Air Mass Characterization of Air Quality and Health Impacts under Current and Future Climate Scenarios Adel Hanna1, Aijun Xiu1, Karin Yeatts1, Richard Smith,1 Zhengyuan Zhu1, Neil Davis1, Kevin Talgo1, Zac Adelman1 Sarav Arunachalam1, Gurmeet

1
Air Mass Characterization of Air Quality and
Health Impacts under Current and Future Climate
ScenariosAdel Hanna1, Aijun Xiu1, Karin
Yeatts1, Richard Smith,1 Zhengyuan Zhu1, Neil
Davis1, Kevin Talgo1, Zac Adelman1 Sarav
Arunachalam1, Gurmeet Arora1,Qingyu Meng2, Scott
Sheridan3, and Joseph Pinto21The University of
North Carolina at Chapel Hill, Chapel Hill, North
Carolina 275992U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina
277113Kent State University, Kent, Ohio 44242
2
Outline

• Motivation and Objectives
• Data and Models
• Concept of Air Mass/Weather Type
• Weather Classification
• Meteorological characteristics of Air Masses
• Air Quality and Air Mass
• Statistical Modeling Approach
• Climate Scenarios and Trends in Air Mass
  Variability
• Summary and Conclusions

3
Objectives

• Define more precisely the interrelationships
  among
• changes in climate and meteorological conditions,
• air pollution, and
• heat- and cold-related morbidity severe enough to
  warrant clinical contact. .
• Examine future climate scenarios in terms of
  potential impacts on air quality and Human health

4
Data and Models

• Nine Years of data (1996 -2004)
• Meteorological Data
• The National Climatic Data Center archives of
  surface and upper-air data over the U.S.
• Air Quality Data
• AQS measurements of ambient concentrations of
  ozone,
• Health Data
• Morbidity measures include asthma and MI hospital
  admissions.
• Models (Years 2001-2003, 2018- 2020, 2048-2050)
• CCSM, WRF, CMAQ
• SMOKE (IPCC)

5
The Concept of Air Mass

• What is an air mass?
• How is it related to basic meteorological
  parameters (temperature, pressure, winds, etc.)?
• How is it different from analysis of basic
  meteoro- logical parameters?
• Source
• Duration
• Spatial coverage

6
Spatial Synoptic Classification

• Sheridan Spatial Synoptic Classification system
  (2001) (sheridan.geog.kent.edu/ssc.html)
• Classification (air mass) types
• DM Dry Moderate (mild and dry)
• DP Dry Polar (very cold temperatures advection
  from Canada)
• DT Dry Tropical (hottest and driest conditions
  at any location)
• MM Moist Moderate (warmer and more humid than
  MP)
• MP Moist Polar (cloudy, humid, and cool)
• MT Moist Tropical (warm and very humid)
• Tr Transition (one air mass giving way to
  another)
• MT Moist Tropical (upper limits of the MT)

7
Monthly frequency of seven air mass types based
on daily meteorological analyses during 1996-2004
8
Characteristics of the air mass types
9
Air Mass Ozone Characteristics

• Probability (expressed as a percentage) of
  finding O3 concentrations above a threshold
  concentration for a given air mass (P(O3AM))

Probability (expressed as a percentage) of having
a particular air mass present when O3
concentrations are above a threshold
concentration (P(AMO3))
10
Trajectory clusters of 72-hour backward
trajectories for Charlotte
DM
DP
DT
MM
MT
MP
11
MT
TR
12
Air Mass/ Air Quality

• Dry Tropical (DT), Dry Moderate (DM), and the
  Moist Tropical (MT), are always among the top
  three circulation patterns associated with the
  high Ozone concentrations. DT shows highest ozone
  concentrations.
• DT has westerly to southwesterly flow (72 hours
  back trajectory
• MT shows air traveling over the Atlantic and the
  Gulf of Mexico
• DM shows air traveling along Northwest and
  Northeast

13
Statistical Analysis-General Linear Models

• Evaluated association of ozone with asthma and MI
  hospitalizations for different air masses
• Modeling strategy
• Joint modeling of ozone and air mass
• Assumed a Poisson distribution of the outcomes,
• Checked for overdispersion
• Used B-spline function with 24 knots to adjust
  for nonlinear seasonal effect and long term
  trend.
• Adjusted for differences in dew point and day of
  the week.

14
Health Data

• Hospitalizations and ER from all of North
  Carolina (North Carolina Center for State Health
  Statistics)
• Asthma (ICD9 493.x)
• Myocardial infarction (ICD 410)

15
Percent Change/10 ppb O3 and 95 CL for
Charlotte, Raleigh and Greensboro by air mass
Asthma ER)
Asthma
16
Percent Change/10 ppb O3 and 95 CL for
Charlotte, Raleigh and Greensboro by air mass
MI
17
Health Associations

• Asthma
• Hospitalization and ER
• Ozone-dry tropical
• Current day and all lags show increase in asthma
  hospitalizations
• Ozone -Transitional and Moist Tropical (MT/MT)
  
• higher lags
• MI
• Hospitalization
• Ozone-Moist Tropical (MT and MT)
• 5 day lag

18
Future Climate Scenarios

• Examine Seasonal and Inter-annual Variability
• How to use our results as a Forecasting Tool to
  provide longer term anticipation of local air
  quality conditions (Ozone Code Red and Code
  Orange days)
• Projection of future climate patterns
• Year (2018-2020 and 2048-2050)
• CCSM/WRF/CMAQ model simulations

Research Question Does Future Climate Air Mass
Type Frequency stay the same as current Climate?
19
WRF model domains
20
Current and Future Climate Modeling Configurations

• May, June, July, and August of the years 2001,
  2002, and 2003, representing current climate
  conditions and the years 2018, 2019, 2020, 2048,
  2049, and 2050, representing future climate
  conditions.
• Dynamical downscaling of the CCSM meteorological
  outputs to provide initial and boundary
  conditions for WRF at the 36-km grid resolution,
• SRES A1B driven CCSM results used for IPCC AR4 on
  a T85 Gaussian grid.
• Constant anthropogenic emissions within each
  period and to develop hourly biogenic emissions
  using simulated meteorology data.
• For Period 1 we used the 2002 National Emission
  Inventory version 3 (NEI2002v3) from EPA for the
  United States, the 1999 National Emission
  Inventory Phase III for Mexico (MNEI99p3), and
  the 2000 National Pollutant Release Inventory
  (NPRI2000) for Canada to represent the
  anthropogenic emissions for each year during the
  period.
• For Period 2, we used the NEI2002v3-based 2020
  NEI (NEI2002v3_2020) from EPA for the United
  States, the 2018 NEI for Mexico (MNEI2018), and
  the 2020 NPRI (NPRI2020) for Canada to represent
  the anthropogenic emissions
• For Period 3, we used the year 2050 inventories
  developed at GaTech for studying how future
  climate change will impact regional air quality
  (Woo et al., 2008).

21
CCSM
22
WRF
WRF 3.0 August 2002, Surface Temperature
WRF 3.0 August 2048, Surface Temperature
23
Isoprene Emissions
24
Grid Resolution
25
Future Frequency of air masses for 108 km (dO1),
36 km (dO2), 12 km (dO3)June-July-August
26
Conclusions

• Specific air masses (DT,DM,MT) are associated
  with episodes of high ozone concentrations in
  North Carolina. Highest levels are associated
  with the DT air mass.
• Each air mass shows a distinctive meteorological
  and air quality characteristics including upwind
  source regions.
• The DT circulation pattern, in conjunction with
  ambient ozone, was most strongly associated with
  increased asthma hospitalizations while MT
• Future Climate simulations show that
  classification of air masses is sensitive to
  model resolution
• The Frequency of the DT air mass tend to increase
  in future decades 2020 and 2050
• The concept of air mass could be useful in public
  health planning by projecting pollution episodes
  and associated health impacts

27
Acknowledgments

• EPA- STAR program (Bryan Bloomer and Barbara
  Glenn), Dr. Ted Russell and Dr. Praveen Amar

R832751010
28
North Carolina Population Map

• Five Cities
• Most cities are within counties in Nonattainment
  areas (8-hour Ozone) and some (PM2.5)

29
DM and DT Air Mass
30
MT and MT Air Mass
31
Percent change Hospital Admissions (NC)
Asthma
MI