Assessment of AquaCrop model for simulating growth and yield of saffron (Crocus sativus L.)

- AquaCrop was calibrated and validated for saffron yield prediction as perennial crop.
- AquaCrop predicted saffron yield and biomass with good and acceptable accuracy.
- AqauCrop predicted the effects of irrigation regimes on saffron yield accurately.

AquaCrop is among the crop growth models that simulates the effect of water stress on yield and biomass of field crops. AquaCrop model has been parameterized for the majority of field crops; however, it still lacks the parameterization of the saffron as a perennial crop. Since saffron growth lasts for several years, therefore, simulating saffron yield and biomass is much more complex than other crops. In this study, the AquaCrop model was calibrated and validated based on the dataset of a 6-year experiment in the semi-arid area of Iran. The data consisted of different combinations of irrigation managements including full irrigation, different levels of deficit irrigation, and rainfed (no irrigation). Primary analysis revealed that unlike former studies on annual crops, it was needed to define a set of normalized water productivity and harvest index for each year of experiment. Therefore a single crop file was defined for each year of simulation. This was due to the fact that saffron yield of each year is a function of corms growth in the preceding year. Results showed that the model could simulate soil water content well in different soil depth with overall normalized root mean square error (NRMSE) = 14%. The model also simulates the potential evapotranspiration fairly with NRMSE = 27%. However, the model could simulate the saffron yield satisfactorily with overall NRMSE = 11%. In addition the NRMSE = 25% was obtained for biomass simulation pooled over the six years. Results suggested that the AquaCrop may be used for simulating the saffron yield for the whole growing seasons, while it was more accurate for simulating biomass for the earlier years and lost its precision in the last years of growth. Finally, it was concluded that the AquaCrop could be also used for simulating perennial crops satisfactorily provided that the model parameterized carefully and calibrated well.