气候变化中高温破纪录事件的蒙特卡罗模拟研究

采用蒙特卡罗方法，分析了南京市1961年—2000年40?a日观测温度资料中的高温破纪录事件的统计规律，并比较了20世纪全球变暖和南京市近40年来区域变暖背景对其统计规律的影响. 理论分析和蒙特卡罗模拟结果均表明：近40年来南京市第ｋ个高温破纪录事件的最概然发生强度与k（k=1,2,3,\:）呈线性增长，而年发生高温破纪录事件的频率随时间t呈1/（t+1）的衰减趋势，且平均温度高的年份，发生高温破纪录事件的概率较大，反之，概率较小. 结果还表明：20世纪的全球变暖速率（ｖ=0.006?℃/a）和南京市区域的变暖速率（ｖ=0.017?℃/a）在短期内还不至于引起高温破纪录事件的发生强度和发生频率有明显变化，但持续变暖最终将会使年发生高温破纪录事件的频率渐渐地收敛于一个常数，近似等于变暖速率的值. 此外，还研究了日温度之间的自相关和方差变化对高温破纪录事件的影响，研究发现异方差和弱的自相关对高温破纪录事件的发生强度和概率的影响基本可以忽略. 关键词： 高温破纪录事件 蒙特卡罗模拟 全球变暖

Monte Carlo simulation of the record-breaking high temperature events of climate changes

Using Monte Carlo method, the law of statistics of record breaking high temperature events has been investigated based on the statistical characteristics of the temperature data registered in 40 years from 1961 to 2000 in Nanjing area, and the influence on the law of statistics of the global warming （ｖ=0.006?℃/a） during 20th century and Nanjing regional warming （ｖ=0.017?℃/a） in last 40 years has also been compared. Both the theoretic analysis and Monte Carlo simulation results show that the most likely probability of its occurrence intensity of the kth record breaking high temperature events takes the form of linear increase with k （k=1,2,3,\:）, and the frequency of the occurrence of record breaking high temperature events in a year tends to 1/（t+1）, decreasing with time t, and in years when the average temperature is high, the probability of the occurrence of record breaking high temperature events is also high, otherwise, the probability is low. The result also indicates that the velocity of global warming （ｖ=0.006?℃/a） in 20th century and the rate of regional warming （ｖ=0.017?℃/a） of Nanjing area in last 40 years are yet insufficient to alter meaningfully the intensity and frequency of record high temperature events. However, such velocity of warming will ultimately make the frequency of the occurrence of record breaking high temperature events decrease gradually towards a constant, being approximately equal to the warming velocities. In addition, we have also studied the effect of variances and autocorrelation between temperatures in two successive days on record breaking high temperature events，finding that the effect of different variance and weak autocorrelation on the intensity and probability of occurrence of the record breaking high temperature events in a year can be neglected.
Keywords： record-breaking high temperature events Monte Carlo simulation global warming