Improving Stream Solute Predictions With a Modified LSTM Model Incorporating Solute Interdependences and Hysteresis Patterns

Abstract Surface runoff and infiltrated water en route to the stream interact with dynamic landscape properties, ranging from vegetation and microbial activities to soil and geological attributes. Stream solute concentrations are highly variable and interconnected due to these interactions, flow pat...

Full description

Saved in:
Bibliographic Details
Main Authors: Tarun Agrawal, Allison Goodwell, Praveen Kumar
Format: Article
Language:English
Published: Wiley 2025-03-01
Series:Journal of Geophysical Research: Machine Learning and Computation
Online Access:https://doi.org/10.1029/2024JH000383
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850278282654646272
author Tarun Agrawal
Allison Goodwell
Praveen Kumar
author_facet Tarun Agrawal
Allison Goodwell
Praveen Kumar
author_sort Tarun Agrawal
collection DOAJ
description Abstract Surface runoff and infiltrated water en route to the stream interact with dynamic landscape properties, ranging from vegetation and microbial activities to soil and geological attributes. Stream solute concentrations are highly variable and interconnected due to these interactions, flow paths, and residence times, and often exhibit hysteresis with flow. Significant unknowns remain about how point measurements of stream solute chemistry reflect interdependent hydrobiogeochemical and physical processes, and how signatures are encapsulated as nonlinear dynamical relationships between variables. We take a Machine Learning (ML) approach to understand and capture these dynamical relationships and improve predictions of solutes at short and long time scales. We introduce a physical process‐based “flow‐gate” into an Long Short‐Term Memory (LSTM) model, which enables the model to learn hysteresis behaviors if they exist. Further, we use information‐theoretic metrics to detect how solutes are interdependent and iteratively select source solutes that best predict a given target solute concentration. The “flow‐gate LSTM” model improves model predictions (1%–32% decreases in RMSE) relative to the standard LSTM model for all nine solutes included in the study. The predictive improvements from the flow‐gate LSTM model highlight the importance of lagged concentration and discharge relationships for certain solutes. It also indicates a potential limitation in the traditional LSTM model approach since flow rates are always provided as input sources, but this information is not fully utilized. This work provides a starting point for a predictive understanding of geochemical interdependencies using machine‐learning approaches and highlights potential improvements in model architecture.
format Article
id doaj-art-6ff3b0f6d6c14f06a9ebac1d7dd4ebb0
institution OA Journals
issn 2993-5210
language English
publishDate 2025-03-01
publisher Wiley
record_format Article
series Journal of Geophysical Research: Machine Learning and Computation
spelling doaj-art-6ff3b0f6d6c14f06a9ebac1d7dd4ebb02025-08-20T01:49:35ZengWileyJournal of Geophysical Research: Machine Learning and Computation2993-52102025-03-0121n/an/a10.1029/2024JH000383Improving Stream Solute Predictions With a Modified LSTM Model Incorporating Solute Interdependences and Hysteresis PatternsTarun Agrawal0Allison Goodwell1Praveen Kumar2Civil and Environmental Engineering University of Illinois at Urbana‐Champaign Champaign IL USAPrairie Research Institute University of Illinois at Urbana‐Champaign Champaign IL USACivil and Environmental Engineering University of Illinois at Urbana‐Champaign Champaign IL USAAbstract Surface runoff and infiltrated water en route to the stream interact with dynamic landscape properties, ranging from vegetation and microbial activities to soil and geological attributes. Stream solute concentrations are highly variable and interconnected due to these interactions, flow paths, and residence times, and often exhibit hysteresis with flow. Significant unknowns remain about how point measurements of stream solute chemistry reflect interdependent hydrobiogeochemical and physical processes, and how signatures are encapsulated as nonlinear dynamical relationships between variables. We take a Machine Learning (ML) approach to understand and capture these dynamical relationships and improve predictions of solutes at short and long time scales. We introduce a physical process‐based “flow‐gate” into an Long Short‐Term Memory (LSTM) model, which enables the model to learn hysteresis behaviors if they exist. Further, we use information‐theoretic metrics to detect how solutes are interdependent and iteratively select source solutes that best predict a given target solute concentration. The “flow‐gate LSTM” model improves model predictions (1%–32% decreases in RMSE) relative to the standard LSTM model for all nine solutes included in the study. The predictive improvements from the flow‐gate LSTM model highlight the importance of lagged concentration and discharge relationships for certain solutes. It also indicates a potential limitation in the traditional LSTM model approach since flow rates are always provided as input sources, but this information is not fully utilized. This work provides a starting point for a predictive understanding of geochemical interdependencies using machine‐learning approaches and highlights potential improvements in model architecture.https://doi.org/10.1029/2024JH000383
spellingShingle Tarun Agrawal
Allison Goodwell
Praveen Kumar
Improving Stream Solute Predictions With a Modified LSTM Model Incorporating Solute Interdependences and Hysteresis Patterns
Journal of Geophysical Research: Machine Learning and Computation
title Improving Stream Solute Predictions With a Modified LSTM Model Incorporating Solute Interdependences and Hysteresis Patterns
title_full Improving Stream Solute Predictions With a Modified LSTM Model Incorporating Solute Interdependences and Hysteresis Patterns
title_fullStr Improving Stream Solute Predictions With a Modified LSTM Model Incorporating Solute Interdependences and Hysteresis Patterns
title_full_unstemmed Improving Stream Solute Predictions With a Modified LSTM Model Incorporating Solute Interdependences and Hysteresis Patterns
title_short Improving Stream Solute Predictions With a Modified LSTM Model Incorporating Solute Interdependences and Hysteresis Patterns
title_sort improving stream solute predictions with a modified lstm model incorporating solute interdependences and hysteresis patterns
url https://doi.org/10.1029/2024JH000383
work_keys_str_mv AT tarunagrawal improvingstreamsolutepredictionswithamodifiedlstmmodelincorporatingsoluteinterdependencesandhysteresispatterns
AT allisongoodwell improvingstreamsolutepredictionswithamodifiedlstmmodelincorporatingsoluteinterdependencesandhysteresispatterns
AT praveenkumar improvingstreamsolutepredictionswithamodifiedlstmmodelincorporatingsoluteinterdependencesandhysteresispatterns