Grantee Research Project Results

2022 Progress Report: A diagnostic package to facilitate and enhance chemical mechanism implementations within regional and global atmospheric chemistry models

EPA Grant Number: R840011
Title: A diagnostic package to facilitate and enhance chemical mechanism implementations within regional and global atmospheric chemistry models
Investigators: Nicely, Julie M , Keller, Christoph , Tong, Daniel , Follette-Cook, Melanie , Ivatt, Peter
Current Investigators: Nicely, Julie M , Tong, Daniel , Follette-Cook, Melanie , Keller, Christoph
Institution: University of Maryland - College Park , Morgan State University , Universities Space Research Association
EPA Project Officer: Chung, Serena
Project Period: August 1, 2020 through July 31, 2023 (Extended to July 31, 2025)
Project Period Covered by this Report: August 1, 2021 through July 31,2022
Project Amount: $796,885
RFA: Chemical Mechanisms to Address New Challenges in Air Quality Modeling (2019) RFA Text | Recipients Lists
Research Category: Air , Air Quality and Air Toxics

Objective:

The objective of this study is to develop 1) a diagnostic software package, based on machine learning, to identify the key elements of a chemical mechanism to aid the future development of air quality models & 2) a chemical mechanism emulator that can be used as an alternative, computationally inexpensive method to simulate atmospheric chemistry.

Progress Summary:

Development of the machine learning framework for diagnosing and emulating air quality model chemical mechanisms is mostly complete and well-vetted for GEOS-Chem. Based in Python XGBoost, the framework has undergone substantive development in the past year to optimize training data sampling, emulation workflow (with differing treatments of long- versus short-lived species), and enable user-specified emulation setups (currently tested on 1- to 50-species simultaneous emulation). Functionality for surrogate (chemical regime-specific) emulators has also been incorporated.

Future Activities:

Testing of surrogate model functionality and chemical regime fingerprinting will be performed. Additional software package development will be undertaken to: maximize speedup gains by running package on GPU nodes, implement error accumulation safeguards by defaulting to the full chemical integrator in certain cases, optimize when relative vs. absolute changes are used to assess error, and implement a new loss function for XGBoost based on this differentiation, among several other advances to improve model stability. Finally, application of the software package to the CMAQ model will be targeted for the upcoming project year.

Journal Articles:

No journal articles submitted with this report: View all 4 publications for this project

Supplemental Keywords:

atmosphere, air quality, ozone, modeling, chemical sensitivity, model performance, model intercomparison, contiguous US, machine learning

Progress and Final Reports:

Original Abstract

The perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.