Recent Advances in the Modeling of Airborne Substances

1
Recent Advances in the Modeling of Airborne
Substances

• George Pouliot
• Shan He
• Tom Pierce

2
Introduction

• In support of air quality modeling, the
  Atmospheric Modeling Division is seeking to
  improve emission estimates by building emission
  models that account for meteorological conditions

3
Improvements to Emission Models in Three Areas

• Biogenic Emissions Inventory System (BEIS)
• Mobile Source Emissions Modeling in an Air
  Quality Forecast System
• Fugitive Dust Emissions for Unpaved Roads

4
Status on BEIS3

• BEIS introduced in 1988 to estimate VOC emissions
  from vegetation and NO emissions from soils.
• BEIS3.09 is the default version in SMOKE 2.0
• 1-km vegetation database by tree species
• Emission factors for isoprene, terpenes, OVOCs
  NO
• NO soil emissions dependent on temperature only
• Only species modulated by solar radiation is
  isoprene
• Supports CBIV, RADM2, and SAPRC99 mechanisms

5
BEIS 3.10

• A research version for CMAQ
• Includes a 1-km vegetation database that resolves
  forest canopy coverage by tree species
• Emission factors for 34 chemicals, including 14
  monoterpenes and methanol
• MBO, methanol, isoprene modulated by solar
  radiation
• a soil NO algorithm dependent on soil moisture,
  crop canopy coverage, and fertilizer application
• support for CBIV, RADM2, and SAPRAC99 mechanisms.

6
BEIS 3.11

• Revises the soil NO algorithm to better
  distinguish between agricultural and
  non-agricultural land, and to limit adjustments
  from temperature, precipitation, fertilizer
  application, and crop canopy to the growing
  season and to areas of agriculture.
• Leaf shading algorithm is added for estimating
  methanol emissions from non-forested areas.

7
BEIS 3.12

• Update to BEIS3.11
• Revises Soil NO algorithm for last half of
  growing season. Reduces the impact of fertilizer
  application during the latter part of growing
  season.
• Available soon on at www.epa.gov/asmd/biogen.html

8
Comparison of BEIS 3.09 3.12

• Annual simulation for 2001
• 36 km continental domain

9
(No Transcript)
10
(No Transcript)
11
Domain total (1000 metric tons/yr)
Compound BEIS3.09 BEIS3.12 change
NO 467 609 30
Total VOC 50,320 48,365 -4
Isoprene 22,141 22,141 0
12
Mobile Source Emissions Modeling for Air Quality
Forecasting

• A National Air Quality Forecast System is being
  developed by EPA and NWS
• Initial Operating Capability for Summer of 2003
• Northeastern U.S domain
• Twice daily forecasts12Z (48 hr) 6Z (30 hr)
• ozone (O3)

13
Mobile Source Emissions Modeling for Air Quality
Forecasting

• Requirements Post-processing of meteorological
  data, emission processing, and the air-quality
  model simulation must be completed in less than
  5.5 hours. Emission processing needs to be
  complete in less than 15 minutes.
• Mobile source processing with Mobile5b requires
  more than an hour. Mobile source processing must
  be faster.

14
Mobile Source Emissions Modeling for Air Quality
Forecasting

1. Separate temperature dependence from MOBILE5B
2. Run Mobile5B with a constant temporal profile
3. Compute coefficients for each species using
   results from (2) and temperature data for a
   representative time period
4. Run Mobile5B with a constant temperature
5. Combine the operational temperature data, results
   from (3) and (4) in a simple loop to calculate
   the mobile source emissions

15
Mobile Source Emissions modeling for Air Quality
Forecasting

• Nonlinear Least-Squares Method can be applied to
  the results from Mobile5B to approximate the
  temperature relationship with a polynomial
  function
• This method of estimating mobile emissions is
  very fast

16
Results from Summer 2003

• July 2003
• Compare retrospective MOBILE5B with real time
  mobile source emission calculation using the
  nonlinear least squares technique
• Domain wide for NO, VOC, CO
• New York State for NO, VOC, CO

17
(No Transcript)
18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
(No Transcript)
22
(No Transcript)
23
Summary of Domain Total Results
Pollutant Real Time AQF system Mobile 5B difference all emissions
NOx (tons/dy) 9,363 9,333 0.3 30
VOC (1000 mol C/dy) 339,096 347,048 -2.3 11
CO (tons/dy) 54,219 55,379 -2.0 56
24
Fugitive Dust Emissions from Unpaved Roads
(Current Method)

• Does not account for transportable fraction near
  the source regions
• Uses road mileage from FHWA
• Uses rainfall data from a single location in each
  state to account for rainfall effects
• Uses AP42 emission factors

25
Fugitive Dust Emissions from Unpaved Roads
(Proposed)

• Use the TIGER road mileage data and grid to the
  county level.
• Model the moisture content of the road surface
  using modeled solar radiation, dew point, wind
  speed and rainfall data for each grid cell (note
  this is an extension of AP-42s documentation).
• Incorporate the transport factor developed by
  Shan He for windblown dust

26
Conclusions

• BEIS3 tested for an annual simulation. Latest
  version is now 3.12
• An efficient method to estimate emissions for an
  air quality forecast system has been used for
  summer 2003
• A module in SMOKE to estimate emissions from
  unpaved roads is being built and tested.