A HydroLSTM‐Based Machine‐Learning Approach to Discovering Regionalized Representations of Catchment Dynamics

Abstract Finding similarities between model parameters across different catchments has proved to be challenging. Existing approaches struggle due to catchment heterogeneity and non‐linear dynamics. In particular, attempts to correlate catchment attributes with hydrological responses have failed due...

Full description

Saved in:
Bibliographic Details
Main Authors: Luis A. De la Fuente, Andrew Bennett, Hoshin V. Gupta, Laura E. Condon
Format: Article
Language:English
Published: Wiley 2025-08-01
Series:Water Resources Research
Subjects:
Online Access:https://doi.org/10.1029/2024WR039008
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846174012911124480
author Luis A. De la Fuente
Andrew Bennett
Hoshin V. Gupta
Laura E. Condon
author_facet Luis A. De la Fuente
Andrew Bennett
Hoshin V. Gupta
Laura E. Condon
author_sort Luis A. De la Fuente
collection DOAJ
description Abstract Finding similarities between model parameters across different catchments has proved to be challenging. Existing approaches struggle due to catchment heterogeneity and non‐linear dynamics. In particular, attempts to correlate catchment attributes with hydrological responses have failed due to interdependencies among variables and consequent equifinality. Machine Learning (ML), particularly the Long Short‐Term Memory (LSTM) approach, has demonstrated strong predictive and spatial regionalization performance. However, understanding the nature of the regionalization relationships remains difficult. This study proposes a novel approach to partially decouple learning the representation of (a) catchment dynamics by using the HydroLSTM architecture and (b) spatial regionalization relationships by using a Random Forest (RF) clustering approach to learn the relationships between the catchment attributes and dynamics. This coupled approach, called Regional HydroLSTM, learns a representation of “potential streamflow” using a single cell‐state, while the output gate corrects it to correspond to the temporal context of the current hydrologic regime. RF clusters mediate the relationship between catchment attributes and dynamics, allowing identification of spatially consistent hydrological regions, thereby providing insight into the factors driving spatial and temporal hydrological variability. Results suggest that by combining complementary architectures, we can enhance the interpretability of regional machine learning models in hydrology, offering a new perspective on the “catchment classification” problem. We conclude that an improved understanding of the underlying nature of hydrologic systems can be achieved by careful design of ML architectures to target the specific things we are seeking to learn from the data.
format Article
id doaj-art-710d87be0a1f4c119f981ab40b96fbc3
institution Kabale University
issn 0043-1397
1944-7973
language English
publishDate 2025-08-01
publisher Wiley
record_format Article
series Water Resources Research
spelling doaj-art-710d87be0a1f4c119f981ab40b96fbc32025-08-26T12:02:53ZengWileyWater Resources Research0043-13971944-79732025-08-01618n/an/a10.1029/2024WR039008A HydroLSTM‐Based Machine‐Learning Approach to Discovering Regionalized Representations of Catchment DynamicsLuis A. De la Fuente0Andrew Bennett1Hoshin V. Gupta2Laura E. Condon3Department of Hydrology and Atmospheric Sciences The University of Arizona Tucson AZ USADepartment of Hydrology and Atmospheric Sciences The University of Arizona Tucson AZ USADepartment of Hydrology and Atmospheric Sciences The University of Arizona Tucson AZ USADepartment of Hydrology and Atmospheric Sciences The University of Arizona Tucson AZ USAAbstract Finding similarities between model parameters across different catchments has proved to be challenging. Existing approaches struggle due to catchment heterogeneity and non‐linear dynamics. In particular, attempts to correlate catchment attributes with hydrological responses have failed due to interdependencies among variables and consequent equifinality. Machine Learning (ML), particularly the Long Short‐Term Memory (LSTM) approach, has demonstrated strong predictive and spatial regionalization performance. However, understanding the nature of the regionalization relationships remains difficult. This study proposes a novel approach to partially decouple learning the representation of (a) catchment dynamics by using the HydroLSTM architecture and (b) spatial regionalization relationships by using a Random Forest (RF) clustering approach to learn the relationships between the catchment attributes and dynamics. This coupled approach, called Regional HydroLSTM, learns a representation of “potential streamflow” using a single cell‐state, while the output gate corrects it to correspond to the temporal context of the current hydrologic regime. RF clusters mediate the relationship between catchment attributes and dynamics, allowing identification of spatially consistent hydrological regions, thereby providing insight into the factors driving spatial and temporal hydrological variability. Results suggest that by combining complementary architectures, we can enhance the interpretability of regional machine learning models in hydrology, offering a new perspective on the “catchment classification” problem. We conclude that an improved understanding of the underlying nature of hydrologic systems can be achieved by careful design of ML architectures to target the specific things we are seeking to learn from the data.https://doi.org/10.1029/2024WR039008regionalizationcatchment attributesstreamflowsimilarityclusteringHydroLSTM
spellingShingle Luis A. De la Fuente
Andrew Bennett
Hoshin V. Gupta
Laura E. Condon
A HydroLSTM‐Based Machine‐Learning Approach to Discovering Regionalized Representations of Catchment Dynamics
Water Resources Research
regionalization
catchment attributes
streamflow
similarity
clustering
HydroLSTM
title A HydroLSTM‐Based Machine‐Learning Approach to Discovering Regionalized Representations of Catchment Dynamics
title_full A HydroLSTM‐Based Machine‐Learning Approach to Discovering Regionalized Representations of Catchment Dynamics
title_fullStr A HydroLSTM‐Based Machine‐Learning Approach to Discovering Regionalized Representations of Catchment Dynamics
title_full_unstemmed A HydroLSTM‐Based Machine‐Learning Approach to Discovering Regionalized Representations of Catchment Dynamics
title_short A HydroLSTM‐Based Machine‐Learning Approach to Discovering Regionalized Representations of Catchment Dynamics
title_sort hydrolstm based machine learning approach to discovering regionalized representations of catchment dynamics
topic regionalization
catchment attributes
streamflow
similarity
clustering
HydroLSTM
url https://doi.org/10.1029/2024WR039008
work_keys_str_mv AT luisadelafuente ahydrolstmbasedmachinelearningapproachtodiscoveringregionalizedrepresentationsofcatchmentdynamics
AT andrewbennett ahydrolstmbasedmachinelearningapproachtodiscoveringregionalizedrepresentationsofcatchmentdynamics
AT hoshinvgupta ahydrolstmbasedmachinelearningapproachtodiscoveringregionalizedrepresentationsofcatchmentdynamics
AT lauraecondon ahydrolstmbasedmachinelearningapproachtodiscoveringregionalizedrepresentationsofcatchmentdynamics
AT luisadelafuente hydrolstmbasedmachinelearningapproachtodiscoveringregionalizedrepresentationsofcatchmentdynamics
AT andrewbennett hydrolstmbasedmachinelearningapproachtodiscoveringregionalizedrepresentationsofcatchmentdynamics
AT hoshinvgupta hydrolstmbasedmachinelearningapproachtodiscoveringregionalizedrepresentationsofcatchmentdynamics
AT lauraecondon hydrolstmbasedmachinelearningapproachtodiscoveringregionalizedrepresentationsofcatchmentdynamics