CMAQCAMx MPE for VISTAS 2002 Episode

1
2018 Visibility Projections Approaches to
Calculations and Results using 2002 and 2018 Base
D Modeling Results ENVIRON, Alpine Geophysics,
and UC Riverside
VISTAS TAWG Meeting Asheville, North
Carolina August 17-18, 2005
2
2018 Visibility Projections

• Several rounds of 2018 visibility projections
  presented to VISTAS TAWG
• April 5-6, 2005 TAWG meeting in RTP
• June 1, 2005 and other TAWG conference calls
• June 9-10, 2005 National RPO meeting in Denver
• Received comments and recommendations
• New methods added, little dropped
• Todays results based on 2002 Typical Base D and
  2018 OTWd modeling results

3
Projection of 2018 Visibility

• 2018 and 2002 modeling results used in a relative
  fashion to project the observed Baseline
  Conditions using Relative Reduction Factors
  (RRFs) that are site- and species-specific ratios
  of 2018 to 2002 model estimates
• Examine several Methods of model derived RRFs
• 2018 visibility projections compared against
  default 2018 Reasonable Progress Goal (RPG)
  obtained through Linear Glide Path from 2004
  Baseline Conditions to 2064 Natural Conditions
• Use EPA Default Natural Conditions effects of
  adding Sea Salt were small (0-3) effects of
  non-US sources expected to be larger
• States/Tribes may select other paths than linear
• Present 2018 visibility projections as percent of
  achieving RPG
• If most methods gt 110 of RPG then achieves RPG
• If most methods lt 90 of RPG then no achieve RPG
• 90-110 may achieve or more analysis needed (20
  fuzz factor)

4
Baseline Conditions 28.9 dv Natural Conditions
11.4 dv 2018 Target 24.9 dv
2018 Reduction Goal 4.1 dv 2018 Modeled
Reduction 5.2 dv GRSM achieves 2018 RPG
5
2018 RPG Calculations

• Baseline Conditions (2004 observed starting point
  for Glide Slope and projections)
• (B) 4-Year average of 2000-2003 Worst 20 Days
• RHR based on 5-yr 2000-2004
• 2004 IMPROVE data expected October 2005
• (C) 2002 Worst 20 days
• Same as modeled year allows use of day-specific
  RRFs
• See similar 2018 visibility projections using
  Baseline Conditions B and C
• Proceed with Baseline Conditions B (2000-2003)
  as more like final Baseline Conditions
  (2000-2004)

6
2018 RPG Calculations

• Methods used in Calculating RRFs
• Method 1 RRFS based on average of Worst 20 2002
  modeled days
• Method 2 Quarterly RRFS based on average of
  Worst 20 modeled days for each Quarter
• Method 3 Day-specific RRFs (Baseline C only)
• Method 4 Weighted RRFS using CART
• Method 4-1 CART weights for all 2002 Modeled
  Days
• Method 4-2 CART weights for 2002 Monitored Days
• Method 4-3 CART weights for 2002 Worst 20 Days

7
2018 RPG Calculations

• Methods used in Calculation RRFs
• Method 5 Quarterly RRFs using all 2002 modeled
  days
• Method 6 RRFS based on extinction
• Dropped from analysis at June 10 Nat RPO Meeting
• Method 7 Concentration specific RRFS
• Added from June 10, 2005 Nat RPO meeting
• For each Class I area and PM species develop
  concentrations specific RRFS using 2002/2018
  modeling results from the 2002 monitored days

8
2018 Visibility Projections using Methods 1-4 and
Baseline B and C
Method 3 (Day-Specific RRFs) looks like an
outlier sometimes unstable Baseline B C
provide similar results
9
Effects of Model Performance on 2018 Visibility
Projections

• Previous used two screening criteria to account
  for model performance in RRFs
• wmpe exclude modeled days/species if 2002
  Actual Base E estimates exceed factor of 2 of
  observed value (0.5 lt .. lt2.0)
• wmpe2 exclude modeled days if modeled
  extinction exceeds factor of 2 of observed
• Added more stringent model performance screening
  within 50 (0.75 lt .. lt 1.5)
• smpe and smpe2
• Added at June 10 Nat RPO meeting

10
Effects of Model Performance on 2018 Projections
Days eliminated by smpe results in less robust
RRFs and eliminates all days from some sites and
Methods (COHU, EVER, CHAS, GRSM, SWAN, BRET)
11
Alternative Extinction Equations

• (a) Standard IMPROVE aerosol extinction equation
• 3 x f(RH) x 1.375 x 3 x S 3 x f(RH) x 1.29
  x NO3_ 10 x EC 4 x 1.4 x OC SOIL
  0.6 x CM
• (b) Alternative with more OC less SO4 (alt_b)
• 2.2 x f(RH) x 1.375 x 3 x S 3 x f(RH) x 1.29
  x NO3_ 10 x EC 4 x 1.7 x OC SOIL
  0.6 x CM
• (c) Extinction equation proposed by Ryan/EPRI
  (alt_c)
• 1.62 x f(RH) x (1.375 x 3 x S 1.29 x
  NO3_)1.28 10 x EC 1.71 x (2.0 x OC)1.42
  SOIL 0.6 x CM
• (d) Extinction equation proposed by Malm
  (alt_d)
• 3.0 x (0.70.02 x Mmix) x f(RH)new x (1.375 x 3
  x S 1.29 x NO3_) 10 x EC 3.63 x
  (0.70.02 x Mmix) x (1.8 x OC) 1.0 x SOIL
  0.6 x CM 1.0 x f(RH)new x Sea Salt
• Where, Mmix (1.375 x 3 x S 1.29 x NO3_
  1.8 x OC)
• Sea Salt not accounted for at this time
• New f(RH) estimates not yet used

12
Effects of Alternative Extinction Equations on
Vis Projections
No Change in W20 Days 1B Std IMPROVE alt_d
modified IMPROVE alt_c Ryan/EPRI alt_d New
Malm/Hand
13
Effects of Alt Equ and Sea Salt (None, Cl- vs.
Na)
14
Expanding List of Days

• Calculations so far used 20 Worst Days based in
  current IMPROVE equation downloaded from VIEWS
• Define new set of 2000-2003 Worst 20 days using
  alternative equations for Baseline Conditions
• Include Methods 2, 5 and 7-2
• Method 1 Average of 2002 Worst 20 Days
• Method 2 Quarterly Avg RRFS using 2002 Worst 20
  Days
• Method 5 RRFs based on 30 Worst days from each
  quarter
• Method 7-2 Concentration specific RRFs using
  2002 Monitored days
• Four Class I Areas (GRSM, ROMA, SAMA SIPS)

15
Visibility Projections Allowing W20 Days to
Change with Alternative Equation _at_ GRSM, ROMA,
SAMA SIPS
Days change w/ Equation 2B cannot be used at GRSM
due to missing Q days Preliminary Conclusions
enhanced with new equations
16
Conclusions 2018 Visibility Projections

• Of alternative approaches studied, Alternative
  Equation makes largest difference
• Ryan/EPRI and Malm/Hand equations achieves RPG
  easier
• Did not use new f(RH) curves and did not include
  Sea Salt
• New results have not been fully analyzed and
  quality assured
• Except for Methods that end up using few days in
  the RRFs, different Methods generally produce
  consistent results
• Method 3c (day-specific) and stringent
  performance test (smpe) sometimes produce
  artifacts usually related to nitrate
• Using new equation (Ryan/EPRI or Malm/Hand) most
  VISTAS Class I areas achieve RPG under 2018 RPG
• OKEF ( WOLF) still not achieving and ROMA EVER
  may achieve RPG

17
.
2018 OTWd Visibility Projections
.
.
.
Current IMPROVE Equation
18
.
.
Hercules Glade, MO
2018 OTWd Visibility Projections
.
.
.
.
.
.
.
.
New Equations???