Response of Rice to Solar Radiation and Temperature Estimated from International Yield Trials1

Abstract

Experimental yields from cooperative rice (Oryza sativa L.) trials of the International Rice Testing Program network are related to major weather variables during reproductive and ripening stages. Based on the grain yield data from irrigated rice variety trials conducted in 40 environments during 1976–1981, regression models were developed and tested for predicting rice yields from total solar radiation and temperature data. The test sites were from eight countries in Asia, two from Latin America and one each from Africa and Oceania. Trials were conducted under conditions of good water control, optimal fertility and adequate plant protection and on sites with no appreciable soil stresses. Estimation of standard errors and evaluation of predictive ability of the formulated models are stressed. Among the models evaluated, a prediction equation based on radiation (RAD) and minimum temperature (MINT) during the ripening stage of 30 days after flowering demonstrated predictive ability. The environments ranged from 17.4 to 29.2°C for MINT and 315 to 637 mWh cm−2 for RAD. Predicted yield increased 0.41 t ha−1 (standard error of 0.092) for an additional 100 mWh cm−2 of total solar radiation, or an increase of 0.47 t ha−1 for an additional 100 langleys. For temperature, the change in predicted yield depended on the value of MINT. The rate of change in predicted yield was estimated as −2.10 + 0.075 MINT, i.e., the predicted yield change decreased 0.075 (0.0068) t ha−1 for each degree increase in MINT. Rates of changes in predicted yield were estimated at −0.68 (0.047), − 0.38 (0.025), and − 08 (0.024) for 19, 23, and 27°C, respectively. Model parameters were estimated by weighted least squares utilizing prior knowledge that yield predictions should decrease throughout the range of MINT from 18 to 28°C. Yield predictions are graphically summarized by yield response curves and isoquant plots.