资源环境科技发展态势分析平台

Entropy is a good index to measure the uncertainty of the precipitation that is a manifestation of complex interactions between water vapour transport and the local land surface processes. However, whether the uncertainty of precipitation time series is highly related to the intensity of these interactions has not been deeply considered before. Thus, sample entropy (SE), the measure of uncertainty based on self-similarity of the time series instead of a probability distribution, is employed to uncover the relationship between them. The spatial distribution of SE is based on the results calculated from 675 daily precipitation time series (1961-2011) in China. Then, the transfer of SE is extracted following the route of water vapour, water vapour source (WVS), obtained from the NCEP/NCAR reanalysis data set. Finally, the relationship between substantial changes of SE along the WVS route and the underlying conditions are discussed, including the digital elevation model (DEM), plant cover and climatic factors. The results reveal that the SE spatial distribution is closely related to the WV source and its interaction with local climatic regions and underlying conditions, which are manifested by the following aspects: (a) SE exhibits lower values in WVS source regions and increases following the WV transfer route because of diverse continental water vapour composition sources, which makes the system more complicated and increases the system entropy. (b) The pronounced SE increases and decreases indicate WV convergence and substantial topographic uplift, respectively. (c) The spatial distribution of SE is comprehensively dominated by climatic conditions since the boundaries of climatic regions correspond well with the specific SE contours. Therefore, SE spatial distribution indicates where WV originates and how WV interacts with local land surfaces along the WV transfer routes, which offers a new information mechanism perspective to understand the hydrological cycle from precipitation time series.