A Deep Learning Approach to Probabilistic Forecasting of Weather

TL;DR

This paper presents a novel deep learning framework combining dimensional reduction and normalizing flows to produce probabilistic weather forecasts, validated on 22 years of wind simulation data.

Contribution

It introduces a two-step machine learning approach that enhances probabilistic weather forecasting by preserving information and preventing overfitting.

Findings

Effective dimensional reduction improves forecast sharpness.

Probabilistic calibration prevents overfitting.

Method validated on 22-year wind data.

Abstract

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.