Dominant variability in terrestrial hydrological processes over the Pearl River basin in South China

Terrestrial hydrological processes are simulated by using the Variable Infiltration Capacity (VIC) model from 1951 to 2000 over the Pearl River in South China. Continuous wavelet transform is performed on monthly time series of the hydrological component in 10 sub-basins. The results reveal that the dominant variability band of precipitation is both 4-8mons and 1-4 year, and the terrestrial processes modulate it into 0.83-4.5 year for the streamflow, 2-5.5 year for the soil moisture, and 1.33-5.5 year for the evapotranspiration. The usefulness of examining scale-averaged wavelet power in dominant variability band in detecting extreme events (flood/drought) is demonstrated. Furthermore, the relationship between the variability of components and basin characteristics are investigated. To capture spatial features simultaneously, coherent regions of exhibiting similar variability distribution across different timescales of each component are obtained by applying principal component analysis on time-averaged wavelet power spectrum. Through analyses of the scale-averaged wavelet power in coherent region, it is found that for 1-4 year variability of precipitation, the El Niño-southern Oscillation (ENSO) and the Indian Ocean Dipole (IOD) are shown to be responsible for eastern and central parts. In the 0.83-4.5 year band, the influences of ENSO and IOD on streamflow in eastern part are obvious. For the 2-5.5 year band, the evidences for a teleconnective influence of ENSO and IOD on the soil moisture in the basin are all found. While in 1.33-5.5 year band, the noticeable association between evapotranspiration and teleconnection patterns is prior to 1980 for the eastern part.