2020 年 58 巻 4 号 p. 123-130
A combination of physical and chemical properties defines the rice quality. During grain filling, kernel location in the panicle and ambient air temperature influence its physicochemical properties thereby affecting the quality of rice. Here, we implemented principal component analysis (PCA) to reduce the dimensionality of 10 characteristics of 719 harvested rice samples, comprised of 10 varieties produced from 2010 to 2017 in regions of Hokkaido, Japan, into latent variables PC-1 and PC-2, which explained 69.7％ of the total variance. The model increased the interpretability of the relationship among physicochemical properties of brown rice and temperatures during grain filling in a visually intuitive manner. Results showed that high temperatures in August lowered the quality of samples produced in some regions in 2013 and 2016 by increasing the percentage of immature kernels and protein. Also, high temperatures in September increased the quality of samples produced in some regions in 2011 and 2012 by increasing the percentage of mature kernels and decreasing amylose and protein levels. Moreover, in 2015 and 2017, low temperatures in August and September increased amylose levels, which negatively affected rice quality attributes. This information could contribute to the production of high-quality and palatable rice demanded by Japanese consumers.
