A Deep Learning Approach to Probabilistic Forecasting of Weather
This work addresses the problem of accurate and reliable probabilistic weather predictions for meteorologists and climate scientists, representing an incremental advancement in forecasting techniques.
The paper tackles probabilistic weather forecasting by introducing a two-step deep learning method combining dimensional reduction and density estimation, achieving improved forecast sharpness and calibration on a 22-year hourly wind simulation dataset.
We discuss an approach to probabilistic forecasting based on two chained machine-learning steps: a dimensional reduction step that learns a reduction map of predictor information to a low-dimensional space in a manner designed to preserve information about forecast quantities; and a density estimation step that uses the probabilistic machine learning technique of normalizing flows to compute the joint probability density of reduced predictors and forecast quantities. This joint density is then renormalized to produce the conditional forecast distribution. In this method, probabilistic calibration testing plays the role of a regularization procedure, preventing overfitting in the second step, while effective dimensional reduction from the first step is the source of forecast sharpness. We verify the method using a 22-year 1-hour cadence time series of Weather Research and Forecasting (WRF) simulation data of surface wind on a grid.