Masashi Okada and colleagues from the University of Tsukuba and Japan's National Institute for Agro-Environmental Studies have developed a model that predicts rice-quality response to changes in temperature and radiation levels. Declines in rice quality have been observed in western Japan, especially Kyushu, since the 1990s. The researchers' model suggests that rice quality will decrease even further if management practices remain the same.

"Policy-makers need to support and encourage the further development of high-temperature-tolerant rice cultivars and management practices so that rice growers can adapt to climate change," Okada told environmentalresearchweb.

The main reason for the decline in rice quality is the occurrence of chalky grains, especially milky-white grains. Chalky grains sharply increase when the mean daily minimum temperature for the 20 days after heading (ear emergence) exceeds 22 °C. The underlying mechanisms for the occurrence of chalky grains in rice plants are: reduced carbohydrates in the plant associated with an increased night-time respiration rate; reduced capacity of stems and leaves for assimilation; insufficient solar radiation during the ripening period; and hits of typhoons during the ripening period.

Okada and his colleagues developed a statistical model that has a medium level of complexity to predict rice quality at broad spatial scales – the model is less complex than field-scale process-based models but more complex than simple regression models.

The model accurately reproduced the temporal trend and interannual variation in observed rice quality. However, the model was inaccurate in the occasional years in which non-climatic factors – such as cultivar and management – dominated the quality results.

"We found that the decreasing rice quality is due to the combination of high-temperature stress as a result of night-time temperature rise and the shortening of the length of the ripening period due to a mean temperature rise," Okada said. "Because rice quality is strongly dependent on local management practices and rice cultivars, it is not appropriate to use our model for other areas of the world. However, the model could be applied to other regions if locally calibrated model parameters from observational data were used."