Integrated Assessments: Using CMAQ/CAMx to Evaluate Health Impacts of Air Pollution in the United States and China

1
Integrated Assessments Using CMAQ/CAMx to
Evaluate Health Impacts of Air Pollution in the
United States and China

• Denise L. Mauzerall
• Xiaoping Wang
• Quansong Tong
• Science, Technology and Environmental Policy
  program
• Woodrow Wilson School
• Princeton University

Presented at Models-3 Workshop, RTP, NC, Oct. 28,
2003
2
Overall Objectives

• Utilize science to inform air quality policy
• Methodology
• Use atmospheric modeling and available data to
  describe air quality
• Evaluate impacts of air pollution on human health
  and agriculture using exposure-response functions
  from literature
• Monetize the costs of the impacts
• Examine alternative energy/pollution control
  technologies and policies for optimal
  cost-effective air quality control strategies

3
Ongoing Integrated Assessment Projects at
Princeton Using Models-3/CMAQ

• Evaluate effectiveness of the current NOx
  emissions cap-and-trade program in the
  North-Eastern United States on reducing surface
  O3 levels and resulting health effects.
• Evaluate impact of air pollution on health in the
  Shandong region of China.
• Evaluate long-term impacts of air pollution on
  health and the environment in the United States.

4
Integrated Assessment Approach
Technology
SMOKE
MM5 / RAMS
Pollutant Emissions
CMAQ / CAMx
Ambient Concentrations
Policy Control Options
Population Distribution
Human Exposure
Epidemiological Exposure-Response Functions
Health Effects
Economic Analysis
Social Benefits
5
Charging NOx Emitters for Health DamagesAn
Exploratory Analysis

• Denise L. Mauzerall,
• Babar Sultan, David Bradford
• Princeton University

6
Outline

• Description of NOx Emissions Trading in NE United
  States
• Evidence that NOx Cap-and-Trade program has
  failed to reduce surface O3 concentrations
• Free trades permitted between May September
  regardless of
• temperature variability
• Biogenic hydrocarbon emissions
• Population density
• Modeling to answer questions of how O3 production
  and mortalities changes with
• high / low temperatures on the days of NOx
  emission?
• Regions of high / low isoprene emission?
• Regions of high / low population density?

7
OTC NOx Budget Cap and Trade Program

• Goal Reduce summer NOx emissions within 13
  north-eastern U.S. states in order to attain
  NAAQS for surface O3 (85 ppb over 8-hour
  average).
• Regulatory Approach
• Limit total NOx emissions from stationary sources
  such as power plants and industrial boilers, but
  permit trading among emission sources.
• Program capped summertime NOx emissions in 1999
  at less than half of the 1990 baseline emissions
  of 490,000 tons.
• Question Do O3 concentrations in the region
  decrease after the emissions cap-and-trade
  program is put in place in summer 1999?

8
EPA-AIRS O3 Data In 13 eastern U.S. OTC states
May - September Black 1995-1998
(pre-cap) Red 1999-2001 (post-cap)
There is no significant reduction in
O3 concentrations after the cap-and-trade
program is in place!
9
Modeling Analysis

• For each scenario, two CAMx simulations are
  conducted for July 7-18, 1995
• Standard simulation with all regional emissions.
• Perturbation simulations (NOx emissions from
• individual power plants are reduced by a fixed
  amount) during conditions of
• high/low temperature,
• high/low isoprene,
• High/low population.
• Calculate DO3 resulting from difference between
  standard run and each perturbation run
• Estimate change in mortality resulting from
  change of O3 distribution and exposed population.
  

10
(No Transcript)
11
High Temperature Period
Total Increase in Mortality 0.17
Total O3 increase 36 ppb
NOx emissions 1.77 x 106 moles from one power
plant in 24-hours results in the following change
in O3 concentration and mortalities. Mean Temp
302 K /- 3.6 K (83.5 F)
12
Low Temperature Period
Total Increase in Mortality 0.0
Total O3 increase 22 ppb
NOx emissions 1.77 x 106 moles from one power
plant in 24-hours results in the following change
in O3 concentration and mortalities. Mean Temp
296 K /- 4.3 K (72 F)
13
High Isoprene Emission Region
Total Increase in Mortality 0.13
Total O3 increase 170 ppb
NOx emissions 1.77 x 106 moles NOx from one
power plant in 24-hours results in the following
change in O3 concentration and mortalities.
14
Low Isoprene Emission Region
Total Increase in Mortality 0.09
Total O3 increase 46 ppb
Change in O3 concentration and resulting change
in Mortality for unit change in NOx emissions
from a single power plant in 24-hours. DNOx
1.77 x 106 moles NOx .
15
High Population Region
Total Increase in Mortalities 0.21
Total O3 increase 40 ppb
Change in O3 concentration and resulting change
in Mortality for unit change in NOx emissions
from a single power plant in 24-hours. DNOx
1.77 x 106 moles NOx
16
Low Population Region
Total Increase in Mortalities 0.19
Total O3 increase 121 ppb
Change in O3 concentration and resulting change
in Mortality for unit change in NOx emissions
from a single power plant in 24-hours. DNOx
1.77 x 106 moles NOx
17
Conclusions on Emissions Trading

• NOx emissions in locations of high temperature
  and high isoprene emissions result in higher O3
  concentrations.
• NOx emissions near regions of high population
  result in greater mortalities.
• Controls on temporal and spatial location of NOx
  emissions are critical, even within the
  May-September O3 season, to reduce damage
  resulting from increased O3 concentrations.

18
Future Work

• Estimate economic costs of mortalities and
  morbidity
• Evaluate feasibility of
• charging emitters for the damage they cause
• adjusting number of pollution permits available
  in locations of high ozone production potential
  and large population density.
• Examine ability of chemical weather forecasting
  system to predict future damage from emissions so
  that emitters may adjust their emissions to
  minimize their total costs.

19
Evaluating Adverse Health Impacts of Air
Pollution in Eastern China-- the Price of Clean
Air
Xiaoping Wang, Denise Mauzerall Princeton
University Yongtao Hu, Armistead Russell Georgia
Institute of Technology
20
Questions to be addressed

• How much health damage does current air
  pollution cause in eastern China based on year
  2000 emissions using conventional energy
  technologies?
• To what extent can alternative energy and
  pollution control technologies mitigate air
  pollution damage to human health?
• What are the conditions and constraints for
  China to adopt a coal gasification-based energy
  system in both the near- and long- term?

21
Integrated Assessment Approach

• Link emissions from a particular energy end-use
  sector/activity to the health effects.
• Four parts
• I. Develop high-resolution regional
  anthropogenic emission inventory
• II. Simulate ambient concentrations of PM and
  gaseous species, including secondary PM using
  Models-3/CMAQ
• III. Estimate physical health impacts
  associated with air pollution exposure
• IV. Quantify costs resulting from health
  effects of ambient air pollution exposure

22
PART I Emission Inventory Development
23
Emission Inventory (1)
Eemissions Aactivity
rate EFemission factor jspecies
kmunicipality lsector mfuel or
activity type nabatement technology
8 species(j) CO, NH3, SO2,NOx,PM10,PM2.5,NMVOC,CO
2 87 municipalities (k), 72 source categories
(lmn)
24
Emission Inventory (2)
25
Emissions by Sector













































26
II Model Simulations
27
Model Boundary and Location of the Case Region
Core region
CMAQ domains (solid squares) and MM5 domains
(dashed squares)
28
Model configuration

• Map projection Lambert Conformal, central
  meridian 116º, latitude of origin 35º, 1st
  standard parallel 25º, 2nd standard parallel
  47º
• Model domains two with 12km and 36km resolutions
• MM5 54 54 grid cells for the 36km domain,
  6072 grid cells for the 12km domain, 34 vertical
  layers from surface to 100mb
• SMOKE and CMAQ 4848 grid cells for the 36km
  domain, 5466 grid cells for the 12km domain, 13
  vertical layers
• Carbon Bond 4 ae3 aq chemical mechanism
• Running periods January, April, July, and
  October 2-18, 2000.

29
Midnight
1pm
Emissions 4/8/2000
1) Mobile
2) Area
3) Total
30
1pm
Midnight
4/8/2000
Air Temp
Wind
ASO4
31
Comparison of Simulated Concentrations with
Observations
32
PART III Health Impact Analysis
33
Methods for Health Impact Analysis

• ?cases Iref Pop exp (??c)-1
• ?Cases annual change in mortalities or
  morbidities
• resulting from air pollution exposure
• Iref annual baseline mortality /morbidity rate
  
• Pop size of affected population
• ? relative risk per unit change in
  concentrations
• ?c annual change in ambient concentrations

34
Changes in PM2.5 and mortality associated with
anthro emissions
Population
DPM2.5 in 2000
Domain Population 281 million Total
PM2.5-related chronic mortality 840,000 deaths
Chronic mortality associated with DPM2.5
35
PM2.5-related chronic mortality by sector





36
Further Work

• A. Conduct an economic analysis
• B. Construct two alternative energy technology
  and emission scenarios for the year 2020 in the
  study region
• C. Conduct a sensitivity analysis for health
  impacts to various input variables such as
  emissions and exposure-response coefficients

37
Summary

• Developed an anthropogenic emission inventory for
  the Shandong region of China.
• Conducted MM5/SMOKE/CMAQ simulations and
  comparison of calculated and observed pollutant
  concentrations for the region
• Calculated health damages associated with the
  year 2000 pollution levels
• Demonstrated the benefit of using an integrated
  approach for examining the environmental
  externalities associated with human activities
• Will examine potential gains in health benefits
  by adopting advanced, cleaner energy technologies
  in 2020

Stay Tuned!
38
An Integrated Assessment ModelArchitecture,
Development and Application
Presented at CMAS Models-3 Workshop, RTP, NC,
Oct. 28, 2003

• Quansong Tong and Denise Mauzerall
• Princeton University
• Robert Mendelsohn
• Yale University

39
Overall Objectives

• Quantify the damage caused by air pollution over
  the continental U.S.
• Monetize these damages so that their impact can
  be included in national economic accounting.
• Develop a user-friendly integrated assessment
  model for community use.

40
Model Framework
Strategy Design Module
Meteorology Module
Emission Module
Chemistry Transport Module
Population Dose Response
Health Module
Socio-economic Module
Reanalysis
Visualization
41
Integrated Assessment ModelModel Flow Chart
Input
Output

• For streamline above,
• Provide interface to utilize state-of-the-art
  models
• Develop needed additional modules based on
  literature review

42
Current Model Components

• Emission Module SMOKE version 1.5
• Meteorology module MM5 version 3.5
• Chemical transport module CMAQ 4.3
• Health module Under construction
• Economic module Yale Microeconomics
• Others SAS/S-Plus, arcGIS, JSP/ Servlet
  container, etc.

43
Health Module

• Current version includes an exposure response
  function for O3
• PM2.5 and SO2 exposure-response functions are
  being added
• Plan to include estimates of visibility losses,
  material damage, ecosystem losses, and crop
  losses in future versions.

44
Hardware Software

• 32-node Beowulf Linux cluster
• 2 x 2.4 GHz / 512K cache Xeon 2GB DDR, 200 MHz
  RAM
• 1 TB SCSI hard disks 1 fast-Ethernet and 2
  Gigabit switches
• Portland and Intel Fortran/C/C compiler
• MPICH/PVM compiled with PGF/Intel
• openPBS job scheduler
• MPIEXEC workload manager

45
Scalability Analysis
46
(No Transcript)
47
Integrated Assessment ModelHardware/Software
Approach
48
Integrated Modeling SystemPotential Benefits to
the Community

• A platform to bridge different scientific
  disciplines with end-users outsides the
  community
• Make use of state of the art models provided from
  different scientific communities
• Provide the full functions to users without
  locally installing the system.
• User friendly and transparent to those not model
  experts
• Centralized data storage and better utilization
  of model data
• Enhanced concept of One-Community modeling!

49
Summary / Future Directions

• We have incorporated the latest research from
  atmospheric science, epidemiology and economics
  into an integrated framework useful for informing
  policy.
• Plan to make the integrated assessment modeling
  system more user-friendly and flexible to
  facilitate additional applications.
• Potentially make the model available for use by
  the community.
• Challenge is to minimize uncertainties involved
  in individual model components in order to
  maximize the utility of the coupled integrated
  assessment model.