Adapting the CROPGRO model to simulate chia growth and yield

Chia (Salvia hispanica L.) seeds are becoming increasingly popular as a superfood in Europe. However, broad experience in growing chia in temperate climates is missing. Crop simulation models can be helpful tools for management and decision-making in crop production systems in different regions. The objective of this study was to adapt the CROPGRO model for simulating growth and yield of chia. Data sets from a field experiment conducted over 2 yr in southwestern Germany (48°74′ N, 08°92′ E, 475 m above sea level) were used for model adaptation. The initial starting point was the CROPGRO–soybean [Glycine max (L.) Merr.] model as a template for parameterizing temperature functions and setting tissue composition. Considerable iterations were made in optimizing growth, development, and photosynthesis parameters. After model calibration, the simulation of leaf area index (LAI) was reasonable for both years, slightly over-predicting LAI with an average d-statistic of 0.95 and root mean square error (RMSE) of 0.53. Simulations of final leaf number were close to the observed data with d-statistic of 0.98 and RMSE of 1.36. Simulations were acceptable for total biomass (d-statistic of 0.93), leaf (d-statistic of 0.94), stem (d-statistic of 0.94), pod mass (d-statistic of 0.89), and seed yield (d-statistic of 0.88). Pod harvest index (HI) showed good model performance (d-statistic of 0.96 and RMSE of 0.08). Overall, the model adaptation resulted in a preliminarily adapted model with realistically simulated crop growth variables. Researchers can use the developed chia model to extend knowledge on the eco-physiology of chia and to improve its production and adaption to other regions.