CALPUFF versus AERMOD comparison

A Comparison of AERMOD and CALPUFF Models for
Regulatory Dispersion Modelling in the Alberta Oil
Sands Region

Garrett Hoeksema, M.Sc., Air Quality Scientist
Koray Önder, P.Eng., Senior Air Quality Engineer
Greg Unrau, M.Sc., Air Quality Meteorologist

Outline

 Background
 Methods
 Results
 Summary
 Questions

2

Background

 Air Quality modelling
 To get ground-level concentrations based on what is being input into
the atmosphere from emissions sources
 Many different models to choose from in order to obtain the ground-level
concentrations
 AERMOD, CALPUFF, SCREEN3
 Air quality models are sensitive to:
 Input Meteorological Data
 Model Switches
 The purpose of this study is:
1. Effect of different model switches or different meteorology on
individual models
2. Compare predictions from different models

3

Background – Air Dispersion Modelling
 AERMOD
 Steady-state plume dispersion model designed to predict near field
 CALPUFF
 Lagrangian Gaussian puff dispersion model for both near and far
field applications
 Both AERMOD and CALPUFF have been approved for regulatory
modelling in Alberta, including near field
 CALPUFF has historically been used in the oil sands region
 It can handle regional emissions and predict potential acid input
 Experience shows that for certain pollutants and for distances less than
10 km, results can vary considerably between the two models

4

Methods

 CALPUFF and AERMOD runs were performed for various types of
emission sources
 Steam Generator (30 m)
 Glycol Heater (9 m)
 Incinerator (100 m)
 Flare (90 m)
 Area
 Modelling was performed for SO2
Unit emission rate of 1 g/s used
 The point source was placed in a valley to observe how each model
handles terrain features
 Stack characteristics were chosen based on those typical for the type of
source found in the oil sands industry

5

Methods

 CALPUFF model runs
 Two CALPUFF runs were completed utilizing two different model
settings for each run
1. U.S. EPA default model settings were used
2. Alternate model settings were used (based on BC and Ontario
modelling guidelines)

Parameter CALPUFF Run 1 CALPUFF Run 2
MDISP 3 – PG coefficients 2 – internally
(rural) and MP calculated coefficients
coefficients (urban)
MPDF 0 – does not use PDF 1 – uses PDF for
for dispersion dispersion

6

Methods

 AERMOD model runs
 Three AERMOD runs were completed based on different
meteorological data

Meteorological AERMOD Run 1 AERMOD Run 2 AERMOD Run 3
data
Upper Air Data MM5 MM5 MM5
meteorological meteorological meteorological
data data data

Surface Air Data MM5 Fort McMurray and Fort McMurray and
meteorological Mannix Tower low Mannix tower all
data level only (20 m) levels (20 m, 45 m,
75 m)

7

Results
Source Averaging Period CALPUFF CALPUFF AERMOD AERMOD AERMOD
Run 1 Run 2 Run 1 Run 2 Run 3
Steam 1-hr 1st highest [µg/m³] 13.0 11.9 12.6 15.0 17.5
Generator
1-hr 9th highest [µg/m³] 4.7 4.3 9.4 11.7 11.0
Stack
height = Distance -direction 1.7 km 0.2 km 1.6 km 1.5 km 1.5 km
30 m NNE NE ENE E E
24-hr 1st highest [µg/m³] 1.6 2.4 2.7 2.7 2.4

Annual [µg/m³] 0.06 0.13 0.15 0.16 0.10

Glycol 1-hr 1st highest [µg/m³] 62.7 47.8 58.0 60.8 124.9
Heater
1-hr 9th highest [µg/m³] 27.1 35.0 49.4 36.9 44.2
Stack
height = Distance -direction 1.5 km 0.1 km 0.1 km 1.8 km 0.9 km
9m NNE NNW NNW S ENE
24-hr 1st highest [µg/m³] 10.8 25.8 28.3 19.9 8.5

Annual [µg/m³] 0.61 2.72 3.16 1.33 0.38

8

Results
Incinerator 1-hr 1st highest [µg/m³] 3.41 4.63 0.87 1.05 1.88
Stack
1-hr 9th highest [µg/m³] 0.87 1.13 0.52 0.64 0.92
height =
100 m Distance -direction 1.6 km 1.5 km 0.7 km 10.0 km 0.5 km
NNW N NW E WNW
24-hr 1st highest [µg/m³] 0.22 0.33 0.23 0.21 0.19

Annual [µg/m³] 0.011 0.022 0.012 0.020 0.013

Flare 1-hr 1st highest [µg/m³] 1.85 2.70 0.62 0.66 1.04
Stack
1-hr 9th highest [µg/m³] 0.54 0.71 0.41 0.66 0.51
height =
90 m Distance -direction 2.0 km 2.1 km 0.8 km 3.6 km 2.5 km
N NNE ENE WSW NW
24-hr 1st highest [µg/m³] 0.18 0.19 0.19 0.15 0.11

Annual [µg/m³] 0.007 0.010 0.008 0.011 0.007

9

Results

Area 1-hr 1st highest [µg/m³] 168.9 130.3 96.0 186.3 189.7

1-hr 9th highest [µg/m³] 147.4 114.0 92.4 162.8 172.9

Distance -direction 0.3 km 0.3 km 0.7 km 0.7 km 0.7 km
ENE ENE NW NW NW
24-hr 1st highest [µg/m³] 68.6 61.0 66.7 81.2 92.8

Annual [µg/m³] 16.7 14.1 13.3 27.9 34.8

10

Results

 CALPUFF concentrations are higher
in close proximity to the source
 Comparing the two CALPUFF runs
 Run 1 results in a higher
distribution of the pollutant
 AERMOD transports the pollutant
farther away from the source
 Comparing the three AERMOD runs
 Including surface data results in
distribution over a larger spatial
area

11

Summary

 CALPUFF and AERMOD are both approved models for use in
regulatory applications in Alberta
 Considerably variable concentrations occur, depending on the model
used and the inputs into the model
 Both models are very sensitive to model inputs and source parameters,
meteorological data and model switches
 AERMOD distributes the pollutant further away from the point source
 AERMOD is not an option for the large regional area and large number
of sources of the oil sands
 Could be used for individual sources
 CALPUFF keeps the pollution within the valley and close to the point
sour
 CALPUFF should continue to be used for large-scale regional modelling
 Alternate switches are more appropriate for near field applications
12

CALPUFF versus AERMOD comparison

More Related Content

Viewers also liked (20)

Similar to CALPUFF versus AERMOD comparison (20)

CALPUFF versus AERMOD comparison