Using Machine Learning to Improve 1-h Probabilistic Low-Level Rotation Forecasts.

One focus of the NOAA Warn-on-Forecast (WoF) project is to provide rapid-update probabilistic guidance to human forecasters for short-term (e.g., 0-1 h) tornado forecasts. Tornadoes are unresolvable with current operational models, but convection-allowing ensembles (CAEs) such as the 3-km NSSL Experimental WoF System for Ensembles (NEWS-e) provide forecasts of low-level rotation that may potentially discriminate between tornadic and non-tornadic storms. However, CAE forecasts often contain large errors in storm intensity, timing, and location. Machine learning methods potentially mitigate these errors by leveraging ensemble uncertainty, incorporating multiple model variables, and accounting for forecast bias to produce calibrated probabilistic forecasts.

This study utilizes real-time NEWS-e forecasts generated during the 2016 -2018 NOAA Hazardous Weather Testbed Spring Forecasting Experiments and low-level azimuthal shear analyses from the NSSL Multi-Radar/Multi-Sensor (MRMS) product suite as input data for training various machine learning algorithms to produce calibrated 1-h probabilistic forecasts of low-level rotation. Given that strong low-level rotation occurs infrequently, we are oversampling the grid points within the MRMS azimuthal shear tracks to produce a more balanced dataset, which has been shown to improve machine-learning performance. Given that traditional, grid point verification statistics are limited in rare events, forecasts are evaluated in an object-matching framework following the methodology of Skinner et al. (2018). Results suggest that while machine learning does not always improve upon raw ensemble forecasts, it sometimes substantially increases forecast probabilities in regions of observed low-level rotation that are missed by the uncalibrated probabilities.