| description abstract | AbstractDiurnal rainfall variation plays an important role in the characterization of regional climate. Most existing research used predefined region boundaries to study the spatial differences of diurnal rainfall variation, which depends on experiential knowledge and is somewhat subjective. In this study, the k-means clustering algorithm was used to mine the spatial distribution of diurnal rainfall variation from gridded precipitation data over China. First, clustering was conducted according to the hourly rainfall frequency at each grid cell. A cluster number large enough to find the main types of diurnal rainfall variation was used. Then similar clusters were merged according to the peak time and amplitude of the diurnal rainfall variation. Each merged cluster corresponds to one type of diurnal variation, and the locations of grid cells in each merged cluster form the spatial distribution. Thus, the classification maps of diurnal rainfall variation can be obtained. From the spatial distribution maps, the conclusions of existing research (e.g., the distribution of well-known nocturnal rainfall in southwestern China, the prevailing afternoon rainfall regions) were confirmed. Some new findings were found, such as the spatial patterns of nocturnal rainfall regions along slopes of the macroterrain and the different types of diurnal rainfall variation in the North China Plain. The diurnal rainfall variation in some regions also shows a close relationship with land cover. The results of this study can provide valuable information for further mechanism studies, and the proposed approach can serve as a useful tool for studies on diurnal rainfall variation in other regions. | |
