降水气候特征的随机模拟试验

A STOCHASTIC SIMULATION TEST FOR CLIMATOLOGICAL FEATURES OF PRECIPITATION

本文利用多状态一阶Markov链,根据随机模拟理伦建立一种用以产生单站逐日降水量模拟记录的随机模式。由该模拟模式可推求降水的各种长年气候统计特征,因而这种模拟模式可为缺少逐日资料的地区提供获取足够长年代资料的有效途径。利用该模式对我国上海、北京、广州、沈阳、荆州等5站不同月份逐日降水的试验给果表明,模拟记录与历史资料所统计的气候特征如累年平均月降水量及其方差、累年平均日降水量及其方差、累年平均各月雨日数、最大日降水量等气候统计参数十分吻合,且具有相当的稳定性和可靠性。由于试验站点分属不同气候区域,因而该模拟方法具有普适性。

In this paper, a stochastic model which can provide simulativerecords of daily precipitation at any single station is constructed bythe use of the multi-state one-order Markov chain based on the theoryof random simulation. Secular climatological-statistical features of dailyprecipitation can be obtained at any single station by means of themodel. Results show that all the climatological-statistical parameters onthe simulative records are highly consistent with those on the historicalrecords for each monthly precipitation at the five representative testingstations. It has been proved that the stochastic simulation is not onlyeffective but also of universal significance.