Simulating yield response of quinoa to water availability with aquacrop
The modeling of yield response to water is expected to play an increasingly important role in the optimization of crop water productivity (WP) in agriculture. During 3 yr (2004–2007), field experiments were conducted to assess the crop response to water stress of quinoa (Chenopodium quinoa Willd.) in the Bolivian Altiplano (4000 masl) under different watering conditions (from rain fed, RF, to full irrigation, FI). Crop physiological measurements and comparisons between simulated and observed soil water content (SWC), canopy cover (CC), biomass production, and final seed yield of a selected number of fields were used to calibrate the AquaCrop model. Subsequently, the model was validated for different locations and varieties using data from other experimental fields and from farmers' fields. Additionally, a sensitivity analysis was performed for key input variables of the parameterized model. AquaCrop simulated well the decrease of the harvest index (HI) of quinoa in response to drought during early grain filling as observed in the field. Further-on, the procedure for triggering early canopy senescence was deactivated in the model as observed in the field. Biomass WP (g m–2) decreased by 9% under fully irrigated conditions compared with RF and deficit irrigation (DI) conditions, most probably due to severe nutrient depletion. Satisfactory results were obtained for the simulation of total biomass and seed yield [validation regression R2 = 0.87 and 0.83, and Nash-Sutcliff efficiency (EF) = 0.82 and 0.79, respectively]. Sensitivity analysis demonstrated the robustness of the AquaCrop model for simulation of quinoa growth and production, although further improvements of the model for soil nutrient depletion, pests, diseases, and frost are also possible.