Estimating Summertime Precipitation from Himawari-8 and Global Forecast System Based on Machine Learning

Abstract

Random forests (RFs), an advanced machine learning (ML) method, was used here to develop a robust and rapid quantitative precipitation estimates (QPEs) algorithm for the new-generation geostationary satellite of Himawari-8. In this algorithm, the global precipitation measurement (GPM) product has been employed to train QPE prediction model. The real-time multiband infrared brightness temperature from Himawari-8, combined with the spatiotemporally matched numerical weather prediction (NWP) data from the global forecast system, have been used as predictor variables for QPE. Among the variables used in RF learning model, total precipitable water and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$K$ </tex-math></inline-formula> -index from NWP data have the highest rankings, indicating the importance of atmospheric environment for QPE. To enhance the accuracy of RF models or to optimize model training, a sample-balance technique has been utilized to adjust the ratios of samples in nonprecipitation/precipitation classification and quantitative precipitation regression data sets. Further sensitivity and validation analyses help determine the optimal RF classification and regression models for predicting nonprecipitation/precipitation pixel and rain rate. The selected RF classification model is found to predict precipitation area with an accuracy of 0.87. For predicted QPE product, the mean-absolute-error and root-mean-square error of RF regression model are 0.51 and 2.0 mm/h, respectively. Overall, the RF ML algorithm has a higher detection rate over homogenous ocean surface as compared with over land. Meanwhile, this RF algorithm tends to underestimate rain rate, especially in the presence of heavy rainfall. Despite this, it still produces a reasonable pattern of rainfall area and intensity, which are highly consistent with GPM observations.