Probabilistic Inverse Modeling: An Application in Hydrology

TL;DR

This paper introduces a probabilistic inverse modeling framework in hydrology that enhances explainability, robustness, and uncertainty estimation, leading to improved streamflow prediction and more trustworthy basin characteristic estimates.

Contribution

It presents a novel probabilistic inverse model that improves uncertainty quantification and robustness in hydrological basin characteristic estimation from noisy data.

Findings

6% improvement in streamflow prediction accuracy

17% reduction in uncertainty with noisy data

4% higher coverage rate for basin characteristics

Abstract

The astounding success of these methods has made it imperative to obtain more explainable and trustworthy estimates from these models. In hydrology, basin characteristics can be noisy or missing, impacting streamflow prediction. For solving inverse problems in such applications, ensuring explainability is pivotal for tackling issues relating to data bias and large search space. We propose a probabilistic inverse model framework that can reconstruct robust hydrology basin characteristics from dynamic input weather driver and streamflow response data. We address two aspects of building more explainable inverse models, uncertainty estimation and robustness. This can help improve the trust of water managers, handling of noisy data and reduce costs. We propose uncertainty based learning method that offers 6\% improvement in $R^2$ for streamflow prediction (forward modeling) from inverse model inferred basin characteristic estimates, 17\% reduction in uncertainty (40\% in presence of noise) and 4\% higher coverage rate for basin characteristics.