A Comparative Performance Evaluation of the AURAMS and CMAQ Air Quality Modelling Systems

1
A Comparative Performance Evaluation of the
AURAMS and CMAQ Air Quality Modelling Systems

• Steven C. Smyth, Weimin Jiang, Helmut Roth, and
  Fuquan Yang
• ICPET, National Research Council of Canada,
  Ottawa, Ontario
• Michael D. Moran and Paul A. Makar
• MSC, Environment Canada, Toronto, Ontario
• Véronique S. Bouchet and Hugo Landry
• CMC, Environment Canada, Dorval, Québec

2
Outline

• Introduction
• AURAMS vs. CMAQ Differences in science, input
  file preparation, etc.
• O3, total PM2.5, and speciated PM2.5 performance
  comparison
• Summary and Conclusions

3
Introduction

• Many aspects of the AURAMS and CMAQ simulations
  were aligned to reduce some of the common
  sources of differences
• same input meteorology from Environment Canadas
  GEM model
• same raw emissions inventories processed by SMOKE
• same biogenic emissions model
• same grid resolution
• Confidence that the differences in model results
  are caused by the AQ models themselves rather
  than by meteorological and/or emissions inputs

4
Modelling Systems

• CMAQ v4.6
• SAPRC-99 chemical mechanism AERO4 NRC PMx
  post-processor
• AURAMS v1.3.1b
• A Unified Regional Air-quality Modelling System
• AQ modelling system with size- and
  composition-resolved PM
• Designed to be a one atmosphere or unified
  model in order to address a variety of
  interconnected tropospheric air pollution
  problems ranging from ground level O3 to PM to
  acid rain

5
AURAMS v1.3.1b (cont.)

• 9 PM species/components sulphate (SU), nitrate
  (NI), ammonium (AM), black carbon (EC), primary
  organic aerosols (PC), secondary organic aerosols
  (OC), crustal material (CM), sea-salt (SE), and
  particle bound water (WA)
• 12 PM size distribution bins 0.01 to 40.96 µm in
  diameter
• Bins 1 thru 8 PM2.5
• Bins 9 and 10 PMC where PM10-PM2.5 PMC
• Bins 11 and 12 PM greater than 10 µm in
  diameter
• Gas phase chemistry modified version of ADOM-II
• Includes sea-salt emissions but not chemistry at
  this time
• Zero-gradient lateral boundary conditions

6
Domains and Simulation period

• AURAMS - Polar Stereographic true scale at 60N
  150 x 106 grid 42-km resolution
• CMAQ - Lambert conformal conic standard
  parallels of 50N and 70N 139 x 99 grid 42-km
  resolution
• 0100 July 1, 2002 to 0000 July 30, 2002 UTC

7
Model Inputs - Meteorology

• GEM v3.2
• AURAMS meteorological pre-processor
• GEM-MCIP (based on MCIP v3.1)
• Overlapping grid cell comparison of surface
  fields NMEs of 0.25 for pressure 0.4 for
  temperature 3.8 for specific humidity (HU)

8
Model Inputs - Emissions

• SMOKE v2.2
• Canadian Emissions
• 2000 CAC inventory
• U.S. Emissions
• 2001 CAIR
• Mexican Emissions
• 1999 inventory
• Biogenic Emissions
• BEISv3.09
• AURAMS online
• CMAQ offline using SMOKE

9
Model Inputs - Emissions (cont.)

• Point source processing
• AURAMS plume-rise of major point sources
  calculated within CTM
• CMAQ meteorological data used to calculate
  plume rise within SMOKE
• Emissions files
• AURAMS grams/sec
• representative week of emissions for each month
  of simulation
• 3 emissions files (non-mobile, mobile,
  minor-point) in RPN format
• 1 emissions file (major-point sources) in ASCII
  format
• CMAQ gaseous - moles/sec PM - grams/sec
• daily emissions files
• single comprehensive file in I/O API format

10
Measurement Data

• O3 - hourly measurements from the EC NAPS
  network (190 sites) and U.S. EPA AQS network
  (1087 sites)
• PM2.5 - hourly measurements from NAPS (92 sites)
  and AQS (262 sites)
• Speciated PM2.5 - daily averaged measurements
  from NAPS (17 sites) and U.S. EPA STN network
  (205 sites)

O3 Measurement Sites
PM Measurement Sites
11
O3 Performance
statistics O3 (ppb) O3 (ppb) daily peak O3 (ppb) daily peak O3 (ppb) daily low O3 (ppb) daily low O3 (ppb)
statistics AURAMS CMAQ AURAMS CMAQ AURAMS CMAQ
meas. mean 35.6 35.6 60.8 60.7 11.7 11.7
mod. mean 42.0 51.4 66.7 70.2 16.4 32.7
MB 6.4 15.8 5.9 9.4 4.6 20.9
NMB () 18 44 10 16 39 178
ME 16.2 18.8 16.4 14.3 11.0 21.9
NME () 46 53 27 24 94 187
r2 0.393 0.438 0.350 0.505 0.104 0.077

• AURAMS lower bias
• Similar levels of error
• CMAQ over prediction mainly due to inability in
  predicting daily lows

12
O3 Performance (cont.)

• Both AURAMS and CMAQ over-predict daily peaks
• AURAMS much better at predicting daily lows
• Both models show correct diurnal patterns and
  overall trends in concentration level

13
Total PM2.5 Performance
statistics total PM2.5 (µg m-3) total PM2.5 (µg m-3) daily peak PM2.5 (µg m-3) daily peak PM2.5 (µg m-3)
statistics AURAMS CMAQ AURAMS CMAQ
meas. mean 14.4 14.4 28.3 28.1
mod. mean 12.2 5.1 21.6 9.1
MB -2.2 -9.3 -6.6 -19.0
NMB - 15 - 64 - 23 - 68
ME 9.7 10.2 16.3 19.5
NME 67 71 58 69
r2 0.074 0.151 0.040 0.086

• AURAMS lower bias
• Similar levels of error

14
Total PM2.5 Performance (cont.)

• Both models under-predict PM2.5
• Forest-fires not included in emissions
  contributes to under-prediction in both models
• Much more PM2.5 sea-salt in AURAMS

15
PM2.5 Species Performance
statistics PM2.5 SO4 (µg m-3) PM2.5 SO4 (µg m-3) PM2.5 NO3 (µg m-3) PM2.5 NO3 (µg m-3) PM2.5 NH4 (µg m-3) PM2.5 NH4 (µg m-3)
statistics AURAMS CMAQ AURAMS CMAQ AURAMS CMAQ
meas. mean 5.0 5.0 0.90 0.90 1.6 1.6
mod. mean 5.3 2.4 2.0 0.26 1.6 0.82
MB 0.3 -2.6 1.1 -0.64 0.0 -0.81
NMB () 6 -51 121 -71 1 -50
ME 3.0 2.8 1.6 0.7 0.9 0.9
NME () 60 56 175 79 54 57
r2 0.367 0.524 0.397 0.437 0.428 0.458

• AURAMS better bias for SO4 and NH4 similar
  levels of error
• CMAQ better correlation

16
PM2.5 Species Performance (cont.)
statistics PM2.5 EC (µg m-3) PM2.5 EC (µg m-3) PM2.5 TOA (µg m-3) PM2.5 TOA (µg m-3)
statistics AURAMS CMAQ AURAMS CMAQ
meas. mean 0.52 0.52 10.6 10.6
mod. mean 0.28 0.34 3.9 0.9
MB -0.24 -0.18 -6.6 -9.6
NMB - 46 - 36 - 63 - 91
ME 0.30 0.30 6.9 9.6
NME 58 58 66 91
r2 0.166 0.204 0.0005 0.007

• CMAQ better bias for EC similar levels of error
• AURAMS much better performance for TOA
• Due to difference in SOA algorithms

• Poor TOA correlation for both models impacts
  overall correlation for total PM2.5 (AURAMS
  0.074 CMAQ 0.151)

17
PM2.5 Species Temporal Comparison
SO4
NO3
NH4
EC
SOA
POA
Other PM2.5
Sea-salt
18
PM2.5 Species Spatial Comparison
19
PM2.5 Species Spatial Comparison (cont.)

• Similar spatial and temporal patterns for most
  species
• Sea-salt aerosols vastly different
• Concentration levels quite different

20
PM Composition

• PM2.5 sea-salt contributes over half to AURAMS
  total PM2.5 mass for all grid cells only 5 in
  CMAQ
• For land grid cells only, PM2.5 sea-salt
  contributes 15 in AURAMS and 2 in CMAQ
• If sea-salt is excluded from model results, total
  PM2.5 performance still better in AURAMS results

PM composition avg. over all grid cells
PM composition avg. over land grid cells only
21
Discussion and Summary

• Similar levels of error for O3, total PM2.5, and
  most PM2.5 species
• AURAMS better bias for all species except PM2.5
  nitrate and elemental carbon
• Enhanced AURAMS bias due to cancellation of
  positive and negative biases
• Sea-salts contribute much more to overall PM
  composition in AURAMS than CMAQ
• Does not impact overall conclusions regarding
  relative PM2.5 performance of the models
• Spatial and temporal patterns similar, but
  overall concentration levels quite different

22
Acknowledgements

• Radenko Pavlovic and Sylvain Ménard of
  Environment Canada
• transfer of AURAMS code and help in understanding
  and compiling various AURAMS related material
• Wanmin Gong of EC
• help in identifying problem in AURAMS land-use
  file
• Pollution Data Division of EC
• 2000 Canadian raw emissions inventories
• U.S. EPA and CMAS
• U.S. emissions data, SMOKE, CMAQ, MCIP
• Colorado State University
• VIEWS database for measurement data
• Meteorological Service of Canada
• NAtChem database for measurement data
• Environment Canada and the Program of Energy
  Research and Development (PERD) for funding
  support

23
(No Transcript)