2003 WESTAR Fall Technical Conference

2003 WESTAR Fall Technical Conference

on PSD Increment Tracking

and Cumulative Effects Modeling

September 15-17, 2003

(9/17 draft for discussion purposes)

Technical Options

Met Data Class I and Class II What type of meteorological data are most appropriate for Class I (long-range) analyses? What type of meteorological data are most appropriate for simple and complex terrain Class II (near-field) analyses?

Met Data Options

/ Advantages-strengths-benefits / Disadvantages-weaknesses-limitations
Class I:
MM5 (3 years) / -EPA/FLM accept / -36 km grid is too coarse
-unproven quality
-limited availability in AK
Other mesoscale models / -smaller, more appropriate scale / -don’t know where the data are
-QA
RUC / -12 km scale / -no consensus about how good it is
Worst-case data set for Calpuff lite (screening) / -speed
-availability / -difficult to pass (conservative)
-not a cumulative analysis
5 years of surface and upper air data/CalMet/ISC/AERMET / -EPA/FLM accept (Appendix W) / -limited representativeness
Class II simple terrain:
On-site (1 year) / -Data is representative
-more sensitive instrumentation can be used
-can be collected at multiple levels / -Time
-$(source and agency review)
-subject to climatic variability
NWS (5 years) / -readily available
-multiple years / -questionable (very for complex terrain) representativeness
Worst-case screening data / -fast and free / -difficult to pass (conservative)
-single source
-poor spatial representation
Class II complex terrain:
Profiler / -improved representativeness / -$$$$$ (initial and operating costs)
On-site (1 year) / -Data is representative
-more sensitive instrumentation can be used
-can be collected at multiple levels / -Time
-$(source and agency)
-subject to climatic variability
NWS (5 years) / -readily available
-multiple years / -very questionable representativeness
Worst-case screening data / -fast and free / -difficult to pass (conservative)
-single source
-poor spatial representation
Off-site monitoring to augment on-site data / -improved representativeness / -$$

Meteorology Class I and Class II How representative are the met data?

Data Representativeness Determination Options / Advantages-strengths-benefits / Disadvantages-weaknesses-limitations
Use of historical met data: / -true representation of base-year conditions / -not all parameters generally available
Consistency of years of data:
-Less than 5 years / -less expensive
-easier to get / -not as ‘stable’ (annual variability
-may need ‘buffer’ to ensure future compliance
-5 or more years / -more stable
-may not need buffer to ensure compliance / -$$
-harder to get
– takes more time

InventoryClass I and Class II Advantages and disadvantages of using actual versus allowable emission rates in the increment analysis?

Actual or Allowable emissions / Advantages-strengths-benefits / Disadvantages-weaknesses-limitations
Actual emissions
-Current / -follows the letter of the law for existing sources
-allows model performance evaluation
-readily available for major sources / -moving target
-contested definition
-may not be protective
-may require permit alteration based on current utilization
-Baseline / -follows the letter of the law
-accurate / -unavailable for non-point and many point sources
Allowable emissions (PTE/permitted)
-Current / -more protective
-readily available in some states
-required for permitted new sources / -conservative, may limit growth (opposite for expansion)
-may cause appeals
-not all states have data
-Baseline / -may be the only thing that is available / -contrary to the law

InventoryClass I and Class II How should missing emissions data and lack of historic data for calculating baseline be accounted for in the inventory?

Options for missing data / Advantages-strengths-benefits / Disadvantages-weaknesses-limitations

Missing emissions data:

-Derived from available data (Census, interpolation, DOE, acid rain data base, national business data, NEPA, NEI)

/ -May be the only way it can be done / -can’t demonstrate accuracy
-inconsistency

-Assume no change

/ -conserve resources / -difficult to defend

-Assume full increment is available (ignore baseline impact)

/ -simple / -can’t demonstrate accuracy
-difficult to defend

-Assume full increment is consumed

/ -simple / -loss of economic opportunities (development requires offsets)

Inventory Class I and Class II How should non-point source emissions be included in the inventory for the analyses?

Non-point source inventorying options

/ Advantages-strengths-benefits / Disadvantages-weaknesses-limitations
Area sources:
-Surrogates such as population to calculate (i.e. dry cleaners) / -data readily available / -surrogate may not be representative
-question of how you distribute it
-Utilize actual inventory (including survey) / -more accurate / -$$$
-NEI / -cheap / -poor spatial resolution (county)
-Assume no change / -easy
-may be appropriate / -would require justification
Mobile sources:
-Utilize actual inventory (Mobile6) / -more accurate
-inventory may already exist (i.e. SIP/TIP) / -$
-difficult to obtain activity data
-NEI / -cheap / -poor spatial resolution (county)
-Assume no change / -easy
-may be appropriate / - would require justification
Temporary sources:
-Exclude sources if less than a fixed time (i.e. 1 or 2 years)
(see 52.21, 51.166) / -varies by state / -varies by state
-must be in SIP (51.166)
Intermittent sources:
-Exclude all / -easy / -may not conform to regulation
-Include / -more accurate / -resource intensive
-Prescribed Fire: / -more accurate / -difficult to distribute spatially
and temporally

Inventory Class I and Class II How should short-term emission rates be calculated from annual emission rates?

Calculation options

/ Advantages-strengths-benefits / Disadvantages-weaknesses-limitations
Calculation of short-term emission rates:
-Annual actual emission rate divided by hours of operation / -better reflection of actual air quality change / -may miss peak short term consumption
-Maximum actual / -reflects short term peak increment consumption / -only applicable to major sources
-may overstate impacts for linked sources
-nth percentile of measured short term (note: has been used for linked power plants which will never all be at maximum at the same time) / -may be more appropriate statistically / -source type inequity

Inventory Class I and Class II What are the database tracking requirements? How can states share databases for increment tracking?

Tracking Options

/ Advantages-strengths-benefits / Disadvantages-weaknesses-limitations
Increment tracking database options:
-NEI / -consistent data content and format
-available to everyone / -no current capability
-State generated / -serves state specific needs / -varying formats
Interstate coordination:
-Include in NEI / -available to everyone
-consistent data content and format / -dependent upon EPA (slow)
-differing baseline years
-RPOs (WRAP) / -may be faster
-state/region controlled / -baseline outside scope of work
-differing baseline years
-Other multi-state coordination (formal or informal) / -more responsive
-less expensive / -duplication of effort
-differing baseline years

Modeling Class I and Class II How should terrain be accounted for in the dispersion modeling analysis for Class I and Class II increment analyses? (Assumption: must use model that handles complex terrain)

Terrain inclusion options

/ Advantages-strengths-benefits / Disadvantages-weaknesses-limitations

Class I:

-FLM provided receptor/elevation information / -consistency / -does not include area outside Class I area
-DEMs / -readily available, except for AK
-consistent / -needs to be QA’d
Class II:
-DEMs / -readily available except for AK
-consistent / -needs to be QA’d

Modeling Class I and Class II What screening techniques are appropriate for Class I and Class II increment analyses? (note: more of a permitting tool than a tracking tool)

Screening technique options

/ Advantages-strengths-benefits / Disadvantages-weaknesses-limitations
Class I:
-Calpuff Lite / -much simpler than Calpuff / -not necessarily worst-case
-single site
-ISC/AERMOD impacts at 50 km or less / -even simpler / -lacks chemistry
Class II:
-SCREEN / -fast
-inexpensive / -single source only
-AERSCREEN / -fast
-inexpensive
-can be used for multiple sources / -sources must be in close proximity
-ISC w/screening met / -fast
-inexpensive
-can be used for multiple sources / -sources must be in close proximity

Modeling Class I and Class II How frequently should the modeling analysis be done for Class I and Class II increment analyses? (note: periodic review required, frequency not defined)

Analysis frequency options

/ Advantages-strengths-benefits / Disadvantages-weaknesses-limitations
Class I and Class II:
-Periodic (i.e. 3-5 years) / -consistent
-seems to meet guidance
-could be linked to other periodic requirements (i.e. RH, inventory) / -$$$$
-On demand (i.e. permit application) / -would not have to do periodic
-applicant bears cost
-address increment problems before permit is issued to avoid a SIP change / -incomplete (usually limited to area of impact)
-might over control
-On demand (regional modeling center) / -could be done by regional modeling center / -regional modeling centers don’t exist
-resistance by applicant
-Triggered by growth or other indicators (i.e. increases in monitored concentrations) / -would not have to do periodic
-provides more knowledge about trends / -threshold has to be defensible
-On a forecast basis / -continuous updates
-stability of solution / -$$$$
-modeling based

Modeling Class I and Class II How should grids be defined in Class I and Class II increment tracking analyses?

Options to define grids (size, receptors, distance)

/ Advantages-strengths-benefits / Disadvantages-weaknesses-limitations
Class I:
-FLM provided receptors (inside Class I area) / -consistency across analysis / -cannot be used for grid models
-Case-by-case to adequately resolve impacts (outside Class I area) / -accurate / -computational resources
Class II:
-Case-by-case to adequately resolve impacts / -accurate / -computational resources
-Regional modeling center for multi-state analysis / -conserve state resources
-consistent / -requires common grid spacing and receptors

Modeling Class I How should the capability of the models to predict small changes in concentration at distant Class I areas be addressed? Under what conditions should different modeling techniques be used in the analyses – how much and when refinement is appropriate?

Accuracy and refinement

/ Advantages-strengths-benefits / Disadvantages-weaknesses-limitations
Options to address model capability:
-Take results at face value / -simple
-EPA approves
-has precedent / -may not be scientifically defensible
-Use monitored detection limits as lower bound / -surrogate for SIL / -may not be defensible
-Case-by-case / -flexible
-attempts reasonableness / -may be viewed as arbitrary
-Model sensitivity testing / -increases confidence in model output / -time and resources to do testing
-must be defensible
Options for refinement of analyses:
-Using observations to nudge met models / -more accurate / -observational data may not be available
-Use more years of met data / -more robust / -more processing time
-observational data may not be available
-Refining or Nesting grids / -more accurate / -more processing time
-Refined source parameters / -more accurate / -more time and resources
-Improved model chemistry / -more accurate / -requires more inputs
-information is not always available
-Use/develop different model / -verification / -time and $$$$
-requires EPA concurrence

Cumulative Effects Class I and Class II When andunder what conditions should cumulative effects analyses be conducted for Class I and Class II increment analyses? (note: in permitting process)

Options for analyses of cumulative effects

/ Advantages-strengths-benefits / Disadvantages-weaknesses-limitations
Class I:
-Always / -represents consumption / -$$$$
-time
-may be unnecessary (distance between source and receptor – 300 km limit)
-state regulatory authority limitations
-Never (when you don’t get a PSD application that affects Class I area) / -no additional work / -minor source effects on Class I areas are not defined)
-allows small source growth impacts
-Use SILs / -guideline
-focus resources on problem areas / -not in regulation, subjects states to challenge
-allows small source growth impacts
-Under conditions of minor or mobile source growth (or based on monitoring) / -focus resources on problem areas
-allows for analysis of Class I impact even if below the SIL / -thresholds are undefined and would have to be defended
-state regulatory authority limitations
Class II:
-Always (when exceeding the SILs) / -legally defensible, but have been challenged
-technically appropriate / -time and resources

Application of FLAG What screening techniques are appropriate? What is the role in permitting of FLAG? How are the threshold levels used? How should states evaluate analyses of AQRVs in Class I areas?

Use of FLAG

/ Advantages-strengths-benefits / Disadvantages-weaknesses-limitations
Screening techniques:
-See above, also Viscreen or Pluvue / -see above / -see above
Use of threshold levels:
-5%/10% for visibility / -guidance, interpreted as regulation
-precedence
-FLM acceptance
-scientifically appropriate / -interpreted as regulation
-not a regulation
-no detailed review of sources below threshold
-0.005 kg/HA/year for deposition / -guidance, interpreted as regulation
-precedence
-FLM acceptance
-scientifically appropriate / -interpreted as regulation
-not a regulation
-no detailed review of sources below threshold
Regulatory role in permitting:
-FLAG / -semi-objective tool
-states understand FLM expectations / -semi-regulatory
-provides single number for case-by-case determinations
Availability of RH:
-Mesoscale modeling / -excellent spatial, temporal and vertical resolution / -not always available
-difficult to evaluate the accuracy
-EPA regional haze guidance / -readily available
-eliminates weather variability / -not hourly
-not paired in time
-no diurnal data
-Surface observations / -readily available, including hourly data / -may not be representative of Class I areas
Inclusion of night hours:
-Include / -simplifies analysis
-protects night sky visibility / -arguable value
-Exclude / -may facilitate permitting / -complicates analysis
-inconsistent with 24 hour analysis
-FLM disapproval
Inclusion of precipitation/fog hours:
-Include / -simplifies analysis
-FLM concurrence / -potential adverse comments on permit
-transport RH differs from background
-Exclude / -consideration under regulation
-may facilitate permitting
-may be reasonable / -complicates analysis
-inconsistent with 24 hour analysis
-difficulty verifying cloud precipitation occurrence
-no scientific basis
-Case-by-case / -flexibility
-FLM practice / -may be considered arbitrary

1