Evaluation of Combination Methods for Garlic Evapotranspiration Estimation

Abstract

ABSTRACT-Different evapotranspiration (ET) estimation equations having different accuracy with different conditions have been developed for ET estimation. This study will firstly focus on the estimation of 13 climatic equations of daily garlic ET estimation whose ET is measured by lysimeter to provide information which can be helpful in selecting an appropriate ET equation. The paper aims at showing the potential for combining the result of the best equation to improve the overall accuracy. The findings showed that the five equations of FAO 56 Penman–Monteith, ASCE Penman–Monteith, Kimberly Penman, Penman, and FAO-24 Blaney-Criddle were the most accurateequations for estimating garlic ET. The results of these five equations were combined using the three combination methods of Simple Average Method (C-SAM), multiple linear regression (C-MLR) and Adaptive Neuro-Fuzzy Interface System (C-ANFIS).The comparison of combination methods at the test stage showed that although C-SAM used simpler equations than C-MLR but its results were more reasonable than C-MLR. Overall, the results of these two combination methods did not significantly surpass those of the best ET estimation equations (FAO 56 PM); however,C-ANFIS combination method estimated ET better than the other techniques. Based on the results of this study, the C-ANFIS combination method is recommended for estimating garlic ET.