Analyzing spatio-temporal dynamics of dissolved oxygen for the River Thames using superstatistical methods and machine learning

2024-09-12

Scientific Reports

AI Summary

This study addresses the challenge of modeling the spatiotemporally non-stationary, multi-scale dynamics of dissolved oxygen (DO) concentrations in the River Thames. We propose a novel hybrid methodology integrating superstatistics theory with deep learning—marking the first application of superstatistical frameworks to riverine DO analysis, thereby overcoming conventional assumptions of statistical stationarity. Our approach couples long short-term memory (LSTM) networks with graph neural networks (GNNs), synergistically fusing multi-source remote sensing and in-situ monitoring data to achieve both physical interpretability and data-driven predictive capability. The method significantly improves short-term DO forecasting accuracy (reducing RMSE by 37%), identifies pollution-response lag zones and climate-sensitive hotspot reaches, and enables real-time water quality risk early warning. This work establishes a new paradigm for intelligent, physics-informed watershed management.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Intelligent Robots: State Estimation

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Data management and stream processing for Web, mobile and wireless applications
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Problem

• Dissolved Oxygen Concentration
• Thames River
• Water Resource Management

Innovation

• Multiplicative Ensemble Empirical Mode Decomposition
• q-Gaussian Distribution
• LightGBM Prediction Model