Simulating the Temporal Evolution and Spatial Distribution of Urban Air Pollutants to 2030

1
Simulating the Temporal Evolution and Spatial
Distribution of Urban Air Pollutants to 2030
Malcolm O. Asadoorian, Ph.D.

• Global Air Pollution Trends Up To 2030
• International Institute for Applied Systems
  Analysis (IIASA)
• Vienna, Austria on January 27-28, 2005

2
Motivation

• Accurately projecting the geographic distribution
  of urban air pollutants is of critical importance
  for the atmospheric chemistry component of global
  climate change models and health impact studies.
• It is essential that a model allow population
  density to change over time to identify growth of
  urban areas and use this information to
  geographically distribute emissions.
• The model discussed in this presentation fulfills
  this goal, using Nitrogen Oxides (NOX) as an
  illustrative example.

3
Outline of Presentation

• Introduction MITs Integrated Global System
  Model (IGSM).
• Methodology - Temporal Evolution
• Overview - Beta Distribution.
• Details - Coupled Beta Distribution and
  Computable General Equilibrium (CGE) Economic
  Model.
• Example - Projected Probability Distribution of
  Population Density for Central European
  Associates (EET) until 2030.
• Application - Spatial Distribution
• Important Remarks Regarding CGE Emissions
  Projections.
• Regional-Aggregate Global Distribution of NOX
  1997 versus 2030.
• 1 X 1 Global Distribution of NOX for 1997.
• 1 X 1 Global Distribution of Change (?) in NOX
• 1997 versus 2030.

4
MIT Integrated Global System Model (IGSM) Version
2
For more info http//web.mit.edu/globalchange/www
/
5
Methodology Overview Beta Distribution

• Need a flexible function form to geographically
  describe the probability distribution of
  population density.
• The Beta Distribution fulfills this need given
  that its inherent flexibility allows the
  distribution to be significantly skewed right,
  left, etc. The Beta probability density function
  (pdf) has two shape parameters, ? and ?, and is
  given by
• I can estimate the shape parameters, ? and ?,
  using Maximum Likelihood estimation and then
  express those estimates as functions of economic
  variables. More specifically, I estimate how they
  change with changes in Gross National Product Per
  Capita (GNPPC) and National Population Per Unit
  of Land Area (POPLAND). In other words,
• ? f ( GNPPC, POPLAND)
• ? f ( GNPPC, POPLAND)

6
Methodology Details Coupled Beta Distribution
CGE Model

1. Estimate a Beta Distribution for each of the 16
   regions in EPPA using a 1990 1 X 1 latitudinal
   spatial population density data set.
2. Estimate ? and ? as functions of economic
   variables (previously mentioned) using a
   cross-sectional log-linear regression.
3. Forecast probability distribution of population
   density for each EPPA region based on EPPA
   projections on GNPPC and POPLAND.
4. Projected probability distribution of population
   density is the basis for distributing emissions
   projections produced by the EPPA CGE Model.

7
Example Projected Probability Distribution of
Population Density for EET until 2030
Legend Black 1997 Green 2010 Blue
2020 Red 2030
8
Spatial Distribution Important Remarks
RegardingEPPA CGE Model Emissions Projections

• Emissions Projections for each of the 16 EPPA
  regions are determined through the interaction of
  a relatively large number of economic variables
  within the CGE framework.
• There exist Emissions Coefficients which vary by
  fuel, sector, region, and over time.
• In general, Emissions Coefficients indicate how
  emissions vary with economic growth.
• Differences in economic growth, energy
  production, energy consumption, energy prices,
  and other factors will collectively determine the
  magnitudes of the Emissions Projections.
• See Babiker, et al. 2001. The Emissions
  Prediction and Policy Analysis (EPPA) Model
  Revisions, Sensitivities, and Comparisons of
  Results. (Report No. 71)
• At http//web.mit.edu/globalchange/www/

9
Spatial Distribution Regional-Aggregate Global
Distribution of NOX (Actual) 1997 versus
(Projected) 2030
10
Spatial Distribution 1 X 1 Global Distribution
of NOX for 1997
-180
-160
-140
-120
-100
-80
-60
-40
-20
0
20
40
60
80
100
120
140
160
180
80
80
60
60
40
40
20
20
0
0
-20
-20
-40
-40
MMT OF NOx
-60
-60
0.00 - 0.09
0.09 - 0.24
-80
-80
0.24 - 0.40
-180
-160
-140
-120
-100
-80
-60
-40
-20
0
20
40
60
80
100
120
140
160
180
0.40 - 9.99
11
Spatial Distribution 1 X 1 Global Distribution
of ?NOX 1997 versus 2030
-180
-160
-140
-120
-100
-80
-60
-40
-20
0
20
40
60
80
100
120
140
160
180
80
80
60
60
40
40
20
20
0
0
-20
-20
-40
-40
MMT OF ?NOx
-60
-60
-0.099 - -0.059
-0.059 - -0.019
-0.019 - 0.020
-80
-80
0.020 - 0.060
0.060 - 0.099
-180
-160
-140
-120
-100
-80
-60
-40
-20
0
20
40
60
80
100
120
140
160
180
0.099 - 2.137