Algorithmic Forecasting of Extreme Heat Waves

TL;DR

This paper develops a forecasting framework for extreme heat waves using AIRS satellite data, genetic algorithms for predictor selection, and conformal prediction for uncertainty quantification, demonstrating high classification accuracy in two regions.

Contribution

It introduces a novel approach combining genetic algorithms and conformal prediction to improve extreme heat wave forecasting accuracy.

Findings

High classification accuracy achieved in Pacific Northwest and Phoenix.

Genetic algorithms effectively identify key predictors.

Conformal prediction provides reliable uncertainty estimates.

Abstract

This paper provides some foundations for valid forecasting of rare and extreme heat waves through a better understanding of the similarities and differences between several consecutive hot days under normal circumstances and rare, extreme heat waves. We analyze AIRS data from the American Pacific Northwest and AIRS data from the Phoenix, Arizona region. A genetic algorithm is used to help determine the most promising predictors. Classification accuracy with supervised learning is excellent for the Pacific Northwest and is replicated for Phoenix. Conformal prediction sets are considered as a way to represent forecasting uncertainty. Complications caused by endogenous sampling are discussed.