Authors: Marshall Rosenhoover, John Beck, John Rushing and Sara Graves – University of Alabama in Huntsville
Title: Super-Resolution of Radar-Based Precipitation Data using Ground Sensors
Abstract: Researchers at the University of Alabama in Huntsville’s (UAH) Information Technology and Systems Center (ITSC) have developed methods for conducting super-resolution of precipitation datasets from the satellite based Global Precipitation Measurement Mission Validation Network (GPM VN) by using deep learning models and radar-based precipitation data as the ground truth. These models, consisting of Convolutional Neural Networks (CNNs), are used to enhance the resolution of GPM Dual-frequency Precipitation Radar (DPR) data (4km) to match the Multi-Radar/Multi-Sensor (MRMS) System data (1km). This work was conducted in Year 1 of a three-year grant from NOAA’s Cooperative Institute for Research to Operations in Hydrology (CIROH). In year 2, we have expanded this super-resolution research to investigate the use of public ground-based precipitation sensors as the ground-truth to develop a neural network for improving the MRMS resolution. Our initial effort focuses on Hawaii as a proof of concept to determine the viability of this approach. If successful, we plan to apply this super-resolution technique to Puerto Rico and the continental United States. The ultimate goal is to integrate our results into NOAA’s National Water Model (NWM) for improved operational hydrologic forecasting. Positive impacts are expected to be the greatest in areas where hydrologic models will benefit from improved spatial resolution, such as boundaries between catch basins.