Enhancing the Performance of Quantitative Precipitation Estimation Using Ensemble of Machine Learning Models Applied on Weather Radar Data

Flash floods are a major weather-related risk, as they cause more than 5000 fatalities annually, according to the World Meteorological Organization. Quantitative Precipitation Estimation is method used approximate rainfall over locations where direct field observations not available. It represents one of most valuable information employed by meteorologists and hydrologists for issuing early warnings concerning flash floods. The current study in line with efforts improve radar-based estimates through use machine learning techniques applied on radar data. With this aim, proof concept, six models evaluated make estimations hourly accumulated using reflectivity data collected lowest elevation angles, we employ new model representing these were WSR-98D weather Romanian Administration, located central region Romania, during 30 non-consecutive days convective seasons, between 2016 2021. We obtained encouraging results stacked model. In terms Root Mean Squared Error evaluation metric, proposed regressor better estimated about 33% also outperform baseline computed Z-R relationship 13%.