Simulating of Potato (Solanum tuberosum L.) Yield under Different Irrigation Conditions using AquaCrop and Cropsyst Models

To evaluate AquaCrop and Cropsyst models  for simulating yield and water use efficiency (WUE), this research was performed at the Research Station in ChaharTakhteh, Shahrekord, during 1998-1999, by considering water stress amount at five levels (E0, E1, E2, E3, and E4 indicating 100, 85,70, 50, and 30 percent of crop water needed according to lysimeter data, respectively) in three periods during potato production (T1, T2, and T3 indicating 50, 100, and 150 days after sowing, respectively) and 15 data in each year. First year data was used to calibrate and the second year was used to validate AquaCrop. The highest and lowest differences between observed and AquaCrop simulated yield were 3.15 (E1T2) and 0.3 (E1T3) ton.ha-1</sup>, respectively, and the highest and lowest WUE were 0.53 (E3T2) and 0.03 (E4T2) kg.m-3</sup>, respectively. The highest and lowest differences between observed and Cropsyst were 2.34 (E3T3) and 0.35 (E1T2), ton.ha-1</sup>, respectively. Corresponding results for WUE were 0.32 (E3T2) and (E1T1) kg.m-3</sup>, respectively. NRMSE results for Yield were 0.9 (AquaCrop) and 0.7 (CropSyst) for calibration periods and 0.5 (AquaCrop) and 0.9 (CropSyst) for validation periods. EF values for yield were 0.99 and 0.95 for AquaCrop and 0.90 and 0.79 for CropSyst in calibration and validation periods, respectively. Both models had good precision; however, AquaCrop had better efficiency for simulating yield. Based on results, it is recommended to use AquaCrop in low water stress (since it is water-driven model and simulates water response to water accurately) and apply Cropsyst in high water stress (since it is radiation-driven model). In addition, using Cropsyst in T2 is better than AquaCrop.