Simulating Growth and Yield of Spring Safflower Using CROPGRO Model in Semiarid New Mexico.

Crop growth models are valuable tools for decision-makers in agriculture. The Decision Support System for Agrotechnology Transfer (DSSAT) models are generic in nature and simulate crop growth and yield based on soil, climate, crop and management information. The objective of the project was to adapt the DSSAT Cropping System Model (CSM-CROPGRO) to simulate growth and seed yield of spring safflower (Carthamus tinctorius L.). The CROPGRO module for soybean (Glycine max L.) was used as the initial reference, and parameters in species and cultivar files were replaced based on safflower literature. The entered base temperatures for photosynthetic, vegetative and reproductive processes of safflower ranged from 0 to 5 °C while corresponding optimum temperatures varied from 19 to 40 °C. Simulated results were compared with observed data collected from irrigated and water-limited treatments in field experiments conducted at Clovis, NM, USA, during 2013 and 2014 summers. The model predicted the crop life cycle (anthesis and harvest maturity date) with root mean square error (RMSE) of 5.4 days. Average plant biomass, head mass, head number and seed number were satisfactorily simulated when compared to observed values. Seed yield, averaged over irrigation treatments and years, was predicted at 1963 kg ha^-1 and was very consistent with observed value of 1902 kg ha^-1 with RMSE of 236 kg ha^-1. Reasonable prediction of phenology, growth and yield by the model adapted for safflower suggested that the CROPGRO-safflower model is promising to simulate safflower production in semi-arid climates. However, further testing of the CROPGRO-safflower model under different environments is needed.