Repository for Computational Science Masters thesis: Citizen-Data-Driven Validation and Acceleration of HYSPLIT Air Pollution Simulations with Physics-Guided Machine Learning
This repository contains code for the following tasks, in the order that the folders are in:
- Processing of citizen reported data of percieved odour from 2016-2023.
- Results of HYSPLIT and application of DCGAN deep learning model to generate more data for these HYSPLIT simulations.
- Supercomputer (Snellius) implementation of HYSPLIT simulations for Pittsburgh.
The hysplit folder also contains the citizen-data-driven validation scheme for air pollution models. Please see READme pages within each folder for more information.