2017 Progress Report: Project 3: Air Quality and Climate Change Modeling: Improving Projections of the Spatial and Temporal Changes of Multipollutants to Enhance Assessment of Public Health in a Changing World

EPA Grant Number: R835871C003
Subproject: this is subproject number 003 , established and managed by the Center Director under grant R835871
(EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).

Center: Solutions for Energy, AiR, Climate and Health Center (SEARCH)
Center Director: Bell, Michelle L.
Title: Project 3: Air Quality and Climate Change Modeling: Improving Projections of the Spatial and Temporal Changes of Multipollutants to Enhance Assessment of Public Health in a Changing World
Investigators: Zhang, Yang , Bell, Michelle L. , Leung, Lai-yung Ruby , Streets, David G.
Institution: North Carolina State University , Argonne National Laboratory , Battelle Memorial Institute, Pacific Northwest Division , Yale University
EPA Project Officer: Callan, Richard
Project Period: October 1, 2015 through September 30, 2020
Project Period Covered by this Report: October 1, 2016 through September 30,2017
RFA: Air, Climate And Energy (ACE) Centers: Science Supporting Solutions (2014) RFA Text |  Recipients Lists
Research Category: Air , Climate Change , Airborne Particulate Matter Health Effects , Particulate Matter , Global Climate Change


The main goal of Project 3 is to make critical improvements to online-coupled air quality models (AQMs) and their inputs and outputs, and apply the improved AQMs to estimate the concentrations resulting from energy and emission scenarios (Project 1) to be used in health risk assessments (Project 4). During this reporting period, we have three specific objectives: (1) Identify the likely causes underlying model biases based on model evaluation of the preliminary simulations from year 1 using one global model (CESM/CAM5) and four regional AQMs during the 5-year period (2008 - 2012) and perform diagnostic analyses and sensitivity simulations to show improved model predictions; (2) Understand the impacts of climate and emission changes on heat waves and stagnation and their compounding effects on high ozone (O3) events in the United States, which will lay the groundwork for further investigating the impacts of weather extremes on fine particulate matter (PM2.5), as both high O3 and PM2.5 episodes have significant health effects; and (3) Develop new methods to generate high-resolution snapshots of several air pollutants (NO2 and PM2.5) from satellites, which will help improve emissions used for AQMs and surface-level air pollutant estimates to be used for health assessment studies.

Progress Summary:

In addition to the 5-year (2008 - 2012) simulations using three regional models (i.e., WRF/Chem, WRF-CAM5, WRF/Chem-ROMS) completed in year 1, we completed the 5-year simulations using online-coupled WRF-CMAQ in this reporting period. To pinpoint the likely causes for model biases, comprehensive diagnostic analyses have been performed based on the evaluation of the preliminary simulations of the 5-year period using CESM/CAM5 and the above four regional models. The diagnostic analyses provide a basis for sensitivity simulation design to pinpoint the actual causes of model biases and improve the model performance. January and July 2011 are selected for sensitivity simulations with improved model inputs, formulation and configurations. Model improvements have been made for each of the four regional models. The major improvements in model inputs include updated chemical boundary conditions (BCONs) of all species from a new set of 5-year simulations using the updated CESM/CAM5, as well as the use of chemical BCONs for several species constrained by their satellite-derived retrievals or surrogate variable for all the four models and corrected emission species mapping of some chemical species in WRF/Chem-ROMS. Improved model formulations include a bug fix in the nucleation scheme of the MADE aerosol module in the National Oceanic and Atmospheric Administration's released version of WRF/Chem v3.7.1 that caused sulfate mass imbalance, adjustment of the emission factor in the dust emission scheme used in WRF-CAM5, an updated cloud fraction calculation in WRF-CMAQ and updated WRF/Chem v3.6.1 to WRF/Chem 3.7.1 in WRF/Chem-ROMS. Improved model configuration is the use of a more advanced Multi-Scale Kain-Fritsch scheme (MSKF) cumulus scheme available in WRF/Chem v3.7.1 than the Grell 3-D (G3) cumulus parameterization in the updated WRF/Chem-ROMS. In addition, the hard-coded conversion factor of POC to POM in WRF-CAM5 is adjusted from 1.4 to 2.1, consistent with the factor used by other models.

The above improvements in model inputs, formulations and configurations help reduce model biases for most chemical species of interest. For example, the normalized mean biases (NMBs) of the maximum 8-hour O3 concentrations simulated by WRF/Chem change from -4.5 percent to 1.2 percent at the AQS sites and from -21 percent to -16.2 percent at the CASTNET sites in January, and from -13.5 percent to -12.3 percent at the CASTNET sites in July, indicating an improved performance for surface O3. The improved WRF/Chem performance for surface O3 concentrations is attributed to updated BCONs from the updated CESM/CAM5 with better agreement for tropospheric ozone residual (TOR) against the OMI-derived TOR than the original CESM/CAM5 simulation. The NMBs of the 24-hour average surface PM2.5 concentrations change from -42.1 percent to -21.8 percent at the CSN sites and from -16.3 percent to -1.7 percent at the CASTNET sites in January for WRF-CAM5, and from -22.8 percent to -9.4 percent at the CASTNET sites in July for WRF/Chem-ROMS. The improved WRF-CAM5 performance for surface PM2.5 is mainly attributed to the adjusted conversion factor from primary organic carbon to primary organic matter and the adjustment of the dust emission factor used in the DEAD dust emission scheme. The improved WRF/Chem-ROMS performance for surface PM2.5 is attributed to the reduced overpredictions in precipitation using a more advanced MSKF cumulus scheme available in WRF/Chem v3.7.1 than the G3 cumulus parameterization in WRF/Chem v3.6.1.

We analyzed decadal regional coupled climate-air quality simulations produced by WRF-Chem for the United States for the present period (2001 - 2010) and a future period (2046 - 2055) under the RCP8.5 emissions scenario. The frequency of heat waves and stagnation was analyzed in

idually and in combination using daily outputs from the simulations. Heat waves were identified as at least three continuous days with daily maximum 2-m air temperature exceeding the 97.5th percentile of the historical period. Stagnation events were identified as days with daily mean 10-m wind speed less than 3.2 ms-1, daily mean 500 hPa wind speed less than 13 ms-1 and daily total precipitation less than 1 mm. High O3 occurrence was identified based on the daily maximum 8-hour O3 concentrations. We compared the frequency of heat waves, stagnation, and their compound events and impacts on the maximum 8-hour O3 concentration between the summer season (June, July, August) of the present and future periods. The simulations showed widespread increases in heat wave frequency in the future, with the largest increase found in the western United States. A dipole change in stagnation frequency is found, showing decreases across the southeastern United States and increases in the southwestern United States resembling the spatial pattern of precipitation changes in the future. Both heat waves and stagnation increase the maximum 8-hour O3 concentration in the eastern United States. Although heat waves have significantly larger impacts on high O3 events than stagnation, the compounded extremes have larger effects than the sum of the in

idual effects. Despite the significant increase in heat wave frequency in the future, its impact on high ozone events is subdued by the reduction in anthropogenic VOC and NOx emissions.

We have developed a new high-resolution NO2 dataset based on the standard National Aeronautics and Space Administration Ozone Monitoring Instrument (OMI) NO2 v3.0 product for the eastern United States. Our new product, derived from a CMAQ 1.33 km simulation and DISCOVER-AQ Maryland observations, yields NO2 columns that are twice as large in urban areas and NO2 columns in rural areas that are 20 - 40 percent lower. Furthermore, the new NO2 retrievals are now able to capture the magnitude of the concentration gradient between urban and rural areas, which is in better agreement with the current U.S. Environmental Protection Agency NO2 monitoring network. We also have developed a preliminary daily surface PM2.5 product at 1-km spatial resolution for 2008 in the eastern United States, which is constrained by MODIS AOD.

Future Activities:

We plan to complete the remaining diagnostic analyses and sensitivity simulations using the four regional models to finalize the models with improved inputs, formulations and configurations. The improved models will be used for final production simulations for the current 5-year period to obtain the best possible performance for air quality and human exposure studies. Those results will be provided to Project 4 to study the health impact of air pollutants under current climate and emission conditions. We will conduct for a future 5-year period using emission change factors projected by the National Energy Model System under future climate conditions to evaluate the impact of projected changes in emissions on future air quality. For the impact of climate extremes, we will compare future simulations following the RCP4.5 and RCP8.5 emissions scenarios to study the effects of emission mitigation on heat waves, stagnation and their compound effects on high O3 and PM2.5 events. We also explore the impacts of wildfires on air quality, as increases in heat waves and droughts in the future may have large influence on the frequency and spatial and temporal distribution of wildfires, with consequential impacts on air quality and human health. Finally, we plan to extend the enhanced daily PM2.5 product to 2009 - 2012 using high-resolution regional model results and compare the spatial variabilities of the enhanced satellite-derived NO2 and PM2.5 products with the high-density NO2 and PM2.5 measurements from Project 2.

Supplemental Keywords:

Global and regional modeling, air quality, O3, PM2.5, diagnostic analysis, model improvement, sensitivity simulation, climate extremes, heat wave, enhanced satellite retrievals.

Relevant Websites:

The SEARCH Center: Solutions for Energy, AiR, Climate, and Health

Progress and Final Reports:

Original Abstract
  • 2016 Progress Report
  • 2018

  • Main Center Abstract and Reports:

    R835871    Solutions for Energy, AiR, Climate and Health Center (SEARCH)

    Subprojects under this Center: (EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
    R835871C001 Project 1: Modeling Emissions from Energy Transitions
    R835871C002 Project 2: Assessment of Energy-Related Sources, Factors and Transitions Using Novel High-Resolution Ambient Air Monitoring Networks and Personal Monitors
    R835871C003 Project 3: Air Quality and Climate Change Modeling: Improving Projections of the Spatial and Temporal Changes of Multipollutants to Enhance Assessment of Public Health in a Changing World
    R835871C004 Project 4: Human Health Impacts of Energy Transitions: Today and Under a Changing World