In this paper, we propose a stochastic model of the conditional time series of the wind chill index. The model is based on the inverse distribution function method and on the normalization method for simulation of the non-Gaussian non-stationary random processes as well as on the method of conditional distributions for simulation of the conditional Gaussian processes. In the framework of the approach considered, two types of conditions (point conditions and interval conditions) are imposed on the time series. The model in question was verified using the real data collected at the weather stations located in West Siberia (Russia). It is shown that the simulated trajectories are close in their statistical properties to the real time series. The model proposed was used for stochastic forecasting of the wind chill index and the results of the numerical experiments have shown that the accuracy of the short-term forecasts is high enough.

在本文中，我们提出了风冷指数的条件时间序列的随机模型。该模型基于逆分布函数方法和用于非高斯非平稳随机过程仿真的归一化方法以及用于条件高斯过程仿真的条件分布方法。在所考虑的方法的框架中，时间序列上施加了两种类型的条件（点条件和间隔条件）。使用在西西伯利亚（俄罗斯）的气象站收集的真实数据验证了该模型。结果表明，模拟轨迹的统计特性与实时序列非常接近。