A Stochastic Time Series Model for Threshold Crossing Statistics of Concentration Fluctuations in non-Intermittent Plumes

A numerical stochastic model is developed for the upcrossing rate across a specified threshold concentration. The model assumes that the concentration time series at a given spatial point within a dispersing plume can be approximated as a first-order Markovian process designed to be consistent with a given time-invariant concentration probability density function (pdf). The model requires only the specification of a concentration pdf with a given mean and variance and a concentration fluctuation integral time scale. Predicted upcrossing rates are compared with atmospheric plume concentration data obtained from a point source near the ground. For this data set, a log-normal pdf is found to give better estimates of the threshold crossing rate than a gamma pdf.     

Simulation of Time Series of Concentration Fluctuations in Atmospheric Dispersion Using a Correlation-distortion Technique

Short-duration fluctuations in the concentration of airborne substances can be important in a variety of atmospheric dispersion problems, especially when assessing the risks posed by harmful materials. This paper discusses a simulation technique for generating individual realisations of fluctuating concentration time series in dispersing plumes based on target probability distributions and spectral functions. The scheme uses a correlation-distortion approach to simulate these time series. Gaussian processes with modified spectral characteristics are generated and then transformed to yield non-Gaussian processes with the desired spectral characteristics. The simulation scheme is initially developed for a single receptor, and is then extended to model pairs of correlated time series at two receptors. In fact, the modelling technique can be generalised to an arbitrary number of receptors and this provides, in principal, an approach that is applicable to a wide class of similar problems (such as the modelling of instantaneous puff releases or the response of line-of-sight detection systems). The simulation technique is illustrated using observations made during recent field experiments, conducted both in the United Kingdom and in the U.S.A., investigating the short-range dispersion of a passive tracer.     

巴州北部山区汛期降水趋势的时间序列预测法

用时间序列法分析了巴音布鲁克、巴仑台两地汛期各阶段的降水规律，并做出预报方程、进行历史检验，结果表明拟合效果较好，说明这种预测方法有一定的实用性。     

基于小波和M-K方法的商丘气温时间序列分析

利用1955-2011年商丘气象站地面逐日气象数据,集成Morlet小波分析法、Mann-Kendall分析法、回归分析法、变异系数等方法,分析商丘气温序列的年、季周期变化特征、突变特征、变化趋势特征及年际变率特征。结果表明：根据回归分析、5 a滑动平均分析,商丘年平均气温呈增加趋势,年际变化倾向率为0.17 ℃/10 a;各季节气温变化趋势差异明显,夏季气温呈下降趋势,年际变化倾向率为-0.08 ℃/10 a,冬春秋季气温呈增加趋势,冬季增温最为显著,冬季气温年际变化倾向率为0.35 ℃/10 a。根据Mann-Kendall分析,年平均气温突变点为1992年,夏季气温不存在明显突变,冬季气温突变点为1986年。根据小波分析,20世纪50-80年代商丘年平均气温存在准17 a周期信号,之后该周期信号消失,80年代和90年代出现准32 a周期信号,在本研究的整个时间序列上存在准5 a和准2 a周期信号,周期信号显示未来几年可能会出现低温年。商丘年平均气温的年际变异系数为0.04,年际变率较小,年际变化较平稳。商丘作为重要的粮食生产基地,集成多种方法的气温变化特征和趋势预测分析,对该区域的农业种植活动具有重要的指导意义。     

大气气溶胶光学厚度与大气污染物及气象因素关系的时间序列研究

为探讨大气气溶胶光学厚度(Aerosol Optical Depth,AOD)与大气污染物及气象因素的关系,进而为AOD的科学预测提供参考,应用时间序列研究分析了2002年1月～2004年12月的杭州市大气AOD与污染物及气象因素数据,并以前35个月的数据建立时间序列ARIMA(Autoregressive Integ...     

诊断天气气候时间序列极值特征的一种新方法

将平稳过程的交叉理论用于天气气候极值分析,提出了一种诊断天气气候时间序列极值特征量的新方法,在正态假设下,推证出天气气候记录中,极值出现频数、持续时间和等待时间的估计公式,论证了极值出现频数与其频谱结构的对应关系及其相互推算方法.实例应用表明,其理论计算值与实测值相当一致,这种方法对于气候变化诊断与预测和天气预报具有很强的实用价值.