Understanding the USEPA

1
Understanding the USEPAs AERMOD Modeling System
for Environmental Managers
Model Evaluation

• Ashok Kumar
• University of Toledo
• akumar_at_utnet.utoledo.edu

2
Evaluation Studies on AERMOD

• USEPA
• Evaluation using field studies
• No evaluation using ambient air monitoring
  network in an urban area

3
Lucas County Sources and Monitoring Stations
4
Input Data Flow in AERMOD
5
Data Requirements for Model Evaluation

• Emission Inventory
• Properties of stacks and super stacks
• Meteorological data
• Receptor data
• Air monitoring data

6
AERMET - Input

1. Meteorological Input Parameters Multi-Level WS,
   WD, and Temperature, Opaque Cloud Cover, Ceiling
   Height, RH, Pressure, Surface Heat Flux,
   Friction Velocity, and Roughness Length, Delta-T
   , Solar Radiation, Upper Air Data
2. Data Formats - CD144, SCRAM, SAMPSON (surface
   data)
3. TD 3280 (surface data )
4. TD6201 (upper air data)
5. On-site (site specific data)

7
AERMET - Output

• Boundary Layer File
• sensible heat flux
• surface friction velocity
• convective velocity scale
• potential temp. gradient above mixing height
• convectively-driven mixing height
• mechanically-driven mixing height
• Monin-Obukhov length
• surface roughness length
• Bowen ratio
• albedo
•  

• Profile File
• Measurement height
• WD, WS
• Temperature
• Standard Dev. of Lateral WD
• Standard Dev. of Vertical WS

8
Atmospheric Stability

• AERMOD uses Monin-Obukhov length as the stability
  parameter
• You will need friction velocity uand the flux of
  sensible heat H to compute L
• L is defined to be negative in convective
  conditions and positive in stable

9
AERMAP

• Input data needs for AERMAP
• DEM formatted terrain data
• User provided receptors and terrain
• Design of receptor grid AERMAP accepts either
  polar, cartesian or discrete receptors

10
Pathways Used in AERMOD Input Runstream

• Control
• Source
• Receptor
• Meteorology
• Output

11
Statistical Evaluation Methods

• Fractional Bias (FB)
• Normalized Mean Square Error (NMSE)

12
Statistical Evaluation Methods

• Coefficient of Correlation (COR)
• Factor of Two (Fa2)
• Fraction of data for which 0.5ltCp/Colt2

13
Statistical Evaluation Methods

• Confidence Limits
• Confidence limits are the lower and upper
  boundaries / values of a confidence interval,
  that is, the values which define the range of a
  confidence interval. The upper and lower bounds
  of a 95 confidence interval are the 95
  confidence limits.
• Q-Q Plots
• The quantile-quantile (Q-Q) plot is a graphical
  technique for determining if two data sets come
  from populations with a common distribution.

14
Results Discussion

• Model is evaluated in the following ways
• Performance measures and confidence limits for
• 3-hr average for SO2
• Stable Condition
• Convective Condition
• 24-hr average for SO2
• Plots of NMSE vs.FB
• Q-Q plots for 3-hr average for S02
• Q-Q plots for 24-hr average for S02

15
NMSE vs FB plots (3-hr average)
16
NMSE vs. FB plots(3-hr average)
17
NMSE vs. FB plots (3-hr average)
18
NMSE vs. FB plots (3-hr average)
19
NMSE vs. FB plots (24-hr average)
20
Q-Q plots (3-hr average)
21
Q-Q plots (3-hr average)
22
Q-Q plots (3-hr average)
23
Q-Q plots (3-hr average)
24
Q-Q plots (3-hr average)

• Observed Concentrations lt 20µg/m3

25
Q-Q plots (3-hr average)

• Observed Concentrations lt 20µg/m3

26
Q-Q plots (3-hr average)

• Observed Concentrations lt 20µg/m3

27
Q-Q plots (3-hr average)

• Observed Concentrations lt 20µg/m3

28
Q-Q plots (24-hr average)
29
Q-Q plots (24-hr average)
30
Confidence Limits (3-hr average)

• The values of NMSE and FB were significantly
  different from zero in the stable case. COR was
  not significantly different from zero.
• The values of NMSE, FB, and COR were not
  significantly different from zero for the
  convective case.

31
Confidence Limits (24-hr average)

• The values of NMSE and FB were significantly
  different from zero. COR, was not significantly
  different from zero.
• Note 24-hr data were not divided according to
  stability classes.