2020 Volume 64 Issue 3 Pages 146-152
To provide a realistic regression equation appropriately expressing the relationship between rice yield and weather variables, multiple regression models were developed using a set of data measured during 2009-2017 at an experimental field at JICA Tsukuba Center, Japan. Correlation, stepwise method, and multiple regression analysis were applied to create regression equations predicting rice yield during a growing season. Yield of IR-28 and/or NERICA 4 was used as the dependent variable, and weather variables were used as the independent variable. A realistic criterion is that the adjusted R2 is comparatively high and the p value of significant value F is p < 0.05, and also p value for coefficient of independent variables is at the level p < 0.05. Applying the above criterion it was found that the regression equation which includes average daily temperature, total precipitation and evapotranspiration during the growing season is favorable for predicting crop yield with “IR-28”, while prediction for crop yield of “NERICA 4” is best estimated using the regression equation which includes maximum daily temperature and total precipitation during the growing season.
