Stochastic Modelling of Monthly Sun Bright in Coffe Growing Areas

Bernardo Chaves

 

 

Abstract

Autorregressive integrated moving average models ARIMA, were adjusted to series of monthly sun bright for 32 meteorological stations of The National Federation of Coffee Growers of Colombia. The structure of the adjusted models was ARIMA (0,1,1)*(0,1,1)₁₂ this is a moving average with a seasonality component each 12 month, the estimated parameters were sufficient to describe the behavior of the series, they were statistically different from zero and non correlated. The estimated forecasts were found very approximated to observed values, they are actualized monthly, this characteristic allow to readjust the model when the pattern series change and to plan activities related with absorption of solar energy. The greatest forecast error was 23% and it is considered acceptable.

 

Palabras clave:  Stochastic model, ARIMA, Time series, Sun Bright.

 

PDF  (English)