Analysis of OTC Modeling Base

1
Analysis of OTC Modeling Base

• RAMMPP Modeling Team
• Review of Application and Assessment of CMAQ in
  OTR
• November 16, 2005
• Outline
• Scatterplots of model-calculated and
  aircraft-measured O3
• Flight-by-flight box plots
• Summary
• Episode Analysis (June 11 and/or August 12) Time
  permitting

2
UMd Research Aircraft

• GPS Position (Lat, Long)
• Meteorology (T, RH, Pr, Palt)
• Carbon Monoxide (CO)
• Ozone (O3)
• Sulfur Dioxide (SO2)
• Aerosol Optical Properties
• Absorption, Bap (565 nm)
• Scattering, Bscat (450,550,700 nm)
• Particle Counts (CPC TSI 3007)
• Aerosol Particle Size (MetOne)
• 6 cuts range 0.31.0µm
• Aerosol Chemistry (PIXE Streaker)

Aircraft operated by University Research
Foundation
3
(No Transcript)
4
(No Transcript)
5
(No Transcript)
6
(No Transcript)
7
(No Transcript)
8
(No Transcript)
9
(No Transcript)
10
(No Transcript)
11
(No Transcript)
12
(No Transcript)
13
(No Transcript)
14
(No Transcript)
15
(No Transcript)
16
(No Transcript)
17
(No Transcript)
18
(No Transcript)
19
(No Transcript)
20
Summary of aircraft comparisons

• Overall (after removing fire flights),
  CMAQ-calculated O3 concentrations are 4 too
  high below 500 meters and 4-7 too low above 500
  meters.
• Model-calculated O3 on August 12th is much too
  low throughout the sampled column. However, the
  shape of column is captured ? . Bias does not
  vary with altitude.

21
Acknowledgements

• Met modeling Da-Lin Zhang, Shunli Zhang
• CMAQ modeling Jeff Stehr, Charles Piety, Gopal
  Sistla NYDEC, and a cast of thousands
• Emissions Sheryl Ehrman and Patti Castellanos
  many others
• This research was supported by Maryland
  Department of the Environment under the RAMMPP
  project

22
Summary of surface layer comparisons

• Extremely good simulation of 1-hr (8-hr) maximum
  O3. Rgt0.9.
• Good simulation of 24-hr avg PM2.5
• CMAQ-calculated O3 concentrations have high bias
  at night leading to a 10 high-bias in 24-hr avg
  O3. Biases in 1-hr and 8-hr maximum O3 are
  small.
• CMAQ-calculated 1-hr maximum O3 concentrations
  over MD are 10-15 too low during major pollution
  events.
• PM2.5 biases are small after removal of July fire
  event.
• Diurnal variation in O3 well simulated. No
  obvious (eyeball norm) temporal shift in timing
  of maxima/minima.
• Spatial pattern of O3 pollution events is well
  captured by CMAQ.
• With a few exceptions, spatial pattern of PM2.5
  events is also well captured.

23
June 11, 2002 Weather and Air Quality
850 mb Geopotential Height (11 June, 2002)
24 Hr Back Trajectory (11 June, 2002)
24
(No Transcript)
25
(No Transcript)
26
(No Transcript)
27
(No Transcript)
28
(No Transcript)
29
August 12, 2002 Weather and Air Quality
850 mb Geopotential Height (12 August, 2002)
24 Hr Back Trajectory (12 August, 2002)
30
(No Transcript)
31
(No Transcript)
32
(No Transcript)
33
(No Transcript)
34
(No Transcript)
35
Comments on Derived Variables and Figures in
CMAQ/Aircraft Comparison Plots
How is Meters Above Sea Level (MASL) calculated
from model output?

• Model-calculated O3 was saved on the lowest 16
  CMAQ layers. Following Kevin Civerolos (NYDEC)
  and Shunli Zhangs (UMD) method, the centers of
  these layers were assumed to be
• 10, 24, 68, 116, 185, 282, 398, 544, 727, 949,
  1212, 1523, 1886, 2312, 2820, 3393 Meters Above
  Ground Level (MAGL)
• To convert to meters above sea level, we added in
  the surface elevation of
• each grid point (HT_SFC). These data are
  archived as part of MM5 output.
• Therefore MASL MAGL HT_SFC

This approximation ignores spatial variations
in temperature. However, the approximation
should be satisfactory during the summer months
covered in this aircraft comparison.
36
Comments on Derived Variables and Figures in
CMAQ/Aircraft Comparison Plots
How are pressures estimated for model output?
-The curtain plots use pressure as the vertical
coordinate. -The model pressures seen in these
plots were calculated assuming a surface pressure
of 1013.24 hPa. -Aircraft pressures were read in
from the RAMMPP data sets.
A more accurate specification of surface
pressure is desirable. In the interim, future
MDE curtain plots may use MASL as the vertical
coordinate.
37
Comments on Derived Variables and Figures in
CMAQ/Aircraft Comparison Plots
How is model output extracted at the location of
aircraft measurements?
1. The longitude, latitude, pressure-altitude,
and time of individual aircaft samples are read
in from RAMMPP data sets. (RAMMPP data sets that
do not include longitude, and latitude are not
included in these plots.) 2. MASL is calculated
for gridded output. 3. The aircraft in then
flown through the model domain (x,y,z,t).
-For each aircraft sample, we determine the model
grid box that is closest in longitude (x),
latitude (y), and sample time (t). For these
values of x, y, and t, we then interpolate
linearly in height in the vertical (z) to obtain
our value.
38
Comments on Derived Variables and Figures in
CMAQ/Aircraft Comparison Plots
What is a "XX-pt data smoother?"

• Model output is extracted at the location of
  aircraft measurements.
• 2. For each flight, model output and aircraft
  measurements are sorted by
• pressure.
• 3. Successive, XX-pt moving averages are applied
  to the model output, the
• aircraft measurements, and the sampling
  pressures until averaged values
• are obtained for the entire profile. These
  values are then plotted.

39
Comments on Derived Variables and Figures in
CMAQ/Aircraft Comparison Plots
How are the "leaning tower" box plots created?

• 1. Model output is extracted at the location of
  aircraft measurements.
• 2. Model output and aircraft measurements are
  binned into 1/4 km bins using
• MASL.

-The measured-mean is shown by an asterisk.
-The mean plus and minus a standard deviation
is shown by the boxes. -The solid line shows
the mean of the model. -The thinner lines show
the model-mean plus and minus one standard
deviation.
40
Comments on Derived Variables and Figures in
CMAQ/Aircraft Comparison Plots
Which flights are flagged as "fire"
flights? Flights 24 and 25 on July 8th were
flagged as "fire" flights.
Which flights did not include longitude/latitude
data? Flights 4 and 31. Output from these
RAMMPP flights were not included in
this analysis.