Building and Environment 42 (2007) 3493–3499 Simulating climatic variables by using stochastic approach Kadri Yu¨rekli, Hu¨seyin Simsek, Bilal Cemek à , Sedat Karaman Department of Agricultural Structures and Irrigation, Faculty of Agriculture, University of Gaziosmanpas - a, 60240 Tokat, Turkey Received 28 July 2006; received in revised form 29 August 2006; accepted 26 October 2006 Abstract In this paper, stochastic approach with Autoregressive Integrated Moving Average (ARIMA) based techniques are applied to generate solar radiation, temperature and relative humidity forecast with the purpose of being incorporated within agricultural facilities predictive control strategy. This study opens a new perspective about light on the characteristics of solar radiation, temperature and relative humidity and may help planners in agriculture and related industries. r 2006 Elsevier Ltd. All rights reserved. Keywords: Solar radiation; Temperature; Relative humidity; ARIMA 1. Introduction The basic justification for most agricultural facilities is to provide some modification of the existing natural climate. Because of important influence of climate on animal and plant production, labor efficiency, and the value of product quality, the expectation concerning with the complex and sophisticated environment of agricultural buildings has become increasingly significant. Control of the environ- mental factors as temperature, moisture, light within agricultural systems is essential for high production, maintenance of quality of stored products, disease control, building and equipment longevity, and safety against hazardous gases. The importance of each factor varies with the specific requirements of each enterprise and with the climate of the area. For example, for greenhouses, storage and livestock facilities, climate is a very complex system in which the factors highly depend on the outside climate conditions and have a considerable effect on the design of agricultural buildings and structures. It is clear that with the scientific understanding of the agricultural buildings and structures, climate mechanism has advanced by the availability of computers that allow simulating the climate by means of static and dynamic agricultural facility models. To achieve optimal agricultur- al production, it is essential to have stochastic or dynamic models to compute short and long-term predictions for outside weather. Temperature, relative humidity and solar radiation have major influences on the crop and livestock production, crop preservation and storage and these factors should be predicted for future design purposes. Most of the statistical methods used in meteorological studies are based on the assumption that the observations are independently distributed in time. The occurrence of an event is assumed to be independent of all previous events. This assumption is not always valid for meteorological time series. In general, high values tend to follow high values and low values tend to follow low values, in this sense there is a tendency for the values to cluster. Thus, meteorological observations are not independently distributed in time. The dependence among monthly observations is less than that between daily observations, and the dependence among annual observations is less than that between monthly observations. Thus, the dependence between meteorological observations decreases with an increase in the time base. The serial correlation between the values of a given time- dependent data which is expressed as stochastic process may be taken as dependence criterion. This serial correla- tion among observations is more effectively considered by autoregressive integrated moving average (ARIMA) model than others (neural network, exponential smoothing, moving average and fuzzy logic, etc.). The popularity of ARTICLE IN PRESS www.elsevier.com/locate/buildenv 0360-1323/$ - see front matter r 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.buildenv.2006.10.046 à Corresponding author. Tel.: +90 356 2521616x2276; fax:+90 356 2521488. E-mail address: bcemek@gop.edu.tr (B. Cemek).