Modelling of the evapotranspiration process is central for efficient management of agricultural water resources in arid regions. Reference evapotranspiration (ETo) computation with the recommended Penman-Monteith (PM) equation is limited in Burkina Faso due to the numerous meteorological data required. Recently, to solve the climatic data unavailability problem, an alternative reference model (RMBF) has been developed for Burkina Faso. In a new approach, this study explores in three production regions, Dori, Bogande and Fada N'Gourma, the potential of using the feed forward backpropagation (BP) neural network algorithm for estimating ETo from temperature data. Furthermore, four temperature-based models including BP, RMBF, Hargreaves (HRG) and Blaney-Criddle (BCR) were employed and compared with the true PM. Based on the statistical evaluation, RMBF, HRG and BCR consistently overestimated the ETo and showed poor performance.Moreover, BP is superior to RMBF, HRG and BCR. Clearly, temperature-based BP is more reliable than the other alternative methods for ETo modelling in Burkina Faso. It is found that both wind velocity and relative humidity improve BP accuracy when integrated into the network input. Relative humidity does not show as strong a correlation to ETo as wind. Wind was found as the key variable of ETo and is highly recommended to be in the BP model for these arid regions of Africa under study.
All Science Journal Classification (ASJC) codes
- Agronomy and Crop Science
- Soil Science