Abstract:Spatiotemporal characteristics of dry-wet change are the key characterization of regional hydrological response under global change. To explore the temporal and spatial distribution characteristics of dry-wet index in Anhui Province under the background of global change, spatiotemporal distribution characteristics of aridity index was comprehensively investigated for 15 meteorological stations during 1957—2016 in Anhui Province. Based on the calculation of potential evapotranspiration (ET0) by the FAO 56 Penman-Monteith model with regional correction mode, the temporal and spatial distribution characteristics, uniformity and stability of the dry—wet index (AI) in Anhui Province in the past 60 years were quantitatively described by the cloud model. The AI and ET0 in Anhui Province showed a downward trend, with propensity rates of -0.006a-1 and -0583mm/a, respectively, and P showed an upward trend of 1155mm/a. The opposite trend of ET0 and P caused the AI to gradually decrease. Anhui Province generally showed a trend of becoming wet. P was the most discrete and had the worst stability compared with ET0 and AI. On the fourseason scale, summerautumn and winter AI, which was dominated by summer (-0.012a-1), was the main feature of dry-wet change in Anhui Province. The AI superentropy value with descending order was summer, autumn, spring and winter, and the uncertainty was gradually reduced. The change entropy of ET0 in the four seasons was lower than the annual average entropy. The ambiguity and randomness of the four seasons’ ET0 were poor. The winter ET0 had the greatest instability. The increase of rain and snow in summer and winter and the decrease of precipitation in spring and autumn were the four seasons’ characteristics in Anhui Province. The main form of the pattern and the summer precipitation were increased significantly (2467mm/a), while showing the greatest unevenness and instability. Spatial scale, AI, and P showed the reference crop evapotranspiration variation gradient of Wannan to Wanbei appeared nonsmooth latitudes phenomenon, the spatial region of each entropy size was higher than the super time series entropy, and the spatial distribution characteristics of AI were more discrete and unstable.