Deep Learning for Detecting Anthropogenic Global Warming Signal in Daily Precipitation

Explore how deep learning techniques can detect anthropogenic global warming signals in daily precipitation patterns. Delve into a comprehensive analysis that utilizes convolutional neural networks trained on climate model simulations to identify emerging climate change indicators. Discover how daily precipitation data serves as an excellent predictor for observed planetary warming, showing clear deviations from natural variability since the mid-2010s. Examine the interpretability framework used to probe the machine learning model, revealing the sensitivity of daily precipitation variability in specific regions to anthropogenic warming. Gain insights into the detection of human interference in daily hydrological fluctuations, even when long-term shifts in annual mean precipitation remain non-emergent above natural background variability. Learn about the methodology, results, and implications of this innovative approach to climate change detection through this informative 29-minute conference talk by Yoo-Geun Ham from the PCS Institute for Basic Science.

Introduction
Impact of Global Warming
Global Climate Models
Long Term
Extreme Events
Climate Change Detection
Linear Detection
Convolutional Neural Network
Detection Results
Results
Emerged Days
Linear Trend
Occlusion Sensitivity
Detail Response
Summary