Modeling reference evapotranspiration by generalized regression neural network in semiarid zone of Africa

  • Authors:

  • Seydou Traore;Yu-Min Wang;Tienfuan Kerh

  • Affiliations:

  • Department of Tropical Agriculture and International Cooperation, National Pingtung University of Science and Technology, Neipu Hsiang, Taiwan;Department of Civil Engineering, National Pingtung University of Science and Technology, Neipu Hsiang, Taiwan;Department of Civil Engineering, National Pingtung University of Science and Technology, Neipu Hsiang, Taiwan

  • Venue:
  • WSEAS Transactions on Information Science and Applications
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper investigates for the first time in Burkina Faso, the potential of using an artificial neural network (ANN) for reference evapotranspiration (ETo) estimation. The ANN algorithm generalized regression neural network (GRNN) was selected for its ability to model the ETo from minimum climatic data. The irrigation management in Burkina Faso still faced to climatic data unavailability for estimating the ETo with the recommended Penman-Monteith (PM) equation. Recently, to overcome the climatic data unavailability difficulty, a reference model for Burkina Faso (RMBF) using only temperature as input has been developed for irrigation management purpose in two production sites, Banfora and Ouagadougou. In this study, four alternative methods to PM, including GRNN, RMBF, Hargreaves (HRG) and Blaney-Criddle (BCR) were employed to study their performances in three production sites, Dori, Bogande and Fada N'gourma. The minimum climatic data were set to the maximum and minimum air temperature as input variables collected from 1996 to 2006. The results revealed that, RMBF, HRG and BCR overestimated the ETo and showed poor performance. In addition, GRNN performance was higher than RMBF, HRG and BCR. Finally, wind has been determined as a sensitive parameter in the ETo estimation for the areas studied. Obviously, using GRNN with minimum climatic variables for ETo estimation is more reliable than the other alternative methods. It is possible to estimate ETo by using ANN in semiarid zone of Africa.