Abstract
A forcasting scheme for daily solar irradiance on agricultural field sis proposed by application of chaos theory to a long term observation data. It was conducted by reconstruction of phase space, attractor analysis, and Lyapunov analysis. Using the methodology , it was determined whether evolution of the five climatic data such as daily air temperature , water temperature , relative humidity, solar radiation, and wind speed are chaotic or not. The climatic data were collected for three years by an automated weather station at Hwasung-gun, Kyonggi-province. The results showed that the evolution of solar radiation was chaotic , and could be predicted. The prediction of the evolution of the solar radiation data was executed by using ' local optimal linear reconstruction ' algorithm . The RMS value of the predicting for the solar radiation evolution was 4.32 MJ/$m^2$ day. Therefore, it was feasible to predict the daily solar radiation based on the chaos theory.