Multi-Quantile Regression for Extreme Precipitation Downscaling

2026-05-12

AI Summary

This study addresses the systematic underestimation of extreme precipitation events by existing deep learning–based downscaling models, which hinders reliable flood risk assessment. To overcome this limitation, the authors propose Q-SRDRN, a multi-quantile super-resolution network that jointly models the conditional distribution of precipitation using quantile regression and a tailored CNN architecture, effectively capturing both median trends and extreme tails. The method introduces an innovative IncrementBound mechanism to enforce quantile monotonicity while preserving gradient flow, and employs quantile-specific output heads to decouple modeling of extreme and moderate rainfall. Trained with pinball loss and cVAE-based data augmentation, Q-SRDRN demonstrates substantial improvements across multiple regions in Florida, California, and Texas, achieving extreme precipitation (P999) detection rates of 75.7%–81.9%, a 63% reduction in KL divergence, and a 3.9% decrease in RMSE.

Abstract

Deep super-resolution networks for precipitation downscaling achieve strong bulk skill yet systematically under-predict the heavy-tail events that drive flood risk. We demonstrate that the primary obstacle is the loss function, not the data: under intensity-weighted MAE, real and synthetic labels at the same input are simply averaged, meaning data augmentation shifts the predicted mean rather than the conditional distribution. We resolve this with Q-SRDRN, a multi-quantile super-resolution network trained with pinball loss at tau in 0.50, 0.95, 0.99, 0.999. Two CNN-specific design choices make this practical: IncrementBound enforces monotonicity while preserving each quantile channel's gradient identity, and separate per-quantile output heads provide independent filter banks for bulk and tail detection. Under this design, data augmentation via cVAE becomes complementary: the median head absorbs synthetic patterns without contaminating upper quantiles. Empirically, on Florida (convective/tropical-cyclone dominated), the un-augmented Q-SRDRN P999 head detects 1,598 of 2,111 events at 200 mm/day versus 88 for the deterministic baseline--an 18x detection-rate gain (4.2% to 75.7%)--with 63% lower KL divergence and 3.9% lower RMSE. Adding cVAE-generated samples lifts the P50 channel from 14 to 1,038 hits at 200 mm/day. On California (atmospheric-river dominated), the architecture reaches near-perfect detection (P999 SEDI >= 0.996 through 300 mm/day). On Texas, the baseline catches only 2 of 10,720 events at 200 mm/day while the P999 head catches 8,776 (81.9%). While the cVAE does not transfer across regions, multi-quantile regression captures extremes wherever the large-scale signal is strong, while augmentation rescues the median where it is not.

Problem

• extreme precipitation
• downscaling
• heavy-tail events
• flood risk
• quantile regression

Innovation

• multi-quantile regression
• extreme precipitation downscaling
• pinball loss
• super-resolution CNN
• conditional VAE