.org/10.3969/2013jms.0028
2013-01-01
Abstract:Based on CMORPH high-resolution satellite precipitation data,three methods including spatial moving average,Fourier transform and wavelet analysis are used respectively to separate the 6-a mean( 04-09) summer precipitation over China region on different scales. The characteristics and possible mechanisms of spatial precipitation distribution on five scales( more than 1 000 km,500-1 000 km, 200-500 km,64-200 km,and less than 64 km) are investigated,especially over Yangtze River Delta region. The advantages and disadvantages of the three methods on scale separation as well as their application range are also analyzed. The results show that scale separation can obtain distribution features of spatial rainfall heterogeneity,based on physical reasoning,which reflect the influences of large-scale landsea distribution,monsoon circulation,topography,oceanic current,city and city groups. The spatial moving average method is easier to operate,but is relatively less precise. Fourier transform can support full spectrum decomposition,but it strongly depends on the stationarity of the signal and generates considerable deviation for small-scale disturbance. While wavelet analysis can deal with nonstationary signal and has the ability to distinguish details.
What problem does this paper attempt to address?