Search for: All records

Abstract India is largely devoid of high‐quality and reliable on‐the‐ground measurements of fine particulate matter (PM_2.5). Ground‐level PM_2.5concentrations 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 PM_2.5estimates 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 PM_2.5over 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 PM_2.5estimates. 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 PM_2.5as 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. 

Abstract. Black carbon (BC) is one of the dominant absorbing aerosol species in the atmosphere. It normally has complex fractal-like structures due to the aggregation process during combustion. A wide range of aerosol–radiation interactions (ARIs) of BC have been reported throughout experimental and modeling studies. One reason for the large discrepancies among multiple studies is the application of the oversimplified spherical morphology for BC in ARI estimates. In current climate models, the Mie theory is commonly used to calculate the optical properties of spherical BC aerosols. Here, we employ a regional chemical transport model coupled with a radiative transfer code that utilizes the non-spherical BC optical simulations to re-evaluate the effects of particles' morphologies on BC shortwave ARI, and the wavelength range of 0.3–4.0 µm was considered. Anthropogenic activities and wildfires are two major sources of BC emissions. Therefore, we choose the typical polluted area in eastern China, which is dominated by anthropogenic emissions, and the fire region in the northwest US, which is dominated by fire emissions in this study. A 1-month simulation in eastern China and a 7 d simulation in the fire region in the northwest US were performed. The fractal BC model generally presents a larger clear-sky ARI compared to the spherical BC model. Assuming BC particles are externally mixed with other aerosols, the relative differences in the time-averaged clear-sky ARI between the fractal model with a fractal dimension (Df) of 1.8 and the spherical model are 12.1 %–20.6 % and 10.5 %–14.9 % for typical polluted urban cities in China and fire sites in the northwest US, respectively. Furthermore, the regional-mean clear-sky ARI is also significantly affected by the BC morphology, and relative differences of 17.1 % and 38.7 % between the fractal model with a Df of 1.8 and the spherical model were observed in eastern China and the northwest US, respectively. However, the existence of clouds would weaken the BC morphological effects. The time-averaged all-sky ARI relative differences between the fractal model with a Df of 1.8 and the spherical model are 4.9 %–6.4 % and 9.0 %–11.3 % in typical urban polluted cities and typical fire sites, respectively. Besides, for the regional-mean all-sky ARI, the relative differences between the fractal model and the spherical model are less than 7.3 % and 16.8 % in the polluted urban area in China and the fire region in the US, respectively. The results imply that current climate modeling may significantly underestimate the BC ARI uncertainties as the morphological effects on BC ARI are ignored in most climate models.