Abstract
Monitoring combines can instantly measure the yield of rough rice and be used to develop yield maps. Yield maps can then be used to estimate nitrogen uptake in mature stages and the amount of nitrogen fertilizer required. The present study aimed to clarify the relationship between the dry yield of rough rice and nitrogen uptake and to develop a predictive formula for nitrogen uptake based on multiple regression analysis. We used a dataset obtained from long-term fertilizer experiments conducted at Daisen research station from 2008 to 2017, which included the rice varieties Akitakomachi, Kiyonishiki, and Bekoaoba. In particular, for Akitakomachi, we analyzed the effect of cropping rotation, namely a continuous paddy and a paddy-upland rotation field, on the predictive formula. The dry yield and nitrogen uptake were strongly correlated for the entire dataset (R2=0.87). The slope of the regression equation among the varieties and the intercept among cropping rotations were significantly different. Multiple regression analysis was conducted with nitrogen uptake as the response variable, and the dry yield of rough rice and meteorological factors, namely the cumulative daily mean temperature and daily radiation from heading to maturity stage, as explanatory variables. Moreover, the least AIC model contained both the dry yield of rough rice and meteorological factors as explanatory variables. However, the differences in R2 and RMSE between the least AIC model and a model containing dry yield as the only explanatory variable were no more than 0.00–0.02 and 0.01–0.07 g-N m−2, respectively. The results indicate that nitrogen uptake at the mature stage was accurately estimated based on the dry yield of rough rice alone.
