Final Report: Combining Empirical Orthogonal Function and Extreme Value Theory Methods to Characterize Observed and Future Changes in Extreme U.S. Air Pollution Events

EPA Grant Number: R835206
Title: Combining Empirical Orthogonal Function and Extreme Value Theory Methods to Characterize Observed and Future Changes in Extreme U.S. Air Pollution Events
Investigators: Fiore, Arlene M , Polvani, Lorenzo M
Institution: Columbia University in the City of New York
EPA Project Officer: Chung, Serena
Project Period: June 1, 2012 through May 31, 2015 (Extended to May 31, 2016)
Project Amount: $749,951
RFA: Extreme Event Impacts on Air Quality and Water Quality with a Changing Global Climate (2011) RFA Text |  Recipients Lists
Research Category: Air Quality and Air Toxics , Water Quality , Climate Change , Air , Water

Objective:

By analyzing observations and a suite of model simulations, we identify and characterize changes in extreme U.S. pollution (ozone and particulate matter) events and some of their underlying drivers, over the past several decades as well as for a range of future climate and emission scenarios. We aim to provide a range of projected 21st century changes in the frequency, intensity, and duration of U.S. pollution events that considers the role of climate variability as well as climate change. We seek to develop approaches to (1) translate coarse scale chemistry-climate model projections of future air pollution levels into policy-relevant metrics that account for known model biases and (2) rapidly screen large datasets generated by ensemble, multi-scenario, multi-century simulations with chemistry-climate models for changes in pollution events. We propose that 1-year return levels, which can be derived using extreme value theory, offer a probabilistic metric useful for air quality planning. We also examine changes in the frequency and intensity of extreme temperature and precipitation events over the United States, including in response to changes in atmospheric abundance of aerosols versus greenhouse gases, to provide information on these two hazards that affect air and water quality.

Summary/Accomplishments (Outputs/Outcomes):

Approaches to quantify past and future changes in extreme events. We developed new approaches to characterize changes in high-ozone events from long-term observations and from transient 21st century chemistry-climate model simulations under different climate and emission scenarios. Specifically, we applied statistical methods from extreme value theory to characterize the decline in high levels of ground-level ozone (maximum daily average 8-hour concentrations) over recent decades in terms of 1-year summertime return levels over the northeastern United States [Rieder, et al., 2013]. After developing a regional-scale bias correction approach and applying it to surface ozone fields archived from the GFDL CM3 chemistry-climate model over the Northeast, we provide a range of projected changes in surface ozone events exceeding past, current, and possible future levels of the National Ambient Air Quality Standard (NAAQS) for ozone, as well as 1-year summertime return levels [Rieder, et al., 2015], also available from our website. We currently are preparing a manuscript documenting a statistical downscaling approach to combine regional-scale information from coarse resolution global models with historical observations at individual monitoring sites across the United States to provide estimates for changes in policy-relevant metrics (the number of days exceeding NAAQS levels, 1-year return levels) throughout the 21st century. Once peer-reviewed, these estimates also will be provided on our website. Another manuscript in preparation describes our proof-of-concept application of empirical orthogonal function analysis as a rapid screening tool to identify regional-scale pollution events and to gauge changes in their frequency and duration in climate model simulations. We expect this approach to become increasingly useful as computational advances permit large ensembles, spanning numerous possible emission and climate states, to be generated from chemistry-climate and air quality models. We also documented a proof-of-concept study indicating that synthetic tracers implemented in physical climate models, which typically perform more scenarios than possible with full chemistry-climate models, can provide information on the response of summertime particulate matter pollution episodes to climate change [Fang, et al., 2013].

Attributing changes in pollution events to underlying drivers. By analyzing a broad set of simulations in the GFDL CM3 model, and conducting additional sensitivity simulations for this project, we parsed separately changes in U.S. ozone pollution induced by regional precursor emissions, global methane abundances, and climate change. Recognizing that interpreting changes in extremes requires a solid understanding of changes occurring throughout the distribution, we broadened our focus to investigate changes occurring throughout the pollutant distributions. We pointed out a possible reversal of the ground-level ozone seasonal cycle over the northeastern United States brought about by continued NOx emission reductions, but with substantial amplification of cold season levels possible if global methane abundances rise, and highlighted evidence for shifting seasonality in the observational record [Clifton, et al., 2014]. We examined the implications of several emission and climate change scenarios for high-ozone events and the overall ozone distribution, and identified a simple relationship between the changes in estimated 1-year return levels for daily maximum 8-hour ozone and regional NOx emission changes [Rieder, et al., 2015]. This simple relationship implies that our results can be generalized to estimate changes in the frequency of EUS pollution events under different regional NOx emission scenarios. We showed that fine particulate matter (PM) increases modestly in some regions as climate warms, but that changes in emissions exert a stronger influence on PM levels over most populated regions [Westervelt, et al., 2016]. We also studied connections between specific meteorological drivers and pollutant levels. We found that the position of the northern mid-latitude jet stream alters the local relationship between ozone and temperature over the eastern United States, providing a cautionary note on projecting future ozone levels directly from modeled changes in temperature based on locally observed relationships [Barnes and Fiore, 2013]. We also examined historical and projected changes in the migratory storm systems (cyclones) tracking from the Great Lakes region to the Northeast that are known to sweep pollution out of the continental boundary layer, and their connection with ozone events [Turner, et al., 2013]. Consistent with earlier work, we found that temperature, precipitation, and wind speed are the major drivers of PM variability, and showed that the GFDL CM3 model captures observed PM-temperature and PM-wind speed relationships over the eastern United States, but fails to represent the observed PM-precipitation relationship [Westervelt, et al., 2016]. Throughout this body of work, we emphasize the role of climate variability, which confounds both the detection of a robust climate change signal [Fiore, et al., 2015 provide a synthesis] and the attribution of observed trends to changes in emissions [Barnes, et al., 2016].

Investigating changes in extreme temperature and precipitation events. We analyzed daily temperature and precipitation fields from several sets of multi-ensemble member historical (1860-2005) simulations with the GFDL CM3 chemistry-climate model. We found that small changes in temperature (summer) and precipitation (all seasons) extremes over the United States during this period reflect cancelation between the contemporaneous rise in aerosols (PM) and greenhouse gases [Mascioli, et al., 2016a]. All changes are evaluated for statistical significance with respect to variability of the extreme temperature and precipitation indices in 800 years from a pre-industrial control simulation with perpetual 1860 conditions. The southeastern United States showed a muted temperature response to forcing from both global aerosols and greenhouse gases, prompting us to conduct a more systematic analysis of observed temperature trends in this region and to investigate the impact of climate variability and aerosol-induced cooling on the observed trends [Mascioli, et al., 2016b]. In 21st century scenarios in which aerosol emissions decline, the influence of greenhouse gases on extreme temperature and precipitation dominates, with daily maximum temperatures on nearly all summer days lying above the 1960-1990 90th percentile by the end of the century under an extreme warming scenario (RCP8.5) [Mascioli, et al., 2016a].

Conclusions:

The Outputs/Outcomes for this research include:

  • Documented approach to quantifying extreme pollution events using methods from extreme value theory;
  • Documented approach for a regional-scale bias correction for summertime ground-level ozone, which can be adapted and refined for future applications to other regions, seasons, and pollutants;
  • Generated a range of estimates for projected changes in policy-relevant metrics (NAAQS exceedances, return levels) for eastern U.S. high-ozone events, including those obtained with a statistical downscaling approach that combines projected regional changes from coarse resolution models with observed pollutant distributions;
  • Attributed changes in ozone and particulate matter under a range of scenarios to underlying drivers including changes in regional and global emissions, climate change, and climate variability, and in some cases particular meteorological conditions;
  • Identified a simple relationship between the changes in estimated 1-year return levels and regional NOx emission changes, implying that our results can be generalized to estimate changes in the frequency of EUS pollution events under different regional NOx emission scenarios;
  • Developed methodology for identifying changes in regional-scale pollution event frequency and duration in 21st century ensemble model simulations;
  • Quantified historical and future changes in extreme temperature and precipitation events over the U.S.A., and attributed a portion of these changes to climate forcing from particulate matter emissions;
  • Published 2015 Air & Waste Management Critical Review Paper describing two-way interactions between climate change and U.S. air quality including relevance for extreme pollution events.


Journal Articles on this Report : 11 Displayed | Download in RIS Format

Other project views: All 48 publications 11 publications in selected types All 11 journal articles
Type Citation Project Document Sources
Journal Article Barnes EA, Fiore AM. Surface ozone variability and the jet position:implications for projecting future air quality. Geophysical Research Letters 2013;40(11):2839-2844. R835206 (2012)
R835206 (Final)
  • Full-text: Wiley Online-Full Text HTML
    Exit
  • Abstract: Wiley Online-Abstract
    Exit
  • Other: Wiley Online-Full Text PDF
    Exit
  • Journal Article Barnes EA, Fiore AM, Horowitz LW. Detection of trends in surface ozone in the presence of climate variability. Journal of Geophysical Research:Atmospheres 2016;121(10):6112-6129. R835206 (Final)
  • Full-text: ResearchGate-Full Text PDF (prepublication)
    Exit
  • Abstract: Wiley-Abstract
    Exit
  • Journal Article Clifton OE, Fiore AM, Correa G, Horowitz LW, Naik V. Twenty-first century reversal of the surface ozone seasonal cycle over the northeastern United States. Geophysical Research Letters 2014;41(20):7343-7350. R835206 (2013)
    R835206 (2014)
    R835206 (Final)
  • Full-text: UCAR-Full Text PDF
    Exit
  • Abstract: Wiley Online-Abstract
    Exit
  • Other: Wiley Online-Full Text PDF
    Exit
  • Journal Article Fang Y, Fiore AM, Lamarque J-F, Horowitz LW, Lin M. Using synthetic tracers as a proxy for summertime PM2.5 air quality over the Northeastern United States in physical climate models. Geophysical Research Letters 2013;40(4):755-760. R835206 (2012)
    R835206 (Final)
  • Full-text: Wiley Online-Full Text HTML
    Exit
  • Abstract: Wiley Online-Abstract
    Exit
  • Other: Wiley Online-Full Text PDF
    Exit
  • Journal Article Fiore AM, Naik V, Leibensperger EM. Air quality and climate connections. Journal of the Air & Waste Management Association 2015;65(6):645-685. R835206 (2014)
    R835206 (Final)
  • Abstract from PubMed
  • Full-text: Taylor & Francis-Full Text HTML
    Exit
  • Abstract: Taylor & Francis-Abstract
    Exit
  • Other: Taylor & Francis-Full Text PDF
    Exit
  • Journal Article Mascioli NR, Fiore AM, Previdi M, Correa G. Temperature and precipitation extremes in the United States: quantifying the responses to anthropogenic aerosols and greenhouse gases. Journal of Climate 2016;29(7):2689-2701. R835206 (2014)
    R835206 (Final)
  • Full-text: Columbia University-Full Text PDF (prepublication)
    Exit
  • Abstract: AMS-Abstract
    Exit
  • Journal Article Mascioli NR, Previdi M, Fiore AM, Ting M. Timing and seasonality of the United States 'warming hole'. Environmental Research Letters 2017;12(3):034008 (10 pp.). R835206 (Final)
    R835878 (2016)
  • Full-text: IOP-Full Text PDF
    Exit
  • Abstract: IOP-Abstract & Full Text HTML
    Exit
  • Journal Article Rieder HE, Fiore AM, Polvani LM, Lamarque J-F, Fang Y. Changes in the frequency and return level of high ozone pollution events over the eastern United States following emission controls. Environmental Research Letters 2013;8(1):014012 (10 pp.). R835206 (2012)
    R835206 (Final)
  • Full-text: IOP Science-Full Text HTML
    Exit
  • Abstract: IOP Science-Abstract
    Exit
  • Other: IOP Science-Full Text PDF
    Exit
  • Journal Article Rieder HE, Fiore AM, Horowitz LW, Naik V. Projecting policy-relevant metrics for high summertime ozone pollution events over the eastern United States due to climate and emission changes during the 21st century. Journal of Geophysical Research:Atmospheres 2015;120(2):784-800. R835206 (2013)
    R835206 (2014)
    R835206 (Final)
  • Full-text: Columbia University-Full Text PDF
    Exit
  • Abstract: Wiley Online-Abstract
    Exit
  • Journal Article Turner AJ, Fiore AM, Horowitz LW, Bauer M. Summertime cyclones over the Great Lakes Storm Track from 1860–2100:variability, trends, and association with ozone pollution. Atmospheric Chemistry and Physics 2013;13(2):565-578. R835206 (2012)
    R835206 (Final)
  • Full-text: ACP-Full Text PDF
    Exit
  • Abstract: ACP-Abstract
    Exit
  • Other: Columbia University-Abstract & Full Text PDF
    Exit
  • Journal Article Westervelt DM, Horowitz LW, Naik V, Tai APK, Fiore AM, Mauzerall DL. Quantifying PM2.5-meteorology sensitivities in a global climate model. Atmospheric Environment 2016;142:43-56. R835206 (Final)
    R835878 (2016)
  • Full-text: ScienceDirect-Full Text HTML
    Exit
  • Abstract: ScienceDirect-Abstract
    Exit
  • Other: ScienceDirect-Full Text PDF
    Exit
  • Supplemental Keywords:

    ambient air, global climate, precipitation, sustainable air quality management, sustainable water management, atmospheric chemistry, air pollution, global change, regional climate change, air pollution episodes, ozone, aerosol, modeling, general circulation models, climate models, atmosphere, tropospheric, statistical downscaling, extreme value theory, CASTNET, AQS, United States, Northeast, Southeast, Midwest, particulate matter, climate variability, extreme temperature, extreme precipitation, warming hole, climate penalty.

    Relevant Websites:

    Related projects are described at Fiore Atmospheric Chemistry Group | Columbia University in the City of New York Exit

    Presentations at Group Presentations | Fiore Atmospheric Chemistry Group | Columbia University in the City of New York Exit

    Publications at Group Papers | Fiore Atmospheric Chemistry Group | Columbia University in the City of New York Exit

    Datasets at Datasets | Fiore Atmospheric Chemistry Group | Columbia University in the City of New York Exit

    Progress and Final Reports:

    Original Abstract
    2012 Progress Report
    2013 Progress Report
    2014 Progress Report