Uncertainty Estimation with Deep Learning for Rainfall-Runoff Modelling

TL;DR

This paper introduces a benchmarking procedure for uncertainty estimation in deep learning models for rainfall-runoff prediction, demonstrating that reliable uncertainty quantification is achievable with current deep learning techniques.

Contribution

It establishes a standardized benchmark for uncertainty estimation in hydrological deep learning models and compares four baseline methods, including Mixture Density Networks and Monte Carlo dropout.

Findings

Deep learning can produce accurate and reliable uncertainty estimates.

The benchmarking procedure enables standardized evaluation of uncertainty methods.

Mixture Density Networks and Monte Carlo dropout are effective baseline approaches.

Abstract

Deep Learning is becoming an increasingly important way to produce accurate hydrological predictions across a wide range of spatial and temporal scales. Uncertainty estimations are critical for actionable hydrological forecasting, and while standardized community benchmarks are becoming an increasingly important part of hydrological model development and research, similar tools for benchmarking uncertainty estimation are lacking. We establish an uncertainty estimation benchmarking procedure and present four Deep Learning baselines, out of which three are based on Mixture Density Networks and one is based on Monte Carlo dropout. Additionally, we provide a post-hoc model analysis to put forward some qualitative understanding of the resulting models. Most importantly however, we show that accurate, precise, and reliable uncertainty estimation can be achieved with Deep Learning.