Neil Wheeler, Kenneth Craig, and Clinton MacDonald

1
Innovative Methods for Evaluating Meteorological
Model Performance during the Central California
Air Quality Studies
Neil Wheeler, Kenneth Craig, and Clinton
MacDonald Sonoma Technology, Inc. Petaluma,
California Presented at the Sixth Annual
Community Modeling and Analysis System (CMAS)
Conference October 1-3, 2007 Chapel Hill, North
Carolina
STI-3229
2
Introduction

• Prior Measurement, Analysis, and Modeling Studies
• The Question
• The Central California Air Quality Studies
  (CCAQS)
• California Regional PM10/PM2.5 Air Quality Study
  (CRPAQS)
• Central California Ozone Study (CCOS)

3
The Central Valley of California
4
Central California Air Quality Studies

• Multi-year
• Meteorological and Air Quality Monitoring
• Quality Assurance and Quality Control
• Data Analysis
• Emission Inventory Development
• Meteorological and Air Quality Modeling
• Back to basics

5
Meteorological Assessment

• Objective Assess the readiness of meteorological
  data and models to drive the air quality
  simulation models
• Issues investigated
• the sufficiency of data precision, accuracy,
  bias, consistency, and time-resolution
• The adequacy and validity of measurement methods
  
• the ability of models to represent important
  processes and phenomena and
• new model evaluation techniques.

6
Model Performance Evaluations

• Typical Operational Evaluations Focus on
  important Parameters
• Statistical
• Graphical Temporal and Spatial Comparisons
  Animations
• Diagnostic and Sensitivity Simulations

7
Innovative Methods (1 of 2)

• Data-Based Analysis Understanding Processes and
  Phenomena
• Community Modeling and Analysis System (CMAS)
  1997 2007
• Analysis Replication
• Derived and Integrated Parameters
• Transport Statistics
• Flux Calculations
• Trajectories and Tracers

8
Innovative Methods (2 of 2)

• Process-Based Analysis
• Assess Meteorology with an AQM
• Assess Processes and Performance between Sources
  and Receptors but
• Synthesis
• Relate Physical and Chemical Processes
• Multi-Parameter Analysis
• Big Picture

9
Examples

• Based on Important Data Analysis Findings
• Tracer Concentration Distribution
• Wildfires and Ozone Aloft
• Flux Calculations and Transports Statistics
• Plume Rise
• Carbon vs. Nitrate Aerosols
• Recirculation
• Nighttime Nitrate Formation Aloft
• Fog and Stratus
• Soil Temperature-Air Temperature-Fog-Mixing
  Heights

10
Tracer Distribution

• CRPAQS MM5-CAMx with 1 ppm initial concentration
• Analysis after 60 hours
• Surface concentration
• Peak tracer concentrations by region
• Mass balance

11
High Ozone Day Temperatures
12
Maximum Predicted Temperatures
September 19, 2000
13
Air Quality Aloft
14
Aircraft Spirals
35.9N 19.5W
37.0N 120.1W
15
Ozonesondes
16
Ozone Aloft
17
Ozone Correlation by Level
7.5 km
1.8 km
0.25 km
0.9 km
18
Ozonesonde Transport?
19
Wild Fire Tracers
16 km
20
Hydrocarbons Aloft
21
Transport Statistics
RWP
MM5
22
Ventilation Index
23
Mixing Depth Growth
24
Vertical Wind Profiles
RWP
CAMx Input
MM5
25
Mass Flux Analysis
26
Concentration Fluxes
27
Plume Rise Experiments
28
Soil Temperature
29
Extent of Fog

• MM5 tends to overestimate the extent of fog and
  stratus.

30
Summary

• Think beyond traditional approaches
• Analysis
• Multi-method
• Multi-parameter
• Phenomena and Processes
• Synthesis
• Challenge models to replicate the synthesis
• Maybe then the atmosphere will behave as models
  predict

31
Acknowledgements

• The evaluation methods discussed in this paper
  were developed over the past decade with funding
  from many agencies. Analyses and evaluations
  specific to the CCAQS were funded by the San
  Joaquin Valleywide Air Pollution Study Agency.
  The statements and conclusions in this paper are
  those of the authors and not necessarily those of
  the California Air Resources Board, the San
  Joaquin Valleywide Air Pollution Study Agency, or
  its Policy Committee, their employees or their
  members. The mention of commercial products,
  their source, or their use in connection with the
  material reported herein is not to be construed
  as actual or implied endorsement of such products.

32
Parting Thought

• Why arent meteorological models instrumented
  with process analysis tools like photochemical
  grid models?