Trends in remote PM2.5 residual mass across the United States: Implications for aerosol mass reconstruction in the IMPROVE network

Abstract

The Interagency Monitoring of Protected Visual Environments (IMPROVE) network collects aerosol samples for gravimetric and composition analysis in support of the Environmental Protection Agency's Regional Haze Rule and for long-term trend studies and model evaluations. Reconstructing PM2.5 mass or extinction from composition measurements requires assumptions of the molecular form of the individual species assumed to compose the bulk of PM2.5 mass. The IMPROVE reconstruction algorithm includes sulfate as ammonium sulfate, nitrate as ammonium nitrate, organic mass calculated with an assumed organic carbon (OC) to organic mass (OM) multiplier (OM/OC) of 1.8, elemental carbon, fine dust assuming common mineral oxides in soil, and sea salt calculated from chloride. Comparisons of reconstructed fine mass (RCFM) to PM2.5 gravimetric fine mass (FM) provide a check on these assumptions as well as help identify possible biases in gravimetric or speciated measurements. Significant changes in aerosol concentration and composition have occurred over time, leading to decreased FM across the United States. However, within the IMPROVE network, annual mean FM and RCFM have decreased at different rates from 2005 through 2016 (−29% versus −43%, respectively), causing the network median residuals (FM − RCFM) to increase by 0.49 μg m−3 over the 12-year period. The residual shifted from mostly negative before 2011 to mostly positive after 2011, with a strong summer peak. A multiple linear regression analysis indicated that FM biases increased due to the presence of particle-bound water (PBW) after 2011, associated with increased laboratory relative humidity during weighing. Results also suggested that the OM/OC ratio increased across the network after 2011, unrelated to the influence of PBW. While temporal behavior in the OM/OC ratio was similar across the network and for all seasons, values were highest in the East and during summer. Fine dust also appeared to be underestimated by ∼20%. Identifying the source of the trends in the FM residual is essential for accurately estimating contributions by individual species to RCFM and visibility degradation.