MCIP2AERMOD: A Prototype Tool for Preparing Meteorological Inputs for AERMOD

1
MCIP2AERMODA Prototype Tool for Preparing
Meteorological Inputs for AERMOD

• Neil Davis and Sarav Arunachalam
• Institute for the Environment
• University of North Carolina at Chapel Hill
• Roger Brode
• U.S. Environmental Protection Agency
• Presented at the
• 7th Annual Models-3 CMAS Users Conference
• October 6-8, 2008

2
Motivation

• Meteorological fields are key inputs for air
  quality modeling
• NWS data typically used in AERMOD modeling have
  some limitations
• Observed sites may be far from source location,
  espl. RAOB sites
• Wind measurements at ASOS locations have large
  number of calm measurements
• Gridded meteorological models potentially helpful
  
• For Hybrid (combining regional and local-scale)
  modeling, a consistent set of meteorology for
  CMAQ and AERMOD simulations is desirable
• Avoids inconsistent meteorological fields
  confounding differences in AQM outputs
• EPA is exploring utilizing gridded met data and
  created MM5 AERMOD tool, which was helpful in
  developing this tool
• Using MCIP outputs helps using either MM5 or WRF
  to drive AERMOD
• No transition needed down the road

3
FAA Modeling Approach
4
Approach

• Created Fortran-based utility with EDSS/Models-3
  I/O API library
• Process 2002 MM5 simulations at 12-km through
  MCIP 2.3
• Use grid cell containing AERMOD source region for
  both surface and upper air fields
• No interpolation is performed
• Make use of METCRO2D, METCRO3D, METDOT3D, and
  GRIDDESC files from MCIP output
• Use all available fields directly from MCIP
  output as these are the values CMAQ will be using
• Only calculate variables which are not in MCIP
  output
• Adjust for AERMOD time requirements (LST, some
  parameters require noon LST values)

5
Met Fields Directly from MCIP

• Sensible Heat Flux
• Surface Roughness Length
• Surface Friction Velocity
• Wind Speed / Direction
• Temperature
• Surface Pressure
• Cloud Fraction
• Monin Obukhov Length
• Convective Velocity Scale
• Convective Mixing Height
• Mechanical Mixing Height

• Notes
• Some massaging of these variables performed
• Maximum / minimum thresholds
• Units conversion
• Mechanical Mixing height is both used directly
  and calculated
• Calculated only for convective conditions

6
Fields Calculated in MCIP2AERMOD

• Mechanical Mixing Height
• Relative Humidity
• Potential Temperature gradient above convective
  mixing height (VPTG), or lapse rate above mixing
  height
• Bowen Ratio
• Albedo

7
AERMOD Study Location

• Red Airport Location
• Blue NWS Surface Site
• Green RAOB Site

• T.F. Green Airport in Providence, Rhode Island

8
Evaluation Simulations

• Developed AERMOD simulations for several
  pollutants using both AERMET and MCIP2AERMOD
  meteorological outputs
• Benzene, Formaldehyde, Primary EC, PM2.5
• Will focus on PEC in this presentation
• Emissions inputs created using the FAA EDMS model
  to provide hourly emissions estimates of aircraft
  activity at airport
• 2002 NWS values were processed through AERMET
  with constant surface characteristics in time and
  space
• Midday Albedo 0.5
• Daytime Bowen ratio 1.0
• Surface Roughness 0.1
• Receptors were placed at the center of every
  census tract within a 50-km radius as well as at
  routine AQ monitor locations
• Our evaluation will look at comparing both
  meteorology fields as well as the AERMOD
  concentrations from both simulations
• Diurnal plots are calculated using averages of
  annual data

9
Evaluation of met. fields (1 of 4)
Surface Temperature

• Very high correlation between NWS and MCIP data
• Diurnal and monthly patterns match very well
• MCIP is slightly cooler overall

10
Evaluation of met. fields (2 of 4)
Mechanical Mixing Height

• MCIP produces lower mixing heights at night than
  NWS, but higher mixing heights in general
• MCIP also produces higher mixing heights during
  summer months
• High correlation, but MCIP results seem to fall
  into discrete bins

11
Evaluation of met. fields (3 of 4)
Wind Speed

• Shows stronger winds with NWS data, both diurnal
  and seasonal
• NWS data is grouped into threshold values
• High correlation overall

12
Evaluation of met. fields (4 of 4)
Wind Rose

• NWS has more calms
• MCIP has fewer high wind values
• Directionally MCIP shows more south westerly flow

13
Meteorology comparison

• Good agreement across most variables
• Comparison of vertical data unavailable, due to
  lack of data in the NWS simulation
• AERMET only calculates the lowest level values
  when onsite data is not included
• Common meteorological parameters (i.e., Temp,
  pressure, winds) show more agreement than other
  parameters
• MCIP precipitation and clouds show large
  discrepancies compared to NWS values (not
  presented here)

14
Evaluation of AERMOD outputs for PEC (1 of 5)

• Only slight differences can be seen here
• Most perceivable changes occur away from the
  airport, deceptive due to log scale

15
Evaluation of AERMOD outputs for PEC (2 of 5)
Zoomed-in domain

• Significant changes in airport vicinity
• MCIP-based AERMOD shows higher concentrations

16
Evaluation of AERMOD outputs for PEC (3 of 5)
Annual Average Absolute Difference

• Maximum change of 0.1 ug/m3
• Largest changes to the North of the airport
• MCIP shown to have larger concentrations

17
Evaluation of AERMOD outputs for PEC (4 of 5)
Comparison of Monthly Means

• MCIP data always higher
• Largest differences in the winter months

18
Evaluation of AERMOD outputs for PEC (5 of 5)
Additional comparisons

• NWS has more lower concentration values
• MCIP has higher maximum concentrations
• Median, 25th and 75th percentiles are similar
• Good correlation overall

19
Discussion

• New prototype tool developed to use gridded
  meteorology (from either MM5 or WRF) for AERMOD
• Evaluation of tool performed for AERMOD study of
  T.F. Green (Providence) airport emissions of
  several pollutants
• Comparison of meteorological fields showed
  reasonable agreement for most variables
• Only limited comparison of upper air data was
  possible
• MCIP2AERMOD meteorology lead to higher
  concentrations throughout the domain for PEC
• Despite magnitude differences, correlations were
  high between model simulations
• Evaluation of outputs for other pollutants showed
  similar patterns
• Evaluation of AERMOD outputs with RIDEM field
  study at T.F. Green is ongoing

20
Future Work

• Complete evaluation of AERMOD inputs and outputs
  using RIDEM field study data for 2005
• Explore sensitivity of AERMOD to different
  physics options in MM5 (or WRF)
• Additional tests in MCIP2AERMOD
• Include only noon time Bowen ratio
• Investigate interpolation
• Allow user to override surface parameters
• Set AERMET surface fields to be closer in
  agreement to the MCIP values and reevaluate
• Develop hybrid calculations of CMAQ and AERMOD
  using consistent meteorology

21
Acknowledgments
This work was funded by the FAA, under Grant
No.03-C-NE-MIT, Amendment No. 027 (w/ UNC-CH
Subaward No. 5710002072) 06-C-NE-MIT,
Amendment No. 002 (w/ UNC-CH Subaward No.
5710002208) 07-C-NE-UNC, Amendment No.
001 The Local Air Quality project is managed by
Mohan Gupta.
Any opinions, findings, and conclusions or
recommendations expressed in this material are
those of the author(s) and do not necessarily
reflect the views of the FAA, NASA or Transport
Canada.