Untangling hybrid hydrological models with explainable artificial intelligence

Abstract

Hydrological models are valuable tools for developing streamflow predictions in unmonitored catchments to increase our understanding of hydrological processes. A recent effort has been made in the development of hybrid (conceptual/machine learning) models that can preserve some of the hydrological processes represented by conceptual models and can improve streamflow predictions. However, these studies have not explored how the data-driven component of hybrid models resolved runoff routing. In this study, explainable artificial intelligence (XAI) techniques are used to turn a ‘black-box’ model into a ‘glass box’ model. The hybrid models reduced the root-mean-square error of the simulated streamflow values by approximately 27, 50, and 24% for stations 17120000, 27380000, and 33680000, respectively, relative to the traditional method. XAI techniques helped unveil the importance of accounting for soil moisture in hydrological models. Differing from purely data-driven hydrological models, the inclusion of the production storage in the proposed hybrid model, which is responsible for estimating the water balance, reduced the short- and long-term dependencies of input variables for streamflow prediction. In addition, soil moisture controlled water percolation, which was the main predictor of streamflow. This finding is because soil moisture controls the underlying mechanisms of groundwater flow into river streams.

Hydrologic similarity based on width function and hypsometry: An unsupervised learning approach

“…Our approach focuses on the efficient use of large volumes of elevation data to find hydrological analogues through dynamical properties of terrains and facilitates large scale applications. This approach is consistent with the growing recognition in the hydrological community regarding the use of explainable AI (XAI) techniques that build upon conceptual and machine learning models to explain hydrological phenomenon (Maksymiuk et al, 2020;Althoff et al, 2021). An application of hydrological similarity study is to assist in improving our understanding of hydrological processes in watersheds (Blöschl et al, 2013) and future works can build upon this study by integrating the width function and elevation-based slope and velocity distribution to create a robust dynamical metric for hydrological response quantification and similarity assessment.…”

Hydrologic Similarity Based on Width Function and Hypsometry: An Unsupervised Learning Approach

“…Our approach focuses on the efficient use of large volumes of elevation data to find hydrological analogues through dynamical properties of terrains and facilitates large scale applications. This approach is consistent with the growing recognition in the hydrological community regarding the use of explainable AI (XAI) techniques that build upon conceptual and machine learning models to explain hydrological phenomenon (Maksymiuk et al, 2020;Althoff et al, 2021). An application of hydrological similarity study is to assist in improving our understanding of hydrological processes in watersheds (Blöschl et al, 2013) and future works can build upon this study by integrating the width function and elevation-based slope and velocity distribution to create a robust dynamical metric for hydrological response quantification and similarity assessment.…”

Can machine learning accelerate process understanding and decision‐relevant predictions of river water quality?

“…implemented for explaining predictions in rainfall-runoff modelling (Althoff et al, 2021;Yang & Chui, 2021), and are applicable for water quality applications as well (Wang et al, 2021). Also in the rainfallrunoff domain, used integrated gradients (Sundararajan et al, 2017) to confirm a theory-consistent influence of precipitation and air temperature on a NN state that correlated with snow water equivalent, building trust in the ability of such models to capture known physical processes.…”

“…Recent advances in explainable AI such as local interpretable model‐agnostic explanation (LIME; Ribeiro et al, 2016) or Shapley additive explanations based on occlusion analysis (SHAP; Lundberg & Lee, 2017) can explain individual predictions by many ML models (Samek et al, 2021). Both methods have been successfully implemented for explaining predictions in rainfall‐runoff modelling (Althoff et al, 2021; Yang & Chui, 2021), and are applicable for water quality applications as well (Wang et al, 2021). Also in the rainfall‐runoff domain, Kratzert et al (2019) used integrated gradients (Sundararajan et al, 2017) to confirm a theory‐consistent influence of precipitation and air temperature on a NN state that correlated with snow water equivalent, building trust in the ability of such models to capture known physical processes.…”