A Volume-to-Point Approach of Radar-Based QPE

In this study, we proposed a 'volume-to-point' framework for quantitative precipitation estimation (QPE) based on the QPESUMS (Quantitative Precipitation Estimation and Segregation Using Multiple Sensor) Mosaic Radar dataset. With a data volume consisting of the time series of gridded radar reflectivities over the Taiwan area, we used machine learning algorithms to establish a statistical model for each weather station. The model extracts spatial and temporal features from the input data volume and then associates these features with the location-specific precipitations. In contrast to QPE methods based on the z-r relation, we leverages the machine learning algorithms to automatically detect the existence and movement of synoptic systems and associate these patterns to a particular location with specific topographic attributes. The proposed framework is considered as a data-driven baseline of QPE without human knowledge. The proposed framework is considered as a data-driven baseline of QPE without human knowledge. We evaluated this framework with the hourly precipitation data of 45 weather stations in Taipei during 2013 ~ 2016. The overall root-mean-squared-error (RMSE) is 1.79 mm/hour, which is lower than the WRF 5km ensemble model output (5.5 mm/hour). However, the proposed approach tends to underestimate the cases of heavy rainfall while it is very accurate while the precipitation is under 10mm/hour. In the future, we will further investigate the characteristics of the proposed framework and its ability for quantitative precipitation forecasting.