Abstract
Greenhouse experiments with controlled irrigation were conducted to study the response of rice (Oryza sativa L.) cultivars IR20, IR46, and IR72 to water deficit during three growth stages, vegetative, reproductive (from panicle initiation to panicle emergence), and grain set and filling. Twelve water production functions, including seven that did not consider growth stage during water deficit and five that did, were tested to quantify the relationship between rice grain yield and evapotranspiration (ET). Values of R2 for water production functions relating grain yield of rice and total season ET were all significant (P < 0.05 or P < 0.01) and ranged from 0.48 to 0.61 for cv. IR20, from 0.79 to 0.86 for cv. IR46, and from 0.75 to 0.82 for cv. IR72. Values of R2 for water production functions that considered growth stage during water deficit were all significant (P < 0.05 or P < 0.01) and ranged from 0.62 to 0.85 for cv. IR20, from 0.82 to 0.90 for cv. IR46, and from 0.71 to 0.94 for cv. IR72. Rice was most susceptible to yield reduction by water deficit during grain set and filling and least susceptible to yield reduction by drought during vegetative growth. Yields were lowest and there were no significant differences among cultivars when water deficit was imposed during grain set and filling. Among cultivars, IR72 produced the highest grain yield under well-watered conditions and under vegetative and reproductive stage stresses, and thus, had the greatest drought resistance among the cultivars. Production functions that considered growth stage during water deficit generally explained observed yield differences better than did those functions that considered only total season ET.
