An Investigation of the Limitations in Plume Rise
Models used in Air Quality Forecast Systems
M. Wade, F.Y.
1
Leung ,
K.
2
Christian
1. Laboratory of Atmospheric Research, Dept of Civil and Environmental Engineering, Washington State University; 2. Dept. of
Geography, University of Idaho
•In ArcMap, used USGS Landfire data to assign
geographic features to WRAP & FEPS matches.
•Landfire data provided on 30 meter grid
•MISR data: 275 meter grid
•Graphed error in plume rise vs. each geographical
feature.
•Error in plume rise was shown using:
Modeled–MISR plume heights, Modeled/MISR
plume heights, & Modeled and Miser plume heights
shown together.
Air Quality forecasting systems are important for
predicting pollutants regulated by National Ambient
Air Quality Standards (NAAQS). NAAQS pollutants,
include CO, NO2, PM2.5, PM10, O3, and SO2, are
considered deleterious to public health and air
quality. When a wildfire occurs, plume rise models
are needed in the air quality forecast systems to
predict the surface concentrations of these
pollutants. Inaccuracies in the plume rise model
lead to forecasting errors.
10000
Chart 4: MISR & WRAP
plume height vs. Elevation
9000
8000
7000
Results
6000
5000
MISR
4000
WRAP
3000
1000
0
11
23
36
47
61
70
90
115
186
255
322
548
793
921
1056
1214
1314
1404
1503
1572
1652
1792
1969
2108
2280
2438
17124
0
Figure2: A 300m by
approximate 300m rectangle
showing Landfire slope data.
Past research showed that there may be a
correlation between slope and plume height error.2
However when more points were analyzed the
linear correlation showed very low R2 values. These
results were also true for aspect.
Method
2
1.8
Plume height (meters)
1.6
y = 0.0135x + 1.1843
R² = 0.0728
1.4
1.2
1
Smoke/MISR
0.8
Chart1: Past research data
showing Smoke/MISR plume
height vs. slope
Chart 4: MISR & FPES plume
height vs. Elevation
FEPS
10000
9000
8000
7000
6000
5000
MISR
4000
FEPS
3000
2000
1000
0
0
11
23
36
47
61
70
90
115
186
255
322
548
793
921
1056
1214
1314
1404
1503
1572
1652
1792
1969
2108
2280
2438
17124
The objective of this project is to find causes for
error in the Western Regional Air Partnership
(WRAP) and Fire Emission Production Simulator
(FEPS) plume rise models, by comparing observed
and modeled plume heights, and investigating the
relationships between model error and various
geographical features.
Geographical Features:
•Slope •Elevation •Aspect •Vegetation •Canopy
Elevation (meters)
Plume height (meters)
The Landfire data is provide in 30m by 30m grid.
The MISR data is provided in a larger 275m by
275m grid. When a MISR fire is located it can be
anywhere in the grid 275m by 275m, meaning
there could be multiple geographical values for that
fire. If these values vary greatly they can add to
error in our analysis.
Objective
Elevation (meters)
The canopy data we found had been normalized and
made it difficult to find variations between multiple
points. The vegetation data provided different
vegetation for most points. This made it hard to
determine which type of vegetation would lead to
an error in plume height.
Future Work
•Using vegetation and canopy data we will look
further into the fuel loading for the wildfires,
because both can be sources of fuel for wildfires.
•Make changes to the models to limit error.
Linear (Smoke/MISR)
0.6
0.4
Acknowledgements
0.2
0
0
5
10
15
20
25
30
slope
3
1.5
1.3
2.5
PLume height (m)
WRAP/MISR
0.9
Linear
(WRAP/MISR)
0.7
0.5
y = -0.0035x + 1.1192
R² = 0.0025
0.3
2
1.5
FEPS/MISR
Linear (FEPS/MISR)
1
y = 0.0045x + 1.3494
R² = 0.0019
0.5
0.1
-0.1 0
Plume height (m)
1.1
Figure 1: A map of
the 163 wildfire
matches’ locations
WRAP
2000
Modeled plume heights are based on Briggs Plume
Rise equations, which were originally created in the
1960’s, to predict plume rise and dispersion for tall
industrial smoke stacks. Errors arise because
wildfires have very different characteristics.
•Obtained wildfire plume height data from both
WRAP and FEPS models from 2005 to 2007.
•Obtained the wildfire elevations with Google Earth .
•Matched WRAP & FEPS wildfire data to MISR
wildfire data, based on date and location.
•Checked matches using ArcMap software
In both the WRAP & FEPS models elevation is
assumed to be equal to 0. To correct for this
elevation was added to the plume heights. Our
data showed that error in plume height did increase
with increase elevation.
Plume height (meters)
Introduction
0
10
20
30
40
Slope (degrees)
Chart 2: WRAP/ MISR plume
height changes graphed against
slope
0
10
20
30
40
Slope (degrees)
Chart 3: FEPS/ MISR plume
height changes graphed against
slope
This work was supported by the National Science
Foundation’s REU program under grant number
ATM-0754990 . A special thank you to Sean Raffuse
for providing the WRAP & FEPS model data and to
Joe Vaughan, W.S.U., for providing AIRPACT data.
Reference
1. Raffuse, S. An Evaluation of Modeled Plume Rise With Satellite
Data. 8th CMAS Confrence, Oct 2009
2. Christian, K. Evaluation of the characteization of smoke plumes
within the AIRPACT model. Summer2009