contaminant-transport

A 2D water quality transporter model to calculate conservative advection and diffusion of constituents from an unstructured grid of flows

Scripts from paper Halloran and Hunkeler (2020), Science of the Total Environment 722, 137749. doi.org/10.1016/j.scitotenv.2020.137749

Advanced framework combining 2D contaminant transport simulation (advection–diffusion) with machine learning to classify ecological risk (Low/Medium/High) and generate dashboards, GIFs, and reports.

Reproducible hybrid mGFD-machine-learning defect correction for two-dimensional reactive contaminant transport, with trained PyTorch models, held-out validation, and an irregular-geometry audit.

Codes and documentation for a few class projects

Machine-learning inference of identifiable reactive-transport parameters from six-sensor ADR1D concentration histories.

Verified physics-constrained PyTorch surrogate for fast prediction of one-dimensional reactive contaminant transport fields.

Reproducible numerical validation of ADR1D-ML and ADR1D-NN using a locked analytical challenge set, a finite-volume reference, persisted evidence, and a DOI-ready technical report.

Bilingual web application for exploring validated one-dimensional reactive-transport simulations and comparing analytical, finite-volume, and neural solutions.

Synthetic 1D transport benchmark and curated water-quality dataset for surrogate modeling, inverse estimation, and hydrology ML tasks.

Precision-preserving optimized inference for one-dimensional reactive contaminant transport, with validated ADR1D-ML-Compact and ADR1D-NN-Fast models and reproducible Python workflows.

Reproducible causal machine-learning system for one-hour early warning of downstream contaminant arrival using synthetic ADR1D sensor histories.