The Influence of Particulate Matter and Methane on Regional Air Quality and Climate in the United States and India

Abstract

Atmospheric pollution is important for both air quality and climate considerations. In what follows, I explore three different atmospheric constituents – sulfate, particulate matter (PM2.5) from fires, methane – and their respective influence on regional (United States, India) air quality and climate. In situ surface observations show that downward surface solar radiation (SWdn) over the central and southeastern United States (U.S.) has increased by 0.58–1.0 Wm-2a-1 over the 2000–2014 timeframe, simultaneously with reductions in U.S. aerosol optical depth (AOD) of 3.3 – 5x10-3a-1. Establishing a link between these two trends, however, is challenging due to complex interactions between aerosols, clouds, and radiation. Here I investigate the clear-sky aerosol-radiation effects of decreasing U.S. aerosols on SWdn and other surface variables by applying a one-dimensional radiative transfer to 2000–2014 measurements of AOD at two Surface Radiation Budget Network (SURFRAD) sites in the central and southeastern United States. Observations characterized as “clear-sky” may in fact include the effects of thin cirrus clouds, and I consider these effects by imposing satellite data from the Clouds and Earth’s Radiant Energy System (CERES) into the radiative transfer model. The model predicts that 2000-2014 trends in aerosols may have driven clear-sky SWdn trends of +1.35 Wm-2a-1 at Goodwin Creek, MS, and +0.93 Wm-2a-1 at Bondville, IL. While these results are consistent in sign with observed trends, a cross-validated multivariate regression analysis shows that AOD reproduces 20-26% of the seasonal (June-September, JJAS) variability in clear-sky direct and diffuse SWdn at Bondville, IL, but none of the JJAS variability at Goodwin Creek, MS. Using in situ soil and surface flux measurements from the Ameriflux network and Illinois Climate Network (ICN) together with assimilated meteorology from the North American Land Data Assimilation System (NLDAS), I find that sunnier summers tend to coincide with increased surface air temperature and soil moisture deficits in the central U.S. The 1990-2015 trends in the NLDAS SWdn over the central U.S. are also of a similar magnitude as our modeled 2000–2014 clear-sky trends. Taken together, these results suggest that climate and regional hydrology in the central U.S. are sensitive to the recent reductions in aerosol concentrations. This work has implications for severely polluted regions outside the U.S., where improvements in air quality due to reductions in the aerosol burden could inadvertently pose an enhanced climate risk. Since at least the 1980s, many farmers in northwest India have switched to mechanized combine harvesting to boost efficiency. This harvesting technique leaves abundant crop residue on the fields, which farmers typically burn to prepare their fields for subsequent planting. A key question is to what extent the large quantity of smoke emitted by these fires contributes to the already severe pollution in Delhi and across other parts of the heavily populated Indo-Gangetic Plain located downwind of the fires. Using a combination of observed and modeled variables, including surface measurements of PM2.5, I quantify the magnitude of the influence of agricultural fire emissions on surface air pollution in Delhi. With surface measurements, I first derive the signal of regional PM2.5 enhancements (i.e., the pollution above an anthropogenic baseline) during each post-monsoon burning season for 2012-2016. I next use the Stochastic Time-Inverted Lagrangian Transport model (STILT) to simulate surface PM2.5 using five fire emission inventories. I reproduce up to 25% of the weekly variability in total observed PM2.5 using STILT. Depending on year and emission inventory, our method attributes 7.0–78% of the maximum observed PM2.5 enhancements in Delhi to fires. The large range in these attribution estimates points to the uncertainties in fire emission parameterizations, especially in regions where thick smoke may interfere with hotspots of fire radiative power. Although our model can generally reproduce the largest PM2.5 enhancements in Delhi air quality for 1-3 consecutive days each fire season, it fails to capture many smaller daily enhancements, which I attribute to the challenge of detecting small fires in the satellite retrieval. By quantifying the influence of upwind agricultural fire emissions on Delhi air pollution, our work underscores the potential health benefits of changes in farming practices to reduce fires. Anthropogenic methane emissions originate from a large number of relatively small point sources, often densely clustered and with a few anomalous sources contributing disproportionately to total emissions. Here I examine the potential of recently launched or planned satellites to promptly identify anomalous emitters among production sites in oil/gas fields through measurements of atmospheric methane, alone or supplemented by a surface observation network. I simulate atmospheric methane over a generic oil/gas field (20-500 production sites of different size categories in a 50x50 km2 domain) for a 1-week period using the WRF-STILT meteorological model with 1.3x1.3 km2 horizontal resolution. The simulations consider many random realizations for the occurrence and distribution of anomalous high-mode emitters in the field by sampling bimodal probability density functions (pdfs) of emissions from individual sites. The atmospheric methane field for each realization is then observed virtually with different satellite and surface observing configurations. Column methane enhancements observed from satellites are relatively small, even for high-mode emitters, so an inverse analysis is necessary. The inverse analysis can be regularized effectively using a L-1 norm to provide sparse solutions for a bimodally distributed variable. I find that the recently launched TROPOMI instrument (low Earth orbit, 7x7 km2 nadir pixels, daily return time) and the planned GeoCARB instrument (geostationary orbit, 2.7x3.0 km2 pixels, 2x or 4x/day return time) are successful at locating anomalous emitters for fields of 20 and sometimes 50 emitters within the 50x50 km2 domain, but unsuccessful for denser fields. GeoCARB does not benefit significantly from more frequent observations (4x/day vs. 2x/day). It performs better with a 5-km error tolerance for localization, but a next-generation geostationary satellite instrument with 1.3x1.3 km2 pixels, hourly return time, and 1 ppb precision can successfully detect and locate the high-mode emitters for a dense field with up to 500 sites in the 50x50 km2 domain. The capabilities of TROPOMI and GeoCARB can be usefully augmented with a surface observation network of 10-20 monitors, and in turn these satellite instruments increase the detection capability that can be achieved from the surface monitors alone.