Multi-gauge Hydrological Variational Data Assimilation: Regionalization Learning with Spatial Gradients using Multilayer Perceptron and Bayesian-Guided Multivariate Regression

TL;DR

This paper introduces a novel regionalization method for hydrological modeling that combines machine learning and Bayesian-guided regression to improve parameter estimation in high-resolution models, especially for ungauged basins.

Contribution

It presents a seamless regionalization technique integrating multilayer perceptrons and Bayesian-guided multivariate regression within differentiable hydrological models for better parameter transfer functions.

Findings

Effective regional transfer functions learned using the proposed methods.

Improved parameter estimation accuracy demonstrated on hydrological models.

Addresses equifinality through Bayesian-guided regression.

Abstract

Tackling the difficult problem of estimating spatially distributed hydrological parameters, especially for floods on ungauged watercourses, this contribution presents a novel seamless regionalization technique for learning complex regional transfer functions designed for high-resolution hydrological models. The transfer functions rely on: (i) a multilayer perceptron enabling a seamless flow of gradient computation to employ machine learning optimization algorithms, or (ii) a multivariate regression mapping optimized by variational data assimilation algorithms and guided by Bayesian estimation, addressing the equifinality issue of feasible solutions. The approach involves incorporating the inferable regionalization mappings into a differentiable hydrological model and optimizing a cost function computed on multi-gauge data with accurate adjoint-based spatially distributed gradients.