NATURAL HAZE REFINEMENT ALTERNATIVES IN THE VISTAS REGION

1
NATURAL HAZE REFINEMENT ALTERNATIVES IN THE
VISTAS REGION

• Ivar Tombach
• VISTAS State and Tribal Air Directors Meeting 17
  August 2005

2
Default Estimates of Concentrations in the East
Under Natural Conditions
3
Default Approach

• Apply annual climatologically representative
  f(RH) to sulfates and nitrates
• Add 10 Mm-1 for Rayleigh scattering to get annual
  bext
• Calculate annual haze index in deciviews
• Deduce haze indices for the 20 worst/best days
  by the statistical procedure of Ames and Malm
• Results for every Class I area are in EPAs
  natural conditions guidance

4
How Did Trijonis Get His Default Concentrations?

• He started with a limited base of data and
  interpreted it with a healthy dose of intuition!
• He estimates that an actual natural
  concentration, C, in the East will be within the
  range Cd/EF C CdEF about 80 of the time,
  where Cd default concentration EF error
  factor

5
Default Estimates Uncertainty(68 probability
that actual value will lie in range given)
6
Summary Concerning Default Uncertainties

• gt There is 1 out of 3 chance that error in dry
  non-Rayleigh extinction could be greater than
  45/-26
• Adding in effects of humidity will not greatly
  change these bounds

7
Estimating Average 20 Worst/Best Conditions

• Ames and Malm calculated standard deviations of
  haze indices using IMPROVE data for 1995-1999
• They scaled sulfate and nitrate concentrations at
  each site to achieve EPAs default annual
  averages and recalculated standard deviations of
  haze indices
• Concentrations of organics, LAC, soil, and coarse
  matter were not rolled back because of probable
  large natural components

8
Natural Standard Deviations vs. Means of Haze
Indices
From Ames Malm, 2001
9
Estimates of Standard Deviations (?) Under
Natural Conditions

• SE Coast ? ranges from 2.5 to 3.0 dv
• ? for rest of E ranges from 2.6 to 3.6 dv
• Ames and Malm chose a single value of 3 dv for
  East

10
Determine HI for Worst 20 of Days

• Ames Malm assumed 90th ile represented average
  of worst 20 of days
• For normal distribution, 90th ile is at 1.28 ?,
  so they said average haze on worst/best 20 of
  days is HI(W20/B20) HI(Ann. Avg.) 1.28 ?
• Assuming ? 3 dv, the default estimates for the
  20 worst and best natural days are
  HI(W20/B20) HI(Ann. Avg.) 3.84 dv

11
Problems with Procedure

• Assumption that 90ile represents average of top
  20 is wrong. For a normal distribution the
  92ile is the correct one
• Changes 1.28 factor to 1.42 and thus increases
  the worst 20 days haze by 0.42 dv (i.e., 3 x
  0.14)
• But, in reality, the distribution of HI is not
  normally distributed. The distribution is
  particularly skewed under natural conditions,
  when particle bext is close to the 10 Mm-1
  Rayleigh add-on.

12
Problems with Procedure (2)

• gt The assumptions used by the EPA to derive the
  worst and best 20 HI are not quite correct.
• Also, assigning ? 3 dv to the entire East
  results in some bias along the SE coast.

13
Recap of Process Uncertainties

• Default concentrations values are inspired
  guesses by John Trijonis.
• In 1/3 of cases, the impact of Trijonis error
  factors is larger than -26 to 45 of dry
  non-Rayleigh extinction (or roughly -0.9 to 1.9
  dv after Rayleigh is added)
• Errors and simplifications in the Ames Malm
  process for estimating best and worst 20 natural
  conditions have potentially large impacts.

14
Potential Refinements to Default(Red Recommend
VISTAS Consider)

• 1. IMPROVE formula
• Add sea salt
• Change OC to POM mass multiplier
• Different current and natural?
• 2. Method for estimating worst/best 20 haze
  from average natural conditions (Ames-Malm)
• Change to 92nd ile to better estimate worst
  conditions
• Different ? for different subregions
• Effect of non-normality of natural HI

15
Potential Refinements (contd)

• 3. Refine default values of average component
  concentrations. Options
• National -- 2 regions, as now
• Regional -- subdivide current regions, e.g., into
  Interior West, Pacific Northwest, Appalachians,
  Southeast Coast
• Local -- account for local area meteorology and
  emissions

16
Contributors to VISTAS Natural Concentrations and
Spatial Variability

• Seasonal
• Oceans, tidal zones, and surf zones
• Sea salt
• POM (primary and secondary)
• Sulfates
• Forests and vegetation
• Sulfates and nitrates
• POM (primary and secondary)
• Coarse organic matter (from plant detritus)
• Ammonia
• Earths crust
• Fine and coarse soil particles from wind and
  natural disturbances

17
Contributors to VISTAS Natural Concentrations and
Spatial Variability (contd)

• Episodic
• Windstorms
• Fine and coarse soil particles
• Wildfires
• POM
• LAC
• Soil
• Emissions from above are influenced by
  meteorology and subject to long range transport

18
Initial Estimates of Component Impacts in the
Southeast
19
Issues with Initial Estimates

• How much of each impact is already included in
  default values?
• Some seasons greater and some less than default
  would be consistent with EPAs default
• Some locations greater and some less than default
  would be consistent with EPAs default
• How to incorporate highly episodic background
  values into refinements?

20
Examples in Southeast

• Inland site -- use GRSM as example
• 1. Default
• 2. Change POM multiplier to 1.8 for present and
  to 2.1 for natural conditions
• 3. Also add 0.3 µg/m3 annual average of biogenic
  POM
• 3. Instead of 3, add African dust and biogenic
  organics during summer only.

21
GRSM Example -- Annual Average Natural Conditions
bext Estimates
22
GRSM Implications -- 20 Haziest Days
23
Examples in Southeast (2)

• Coastal site -- use ROMA as example
• 1. Default
• 2. Change POM multiplier to 1.8 for present and
  to 2.1 for natural conditions
• 3. Also add 1.3 µg/m3 annual average of sea salt
• 4. Also add 0.2 µg/m3 annual average of oceanic
  POM
• 5. Also add annual average of 0.3 µg/m3 of
  African dust

24
ROMA Example -- Annual Average Natural Conditions
bext Estimates
25
ROMA Implications -- 20 Haziest Days
26
Conclusions

• Typical default slope is -2 to -3 dv/decade in
  SE.
• Estimated error in default estimates will alter
  glide path slope by greater than -0.2 to 0.4 dv
  per decade in about 1/3 of cases
• Positive is less steep
• Conservative refinements to formulas and default
  concentrations can affect glide path slope by
  0.4 to 0.5 dv per decade