关键词: 气候变化, 日最高气温, 日最低气温, 日雨量, 随机模拟

Abstract: The changes of daily maximum, minimum temperature and daily precipitation are the basis of evaluating the climate change impacts on various extreme events in a CO₂ doubling environment. In this paper, two stochasticsimulation models are developed for daily maximum and minimum temperature and daily precipitationrespectively.Based on present correlative equations between monthly averages of mean temperature and maximum or minimum temperature and scenario predicted by UKMO, monthly averages of maximum and minimum temperature are calculated in CO₂ doubling situation. After that, A new set of annual cycle parameters are estimated while other two component's parameters are unchanged in CO₂ doubling condition.As to daily precipitation simulation, monthly one-order Markovian chain transition matrices for wet and dry transition probabilities are derived from historical daily precipitation records. Then, precipitation values aregenerated by daily precipitation distributions. To reflect the climate change scenario, mean monthly precipitationis changed according to UKMO GCM outputs in daily precipitation distributions.The 100 year daily maximum and minimum temperatures can be decomposed three components. They areseasonal cycle, a sequence of several-day-time scale waves and random fluctuation components. Seasonal cyclecan be derived with a Fourier series. When higher harmonics are excluded, seasonal cycle is represented by theannual cycle. When the seasonal cycle is removed from the temperature data, short-time scale waves are apparent.A wave is approximated by a linear rising limb and a falling limb. After the annual cycle and short-time scalewaves are removed from temperature data, the residual value is random component. Each component can begenerated stochastically by producing various new parameter according to their probability distributions whichobtained from daily maximum or minimum temperature records of many years. After three components aresummed. the daily temperatures are produced. At present, GCMs only predict with some certainty a change in themonthly mean temperatures.100 year daily maximum, minimum temperature and precipitation values of seven stations in China aregenerated stochastically in present and CO₂ doubling situations. The results have showed in CO₂ doublingenvironment daily maximum and minimum temperature will increase 5 ℃～10℃ in seven places. The extremehigh temperature events will increase greatly while extreme low temperature events decrease obviously. Dailyprecipitation will change a little.

Key words: climate change, daily maximum temperature, daily minimum temperature, daily precipitation, stochastic simulation