ABSTRACT: A stochastic climatological model is presented to describe the space-time behavior of daily global solar radiation measured by pyranometers. Due to the difficulty of describing daily global radiation by common probability distributions, a nonparametric technique, the Abramson kernel estimator, is used. The space-time climatological model of daily global radiation consists of a transformed multivariate autoregressive (AR) process conditioned on large-scale atmospheric circulation patterns. The transformation is necessary to establish a relationship between the probability distribution of global radiation and a normal variable used in the AR model. In order to check the stochastic model a simulation study was performed for 7 locations in Nebraska, USA. Although a split sampling approach could not be used due to the relatively small sample size the simulated probability distributions reproduced the empirical distributions quite correctly. Also, the space-time dependency was maintained. The methodology may be a useful tool to estimate local/regional radiation under climate change.

KEY WORDS: Global solar radiation . Probability distribution . Kernel estimator . Space-time model . Circulation patterns . Downscaling

Published in CR Vol. 7, No. 1 (1996) on August 22
ISSN: 0936-577X. Copyright © Inter-Research, Oldendorf/Luhe, 1996