RAIMI A Practical Approach for Implementing Cumulative Type Assessments on a Localized Scale

1
RAIMI - A Practical Approach for Implementing
Cumulative Type Assessments on a Localized Scale

• U.S. EPA Region 6
• Multimedia Planning and Permitting Division
• Chicago 2002

2
1.0 INTRODUCTION AND OVERVIEW

• 1.1 Introductions
• U.S. EPA Region VI
• U.S. EPA Region V
• Participants

3
1.0 INTRODUCTION AND OVERVIEW

• 1.2 Seminar Objectives and Methods
• Transfer EPA Region 6 experience in localized
  assessment
• Review details in performing the localized
  assessment for Port Neches, Texas
• Provide for participant interface on
  defining project objectives, resources, and
  constraints of a localized assessment

4
1.0 INTRODUCTION AND OVERVIEW

• 1.3 Content and Materials
• Stepwise methods with RAIMI case study example
• Notebook of presentations and handouts by
  presenters
• Demonstrations
• Group discussions

5
2.0 COMMUNITY LEVEL ASSESSMENTS REGIONAL
PERSPECTIVE

• Why Are Cumulative Type Risk-based Assessments
  Becoming So Popular
• Reality - permitting/enforcement actions need to
  consider bigger picture as opposed to source by
  source permitting
• we blindly focus on RCRA-regulated units, often
  in a forest of others are we making a
  difference with our resources?
• Internal Agency Pressures challenged to find
  mechanisms that support cross program cooperation
  and sharing of resources

6
2.0 COMMUNITY LEVEL ASSESSMENTS REGIONAL
PERSPECTIVE

• Why Are Cumulative Type Risk-based Assessments
  Becoming So Popular (cont.)
• Public Simply Wants to Know Am I safe
  breathing/eating/drinking where me and my family
  live?
• Regulatory and National Program Pressures
  challenged to develop localized assessment
  capabilities (e.g., Urban Air Toxics Strategy)
• Simply a logical approach to a complex and
  often overwhelming issue

7
2.0 COMMUNITY LEVEL ASSESSMENTS - REGIONAL
PERSPECTIVE

• Significant Intermingling of Industry and
  Neighborhoods
• National Scale Studies Continue to Flag Problems
  in Several Region 6 States
• Scale of Issues Too Large to Ignore
• Holistic Answer Needed to Bottom-Line Questions
• Multi-program/Cross Media
• Receptor-based Approach

8
3.0 REGION 6'S APPROACH FOR DEVELOPING
CAPABILITIES TO DO LOCALIZED ASSESSMENTS

• Conducted detailed review of national scale
  studies (NATA, CEP, Air Toxics Strategy, etc.)
• Established practical implementation objectives
  conduct assessments that provide usable results
  that help us as regulators do our job (that
  includes being in a position to effectively, and
  with complete competence, work with industry,
  public, and other stakeholders)

9
3.0 REGION 6'S APPROACH FOR DEVELOPING
CAPABILITIES TO DO LOCALIZED ASSESSMENTS

• Test technical approach through pilot studies of
  actual areas within region
• Make real-time improvements to approach and tools
  based on what is learned from pilots
• Develop risk management plan and communication
  strategy to facilitate use of results

10
4.0 RAIMI - Regional Air Impact Modeling
Initiative

• Risk-based Prioritization Tool and Platform to
  Develop Multi-media Solutions to Environmental
  Problems

11
4.0 REGIONAL AIR IMPACT MODELING INITIATIVE
(RAIMI)

• EPA Region 6 initiated design of the RAIMI
  Program in 1999 to evaluate
• Region-wide estimation of potential health risks,
  on a community level of resolution, as a result
  of exposure to multiple contaminants/sources/expos
  ure pathways.
• As a test of the RAIMI methods and approach, a
  Pilot Study was designed and implemented in a
  small area in southeast Texas.

12
4.0 REGIONAL AIR IMPACT MODELING INITIATIVE
(RAIMI)

• Establishment of the RAIMI Program is being
  driven in part by national initiatives that
  identify the need for a more refined assessment,
  including
• CUMULATIVE EXPOSURE PROJECT
• RESIDUAL RISK REPORT TO CONGRESS
• INTEGRATED URBAN AIR TOXICS STRATEGY
• NATIONAL AIR TOXICS ASSESSMENT

13

• CEP FINDINGS
• HAPs Are Broadly Distributed
• In Specific Locations, Concentrations May Result
  In Unacceptable Risk, Sometimes By An Order Of
  Magnitude Or More

CEP OBJECTIVES Estimate Exposure Levles For A
Wide Variety Of Toxic Pollutants Characterize The
National Distribution Of These Estimated Exposure
Levels Across Communities And Demographic
Groups Identify The Types Of Communities And
Demographic Groups Which Appear To Have The
Highest Exposure Levels Identify Potentially
Important (Based On Risk) Emissions Sources And
Pollutants For Which Information Is Most
Uncertain
14
RESIDUAL RISK REPORT TO CONGRESS - AS REQIRED BY
SECTION 112(f) OF THE CAA
The Residual Report To Congress Identifies Two
Objectives For Residual Risk Activities
Assess Any Risk Remaining After MACT Standard
Compliance
Set Additional Standards, If Necessary, To
Provide An Ample Margin Of Safety To Be Protective
Follows Benzene NESHAP In That Standards Are To
Be 10-4 For Max Individual And 10-6 For
Population (How Defined ??)
15
INTEGRATED URBAN AIR TOXICS STRATEGY AS REQUIRED
BY SECTION 112(k) OF THE CAA
The Strategy Consists Of Four Main Components
Source-specific And Sector-specific Standards
(Regulations)
National, Regional, And Community-based
Initiatives To Focus On Multi-media And
Cumulative Risk, Including Pilot Studies
NATA To Develop Analytical Tools (Models) And
Support Activities (Monitoring) To Identify
Risks, Track Progress Toward Risk Goals, And
Prioritize Efforts To Address Emissions And Risks
From Air Toxics

Educational And Outreach Activities To Reduce
Toxics Emissions
16
NATIONAL-SCALE AIR TOXICS ASSESSMENT
The Goal Is To Identify Those Air Toxics Which
Are Of Greatest Potential Concern, In Terms Of
Contribution To Population Risk
National Scale Study That Includes 33 Air Toxics
17
4.0 RAIMI COMPLEMENTS NATIONAL INITIATIVES

• Complements national scale studies by providing
  neighborhood-level assessment of potential risk
• Is a means to rapidly assess source category
  specific risk before and after the
  implementation of MACT standards
• Provides a scalable and flexible risk management
  tool, as envisioned by the Integrated Urban Air
  Toxics Strategy, to
• Support monitoring activities,
• Identify risks,
• Focus resources,
• Prioritize corrective actions, and
• Track progress toward risk reduction goals.

18
4.0 RAIMI DESIGN STRATEGY
Provide A Consistent Means By Which Permitting
Authorities Could Account For And Assess
Potential Health Effects To Multiple Contaminants
From Multiple Sources, Which Are Often Subject To
Multiple Permitting Schemes, But Cumulatively
Impacting The Same Receptor Neighborhoods
Useful As A Permitting Tool To Support EPA,
State, And Local Permitting Authorities-independen
tly Or Combined-Evaluate And Demonstrate
Protectiveness Of Cross Program (e.g., RCRA, CAA,
Exempt) Permitting Decisions And Support
Holistic, Tailored Permit Strategies
Calculate And Track Risks From Literally Hundreds
Of Sources And Contaminants Based On Actual
Emissions Data. As New Or Refined Data Become
Available, It Can Be Directly Incorporated Into
The Assessment To Obtain Revised Risk Estimates
On A Real Time Basis
Provide Necessary Information To Prioritize And
Identify Solutions, For Sources Resulting In
Unacceptable Risks, At A Community Level Of
Resolution, And Generated In A Fully Transparent
Fashion Such That Risk Levels Are Traceable To
Each Contaminant, Each Pathway, And Each Source
19
5.0 DETERMINING PROJECT OBJECTIVES

• Guiding Factors in Determining Project Objectives
  - USABILITY and UTILITY
• Defensible-relies on approved methods
• Numerically correct and consistent
• Time efficient (month vs. year timeframe to
  complete)
• Cost efficient (tens of thousands vs. hundreds of
  thousands)
• Flexible-analyze variations/what ifs
• Provides interim utility (useful data for
  trending, flags potential problems, etc.)
• Directly applicable to end users needs
• Directly supports solution implementation

20
5.0 DETERMINING PROJECT OBJECTIVES

• Start With the End in Mind

21
Conceptual Model
22
5.0 DETERMINING PROJECT OBJECTIVES

• Some general examples
• Results generated in a manner to support
  consideration of a range of risk management
  alternatives (supports solutions)
• Provides a standardized and consistent means
  to prioritize sources based on risk impacts
  (apples to apples across different project
  areas)
• Each elemental component of risk or modeled
  air concentration is fully traceable to the
  culpable source (attribution can be determined)
  and
• Flexibility to be revised as new and more
  complete emissions data sets become available
  (future utility).

23
Risk - Management and Analysis Platform (RiskMAP)
Project Database

• Emissions Characterization Component
• Integration of Primary Emission Databases
  (DataMiner)
• Data Quality Review and Data Extraction
  (ACCESS)

• Air Modeling Component
• Preprocessing of Air Modeling Inputs (AMP)
• Execute Air Dispersion Model (ISCST3)

• Risk Modeling Component
• Site Assessment and Data Input Processing
  (Risk-MAP)
• Generate Risk Results and Perform Data
  Analysis (Risk-MAP)
• Map and Report Generation (Risk-MAP)
• Solutions Implementation and Performance Review

24
6.0 RAIMI COMPONENTS

• Interactive GIS Platform Architecture
• Architecture is a good match for
  managing/analyzing/presenting multi-layer spatial
  data
• Links major components - emissions
  characterization/air modeling/risk modeling
  components
• Emissions Characterization
• Hierarchy of databases to meet DQOs
• Driven by data management capabilities -
  Dataminer
• Air Modeling on a Universal Grid
• Source specific single-pass approach
• Model flexible ISCST3/Calpuff/Aermod
• Receptor-Based Risk Modeling
• Multi-pathway simplified exposure scenario
• Attribution profiling (source/contaminant/pathway)
  
• Oriented to supporting solutions

25
7.0 GIS PROJECT PLATFORM

• 7.1 Advantages of the GIS Architecture as a
  Project Platform
• Natural match for the types of spatially related
  data used in these types of assessments
• Platform to link emissions characterization/air
  modeling/risk modeling
• Capability to visually manage, analyze, and
  present data and results
• Capacity to handle cumulative type assessments

26
7.0 GIS PROJECT PLATFORM

• 7.2 GIS Platform Construct
• ArcViewTM Version 8.2 core platform to that
  provides operating environment for required
  databases, data management operations and risk
  modeling module (RiskMAP)
• ETD string of inter-linked database tables
  database structure and functionality designed to
  support cumulative type assessments requiring
  large capacity and high resolution of results for
  solution management
• Risk - MAP risk modeling module calculates
  exposure pathway specific values in a spatially
  layered data environment supports capacities
  typically required of cumulative type studies
  and custom visual displaying of interim and final
  results in traditional (tabular, etc.) and mapped
  (isopleths, spatial attributes, attribution
  tracking, etc.) formats to support solution
  consideration, implementation, and tracking

27
8.0 EMISSIONS CHARACTERIZATION

• 8.1 Example Objectives Specific to Emissions
  Characterization
• Obtain necessary data as inputs to complete air
  and risk modeling
• Obtain resolution and quality of data to support
  source-specific prioritization and decision
  making
• Identify and track key source attributes to
  support trending analysis
• Support attribution profiling (source/contaminant/
  exposure pathway)

28
8.0 EMISSIONS CHARACTERIZATION

• 8.1 Approach to Emissions Characterization
• Utilize and cross reference all available
  emissions data sources (national, state, hard
  copy files) to obtain the most complete and
  accurate emissions characterization possible
• Thoroughly research emissions databases to gain a
  better understanding of their usability
• What are the reporting requirements and
  specifications?
• What QC checks are performed, and who conducts
  them?
• For what purpose is the data typically used?
• How is missing data handled?

29
8.0 EMISSIONS CHARACTERIZATION

• 8.1 Approach to Emissions Characterization
• Digest available electronic emissions databases
  to evaluate if data meet completeness and quality
  requirements
• Data needs for source-specific air and risk
  modeling parameters
• DQOs for source-specific air and risk modeling
  parameters
• Effectively manage data to ensure integrity of
  modeled results that are traceable to unique
  modeled sources
• Link emissions data to GIS capabilities of air
  and risk modeling components
• Track emission source attributes (e.g. permit
  status, source type, permit limits, enforcement
  history, etc.) to support trending analysis

30
8.1 EMISSIONS CHARACTERIZATION

• 8.2 Emissions Data Needs
• Source-Specific Parameter Values Required to
  support completion of air modeling component on a
  source-specific basis. Must comply with DQOs for
  air modeling inputs.
• Speciated Emission Rates Required for accurate
  risk modeling, solution implementation, and
  attribution profiling. Generic contaminant
  groupings (total VOCs, gasoline, crude oil, etc.)
  may compromise project objectives.
• Source-Specific Attributes Required to support
  trending analysis and possibly solution
  implementation. Examples may include permit
  status, source type, industry type, enforcement
  history, facility ID, SCC, etc.

31
8.0 EMISSIONS CHARACTERIZATION
8.2 Emissions Data Needs
32
8.0 EMISSIONS CHARACTERIZATION

• 8.3 Emissions Data Sources
• Important Consideration for Use Most, if not
  all, available emissions databases are not
  designed to support risk analysis
• Emissions databases are designed to meet specific
  regulatory reporting requirements, and therefore,
  vary with respect to structure, content, and
  terminology.
• Data source formats
• Digital large data sets can be readily
  incorporated into the study
• Hard copy individual regulatory files or
  facility records may provide critical missing
  data for some sources

33
8.0 EMISSIONS CHARACTERIZATION

• 8.3 Emissions Data Sources
• To combat terminology issues across data sources,
  the following definitions can be used
• Individual Sources those sourcesgenerally
  industrial stack or fugitive sourcesfor which
  the available emissions inventories DO provide
  complete data sets to support source-specific air
  dispersion and risk modeling. (Typically subject
  to regulatory reporting requirements)
• Grouped Sources those sourcesgenerally small
  stack, fugitive, and mobile sourcesfor which the
  available emissions inventories DO NOT provide
  complete data sets to support source-specific air
  dispersion and risk modeling. (Typically not
  subject to regulatory reporting requirements)

34
8.0 EMISSIONS CHARACTERIZATION
35
8.0 EMISSIONS CHARACTERIZATION

• 8.3 Emissions Data SourcesIndividual Sources
• State - Texas Point Source Database (PSDB)
• Repository for an annual survey of facilities
  that meet the TNRCC emissions inventory rule.
  1997 PSDB data used in RAIMI Pilot Study.
• Contains the source-specific values necessary for
  air and risk modeling, meeting RAIMI
  requirements.
• Includes many other pertinent source attributes
  of interest (SCC, actual and allowable emission
  rates, emissions history).
• Reviewed by State.
• NTI incorporates a version of the Texas PSDB as
  its source of Texas point source emissions data.

36
8.0 EMISSIONS CHARACTERIZATION

• 8.3 Emissions Data SourcesGrouped Sources
• National Toxics Inventory (NTI) Area/Mobile
• Incorporates emissions estimates of sources that
  are too small or diffuse to fall under CAA
  emissions reporting requirements (e.g., dry
  cleaners, gas station)
• Organizes data by area and mobile source
  subcategories, which can be prioritized for
  evaluation on the basis of occurrence of the
  surrogate (e.g. railroad miles) within an
  assessment area.
• Emissions factors used to develop estimates are
  well documented.
• NTI data is reviewed by States, industry, and
  other federal agencies.

37
8.0 EMISSIONS CHARACTERIZATION

• 8.3 Emissions Data Sources
• Other Sources
• In addition to the State and NTI databases,
  emissions characterization data can be
  supplemented through other sources, including
• EPA files Permit applications, trial burn
  reports, etc.
• State files Check confidential files.
• On Paper Emission rate data is public
  information
• In Reality Once a submittal is stamped
  Confidential, the emissions rate most likely
  does not make it into the database
• Resolution Manual review of confidential files
  in State offices

38
8.0 EMISSIONS CHARACTERIZATION

• 8.4 Database Tool A Helping Hand
• DATAMINER A large database client-server
  processing system that facilitates the assembly
  of multi-source emissions inventories for air and
  risk modeling by
• Enabling the creation and editing of database
  table relationships and views for complete access
  to all emissions attributes maintained in the
  database
• Linking source-specific parameters necessary for
  air and risk modeling from multiple database
  tables through the Data Organizer function
• Extracting the source-specific data sets through
  the construction and execution of simple or
  complex data queries in the Query Builder
  function

39
8.0 EMISSIONS CHARACTERIZATION

• 8.5 Database Issues
• Generally will not prevent completion of a
  successful localized assessment
• Awareness of critical issues is important
• Maintain focus on project objectives
• Various options (e.g., bounding assessments) can
  be employed to get a idea as to risk-based
  significance

40
8.0 EMISSIONS CHARACTERIZATION

• 8.5 Database Issues
• Inventory Completeness Example
• Different reporting requirements can result in
  substantially different content
• For example, the 1997 PSDB and TRI indicate the
  following
• Do not conclude that PSDB is more complete than
  TRI!

41
8.0 EMISSIONS CHARACTERIZATION

• 8.5 Database Issues
• Inventory Completeness (cont.)
• PSDB butadiene emissions at Ameripol Synpol are
  less than the TRI emissions
• Applying the TRI emission value to the source(s)
  at Ameripol Synpol is one option for evaluating
  the risk-based significance of this data gap

42
8.0 EMISSIONS CHARACTERIZATION

• 8.5 Database Issues
• Speciation Example
• 1997 PSDB reports in the RAIMI Pilot Study
  assessment area
• 5,668 tpy of speciated emissions (58)
• 4,135 tpy of unspeciated emissions (42)
• Pilot Study results based on speciated emissions
• Not a Pilot Study objective to speciate emissions
• To determine significance of the data gap, the
  speciated emissions could be proportionately
  increased to account for the unspeciated
  emissions
• Applying this bounding emission rate to the risk
  model may help to establish the risk-based
  significance of unspeciated emissions.

43
8.0 EMISSIONS CHARACTERIZATION

• 8.5 Database Issues
• Speciation (cont.)
• TNRCC emissions inventory program instructs
  facilities to provide speciation data for 90
  percent of HAP emissions for those sources with
  emissions rates greater than 1 tpy, or 0.1 tpy
  for any individual HAP.
• Adherence to these instructions by facilities
  reporting emissions to the PSDB would
  significantly improve the speciationand thus
  usabilityof the PSDB for risk management
  purposes

44
8.0 EMISSIONS CHARACTERIZATION

• 8.6 Source Selection
• Identifying emissions sources for modeling
• Individual Sources
• In Pilot Study, 113 of 1,529 individual sources
  were modeled (generally, sources with greater
  than 1 tpy of a speciated contaminant).
• Sources were prioritized based on mass of toxic
  emissions.
• Air and risk modeling was conducted in groups of
  10-20 sources.
• Prioritization refined based on toxicity and
  proximity of receptor neighborhoods through
  iterative modeling of prioritized groups.
• Based on pilot study, prioritization of
  individual sources discontinued in favor of using
  more robust modeling capabilities to evaluate all
  sources in the assessment area.

45
8.0 EMISSIONS CHARACTERIZATION

• 8.6 Source Selection
• Grouped Sources
• For each grouped source subcategory, construct a
  worst-case hypothetical emissions scenario that
• County-wide emissions estimates are allocated to
  the smallest census tract
• Air and risk modeling is conducted specific to
  each subcategory
• Those subcategories that exceed the screening
  threshold are prioritized for additional modeling
• For prioritized grouped subcategories
• Allocate emissions estimates to census
  tract-level estimates for each census tract,
  utilizing the appropriate allocation scheme
• Prioritize these results based on air and risk
  modeling for each grouped emissions subcategory

46
8.0 EMISSIONS CHARACTERIZATION

• 8.6 Special Topics
• Actual vs. Allowable Emission Rates
• Modeling only allowable emissions will not
  provide a worst-case result, as most emissions
  sources are not permitted and allowable rates are
  generally not provided for grandfathered or
  exempt sources.
• Modeling only actual rates may underestimate
  potential risk in some cases by not including
  emissions from permitted sources identified in
  the PSDB known to be operating, but with no
  reported actual emissions
• Actual and allowable rates may be used to fill
  emission rate gaps.
• Modeling of allowable rates for permitted sources
  may provide important information to better
  support or assess regional permitting and
  cross-program permitting

47
8.0 EMISSIONS CHARACTERIZATION

• 8.6 Special Topics
• Accounting for Process Upsets Maintenance
• PSDB Actual emissions do not include Upset
  Maintenance event emissions
• State may require UM emissions to be reported,
  but as a separate emission rate
• In 1997 in Jefferson County, there were 38,605
  tons of pollutants released during UM events
  that are not listed as actual emissions in the
  State PSDB
• Applying UM emissions .?

48
8.0 EMISSIONS CHARACTERIZATION

• 8.6 Special Topics
• Sensitivity of Source Location on Modeled Impacts
• Example Monitoring station T-136, located in
  the Port Neches/Nederland neighborhood, is
  located about 750 meters west of a modeled
  1,3-butadiene source (wastewater pond).
• Result A RAIMI modeling analysis of this source
  indicates that if the source location was
  actually adjusted by as little as 200 meters, the
  resulting impact at the T-136 location increases
  by a factor of two.

49
8.0 EMISSIONS CHARACTERIZATION

• 8.6 Special Topics
• What If TRI Was the Only Data Available?
• TRI data is facility-specific as opposed to
  source-specific
• Limiting assumptions must be made regarding
  source locations and source characteristics
• Receptor impacts are highly dependent upon
  correlation of source and receptor locations
• Potential for false negatives very high and
• Not possible to determine numerically reliable or
  source-specific risk.
• TRI data can be used in a generalized screen
  approach by assuming release point at facility
  boundary closest to target receptors
• Useful to obtain better emissions data and
• Can be used as comparison or QC check of other
  emissions databases.

50
RESULTING ENHANCEMENTS Data Miner Database
Mining Tool

• Large capacity datasets (gigabytes)
• Multiple source attribute handling
• Ingest various database formats
• Sophisticated data exporting

51
Data Miner Example Capabilities
52
9.0 AIR MODELING COMPONENT

• 9.1 Example Objectives Specific to Air Modeling
• Ensure adequate data are available to support
  risk modeling
• Minimize the production of unnecessary data so
  that data management resources are not strained
• Accommodate flexibility in the design of
  site-specific risk evaluation and management
  without the need for repetitive air modeling
  events

53
9.0 AIR MODELING COMPONENT

• 9.2 Overview of Air Modeling Approach
• Applies current procedures, regulatory guidance,
  data requirements, and default parameter values
• Draft Human Health Risk Assessment Protocol for
  Hazardous Waste Combustion Facilities (HHRAP)
• 40 CFR Part 51, Appendix W, Guideline on Air
  Quality Models (GAQM)
• Meteorological Processor for Regulatory Modeling
  (MPRM)
• Industrial Source Complex Short Term, version 3
  (ISCST3)
• National Weather Service meteorological data
• USGS/EPA Land Use (LULC) and USGS Digital
  Elevation Mapping (DEM) data

54
9.0 AIR MODELING COMPONENT

• 9.2 Overview of Air Modeling Approach
• Applies single-pass air modeling for each
  source
• Up-front production of all necessary air
  modeling data to support current and anticipated
  future risk modeling needs
• Use of unit emission rates enables one set of
  model runs for modeling each emission source to
  accommodate any combination of emissions
  scenarios (e.g., reported actual emissions,
  permitted allowable emissions, revised quantities
  of emissions due to operational changes, or
  inclusion of new contaminants in the emissions
  profile)
• Phase-specific modeling runs for emissions
  partitioning
• (vapor, particle, particle-bound, mercury)

55
9.0 AIR MODELING COMPONENT

• 9.2 Overview of Air Modeling Approach
• Special considerations of single-pass approach
• No chemical transformation is addressed in the
  single-pass approach. However, based on a
  review of risk results for certain highly
  reactive pollutants using the single-pass
  approach, additional pollutant-specific air
  modeling may be required to implement chemical
  decay (transformation) or secondary formation of
  derivative pollutants.
• Building downwash effects may be significant
  and should be included in the single-pass
  approach for specific sources with nearby
  receptors (within 500 meters), requiring
  gathering of additional information on the
  source and the facility not typically included
  in the emission inventory

56
9.0 AIR MODELING COMPONENT

• 9.2 Overview of Air Modeling Approach
• Applies universal grid
• All data converted and stored in standardized
  geographic coordinate system (NAD 83
  Latitude/Longitude curvilinear) for quantitative
  data processing and qualitative display
• Real-time conversion to any standard coordinate
  system for component-specific data processing
  (e.g., air model component requires UTM
  rectilinear coordinates)
• Point-to-point spatial and scale alignment of
  layered datasets from various database caretakers
  in various coordinate systems unified into single
  universal system

57
RAIMI Air Modeling Approach
58
9.0 AIR MODELING COMPONENT

• 9.2 Overview of Air Modeling Approach
• Modeling of Individual Sources
• Separate air modeling runs conducted specific to
  each individual source modeled
• Source locations reviewed for accuracy through a
  stepwise geo-correction process (small variance -
  dramatic effect on modeled results)
• Source-specific inputs obtained from emissions
  inventory data assignment of surrogate values
  missing physical characteristics data an option
  (except for particle density and size
  distribution)
• A facility may define a fugitive source either as
  a stack or an area source, but in such cases, the
  physical characteristics for the source as
  reported in the emissions inventory are used in
  the air modeling without redefinition

59
9.0 AIR MODELING COMPONENT

• 9.2 Overview of Air Modeling Approach
• Modeling of Grouped Sources
• Define a circle about the centroid of each
  census tract, with the area of the circle equal
  to the area of the tract
• Place five pseudo-stack sources within each
  census tract, one at the center and one to the
  North, East, South and West at a distance of
  one-half of the radius of the circle.
• For each pseudo-source, set ISCST3
  source-specific inputs to represent conditions
  where no plume rise occurs
• Proportionally allocate emission rates to each
  pseudo- source (1/9th at center and 2/9th for
  four other points)

60
9.0 AIR MODELING COMPONENT

• 9.3 Criteria for Selection of Air Model
• Method is flexible to accommodate any air model
  that
• Represents many different source types,
  including stacks, fugitives, flares, area and
  mobile sources
• Supports large grid arrays covering potential
  receptors
• Accounts for variations in local site
  characteristics, including land use and terrain
  elevations
• Accounts for variations in hour-to-hour weather
  conditions for up to 5 years of data near the
  sources
• Addresses dispersion, deposition and removal
  processes
• Is widely-accepted by agencies, industry and
  public

61
9.0 AIR MODELING COMPONENT
9.3 Criteria for Selection of Air
Model Comparison of Model Capabilities
62
9.0 AIR MODELING COMPONENT
9.3 Criteria for Selection of Air Model Modeled
to Monitored Comparison
63
9.0 AIR MODELING COMPONENT
9.3 Criteria for Selection of Air Model

• How long does it take to complete the air
  modeling?
• Pre-processing (ISCST3/AERMOD, CALPUFF 10t)
• 5 days Area-wide weather data acquisition and
  processing
• 5 days Area-wide site characteristics around
  sources
• 5 - 60 min Parameter QC and preparation for each
  source
• Run air model (ISCST3 1 GHz P3 CALPUFF 10t)
• 20 min Each Vapor (v,h) phase for each source
• 2 - 4 hrs Each Particle (p,b) phase for each
  source
• Post-processing
• 5 min Each source for all four phases

64
9.0 Air Modeling Component
9.3 Criteria for Selection of Air Model
What about projects with different number of
sources?
65
9.0 AIR MODELING COMPONENT

• 9.4 Air Modeling Tools
• AMP automates labor-intensive, site-specific
  data preparation and pre-processing to facilitate
  completion of air modeling on multiple sources in
  a reasonable time period by
• Importing GIS data for human geo-location of
  source inventory (source/facility plot files,
  land use, topo map, aerial photos)
• Computing source-specific site parameters
  (surface roughness for 12 30-degree sectors and
  urban/rural land use to 3 km for each source)
• Auto-generates and processes met files for 5
  years of hourly weather data for each source
• Auto-generates ISCST3 input files for all
  four phases (v,p,b,h), including grid node array
  with terrain elevations, wet/dry deposition
• AIR2GIS executes ISCST3 in batch mode, and
  converts plot files to format for import into GIS
  Platform for risk modeling

66
Air Model Pre-processor (AMP)

• Windows Start Programs

67
AMP
68
9.0 AIR MODELING COMPONENT

• 9.5 Air Modeling Inputs What is Needed and Why?
• Source parameters location, size and buoyancy of
  gases and particles at time of release from
  source
• Stacks predominant source type (60-90),
  requiring stack height, temperature, velocity,
  diameter, particle size distribution and density
• Fugitives up to 1/3rd of inventory, but
  inventory typically lacks source definition, so
  assumptions are required on source size, shape
  and phase (volatiles as vapors)
• Flares usually less than 5 of inventory,
  modeled as stacks with pseudo-diameter computed
  from heat of combustion (requires release height,
  low heat value and molecular weight)
• Area may have significant emissions from
  estimated releases over a defined area (e.g.,
  census tracts) based on geo-political data
• Mobile significant emissions for dense
  populations that may be treated as area sources,
  or roadway-specific line segments

69
9.0 AIR MODELING COMPONENT

• 9.5 Air Modeling Inputs What is Needed and Why?
• Meteorological Data
• Surface data hourly measurements of wind
  direction, wind speed, air temperature, cloud
  cover, cloud height, precipitation amount and
  type, solar radiation and leaf area index
• Used in air model to compute spatial and
  temporal distribution, and magnitude, of ambient
  concentrations in air (inhalable), and dry
  deposition and wet deposition at ground
  (impaction, adsorption)
• Typically extract SAMSON data obtained from NCDC
  on CD-rom
• Upper air data (aka, mixing height) twice daily
  measurements of wind speed and temperature
  profile in air above the surface
• Used in air model to compute the top of the
  mixing layer above the ground, or boundary layer,
  where dispersion processes occur
• Typically download from US EPA SCRAM web site

70
9.0 AIR MODELING COMPONENT

• 9.5 Air Modeling Inputs What is Needed and Why?
• GIS Data
• Land Use Land Cover (LULC) best available
  spatial distribution of LULC should be obtained
  electronically (shapefiles) within 3 km of each
  source used for surface roughness and urban/rural
  factors
• Source plot file shapefile with associated
  database file for every source location in UTM 83
  (and Lat/Long 83), including source parameters
  for input into Charview for geo-location and
  input files
• USGS topographic files shapefile of topographic
  features for use in source geo-location using AMP
• USGS DEM 90-meter (aka, 1-degree, 3 arc-second)
  files - covering maximum extent of grid node
  arrays for elevations of grid nodes
• Aerial photographs may be useful in source
  geo-location or extracting building dimensions
• Facility boundary files used to geo-locate
  sources on-property

71
9.0 AIR MODELING COMPONENT

• 9.5 Air Modeling Inputs What is Needed and Why?
• Grid Node Array (Universal Grid)
• Coordinate System ISCST3 requires UTM
  coordinates (CALPUFF also allows
  latitude/longitude). Recommend using North
  American Datum 1983 (NAD83) for consistency
  with grid locations in USGS DEM data. Most USGS
  topographic maps are printed hardcopy in NAD27,
  which differ from NAD83 locations by about 200
  meters northerly and 50 meters easterly.
• Grid centroid Recommend a universal grid
  centered on the grid node in the USGS DEM
  90-meter data (aka, 3 arc second, 1-degree)
  closest to the source location (NAD83). Grid
  center will be within 45 meters of the source
  location and produce air model results at the
  same locations for overlapping grid arrays from
  other sources within 20 km.
• Coverage use every DEM grid location out to 3
  km (spacing from 6090 meters) and every 5th DEM
  grid location from 3-10 km (spacing 300-500
  meters)

72
UTM Zones (US)
73
9.0 AIR MODELING COMPONENT

• 9.6 Special Topics on Air Modeling
• What is the impact of chemical decay and
  secondary formation?
• Based on chemical decay rates identified in the
  Cumulative Exposure Project (CEP), only highly
  reactive pollutants (e.g., 1,3-butadiene) decay
  (by up to 10) within 10 km. Similarly,
  secondary formation (e.g., formaldehyde) of
  pollutants only occurs within 10 km for highly
  reactive precursors (amounts up to 10). These
  values may be significant for highly toxic
  pollutants, or where large quantities of highly
  reactive pollutants are emitted.
• What about impacts further than 10 kilometers?
• The highest impacts from a source or group of
  sources occur within 3 km of the release and will
  be identified by the methods described. For
  impacts further than 10 km, individual grid nodes
  may be put into the ISCST3 model for specific
  Receptor scenarios of concern.

74
9.0 AIR MODELING COMPONENT

• 9.6 Special Topics on Air Modeling
• If the inventory only reports the facility
  location and not individual source locations,
  should we use the facility location for all
  sources and sum the emissions?
• Only if errors of 101-103 are acceptable.
• Results are very dependent on the source-receptor
  relationship.
• Dilution increases exponentially from the point
  of release to the receptor. Most dilution occurs
  in the first 100-300 meters after release with
  reductions up to 1/10th of initial concentration.
  By 1-3 km from the release, air concentrations
  and deposition rates typically decrease by 10-2
  to 10-4.
• Most facilities have property dimensions of
  several hundred meters to several kilometers.
  Errors in results will be several orders of
  magnitude due solely to the source location input
  to the model.

75
9.0 AIR MODELING COMPONENT

• 9.6 Special Topics on Air Modeling
• What should I do if I dont have building
  dimensions?
• Building dimensions are almost never available in
  inventories. But, always use building data in the
  air model when it is available.
• Building downwash affects air model results most
  significantly within 500 meters from the source.
  The impacts typically increase for taller stacks
  (about 20 meters or higher), but impacts may
  decrease for shorter stacks (less than 20
  meters).
• When Receptors of concern are located within 500
  meters from a stack (only stacks have downwash in
  ISCST3), the air model could be run without
  building dimensions. If risk would be adverse at
  levels 101 to 102 higher than calculated without
  building downwash, and there is potential for
  building downwash based on structures near the
  stack, then building data should be obtained to
  assess impacts with downwash.

76
9.0 AIR MODELING COMPONENT

• 9.6 Special Topics on Air Modeling
• When should I consider using CALPUFF instead of
  ISCST3?
• CALPUFF is an advanced, state-of-the-science air
  dispersion model that requires significant
  expertise in pre-processing the meteorological
  data prior to performing air modeling.
• It is a non-steady-state model that tracts
  individual air parcels spatially and temporally
  for better representation of actual flows in the
  air shed.
• Pre-processing and model execution times are
  10-100 times longer than ISCST3.
• However, if stagnate winds (calm conditions
  occurring for extended periods), long range
  transport beyond 20-50 km, recirculation of air
  flow due to terrain or sea breeze, or regions of
  terrain above stack top are of concern, CALPUFF
  is the recommended US EPA model to address these
  issues and is compatible with the HHRAP methods.

77
9.0 AIR MODELING COMPONENT

• 9.6 Special Topics on Air Modeling
• What other resources are available on air
  modeling?
• US EPA Region 6 Air Modeling Audit Checklist
• ww.epa.gov/earth1r6/6pd/rcra_c/protocol/audit.p
  df.
• US EPA Region 6 Sensitivity Analysis of Air
  Modeling Parameters
• www.epa.gov/earth1r6/6pd/rcra_c/protocol/analys
  is.pdf
• US EPA Air Quality Models (and mixing height
  data)
• www.epa.gov/ttn/scram/

78
10.0 RISK MODELING COMPONENT

• 10.1 Example Objectives Specific to Risk Modeling
• Results generated in a timely and interactive
  manner so as to actually be useful in day-to-day
  permitting, planning, and enforcement activities
  (being in a position to effectively work with
  regulators, industry, and public to address
  exposure priorities and implement solutions)
• Results generated at a level of resolution and
  traceability to serve as an asset to stakeholders
  needing to evaluate and implement solutions
  (source-specific decision making)
• Provide a standardized and consistent means to
  conduct risk-based assessment and prioritization
  of multiple emissions sources of multiple
  contaminants from multiple facilities
• Provide readily accessible risk-based
  prioritization tools and project platform
  designed to address multi-media solutions to
  environmental problems
• Capacity to calculate and track potential risks
  from hundreds of sources in a fully transparent
  manner

79
10.0 RISK MODELING COMPONENT

• 10.2 Overview of Risk Modeling Approach
• Applies existing and current risk procedures,
  equations, and default parameter values
• Draft Human Health Risk Assessment Protocol for
  Hazardous Waste Combustion Facilities (HHRAP)
• Methodology for Assessing Health Risks Associated
  with Multiple Pathways of Exposure to Combustor
  Emissions
• Risk Assessment Guidance for Superfund Volume
  IHuman Health Evaluation Manual (RAGS)
• Exposure Factors Handbook
• Integrated Risk Information System (IRIS)

80
10.0 RISK MODELING COMPONENT

• 10.2 Overview of Risk Modeling Approach
• Implements true receptor-based approach
• Focuses on defining impacts to target receptor
  locations or neighborhood area.
• Each target neighborhood can have customized
  exposure inputs that may influence results,
  management decisions, and communication.
• Structure allows data to be generated and managed
  at the neighborhood or receptor level while
  maintaining coverage over large geographic areas
  (e.g., county, state, region).
• Foundation of the dynamic project platform.

81
10.0 RISK MODELING COMPONENT

• 10.2 Overview of Risk Modeling Approach
• Applies simplified exposure scenarios
• For example, default exposure assumptions and
  inputs for inhalation pathway based on an
  individual breathing outdoor air concentrations
  24 hours per day, 350 days per year
• Pathway driven (e.g., inhalation)
• Does not include indoor air
• Does not include micro-exposure activity patterns

82
10.0 RISK MODELING COMPONENT

• 10.2 Overview of Risk Modeling Approach
• Utilizes true GIS Architecture
• Risk Modeling component fully integrated within
  GIS environment direct linkage to Emissions
  Characterization and Air Modeling Components
• Capability of tracking source attributes e.g.,
  emission scenario, location, source type, permit
  status, etc.
• Capability of attribution profiling e.g.,
  contaminant, pathway, source
• Effective management of macro and micro data sets
• Increased capacity through efficiency and
  automation
• Significant flexibility to modify or replace risk
  modeling inputs (pathway, receptor, source, FT,
  toxicity) to tailor the risk modeling to
  site-specific conditions and concerns

83
10.0 RISK MODELING COMPONENT

• 10.3 Criteria for Selection of Risk Modeling
  Tools
• Risk model selection focuses on the ability to
• Directly import air parameter values from air
  model
• Implement multi-pathway exposure setting
  (scenarios) with flexibility to add or delete
  pathways
• Provide attribute tracking capabilities
• Capacity to calculate and track risks from
  multiple (hundreds) sources and contaminants
  specific to each exposure pathway and location
  (attribution profiling)

84
10.0 RISK MODELING COMPONENT

• 10.3 Criteria for Selection of Risk Modeling
  Tools
• Risk model selection focuses on the ability to
• Interchange source-specific emission rate
  scenarios
• Customize exposure parameter input values
  specific to each receptor location or
  neighborhood
• Manage changes to fate and transport or toxicity
  parameter input values specific to site and
  contaminant
• Support expanded presentation and graphical
  evaluation of risk results
• Flexible, integrates with other software
  components and tools capacity to handle very
  large amounts of data (gt2GB)

85
10.0 RISK MODELING COMPONENT
Comparison of Risk Model Capabilities
86
10.0 RISK MODELING COMPONENT

• 10.4 Risk Modeling Tools Being Implemented
• Project Platform - RiskMAP (ArcView V8.2)
• Import air dispersion modeling data
• Support receptor based approach
• Perform multi-pathway risk calculations
• Direct and full GIS functionality and
  presentation
• Confirm source locations and other spatial
  queries
• Present risk results/map generation
• Manage and update supporting background maps
• Supporting Tools - ACCESS/Excel
• Interact with smaller databases
• Perform non-complex calculations and data queries
• Format data outputs or extract specific results

87
10.0 RISK MODELING COMPONENT
10.4 Risk Modeling Tools

• How long does it take? (Based on 100 emission
  sources)
• Pre-processing (Set up project files and import
  data)
• 1 d Import air modeling output
• 2 d Import emissions data/background maps
• 5 d Receptor analysis/define receptor target
  areas
• 1 d Import site characterization data
• Run risk model (800 MHz P3 256K RAM)
• seconds Total run time
• Post-processing for GIS presentation
• seconds Export all risk reports and mapping

88
10.0 RISK MODELING COMPONENT

• 10.5 Risk Modeling Inputs What is Needed and
  Why?
• Emissions Characterization Data
• Facility and source-specific attributes
• Examples include Facility SIC, source type,
  production description, process type, permit
  status, permit type, etc.
• Required to support trending analysis,
  attribution profiling, risk management
  objectives, source prioritization
• Speciated source-specific emissions data
• Units - mass per unit time requirement of one
  pass air modeling approach
• Emission rate speciated to individual chemical or
  CAS level (i.e., no generic groupings or
  mixtures)

89
10.0 RISK MODELING COMPONENT

• 10.5 Risk Modeling Inputs What is Needed and
  Why?
• Air Modeling Output Data
• Pre-processed AIR2GIS file
• Source-specific unitized air dispersion modeling
  parameters and facility and source-specific
  attributes
• Source-specific universal receptor grid node
  array

90
10.0 RISK MODELING COMPONENT

• 10.5 Risk Modeling Inputs What is Needed and
  Why?
• Exposure and Site-Specific Parameters
• Implementation of HHRAP defaults
• Examples inhalation and consumption rates,
  exposure duration, direct and indirect exposure
  pathways, quantity and type of food ingested by
  livestock, percent contaminated
• Parameters without HHRAP defaults
• Average annual evapotransporation, irrigation,
  rainfall, runoff and windspeed
• Flexibility to include site-specific data at a
  regional, neighborhood or location specific level.

91
10.0 RISK MODELING COMPONENT

• 10.5 Risk Modeling Inputs What is Needed and
  Why?
• Fate and Transport Parameters
• Implementation of HHRAP defaults
• Flexibility to incorporate site-specific data
• Flexibility to add additional contaminants

92
10.0 RISK MODELING COMPONENT

• 10.5 Risk Modeling Inputs What is Needed and
  Why?
• Toxicity Parameters
• Incorporation of IRIS, NCEA Values, HEAST
• Flexibility to add new or updated values
• Inhalation Pathway exposure air concentrations
  are combined with inhalation toxicity factors
  (RfC or URF) to estimate noncarcinogenic hazards
  and carcinogenic risks, respectively.

93
10.0 RISK MODELING COMPONENT

• 10.5 Risk Modeling Inputs What is Needed and
  Why?
• GIS Data Background Maps
• Land Use Land Cover (LULC) provide land use
  characteristics for use in identifying
  neighborhoods, exposure scenarios, and exposure
  scenario locations.
• USGS topographic files provide additional
  attribute information including elevations,
  roads, water features, and general facility
  boundaries.
• Aerial photographs used in combination with
  LULC and topographic maps. Aerials are usually
  more current and are used to confirm and update
  information from other background maps.
• Facility boundary files used to identify
  facilities, boundaries and helpful when verifying
  source locations.

94
10.0 RISK MODELING COMPONENT

• 10.6 Special Topics on Risk Modeling
• When should I consider using ACCESS or EXCEL as a
  risk model?
• Building simple exposure equations/models or
  analyzing results from small projects
• Performing validation or QC checks
• Plotting small scale isopleths
• Consideration of specialty exposure scenarios
  (e.g., dermal exposure)

95
How to Use Results
10.0 RISK MODELING COMPONENT

• 1. Conduct Risk-Based Prioritizations
• RAIMI can be used to conduct risk-based
  prioritizations specific to contaminants,
  facilities, emission sources, and data gaps.
• For example, Pilot Study results indicate for a
  profiled neighborhood
• 1,3-Butadiene is a risk concern
• Identified emission sources at the Huntsman and
  Ameripol Synpol facilities are the most
  significant known contrib