Benchmarking Artificial Intelligence Models for Daily Coastal Hypoxia Forecasting

2026-02-05

AI Summary

Existing seasonal models struggle to capture the daily-scale dynamics of coastal hypoxia, limiting the effectiveness of fine-grained ecological management. To address this gap, this study develops and systematically evaluates four deep learning architectures—BiLSTM, Medformer, ST-Transformer, and TCN—trained on hindcast hydrodynamic–biogeochemical data from 2009 to 2020 to enable daily hypoxia classification, with validation on 2020–2024 observations. We present the first reproducible, AI-driven framework for daily-scale coastal hypoxia prediction, explicitly incorporating key processes such as water column stratification, sediment oxygen consumption, and temperature-dependent organic matter decomposition, thereby supporting operational forecasting. Experimental results demonstrate consistently high performance across all models, with the ST-Transformer achieving state-of-the-art results (AUC-ROC ranging from 0.982 to 0.992), significantly outperforming alternative approaches.

Technology Category

• Machine Learning: Deep Neural Architectures and Foundation Models
• Computer Vision: Remote Sensing / Geospatial AI
• Humans and AI: Other Foundations of Human Computation & AI

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Coastal hypoxia, especially in the northern part of Gulf of Mexico, presents a persistent ecological and economic concern. Seasonal models offer coarse forecasts that miss the fine-scale variability needed for daily, responsive ecosystem management. We present study that compares four deep learning architectures for daily hypoxia classification: Bidirectional Long Short-Term Memory (BiLSTM), Medformer (Medical Transformer), Spatio-Temporal Transformer (ST-Transformer), and Temporal Convolutional Network (TCN). We trained our models with twelve years of daily hindcast data from 2009-2020 Our training data consists of 2009-2020 hindcast data from a coupled hydrodynamic-biogeochemical model. Similarly, we use hindcast data from 2020 through 2024 as a test data. We constructed classification models incorporating water column stratification, sediment oxygen consumption, and temperature-dependent decomposition rates. We evaluated each architectures using the same data preprocessing, input/output formulation, and validation protocols. Each model achieved high classification accuracy and strong discriminative ability with ST-Transformer achieving the highest performance across all metrics and tests periods (AUC-ROC: 0.982-0.992). We also employed McNemar's method to identify statistically significant differences in model predictions. Our contribution is a reproducible framework for operational real-time hypoxia prediction that can support broader efforts in the environmental and ocean modeling systems community and in ecosystem resilience. The source code is available https://github.com/rmagesh148/hypoxia-ai/

Problem

• coastal hypoxia
• daily forecasting
• ecosystem management
• fine-scale variability
• oxygen depletion

Innovation

• Spatio-Temporal Transformer
• daily hypoxia forecasting
• deep learning benchmarking
• reproducible AI framework
• coastal ecosystem modeling