Automated Machine Learning to Evaluate the Information Content of Tropospheric Trace Gas Columns for Fine Particle Estimates Over India: A Modeling Testbed. (5th March 2023)

• Record Type: Journal Article
• Authors: Zheng, Zhonghua; Fiore, Arlene M.; Westervelt, Daniel M.; Milly, George P.; Goldsmith, Jeff; Karambelas, Alexandra; Curci, Gabriele; Randles, Cynthia A.; Paiva, Antonio R.; Wang, Chi; Wu, Qingyun; Dey, Sagnik
• Is Part Of: Journal of advances in modeling earth systems. Volume 15:Number 3(2023)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2023-03-05
• Subjects: Geological modeling -- Periodicals; Climatology -- Periodicals; Geochemical modeling -- Periodicals; 551.5011
• DOI: 10.1029/2022MS003099
• Languages: English
• ISSNs: 1942-2466

Abstract

Abstract: India is largely devoid of high‐quality and reliable on‐the‐ground measurements of fine particulate matter (PM2.5 ). Ground‐level PM2.5 concentrations are estimated from publicly available satellite Aerosol Optical Depth (AOD) products combined with other information. Prior research has largely overlooked the possibility of gaining additional accuracy and insights into the sources of PM using satellite retrievals of tropospheric trace gas columns. We evaluate the information content of tropospheric trace gas columns for PM2.5 estimates over India within a modeling testbed using an Automated Machine Learning (AutoML) approach, which selects from a menu of different machine learning tools based on the data set. We then quantify the relative information content of tropospheric trace gas columns, AOD, meteorological fields, and emissions for estimating PM2.5 over four Indian sub‐regions on daily and monthly time scales. Our findings suggest that, regardless of the specific machine learning model assumptions, incorporating trace gas modeled columns improves PM2.5 estimates. We use the ranking scores produced from the AutoML algorithm and Spearman's rank correlation to infer or link the possible relative importance of primary versus secondary sources of PM2.5 as a first step toward estimating particle composition. Our comparison of AutoML‐derived models to selected baseline machine learning models demonstrates that AutoML is at least as good as user‐chosen models. The idealized pseudo‐observations (chemical‐transport model simulations) used in this work lay the groundwork for applying satellite retrievals of tropospheric trace gases to estimate fine particle concentrations in India and serve to illustrate the promise of AutoML applications in atmospheric and environmental research. Plain Language Summary: Ground‐level fine particle (PM2.5 ) concentrations are frequently estimated with freely available satellite Aerosol Optical Depth (AOD) products. We focus on India where sparse ground‐based monitoring leaves gaps in our understanding of particle concentrations and the relative importance of different sources. We use an atmospheric chemistry model to test whether satellite retrievals of tropospheric trace gas columns can provide information on the origins of PM2.5 and improve satellite‐derived PM2.5 . We created an Automated Machine Learning workflow to evaluate the utility of incorporating multiple trace gas columns in PM2.5 estimates, which represents nonlinear relationships between predictands and predictors while freeing users from selecting and tuning a specific machine learning model. On daily and monthly time scales, we quantify the relative information content of trace gas columns, AOD, meteorological fields, and emissions. We find that incorporating trace gas columns improves PM2.5 estimates and may also enable inference of broad characteristics of particle composition. Key Points: We developed an Automated Machine Learning workflow to evaluate the utility of incorporating multiple trace gas columns in PM2.5 estimates Tropospheric trace gas columns contain signatures of PM2.5 precursors and improve PM2.5 estimates We infer or link the possible relative importance of primary versus secondary sources of PM2.5 using Automated Machine Learning and Spearman's rank correlation