Spatial and Temporal Distributions of Probability Classification of Precipitation and Temperature Anomalies over China

Based on the standard of the probability classification definition and scoring method in short term climate prediction operation, analysis is conducted on six-level probability classification of monthly precipitation and temperature anomalies in January and July. Spatial and temporal distributions are obtained through the monthly precipitation and temperature data at 160 stations in China, which are operationally used by National Climate Center of CMA. The six levels are defined as much more than normal (L1), moderately more than normal (L2), slightly more than normal (L3), slightly less than normal (L4), moderately less than normal (L5), much less than normal (L6).The results indicate that the issued six-level probability classification is suitable for symmetrical distribution cases for positive and negative anomalies but neglecting spatial inhomogeneous distributions and inter-decadal variations of monthly temperature and precipitation. During the period of 1980—2009, the probability of L1 and L6 for precipitation in North China is high in January whereas that of L6 and L5 is elevated in South China in July. The six-level probability for precipitation in January and July is generally similar in South China. The probability of L4, L3, and L2 temperature is high whereas that of L6, L5, and L1 is low for temperature in China in both January and July. Compared to those in the period of 1951—1979, the station numbers of L1 and L2 in January and L5 for precipitation in July have significantly increased but those of L6 precipitation in January and L6 and L4 for precipitation in July have remarkably decreased in the period of 1980—2009. Meanwhile, the station numbers of L4, L5, L6 for temperature in January have substantially decreased but those of L1, L2, L3 for temperature in January increases significantly and the six-level temperature probability in July shows no variability since 1980.The above results could provide an important reference for climate forecasters to fully consider inter-decadal, inter-annual and inter-seasonal variability. The standard of the scoring method for the climate prediction focuses on the accurate rate of classification prediction, and especially emphasizes the abnormal level of precipitation and temperature. Therefore, the scoring method will help promote climate prediction services. The six-level scoring method for precipitation is more reasonable, while for temperature the method needs appropriate improvements.