CARICE - A RICE MODEL FOR SCHEDULING AND EVALUATING MANAGEMENT ACTIONS

Crop growth modeling can be useful in crop management, but previous ri ce crop models lack the morphologically accurate phenological detail n eeded for practical application. We designed a rice crop management mo del, CARICE, for scheduling management actions on the farm and evaluat ing consequences of alternative management strategies. A simple rice c rop productivity model was expanded to include (i) phenology, based on leaf stage; (ii) tiller development, to establish yield components; ( iii) assimilate partitioning by developmental stage, (iv) assimilate p artitioning patterns of California cultivars; and (v) a direct-seeded cultural system. The key phenological events were accurately simulated : leaf stage development followed the expected pattern; predicted 50% heading was within 6 d of the actual heading dates; and grain filling predictions were within 1 to 11 d of field observations. Cultural mana gement strategies simulated included delayed planting, and managing ba rnyardgrass (Echinochloa spp.; BYG) competition by increasing water de pth and seeding rate. Simulated yield results were within the range of field observations, although total aboveground biomass was overestima ted. The model simulated a 27% loss in grain yield from a 21-d delay i n planting, which is close to the 21% loss measured in the field. At 1 1 and 54 BYG plants m-2, the model simulated 20.6 and 54.6% yield redu ctions relative to the control, as compared with 32.5 and 57.5% report ed in the literature. The model predicted the interactive effect of ri ce seeding rate and water depth management for reducing barnyardgrass competition. The morphologically accurate phenology submodel and the y ield-component approach to calculating yield are keys to using CARICE for scheduling and evaluating management actions and strategies.