Search

The shifting frontiers of air pollution emission sources contribute to stagnation or reversal of air quality gains across the United States (US). The frequency and possible duration of Exceptional Events – driven primarily by wildfires and dust storms – have significantly increased in the US over the past decade. Combined with the US Environmental Protection Agency (EPA) final rule strengthening primary annual National Ambient Air Quality Standards (NAAQS) for PM2.5 by 25%, communities will need to reevaluate domestic and international sources of PM2.5. This study applies the Isolation Forest methodology to Exceptional Event demonstrations to flag and evaluate sources of anomalies in large PM2.5 measurement datasets. Focusing on a decade of hourly PM2.5 data measured in seven regions across Texas from 2012 to 2021 (>3 million data points), we present methods to efficiently flag hourly PM2.5 anomalies with compute times of ~minutes and characterize their spatial impacts as local or (multi-) regional; subsequent evaluation of potential sources of the increase can then be conducted more efficiently in a targeted manner. For a subset of anomalies, we incorporate air mass back trajectories, surface influences, and positive matrix factorization to evaluate potential sources. Our anomaly characterization method separated statistically normal PM2.5 data and enabled differentiation of localized versus larger-scale PM2.5 sources. In addition, our method successfully characterized the Summer 2020 severe Saharan dust intrusions into Texas, as well as the influence of international smoke from Mexico on El Paso's regional air quality. This anomaly flagging and characterization method is promising for assessing the relative importance of sources to anomalies in PM2.5 and other criteria air pollutants for multiple purposes; while this work focuses on its capacity for exceptional event demonstrations, the applicability includes long-term trend analyses from environmental justice analyses of air pollutant exposure to air quality attainment demonstrations. Implications: The shifting frontiers of air pollution emission sources contribute to stagnation or reversal of air quality gains across the United States (US). The frequency and possible duration of Exceptional Events – driven primarily by wildfires and dust storms – have significantly increased in the US over the past decade. Combined with the US Environmental Protection Agency (EPA) final rule strengthening primary annual National Ambient Air Quality Standards (NAAQS) for PM2.5 by 25%, communities will need to reevaluate domestic and international sources of PM2.5. This study presents a robust methodology to rapidly flag and evaluate sources of anomalies in PM2.5 measurements. This anomaly flagging and characterization method is promising for assessing the relative importance of sources to anomalies in PM2.5 and other criteria air pollutants for multiple purposes; while this work focuses on its capacity for exceptional event demonstrations, the applicability includes long-term trend analyses from environmental justice analyses of air pollutant exposure to air quality attainment demonstrations. [ABSTRACT FROM AUTHOR]

Ammonia is a significant precursor of PM 2.5 particles and thus contributes to poor air quality in many regions. Furthermore, ammonia concentrations are rising due to the increase of large-scale, intensive agricultural activities, which are often accompanied by greater use of fertilizers and concentrated animal feedlots. Ammonia is highly reactive and thus highly variable and difficult to measure. Satellite-based instruments, such as the Atmospheric Infrared Sounder (AIRS) and the Cross-Track Infrared Sounder (CrIS), have been shown to provide much greater temporal and spatial coverage of ammonia distribution and variability than is possible with in situ networks or aircraft campaigns, but the validation of these data is limited. Here we evaluate MUSES (multi-spectra, multi-species, multi-sensors) ammonia retrievals from AIRS and CrIS against ammonia measurements from aircraft in the California Central Valley and in the Colorado Front Range. These are small datasets taken over high-source regions under very different conditions: winter in California and summer in Colorado. Direct comparisons of the surface values of the retrieved profiles are biased very low in California (∼ 40 ppbv) and slightly high in Colorado (∼ 4 ppbv). This bias appears to be primarily due to smoothing error, since applying the instrument operator effectively reduces the bias to zero; even after the smoothing error is accounted for, the average uncertainty at the surface is in the 20 %–30 % range. We also compare 3 years of CrIS ammonia against an in situ network in the Magic Valley in Idaho We show that CrIS ammonia captures both the seasonal signal and the spatial variability in the Magic Valley, although it is biased low here also. In summary, this analysis substantially adds to the validation record but also points to the need for more validation under many different conditions and at higher altitudes. [ABSTRACT FROM AUTHOR]

Ammonia is a significant precursor of PM2.5 particles and thus contributes to poor air quality in many regions. Furthermore, ammonia concentrations are rising due to the increase of large scale, intensive agricultural activities, which are often accompanied by greater use of fertilizers and concentrated animal feedlots. Ammonia is highly reactive, and thus highly variable and difficult to measure. Satellite based instruments, such as the Atmospheric Infrared Sounder (AIRS), and the Cross-Track Infrared Sounder (CrIS) sensors, have been shown to provide much greater temporal and spatial coverage of ammonia distribution and variability than is possible with in situ networks or aircraft campaigns, but the validation of these data is limited. Here we evaluate ammonia retrievals from AIRS and CrIS against ammonia measurements from aircraft in the California Central Valley and in the Colorado Front Range. The satellite datasets were small and in California were obtained under difficult conditions. We show that the surface values of the retrieved profiles are biased very low in California and slightly high in Colorado, and that the bias appears to be primarily due to smoothing error. We also compare three years of CrIS ammonia against an in situ network in the Magic Valley in Idaho We show that CrIS ammonia captures both the seasonal signal and the spatial variability in the Magic Valley, though it is biased low here also. In summary, analysis adds to the validation record but also points to the need for more validation under different conditions. [ABSTRACT FROM AUTHOR]

The shifting frontiers of air pollution emission sources contribute to stagnation or reversal of air quality gains across the United States (US). The frequency and possible duration of Exceptional Events - driven primarily by wildfires and dust storms - have significantly increased in the US over the past decade. Combined with the US Environmental Protection Agency (EPA) final rule strengthening primary annual National Ambient Air Quality Standards (NAAQS) for PM 2.5 by 25%, communities will need to reevaluate domestic and international sources of PM 2.5 .This study applies the Isolation Forest methodology to Exceptional Event demonstrations to flag and evaluate sources of anomalies in large PM 2.5 measurement datasets. Focusing on a decade of hourly PM 2.5 data measured in seven regions across Texas from 2012 to 2021 (>3 million data points), we present methods to efficiently flag hourly PM 2.5 anomalies with compute times of ~minutes and characterize their spatial impacts as local or (multi-) regional; subsequent evaluation of potential sources of the increase can then be conducted more efficiently in a targeted manner. For a subset of anomalies, we incorporate air mass back trajectories, surface influences, and positive matrix factorization to evaluate potential sources.Our anomaly characterization method separated statistically normal PM 2.5 data and enabled differentiation of localized versus larger-scale PM 2.5 sources. In addition, our method successfully characterized the Summer 2020 severe Saharan dust intrusions into Texas, as well as the influence of international smoke from Mexico on El Paso's regional air quality.This anomaly flagging and characterization method is promising for assessing the relative importance of sources to anomalies in PM 2.5 and other criteria air pollutants for multiple purposes; while this work focuses on its capacity for exceptional event demonstrations, the applicability includes long-term trend analyses from environmental justice analyses of air pollutant exposure to air quality attainment demonstrations. Implications : The shifting frontiers of air pollution emission sources contribute to stagnation or reversal of air quality gains across the United States (US). The frequency and possible duration of Exceptional Events - driven primarily by wildfires and dust storms - have significantly increased in the US over the past decade. Combined with the US Environmental Protection Agency (EPA) final rule strengthening primary annual National Ambient Air Quality Standards (NAAQS) for PM 2.5 by 25%, communities will need to reevaluate domestic and international sources of PM 2.5 . This study presents a robust methodology to rapidly flag and evaluate sources of anomalies in PM 2.5 measurements. This anomaly flagging and characterization method is promising for assessing the relative importance of sources to anomalies in PM 2.5 and other criteria air pollutants for multiple purposes; while this work focuses on its capacity for exceptional event demonstrations, the applicability includes long-term trend analyses from environmental justice analyses of air pollutant exposure to air quality attainment demonstrations.