This article investigates the impacts of climate change on per hectare rice production in Korea, using a 32 year data set. Both nonparametric kernel regression and semiparametric partially linear models are estimated. The article employs several newly-developed methos that incorporate not only continuous but also (ordered or unordered) discrete explanatory variables such as regional dummies or time indices. By constructing and using regional-level data, this study incorporates much more observations than the existing works. It is shown that there is a positive relation between the average temperature and land productivity, but the impact is very small when the temperature is relatively high. There is a negative relation between the average rainfall and land productivity. It is shown that weather shocks in a harvest season have higher impacts on the productivity than those in other season.