Climate Change and Air Pollution Modeling

1
Climate Change and Air Pollution Modeling

• NCAR Summer Colloquium on Climate and Health
• July 27, 2004

Michelle L. Bell Yale University, School of
Forestry and Environmental Studies Quansong
Tong Princeton University, Woodrow Wilson School
2
Air Quality Modeling and Climate Change

• Purposes of air quality modeling
• Overview of air quality models
• Relationship to climate change research
• Examples

3
(No Transcript)
4
Air Quality Model
5
Why not use monitors?

• Measurements
• Only available at certain times and locations
• Typically geared to measuring compliance with
  regulatory standards
• Hard to explore some research questions
• Other places and time periods
• Other scenarios for emissions, meteorology, etc.

6
Comparison of Maximum O3 Concentrations
24 Monitor Measurements
2880 Model Estimates
7
Uses for Air Quality Modeling

• Regulatory purposes
• Generate concentration estimates in places
  without monitors
• Estimate public health impacts from air pollution
  policies
• Explore what-if scenarios, such as climatic
  changes

8
Gridded Atmosphere (Box Model)

• Incorporates chemical reactions, decay, physical
  transformation
• Requires initial and boundary conditions
• Assumptions include homogeneous mixing within
  the box

9
Domains and Nests
Domain 1
Domain 2
Domain 3
Domain 4
Domain 1 108-km grid cell resolution Domain 3
36-km grid cells Domain 2 12-km grid
cells Domain 4 4-km grid cells
10
Gridded Results Can Be Connected to Other Data
Example Population in areas that exceed the O3
NAAQS (using Census)
June 27-29, 1990 July 13-15, 1995
1-Hour NAAQS 1-Hour NAAQS
Population 54,126 1,685,355
of Total Population 0.66 20.7
8-Hour NAAQS 8-Hour NAAQS
Population 6,563,049 8,024,040
of Total Population 80.5 98.4
Bell and Ellis (2003), Journal of the Air Waste
Management Association 531531-40
11
Meteorology (MM5)
12
Features of Models-3

• Multi-scale can be run at different resolutions
• Multi-pollutants 95 species including O3, PM
  (including PM2.5), SO2, CO, NOx, and VOCs
• Modular
• Active development (state of the science)

13
Models-3 Framework
MM5 (Meteorological Modeling System)
CMAQ (Community Multi-scale Air Quality)
LUPROC (Land-Use Processor)
MEPPS (Models-3 Emissions Processing and
Projection System)
MCIP (Meteorology-Chemistry Interface Processor)
INPRO (Input Processor)
EMPRO (Emission Processor)
ECIP (Emissions-Chemistry Interface Processor)
CCTM (CMAQ Chemical Transport Model)
OUTPRO (Output Processor)
JPROC (Photolysis Rate Processor)
SAS
MEPRO (Models-3 Emissions Projections Processor)
ICON (Initial Conditions Processor)
Arc/Info
BCON (Boundary Conditions Processor)
IDA (Inventory Data Processor)
14
PSU/NCAR Mesoscale Model Version 5 (MM5)

• Generates gridded estimates of meteorological
  variables (e.g., temperature) to be used as input
  in the air pollution model

15
Air Quality Modeling Component I Emission Model
(SMOKE)
(SMOKE Sparse Matrix Operator Kernel Emissions
modeling system)
16
Model Output

• Estimates for each gridcell for every time period
  (e.g., hour) for each pollutant
• Ozone, particulate matter (including PM2.5),
  sulfur dioxide, others

17
Example Modeling Application
All VOC and NOx emissions increased 25
Increase in 1-hour max O3
All anthropogenic VOC and NOx emissions increased
25
18
Model Evaluation

• How good are the concentration estimates?
• What do we care about?
• Several graphical and numerical techniques
  available
• Model developers, policy-makers, and EPA
  generally do not promote rigid criteria for
  accepting or rejecting modeling results

19
Some Statistical Measures

• Mean Bias

Normalized Bias
Gross error
Unpaired highest-prediction accuracy
20
observed (monitor) concentration at location i
and time j
predicted (model) concentration at location i
and time j
number of hourly prediction-observation pairs
for location i
number of monitoring stations
number of total hours for all
prediction-observation pairs across all monitors
peak observed concentration
peak predicted concentration
21
Evaluation of Models-3 for O3
Bell and Ellis (2004) Tong (2004) EPA recommendations
Normalized bias (D) 18 6 5 to 15 ()
Normalized gross error (Ed) 28.7 21 to 25 30 to 35
Unpaired highest prediction accuracy (Au) 7.7 12 to 17 15 to 20 ()
Bell and Ellis (2004), Atmospheric Environment
381879-89
22
Comparison of Model Estimates and Measurements
(O3)
23
Model-Estimates and Measurements (O3 8-hr
average)
Location Millington, Maryland
8-hr Average O3 (ppb)
Local Time
24
PM10 Model-Estimates and Measurements (24-hr
avg. July 14, 1995)
25
More Information on Air Quality Modeling

• Models-3 (Air quality model)
• http//www.epa.gov/asmdnerl/models3/index.html
• Community Modeling Analysis System (CMAS)
• http//www.cmascenter.org/html/models.html
• MM5 (Meteorological model)
• http//www.mmm.ucar.edu/mm5/

26
Connecting Air Quality Modeling to Climate Change
and Human Health

• Can explore the impacts of future climatic
  conditions (or other future conditions) on
  ambient air quality
• Can use model results to relate changes in air
  quality to human health (e.g., epidemiological
  concentration-response functions)

27
Air Quality Model
28
Example Ozone and Climate

• Elevated mortality and hospital admissions from
  climate change induced increases in ambient ozone
  concentrations
• M.L. Bell, C. Hogrefe, C. Rosenzweig, P. Kinney,
  J. Rosenthal, K. Knowlton, B. Lynn, J. Patz
• Builds on research of the New York Climate
  Health Project (Columbia University, P. Kinney)
• http//eiwall.ei.columbia.edu/directory/displaypro
  ject.php?projectid150
• Related project for New York City region
• (Kim Knowlton)

29
How could climate change affect O3 and thereby
health?

• Hourly ambient concentrations of ground-level
  ozone
• Current climate summers of 1995 to 1997
• Future climate summers of 2053 to 2057
• Connect changes in concentration to health using
  epidemiological studies
• Only examines a piece of how climate change could
  affect (ozone air quality only)
• Only examines a piece of how elevated ozone could
  affect health (a subset of the adverse impacts)

30
Changes in Daily O3
31
Changes in Daily Ozone (ppb)
Avg Increase Min Increase Max Increase
Daily average 2.9 -0.01 6.4
Daily 1-hour max 4.8 0.51 9.6
Daily 8-hour max 4.4 0.45 9.0
32
(No Transcript)
33
Changes in O3 Air Quality Index Category
Future Climate
Current Climate
34
Human Health Impacts

• Percent Increase in Total Mortality
• (Stieb et al. 2003)
• Overall 0.28 (0.17, 0.39)
• Largest 0.43 (0.12, 0.73)
• Percent Increase in asthma-related hospital
  admissions for those lt 64
• (Sheppard et al. 2003)
• Overall 2.1 (0.6, 3.6)
• Largest 4.7 (1.4, 8.1)

35
Current Research on Climate Change Using Air
Quality Modeling

• Pacific Northwest Regional Climate Modeling
  Project
• Focuses on the Pacific Northwest and the Northern
  Midwest
• Univ. of Washington
• (Jack Chen, et al.)

www.atmos.washington.edu/salathe/reg_climate_mod/
STARCCAQ/
36
Recent Research on Climate Change Using Air
Quality Modeling

• Carnegie Mellon University
• Peter Adams, Spyros Pandis
• Development of an integrated modeling framework
  and sensitivity assessment for the effects of
  climate change and global emissions on U.S. air
  quality

http//cfpub2.epa.gov/ncer_abstracts//index.cfm/fu
seaction/display.abstractDetail/abstract/6238/repo
rt/0
37
Recent Research on Climate Change Using Air
Quality Modeling

• Georgia Institute of Technology, MIT, Northeast
  States for Coordinated Air Use Management
• Russell Armistead, M. Talat Odman, Ronald Miller,
  et al.
• EPA-funded project for sensitivity and
  uncertainty assessment of climate change on ozone
  and particulate matter

http//cfpub2.epa.gov/ncer_abstracts//index.cfm/fu
seaction/display.abstractDetail/abstract/6238/repo
rt/0
38
Recent Research on Climate Change Using Air
Quality Modeling

• New York Climate Health Project
• Patrick Kinney, Christian Hogrefe et al.,
  Columbia University
• Focuses on NY
• metropolitan region
• Estimating health
• impacts associated
• with climate change

http//eiwall.ei.columbia.edu/directory/displaypro
ject.php?projectid150 http//www.earth.columbia.e
du/events/2004/nycch.htm
39
Concluding Thoughts on Air Quality Modeling and
Climate Change

• Useful way to link changes in emissions,
  land-use, and meteorology to ambient air quality
  and then to health
• Highly resolved estimates, temporally and
  spatially
• Generally reflects state of the science
  knowledge regarding the physical and chemical
  transformation of pollutants

40
Concluding Thoughts on Air Quality Modeling and
Climate Change

• Requires time, money, and expertise
• Data input requirements
• Uncertainties
• Model evaluation
• Modeling system under different conditions
• Input of climate change scenarios into the air
  quality modeling system