Regional 12km Screening

1
Modeling Flow Chart
Regional 12-km Screening CALPUFF Simulations for
2001-2003
High ?dv gt 0.5
No
Source not subject to BART controls
Yes
Subregional fine-scale refined CALPUFF
simulations for 2001-2003
98th ?dv gt 0.5
Source not subject to BART controls
No
Yes
Source subject to BART controls
2
VISTAS Regional Domain
3
CALMET Overview and Unresolved issues

• Grid resolution
• Regional screening domain
• Subregional refined modeling domain
• Observations/hybrid mode vs. no-observations mode
• Definition of subregional domains
• MM5 data problems

4
Terrain (1-km) in Great Smoky Mtns Area

5
Terrain Effects

• Coarse resolution flattens terrain features and
  fills valleys

Great Smoky Mountains NP domain
6
Terrain Cross Section through Great Smoky Mtn NP

7
Valley Becomes Discontinuous
8
Grid Resolution

• Grid resolution
• 12-km for regional screening modeling
• 4-km recommended by VISTAS for subregional runs
• 1-km recommended by Earth Tech for subregional
  runs

9
No-Obs vs. Hybrid/Obs Mode

• Proposed
• No-Observations model for 12-km for regional
  screening modeling
• Hybrid mode including both observations and MM5
  data for subregional refined modeling

10
Definition of Subdomains

• Factors include
• Source locations
• Class I area locations
• Grid size
• Minimize number of domains
• Reduction in number of ALM boundary condition
  runs, which are required once for each
  subregional domain

11
MM5 Data Precipitation Problem

• MM5 datasets contain errors in precipitation
  fields associated
• Associated with MM5s tracking of precipitation
  as cumulative from start of run and use of warm
  start continuation runs
• Data can be fixed, but postprocessing needed
• May be beneficial to switch to new 3D.DAT data
  format

12
Time Series of MM5 Precipitation
13
Other CALMET Issues?

• ???

14
CALPUFF Overview and Unresolved issues

• Methods to address SO2, NOx, H2SO4, PM10, VOC,
  NH3
• Puff Splitting
• Ammonia Limiting/BC simulations
• Other questions?
• Dispersion options
• Building downwash

15
Consideration of Ammonia in BART Analyses

• Essential to evaluate time and space dependent
  NH3 in BART analyses
• Option A to use empirical observational data
  (CASTNet and IMPROVE) to estimate background NH3
• Option B Use time-averaged CMAQ NH3 predictions
• Confirm estimates using NH3 measurements and
  compare Obs vs Pred NO3

16
CALPUFF Chemistry - ALM

• HNO3/NO3 equilibrium based on T and RH
• Formation of NH4NO3 may be limited by the
  availability of NH3
• NH3-limiting condition shifts split of TNO3
  toward HNO3
• Sulfate preferentially scavenges NH3 over NO3
• Background SO4 and NO3/HNO3, affect availability
  of NH3 and the formation of NO3

17
NO3/HNO3 Equilibrium Depends on Temperature,
Relative Humidityhere, total TNO3 NO3 HNO3
18
How is NH3 used in CALPUFF ?

• Ammonia available on puff-by-puff basis
• Cumulative ammonia consumption modeled in
  postprocessor (POSTUTIL)
• CALPUFF ability to treat non-linear effects of
  ammonia limitation using Ammonia Limiting Method
  (ALM)

19
CALPUFF Validation Results with ALM

• Explicit modeling of 1776 sources in Wyoming
  using actual emissions, including CEM data, when
  available
• Boundary conditions based on empirical CASTNet
  and IMPROVE data
• Excellent model performance in predicting both
  sulfate and nitrate if using ALM

20
CALPUFF Performance Predicting SO4
W/ BCs and ALM (Bridger IMPROVE site)
2x overprediction
Pred Obs
2x underprediction
21
CALPUFF Performance Predicting NO3
W/ BCs and ALM (Bridger IMPROVE site)
2x overprediction
Pred Obs
2x underprediction
22
Modeling ResultsTesting of ALM/BC vs FLAG
Constant AmmoniaAssumption(Scire et al., 2001)

23
CALPUFF Model Performance NO3 w/ BCs/ALM -
Bridger IMPROVE Site
24
NO3 w/ Constant 0.5 ppb NH3 - FLAG (2000) -
Bridger IMPROVE Site
25
NO3 w/ Constant 0.5, 1.0 ppb NH3 and time-varying
NH3 -Bridger IMPROVE Site
26
Conclusions

• Consideration of distant sources (BCs) and time-
  space-varying ammonia is essential to good model
  performance for NO3
• ALM/BCs should be used for refined BART
  visibility analyses
• Tests of CMAQ vs empirically-derived NH3 will be
  conducted before final recommendation is made

27
CALPUFF PUFF-SPLITTING

Side View
3
2
1
Top View
1
2
3
28
Summary of Critical Factors in Visibility Modeling

• (1) Ammonia Limiting Method
• Time and space variability in ammonia
  concentrations
• Account for non-linear effects associated with
  ammonia limiting conditions
• (2) Grid Resolution
• Important to obtain correct terrain influences on
  wind field, transport
• Proper characterization of land use and
  land/water boundary in coastal areas
• (3) CALMET model configuration
• Correct selection of CALMET diagnostic wind
  parameters
• (4) Accounting for large changes in natural
  background light extinction due to weather events
• Fog, precipitation can dramatically affect
  background light extinction