Evaluation of Stochastic Models for Estimating the Persistence Probability of Cloud-Free Lines-of-Sight.

摘要

Stochastic models, based upon the Ornstein-Uhlenbeck (O-U) class of the simple Markov process have been used effectively to estimate the conditional probability of a variety of weather events (Gringorten, 1972). This study is directed toward extension of the analytical form of the O-U Markov model to yield estimates of the joint occurrence probability and duration of sky cover and cloud free lines of sight (CFLOS). As a first step, model estimates are made of the recurrence probability of CFLOS for specific categories of sky cover in tenths as a function of time. Then model calculations are made for the persistence probability of CFLOS and sky cover as a function of sky-cover category and time using an analytical representation of the mathematic solution of persistence probability for the O-U Markov process given by Kielson and Ross (1975). The model calculations of CLFOS recurrence and persistence are compared with the Columbia, MO, data base (Lund, 1973). Again using an analytical form of the O-U Markov model, calculations are made of the joint occurrence probability of sky-cover at two or more sites and compared with the observations from the central United States. Finally, analytic techniques are presented for calculating probability estimates of the duration of cloud-free or cloudy lines of sight from one or multiple ground sites to points in space. Trial calculations based upon climatic summaries of sky-cover are made for a selected group of sites in south western United States. Keywords: Cloud-free-line-of-sight, Whole-sky imagery, Stochastic modeling, Markov process, Recurrence and persistence probability, Sky cover, Weather, CFLOS, Probability, Climate. (JG)