The scientists compared two methods to fix errors in predicting summer rainfall in South Korea. They used a computer model and data from ERA5 to test the methods. The corrected models showed less error in predicting rainfall compared to the uncorrected model. One method was better at predicting average rainfall, while the other was better at showing yearly rainfall patterns. However, the method using machine learning struggled with very heavy rainfall.