2016 Progress Report: Center for Air, Climate, and Energy Solutions (CACES)

EPA Grant Number: R835873
Center: Center for Air, Climate, and Energy Solutions
Center Director: Robinson, Allen
Title: Center for Air, Climate, and Energy Solutions (CACES)
Investigators: Robinson, Allen , Adams, Peter , Apte, Joshua S. , Azevedo, InĂªs L , Boies, Adam M. , Brauer, Michael , Burnett, Richard T , Coggins, Jay S. , Donahue, Neil , Ezzati, Majid , Hankey, Steve , Hill, Jason , Jaramillo, Paulina , Marshall, Julian D. , Matthews, H. Scott , Michalek, Jeremy J. , Millet, Dylan B , Muller, Nicholas , Pandis, Spyros N. , Polasky, Stephen , Pope, Clive Arden , Presto, Albert
Institution: Carnegie Mellon University , Brigham Young University , Health Canada - Ottawa , Imperial College , Middlebury College , The University of Texas at Austin , University of British Columbia , University of Minnesota , Virginia Polytechnic Institute and State University
EPA Project Officer: Chung, Serena
Project Period: May 1, 2016 through April 30, 2021
Project Period Covered by this Report: May 1, 2016 through April 30,2017
Project Amount: $10,000,000
RFA: Air, Climate And Energy (ACE) Centers: Science Supporting Solutions (2014) RFA Text |  Recipients Lists
Research Category: Airborne Particulate Matter Health Effects , Air , Climate Change , Health Effects


The Center for Air, Climate, and Energy Solutions (CACES) is a multidisciplinary, multi-institutional research center that is addressing critical questions at the nexus of air, climate, and energy. The Center has overarching themes of regional differences, multiple pollutants, and development and dissemination of tools for air quality impact assessment. Novel measurement and modeling approaches are being applied to understand spatial and temporal differences in human exposures and health outcomes. We are investigating a range of technology and policy scenarios for addressing our nation’s air, climate, and energy challenges, and testing their potential ability to meet policy goals such as improved health outcomes and cost-effectiveness.

The Center is comprised of five thematically and scientifically integrated research projects and one support center. Project 1 is extending existing chemical transport models to high spatial resolution (1 km) with tagged source apportionment and developing a new class of reduced complexity models for air quality and exposure assessment. Project 2 is conducting comprehensive measurements in three cities (Austin, TX; Oakland, CA; and Pittsburgh, PA) to quantify factors influencing gradients in pollutant concentrations and develop mechanistic understanding of how pollutant transformations affect population exposures. Project 3 is developing multi-pollutant land-use regression (LUR) models at high spatial resolution (~0.1 km), national-scale and over multiple decades. Project 4 is using tools developed in other projects to investigate key air, climate, and energy challenges and their interactions focusing on four main elements: electricity generation; transportation; land use; and climate-dependent emissions, transport and chemistry. Project 5 is analyzing nationally representative population-based health data, combined with novel multi-pollutant exposure estimates and source contributions (Projects 1 and 3), to derive new knowledge on multi-pollutant mortality risk and its variability across the United States.

Progress Summary:

Project 1: Mechanistic air quality impact models for assessment of multiple pollutants at high spatial resolution

Project 1 is focused on the development, evaluation and application of mechanistic air quality models, both chemical transport models (CTMs) and reduced-complexity models (RCMs).  Major activities in the past reporting period included:

  • Historical modeling of exposure to PM2.5 and related pollutants (1980-2015) for the continental United States.  During this project period research focused on the development of emissions inventories and meteorological inputs for the historical simulations.  Records from 1980 to present of dozens of spatially-resolved activity indicators (e.g., vehicle-miles traveled, acres of agricultural tillage, BTUs of power plant output) have been compiled. Published and measured emission factors have been gathered and compared. 
  • High-resolution (1 km) modeling of present-day air quality. During this project period research focused on developing high spatial resolution inventories for mobile and cooking sources for Pittsburgh. Using restaurant location data from the Google Places API, we have obtained restaurant data for the Pittsburgh Combined Statistical Area (CSA). Using these data, we demonstrated that the population density (the standard activity surrogate for cooking) is not well correlated with restaurant density at high spatial scales.
  • Development of Reduced-Complexity Models (RCMs).  During this project period significant progress was made on the development of three RCMs: AP2, EASIUR, and InMap. This included improved treatments of nitrate aerosol in AP2, creation of a source-receptor version of EASIUR, and development of a neural-network based chemical mechanism emulator for InMAP. 
  • Evaluation of Reduced-Complexity Models (RCMs).  During this project period, we compared among the three models the marginal social costs of SO2, NOx, NH3, and inert primary PM2.5 ground-level emissions in each county in the continental United States.  We also evaluated all three models using ambient air quality data. Despite fundamental structural differences among the three models, predicted marginal social costs are generally within the same order of magnitude and usually within a factor of 2 or 3. The agreement varies with the complexity of the chemistry that links the emissions to their equivalent ambient PM2.5 concentrations; predictions are most similar for primary PM2.5 and most different for NOx and NH3. Even with these differences, the three models generate robust rankings of national-level air quality policies based on social costs and benefits summed across pollutants and geographical locations.

Project 2: Air quality observatory 

Project 2 is collecting and analyzing air quality observations to characterize spatial (intra-city, urban-to-rural, and inter-city) and temporal distributions of multiple air pollutant species in three cities.  Major activities in the past reporting period included:

  • Evaluate performance of low-cost RAMP monitors.  Real-time Affordable Multi-Pollutant (RAMP) monitors measure CO, NO2, SO2, O3, and CO2 using low cost-sensors. Nineteen RAMPs were tested via co-location with the supersite on the Carnegie Mellon University campus and through laboratory calibration. We developed and evaluated three calibration strategies: a standard laboratory calibration; an empirical ambient calibration using a multi-linear regression that included sensor voltage, temperature, and RH as independent variables; and a machine-learning based calibration.  The machine learning algorithm performs the best; using it, every RAMP monitor evaluated meets the U.S. EPA Air Sensors Guidebook recommendations of minimum data quality for personal exposure measurement. We also demonstrated that a 4-week co-location period prior to deployment provides sufficient data for calibration model building.
  • Perform case studies of modifiable factors. We deployed RAMPs and additional instrumentation to create a monitoring network to investigate spatial and temporal patterns of air pollution in Pittsburgh. This network was used to conduct two multi-month case studies: (i) urban-rural transect and (ii) downtown business district (high traffic area with street canyons). 
  • High spatial mobile monitoring around air pollution sources. To complement the network of fixed sites, in-motion sampling with a mobile laboratory equipped with an aerosol mass spectrometer (AMS) and other instrumentation was used to characterize concentrations of a suite of air pollutants at high spatial resource around roads, restaurants, and other sources in Pittsburgh. This monitoring shows greatly elevated (up to a factor of 10) organic aerosol concentrations up to several hundred meters downwind of many restaurants. In contrast, organic aerosol concentrations on heavily-congested highways and other traffic dominated areas are only modestly higher than the urban background. This highlights the importance of cooking as a source for local exposures, especially given that restaurants often are located in residential neighborhoods.

Project 3: Next generation LUR models: Development of nationwide modeling tools for exposure assessment and epidemiology 

Project 3 is developing national scale, high spatial resolution (1 km), multi-pollutant (PM2.5, NO2, O3, CO, and subspecies of PM2.5) empirical models of air pollutant concentrations for use in health analysis and investigation of the influence of modifiable factors on human exposure.  Major activities in the past reporting period included:

  • Data compilation: We completed the assembly and processing of the necessary air pollution and geographic data for the development of empirical air pollution models.
  • Development of modeling framework:  Our modeling approach employs 2-stage partial least squares (PLS) + Universal Kriging to estimate annual average concentrations. PLS leverages predictive information from a large number of geographic covariates with less concern for model overfitting, while also limiting the impact of geographic covariate outliers. Making predictions at ~8-10 million Census block centroids for 6+ pollutants and 36 years (1980-2015) is a computationally intensive task.  Employing PostGIS and parallel processing, we are able to calculate our geographic covariates at all Census block centroids in ~20 days on a 10-node server. The improvement in processing of covariates with PostGIS (~100× faster than our previously published models using Python and ArcGIS) has dramatically improved our (and other researchers) ability to make fine-scale spatial predictions over very large geographic scales. 
  • Preliminary model building:  Preliminary national-scale land use regression (LUR) models have been developed for PM2.5 (1999-2015), NO2 (1979-2015), SO2 (1979-2015), O3 (1979-2015), and CO (1990-2015).  The preliminary PM2.5 models exhibit good performance across all years (CV-R2: 0.72-0.90). Preliminary NO2 models exhibit good performance for years 1981-2014 (CV-R2: 0.73-0.89, CCV-R2: 0.45-0.76). Preliminary O3 models exhibit good performance for years 1990-2014 (CV-R2: 0.69-0.79, CCV-R2: 0.52 -0.71), but much poorer performance prior to 1989 (CV-R2: 0.51-0.66, CCV-R2: 0.09-0.39). Preliminary SO2 and CO models exhibit poor-to-moderate model performance. Initial model results for PM2.5 and NO2 have been provided to Project 5 to begin working through linking exposure estimates with health data. 

Project 4: Air pollutant control strategies in a changing world 

Project 4 is applying chemical transport and reduced-form air quality models to assess the air quality and health impacts of various technology, policy, land-use, and climate scenarios.  Major activities in the past reporting period included:

  • United States economy-wide PM2.5 damages: We estimated air quality-related health effects for each of 428 sectors of the U.S. economy, the largest fractions of which were physically produced by electricity generation but induced by demand for manufactured goods. This alternative framing of air-quality related health impacts, which reveals the embodied health impacts of economic consumption, offers novel opportunities for strategies of air quality improvement. We used this new framework to explore health equity effects by economic sector, finding that Hispanic and Black populations are disproportionately impacted by electricity generation for manufacturing. 
  • Greenhouse gas and criteria air pollutant emissions under future policy scenarios:  We are comparing the economic efficiency of homogeneous versus heterogeneous air pollution regulations in the presence and absence of greenhouse gas regulations in the United States. Homogeneous regulations treat all emissions the same (what is currently in the United States) versus heterogeneous regulations that vary according to the magnitude of damage caused by the emission of a specific species in a specific location.
  • Air pollution impacts of corn production:  Using a life-cycle impact analysis and reduced complexity models, we performed spatially explicit analysis to estimate the damages of corn production. We estimate mean damages of $3.27/bu of corn produced in the United States, with 65% from ammonia emissions. Ammonia damages are more than six times larger than damages from GHG emissions. Spatial variation of damages is large, with the least damaging 5% of corn produced with damages less than $1.56/bu, and the most damaging 5% of corn produced with damages more than $6.17/bu.
  • VSL and mortality risk age adjustments:  We estimated the impact of including age differences in both the value of statistical life (VSL) and the risk of mortality. When we adjust both the VSL and mortality risk by age the total damages are lower than without an age adjustment, but not dramatically lower as conventional wisdom and past estimates indicate.
  • Controlling secondary organic aerosol (SOA) production from gasoline vehicle emissions: We found a strongly nonlinear relationship between SOA formation from gasoline vehicle exhaust and the atmospheric ratio of non-methane-organic-gas-to-NOx (NMOG:NOx). We investigated the implications of this relationship for the Los Angeles area. Although organic gas emissions from gasoline vehicles in Los Angeles are expected to fall by almost 80% over the next two decades, we predict no reduction in SOA production due to the effects of rising NMOG:NOx on SOA yields. This highlights the importance of integrated emission control policies for NOx and organic gases.

Project 5: Health effects of air pollution and mitigation scenarios 

Project 5’s specific aims include (1) estimate multi-pollutant mortality risk surfaces using two large, unique, population-based U.S. datasets and (2) explore regional and temporal variability in those risk surfaces.  Major activities in the past reporting period included:

  • Analysis of National Health Interview Survey (NHIS) data:  We have conducted a preliminary/preparatory study that evaluated associations between long-term PM2.5 exposure and mortality risk using cohorts of the U.S. adult population constructed from public-use NHIS data. Mortality hazard ratios (HRs) were estimated using Cox proportional hazards regression models, controlling for age, race, sex, income, marital status, education, body mass index, and smoking status. Estimated HRs for all-cause and cardiovascular mortality, associated with a 10 µg/m3 exposure increment of PM2.5, were 1.06 (1.01-1.11) and 1.34 (1.21-1.48), respectively, in models that controlled for various individual risk factors including smoking. This preliminary study demonstrates that the NHIS survey data with mortality linkage can be effectively used to evaluate mortality associations with air pollution.   
  • County-level mortality space-time study: We have established a time consistent set of data for 3,082 counties (essentially the entire continental U.S.) and carried out test analyses for a suite of models using age- and county-specific death rates based on national mortality and population data.

The Administrative Core provides overall oversight, coordination, and integration of the Center. Since initial funding of the Center, the Administrative Core has developed a quality management structure, which is detailed in the EPA-approved Quality Management Plan. An 11 member Science Advisory Committee was selected and the first annual meeting was held in January 2017 in Pittsburgh. An in-person center meeting was held in September 2016 in Pittsburgh. Finally, the Administrative Core has organized monthly conference calls of the project Executive Committee and weekly to monthly calls for groups of investigators for project-specific meetings.

Future Activities:

Project 1: Mechanistic air quality impact models for assessment of multiple pollutants at high spatial resolution

  • Complete and evaluate historical modeling of criteria pollutants for the 1980 to 2015 time period. The results will be passed to Projects 3 and 5. 
  • Complete the 1 km emissions inventories and modeling for Pittsburgh, evaluating the ability of CTMs to predict intraurban pollution variability against observations collected as part of Project 2. 
  • Continue enhancements of RCMs including extension of EASIUR to include social cost estimates for VOCs and enhancements to the treatment of difficult species like nitrate PM2.5 will be completed for AP2 and InMAP.

Project 2: Air quality observatory

  • Deploy a 50-location, low-cost monitor network in Pittsburgh.
  • Conduct mobile sampling in a high exposure, environmental justice area near the port of Oakland in Oakland, CA to quantify magnitude and sources of hotspots of fine particulate matter.
  • Measure emission factors from restaurants in Pittsburgh area using mobile measurements and tracer flux techniques. 
  • Provide quality assured high resolution air quality data collected in Pittsburgh to Projects 1 and 3 for model evaluation.

Project 3: Next generation LUR models: development of nationwide modeling tools for exposure assessment and epidemiology 

  • Continue model development to assess the geographic covariates selected into each model, particularly for poorer performing pollutants and years.
  • Provide Project 5 census block centroid estimates for all years (1980-2015), and county-level population-weighted estimates for years 1982-2013 to correspond with National Center for Health Statistics (NCHS) county mortality data.

Project 4: Air pollutant control strategies in a changing world 

  • Expand on the EPA US-TIMES model by incorporating region and sector-specific emissions damage values derived from the Estimating Air pollution Social Impact Using Regression (EASIUR) and the Air Pollution Emission Experiments and Policy analysis (AP2) models. 
  • Evaluate air quality scenarios associated with climate-dependent biogenic and wildfire emissions.
  • Perform simulations with PMCAMx for 2050 to evaluate applicability of future-day EASIUR to future scenarios.

Project 5: Health effects of air pollution and mitigation scenarios 

  • Complete proposal/application to Research Data Center (RDC) to access NHIS data.
  • Link CACES generated estimates for census block of residence to health data allowing for much greater spatial resolution.  
  • Perform preliminary health analyses using linked health and CACES exposure estimates.

Journal Articles: 9 Displayed | Download in RIS Format

Other center views: All 21 publications 9 publications in selected types All 9 journal articles
Type Citation Sub Project Document Sources
Journal Article Clark LP, Millet DB, Marshall JD. Changes in transportation-related air pollution exposures by race-ethnicity and socioeconomic status: outdoor nitrogen dioxide in the United States in 2000 and 2010. Environmental Health Perspectives 2017;125(9):097012 (10 pp.). R835873 (2016)
R835873C001 (2016)
R835873C003 (2016)
  • Abstract from PubMed
  • Full-text: EHP-Full Text PDF
  • Abstract: EHP-Abstract & Full Text HTML
  • Journal Article Heo J, Adams PJ, Gao HO. Public health costs accounting of inorganic PM2.5 pollution in metropolitan areas of the United States using a risk-based source-receptor model. Environment International 2017;106:119-126. R835873 (2016)
    R835873C001 (2016)
  • Abstract from PubMed
  • Full-text: ScienceDirect-Full Text HTML
  • Abstract: ScienceDirect-Abstract
  • Other: ScienceDirect-Full Text PDF
  • Journal Article Tessum CW, Hill JD, Marshall JD. InMAP: a model for air pollution interventions. PloS ONE 2017;12(4):e0176131 (26 pp.). R835873 (2016)
    R835873C001 (2016)
  • Full-text from PubMed
  • Abstract from PubMed
  • Associated PubMed link
  • Full-text: PLoS ONE-Full Text HTML
  • Abstract: PLoS ONE-Abstract
  • Other: PLoS ONE-Full Text PDF
  • Journal Article Weis A, Jaramillo P, Michalek J. Consequential life cycle air emissions externalities for plug-in electric vehicles in the PJM interconnection. Environmental Research Letters 2016;11(2):024009 (12 pp.). R835873 (2016)
    R835873C001 (2016)
    R835873C004 (2016)
  • Full-text: IOP Science-Full Text PDF
  • Abstract: IOP Science-Abstract & Full Text HTML
  • Other: Research Gate-Abstract & Full Text PDF
  • Journal Article Zhao Y, Saleh R, Saliba G, Presto AA, Gordon TD, Drozd GT, Goldstein AH, Donahue NM, Robinson AL. Reducing secondary organic aerosol formation from gasoline vehicle exhaust. Proceedings of the National Academy of Sciences of the United States of America 2017;114(27):6984-6989. R835873 (2016)
    R835873C001 (2016)
    R835873C004 (2016)
  • Abstract from PubMed
  • Abstract: PNAS-Abstract
  • Other: Harvard University-Abstract
  • Journal Article Muller NZ. Environmental benefit-cost analysis and the national accounts. Journal of Benefit-Cost Analysis 2017;1-40 [Epub ahead of print]. R835873C004 (2016)
  • Abstract:
  • Journal Article Muller NZ, Jha A. Does environmental policy affect scaling laws between population and pollution? Evidence from American metropolitan areas. PLOS One 2017;12(8):e0181407 (15 pp.). R835873C004 (2016)
  • Full-text from PubMed
  • Abstract from PubMed
  • Associated PubMed link
  • Full-text: PLOS ONE-Full Text PDF
  • Abstract: PLOS ONE-Abstract & Full Text HTML
  • Journal Article Zimmerman N, Presto AA, Kumar SPN, Gu J, Hauryliuk A, Robinson ES, Robinson AL, Subramanian R. Closing the gap on lower cost air quality monitoring:machine learning calibration models to improve low-cost sensor performance. Atmospheric Measurement Techniques Discussions August 2017 [In review]. R835873 (2016)
    R836286 (2016)
  • Full-text: AMT-Prepublication PDF
  • Abstract: AMT-Abstract
  • Journal Article Tessum, C.W., J. D. Hill, J. D. Marshall, InMAP:A model for air pollution interventions. PLoS ONE 12, e0176131, 2017. R835873C001 (2016)
    not available

    Supplemental Keywords:

    air pollution, climate, energy, health effects, social cost, impact assessment

    Relevant Websites:

    The Center for Air, Climate, and Energy Solutions Exit

    Subprojects under this Center: (EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
    R835873C001 Mechanistic Air Quality Impact Models for Assessment of Multiple Pollutants at High Spatial Resolution
    R835873C002 Air Quality Observatory
    R835873C003 Next Generation LUR Models: Development of Nationwide Modeling Tools for Exposure Assessment and Epidemiology
    R835873C004 Air Pollutant Control Strategies in a Changing World
    R835873C005 Health Effects of Air Pollution and Mitigation Scenarios