Environmental Benefits Mapping and Analysis Program BenMAP User's Manual 2008


User's Manual

September 2008

Prepared for
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, NC
Neal Fann, Project Manager

Prepared by
Abt Associates Inc.

-------
Table of Contents

1	Welcome to BenMAP	6

1.1	Overview of BenMAP & Benefits Assessment	8

1.2	How to Use this Manual,	10

1.3	Computer Requirements	11

1.4	Installing BenMAP,	12

1.4.1	MSDE Installation Issues	13

1.4.2	Upgrading to a new version of BenMAP	13

1.5	Uninstalling BenMAP	14

1.6	Contacts for Comments, Questions & Bug Reporting	15

1.7	Sources for More Information	15

1.8	Frequently Asked Questions (General)	16

2	Terminology and File Types	21

2.1	Common Terms	22

2.2	File Types 	30

3	Overview of BenMAP Components	31

3.1	One-Step Analysis	33

3.1.1	One-Step Reports	34

3.1.2	Questions Regarding the One-Step Analysis	38

3.2	Custom Analysis	42

3.2.1	Create Air Quality Grids	43

3.2.2	Create and Run Configuration 	44

3.2.3	Aggregation, Pooling, and Valuation	46

3.2.4	Generate Reports	49

3.3	Menus 	52

3.3.1 Tools 	52

4	Loading Data	55

4.1 Add, Modify, and Delete Setups	56

4.1.1	Grid Definitions	58

Regular Grid	60

Shapefile Grid	61

4.1.2	Pollutants 	63

Hourly Metrics	66

Seasonal Metrics	70

Advanced Pollutant Options	72

4.1.3	Monitor Data	73

Format Monitor Data	76

2

September 2008

-------
Contents

4.1.4	Incidence/Prevalence Data	79

Format Incidence/Prevalence Data	83

4.1.5	Population Data	84

Format Population Data	91

4.1.6	Health Impact Functions	92

Format Health Impact Functions	96

4.1.7	Variable Data	99

Format Variable Data	103

4.1.8	Inflation Data	104

Format Inflation Data	105

4.1.9	Valuation Data	106

Format Valuation Data	109

4.1.10	Income Growth Data	110

Format Income Growth Data	114

4.1.11	QALY Data	115

Format QALY Data	117

4.2	Export and Import Setups	118

4.2.1	Export Setups	118

4.2.2	Import Setups	121

4.3	Questions Regarding Loading Data	122

5	Air Quality Grid Creation	124

5.1	Model Direct	126

5.2	Monitor Direct	128

5.2.1	Closest Monitor for Monitor Direct	130

5.2.2	Voronoi Neighbor Averaging (VNA) for Monitor Direct	130

5.2.3	Fixed Radius for Monitor Direct	133

5.2.4	Custom Monitor Filtering for EPA Standard Monitors	133

5.3	Monitor and Model Relative	138

5.3.1	Spatial Scaling	139

5.3.2	Temporal Scaling	140

5.3.3	Spatial and Temporal Scaling 	140

5.3.4	Examples 	141

5.4	Monitor Rollback	143

5.4.1 Example: Combining Three Rollback Approaches in Different Regions	148

5.5	Questions Regarding Creating Air Quality Grids	151

6	Incidence Estimation	153

6.1	Introduction to Estimating Health Incidence	154

6.2	Create Health Impact Configuration,	155

6.2.1	Configuration Settings	155

6.2.2	Select Health Impact Functions	157

6.3	Run Health Impact Configuration	160

6.4	Open & Modify Existing Health Impact Configuration	161

6.5	Questions Regarding Health Impact Configurations	161

3

September 2008

-------
7	Aggregation, Pooling, and Valuation	163

7.1	Introduction to Valuation, Discounting, and Pooling	164

7.1.1	Overview of Valuation	164

Valuing Reductions in Premature Mortality	166

7.1.2	Overview of Discounting 	166

7.1.3	Overview of Pooling	171

7.2	Create Aggregation, Pooling, and Valuation (APV) Configuration	171

7.2.1	Pooling Incidence Results	172

Example: Simple Sorting & Pooling	178

Example: Multiple Pooling Windows	181

7.2.2	Valuing Pooled Incidence Results	182

7.2.3	Using QALYs Weights	187

7.2.4	APV Configuration Advanced Settings	191

Aggregation & Pooling	192

Currency & Income	194

7.3	Run APV Configuration	196

7.4	Open & Modify Existing APV Configuration	196

7.5	Questions Regarding APV Configurations,	197

8	Reports	199

8.1	One-Step Reports,	200

8.1.1	One-Step Incidence Reports	201

Table of Mortality Incidence	201

Table of Morbidity Incidence	202

Box Plot of Mortality Incidence	203

8.1.2	One-Step Valuation Reports	204

Table of Mortality Valuation	204

Table of Morbidity Valuation	205

Box Plot of Mortality Valuation	206

Cumulative Distribution Functions	207

Bar Chart of Valuation	208

8.1.3	One-Step Audit Trail	209

8.1.4	Load One-Step Results to Generate Reports 	210

8.1.5	Export One-Step Reports	212

8.2	Incidence, Valuation, & QALY - from APVR file	213

8.2.1	Incidence Results	215

8.2.2	Aggregated Incidence Results	217

8.2.3	Pooled Incidence Results	218

8.2.4	Valuation Results	219

Add Sums Button	220

8.2.5	Aggregated Valuation Results	221

8.2.6	Pooled Valuation Results	222

8.2.7	QALY Results	223

8.2.8	Aggregated QALY Results	224

8.2.9	Pooled QALY Results	225

8.2.10 Export Incidence, Valuation, & QALY Reports	226

8.3	Raw Incidence - from CFGRfile	227

8.3.1 Export Raw Incidence	228

4

September 2008

-------
Contents

8.4	Audit Trail Reports - from all BenMAP files	229

8.4.1 Export Audit Trail	230

8.5	Questions Regarding Creating Reports	231

9	GIS/Mapping	233

9.1	Overview of Mapping	234

9.1.1	Display Options	235

9.1.2	Taskbar Buttons	237

9.2	Mapping from the Tools Menu	239

9.2.1	Air Quality Grids	241

9.2.2	APV Configuration Results	243

9.2.3	Configuration Results	247

9.2.4	Modeling Data	251

9.2.5	Monitoring Data	252

9.2.6	Population Data	256

9.2.7	Mapping Multiple Layers of Data	257

9.3	Viewing Maps in a BenMAP Analysis	259

9.3.1	Mapping Air Quality Deltas	260

9.3.2	Mapping Monitor Direct Air Quality Grids	261

9.3.3	Mapping Air Quality Monitors from the Advanced Monitor Filter	263

9.3.4	Mapping Monitor Rollback Inputs & Outputs 	266

9.4	Questions Regarding Mapping	268

10	Tools Menu	270

10.1	Air Quality Grid Aggregator	272

10.2	Model File Concatenator	274

10.3	Export Air Quality Grid	278

10.4	Neighbor File Creator	280

10.5	One-Step Setup	282

5

September 2008

-------
Welcome to BenMAP

CHAPTER 1

Welcome to
BenMAP

In this chapter...

>Find an overview of the model.

>Find installation instructions.

>Find contacts, sources for more information and answers
to frequently asked general questions.

6

September 2008

-------
Welcome to BenMAP

HP*

Tools Help

Two Wavs to Use BenMAP: Which Analvsis Meets vour Needs?

One-Step Analysis

Custom Analysis

After you import the air quality data
for your area, use this tool to apply
default settings and create a report.

Step 1 — Import air quality data

Air Quality Grid Creation

Step 2 - Set custom parameters

Preloaded
EPA paramslers

-0-

;"h.

Incidence Estimation

Step 3 — Use results from Step 2
to set custom parameters

'.illl

Pooling. Aggregation
and Valuation

Step 4 — Run report

Report

Lull

Active Setup:

[IMt3!laEll33 t

BenMAP is a powerful, yet easy-to-use program that helps you estimate human health
benefits resulting from changes in air quality. The U.S. Environmental Protection Agency
(U.S. EPA) originally developed this tool to analyze national-scale air quality regulations,
including the National Ambient Air Quality Standards for Particulate Matter (2006) and
Ozone (2008) as well as the Locomotive Marine Engine Rule (2008). To accommodate a
wider range of potential users interested in simplified analyses, local/regional analyses, or
international analyses, EPA developed a more flexible version of BenMAP that includes
options for a one-step analysis or a custom analysis.

A wide range of users, including scientists, policy analysts, and decision makers, can
utilize BenMAP to answer many policy questions, such as estimating the benefits from
previous reductions in air pollution emissions or the benefits of new requirements to

September 2008

-------
Welcome to BenMAP

control air pollution. Advanced users can explore a wide range of options in the custom
analysis, such as adding new health impact and valuation functions, mapping results, and
performing sensitivity analyses. Beginning users can take advantage of EPA's
pre-programmed settings and reports in the one-step analysis.

1.1 Overview of BenMAP & Benefits Assessment

BenMAP is primarily intended as a tool for estimating the human health effects and
economic benefits associated with changes in ambient air pollution. The improvements in
human health as a result of air pollution control regulations are typically referred to as the
benefits of the regulations. As part of the process of developing new regulations,
government agencies are typically required to assess the benefits and the costs of that
regulation. Essentially, benefit analysis develops monetary values to inform the policy
making process and allows decision makers to directly compare costs and benefits using
the same measure (i.e., dollars). BenMAP is a tool that can support these types of benefit
analysis.

BenMAP estimates benefits from improvements in human health, such as reductions in
premature mortality, heart attacks, chronic respiratory illnesses, and other adverse health
effects. Other benefits of reducing air pollution (i.e., visibility and ecosystem effects) are
not quantified in the current version of BenMAP. After estimating the reductions in
adverse health effects, BenMAP calculates the monetary benefits associated with those
reductions.

How does BenMAP estimate human health effects?

First, BenMAP determines the change in ambient air pollution. Because BenMAP does
not include an air quality model, this data must be input into BenMAP as modeling data or
generated from air pollution monitoring data pre-loaded in BenMAP. Next, BenMAP
calculates the relationship between the pollution and certain health effects (also known as
health endpoints). This step is often referred to as the health impact function or the
concentration-response (C-R) function in epidemiology studies. Finally, BenMAP applies
that relationship to the exposed population.

A simplified example is shown below.

Health Effect = Air Quality Change * Health Effect Estimate * Exposed Population *
Health Baseline Incidence

> Air Quality Change. The air quality change is the difference between the starting air
pollution level, (i.e., the baseline), and the air pollution level after some change, such as
a new regulation (i.e., the control).

/'Health Effect Estimate. The health effect estimate is an estimate of the percentage
change in an adverse health effect due to a one unit change in ambient air pollution.
Epidemiological studies provide a good source for effect estimates.

September 2008

-------
Welcome to BenMAP

/'Exposed Population. The exposed population is the number of people affected by the air
pollution reduction. The government census office is a good source for this information.
In addition, private companies may collect this information and offer it for sale.

/'Health Baseline Incidence. The health incidence rate is an estimate of the average
number of people that die in a given population over a given period of time. For
example, the health incidence rate might be the probability that a person will die in a
given year. Health incidence rates and other health data are typically collected by the
government. In addition the World Health Organization is a good source for this. See
World Health Organization

How does BenMAP estimate the economic value of human health effects?

BenMAP also calculates the economic value of avoided health effects. After calculating
the health changes, you can estimate the economic value by multiplying the reduction of
the health effect by an estimate of the economic value per case:

Economic Value = Health Effect * Value of Health Effect

There are several different ways of calculating the value of the health effect. For example,
the value of an avoided premature mortality is generally calculated using the Value of
Statistical Life (VSL). The value of a statistical life is the monetary value that people are
willing to pay to slightly reduce the risk of premature death. For other health effects, the
medical costs of the illness may be the only valuation data available. The BenMAP
database includes several different functions for VSL and valuation functions for other
health effects for you to choose, or you can rely on EPA's preloaded selections.

Figure 1-1 summarizes the basic steps in BenMAP. This figure shows the types of choices
that you make regarding the modeling of population exposure, the types of health effects
to model, and how to place an economic value on these health effects. Please note that
BenMAP does not have air modeling capabilities, and instead relies on modeling and
monitoring inputs.

September 2008

-------
Welcome to BenMAP

Figure 1-1. BenMAP Flow Diagram

Census
Population
Data

Population
Projections

Air Quality
Monitoring

Population
Estimates

Air Quality
Modeling

Population
Exposure

Health
Functions

Incidence and
Prevalence
Rates

Valuation
Functions

Adverse
Health Effects

Economic
Costs

Black text represents user input
Blue text represents result from inputs

What else can BenMAP do?

BenMAP incorporates a Geographic Information System (GIS), allowing users to create,
utilize, and visualize maps of air pollution, population, incidence rates, incidence rate
changes, economic valuations, and other types of data.

BenMAP can thus be used for a variety of purposes, including:

>	Generating population/community level ambient pollution exposure maps;

>	Comparing benefits associated with various regulatory programs;

>	Estimating health impacts and costs of existing air pollution concentrations;

>	Estimating health benefits of alternative ambient air quality standards; and

>	Performing sensitivity analyses of health or valuation functions, or of other inputs.

Chapters 1 through 9 of this manual provide step-by-step instructions on how to use
BenMAP.

New users should start with Chapters 1 and 2, which are both very short, but provide a
good overview of the model and how it works, and explain some potentially confusing
terminology. In addition, Appendix A has some useful training exercises. Use the rest of

1.2 How to Use this Manual

10

September 2008

-------
Welcome to BenMAP

the manual to answer any specific questions you may have, or to walk you step-by-step
through the various model functions. Chapter 3 discusses how to enter data into BenMAP,
Chapters 4 through 7 cover each of the four main buttons, and Chapter 8 covers mapping.

Each chapter is introduced by a short section which describes what you can find within the
chapter, and provides an outline of the chapter's contents. This is a good place to go if the
Table of Contents does not provide enough detail for you to find the section you need.
The end of most chapters has a series of "Frequently Asked Questions," which may also
be helpful in answering specific questions. In sections that provide instructions on
navigating the model, the following conventions are observed: menu items, buttons, and
tab and selection box labels are in bold type; prompts and messages are enclosed in
quotation marks; and drop-down menu items, options to click or check, and items that
need to be filled in or selected by the user are italicized. Throughout the chapters you will
also see boxes presenting common mistakes and important things to remember when
working with BenMAP.

There is also a set of Technical Appendices to provide more detailed information on model
functions, data, and underlying assumptions. In particular note Appendix A, which has two
training modules to help you get started.

Appendix A: Training Courses

Appendix B: Monitor Rollback Algorithms

Appendix C: Air Pollution Exposure Estimation Algorithms

Appendix D: Deriving Health Impact Functions

Appendix E: Health Incidence & Prevalence Data in U.S. Setup

Appendix F: Particulate Matter Health Impact Functions in U.S. Setup

Appendix G: Ozone Health Impact Functions in U.S. Setup

Appendix H: Health Valuation Functions in U.S. Setup

Appendix I: Population & Other Data in U.S. Setup

Appendix J: Uncertainty & Pooling

Appendix K: Command Line BenMAP

Appendix L: Function Editor

References

1.3 Computer Requirements

The computer hardware requirements for BenMAP are typically modest, though this will
again vary depending on the complexity of the analysis. BenMAP requires a Windows
platform, and can be used on machines running Windows 2000, as well as more recent
versions of Windows. In particular, BenMAP requires a computer with:

> Windows 2000 or greater.

> Adobe Acrobat Reader

>128 megabytes of RAM or greater. In general, the amount of RAM required depends on the
scope of the Grid Definitions defined for the model (specifically, the number of cells within the
Grid Definitions). If the largest grid contains hundreds of cells, 128 MB is required. If the

September 2008

-------
Welcome to BenMAP

largest grid contains thousands of cells, 256-512 MB is required. If the largest grid contains tens
of thousands of cells, 1 GB is required.

> Intel® or compatible processor, Pentium 166 MHz or higher. 1 GHz processor or greater
recommended for optimum performance.

> A CD-ROM drive for CD based installation. Alternatively, an internet connection might be used
to download the (-75 MB) installer.

>75 MB free hard disk space, plus room for all data loaded into the application. At least 2 GB
free space recommended for U.S. version.

1.4 Installing BenMAP

BenMAP 3.0 uses the Microsoft SQL Server Desktop Edition (MSDE) as its underlying
database technology. As such, the installation process for BenMAP 3.0 has two steps -
installing MSDE, and installing BenMAP 3.0.

Get started by double clicking Setup.exe in your installation directory. This will bring up
the following dialog:

£ BenMAP 3.0 US/Regional Version Insta... ^ D| X

tnMrcfmenfal Benefits M3[j0rrq and Analysts fto^win

Install BenMAP 3.0

Database

Install BenMAP 3.0

Click on the Install BenMAP 3.0 Database button. This will bring up the MSDE installer.
Normally this installer will complete its work without any input from you. In certain
cases, however, it may require you to reboot your computer after it is finished. If this
occurs, simply reboot your computer and then return to the BenMAP 3.0 installation
process by double clicking Setup.exe in your installation directory again.

Additionally, there are two known issues with the MSDE installer. If you have trouble
with this installer, see the additional notes below.

Once MSDE is installed, the installation dialog will return, with the second button
enabled.

September 2008

-------
Welcome to BenMAP

w BenMAP 3.0 US/Regional Version Insta... _><_

En^r
-------
Welcome to BenMAP

1.5 Uninstalling BenMAP

To uninstall BenMAP, go to Control Panel, Add/Remove Programs and remove both
BenMAP 3.0 and Microsoft SQL Server Desktop Edition (BenMAP30). Note that
uninstalling BenMAP does not also remove any files that you have created with BenMAP.

When choosing to remove BenMAP 3.0, the Add or Remove Programs window should
look something like the following:

a Add or Remove Programs

Change or
Remove
Programs

Add New
Programs

Add/Remove

Windows
Components

Set Program
Access and
Defaults

Currently installed programs:

~ Show updates

Sort by: Name

** Apple mod lie Device Support

Size

34.03MB

^ Apple Software Update

Size

2.14MB

® ArcView GIS 3.2

Size

52.55MB

I® BenMAP 3.0.4

Size

107.00MB

Click here for support information,

Used

frequentlv

To remove this program from your computer, click Remove.

j§' Broadcom Driver Installer

& Check Point VPN-1 Securedient NGX R60 HFA1

Size

2.43MB

® CloneCD

Size

5.44MB

# Conexant D850 56K V.9x DFVc Modem

Size

0.51MB

¦" Dell Media Experience

Size

65.25MB

Dell ResourceCD

Size

2.74MB

» . .. —.

When choosing to remove Microsoft SQL Server Desktop Edition, the Add or Remove
Programs window should look something like the following:

September 2008

-------
Welcome to BenMAP

» Add or Remove Programs

Change or
Remove
Programs

Add New
Programs

Add/Remove

Windows
Components

Set Program
Access and
Defaults

Currently installed programs:

~ Show updates

Sort by: Name

"7 Microsoft Compression Client Pack 1.0 tor Windows XP

® Microsoft Office XP Standard

Size

477.00MB

0 Microsoft SQL Server Desktop Engine (BENMAP30)

Size

69.07MB

Click here for suooort information,

Used

rarelv

To remove this program from your computer, click Remove,

$ Microsoft User-Mode Driver Framework Feature Pack 1.0

M MSN Music Assistant

# MSXML 4,0 SP2 (KB927978)

Size

2.56MB

# MSXML 4,0 SP2 (KB936181)

Size

2.62MB

O POPGRID

Size

2,225.00MB

& PowerDVD 5.1

Size

1.78MB

@ Quicken 2006

Size

74.37MB

Quicken WillMaker Plus 2006

Si7R

16.45MR

1.6 Contacts for Comments, Questions & Bug Reporting

For comments and questions, please contact Neal Fann at the United States Environmental
Protection Agency.

Address: C339-01, U.S. EPA Mailroom, Research Triangle Park, NC 27711
Email: fann.neal@epa.gov
Telephone: 919-541-0209

Alternatively, you can send a message at the BenMAP website:

http://www.epa.gov/air/benmap/contact.html, or by simply emailing benmap@epa.gov.
To report programming bugs in BenMAP:

>Community Modeling and Analysis System (CMAS) Center at the University of North
Carolina at Chapel Hill, BenMAP Bugzilla website. Available at:
http://benmap-model.org/

1.7 Sources for More Information

For files that you can use in BenMAP:

>U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards
(OAQPS), BenMAP website. Available at: http://www.epa.gov/air/benmap/

September 2008

-------
Welcome to BenMAP

For more information on conducting benefit analysis, see the following documents:

>U.S. EPA (1999). OAQPS Economic Analysis Resource Document. Office of Air
Quality Planning and Standards, Innovative Strategies and Economics Group. Available
at: http://www.epa.gov/ttn/ecas/analguid.html

>U.S. EPA (2000). Guidelines for Preparing Economic Analyses. Office of the
Administrator, National Center for Environmental Economics. EPA 240-R-00-003.
September. Available at:

http://yosemite.epa.gov/ee/epa/eed.nsf/webpages/Guidelines.html/$file/Guidelines.pdf

>U.S. EPA (various years). Costs and Benefits of the Clean Air Act. Available at:
http://www.epa.gov/oar/sect812/

>U.S. EPA (2004). Final Regulatory Analysis: Control of Emissions from Nonroad Diesel
Engines. EPA420-R-04-007. May. See: Chapter 9. Available at:
http://www.epa.gOv/nonroad-diesel/2004fr/420r04007a.pdf

>U.S. EPA (2006). Final Regulatory Impact Analysis: 2006 National Ambient Air Quality
Standards for Particulate Matter. Office of Air Quality Planning and Standards. See :
Chapter 5. Available at:

http://www.epa.gov/ttn/ecas/regdata/RIAs/Chapter%205~Benefits.pdf

>U.S. EPA (2008). Final Ozone NAAQS Regulatory Impact Analysis. Office of Air
Quality Planning and Standards. March. See: Chapter 6. Available at:
http://www.epa.gov/ttn/ecas/regdata/RIAs/6-ozoneriachapter6.pdf

>U.S. EPA (2008). Regulatory Impact Analysis: Control of Emissions of Air Pollution
from Locomotive Engines and Marine Compression Ignition Engines Less than 30 Liters
Per Cylinder. Office of Transportation and Air Quality. EPA420-R-08-001a. May. See:
Chapter 6. Available at: http://www.epa.gov/otaq/regs/nonroad/420r08001a.pdf

To get email updates about BenMAP:

>U.S. EPA, Office of Air Quality Planning and Standards, BenMAP website - Get Email
Updates. Available at http://www.epa.gov/air/benmap/regis.html

1.8 Frequently Asked Questions (General)

>Why do I get different results than someone else?

There are a lot of possible reasons why your results might differ from someone else's
results. One good place to start is the Audit Trail reporting option. With the Audit Trail
you can examine the assumptions that you have made to generate your results and
compare your assumptions with those made in another analysis.

>How do I know which version of BenMAP I am using? How do I know if I have the
most current version of BenMAP? How do I get the most current version?

September 2008

-------
Welcome to BenMAP

You can identify the version of BenMAP you are using by going to the Help menu and
choosing About. Here you will see the version number and contact information. To
determine whether you have the most recent version of BenMAP, you can check the
BenMAP website (http://www.epa.gov/air/benmap/), which should have the latest
version that is publicly available. Alternatively, you can use the contact information to
inquire about any upcoming versions of the model.

S About BenMAP

Al >1

Environmental Benefits Mapping and Analysis Program (BenMAP).
Version 3.0.6 (6/3/2008)

Albfl .Wn#L&V«-]nr.

For comments and questions, please contact Neal Fann at the U.S. Environmental Protection
Agency:

Address: C439-02, USEPA Mailroom, Research Triangle Park, NC 27711
Email: fann.neal@epa.gov
Telephone: (919) 541-0209
Suggested report citation for BenMAP:

Abt Associates Inc (2005). Environmental Benefits Mapping and Analysis Program (Version
3.0). Bethesda, MD. Prepared for Environmental Protection Agency, Office of Air Quality
Planning and Standards, Innovative Strategies and Economics Group. Research Triangle Park,
NC.

DBVersion

>Why are my pop-up windows too small? Why are buttons missing?

You need to check the display properties for your computer. If you right-click on the
desktop, you will get the Display Properties window.

September 2008

-------
Welcome to BenMAP

Display Properties

Themes Desktop ScreenSaver Appearance Settings

Drag the monitor icons to match the physical arrangement of your monitors.

Display:

1. (Multiple Monitors) on NVIDIA Quadro MVS 135M

Screen resolution
Less

Color quality

Highest (32 bit)

1280 by 1024 pixels

[ Use this device as the primary monitor.

E xtend my Windows desktop onto this monitor.

I dentify

T roubleshoot...

Advanced

Cancel

Apply

At the Settings tab, click on the Advanced button and try a different DPI setting. If this
does not resolve the problem, try different Screen resolution and Color quality options
back on the Settings tab.

September 2008

-------
Welcome to BenMAP

(Multiple Monitors) and NVIDIA Quadro NVS 135M Pro... [~?~|["x"l

Color Management

^ Quadro NVS 135M

General

Adapter

Monitor

T roubleshoot

Display

If your screen resolution makes screen items too small to view
comfortably, you can increase the DPI to compensate. To change
font sizes only, click Cancel and go to the Appearance tab.

DPI setting:

Normal size (96 DPI)

Normal size (96 dpi)

Compatibility

Some programs might not operate properly unless you restart the
computer after changing display settings.

After I change display settings:

O Restart the computer before applying the new display settings
® Apply the new display settings without restarting
(._) Ask me before applying the new display settings

Some games and other programs must be run in 256-color mode.
Learn more about running programs in 256-color mode.

Cancel

Apply

>What do I need to be aware of if I use BenMAP for a local scale analysis?

Perhaps the most important issue is to make sure that you have identified the resolution of
your analysis and have the appropriate grid definitions loaded into BenMAP. See here
about loading grid definitions. The next key step, which is closely connected to the grid
definitions, is to determine the data that you want to use. Data such as air quality
modeling, incidence data, and population data need to match the grid definitions that you
are using. You also need to be careful about the formatting of your data when loading it
into BenMAP. The chapter describing how to Load Data goes into this in detail.

^How can I get training for BenMAP?

Contact benmap@epa.gov for the latest information on training options.

>-ls BenMAP free? Is there a Terms of Use agreement? Are there any restrictions

September 2008

-------
Welcome to BenMAP

on using BenMAP?

BenMAP is free. There is no Terms of Use agreement and there are no restrictions on
using BenMAP. Feel free to share it with others.

>Does BenMAP estimate effects of air pollution that are not related to human health
(i.e., ecological effects)?

No. BenMAP does not currently have impact functions to estimate other than human
health effects. In principle it would be possible to estimate ecological effects, as BenMAP
is designed to combine different types of geographically variable data. To do so, you
would need to develop and load data and impact functions appropriate to estimating
ecological effects of interest.

> Why don't my files created with an older version of BenMAP work with the newer
version of BenMAP?

Files created with an older version of BenMAP will not, in some cases, work with a newer
version of BenMAP because of changes to the program. For example, later versions of
BenMAP have the capability to handle population data differentiated by ethnicity. For this
reason, after completing an analysis with BenMAP, it is always good to archive the
BenMAP installer along with the files used in your analysis, so that you will always be
able to reproduce your work in the future.

September 2008

-------
Terminology and File Types

CHAPTER 2

Terminology and

File Types

In this chapter...

>Find definitions for terms used in the model and in this
manual that may be confusing.

>Find descriptions of the various types of files used in
BenMAP.

September 2008

-------
Terminology and File Types

The first section of this chapter explains common terms used in this user's manual and in
the model. The second section describes in detail the necessary format for
externally-generated model and monitor data files that can be read into BenMAP and used
to generate the air quality grid files.

2.1 Common Terms

> Active Layer. In the GIS window, the active layer is the top-most data layer. All queries or
statistical analysis of the map will act upon this top-most layer.

> Aggregation. The summing of grid cell level results to the county, state and national levels.

> Aggregation, Pooling, and Valuation (APV) Configuration. APV Configurations store the
aggregation levels, pooling options, and valuation methods used in the analysis. In particular,
you may specify the aggregation level for incidence estimates, how to pool incidence estimates,
the valuation estimates to use, the aggregation level for the valuation estimates, and how to pool
the valuation estimates. APV Configurations are stored in files with an "apv" file extension.
The results derived from an APV Configuration have an ".apvr" file extension. APV files are
typically stored in the Configurations folder, and APV Results files are typically stored in the
Configuration Results folder.

> Air Quality Grid. An air quality grid contains air pollution data. BenMAP uses one air quality
grid for the baseline scenario and a second for the control scenario, in order to estimate the
change in the number of adverse health effects between the two scenarios. Air Quality Grids are
stored in files with an "aqg" file extension. AQG files are typically stored in the Air Quality
Grids folder.

> Air Quality Metric. One of the measures typically used for air pollution. These are daily values
calculated directly from daily observations, or through various mathematical manipulations of
hourly observations. Typical ozone metrics include the highest hourly observations during the
course of each day, such as the average of all twenty four hourly observations. These daily
values may be summarized to characterize seasonal and annual values.

> Air Quality Model. Air quality models are valuable air quality management tools. Models are
mathematical descriptions of pollution transport, dispersion, and related physical and chemical
processes in the atmosphere. Air quality models estimate the air pollutant concentration at many
locations, which are referred to as receptors, or over a spatial area that has been divided into
uniform grid squares. The number of receptors or grid-cells in a model far exceeds the number of
monitors one could typically afford to deploy in a monitoring study. Therefore, models provide a
cost-effective way to analyze pollutant impacts over a wide spatial area where factors such as
meteorology, topography, and emissions from both local and remote sources could be important.
Some examples include CMAQ and CAMx.

September 2008

-------
Terminology and File Types

>APV Configuration. See Aggregation, Pooling, and Valuation (APV) Configuration.

> Attainment. The state of meeting the National Ambient Air Quality Standard (NAAQS) for a
pollutant. A geographical area that meets the NAAQS is called an "attainment area."

> Audit Trail. This is a report that contains a record of all the choices involved in creating a
particular file. Audit trails can be created for any file that BenMAP creates.

> Background Concentration. The concentration of a pollutant irrespective of local,
anthropogenic sources. In other words, the background concentration is the level of the pollutant
due to long-range transport, natural sources, and unidentified sources.

> Background Incidence. The incidence of a given adverse effect due to all causes including air
pollution.

> Baseline Scenario. The base air quality conditions where emissions are not adjusted or
controlled. The baseline scenario is usually considered to be the reference scenario against which
to compare a potential "control scenario", in which emissions levels are changed from the
baseline levels through the application of various control measures.

>Beta. The coefficient for the health impact function. The value of beta (B) typically represents
the percent change in a given adverse health impact per unit of pollution.

> Binning. The process of summarizing data by sorting the data values from low to high, dividing
the sorted data into a pre-determined number of groups, and then choosing a representative value
for each group, by either averaging the values in each group, or picking the mid-point of each
group.

> Closest Monitor. The procedure by which data from a nearest monitor are used to represent air
pollutant levels in a population grid cell. BenMAP can also scale the data from the nearest
monitor with air pollution modeling data. BenMAP includes two types of scaling - "temporal"
and "spatial". See "Scaling" for additional explanation.

> Community Multi-scale Air Quality (CMAQ) Model. The CMAQ modeling system is the
U.S. EPA 3rd generation air quality model that has been designed to approach air quality as a
whole by including state-of-the science capabilities for modeling multiple air quality issues,
including tropospheric ozone, fine particles, toxics, acid deposition, and visibility degradation.

> Concentration-Response (C-R) Function. A C-R function estimates the relationship between
adverse health effects and ambient air pollution, and is used to derive health impact functions
(defined below). You will often see that the term C-R function and health impact function are
used interchangeably.

> Configuration. A Configuration stores the health impact functions and model options used to
estimate adverse health effects. Configurations are stored in files with a "cfg" file extension.
CFG files are typically stored in the Configurations folder. The results derived from a
Configuration have a ".cfgr" file extension. CFGR files are typically stored in the Configuration
Results folder.

September 2008

-------
Terminology and File Types

> Contingent valuation. A survey-based economic technique for the valuation of non-market
resources, such as environmental preservation or avoidance of air pollution health risk.

> Control scenario. In a modeling study, this is a sensitivity scenario in which emissions from one
or more source sectors are changed (increased or decreased), from a given "baseline scenario".
The subsequent modified emissions are then used within an air quality model to derive new
ambient concentrations. Control scenarios are used to assess the efficacy of proposed emissions
controls for sources of pollution.

>Cost of Illness (COI). The cost of illness includes the direct medical costs and lost earnings
associated with illness. These estimates generally understate the true value of reductions in risk
of a health effect, as they include just the direct expenditures related to treatment and lost
earnings, but not the value of avoided pain and suffering from the health effect.

>C-R Function. See Concentration-Response (C-R) Function.

> Currency year. The value of the dollar based on the year specified. Typically, you would
adjust the valuation estimate to a consistent currency year. For example, you might want to
report the valuation estimate in 2000 dollars to make it easier to compare with your cost analysis,
which uses that same currency year.

> Custom analysis. This type of BenMAP analysis allows you to perform each part of an analysis
separately: air quality estimation; calculation of health impact functions; aggregation, pooling,
and valuation; and reporting.

> Deltas. The difference between two pieces of data. As used in BenMAP, mapping the deltas
shows the reduction in air pollution from the baseline air quality grid to the control air quality
grid.

> Discount Rate. In this context, the discount rate is a quantitative method to account for the fact
that people generally value future benefits and costs less than current costs and benefits.
Typically, if a benefit occurs over multiple years, the benefit should be discounted.

> Endpoint. An endpoint is a subset of an endpoint group, and represents a more specific class of
adverse health effects. For example, within the endpoint group Mortality, there might be the
endpoints Mo rial i iy, Long Term, All Cause and Mortality, Long Term, Cardiopulmonary.

> Endpoint Group. A endpoint group represents a broad class of adverse health effects, such as
premature mortality, chronic bronchitis, and hospital admissions. BenMAP only allows pooling
of adverse health effects to occur within a given endpoint group, as it generally does not make
sense to sum together the number of cases of disparate health effects, such as premature
mortality and chronic bronchitis.

> Epidemiology. The study of factors affecting the health and illness of populations.
Epidemiological studies cannot prove that a specific risk factor actually causes the disease being
studied, and instead can only show that a risk factor is associated (correlated) with a higher
incidence of disease in the population exposed to that risk factor. The higher the correlation the
more certain the association, but it cannot prove the causation.

September 2008

-------
Terminology and File Types

>FIPS Code. Federal Information Processing Standard codes. Each state in the U.S. is assigned a
2-digit code. For example, "01" refers to Alaska, "37" refers to North Carolina, and "56" refers
to Wyoming.

> Fixed Effects Pooling. Fixed effects pooling is used to combine two or more distributions
(represented by Latin Hypercube points) into a single new distribution. Fixed effects pooling
assumes that there is a single true underlying relationship between these component
distributions, and that differences among estimated parameters are the result of sampling error.
Weights for the pooling are generated via inverse variance weighting, thus giving more weight to
the input distributions with lower variance and less weight to the input distributions with higher
variance.

>Fixed Radius. An air quality monitor interpolation option that uses all monitors within a fixed
radius (or distance) of a given point of interest. All monitors are used and weighted by their
relative distance.

>GIS. Geographic Information System. A GIS is a system of hardware and software used for
storage, retrieval, mapping, and analysis of geographic data.

> Grid Cell. One of the many geographic, or spatial components within Grid Definition.

r- Grid Definition. A BenMAP Grid Definition provides a method of breaking a geographic
region into areas of interest (Grid Cells) in conducting an analysis. This can be done in two
ways - by loading a Shapefile (a particular type of Geographic Information System file) or by
specifying a regularly shaped grid pattern. These are referred to as Shapefile Grid Definitions
and Regular Grid Definitions, respectively. Typically a Shapefile Grid Definition is used when
the areas of interest are political boundaries with irregularly shaped borders, while a Regular
Grid Definition is used when the areas of interest are uniformly shaped rectangles.

> Growth Income Adjustment. See Income Growth Adjustment.

> Health Impact Function. A health impact function calculates the change in adverse health
effects associated with a change in exposure to air pollution. Based on a C-R function, a typical
health impact function has inputs specifying the air quality metric and pollutant, the age, race
and ethnicity of the population affected, and the incidence rate of the adverse health effect.

> Incidence. The total number of adverse health effects in a geographic region in a given unit
time. In BenMAP, this is the total number of adverse health effects avoided due to a change in
air pollution.

> Incidence Rate. The background rate of a health effect per person in a given geographic region
(a.k.a. background incidence rate). The average number of adverse health effects per person per
unit of time, typically a day or a year.

> Income Growth Adjustment. Adjusting certain valuation functions to reflect increases in real
income. Generally, an increase in real income implies an increase in the willingness to pay
(WTP).

> Interpolation. The process of estimating the air quality level in an unmonitored area by using

September 2008

-------
Terminology and File Types

one or more nearby air quality monitors. BenMAP uses two types of interpolation procedures:
one is to simply choose the closest monitor, the other is to use a technique called Voronoi
Neighbor Averaging. These interpolation methods are discussed in more detail in Appendix C.

>Lat/Long. Latitude and longitude information to specify the geographic coordinates of a spatial
location. The CMAQ model data are usually provided for each grid cell identified by the latitude
and longitude of the grid cell's center point. Latitude identifies the north-to-south location of a
point on the Earth. Latitude can be defined precisely as the angle between the vertical at a
location, and the equatorial plane of the Earth. Longitude identifies the east to west location of a
point on the Earth, by measuring the angular distance from the Greenwich meridian (or Prime
meridian, where longitude is 0), along the equator.

> Latin Hypercube. A series of points generated by using specified percentiles in a given
distribution, such as that of a health impact coefficient. It is a short-cut method designed to
represent a distribution, while at the same time saving on computation time. For example, when
using 20 Latin Hypercube points, BenMAP would use the 2.5th, 7.5th, 12.5th, ..., and 97.5th points
from the distribution. The Latin Hypercube points are used when combining the results of
different health impact functions (discussed in Chapter 7), and in presenting confidence intervals
for the incidence estimates (discussed in Chapter 8).

> Layer. In geographic information systems (GIS), a layer represents a logical separation of
mapped data usually representing a theme, such as political boundaries, roads, ozone data,
number of mortalities avoided, etc. In BenMAP, "active layer" refers to the top-most layer with
which the user is currently working.

> Layer Statistics. The summary statistics that correspond to the active layer in BenMAP. For
example, "mean", "standard deviation" or "max" of PM2.5 air quality grid.

> Micrograms per Cubic Meter (ug/m3). The unit of measure for particulate matter in the
NAAQS. This unit represents the mass of PM found in a cubic meter of air.

> Model Data. Pollutant concentration data that are generated by running an air quality model
such as CMAQ. This is different from "monitor data," which are based upon measurements.

>Model Delta. Differences in pollutant concentrations from applying an air quality model (e.g.
CMAQ) for two different emissions scenarios - a baseline scenario and an emissions control
scenario.

> Modeling. Estimating air pollution levels using air quality models. The EPA website discusses
a wide range of air quality models: http://www.epa.gov/ebtpages/airairquaairqualitymodels.html
and at http://www.epa.gov/ttn/scram/.

>Monetize. In the context of human health benefits assessment, this is the practice of expressing
society's preferences for avoiding certain health effects as a dollar value. In BenMAP we
estimate monetized benefits by using either Willingness-to-Pay or Cost-of-Illness valuation
functions (see above and below).

> Monitor Data. Pollutant concentration data that are based upon measurements. "Raw" monitor
data usually refers to data that are taken directly from measurement networks, with no additional

September 2008

-------
Terminology and File Types

processing of the data. Monitor data are different from "model data," which are based upon
numerical predictions.

> Monitoring. Actual measurements of air pollution levels. The U.S. Environmental Protection
Agency has monitoring data, as well as other information related to monitoring, available
through its Air Quality System (AQS): http://www.epa.gov/air/data/aqsdb.html.

> Morbidity. A measure of being diseased or afflicted by an illness (non-fatal).

> Mortality. A measure of the number of deaths in a given population.

>National Ambient Air Quality Standards (NAAQS). The U.S. EPA establishes levels for
pollutants that are considered harmful to public health and the environment. The Clean Air Act
established two types of national air quality standards. Primary standards set limits to protect
public health, including the health of "sensitive" populations such as asthmatics, children, and
the elderly. Secondary standards set limits to protect public welfare, including protection against
decreased visibility and against damage to animals, crops, vegetation, and buildings. The U.S.
EPA has set NAAQS for six principal pollutants, which are called "criteria" pollutants. (Carbon
Monoxide, Lead, Nitrogen Dioxide, Ozone, Particulate Matter (PM2.5, PM10), and Sulfur
Dioxide).

> One-Step Analysis. An easier and shorter approach to performing analyses with BenMAP,
targeted for beginning users. This combines all three major steps of BenMAP, i.e., a) calculate
changes in pollutant concentrations, b) use concentration-response (C-R) function to estimate
health incidences, and c) calculate economic benefits due to changes in morbidity and mortality
into a single process. However, the choices of analyses are limited. For example, the health
incidences and valuation functions are predefined, and the user cannot change them.

> Ordinality. In relation to AQ monitors, ordinality refers to the number of monitor values in the
season that can exceed your standard. For example, if we had set the ordinality to 4, then a
monitor can have as many as three daily averages (assuming that we are using the daily average
metric to define our standard) greater than your standard without violating the standard. In terms
of rollback, if it has more than 3 daily averages in exceedance of the standard, then the rollback
technique will be applied to that monitor.

> Ozone (03). A gas composed of three oxygen atoms. It is not usually emitted directly into the
air, but at ground-level is created by a chemical reaction between oxides of nitrogen (NOx) and
volatile organic compounds (VOC) in the presence of sunlight. Ozone has the same chemical
structure whether it occurs miles above the earth (in what is sometimes referred to as "the ozone
layer") or at ground-level, and can be "good" or "bad," depending on its location in the
atmosphere. For the purposes of BenMAP, the focus is on ground-level ozone, which can be
breathed in. Breathing ozone can trigger a variety of health problems, including chest pain,
coughing, throat irritation, and congestion. It can worsen bronchitis, emphysema, and asthma.
Ground-level ozone also can reduce lung function and inflame the linings of the lungs. Repeated
exposure may permanently scar lung tissue.

> Particulate Matter. "Particulate matter," also known as particle pollution or PM, is a complex
mixture of extremely small particles and liquid droplets. Particle pollution is made up of a
number of components, including acids (such as nitrates and sulfates), organic chemicals, metals,

September 2008

-------
Terminology and File Types

and soil or dust particles. Once inhaled, these particles can affect the heart and lungs and cause
serious health effects. Includes PM25 (particles less than 2.5 microns in aerodynamic diameter),
PM10 (particles less than 10 microns in aerodynamic diameter), and PMC (particles between 2.5
and 10 microns in aerodynamic diameter).

> Parts per Million (ppm). This unit represents the concentration of the pollutant in a million
parts of air. Carbon monoxide is often measured in units of ppm.

> Parts per Billion (ppb). This unit represents the concentration of the pollutant in a billion parts
of air. Ozone concentrations in BenMAP are reported in units of ppb.

> Pooling. The combining of different sets of data. BenMAP has several pooling methods,
including fixed effects, fixed/random effects, and subjective weighting. Appendix I discusses
the pooling approaches available in BenMAP.

> Point Mode. When defining the configuration, you may choose to either estimate adverse
health effects in point mode or using a Latin Hypercube. The point mode simply means that
BenMAP will use the mean value of the coefficient in the health impact function.

> Population Exposure versus Personal Exposure. Population (or ambient) exposure refers to
the average air pollution level measured in a grid cell. In contrast, personal exposure keeps track
over the course of a day the exposure individuals encounter in different micro-environments,
such as the freeway, outdoors and indoors. BenMAP only keeps track of population exposure.

> Post-processing. Additional data modification steps that are undertaken to prepare either raw
monitor data or raw model data for subsequent import into BenMAP. Postprocessing may be as
simple as reformatting the data, or it may involve more complex operations such as performing
additional numerical calculations to aggregate or disaggregate data temporally or spatially.

> Prevalence Rate. The percentage of individuals in a given population who already have a given
adverse health condition. Used to calculate changes in health conditions among those who
already have a health condition, such as asthmatics.

> Quality Adjusted Life Years (QALYs). The number of "healthy" life years gained or lost due
to changes in air quality. The QALY weights each life year gained according to the severity of
the health effect.

> Random Effects Pooling. Random effects pooling is an alternative to the fixed effects pooling
(see Fixed Effects Pooling, above), and allows the possibility that the estimated parameter from
different studies may in fact be estimates of different parameters, rather than just different
estimates of a single underlying parameter.

> Regulatory Impact Analysis (RIA). A policy tool used to assess the likely effects of a proposed
regulation or regulatory change. It usually involves detailed analyses to quantify the costs and
benefits of the regulation.

> Relative Risk. Relative risk typically is used as a measure of the change in risk of an adverse
health effect associated with an increase in air pollution levels. More specifically, it is the ratio

September 2008

-------
Terminology and File Types

of the risk of illness with higher pollution to the risk of illness with a lower pollution level,
where the "risk" is defined as the probability that an individual will become ill.

>Rollback. The process by which monitor data are reduced to a different level. BenMAP
facilitates three different types of monitor data rollback. Percentage rollback reduces all monitor
observations by the same percentage. Incremental rollback reduces all observations by the same
increment. Rollback to a standard reduces monitor observations so that they just meet a specified
standard.

> Scaling. The process in which monitor data are used in conjunction with model data to improve
interpolation and/or forecast future air quality trends. BenMAP has 3 types of scaling - temporal
only, spatial only, temporal and spatial scaling.

> Setup. A BenMAP Setup encapsulates all of the data needed to run analyses for a particular
geographic area - a city, an entire country, etc. These data consist of grid definitions, pollutants,
monitor data, incidence and prevalence rates, population data, health impact functions, variables,
inflation rates, and valuation functions.

> Shapefile. A shapefile is a particular type of Geographic Information System (GIS) file, and has
a .shp extension. These files are accompanied by companion files with .shx and .dbf extensions,
and can be used to create Shapefile Grid Definitions. See
http://www.esri.com/library/whitepapers/pdfs/shapefile.pdf for more information.

> Subjective Weighting Pooling. Subjective weights let you specify the weights that you want to
use when combining two or more distributions of results. The weights should sum to one. If
not, BenMAP normalizes the weights so that they do.

>Sum Dependent Pooling. Summing two or more incidence or valuation results, assuming the
underlying functions are correlated.

> Sum Independent Pooling. Summing two or more incidence or valuation results, assuming the
underlying functions are independent (uncorrelated).

> Threshold. An air quality level below which benefits are not calculated. For example, if the
threshold is 10 (ig/m3, then only areas with PM2 5 concentrations equal to or greater than 10 (ig/m
3 will be included in estimating health incidence results.

>Unit Value. A unit value is the estimated mean value of avoiding a single case of a particular
health effect.

> Valuation Function. Valuation functions are used by BenMAP to estimate the economic values
of changes in the incidence of health effects. These are selected within an Aggregation, Pooling,
and Valuation Configuration (APV Configuration).

> VNA (Voronoi Neighbor Averaging). An algorithm used by BenMAP to interpolate air quality
monitoring data to an unmonitored location. BenMAP first identifies the set of monitors that
best "surround" the center of the population grid cell, and then takes an inverse-distance
weighted average of the monitoring values. This is discussed in detail in Appendix C.

September 2008

-------
Terminology and File Types

> WTP (Willingness to Pay). The willingness of individuals to pay for a good, such as a
reduction in the risk of illness. In general, economists tend to view an individual's WTP for a
improvement in environmental quality as the appropriate measure of the value of a risk
reduction. An individual's willingness-to-accept (WTA) compensation for not receiving an
improvement is also a valid measure. However, WTP is generally considered to be a more
readily available and conservative measure of benefits.

2.2 File Types

BenMAP has a number of file types that you can use to store the settings used in a
BenMAP analysis, the results of an analysis, as well as maps and reports. Table 2-1
presents the names of the different file types, their functions, and their default folder
locations.

Table 2-1. File Types Generated by BenMAP

File Extension

Description

Default Folder Location

*.aqg

Air quality grid.

Air Quality Grids

*.cfg

Configuration specifying the health impact functions and
other options used to generate incidence estimates.

Configurations

*.cfgr

Configuration results, containing incidence results at the grid
cell level.

Configuration Results

*.apv

Aggregation, Pooling, and Valuation configuration specifying
the aggregation levels, pooling options, and valuation
methods used to generate aggregated incidence estimates,
pooled incidence estimates, valuation estimates, aggregated
valuation estimates, and pooled valuation estimates.

Configurations

*.apvr

Aggregation, Pooling, and Valuation configuration results,
containing incidence results at the grid cell level, aggregated
incidence results, valuation results, aggregated valuation
results, and pooled valuation results.

Configuration Results

*.shp

Shape files generated by BenMAP's mapping capabilities.
These files can be viewed within BenMAP or within shape
file viewers, such as ArcView.

Maps

*.csv

Reports are exported as *.csv files, which may be viewed in a
text editor, or in programs such as Excel.

Reports

September 2008

-------
Overview of BenMAP Components

CHAPTER 3

Overview of
BenMAP
Components

In this chapter...

>Learn about the One-Step Analysis.

>Get an overview of the functions available with the Custom
Analysis.

>Learn about the Tools and Help menu options.

September 2008

-------
Overview of BenMAP Components

Upon starting BenMAP for the first time, you will see the following main window.

1 Tools Help

Two Wavs to Use BenMAP: Which Analysis Meets vour Needs?

One-Step Analysis

Custom Analysis

After you import the air quality data
for your area, use this tool to apply
default settings and create a report.

Step 1 — Import air quality data

Air Quality Grid creation

|- f"?\

Air Quality Grid Creation

^ |

Step 2 — Set custom parameters

Incidence Estimation

Preloaded
EPA param Biers

¦^>1 k

Step 3 — Use results from Step 2
to set custom parameters

'..111

Pooling. Aggregation
and Valuation

Step 4 — Run report

,,11

Active Setup:

The opening screen is broken into two sections. On the left hand slide is the One-Step
Analysis button. This feature enables you to perform a quick benefit analysis while
selecting a minimum of analysis options. On the right hand side is the Custom Analysis
section, which contains four large buttons. This feature allows you to perform a highly
customized benefit analysis. The Tools menu at the top of the screen is for less frequently
used functions, such as importing and exporting data and mapping.

The first section in this Chapter describes the One Step Analysis. The second section
describes the Custom Analysis. And the final section describes the functions found in the
Tools and Help menu options. All of these topics are covered in greater detail in

September 2008

-------
Overview of BenMAP Components

subsequent chapters of this manual.

3.1 One-Step Analysis

The One-Step Analysis allows you to run a BenMAP analysis quickly. This option can be
useful when you are time constrained, when you don't have a strong preference regarding
the benefit analysis options, or if you are new to the program. This feature uses an EPA-
developed benefit analysis configuration to estimate the human health benefits of the air
quality scenario that you specify in the model.

The One-Step Analysis requires a minimal number of inputs. At minimum, you should:

>Give your run a name. Very descriptive titles can help you remember more about your
run when you want to review your results in the future.

>Choose an output directory. The output directory is the location on your computer in
which you will save your results. (The default is to put the output in the Configuration
Results folder.)

> Select air quality grids. BenMAP estimates the benefits of some change in air quality for
a particular area of the country. Here, you're telling BenMAP which air quality files to
use to calculate benefits. You can either import an air quality file that has already been
created, or you can generate your own. If you want to create your own air quality file,
you will want hit the "create" button and follow the instructions under the custom
analysis page below. For additional details on air quality grid creation, see the chapter
on Air Quality Grid Creation.

>Choose your aggregation level. Select the geographic level at which you want to
summarize your results.

September 2008

-------
Overview of BenMAP Components

v One Step Analysis

Run Name' |PM2.5 2000 County ClosestMonitor 10PctRollback

Output Director^': |C:\Progrann Files\BenMAP 3.0\Configuration Results

Select

1. Air Quality Grids
Baseline File:

|MAP 3.0V\ir Quality Grids\Baseline PM25 2000 ClosestMonitor County.aqg

Open

Create

Control File:

Map Deltas j

uality Grids\Control PM25 2000 ClosestMonitor County 10pctRollback.aqg

Open

Create

r2. Incidence

At what level would you like to aggregate your incidence results?

Nation

3. Valuation

At what level would you like to aggregate your valuation results?

Valuation Aggregation must be at a level the same as or higher than the Incidence aggregation.

Nation

Open Results

Cancel

After you have selected these options, just hit the "Go!" button, and BenMAP will run the
analysis in a single step. The One-Step run time generally depends on the complexity of
the air quality data. More complex air quality data requires more time to run.

3.1.1 One-Step Reports

When BenMAP completes its run a new window will appear that gives you an array of
graphical and tabular reporting options. The top part of this window provides incidence
report options, the middle part of the window contains options for creating valuation
reports and in the bottom part of the window is the audit trail report button. (For more
details on One-Step reports, see the chapter on Reports.)

September 2008

-------
Overview of BenMAP Components

One Step Results

Incidence Results-

Table of Mortality
Incidence

Table of Morbidity
Incidence

Box Plot of
Mortality Incidence

["Valuation Results-

Table of Mortality
Valuation

Table of Morbidity
Valuation

Box Plot of
Mortality Valuation

Cumulative
Distribution
Functions

Bar Chart of
Valuation

Audit T rail-

Audit Trail Report

Results File:

uration Results\Test One Step.apv

BenMAP can generate tables of incidence and valuation results for each health impact
function selected. This example shows valuation estimates for cases of premature
mortality avoided. These estimates are categorized by region of the country. You can save
these tables as .jpeg files and then import them into reports, or print them directly from
this screen.

September 2008

-------
Overview of BenMAP Components

Estimated Reduction in PrematureMortality
dOthPercentileConfidencelntervalsProvidedinParentheses

Western U.S.

Eastern U.S. ExcludingCA California NationalTotal

Mortality Impact Functions Derived from Epidemiology Literature

Harvard Six-City
Study

1210,000
($59,0004410,000)

17,400
($2,000-$14,000)

$30,000
($8,400-$58,000)

$250,000
($70,000-$480,000)

ACS Study

195,000
($22,000- $190,000)

13,200
($770-516,600)

$13,000
($3,200-$27,000)

$110,000
($26,000-$230,000)

Woodruff et al.,
1997 (infant
rnortalitvl

1200
(154-1400)

$17
($4-$33)

$30
($8-$59)

$250
($66-$490)

Mortality Impact Functions Derived from Expert Elicitation

Expert A

$250,000
($36,000-$560,000)

$22,000
($3,100-$49,000)

$36,000
($5,200-$81,000)

$310,000
($45,000-$690,000)

Expert B

$210,000
($22,000-$530,000)

$17,000
($1,400-$46,000)

$30,000
($3,300-$76,000)

$250,000
($27,000-$650,000)

Expert C

$210,000
($27,000-$490,000)

$18,000
($2,400-$43,000)

$30,000
($3,900-$71,000)

$260,000
($33,000-$610,000)

Expert D

$140,000
($22,000-$280,000)

$12,000
($1,900-$25,000)

$20,000
($3,200-$41,000)

$170,000
($27,000-$350,000)

Expert E

$320,000
($85,000-$620,000)

$28,000
($7,400-$54,000)

$46,000
($12,000-$89,000)

$390,000
($100,000-$760,000)

You can also generate a variety of graphics as well. In this example, different estimates of
monetized benefits are plotted as bar charts and matched to the area of the country
receiving the benefit.

September 2008

-------
Overview of BenMAP Components

Total Monetized Benefits
(Millions of 1999$)

| 3210,000 |

I 195,000 I | $92,000 | I 393,000 |

[Bp "owf "\mi/ "[131] ¦ M M
& & 0

n (3) CA
¦ (2) East
~ (1) West

I $13,000 I | | $13,000 | ^ | $13,000 | ^ | $11000.'» 0Q ^9£Qo1

(3) CA

(2) East

(1) West

Laden Pope Expert Expert Expert

et al et al L K J

Expert

BenMAP will report an "audit trail" that reports all of your data inputs and analysis
options. This report can be very useful when someone wants to replicate your analysis, or
when you would simply like to remember all of the data and options you used in your
analysis. The audit trail is discussed in further under "generate reports" in the Custom
Analysis section below.

September 2008

-------
Overview of BenMAP Components

13 Configuration Results: CAProgram Files\BenMAP 3.0\Configuration Results\PM2.5 2000 County ClosestMonitor 10PctRollback.ofgr
Latin Hypercube Points: 10

Population Dataset: United States Census - CMAQ 3Gkm
Year: 2020
Threshold: 0
- Grid Definition

Name: County
ID: 0

Columns: 5E
Rows: 840
Grid Type: Shapefile
Shapefile Name: County
+ Selected Studies

+ Baseline Air Quality Grid: CAProgram Files^BenMAP 3.0\Air Quality Grids\Baseline PM25 2000 ClosestMonitor County.aqg
+ Control Air Quality Grid: C:\Program Files\BenMAP 3.0VAir Quality GridsVControl PM25 2000 ClosestMonitor County 10pctRollback.aq

>How do I know what currency year is used in the One-Step Analysis?

You can find the answer in the Audit trail for the APVR file that you generated.
Alternatively, go to the Tools menu and choose One-Step Setup. Then choose a
Pollutant (e.g., PM2.5) to determine the currency year.

+ Advanced

+ Incidence Pooling Windows

El Valuation Pooling Windows

El QALY Pooling Windows

Export | OK

3.1.2 Questions Regarding the One-Step Analysis

September 2008

-------
Overview of BenMAP Components

S BenMAP 3.0.6

Help

Air Quality Grid Aggregation
Model File Concatenator
Database Export
Database Import
Export Air Quality Grid
GIS/Mapping
Modify 5etup
Neighbor File Creator

One-Step Setup

/s to Use BenMAP: Which Analysis Meets your Needs?

SIS

air quality data
s tool to apply
create a report.

Air Quality Grid Crealion

J3L

Preloaded
EPA parameters

Report

.11

Custom Analysis

Step 1 — Import ait quality data

Air Quality Grid Crealion

Step 2 - Set custom parameters

Incidence Estimation

Step 3 — Use results from Step 2
to set custom parameters

Pooling. Aggregation
and Valuation

Step 4 - Run report

Report

S BenMA... / B ^ ® _J

September 2008

-------
Overview of BenMAP Components

v One Step Setup Parameters

For each pollutant specify the standard cfg and apv files you wish used for the One-Step Analyses.

Pollutant:

CFG File Name: |C:\Progrann Files\BenMAP 3.0\Configurations\PM2.5 Wizard CFG.cfg Browse |
APv' File Name' jCAProgram Files\BenMAP 3.0\Configurations\PM 2.5 Wizard APV.apv Browse

Currency Year:

2000

Cancel

>How do I know what population year is used in the One-Step Analysis?

You can find this answer in the Audit Trail.

September 2008

-------
Overview of BenMAP Components

v Audit Trail Report

E CAProgram Files\BenMAP 3.0\Configuration Results\PM2.5 2000 County ClosestMonitor 1OPotRollback.apvr

&• Configuration Results: CAProgram Files\BenMAP 3.0\Configuration Results\PM2.5 2000 County ClosestMonitor lOPctRollbaok.cfgr
Latin Hyperoube Points: 10

Population Dataset: United States Census - CMAQ 36km

Year: 2020

Threshold: 0
+ Grid Definition
+ Selected Studies

+ Baseline Air Quality Grid: CAProgram Files^BenMAP 3.0\Air Quality Grids\Baseline PM25 2000 ClosestMonitor County.aqg
+ Control Air Quality Grid: CAProgram Files\BenMAP 3.0\Air Quality Grids\Control PM25 2000 ClosestMonitor County lOpctRollback.aq
+ Advanced

+ Incidence Pooling Windows
+ Valuation Pooling Windows
+ QALY Pooling Windows

Export

>Why are the reporting options limited to just certain types of PM2.5 reports in the
One-Step Analysis?

The reports are custom-made and time-consuming to develop. Initially, we focused on
PM2.5, since it is most closely linked to adverse health impacts. Over time, we will
develop additional reporting options. Note that the One-Step option will allow you to
generate .apvr files for ozone and other pollutants, you just do not have reports
specifically designed for them. If you estimate effects for other than PM2.5, then you
should use the Report options available under the Custom Analysis.

>How do I change the .cfg or .apv files used in the One-Step Analysis? Why do my
tables and graphs look strange for analysis completed in the Custom Analysis?

Go to the Tools menu and choose One-Step Setup. At the One Step Setup Parameters
window, you can browse for alternative .cfg and .apv files. If you choose to use .cfg or .
apv files for a PM2.5 analysis that differ from the defaults installed with BenMAP, then
the reports may well look strange, because the reports are specifically designed for the
default .cfg and .apv files. In this case, you should use the Report options available
under the Custom Analysis.

September 2008

-------
Overview of BenMAP Components

>Can I use the pre-made tables and graphs in the One-Step Analysis if I completed
the analysis using the Custom Analysis? Why do my tables and graphs look
strange for an analysis completed in the Custom Analysis?

No, generally you should not use the pre-made tables and graphs in the One-Step with
results generated from the Custom Analysis. If you generate an .apvr file using the
Custom Analysis, in most cases the tables and graphs will look strange because the
reports are specifically designed for the default .cfg and .apv files that come with the
BenMAP installation.

>How do I find out which functions are used in the One-Step Analysis?

You can identify the functions used in the One-Step Analysis at least two ways. First,
you can use the Audit Trail option to identify the functions used in an existing set of
One-Step results. Alternatively, if you want to look at the functions prior to the
analysis, then you might want to look at the .cfg and .apv files using the Audit Trail
available with the Report options under the Custom Analysis.

>How do I create the Air Quality Grids needed for the One-Step Analysis?

There are several ways to create Air Quality Grids needed for a One-Step Analysis,
using monitor data, model data, or both monitor and model data. These options are
briefly discussed in the Create Air Quality Grids section of this Overview chapter and in
detail in the Air Quality Grid Creation chapter.

>Why are the aggregation levels limited in the One-Step Analysis?

The aggregation levels are limited in the One-Step Analysis in order to simplify the
analysis and to make the results compatible with the pre-made tables and graphs.

>Can I calculate QALYs with a One-Step Analysis?

Yes. However, the reporting possibilities for the One-Step Analysis are geared toward
incidence and valuation. As a result, you will need to use the reporting options in the
Custom Analysis to view and export your results.

3.2 Custom Analysis

Returning to the main BenMAP window, you can choose the Setup you want to use for an
analysis by selecting it from the Active Setup drop-down list. You are then ready to begin
using the functions available through the four main buttons.

September 2008

-------
Overview of BenMAP Components

These four buttons take you through the steps of an analysis. The first button allows you
to create air quality grids which contain estimates of population-level exposure to air
pollution. The second button lets you choose the air quality grids for a particular analysis,
and then to choose the health impact functions to estimate the incidence of adverse health
effects. The third button gives you different options for combining the health effects
estimates and choosing economic valuation functions. Using the fourth button, you can
generate several different kinds of reports.

3.2.1 Create Air Quality Grids

BenMAP is not an air quality model, nor can it generate air quality data independently.
Instead it relies on the air quality inputs given to it. To estimate population exposure to air
pollution, BenMAP combines population data with an Air Quality Grid, which it generates
using some combination of air quality modeling and/or monitoring data. The Create Air
Quality Grids button allows you to put your air quality data into the format that is used by
BenMAP, the first step in conducting an analysis within BenMAP.

Grid Definitions

Air quality grids contain air pollution exposure estimates for a particular Grid Definition,
as defined in the Setup Manager window. Grid Definitions are typically comprised of
either regularly shaped rectangles covering the region of analysis, or irregularly shaped
polygons corresponding to political boundaries.

Modeling and Monitoring Data

To generate air quality grids, you can use air quality modeling data and air quality
monitoring data in three different ways, as discussed below. However, once generated, all
air quality grids have the same structure, and have the same .aqg extension that BenMAP
uses to designate these file types.

> Model Direct. Model Direct grid creation simply takes raw model data and converts it
into a file that BenMAP recognizes as an air quality grid. This type of grid definition
allows you to directly specify the air pollution values for each grid cell in a Grid
Definition.

September 2008

-------
Overview of BenMAP Components

« Air Quality Grid Creation ... [V][~][><]

Choose Grid Creation Method
(I- iModei Direcij

r Monitor Direct

C Monitor and Model Relative

C Monitor Rollback

Cancel Go!

> Monitor Direct. Monitor Direct grid creation uses air pollution monitoring data to estimate air
pollution levels for each grid cell in the selected Grid Definition. This may be done using one of
two interpolation procedures - Closest Monitor, or Voronoi Neighbor Averaging (VNA). With
closest monitor, BenMAP simply uses the data of the monitor closest to each grid cell's centroid.
With VNA, BenMAP first identifies the set of monitors that most closely "surround" each grid
cell, and then calculates an inverse-distance weighted average of the data from these neighboring
monitors.

> Monitor and Model Relative. Monitor and Model Relative grid creation allows you to combine
information from both monitor and modeling data files. BenMAP uses the modeling data to
scale the monitoring data, to compensate for incomplete monitoring data coverage, and the
inability of monitor data to predict future conditions. Like the Monitor Direct approach, you
choose an interpolation method (Closest Monitor or VNA), but a scaling approach is also
chosen, either temporal, spatial, or both. This approach is described in more detail in the Air
Quality Grid Creation chapter, as well as in the appendix on Air Pollution Exposure Estimation
Algorithms.

> Monitor Rollback. Monitor Rollback grid creation allows you to reduce, or roll back, monitor
data using three methods: percentage rollback, incremental rollback, and rollback to a standard.
Percentage rollback reduces all monitor observations by the same percentage. Incremental
rollback reduces all observations by the same increment. Rollback to a standard reduces monitor
observations so that they just meet a specified standard. After the monitor data is rolled back, it
may be directly interpolated (as in Monitor Direct grid creation) or combined with modeling data
(as in Monitor and Model Relative grid creation). This approach is described in more detail in
the Air Quality Grid Creation chapter, as well as in the appendix on Monitor Rollback
Algorithms.

3.2.2 Create and Run Configuration

The Create and Run Configuration button allows you to calculate the change in the
incidence of adverse health effects associated with changes in air quality. There are
several steps in the process.

>Step 1. Specify the baseline and control air quality grids that you created using the Create Air

September 2008

-------
Overview of BenMAP Components

Quality Grids button.

» Configuration Settings

~ HI®

Select Air Quality Grids
Baseline File:

|C:\Program Files\BenMAP 3.0\Air Quality Grids\Baseline PM25 2000 ClosestMonitor County.aqg
Control File:

|C:\Program Files\BenMAP 3.O^-Air Quality Grids\Control F'M25 2000 ClosestMonitor County 10pctRollback.aqg

~pen

Create

~pen

Create

Map Grids

Settings—

Latin Hypercube Points:

|10

Population DataSet:

IUnited States Census
Run In Point Mode:

Threshold:

Population Year:

Cancel

> Step 2. Specify whether BenMAP should make a "point estimate", or a set of "Latin Hypercube
" points. The point estimate is the change in incidence, generated using the mean value of the
coefficient in the health impact function. The Latin Hypercube points are a series of points
generated by using evenly spaced percentiles in the distribution of the health impact coefficient -
these points represent the distribution of incidence values. (The Latin Hypercube points can
later be used when combining the results of different health impact functions, and in presenting
confidence intervals for the incidence estimates.)

> Step 3. Specify the threshold, or a lowest value for air quality data. Any observations which
fall below this threshold will be replaced with the threshold value in all calculations.

September 2008

-------
Overview of BenMAP Components

v Configuration Settings

Available CR Functions:

~111®

1T ree

DataSet

Endpoint Group

Endpoint

Year

Location

Other Pollutants

Qualifier

BenMAP Health

Mortality

Asthma Exacerbatio

+ ~

Hospital Admissions.

—

Lower Respiratory S

Upper Respiratory S

Upper Respiratory S>

1991

Utah Valley

Adjusted CoelH

1991

Utah Valley

•

Acute Myocardial In

J '

Selected CR Functions:

Function Identification

Function Parameters

DataSet

Endpoint

Race

Ethnicity

Gender

Start Age

End Age

Incidence DataSet

Prevalence Data...

Variable DataSet

BenMAP He

Mortality

BenMAP He

Mortality

BenMAP He

Hospital /=

2000 Incidence anc

BenMAP He

Upper Re

Step 4. Choose the health impact functions that will be used in the estimation, and hit the Go!
button to start estimating the change in incidence.

BenMAP can store configuration choices in a user-named file with a .cfg extension, and
can store incidence change estimates in a user-named file with a .cfgr extension.

3.2.3 Aggregation, Pooling, and Valuation

The Aggregation, Pooling, and Valuation button allows you to specify an aggregation
level for previously calculated incidence estimates, pool these aggregated incidence
estimates, place an economic value on these pooled and aggregated incidence estimates,
aggregate these economic values, and finally pool these aggregated economic values.

There are several steps in this process.

>Step 1. Choose whether to create a new .cipv configuration or use an existing .cipv configuration.
(For this description we will choose to create a new .cipv configuration.)

September 2008

-------
Overview of BenMAP Components

w APV Configuration Creation Method

EMI®

(* jCreate New Configuration for Aggregation, Pooling, and Valuation.:

C Open Existing Configuration file for Aggregation, Pooling, arid Valuation P.apv file).

Cancel

Go!

Step 2. Choose a set of incidence estimates to use. These will be loaded from a configuration
results file made with the Create and Run Configuration button, which is stored by BenMAP
with a .cfgr extension.

Open ?

My Recent
Documents

Desktop

My Documents

My Computer

Look in: Configuration Results

S cj !!t

i^jBond Rule
ljjj|EJ NCEE
tjJEJ Spring 2008

['5+03 2000 County ClosestMonitor lOPctRollback.cfgr
¦0-RM2.5 2000 County ClosestMonitor lOPctRollback.cfgr

My Network File name:
Places

Files of type:

| Configuration Results (K. cfgr I

"3
"3

Open
Cancel

r Step 3. Choose the desired pooling and aggregation options for the incidence results.

September 2008

-------
Overview of BenMAP Components

v Incidence Pooling and Aggregation | - ||n |[X |

Available Incidence Results —

Select Pooling Methods

H- PM2.5

Pooling Window Name: Basic Functions

A
V

Endpoint Group

Start Age Author Endpoint

Pooling Method

Mortality

None

Laden et al.

Mortality, All CaL

Pope et al.

Mortality, All Cat

Pooling Window Name: |AMI 7% Calcs

Endpoint Group

Endpoint Author Qualifier

Pooling Method

Acute Myocardial In

Acute Myocardial Inl

Peters et al.

Adjusted Coeffic

None

Window to Delete --

Delete

Add

Target Grid Type:

Configuration Results File Narne(s): |CABenMAP_Developrnent\StandardWizard\PM_Wizard_20080111 \County_ t

Browse

Advanced

Cancel

| Next I

r Step 4. Choose the economic valuation functions to apply to the pooled and aggregated
incidence results.

48 September 2008

-------
Overview of BenMAP Components

v Select Valuation Methods, Pooling, and Aggregation

i®

Valuation Methods

Variable DataSet:

0 EPA Standard Valuation F

EPA Standard Variables

Pooling Window Name: Basic Functions

.13.1
m

•v-|

Endpoint Group

Start Age Author

Valuation M ethod Pooling M ethod

Mortality

None

t
-

VSL, based on rang

Nnnp

Pooling Window Name: AMI 1% Calcs

Endpoint Group

Endpoint Author

Valuation Method Pooling Method

Acute Myocardial In

Acute Myocardial Inl

Peters et al.

Sum (Dependent)

•

Subjective Weights

COI: 5yrs med, 5yr:

COI: 5yrs rned, 5yrs

Subjective Weights

rni- R nrfl- morl R nr<

"" I'l >|

Advanced p SkipQ,

^LY Valuation

Cancel Previous Next

r Step 5. Choose the desired pooling and aggregation options for the economic valuations, and hit
the Next button. If the Skip QALY Valuation box is checked, then BenMAP is ready to save
your APV Configuration and generate results.

r Step 6. If desired, you can uncheck the Skip QALY Valuation box and proceed to choosing
your desired QALY functions. Why have Skip QALY Valuation box? It takes a bit of time for
the QALY functions to load, and since QALY functions are not always used, this box is intended
to speed up the time it takes to develop the APV Configuration.

BenMAP can store APV Configuration choices in a user-named file with an "apv" extension, and
can store APV Configuration results in a user-named file with an ".apvf' extension. As needed,
you can access both files for later use.

3.2.4 Generate Reports

The Generate Reports button allows you to generate several types of reports.

September 2008

-------
Overview of BenMAP Components

w Select Report Type

~II®

C Incidence and Valuation Results: Raw, Aggregated, and Pooled. (Created from ".apvr files)
C Raw Incidence Results. (Created from "cfgr files)

C Audit Trail Reports (Created from K.aqg files, K.cfg files, "cfgr files, K.apv files, or K.apvr files)

Cancel OK

>Incidence and Valuation Results reports use an Aggregation, Pooling, and Valuation
Results file (with the ".apvr" extension) to create reports for incidence, aggregated
incidence, pooled incidence, valuation, aggregated valuation, or pooled valuation results.
These reports are comma separated values (CSV) files (*.csv) which can be read using a
text editor, or by various spreadsheet and database programs, such as Microsoft Excel.

v Choose a Result Type

Result Type
C Incidence Results
C Aggregated Incidence Results
C Pooled Incidence Results
C Valuation Results
C Aggregated Valuation Results
r Pooled Valuation Results
C QALY Valuation Results
C Aggregated QALY Valuation Results
C Pooled QALY Valuation Results

Cancel OK

rRaw Incidence Results use a Configuration Results file (with the ".cfgr" extension) to
create reports for incidence results. These reports are CSV files.

September 2008

-------
Overview of BenMAP Components

w Configuration Results Report

File

Column Selection

Grid Fields:

3-R Function Fields:

Result Fields:

B Column

~ DataSet

Reference

~ P2Beta

ly Point Estimate

0 Row

~ Endpoint Group

Race

~ A

y Population

Endpoint

Ethnicity

~ NameA

0 Delta

Pollutant

Gender

~ B

S Mean

Metric

Start Age

NameB

B Baseline

~ Seasonal Metric

End Age

Is? Percent of Baseline

~ Metric Statistic

Function

~ NameC

0 Standard Deviation

Author

Incidence DataSet

0 Variance

Year

C Prevalence DataSet

S Latin Hypercube Points

Location

Beta

Other Pollutants

~ DistBeta

~ Qualifier

~ PIBeta

Grouping Options

f* Group by Gridcell, then by C-R function.
Group by C-F! function, then by Gridcell.

Display Options
Digits After Decimal Point: 4

Advanced Options

Population Weighted Deltas: f

Elements in Preview: 25

Aggregation Level: |

Preview

Column

Row

Point Estimate

Population

Delta

Mean

Baseline

25.5059

76,404.3125

1.7215

25.4841

1,011.0748

Done

r Audi I Trail Reports provide a summaries of the assumptions underlying each of five types of
files generated by BenMAP: Air Quality Grid (".aqg"). Configuration (".cfg"). Configuration
Results ( "cfgr ), Aggregation, Pooling, and Valuation ("apv'), and Aggregation, Pooling,
and Valuation Results ("".apvr"). These reports can be viewed within BenMAP in an
expandable tree structure, or can be exported to text files.

September 2008

-------
Overview of BenMAP Components

v Audit Trail Report

C:\Program Files\BenMAP 3.0\Configuration Results\PM2.5 2000 County ClosestMonitor "lOPotRollback.apvr

~ Configuration Results: CAProgram Files\BenMAP 3.0\Configuration Results\PM2.5 2000 County ClosestMonitor 10PctRollback.ofgr
Latin Hypercube Points: 10

Population Dataset: United States Census - CMAQ 3Gkm
Year: 2020
Threshold: 0
- Grid Definition

Name: County
ID: 0

Columns: 5E
Rows: 840
Grid Type: Shapefile
Shapefile Name: County
+ Selected Studies

+ Baseline Air Quality Grid: CAProgram Files^BenMAP 3.0\Air Quality Grids\Baseline PM25 2000 ClosestMonitor County, aqg
+ Control Air Quality Grid: C:\Program Files\BenMAP 3.0\Air Quality GridsVControl PM25 2000 ClosestMonitor County lOpctRollback.aq
+ Advanced

+ Incidence Pooling Windows
El Valuation Pooling Windows
El QALY Pooling Windows

Export

3.3 Menus

There are two menu options found at the top of the main window: Tools and Help. The
Tools menu provides access to data import and export functionality and mapping. The
Help menu provides access to information about BenMAP - the version, contact
information, and a suggested citation.

3.3.1 Tools

The Tools menu has several options: Database Export, Database Import, GIS Mapping
,and Modify Setup.

September 2008

-------
Overview of BenMAP Components

3 BenMAP 3.0.6 [7][n][x]

9 Help

Air Quality Grid Aggregation

/s to Use BenMAP: Which Analvsis Meets vour Needs?

Database Export
Database Import
Export Air Quality Grid
GIS/Mapping
Modify Setup
Neighbor File Creator
One-Step Setup

sis

air quality data
s tool to apply
sreate a report.

Custom Analysis

Step 1 — Import air quality data

Air Quality Grid Creation

Air Quality Grid Credion

Step 2 — Set custom parameters

5rX

Incidence Estimation

Preloaded
r EPA parameters

s>\ k

-a

Step 3 — Use results from Step 2
to set custom parameters

Report

'¦Ill

\!T|SrJ=^ Pooling. Aggregation
and Valuation

!¦¦¦¦¦

Step 4 — Run report
Report I

Lull

Active Setup:

3Mz> Quality> Grid Aggregation. Create a new air quality for a new grid definition (e.g., State)
from an existing air quality grid created with a different grid definition (e.g., County).

> Model File Concatenator. Combine model files for use in the creation of air quality grids.

>•Database Export. Export all or part of BenMAP's internal database to a file which can later be
used (on another computer, for example) with the Database Import tool. Manually loading data
into BenMAP can be time and labor intensive, so this tool can be quite useful in sharing data
with other users or computers.

>•Database Import. Import data exported using the Database Import tool.

y Export Air Quality Grid. Generate a text file (.csv) with all of the data in air quality grid,
including summary statistics such as mean, median, minimum, and maximum.

> GJS ¦ Mapping. Generate a wide variety of maps, including maps of monitor data, air quality

September 2008

-------
Overview of BenMAP Components

grids, population data, incidence rates, and both incidence and valuation results. In addition, you
can export the maps you have generated and view them in a shapefile viewer, such as ArcView.
Note that BenMAP also has context specific mapping capabilities in various parts of the
application. See the GIS/Mapping chapter for details on BenMAP's mapping capabilities.

yModify Setup. View, edit, add, and delete data from BenMAP's internal database. For details see
the Load Data chapter.

>Neighbor File Creator. Create a text file (.1x1) identifying "neighbor" monitors and associated
interpolation weights for each grid cell in an air quality grid.

> One-Step Setup. Specify the CFG and APV Configuration files and currency year for the
One-Step Analysis.

September 2008

-------
Loading Data

CHAPTER 4
Loading Data

In this chapter...

>Create new data setups to load and store your input data.
>Find details on the file structures for data inputs.

>Learn how to export and import setups.

September 2008

-------
Loading Data

BenMAP encapsulates in a single dataset all of the data needed to run analyses for a
particular geographic area, such as a city, region, or nation. This dataset is called a "Setup
" and consists of grid definitions, pollutants, monitor data, incidence and prevalence rates,
population data, health impact functions, variables, inflation rates, valuation functions,
income growth data, and QALY data. Grouping the data in this way has a number of
advantages. It makes it easy to organize and view the data, export the data (either a whole
setup or a portion of a setup), and import setups generated by others.

In this Chapter we discuss how to add, modify, and delete setups, as well as how to export
and import existing setups. There are a number of steps involved in loading the data,
particularly in the formatting of the data, so it is important to carefully review the steps in
this Chapter. Also, it is important to keep in mind that if you delete one part of a setup,
you may be affecting other parts of the setup. We discuss this further below.

Finally, it's worth noting that many users will never need to modify the setup. If you are
performing an analysis with the U.S. setup, you may find that BenMAP contains all of the
data you need to perform your analysis, and no additional modifications are necessary.

4.1 Add, Modify, and Delete Setups

To add a new setup, modify an existing setup, or delete a setup, go to the Tools drop-down
menu, and choose the Modify Setup option. This will bring up the Manage Setups window.
The United States setup, which comes preinstalled with BenMAP comes with a variety of
datasets and might look something like this:

September 2008

-------
Loading Data

$ Manage Setup

Available Setups:

United States

Delete

Add

Grid Definitions

CMAQ 12km Nation
CMAQ 12km Nation - Clipped
CMAQ 36km

CMAQ 36km Nation Overlap
County

NaHnn

Edit

Incidence/Prevalence DataS...

2000 Incidence and Prevalence
2005 Mortality Incidence
2010 Mortality Incidence
2015 Mortality Incidence
2020 Mortality Incidence
?0?Fi Mortality Incidence

Edit

Variable DataSets

Education

EPA Standard Variables
Poverty

Edit

Income Growth Adjustments
Income Elasticity (3/21 /07)

Pollutants

Ozone
PM10
PM2.5
Temp

Monitor DataS ets

EPA Standard Monitors

Edit

Population DataS ets

United States Census - County
Woods And Poole
United States Census ¦ CMAQ 36krr
CMAQ 12km Nation - Clipped

C-R Function DataS ets
BenMAP Health Impact Functions

Edit

Inflation DataS ets

EPA Standard Inflators

Valuation DataS ets

EPA Standard Valuation Functions

Edit

QALY Distribution DataSets

AMI DF! 0%

AMI DF! 2%

AMI DR 1%

AM HE DR 0%

AMILE DR 3%

AMII F DR IX

Edit

Cancel

Add a Setup. To add a setup, click the Add button. The New Setup window will appear,
where you can type the name of the setup.

September 2008

-------
$ New Setup

3@[x

Loading Data

After naming a new setup, you can then define anew the elements that comprise a setup:
Grid Definitions, Pollutants, Monitor DataSets, Incidence and Prevalence DataSets,
Population DataSets, C-R Function DataSets, Variable DataSets, Inflation DataSets,
Valuation DataSets, Income Growth Adjustments, and QALY Distribution DataSets.

Some of the elements of a setup are fundamental and should be entered before the others,
namely Grid Definitions and Pollutants. The Incidence and Prevalence, Population,
and Variable DataSets depend on the Grid Definitions, and the Monitor and C-R
Function DataSets depend on the Pollutants that you have defined. As a result, it is best
to start by defining your Grid Definitions and Pollutants, and then defining the other
elements of the setup.

Modify a Setup. To modify a setup, start by choosing the setup for modification from the
drop-down menu in Available Setups. Click on the Edit button under one of the eleven
components comprising a setup. The sections below describe each of these components.

Delete a Setup. To delete a setup, choose the setup for deletion from the drop-down menu
in Available Setups. Click the Delete button. You will then be asked to confirm your
decision.

4.1.1 Grid Definitions

A BenMAP Grid Definition provides a method of breaking a geographic region into areas
of interest (Grid Cells) in conducting an analysis. This can be done in two ways: by
loading a Shapefile (a particular type of Geographic Information System file) or by
specifying a regularly shaped grid pattern. These are referred to as Shapefile Grid
Definitions and Regular Grid Definitions, respectively. Typically a Shapefile Grid

58 September 2008

-------
Loading Data

Definition is used when the areas of interest are political boundaries with irregularly
shaped borders, while a Regular Grid Definition is used when the areas of interest are
uniformly shaped rectangles.

Ideally, one Shapefile Grid Definition will be created using a shapefile containing an
outline of the area of interest (a city boundary, for example). Additional grid definitions
will then be created for each subdivision of that area for which (a) data is available (see
the Air Monitoring, Population, Incidence and Prevalence, and Variables sections below),
or (b) reports or maps are desired.

For example, to conduct analyses for the United States, one might use the following grid
definitions:

>Nation - this Shapefile Grid Definition contains an outline of the United States (just the
lower forty-eight states), defining an overall area of interest.

>County - this Shapefile Grid Definition contains county borders, for use with
county-based population and incidence rate data.

>CMAQ 12km Nation - this Shapefile Grid Definition contains grid cells that are roughly
12 kilometers on each side, for use with air quality modeling data.

> State - this Shapefile Grid Definition contains state borders, for use in generating reports
and maps with results aggregated to the state level.

To start adding or modifying grid definitions, click on the Edit button below the Grid
Definitions box. The Manage Grid Definitions window will appear.

Manage Grid Definitions

Available Grid Definitions

Delete

Add

Grid Type:

Edit

Cancel

Click on the Add button below the Available Grid Definitions box,. The Grid
Definition window will appear, with two tabs showing the two main types of grid
definitions that you may use in BenMAP: Shapefile Grid or Regular Grid.

September 2008

-------
Loading Data

« Grid Definition

~in®

Grid ID:

GridDefinition 0

Shapefile Grid Regular Grid
Load Shapefile:

Preview

Cancel

4.1.1.1 Regular Grid

Regular Grid Definitions are defined by a lower left corner (specified as decimal degree
latitude and longitude, with West and South having negative values and East and North
having positive values), a total number of columns and rows, a number of columns per
degree longitude, and a number of rows per degree latitude. Individual cells within the
resultant grid are numbered in sequential order (columns from left to right, rows from
bottom to top) starting at (1, 1). These field values will be used to link the Regular Grid
Definition with other sources of data, as discussed in more detail below.

To define a Regular Grid, start by typing the name of the grid definition in the Grid ID
box, and then defining the number of Columns and Rows in the grid. To locate this grid
geographically, provide the decimal degree coordinates for the lower left-hand corner of
the grid in the Minimum Longitude and Minimum Latitude boxes.

To give the overall geographic size of the grid, provide the number of Columns Per
Longitude and Rows Per Latitude. For example, if there are 16 columns and 2 columns
per degree longitude, then the grid will span 8 degrees longitude. And if there are 25 rows
and 4 rows per degree latitude, then the grid will span 6.25 degrees of latitude.

Combining the numbers in this example, if the minimum longitude and latitude are -81 and
38 and the grid spans 8 degrees longitude and 6 degrees latitude, then the grid will run
between -81 and -73 degrees longitude and between 38 and 44.25 degrees latitude.

After defining the grid, click the Preview button to see what the grid looks like. You may
change the parameters and click the Preview again to see how the grid changes. When

September 2008

-------
Loading Data

you are satisfied with the grid definition, click the OK button.

« Grid Definition

~in®

I Grid ID:

d |
J

I |Rectangular Grid)

Columns:

Rows:

l1G 2d

I25 2d

Minimum Longitude:

I -81

Minimum Latitude:

Columns Per Longitude:

|2 2il

Rows Per Latitude:

l< 2&

I Preview |

Cancel

The name of your newly defined grid will then appear in the Manage Grid Definitions
window. You may click Edit to change the grid definition, Delete to permanently remove
the grid that you just defined, Add to define a new grid definition, or OK to get back to
the Manage Setups window.

Manage Grid Definitions

Available Grid Definitions

Rectangular Grid

Delete

Add

Grid Type:

Regular

Edit

Cancel

4.1.1.2 Shapefile Grid

Shapefiles used to create Shapefile Grid Definitions should be of the ESRI® Shapefile

September 2008

-------
Loading Data

format. Details on this format can be found at

http://downloads.esri.com/support/whitepapers/other_/shapefile.pdf. Shapefiles used
must be unprojected - that is, they should store their geographic information as decimal
degree latitude and longitude values. If your shapefile is already projected, you will want
to use a program like ArcGIS® to reproject the shapefile using the NAD 1983 datum.
Finally, any shapefiles used must contain integer fields named Column (or Col) and Row,
and each shape within the shapefile must contain a unique combination of values for these
two fields. These column and row values are used, just as the Column and Row field
values in Regular Grid Definitions, to link the Shapefile Grid Definition with other
sources of data, as discussed in more detail below.

To add a Shapefile Grid, click on the Add button in the Manage Grid Definitions

window, choose the Shapefile Grid tab, name the grid in the Grid ID, and browse for the
correct shapefile by clicking on the small open-file icon just to the right of the Load
Shapefile box.

After locating the file in the Browse for a Shapefile window, click Open. This will
choose the file, and bring you back to the Grid Definition window. To view the shapefile,
choose Preview.

« Grid Definition

0H)B

Grid ID:

Metropolitan Area

Shapefile Grid | Regular Grid | Regular Grid Wizard j

Load Shapefile:

\Shapefiles\Metropolitan Area CityOne.shp" i^|

^ \,

Preview

Cancel OK

When you are satisfied that the shapefile looks correct, click OK. This will bring you
back to Manage Grid Definitions window. Note that the Grid Type box displays the
type of grid for each of your grid definitions.

Click OK when you are finished loading grid definitions. The Manage Setups screen will
now list the Grid Definitions that you have just created. At any time, you may click the

September 2008

-------
Loading Data

Edit button to add, modify, or delete grid definitions.

Note that if you delete a Grid Definition you will delete any data that is dependent on it,
such as any Incidence, Prevalence, Population, and Variable DataSets that use this
particular Grid Definition. As we discuss below each of these other elements of a setup,
we will describe how this might happen.

4.1.2 Pollutants

The Pollutants section of a setup specifies the pollutants that BenMAP will analyze and
defines the air quality metrics to be used by BenMAP. You are not importing air pollution
data, but rather naming your pollutants and defining the measures or metrics BenMAP will
use when performing a benefit analysis for each pollutant. You may include any pollutant,
though typically air pollutants such as particulate matter, ozone, sulfur dioxide, and carbon
monoxide are used in a BenMAP analysis.

A key concept for pollutants is the Metric. Air quality metrics are daily values calculated
directly from daily observations, or through various mathematical manipulations of hourly
observations. Typical ozone metrics include the highest hourly observations during the
course of each day, the average of all twenty four hourly observations, etc. In a benefit
analysis, the air quality change must be expressed in a metric that matches the metric used
by the health impact function; this concept is discussed further below.

In general, air pollution data in BenMAP is hierarchical - a pollutant can have multiple
Metrics, each of which has multiple Statistics (these are automatically calculated by
BenMAP) and which can have multiple Seasonal Metrics. Similarly, Seasonal Metrics
have multiple Statistics. Furthermore, air pollution data can be provided to BenMAP at
any of these levels, in addition to the daily/hourly observation level, as described in more
detail in Section 4.3.

To start adding pollutant definitions to BenMAP, click on the Edit button below the
Pollutants box of the Manage Setups window. The Manage Pollutants window will
appear. Here you may click Add to add a new Pollutant, Delete to remove a previously
defined pollutant, or Edit to modify an existing pollutant.

September 2008

-------
Loading Data

$ Manage Pollutants f-l|P |[x

Available Pollutants

Pollutant Metrics

Delete Add

Edit

Cancel 1 OK

Add a Pollutant

Air pollution data in BenMAP is of two types - point source monitoring data, and Grid
Definition based modeling data. For both types, the data must be associated with a
particular pollutant. Table 4-1 describes these variables used to define a pollutant in
BenMAP.

Table 4-1. BenMAP Pollutant Definitions

Pollutant Field .. ,

t»t Notes

Name

Pollutant ID

Unique name for the pollutant which will be referenced in health impact functions,
associated with monitoring and modeling data, etc.

Observation Type

Pollutants may have hourly observations or daily observations. In the United States,
Ozone has hourly observations, while PM10 and PM2.5 have daily observations.

Metrics

Daily values calculated directly from daily observations, or through various
mathematical manipulations of hourly observations. Typical ozone metrics include
the highest hourly observations during the course of each day, the mean of all twenty

four hourly observations, etc.

Seasonal Metrics

Seasonal values calculated from metric values. In the United States, for example,
quarterly means are calculated forPM2.5 from daily means.

To start defining a pollutant, click the Add button and the Pollutant Definition window
will appear. In the Pollutant ID box, you give a unique name for the pollutant (e.g.,

September 2008

-------
Loading Data

PM2.5), and then define the characteristics of this pollutant - the Observation Type and
Metrics.

Pollutant Definition

Pollutant ID:
|~PM25

Observation Type:

| Daily

Metrics

D24HourMea

Delete

Add

Advanced Pollutant Uptions

Hourly Metric Generation:

Fixed Window I Moving Window | Custom

Start Hour:
End Hour:

[23~

Statistic: |Mean

Seasonal Metrics

~3
"3

Edit

Uancel

The Observation Type identifies whether a pollutant is measured hourly or daily. In the
United States, ozone, sulfur dioxide, carbon monoxide, and others have hourly
observations, while particulate matter has daily observations. Use the Observation Type
drop-down menu to choose either the Hourly ox Daily observation type.

Next you need to define a pollutant's Metrics. A pollutant has to have one or more
metrics, which are daily values calculated directly from daily observations, or through
various mathematical manipulations of hourly observations.

To add a Metric, click on the Add button below the Metrics box. A default name Metric
0 will appear in the box. Since the default name is not very descriptive of a metric, it is
best to change the name. Typical names used for metrics given in Table 4-2. These are
provided just as an example, and you may use any names that you like. However, keep in
mind that the names that you use for your metrics need to be consistent with the metric
names that you include in your air pollution monitoring and modeling data, as well as your
health impact functions. (We will discuss this further below.) Additionally, metric names
are used to display pollutant concentrations in BenMAP's mapping window. As such,
they must be consistent with GIS naming conventions, meaning they must begin with a
letter, and may only contain letters, numbers, and underscores.

September 2008

-------
Loading Data

Table 4-2. Examples of Metric Names
Name Description

DIHourMax Highest hourly value from 12:00 A.M. through

11:59 P.M.

D24HourMean Average of hours from 12:00 A.M. through 11:59

P.M.

D8HourMax Highest eight-hour average calculated between

12:00 A.M. and 11:59 P.M.

4.1.2.1 Hourly Metrics

Pollutants that are measured hourly (Observation Type = Hourly), such as ozone, sulfur
dioxide, carbon monoxide, and others, must be characterized by a daily metric, which
mathematically summarizes the hourly observations.

Table 4-2 lists some of the ways that metrics can be generated from hourly values. Note
that these metrics are not arbitrarily chosen, and instead match the metrics used in
epidemiological studies.

The Hourly Metric Generation portion of the Pollutant Definition window lets you
define the metrics that you want to use. There are three tabs that you may choose: Fixed
Window, Moving Window, and Custom.

The Fixed Window tab lets you define simple metrics which are calculated as statistics
over a fixed window of hours (Start Hour and End Hour) within each day. The Start
Hour should be less than or equal to the End Hour, and both can range from 0 to 23,
where 0 stands for the period 12:00 am to 12:59 am, and 23 stands for 11:00 pm to 11:59
pm. The Statistic includes the Mean, Median, Max, Min, and Sum. Some examples
follow:

D24HourMean: The mean of the observations from 12:00 am through 11:59
pm. Start Hour = 0. End Hour = 23. Statistic = Mean.

DIHourMax: The highest hourly value of the observations from 12:00 am
through 11:59 pm. Start Hour = 0. End Hour = 23. Statistic = Max.

D12HourMean: The mean of the daylight observations, defined as the period
from 8:00 am through 7:59 pm. Start Hour = 8. End Hour = 19. Statistic =
Mean.

September 2008

-------
Loading Data

Pollutant Definition

Pollutant ID:

~ zone

Observation Type:
[Hourly

Metrics

D24HourMean

D1 Hour Max
DGHourMax

Delete

Add

Advanced Pollutant Options

Hourly Metric Generation:

Fixed Window I Moving Window I Custom

Start Hour: |0
End Hour: 123

Statistic: |

Seasonal Metrics

IBEEBI t

Mean

Median

Max

Min

Sunn

Edit

Cancel

The Moving Window tab lets you consider metrics that are not based on the same set of
hours each day. The Window Size defines the number of hours that will be considered
together. The Window Statistic defines how the hours in the Window Size will be
characterized. And the Daily Statistic defines how BenMAP will use the statistics
generated for each window.

For example, consider the highest eight-hour mean (D8HourMax) over the course of a day.
You would have the following settings: Window Size = 8. Window Statistic = Mean.
Daily Statistic = Max. BenMAP would calculate every possible eight-hour mean, starting
with the eight-hour mean from 12:00 am through 7:59 am, and ending with the eight-hour
mean from 4:00 pm through 11:59 pm. This would generate 17 possible eight-hour means.
BenMAP would then choose the eight-hour mean that has the highest value.

September 2008

-------
Loading Data

Pollutant Definition

Pollutant ID:

~ zone

Observation Type:
[Hourly

Metrics

D24HourMean
D1 Hour Max

DSHourMax

Delete

Add

Advanced Pollutant Options

Hourly Metric Generation:

Fixed Window Moving Window | Custom

Window Size: [fT

Window Statistic: (Mean
Daily Statistic:

Seasonal Metrics

IBEB1

Mean

Median

Max

Min

Sunn

Edit

Cancel

The Custom tab lets you define Metrics using customized functions. These functions can
include measures such as the sum of the number of hours of ozone exposure about 60 parts
per billion. The possibilities are quite diverse, as evidenced by the range of functions and
variables available for use in Table 4-3. However, the syntax for using these functions is
somewhat involved, so we have reserved discussion of this for the Function Editor
appendix.

September 2008

-------
Loading Data

Pollutant Definition

Pollutant ID:

~ zone

Observation Type:
[Hourly

Metrics

D24HourMean
D1 HourMax
DGHourMax

Delete

Add

Advanced Pollutant Options

Hourly Metric Generation:

Fixed Window | Moving Window Custom |

Function:

Seasonal Metrics

Edit

Cancel OK

Table 4-3. Available Functions and Variables for Custom Metrics

Available
Functions

Description

Available Variables

Description

ABS(x)

Returns the absolute value

of X.

Observations^]

All hourly observations for the
year (index begins at zero,
typically ranging to 8,760).

EXP(x)

Returns e the power x,
where e is the base of the
natural logarithm.

DailyObservations[i]

All hourly observations for the
day (indexed zero to
twenty-three).

IPower(x,y)

Returns x to the power y (y
an integer value).

SortedObservations [i]

All hourly observations for the
day, sorted from low to high
(indexed zero to twenty-three).

LN(x)

Returns the natural
logarithm of x.

Day

Index of the day whose metric
value is being generated (index
begins at zero).

POWER(x.y)

Returns x to the power y (y
a floating point value).

Mean

Mean of the daily observations.

September 2008

-------
Loading Data

SQR(x) Returns the square of x.

Median

Median of the daily
observations.

SQRT(x) Returns the positive square
root of x.

Min

Minimum of the daily
observations.

Max

Maximum of the daily
observations.

Sum

Sum of the daily observations.

NoObservation Flag value indicating a missing

observation (-345)

4.1.2.2 Seasonal Metrics

Seasonal Metrics allow users to aggregate daily Metric values. This has a number of
uses. For example, if pollutant values vary greatly by season, you can calculate separate
pollutant measures for each season of interest. You might be interested in dry season
versus wet season, or differences between Winter, Spring, Summer, and Fall.

To add seasonal metrics, click on the Edit button below the Seasonal Metrics box. The
Manage Seasonal Metrics window will appear.

To add a Seasonal Metric, click on the Add button below the Seasonal Metrics box. A
default name Seasonal Metric will appear in the box. Since the default name is not very
descriptive, it is best to change the name to something more informative such as the
QuarterlyMean. As with the Metric names, keep in mind that the Seasonal Metric names
that you use need to be consistent with the metric names that you include in your air
pollution monitoring and modeling data, as well as your health impact functions.

The next step is to define the seasons that you want associated with your Seasonal Metric
name. For example, in the case of a Quarterly Mean, you would want to define four
seasons. To start this process, click on the Add button below the Seasons box. Then, in
the far right side of the window under Selected Season Details, give the Start Date and
End Date for each season.

September 2008

-------
Loading Data

Manage Seasonal Metrics

Metric ID:
|D24HourMeari

Seasonal Metrics

Uuarterly Mean

Delete

Add

Seasonal Metric Seasons:

Seasons

Season 1
Season 2
Season 3

Season 4

Delete

Add

Selected Season Details:
Start Date:

October 01
End Date: | December 31

Statistic | Custom I

Statistic:

M ean

"3]

Cancel

Next, you need to choose the Statistic tab or the Custom tab to determine how the daily
Metrics will be combined in each season. For example, you might choose the Mean from
the drop-down list on the Statistics tab. This would calculate the mean of the daily
metrics in each season. The Custom tab allows seasonal metric values to be calculated
using customized functions, similar to those used to calculate daily metric values from
hourly observations. See the Function Editor appendix for more detail on this topic.

Once you have finished defining the Seasonal Metrics, click OK to return to the
Pollutant Definition window.

Click OK after defining each Pollutant - this will return you to the Manage Pollutants
window.

September 2008

-------
Loading Data

$ Manage Pollutants f-1|n ||x

PM2.5

0 zone
PM10

Delete

Add

Pollutant Metrics

D24HourMean
Quarterly Mean

Edit

Cancel

4.1.2.3 Advanced Pollutant Options

The Advanced Pollutant Options button allows you to associate Seasons with a
Pollutant. These seasons differ somewhat from the Seasonal Metrics discussed above.
They are used to define:

The portion of the year for which benefits are calculated for a Pollutant. For
example, in the United States ozone benefits are only calculated for the ozone
season, which is assumed to be May 1 through September 30.

The portion(s) of the year for which missing pollutant concentrations are filled
in by BenMAP. That is, in order to calculate benefits, BenMAP in certain
cases needs to generate complete sets of metric values by estimating
concentrations for those days which have missing observations. This can be
important if certain seasons tend to have more missing values than others.

The portion(s) of the year for which monitor data is scaled by modeling data
when creating Monitor Model Relative air quality grids (discussed in the Air
Quality Grid Creation chapter).

To define Seasons for a Pollutant, click the Advanced Pollutant Options button. This
will bring up the Define Seasons window. For each season desired, click the Add button,
select the appropriate Start Date and End Date, which define the days included in the
season; the appropriate Start Hour and End Hour, which define the hours included in
Monitor Model Relative scaling; and the appropriate Number of Bins, which define the
number of bins used in Monitor Model Relative scaling. The advanced options for PM2.5
look like the following:

September 2008

-------
Loading Data

$ Define Seasons

Season 1
S eason 2
Season 3

Season 4

Delete

Add

Selected Season Details:

Start Date:

[Octobefl]^

-
-

End Date:

December 31

Start Hour:

End Hour:

Number of Bins:

Cancel

Once you have finished defining the Seasons, click OK to return to the Pollutant
Definition window.

If you later wish to View or Edit a particular Pollutant definition, simply select the
appropriate Pollutant within the Available Pollutants box and click the Edit button.
When you are done, click OK to return to the Manage Setups window.

After defining all of the pollutants that you want to consider, click OK. This will return
you to the Manage Setups window.

4.1.3 Monitor Data

The Monitor DataSets section of the Manage Setups window allows to you to add air
pollution monitoring data to your setup. Air pollution monitoring data may be used alone,
or in conjunction with air pollution modeling data, to estimate ambient pollution levels in
each grid cell defined by a Grid Definition. BenMAP uses a variety of procedures (such
as Voronoi Neighbor Averaging, discussed later) to interpolate the monitor data points
across the area of interest.

NOTE: Air pollution data in BenMAP is of two types - point source monitoring data, and
Grid Definition based modeling data. Both types of data must be associated with a
particular pollutant that you have defined. Only the point source monitoring data is stored
in the setup database. The modeling data are loaded into BenMAP as you need them for a
particular analysis.

To start, click on the Edit button below the Monitor DataSets box. The Manage
Monitor DataSets window will appear. From this window you may Add monitoring
data, view and Edit existing datasets, as well as Delete them. The section on the left
under Available DataSets lists the monitor datasets that are currently in the setup. The
section on the right under the DataSet Contents identifies the number of monitors in each
dataset by pollutant and by year.

September 2008

-------
Loading Data

$ Manage Monitor DataSets

Available DataSets

Delete

Add

DataSet Contents (Number of Monitors by Pollutant by Year):

Edit

Cancel

To start adding data, click the Add button. This will bring up the Monitor DataSet
Definition window. Give the dataset a name in the DataSet Name box, choose the
appropriate pollutant from the Pollutant drop-down menu, and then type the year of the
data in the Year box.

NOTE: The DataSet that you define can have one or more pollutants and multiple years of
data. If your data come from a single source, for simplicity it probably makes sense to
combine these data into a single DataSet.

$ Monitor DataSet Definition [_J[nJ[xJ

DataSet Name:

|CityOne Monitors

DataSet Contents (Number of Monitors by Pollutant by Year):

Pollutant

MJft

Pollutant:

Year:

"31

Database, Columns | Database, Rows | Text File
Monitor Data File:

Browse

Monitor Definition File:

Browse

Load Monitor Data

Cancel

Now choose one of the three tabs (Database, Columns; Database, Rows; or Text File)
that matches the format of your data. Monitor data may be formatted in three different
ways: (1) in two database files/tables, with monitor definition information in rows in one
file/table and monitor values in a single column in the other file/table (Database,

September 2008

-------
Loading Data

Columns format), (2) in a database file, with monitor definition information and monitor
values in a single line (Database, Rows format), and (3) in a text file, with monitor
definition information and monitor values in a single line (Text File format). There is no
preferred approach, and instead use the approach with which you are most comfortable.
(We define these formats in detail below.)

Click on the Browse button next to the Monitor Data File box. This will bring up a
window from which you can browse the BenMAP Data directory to find the desired data
file. The Files of type drop-down menu allows you to choose from a variety of file types.

Click Open. This will choose the file, and bring you back to the Monitor DataSet
Definition window. If you are using the Database, Columns format, you will need to
repeat this procedure with the Monitor Definition File. To add monitors from the
selected file(s), click Load Monitor Data.

* Monitor DataSet Definition EJJ=L)I ®

DataSet Name:

|CityOne Monitors

DataSet Contents (Number of Monitors by Pollutant by Year):

Pollutant

Year

Count

Ozone

1994

Pollutant: Year:

|Ozone T] 11994

Database, Columns | Database, Rows | Tent File
Monitor Data File:

|onitor Data 03 Hourly 1994 CityClne Data.xls Browse
Monitor Definition File:

|or Data 03 Hourly 1994 CityOne Definition.xls Browse

I Load Monitor Data i

Cancel | OK

Follow this same procedure to load all of your monitoring data. Select the pollutant from
the Pollutant drop-down menu, edit the Year box, choose the tab that matches the
database format that you are using, and select the appropriate file(s) by clicking the
Browse button(s). When you have clicked Load Monitor Data for your final dataset,
click OK. This will bring you back to the Manage Monitor DataSets window.

September 2008

-------
Loading Data

$ Manage Monitor DataSets [T][n][>r

Available DataSets
CityOne Monitors

Delete

Add

DataSet Contents (Number of Monitors by Pollutant by Year):

Pollutant

Year

Count

~ zone

1994

~ zone

2000

~ zone

2003

Edit

Cancel OK

To see the years of data and the number of monitors each year, use the scrollbars on the
bottom and on the right of the DataSet Contents box.

If desired you can add additional monitor datasets (click the Add button), delete existing
datasets (select the dataset in the Available DataSets list and click the Delete button), or
edit existing datasets (select the dataset in the Available DataSets list and click the Edit
button).

When you have finished loading your monitor data, click OK in the Manage Monitor
DataSets window. This will take you back to the Manage Setups window, which will
show the name of the DataSet(s) that you just entered.

4.1.3.1 Format Monitor Data

Monitor data may be formatted in three different ways. There is no preferred option.
Simply use whichever method seems most appropriate.

Database, Columns format. Two database files/tables, with monitor definition information in
rows in one file/table and monitor values in a single column in the other file/table.

Database, Rows format. One database file, with monitor definition information and monitor
values in a single line.

Text File format. One text file, with monitor definition information and monitor values in a single
line.

Tables 4-4a through 4-4d present the variables in each of the datasets when using the
columns format and a sample of what the files might look like. Tables 4-5a and 4-5b
present the variables in the dataset when using the rows format and a sample of what a
data file might look like.

NOTE: The monitor data files do not specify the pollutant with which the data is

76 September 2008

-------
Loading Data

associated - this is specified by the user when loading the monitor data into BenMAP.

Table 4-4a. Column Format, Air Monitor Definition Variables

Variable

Type Required

Notes

Monitor
Name

text

yes

Unique name for the monitor.

Description text

Description of the Monitor.

Longitude numeric
(double)

yes Values should be in decimal degree format. Values in the eastern
hemisphere are positive, and those in the western hemisphere are

negative.

Latitude numeric
(double)

yes Values should be in decimal degree format. Values in the

northern hemisphere are positive, and those in the southern
hemisphere are negative.

Table 4-4b. Column Format, Sample Air Monitor Definition File

Monitor Name

Description

Latitude

Longitude

M o n it o r1

City Park

40.107222

-74.882222

M o n it o r2

Industrial Site

39.835556

-75.3725

M o n it o r3

Suburan

40.112222

-75.309167

M o n it o r4

Rural

39.94

-75.6

Table 4-4c. Column Format, Air Monitoring Data Set Variables

Variable Type Required Notes

Metric

text

This variable is either blank (signifying that the Values are
Observations, rather than Metric values), or must reference an
already defined Metric for the appropriate Pollutant.

Seasonal
Metric

text

This variable is either blank (signifying that the Values are not
Seasonal Metric values) or must reference an already defined
Seasonal Metric for the Metric.

Annual
Metric

text

This variable is either blank (signifying that the values are not
annual statistics) or must be one of: None, Mean, Median,
Max, Min, Sum.

Monitor
Name(s)

numeric
(double)

yes

Each of these field names should match up with a value in the
Monitor Name field in the Definitions file. The values will be

September 2008

-------
Loading Data

observations, metric, seasonal metric, or annual metric values,
as determined by the values of the other fields.

Table 4-4d. Column Format, Sample Air Monitoring Data File

Metric

Seasonal Metric

Annual Metric

M o n it o r1

Munitor2

Monitors

Monitor4

Table 4-5a. Row Format, Air Monitoring Data File Variables

Variable

Type Required

Notes

Monitor
Name

text

yes

Unique name for each monitor in a particular location.

Description text

Description of the Monitor.

Longitude numeric
(double)

yes Values should be in decimal degree format. Values in the

eastern hemisphere are positive, and those in the western
hemisphere are negative.

Latitude numeric
(double)

yes Values should be in decimal degree format. Values in the

northern hemisphere are positive, and those in the southern
hemisphere are negative.

Metric text no This variable is either blank (signifying that the Values are

Observations, rather than Metric values), or must reference an
already defined Metric (e.g., 1-hour daily maximum) for the
appropriate Pollutant.

Seasonal text no This variable is either blank (signifying that the Values are not

Metric Seasonal Metric values) or must reference an already defined

Seasonal Metric for the Metric (e.g., mean of the 1-hour
maximum for the months of June through August).

Annual text no This variable is either blank (signifying that the values are not

Metric annual statistics) or must be one of: None, Mean, Median, Max,

Min, Sum. (e.g., mean of the 1-hour maximum for the year)

Values comma yes If Metric is blank, either 365 or 8,760 comma separated

separated observations (depending on whether the Pollutant has daily or

values hourly observations). If Metric is defined, but Seasonal Metric

(text) and Annual Metric are blank, 365 metric values. If Seasonal

Metric is defined, but Annual Metric is blank, n seasonal metric
values. If Statistic is defined, one annual statistic value (for
either the Metric (if Seasonal Metric is blank) or the Seasonal

September 2008

-------
Loading Data

Metric. Missing values are signified with a period

Table 4-5b. Row Format, Sample Air Monitoring Data File

Monitor Name

Description

Latitude

Longitude

Metric

Seasonal Metric

Annual Metric

Values

10030010881011

Urban center near park

28.3501

77.1200

5.90,.,., 11,.,..

10270001881011

Urban center near highway

28.4502

77.3113

19.80,., .,3.5,.,., 14.70 4.9C

10331002881011

Northern suburb

28.3524

77.6579

11.5,.,.,1.5,.,.,10.20,.,.,2.5,.,.,2.(

10491003881011

Eastern suburb near highway

28.5432

76.4544

12.60,., .,4.60,., .,7.30 9.2C

10530002881011

28.3511

77.7762

D24HourMean

.,., 5.90,.,.,11,.,..

10550010881011

Industrial area southwest of city

28.9054

77.9241

D24HourMean

19.80 3.5,. ,.,14.70; ,.,4.9C

10690002881011

28.1123

77.3425

D24HourMean

11.5,.,.,1.5 10.20,.,.,2.5,.,.,2.f

10730023881011

28.5834

78.0121

D24HourMean

12.60,.,.,4.60,.,.,7.30 9.2C

10731005881011

28.2351

78.4237

D24HourMean

QuarterlyMean

11.82,13.24,18.79,14.25

10732003881011

28.1598

77.9453

D24HourMean

QuarterlyMean

12.31,13.89,18.27,21.19

10732006881011

28.4286

78.0042

D24HourMean

QuarterlyMean

12.33,15.68,18.49,15.96

10735002881011

28.3469

76.8953

D24HourMean

QuarterlyMean

13.52,13.69,19.04,22.43

10890014881011

Rural area west of city

28.3326

77.4683

D24HourMean

Mean

14.52

10970002881011

Rural area near large farms

28.9985

77.8531

D24HourMean

Mean

16.41

10972005881011

Rural area near national park

28.0014

78.9435

D24HourMean

Mean

15.62

11010007881011

28.1318

77.6289

D24HourMean

Mean

17.17

11130001881011

28.1689

77.4689

D24HourMean

QuarterlyMean

Mean

14.29

11170006881011

27.9845

77.8492

D24HourMean
D24HourMean
D24HourMean

QuarterlyMean

Mean

18.97

11190002881011

27.8934

77.1142

QuarterlyMean

Mean

20.14

11210002881011

27.9923

76.8994

QuarterlyMean

Mean

16.46

4.1.4 Incidence/Prevalence Data

Most health impact functions, such as those developed from log-linear or logistic health
impact functions, estimate the percent change in a health effect associated with a pollutant
change. In order to estimate the absolute change in incidence using these functions, the
baseline incidence rate (and in some cases the prevalence rate) of the adverse health effect
are needed.

The incidence rate is the number of health effects per person in the population per unit of
time, and the prevalence rate is the percentage of people that suffer from a particular
chronic illness. For example, the incidence rate for asthma attacks may be 25 cases per
asthmatic individual per year, and the prevalence rate (measuring the percentage of the
population that is asthmatic) might be six percent.

NOTE: For both incidence and prevalence rates, BenMAP allows the user to have rates
that vary by race, ethnicity, gender, and age group. BenMAP can support multiple sets of
incidence and prevalence rates, if the rates differ by year or by grid definition.

To start adding incidence and prevalence data files, click on the Edit button below the
Incidence/Prevalence DataSets box. The Manage Incidence DataSets window will
appear.

September 2008

-------
Loading Data

$ Manage Incidence DataSets

Available DataSets

Delete

Add

DataSet Incidence Rates:

Endpoint Group Endpoint

Type

Edit

Cancel

In this window you may Add, Edit, and Delete datasets. The section on the left under
Available DataSets lists the incidence/prevalence datasets that are currently in the setup.
The section on the right under the DataSet Incidence Rates identifies the rates in the
selected dataset.

To add a dataset, click the Add button. This will bring up the Incidence DataSet
Definition. Give a name to the dataset that you are creating by typing a name in DataSet
Name box.

September 2008

-------
Loading Data

$ Incidence DataSet Definition

DataSet Name:

|CityOne Incidence

DataSet Incidence Rates:

NOTE: If you have multiple incidence/prevalence datasets that vary by, say, year and grid
definition, then use the name to provide a reference to the year and grid definition.

Click the Load From Database button. In the Grid Definition drop-down list choose the
grid definition that matches the grid definition used to develop the incidence/prevalence
dataset. Then click on the Browse button, to browse for the dataset file. (The format for
the dataset is detailed in the next sub-section.)

Load Incidence/Prevalence Database

Grid Definition:

[County T]

Database:

|3.0\CityOne\CityOne Incidence and Prevalence.xlsj Browse

Cancel

After finding the file, click OK. The Incidence DataSet Definition window will appear,
displaying the rates in the data file that you just loaded.

September 2008

-------
Loading Data

$ Incidence DataSet Definition

DataSet Name:

| City One Incidence

DataSet Incidence Rates:

Endpoint Group
|Acute Bronchitis

Endpoint

Acute Bronchitis Inc

Acute Bronchitis
Acute Myocardial
Acute Myocardial
Acute Myocardial
Acute Myocardial
Acute Myocardial

Pre

Inf Acute Myocardial Inf Inc
Inf Acute Myocardial Inf Inc
Inf Acute Myocardial Inf Inc
Inf Acute Myocardial Inf Inc
Inf Acute Myocardial Inf Inc

Load From Database

Delete

Grid Definition:

I County

Column

Row

Value

0.043

101

0.043

Cancel

If the data look correct, click OK. This will return you to the Manage Incidence
DataSets window.

$ Manage Incidence DataSets

Available DataSets

City One Incidence

Delete

Add

DataSet Incidence Rates:

Endpoint Group Endpoint

Acute Bronchitis Acute Bronchitis

Ty -

Inc

Acute B ronchitis Pre

Acute Myocardial Inf Acute Myocardial Inf Inc
Acute Myocardial Inf Acute Myocardial Inf Inc
Acute Myocardial Inf Acute Myocardial Inf Inc
Acute Myocardial Inf Acute Myocardial Inf Inc
Acute Myocardial Inf Acute Myocardial Inf Inc „

^ i " i Pn

Edit

Cancel

Follow the same procedure for any additional incidence/prevalence datasets that you want
to add to the setup database. When you have finished adding data, click OK in the
Manage Incidence DataSets window. The Incidence/Prevalence DataSets box in the

September 2008

-------
Loading Data

Manage Setups window will show the datasets that you have entered.

4.1.4.1 Format Incidence/Prevalence Data

Table 4-6a presents the variables that can be used in incidence and prevalence datasets,
and Table 4-6b presents a sample dataset that follows this format.

Table 4-6a. Health Incidence and Prevalence Data Set Variables

Field Name

Type

Required

Notes

Endpoint Group text

yes If this does not reference an already defined Endpoint Group,

one will be added.

Endpoint text yes If this does not reference an already defined Endpoint for the

Endpoint Group, one will be added.

Race

text

no Should either be blank (signifying All Races) or reference a
defined Race, such as "Black" (from one or more Population
Configurations).

Ethnicity

text

no Should either be blank (signifying All Ethnicities) or

reference a defined Ethnicity, such as "Hispanic" (from one
or more Population Configurations).

Gender

text

no Should either be blank (signifying All Genders) or reference
a defined Gender (from one or more Population
Configurations).

Start Age
End Age
Column
Row
Value

integer

numeric
(double)

yes
yes
yes
yes
yes

Specifies the low and high ages, inclusive. For example,
Start Age of "0" and End Age of "1" includes infants through
the first 12 months of life and all one-year old infants.

The column and the row link the incidence/prevalence data
with cells from a Grid Definition.

The incidence/prevalence rate for the specified demographic
group for this location.

Type

text

no If value is a prevalence rate, then "Prevalence" should be

specified. Otherwise BenMAP assumes that the value is an
incidence rate.

September 2008

-------
Loading Data

Table 4-6b. Sample Health Incidence and Prevalence Data Set

Endpoint Group

Endpoint

Year

Race

Ethnicity

Gender

Start Age
0

End Age
17

Column

Row

Type

Value

Acute Bronchitis

2000

Black

Hispanic

0.065

Acute Bronchitis

2000

Black

0.058

Acute Bronchitis

2000

0.043

Acute Bronchitis

2000

Black

Hispanic

Prevalence

0.08

Acute Bronchitis

2000

Black

Prevalence

0.07

Acute Bronchitis

2000

Prevalence

0.0587

4.1.5 Population Data

The population data is used to estimate population exposure and in turn any adverse health
effects associated with a change in air pollution. BenMAP allows the user to specify race,
ethnicity, gender, and age of the population, as well as the year of the population estimate.

Population data loaded into BenMAP must be associated with a Population
Configuration, which defines the races, ethnicities, genders, and age ranges present in the
data. Race, ethnicity, and gender are unique text values representing population
subgroups. Age ranges are defined by integer values for starting age and ending age
(inclusive), and a unique text value representing the name of the age range. For example,
"0TO1" might be used as a name for the age range defined by a start age of zero and an
end age of one, thus consisting of infants through the first twelve months of life and all
one-year old infants. The population data provided to BenMAP should then contain
population values for all combinations of race, ethnicity, gender, and age range. The
population values may be real numbers (e.g., non-integer values).

Population data must also be associated with a Grid Definition which specifies the
geographic areas for which the data is available (see for more details the section on Grid
Definitions). If population data is available for multiple grid definitions (cities and
neighborhoods, for example), users can have the option of using different sets of
population data for different analyses. BenMAP can also estimate populations for Grid
Definitions for which no population data is available by calculating spatial overlap
percentages with Grid Definitions for which data is available.

To add population data to BenMAP, click on the Edit button below the Population
DataSets box in the Manage Setups window. The Manage Population DataSets

window will appear.

September 2008

-------
Loading Data

$ Manage Population DataSets Mln ||x

Available DataSets

Delete

Add

Grid Definition:

Population Configuration:

Values:

Race

Ethnicity

Gender

Cancel

Click on the Add button.

This brings up the Load Population DataSet window, which has a number of important
features. The Population DataSet Name box allows you to name the dataset.

The Population Configuration Definition window allows you to define the variables that
are in the population data file to be loaded into BenMAP. Use the drop-down list to
choose an existing population configuration and then view it by clicking the View button,
or you may click the Add button and define a new population configuration.

The Grid Definition drop-down list provides the list of existing grid definitions, and you
need to choose the grid definition that matches that in the population dataset. The Browse
button to the right of the Database box allows to find the datafile that you want to load into
BenMAP.

September 2008

-------
Loading Data

Load Population DataSet

Population DataSet Name:

| City One Tract
Population Configuration:

| City One

Delete Add

View

Grid Definition:

Tract

Database:

BenMAP 3.0\City~ ne\CityOne Population 2000.xls

Browse

P Use Population Growth Weights

Browse |

Cancel OK

The Use Population Growth Weights checkbox should be checked when using
population data generated by the PopGrid software application. The population weights
file assists in forecasting population levels. This topic is fairly involved, so we have
deferred discussion until the appendix.

Defining a Population Configuration

The Population Configuration defines the age range, race, and gender of the variables in
your population database. It is critical that the definitions in the population configuration
match those used in the development of the database. If they do not match exactly,
BenMAP will fail to load the population data. (If this occurs, you need to go back and
either develop a new population configuration to match your data, or you need to revise
your population database so that it matches the population configuration.)

To add a population configuration, click the Add button in the Load Population DataSet
window.

September 2008

-------
Loading Data

Population Configuration Definition

Population Configuration Name:
PopulatioriConfiguration 0

Races

Remove

Genders

R emove

Remove

New

Available Races

ASIAN

BLACK

NATAMER

WHITE

ALL

Available Genders

FEMALE

MALE

ALL

Available Ethnicity

NON-HISPANIC

HISPANIC

ALL

Age Ranges

Low Age

High Age

Delete

Add

Cancel

~ K

In the Population Configuration Name, replace "PopulationConfiguration 0" with a name
of your choosing. Under the Races box click on New and type in the name for any races
present in your population data. The names appear in both the Races list box and the
Available Races list box. (If you later create alternative population configurations, you
can simply drag the relevant names from the Available Races list box into the Races list
box.) Similarly, under the Available Genders and Available Ethnicity list boxes click on
New and type in the name for any ethnicity and gender identifiers present in your
population data.

If you want to remove a category in Races, Genders, and Ethnicity list boxes, then
highlight the entry that you want to remove and click the Remove button. (It is not
possible to remove entries from either the Available Races, Available Genders, or
Available Ethnicity list boxes.)

The next step is to create the age groups that match the age groups in your population file.

September 2008

-------
Loading Data

To start click on the Add button below the Age Ranges list box. The AgeRange
Definition window will appear. Type in the name of the age variable in the Age Range
ID box and the upper bound of the age range in the High Age box. (BenMAP
automatically fills in the value for the Low Age box.) For example, the age range names
(with corresponding low and high ages) might include the following: OtoO, Ilo4, 5to9,
lOtol-t, 15tol9, 20to24, 25to29, 30to34, 35to39, 40to44, 45to49, 50to54, 55to59, 60to64,
65to69, 70to74, 75to79, 80to84, and 85up. The choice of the names is up to you.
However, you must be sure that the names exactly match those in your population datafile.

« AgeRange Defini... |JnJ|_x

Age Range ID:

|0to9

Low Age:

Hiqh Age:

o r

OK |

Click OK when you have defined the age range. If you make a mistake and want to delete
an age definition after you have entered it, click on the Delete button. This will remove
the last age group that you have entered. (Click on it twice if you want to remove the last
two age groups that you entered.) The population configurations can be quite detailed as
in the case of the United States Census population configuration that comes loaded with
BenMAP.

September 2008

-------
Loading Data

Population Configuration Definition

Population Configuration Name:

United States Census

Races

ASIAN
BLACK
NATAMER
WHITE

Remove

Genders

FEMALE
MALE

Remove

New

Available Races

ASIAN

BLACK

NATAMER

WHITE

ALL

Available Genders

FEMALE

MALE

ALL

Ethnicity

Available Ethnicity

NON-HISPANIC

HISPANIC

ALL

Remove | New

Age Ranges

Low Age

High Age

OTOO

1T04

5T09

10T 014

15T019

20T024

25T02S

¦-if-

Delete

Add

Cancel

Or the population configuration can be somewhat simpler as in the case of the population
configuration for CityOne — also pre-loaded with BenMAP:

September 2008

-------
Loading Data

Population Configuration Definition

Population Configuration Name:
| City One

Races

ALL

Remove

Genders

FEMALE
MALE

Remove

New

Available Races

ASIAN

BLACK

NATAMER

WHITE

ALL

Available Genders

FEMALE

MALE

ALL

Available Ethnicity

N ON-HIS PANIC

HISPANIC

ALL

Age Ranges

Low Age

High Age

1to4

5to9

10to19

20to39

40to64

65up

—

Delete

Add

Cancel

Click OK to return to the Load Population DataSet window. From the Grid Definition
drop-down list choose the grid definition (e.g., Tract). Then click on the Browse button to
the right of the Database box to choose the population data file to be loaded into
BenMAP.

To finish loading the population datafile, click OK. The Manage Population DataSets
window will appear displaying the characteristics of the available population datafile that
you just loaded and any other datasets that have already been loaded into BenMAP.

Click OK. In the Population DataSets box of the Manage Setups window you should
see an entry for the population dataset that you just loaded.

September 2008

-------
Loading Data

$ Manage Population DataSets |^||n ||x

Available DataSets

City One T ract

Delete

Add

Grid Definition:

| Tract

Population Configuration:

|C%0ne
Values:

Race

ALL

ALL
ALL
ALL
Ail

Ethnicity

NON-HISPANIC
NON-HISPANIC
NON-HISPANIC
NllN-HRPiMir

Gende "

NON-HISPANIC FEMA

MALE
FEMA
MALE
FF

Cancel

4.1.5.1 Format Population Data

Table 4-7 presents the variables that can be used in population datasets. Note that the
names used for the age ranges do not need to follow this same pattern, and instead should
be based on what seems most appropriate for you.

Table 4-7. Population Data Set Variables

Variable

Type

Required

Notes

Age Range

text

yes

References a defined age range in the associated
Population Configuration.

Column
Row
Year

Race

integer
integer
integer

text

yes
yes
yes

The column and the row link the population data with
cells in a Grid Definition.

The year of the data. Note that this may include historical
population estimates (such as from a census), as well as
population forecasts.

Population numeric yes Population estimate. Note that the estimate is not

(double) restricted to integers.

yes References a defined race in the associated Population

Configuration.

September 2008

-------
Loading Data

Ethnicity

text

References a defined ethnicity in the associated
Population Configuration.

Gender

text

yes References a defined gender in the associated Population

Configuration.

4.1.6 Health Impact Functions

Health impact functions calculate the change in adverse health effects associated with a
change in exposure to air pollution. A typical health impact function has inputs specifying
the pollutant; the metric (daily, seasonal, and/or annual); the age, race, ethnicity, and
gender of the population affected; and the incidence rate of the adverse health effect.

Health impact functions are subdivided by user-specified types of adverse health effects.
The broadest category is the Endpoint Group, which represents a broad class of adverse
health effects, such as premature mortality, cardiovascular-related hospital admissions, and
respiratory-related hospital admissions, among other categories. (BenMAP only allows
pooling of adverse health effects to occur within a given endpoint group, as it generally
does not make sense to sum or average together the number of cases of disparate health
effects, such as premature mortality and chronic bronchitis.) The Endpoint Group may
then be subdivided by user-specified Endpoints. For example, the respiratory-related
hospital admission Endpoint Group, may have separate Endpoints for Asthma-related
hospital admissions and chronic bronchitis-related hospital admissions.

There are a wide range of variables that can be included in a health impact function, to
specify the parameters of the function and to identify its source, such as the author, year,
and location of the study, as well as other pollutants used in the study. The bibliographic
reference for the study may be included, as well as any additional information needed to
identify a particular impact function. (The Reference and Qualifier variables are useful
for this.) A number of health impact functions have been developed based on
epidemiological studies in the United States and Europe. However, researchers have
conducted an increasing number of epidemiological studies in Asia and Latin America that
can be used to develop more location-specific impact functions.

There are a number of issues that arise when deriving and choosing between health impact
functions that go well beyond this user manual. Hence, it is important to have a trained
health researcher assist in developing the impact function data file.

To add health impact functions to BenMAP, click on the Edit button below the C-R
Functions DataSets box in the Manage Setups window. The Manage C-R Functions
DataSets window will appear.

September 2008

-------
Loading Data

$ Manage C-R Function DataSets

Available DataSets

Delete

Add

C-R Functions In DataSet:

Tree

Data

Endpoint Group

Endpoint

Pollutant

Metric

±u

Edit

Cancel

In this window you may Add, Edit, and Delete datasets. The section on the left under
Available DataSets lists the health impact function datasets that are currently in the setup
database. The section on the right under the C-R Functions in DataSet lets you view the
functions in a selected dataset.

To add a new dataset, click the Add button. The C-R Function DataSet Definition
window will appear. Type the name that you want to use for the dataset in the C-R
Function DataSet Name box.

You may then enter functions into this dataset through an externally created database by
clicking the Load From Database button. Alternatively, you may Add, Delete, and Edit
individuals functions within BenMAP.

To add a database click the Load From Database button. In the Load a C-R Function
Database window, click the Browse button and then find and select the health impact
function database that you want to load into your setup.

The C-R Function DataSet Definition window will reappear, and you can then view the
health impact functions that you have loaded into your dataset. Note that the central boxed
area in this window has two parts. On the left is the Tree and on the right is the Data.

September 2008

-------
Loading Data

« C-R Function DataSet Definition

C-R Function DataSet Name:
|CityOne Functions

Tree

Data

Endpoint Group

Endpoint

Pollutant

Metric

Seasonal Metric

Metric Statistic

Hospital Admissit

Acute Bronchitis

Acute Myocardi;

Acute Respirator

Chronic Bronchit

Emergency Roor

Hospital Admissit

Lower Respiratoi

Mortality

Work Loss Days

+
<

Schnnl 1 nss

Load From Database

Delete

Cancel

Add

Edit

The columns within each of the list boxes can be rearranged in order to provide the most
useful display. By clicking and holding the cursor on a column, you may move it within
either the Tree or Data boxes, or you may move columns between boxes. For example,
by clicking and holding on the Pollutant column and then dragging it to the far left of the
tree, you can sort all of the health impact functions by Pollutant. (Note rearranging the
columns is only for display and has no effect on the underlying health impact function
database.)

Clicking OK brings you back to the Manage C-R Function DataSets window. Here you
have the opportunity to view the functions in your dataset. If you have more than one
dataset, you can choose the dataset that you want to view in the Available DataSets in the
left-hand section of the window. In the C-R Functions in DataSet in the right-hand
section of the window, you have the same Tree and Data structure that you have in the
C-R Function DataSet Definition window.

September 2008

-------
Loading Data

$ Manage C-R Function DataSets

Available DataSets

CityQne Functions

Delete

Add

G-R Functions In DataSet:

Tree 1

Data 1

Endpoint Group

Endpoint

Pollutant

Metri

Hospital Admissit

HA, Congestive He;

PM2.5

D24I-

IS'

HA, Dysrhythmia

Edit

Cancel

You may edit this dataset by clicking the Edit button. This will bring up the C-R
Function DataSet Definition window. You may load an additional database to your
dataset, or you may Add, Edit, and Delete individual functions directly in BenMAP.

To edit an existing function, highlight a particular function in the C-R Function DataSet
Definition window, and then click the Edit button. The C-R Function Definition

window appears, and you may change any of the values in the boxes and the drop-down
lists. When you are finished click OK.

September 2008

-------
Loading Data

C-R Function Definition

DataSet Name:

Function:

|BenMAP Health Impact Functions

|(1 -(1 /EXP(B etaxD E LT AQ )))xl ncidenceKPO P

Edit

Baseline Incidence Function:

Endpoint Group: Endpoint:

|lBPlig|lin 1 |Mortality, All Cause

|lncidence*POP

Edit

Beta Distribution: Beta:

|Normal | 0.005826890812
Beta Parameter 1: Beta Parameter 2:

0.002157076225 | 0

Pollutant: Metric: Metric Statistic:

|PM2.5 J |D24HourMean |Mean -1
Seasonal Metric:

|QuarterlyMean

Race: Ethnicity: Gender: Start Age: End Age:

I A \ A I A I30 n I" 24

Constant D escription: Constant Value:
A:| |

Author: Year: Location:

jPopeetal. || 2002 151 cities
Qualifier: Other Pollutants:

I I

Reference:

B:| |

C: | 0
Incidence DataSet: Prevalence DataSet:

| Pope, CA, 3rd, R.T. Burnett, M.J. Thun, E.E. Calle, D. Krewski, K. I to and G.D. Thurston. 20

1 A 1 -i

Variable DataSet:

J A

Cancel | OK

J] > i

¦¦

From the C-R Function DataSet Definition window you can also add health impact
functions to the ones that are already in your dataset. Click the Add button and a blank
C-R Function Definition window will appear, and you can then create new health impact
functions. (We will discuss this in more detail in subsequent examples.)

After defining the new health impact function, click OK. This will take you back to the
C-R Function DataSet Definition window. When you are finished with any editing or
adding of health impact functions click OK. The Manage Setups window will appear.
Here in the C-R Function DataSets box you should see an entry for any health impact
function datasets that you have loaded.

4.1.6.1 Format Health Impact Functions

Table 4-8 presents the variables that can be used in health impact function datasets.

Table 4-8. Health Impact Function Data Set Variables

Variable Type Required Notes

Endpoint text yes If this does not reference an already defined Endpoint Group,

Group one will be added.

Endpoint text yes If this does not reference an already defined Endpoint for the

September 2008

-------
Loading Data

Variable

Type

Required

Notes

Endpoint Group, one will be added.

Pollutant text yes Should reference an already defined Pollutant.

Metric text yes Should reference an already defined Metric for the Pollutant.

Seasonal text no Should either be blank (signifying no Seasonal Metric value)

Metric or reference an already defined Seasonal Metric for the

Metric.

Metric Statistic text

no Should either be blank (signifying no Metric Statistic value)
or be one of: None, Mean, Median, Min, Max, Sum.

Race

text

no Should either be blank (signifying All Races) or reference a

defined Race.

Ethnicity

text

Should either be blank (signifying All Ethnicities) or
reference a defined Ethnicity.

Gender

text

no Should either be blank (signifying All Genders) or reference a

defined Gender.

Start Age integer yes Specifies the low and high ages, inclusive. For example, Start

Age of "0" and End Age of "1" includes infants through the
End Age first 12 months of life and all one-year old infants

Author text no The author(s) of the study from which the function is derived.

Year integer no The year of publication of the study.

Location text no The location of the study.

Qualifier text no Provides additional information to identify a particular health

impact function, such as when a particular study has multiple
functions.

Other
Pollutants

text

no Identifies other pollutants that were included simultaneously
in the estimation equation for the pollutant of interest.

Reference

text

no Bibliographic reference, included to identify the source in the

health literature.

Function text yes The functional form, interpreted (executed) by BenMAP

when running an analysis to estimate air pollution-related

health impacts. For example, the log-linear form is as
follows: "(l-(l/EXP(Beta*DELTAQ)))*Incidence*POP".

September 2008

-------
Loading Data

Variable Type Required Notes

Baseline text no The functional form, interpreted (executed) by BenMAP to

Incidence estimate health impacts due to all causes. This typically has

Function the form: "Incidence*POP".

Beta

numeric
(double)

Mean value of the Beta distribution.

Distribution
Beta

text

no If the Beta has no distribution, any value is acceptable.

Otherwise, should be one of: Normal, Triangular, Poisson,
Binomial, LogNormal, Uniform, Exponential, Geometric,
Weibull, Gamma, Logistic, Beta, Pareto, Cauchy, Custom.

Parameter 1 numeric no Parameter 1 of the Beta distribution (meaning depends on the

Beta (double) distribution - for Normal distributions this represents the

standard deviation).

Parameter 2 numeric no Parameter 2 of the Beta distribution (meaning depends on the

Beta (double) distribution - for Normal distributions this is not required).

Name A
A

text

numeric
(double)

Description of variable A.

A constant value which can be referenced by the Function
(see below).

Name B
B

text

numeric
(double)

Description of variable B.

A constant value which can be referenced by the Function
(see below).

Name C
C

text

numeric
(double)

Description of variable C.

A constant value which can be referenced by the Function.

Incidence
DataSet

text

no Specifies the dataset from which incidence data will be

derived. The user may choose from multiple datasets.
(Initially this field may be left blank.)

Prevalence
DataSet

text

no Specifies the dataset from which prevalence data will be

derived. The user may choose from multiple datasets.
(Initially this field may be left blank.)

Variable
DataSet

text

no Specifies the dataset from which "variable" data (e.g., income
data) will be derived. The user may choose from multiple
datasets. (Initially this field may be left blank.)

September 2008

-------
Loading Data

4.1.7 Variable Data

Health Impact functions and/or Valuation functions may sometimes refer to variables other
than those for which BenMAP automatically calculates values. For example, some
valuation functions reference the median income within each area of analysis. To
facilitate this, BenMAP allows you to load datasets of variables, which may either be
global values or may vary geographically (meaning they are associated with a Grid
Definition).

To add variables to BenMAP (such as income and other miscellaneous variables that
might be needed in the analysis), click on the Edit button below the Variables DataSets
box in the Manage Setups window. The Manage Setup Variables DataSets window
will appear.

$ Manage Setup Variable Dat... f^~1|n |[x |

Available DataSets

DataSet Variables

Delete Add

Edit

Cancel OK

In this window you may Add, Edit, and Delete datasets. The section on the left under
Available DataSets lists the variables datasets that are currently in the setup database.
The section on the right under the DataSet Variables lets you view the variables in a
selected dataset.

To add variables click the Add button. This will take you to the Setup Variable DataSet
Definition window. In this window you may add externally created variables through the
Load From Data button for any of your predefined grid definitions. Alternatively, by
clicking the Add button you may create global variables to your dataset that are not tied to
a specific geographic region.

September 2008

-------
Loading Data

« Setup Variable DataSet Definition |T]fn](x]

DataSet Name:
| CityO ne Variables)

DataSet Variables

Delete

Add

Load From Database

G lobal G eographically Variable
Grid Definition:

Values:

Cancel

To start, type the name that you want to use for the dataset in the DataSet Name box.

(This is a name that is internal to BenMAP and used just for identification.)

To add an externally created variable database, click the Load From Database button.
This will bring up the Load Variable Database window choose. Here you need to choose
the grid definition from the Grid Definition drop-down list that matches the level of
aggregation in your variable database. Then you may use the Browse button to find and
select the desired data file.

Load Variable Database

Grid Definition:

| County

Database:

Browse

|m Files\BenMAP 3.0\CitvO ne\CitvO ne Variables.xls

Cancel

After choosing the database, click OK. This takes you back to the Setup Variable
DataSet Definition window. This allows you to see the variables in the dataset and their
values.

100

September 2008

-------
Loading Data

$ Setup Variable DataSet Definition I-J

DataSet Name:
| City One Variables

DataSet Variables
median income

Delete

Add

Load From Database

Global Geographically Variable
Grid Definition:

| County
Values:

Column

Row

Value

101

31261.071

38186.77

43760.31

39871.19

iYl

Cancel OK

You may also add Global variables to the dataset by clicking on the Add button in the
Setup Variable DataSet Definition window. After clicking Add, type the name of the
variable where it says Setup Variable 0 in the DataSet Variables list box, and then in
Value box type the value of the global variable.

101

September 2008

-------
Loading Data

« Setup Variable DataSet Definition [-J (~](><]

DataSet Name:

| CityO ne Variables

Global I Geographically Variable |

DataSet Variables

Value:

medianjncome

I 1l

income elasticity

Delete Add

Load From Database

Cancel OK

When finished adding variables, click OK. This will take you to the Manage Setup
Variable DataSets window. In the Available DataSets list box there is an entry for the
dataset that you just added. And in the DataSet Variables list box are the variables in the
highlighted dataset.

102

September 2008

-------
Loading Data

$ Manage Setup Variable Dat... - n x

Available DataSets

DataSet Variables

City One Variables

| medianjncorne

income elasticity

Delete Add

Edit

Cancel 1 OK

'¦

At this point you may click Add to load an additional dataset, click Edit to edit the
selected dataset, click Delete to delete the selected dataset, or complete this variable
management step by clicking OK.

Clicking OK returns you to the Manage Setups window, where in the Variable DataSets
box, you should see an entry for the variable dataset that you just created.

4.1.7.1 Format Variable Data

Table 4-9a presents the variables that can be used in variable datasets, and Table 4-9b
presents a sample of what a dataset might look like.

Table 4-9a. Variable Data Set Variables

Variable

Type

Required

Notes

Column

integer

yes

The column and the row link the population data with

Row

integer

yes

cells in a Grid Definition.

Variable Name

numeric
(double)

103

September 2008

-------
Loading Data

Table 4-9b. Sample Health Impact Function Data Set

COL

ROW

median income

39871.19

43760.31

38186.77

42298.49

101

31261.07

4.1.8 Inflation Data

It may be desirable for the dollar values generated by Valuation functions to account for
inflation, and generate economic benefits for years other than the year initially specified
by your valuation data. To do this, you can load inflation data into BenMAP, and then
include a reference to your inflation data when developing valuation functions. (We give
an example of this below.)

The valuation functions should have a consistent dollar year, and this dollar year has to be
kept in mind when developing the inflation datasets. That is, whichever dollar year is used
for your valuation functions, then the inflation values for that year should be set to 1. For
example, in the U.S. setup, the valuation functions have a dollar year of 2000, so the
inflation datasets have a value of 1 for the year 2000. (Values for years earlier than 2000
are less than 1, and values for years after 2000 are greater than 1, because inflation has
increased from one year to the next.)

The U.S. setup in BenMAP provides inflation factors for three different types of values:

All Goods Index can be used to adjust the value of generic goods.

Medical Cost Index can be used to adjust the value of medical expenses.

Wage Index can be used to adjust the value of wages.

If a valuation function includes an estimate of wage income, for example, this value could
be multiplied by the Wage Index adjustment factor to get the specified currency year. For
example, in valuing workloss days, the U.S. setup uses a function like the following:
DailyWage*WageIndex, where the DailyWage is specified in year 2000 dollars. The
Wagelndex scales this DailyWage value up or down depending on the currency year you
have chosen. (The choice of currency year can be made with the One-Step Setup,
discussed here, and with the Advanced button when creating an APV file, discussed here.)
If the currency year is 2000, then the Wagelndex has a value of 1 and no change is made
to the DailyWage. If the currency year is specified to, say, 2005, then the Wagelndex will
have a value greater than 1 because of the inflation that has occurred between 2000 and
2005.

To add inflation data to BenMAP, click on the Edit button below the Inflation DataSets
box in the Manage Setups window. The Inflation DataSet Manager window will
appear. In this window you may Add, Edit, and Delete datasets. The section on the left
under Available DataSets lists the inflation datasets that are currently in the setup
database. The section on the right presents the inflation factors in a selected dataset.

104

September 2008

-------
Loading Data

Click on the Add button. In the Load Inflation DataSet window, type in the name of the
dataset in the Inflation DataSetName box, and then click on the Browse button to the
right of the Database box to choose the dataset that you want to import. Click OK. The
Inflation DataSet Manager window will appear. Here you may view the data that you
just loaded.

$ Inflation DataSet Manager [_J[n)[xJ

Available DataSets

CityOne Inflators

Delete

Add

Year
11930

All Goods Index Medical Cost Index Wage Index

1981

1982

1983

1934

1935
193G

1937

1938

1989

1990

1991

0.46999999880790r 0.28000000119209: 0.43999999761581'

0.51999998092651'
0.56000000233418E
0.56999999284744:
0.60000002384185E
0.62000000476837:
0.62999999523162E
0.64999997615814:
0.68000000715255
0.72000002861023
0.75

0.79000002145767:

0.31000000238418E
0.34999999403953E
0.37999999523162E
0.40000000596046'
0.43000000715255;
0.46000000834465
0.49000000953674:
0.52999997138977
0.56999999284744:
0.62000000476837:
0.6700000166893

0.47999998927116'
0.51999998092651'
0.54000002145767:
0.56999999284744:
0.58999997377395E
0.61000001430511E
0.62999999523162E
0.66000002622604'
0.68000000715255;
0.70999997854232E
0.74000000953674:

Cancel

~ K

At this point you may add more data by clicking Add, or you may complete this step by
clicking OK. Clicking OK takes you to the Manage Setups window, where in the
Inflation DataSets box, you should see an entry for the inflation dataset that you just
loaded.

4.1.8.1 Format Inflation Data

Table 4-10a presents the variables that can be used in inflation datasets, and Table 4-10b
presents a sample of what a dataset might look like. Note that if you are loading your own
inflation data, you can use different names than the ones specified below. Instead of
specifying "AllGoodsIndex" you could have a variable called "General Index" — this is
fine so long as you make sure that your valuation functions properly reference these
inflation variables.

Table 4-10a. Inflation Data Set Variables in U.S. Setup

Variable Type Required Notes

Year integer yes The year of the data. Note that this will typically only

include historical estimates.

105

September 2008

-------
Loading Data

AllGoodsIndex

integer

All goods inflation index value.

MedicalCostlndex

integer

Medical cost inflation index value.

Wagelndex

integer

Wage inflation index value.

< variable name >

integer

Additional indices can be specified.

Table 4-10b. Sample Inflation Data Set

YEAR

AllGoodsIndex

MedicalCostlndex

Wagelndex

1995

0.75

0.62

0.71

1996

0.79

0.67

0.74

1997

0.81

0.72

0.76

1998

0.83

0.77

0.78

1999

0.86

0.8

0.81

2000

0.88

0.84

0.83

2001

0.91

0.87

0.86

2002

0.93

0.89

2003

0.94

0.92

2004

0.96

2005

2006

1.02

1.04

1.03

4.1.9 Valuation Data

BenMAP allows the valuation estimates to vary by Endpoint Group, Endpoint, and Age
(note that multiple estimates may be provided for a particular combination). BenMAP
allows the valuation function to be quite detailed, and allows an uncertain parameter (A)
with a user-specified distribution. The user can modify the valuation function with a
number of constant values (B, C, and D) that might represent an adjustment for inflation,
income growth, income elasticity, or, say, purchasing power parity. Finally, BenMAP has
two fields to more clearly identify the valuation function (i.e., Qualifier and Reference).

When reviewing the economic literature to develop a valuation database or simply add
additional valuation functions to an existing database, it is important to have an economist
assist. For an overview of valuation, see the Overview of Valuation section in the chapter
on Aggregation, Pooling, and Valuation.

To add valuation functions to BenMAP, click on the Edit button below the Valuation
DataSets box in the Manage Setups window. The Manage Valuation Function
DataSets window will appear.

106

September 2008

-------
Loading Data

$ Manage Valuation Function DataSets |-J(n](>r

Available DataSets

Delete

Add

Valuation Functions In DataSet:

Tree

Data

Endpoint Group

Endpoint

Qualifier

Referen

< I I

Edit

Cancel

In this window you may Add, Edit, and Delete datasets. The section on the left under
Available DataSets lists the valuation datasets that are currently in the setup database.
The section on the right under the Valuation Functions in DataSet lets you view the
valuation factors in a selected dataset.

To add a dataset click on Add. In the Valuation Function DataSet Name type the name
of the dataset that will be added.

You may load valuation data with an externally-created database, or you may add
individual valuation functions from within BenMAP. To load an externally-created
valuation database, click on the Load From Database button. This will bring you to the
Load a Valuation Function Database window.

In the Load a Valuation Function Database window use the File of Type drop-down list
to choose the type of file (e.g., MS Access Database). Choose the valuation database and
click Open. This will bring you back to the Valuation Function DataSet Definition

window. Here you can view the valuation functions that you have in your database.

Note that the large box is divided in two. On the left is the Tree and on the right is the
Data. The tree can be expanded by clicking on the "+" sign, or contracted by clicking on
the sign.

The columns within each of the list boxes can be rearranged in order to provide the most
useful display. By clicking and holding the cursor on a column, you may move it within
either the Tree or Data boxes, or you may move columns between boxes. (Note that
rearranging the columns is only for display and has no effect on the underlying Valuation
function dataset.)

107

September 2008

-------
Loading Data

$ Valuation Function DataSet Definition

Valuation Function DataSet Name:
CityOne Valuation

T ree

Data

Endpoint Group Endpoint

Qualifier

Reference

Start Age

End Age

Acute Bronchitis |

Acute Bronchitis

WTP: G day illness, I

Acute Respirator

Chronic Bronchit

Emergency Floor

Hospital Admissit

—

Hospital Adrnissk

Lower Respiratoi

Mortality

Work I nss Daus

Load From Database

Delete

Cancel

Add

Edit

In the Valuation Function DataSet Definition window you can also edit the functions
already existing in your data set by highlighting a particular Valuation function and then
clicking the Edit button.

The Valuation Function Definition window allows you to change any of the values in the
boxes and the drop-down lists. When you are finished click OK.

108

September 2008

-------
Loading Data

Valuation Function Definition

DataSet Name:

Function:

| CityO ne Valuation

|A"AIIGoodslridex

A Description:

Endpoint Group: Endpoint:

t| I Acute Bronchitis »|

|WTP in 2000$

355.849243164063

A Distribution:

A Parameter 2:

Qualifier:

| Uniform
A Parameter 1:

|WTP: 6 day illness, CV studies

105.059127807617

606.639404296875

Reference:

Constant Description:

Constant Value:

B: adjustment for household inc

0.973810970783

Start Age: End Age:

C:| |

o % i17 n

I 0

Cancel OK

From the Valuation Function DataSet Definition window you can also add Valuation
functions to the ones that are already in your dataset. Click the Add button and a blank
Valuation Function Definition window will appear, and you can then create a new
Valuation function.

After defining the new Valuation function, click OK. This will take you back to the
Valuation Function DataSet Definition window. When you are finished with any
editing or adding of valuation functions click OK. The Manage Setups window will
appear. Here in the Valuation Function DataSets box you should see an entry for the
valuation function dataset that you just loaded.

4.1.9.1 Format Valuation Data

Table 4-11 presents the variables that can be used in valuation datasets.

Table 4-11. Valuation Data Set Variables
Field Name Type Required Notes

Endpoint Group text yes If this doesn't reference an already defined Endpoint Group,

one will be added.

Endpoint text yes If this doesn't reference an already defined Endpoint for the

Endpoint Group, one will be added.

Qualifier text no Provides additional information to identify a particular

109

September 2008

-------
Loading Data

valuation function.

Reference text no Bibliographic reference, included to identify the source in the

economic literature.

Start Age
End Age
Point Estimate

Function

integer

numeric
(double)

text

no Specifies the low and high ages, inclusive. For example, Start

Age of "0" and End Age of " 1" includes infants through the
no first 12 months of life and all one-year old infants.

yes Central estimate of the unit value.

yes The functional form, interpreted (executed) by BenMAP when
running an analysis. The BenMAP User Manual will contain
more information on the format of this field.

A Description
A

A Distribution

Constant
Description

text

numeric
(double)

text

A Parameter 1 numeric
(double)

A Parameter 2 numeric
(double)

text

Constant Value numeric
(double)

no Description of variable A.

no Mean of the A distribution.

no If A has no distribution, any value is acceptable. Otherwise,
should be one of: Normal, Triangular, Poisson, Binomial,
LogNormal, Uniform, Exponential, Geometric, Weibull,
Gamma, Logistic, Beta, Pareto, Cauchy, Custom.

no Parameter 1 of the A distribution (meaning depends on the

distribution - for Normal distributions this represents the
standard deviation).

no Parameter 2 of the A distribution (meaning depends on the

distribution - for Normal distributions this is not required).

no Description of variables, B, C, and D.

NO A constant value (denoted by B, C, and/or D) which can be
referenced by the Function.

4.1.10 Income Growth Data

As real income increases, willingness-to-pay (WTP) to avoid air-pollution related
morbidity effects and premature mortality should grow. BenMAP allows users to adjust
the WTP estimates to account for the growth in income over time. This adjustment is a
combination of data on income growth and estimated income elasticity of demand, which
measures the responsiveness of the quantity demanded of a good to the change in the
income of the people demanding the good; this is distinct from elasticity of demand, which
quantifies the change in demand for goods and services as a result of changes in price for
those goods and services. This section describes how to load the data adjusting for income
growth and how EPA developed these adjustment factors.

110

September 2008

-------
Loading Data

Note that the WTP estimates in the default valuation functions in the U.S. setup are
assumed to be based on 1990 income, so the income growth adjustments are all relative to
1990. That is, the income growth data has a value of 1 in 1990, and because income has
generally increased over time in the U.S., the income growth values are typically greater
than 1 after 1990. (An exception is 1991, when incomes declined slightly in the U.S.)

If you load in your own valuation functions and/or income growth adjustment factors, be
sure that you have carefully considered the income year. For example, if your valuation
functions are based on income in the year 2005, then the income growth adjustment should
have a value of 1.0 in 2005, because no adjustment is necessary. As you forecast into the
future, under the assumption that incomes go up over time, then your income growth
adjustment factors would have values greater than 1.0 for years past 2005, and would have
values less than 1.0 for years prior to 2005.

To add income growth data to BenMAP, click on the Edit button below the Income
Growth Adjustments box in the Manage Setups window. The Income Growth
Adjustment DataSet Manager window will appear.

In this window you may Add and Delete datasets. The section on the left under Available
DataSets lists the income growth datasets that are currently in the setup database. The
section on the right presents the income growth adjustment factors in a selected dataset.

Click on the Add button. In the Load Income Growth Adjustment Factors DataSet

window, type in the name of the dataset in the Income Growth Adjustment
DataSetName box, and then click on the Browse button to the right of the Database box
to choose the dataset that you want to import.

Click OK. The Income Growth Adjustment DataSet Manager window will appear.

111

September 2008

-------
Loading Data

Here you may view the data that you just loaded.

Developing Income Growth Adjustment Factors

Adjusting WTP to reflect growth in real income requires three steps:

1. Identify relevant income elasticity estimates from the peer-reviewed literature

2. Calculate changes in future income

3. Calculate adjustments to WTP based on changes in future income and income
elasticity estimates

1. Identifying income elasticity estimates

Income elasticity estimates relate changes in demand for goods to changes in income.
Positive income elasticity suggests that as income rises, demand for the good also rises.
Negative income elasticity suggests that as income rises, demand for the good falls.
BenMAP does not adjust Cost-of-Illness (COI) estimates according to changes in income
elasticity due to the fact that COI estimates the direct cost of a health outcome; instead we
adjust this metric using inflation factors described above. BenMAP includes income
elasticity estimates specific to the type of health endpoint associated with the WTP
estimate. BenMAP contains elasticity estimates for three types of health effects: minor,
severe and premature mortality. Minor health effects are those of short duration. Severe, or
chronic, health effects are of longer duration. Consistent with economic theory, the peer
reviewed literature indicates that income elasticity varies according to the severity of the
health effect. Below we summarize the health endpoints considered minor and severe
within BenMAP.

Minor

> Asthma exacerbation

> Acute bronchitis

> Acute respiratory symptoms (minor restricted activity days)

> Lower respiratory symptoms

> Upper respiratory symptoms

Severe

> Chronic bronchitis

> Chronic asthma

A review of the literature revealed a range of income elasticity estimates that varied across
the studies and according to the severity of health endpoint. Table 4-12 summarizes the
income elasticity estimates found in BenMAP to adjust minor health effects, severe health
effects and premature mortality. Here we have provided a lower-, upper- and
central-estimate for each type of health endpoint.

112

September 2008

-------
Loading Data

Table 4-12. Income Elasticity Estimates

Health Endpoint

Lower Bound

Central Estimate

Upper Bound

Minor Health Effect

0.04

0.15

0.30

Severe and Chronic
Health Effects

0.25

0.45

0.60

Premature Mortality

0.08

0.40

1.00

2. Calculating changes in future income

The next input to the WTP adjustment is annual changes in future income. To estimate
changes in future income, EPA used the Standard & Poor's projections of future changes
in Gross Domestic Product (GDP) occurring after the year 2010. EPA then divided the
projected change in GDP by the Woods & Poole projected change in total US population
to produce an estimate of the future GDP per capita.

3. Calculating changes in WTP

The income elasticity estimates from Table 4-12 and the estimated changes in future
income may then be used to estimate changes in future WTP for each health endpoint. The
adjustment formula follows four steps:

A WTP

c_ ~wtp _
-------
Loading Data

Table 4-13. Income-Based WTP Adjustments by Health Effect and Year

M incr Hes/th Bxftq'rt

S&ere Heath Brfoart

Mataiy

Year

Low

Mid

Upper

Low

Mid

Upper

Low

Mid

Upper

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

2018

2019

2020

2021

2022

2023

2024

0.9992
0.999818
1.000308
1.001277
1.001686
1.00241
1.003365
1.003817

1.0042
1.003788
1.004299
1.004889
1.005484
1.006086
1.006653
1.007225

1.0078
1.008379
1.008959
1.009541
1.010619
1.011646
1.012625
1.013561
1.014457
1.015315
1.016137
1.016928
1.017687
1.018419
1.019122
1.019798
1.020447
1.021072

0.997201
0.999363
1.001079
1.004476
1.005913
1.00846
1.011828
1.013425
1.014776
1.013323
1.015126
1.017218
1.019328
1.021464
1.023482
1.025519
1.027571
1.029635
1.03171
1.033796
1.037665
1.041361
1.044897
1.048282
1.05153
1.054647
1.057644
1.060527
1.063305
1.065985
1.068567
1.071051
1.073442
1.075746

0.994013
0.998636
1.002314
1.009617
1.012715
1.018216
1.025519
1.028991
1.031932
1.028768
1.032696
1.037261
1.04188
1,046565
1.051003
1.055492
1.060025
1.064598
1.069205
1.073846
1,082486
1,090777
1.09874
1.106398
1.113771
1.120876
1.127729
1.134347
1.140745
1.146937
1,15292
1,158694
1,164269
1,169656

0.995008
0.998863
1.001928
1.008008
1,010584
1.015157
1.021221
1.024101
1.02654
1.023916
1.027173
1.030955
1.034778
1.038654
1.042323
1.046031
1.049772
1.053544
1.05734
1.061162
1.068269
1.07508
1,081612
1.087886
1.093919
1.099725
1.10532
1.110717
1.115928
1.120966
1.12583
1.130519
1.135042
1.139409

0.991033
0.997955
1.003474
1.01446
1.019133
1.027449
1.038525
1.043803
1.048284
1.043465
1.049448
1.056418
1.063484
1.07067
1.077493
1.08441
1.09141
1.098488
1.105635
1.112853
1.126335
1.139328
1.151859
1.163958
1.175653
1.186964
1.197914
1.208525
1.218817
1.22881
1.238497
1.247875
1.256956
1.265756

0.988061
0.997274
1.004634
1.019327
1.025592
1.036768
1.051698
1.058834
1.064901
1.058376
1.066479
1.075937
1.08555
1.09535
1.104675
1.114152
1.123765
1.133508
1.14337
1.153355
1.17207
1.190187
1.207737
1.224753
1.241267
1.257303
1.272886
1.288044
1.3028
1.317176
1.33116
1.344743
1.357937
1.370763

0.9984
0.999636

.000617
.002556
.003375
.004825
.006742
.007649
.008417
.007591
.008616
.009802
.010999
.012209
.013351
.014503
.015662
.016828
.017998
.019173
.021351
.023427
.025411
.027307
.029124
.030866
.032537
.034144
1.03569
1.03718
1.038614
1.039991
1.041 316
1.042592

992025
998182
003087
012843
016989
024362
034171
038842
042804
038542
043834
049992
056232
062572
068587
074681
080843
087068
1.09335
099688
111515
122896
133857
144425
154627
164482
1.17401
183233
192168
200834
209226
217341
225191
1.23279

4.1.10.1 Format Income Growth Data

Table 4-14a presents the variables that can be used in income growth adjustment datasets,
and Table 4-14b presents a sample of what a dataset might look like.

Table 4-14a. Income Growth Data Set Variables

Variable Type Required Notes

114

September 2008

-------
Loading Data

Year integer yes The year of the data. Note that this will include

historical estimates as well as forecasts.

Mean integer yes Mean income growth adjustment factor.

EndPoint Group text yes

Table 4-14b.

Sample Income Growth Data

Year

Mean

EndpointGroup

1990

1.0000

Acute

Bronchitis

1991

0.9910

Acute

Bronchitis

1992

0.9980

Acute

Bronchitis

1993

1.0035

Acute

Bronchitis

1994

1.0145

Acute

Bronchitis

1995

1.0191

Acute

Bronchitis

1996

1.0274

Acute

Bronchitis

1997

1.0385

Acute

Bronchitis

1998

1.0438

Acute

Bronchitis

1999

1.0483

Acute

Bronchitis

2000

1.0435

Acute

Bronchitis

2001

1.0494

Acute

Bronchitis

2002

1.0564

Acute

Bronchitis

2003

1.0635

Acute

Bronchitis

2004

1.0707

Acute

Bronchitis

2005

1.0775

Acute

Bronchitis

4.1.11 QALY Data

To add QALY data to BenMAP, click on the Edit button below the QALY Distribution
DataSets box in the Manage Setups window. The QALY DataSet Manager window
will appear.

115

September 2008

-------
Loading Data

$ QALY DataSet Manager

Endpoint Group- (Acute Myocardial Infarction Endpoint' |Acute Myocardial Infarction, Nonfatal
Qualifier- |Acute Myocardial Infarction I

¦i DR OX

Available Datasets

AMI - Acute Myocardial Infarction, Nonfatal

AMI DR OX

Undiscounted

Base Weight is T[0.9-0.1] centered at 0.35
Weight during AMI is 0.605
Duration of AMI is U[5.5-22] days
Weight after AMI:
wo/CHF wo/ANG (p=0.39): 0.93
wo/CHF w/ANG (p=0.41): U[0.70-0.B9]

AMI DR 3X
AMI DR IX
AMILE DR OX
AMILE DR 3X
AMILE DR IX

MILYAMI DR OX

MILYAMI DR 3X

Start Age

EndAge

QALY

MILYAMI DR 7X

2.835171175003

CB DR OX

5.200510501861

CB DR 3X

2.949071884155

CB DR IX

2.349098205566

MILY CB DR OX

3.400484561920

MILY CB DR 3X

5.939572334289

MILY CB DR IX

6.422775745391

MILY Mort DR OX

6.769740104675

1.824124455451

Delete

Add

2.391076326370

Cancel

In this window you may Add and Delete datasets. The section on the left under Available
DataSets lists the QALY datasets that are currently in the setup database. The sections on
above and to the right describe the data and let you view the QALY factors in a selected
dataset.

To add a dataset click on Add. This will bring up the Load QALY Dataset window.

116

September 2008

-------
Loading Data

toad QALY DataSet

QALY DataSet Name

UALYDataSet 0

Database:

Endpoint Group:

Browse

Endpoint:

Qualifier: I
Description:

Cancel

In the QALY DataSet Name type the name of the dataset that will be added. Then click
the Browse button to find the QALY database that you want to load. This will bring up
the Open window. Use the File of Type drop-down list to choose the type of file (e.g.,
MS Access Database). Choose the QALY database and click Open. This will bring you
back to the Load QALY DataSet window.

Each QALY dataset is assumed to be specific to an Endpoint Group and Endpoint, so use
the Endpoint Group and Endpoint drop-down menus to identify the appropriate
selections.

Next use the Qualifier box to provide a brief description of the datafile. When creating an
APV Configuration, you will see this Qualifier description, so choose a description that
adequately captures what you have in your QALY datafile. There can be multiple QALY
datafiles, in which case the Qualifier can be quite useful.

Finally, include a detailed description in the Description box. This is not required, but
can be useful when comparing different QALY datasets.

4.1.11.1 Format QALY Data

Table 4-15a presents the variables that can be used in QALY datasets, and Table 4-15b

117

September 2008

-------
Loading Data

presents a sample of what a dataset might look like.

Table 4-15a. QALY Data Set Variables

Variable

Type

Required

Notes

Start Age
End Age

integer

yes

Specifies the low and high ages, inclusive. For
example, Start Age of "0" and End Age of "1"
includes infants through the first 12 months of life and
all one-year old infants

QALY

integer

yes

QALY value

Table 4-15b. Sample QALY Data Set

Start Age

EndAge

QALY

5.666993

2.758569

5.485653

3.647018

4.196073

3.501245

6.018638

2.057168

4.2 Export and Import Setups

BenMAP allows you to export and import entire databases (all Setups), individual Setups,
and parts of individual Setups (e.g. all GridDefinitions, or individual health impact
Function DataSets). This functionality can be used to share data with other BenMAP
users, as well as to allow you to move databases between computers. In particular, all of
the steps involved in creating a Setup can be done just once, after which the data can be
exported and then imported on other computers.

4.2.1 Export Setups

To export part or all of an existing setup, go to the Tools drop-down menu, and choose
the Database Export option.

118

September 2008

-------
Loading Data

Help

Air Quality Grid Aggregation
Model File Concatenated

Database Export

Database Import
Export Air Quality Grid
GIS/Mapping
Modify Setup
Neighbor File Creator
One-Step Setup

Use BenMAP: Which Analysis Meets your Needs?

uality data
ol to apply
te a report.

-nirxmantyer d Creation

JUL

Preloaded
EPA parameters

J3-

Report

Custom Analysis

Step 1 — Import ail quality data

Air Quality Grid Creadon

Step 2 — Set custom parameters

Incidence Estimation

Step 3 — Use results from Step 2
to set custom parameters

Pooling. Aggregation
and Valuation

Step 4 — Run report

Report

Active Setup:

This will bring up the DatabaseExportForm window. Initially, all of the setups are in a
data tree, which is initially collapsed into the dataset Available Setups. To view the
available Setups, click on the "+" sign to the left of Available Setups. This will expand
the data tree. To contract the tree simply click on the sign.

Choose an dataset to export by clicking on it. In the screenshot below, we are have chosen
to export EPA Standard Monitors.

119

September 2008

-------
Loading Data

$ DatabaseExportForm

Select Object to Export:

El- Available Setups
- United States
EE- Grid Definitions
EE- Pollutants
|- Monitor DataSets

EPA Standard Monitors

EE - Incidence/Prevalence DataSets

EE - Population DataSets

EE - C-R Function DataSets

EE-Variable DataSets

+ Inflation DataSets

EE- Valuation DataSets

EE Income Growth Adjustments

Cancel OK

You may choose all available Setups, individual Setups, the various collections within a
Setup (all GridDefinitions, all Pollutants, all Population DataSets, etc.), and individual
datasets within the various collections within a Setup. Click OK when you have selected
the dataset to export.

This will bring up the Export Database Objects To File window. From here you may
name the dataset and choose the folder into which to export it. Type the dataset name the
Filename box, and in the Save in drop-down menu locate the folder that you want to use.

120

September 2008

-------
Loading Data

Export Database Object To File

Save in: fcD BenMAP 3.0

My Recent
Documents

IF7~3

Desktop

My Documents

My Computer

My Network
Places

IpfrAir Quality Grids
£3 City One

QConfiguration Results

Q Configurations

QMaps

El Reports

QShapefiles

j®]Ad Hoc Functions.bdb

H)DA_Mi.bdb

H] DA_Mort.bdb

@]DA-ER Functions.bdb

Q Te mperature. bdb

File nanne:

"3] hd c? iii'r

Save

Save as type: BenMAP Database Files (K.bdb)

Cancel

NOTE: Exported BenMAP database files have a .bdb extension, and are a binary format
not suitable for viewing in external applications.

4.2.2 Import Setups

To import part or all of an existing setup, go to the Tools drop-down menu, and choose
the Database Import option. This will bring up the Import Object Into Database
window.

121

September 2008

-------
Loading Data

$ Import Object Into Database (TlfcT [x~

Database Object File:

Browse

Target Setup:

1 J

— H

Cancel

[ °K. i

i 1

The Database Object File box identifies the file that you want to import. Click on the
Browse button to locate the file to import. This will bring up the Import Database
Object From File window.

Select the file that you want to import, and then click Open. This will bring you back to
the Import Object Into Database window. Click OK to finish the import process.

« Import Object Into Database - (n"[x

Database Object File:

Browse

|C:\Program Files\BenMAF' 3.0V\d Hoc Functions.bd

Target Setup:

[itminr;

Cancel

" "

If the file contains a sub-Setup item, such as a collection (e.g., a set of grid definitions) or
an individual item (e.g., a single grid definition from among many available), select the
Setup into which it should be imported. Select the Setup from the drop-down list below
Target Setup. Click OK to finish.

NOTE: Duplicates of datasets (typically identified by their names - e.g. "Population
CityOne") will overwrite existing data in a Setup. New datasets (i.e., non-duplicated) will
be added to the Setup.

4.3 Questions Regarding Loading Data

>I've loaded new baseline incidence data, but BenMAP won't let me select it in the
configuration stage

When formatting these data for importation to BenMAP, take special care to ensure that

722

September 2008

-------
Loading Data

you have specified the health endpoints correctly. The baseline incidence rate must be
associated with a specific health endpoint and endpoint group in BenMAP. Be sure that
you have recorded the endpoint group and endpoint exactly as it is recorded in
BenMAP. For example, if the baseline incidence rate is for asthma-related hospital
admissions, be sure you have recorded the endpoint group as "Hospital Admissions,
Asthma" and the endpoint as "HA, Asthma".

>I've loaded a new grid and new population data into BenMAP but I can't seem to
use these new data

Be sure to load the new grid definition first. Next, load the population dataset and be
sure to select your new grid definition.

>Why doesn't BenMAP have health impact functions for pollutants other than
particulate matter and ozone?

EPA has included health impact functions for particulate matter and ozone because
these two pollutants have the strong epidemiological evidence suggesting a variety of
serious adverse health impacts, including premature mortality and hospital admissions.
For other pollutants, the evidence is less persuasive of serious adverse health effects,
independent of the effect of other air pollutants.

>How do I generate a population dataset for a new grid definition?

You can generate a population dataset using a variety of approaches. The key is that
you need to have a shapefile of your area of interest (e.g., census tracts in a city) and
you have census data matching your area of interest. One source for both a shapefile
and the associated population data is the U.S. Census Bureau. (A variety of other
agencies have census data, and you need to check around for your area of interest.)
Another option for U.S. population data is to use the PopGrid software application,
described in appendix on Population Data for U.S. Setup. Using PopGrid, you still need
to have a shapefile for your area of interest.

>Can I edit a population configuration?

No, you cannot edit a population configuration. You can only view a population
configuration. If the population configuration does not match your data, then you need
to either create a new population configuration to match your data, or reshape your data
so that it matches the population configuration.

123

September 2008

-------
Air Quality Grid Creation

CHAPTER 5

Air Quality Grid

Creation

In this chapter...

>Create air quality grids using different methods.

>Find details on the file structures for data inputs.

Hnterpolate with Closest Monitor or Voronoi Neighbor
Averaging.

>Scale monitor data with modeling data.

>Learn about the Monitor Rollback feature.

124

September 2008

-------
Air Quality Grid Creation

BenMAP creates air quality grids to estimate the average exposure to ambient air pollution
of people living in some regularly-shaped area, or domain, such as that delineated by air
quality models, as well as more irregular shapes, such as cities or nations. However,
BenMAP does not estimate personal exposure. Instead, the air quality grids provide the
average population exposure for each grid cell that BenMAP can then use in health impact
functions.

To create air quality grids, BenMAP uses a number of inputs, including modeling data,
monitoring data, or both. You may enter your own modeling and monitoring data,
provided that the data are in a format recognized by BenMAP.

To start the grid creation process, click on the Create Air Quality Grids button. BenMAP
will then ask which of the following types of air quality data you wish to use:

>Model Direct. Choose this option if you have air quality modeling data that you wish to
use directly. Table 5-1 below describes the input format for modeling data.

>Monitor Direct. Choose this option if you wish to use monitoring data, without
additional modeling data.

>Monitor and Model Relative. Choose this option if you wish to use a combination of
monitor and model data.

>Monitor Rollback. Choose this option if you want to reduce monitor levels by a
specified amount.

$ Air Quality Grid C... pjn_|[x_

Choose Grid Creation Method
(* IModei Direclj
C Monitor Direct
C Monitor and Model Relative
C Monitor Rollback

Cancel Go!

Select your option and then click Go!. BenMAP will direct you through the necessary
steps for each option.

125

September 2008

-------
Air Quality Grid Creation

5.1 Model Direct

After choosing the Model Direct option, you need to specify the grid definition to which
the data file corresponds (Grid Type), the pollutant that the data is modeling (Pollutant),
and the location of your data file (Model Database).

$ Model Direct Settings

Grid Type: |T

Pollutant:

Generic Model Databases US Model Data Files |
Model Database:

Browse

Map

Cancel

~ K

The Model File specifies the location of the air quality model results that you want to
import. Table 5-1 presents the structure that these files must have, and Table 5-2 presents
a sample data file with a variety of metrics. (For more information on air quality models,
the EPA website has detailed descriptions of a variety of models:
http://www.epa.gov/ttn/scram/.)

Table 5-1. Air Modeling Data File Variables

Variable

Type

Required

Notes

126

September 2008

-------
Air Quality Grid Creation

Column integer yes

Row integer yes

Metric text no

Seasonal Metric text no

Annual Metric text no

Values comma yes

separated
values (text)

The column and the row uniquely identify each set of
modeling values, and link the modeling data with cells in a
Grid Definition.

This variable is either blank (signifying that the Values are
Observations, rather than Metric values), or must reference
an already defined Metric (e.g., 1-hour daily maximum) for
the appropriate Pollutant.

This variable is either blank (signifying that the Values are
not Seasonal Metric values) or must reference an already
defined Seasonal Metric for the Metric (e.g., mean of the
1-hour maximum for the months of June through August).

This variable is either blank (signifying that the values are
not an Annual Metric) or must be one of: None, Mean,
Median, Max, Min, Sum. (e.g., mean of the 1-hour
maximum for the year)

If Metric is blank, either 365 or 8,760 comma separated
observations. If Metric is defined, but Seasonal Metric and
Annual Metric are blank, 365 metric values. If Seasonal
Metric is defined, but Annual Metric is blank, n seasonal
metric values. If Annual Metric is defined, one annual
metric value (for either the Metric (if Seasonal Metric is
blank) or the Seasonal Metric. Missing values are signified
with a period ('.').

127

September 2008

-------
Air Quality Grid Creation

Table 5-2. Sample Air Modeling Data File

Column

Row

Metric

Seasonal Metric

Annual Metric

Values

101

5.90,.,.,H,,

102

19.80,.,.,3.5,.,.,14.70,.,.,.,.,.,4.9[

103

11.5,.,.,1.5,.,.,10.20,.,.,2.5,.,.,2.e

104

12.60,.,.,4.60,.,.,7.30,.,.,.,.,.,9.2[

105

Daily Average

5.90,.,.,H,,

106

Daily Average

19.80,.,.,3.5,.,.,14.70,.,.,.,.,.,4.9[

107

Daily Average

11.5,.,.,1.5,.,.,10.20,.,.,2.5,.,.,2.e

108

Daily Average

12.60,.,.,4.60,.,.,7.30,.,.,.,.,.,9.2[

109

Daily Average

Quarterly Average

11.82,13.24,18.79,14.25

110

Daily Average

QuarterlyAverage

12.31,13.89,18.27,21.19

101

Daily Average

Quarterly Average

12.33,15.68,18.49,15.96

102

Daily Average

Quarterly Average

13.52,13.69,19.04,22.43

103

Daily Average

Mean

14.52

104

Daily Average

Mean

16.41

105

Daily Average

Mean

15.62

106

Daily Average

Mean

17.17

107

Daily Average

Quarterly Average

Mean

14.29

108

Daily Average

Quarterly Average

Mean

18.97

109

Daily Average

Quarterly Average

Mean

20.14

110

Daily Average

Quarterly Average

Mean

16.46

5.2 Monitor Direct

Using the Monitor Direct grid creation option, you create an air quality grid directly from
air pollution monitoring data. At the top left of the Monitor Direct Settings window, you
are asked to select a previously defined grid definition from the Grid Type drop-down
list, and a previously-defined pollutant from the Pollutant drop-down list. And at the top
right of the Monitor Direct Settings window, you are asked to select an interpolation
method. The interpolation method is used to move from point-based monitor data to grid
cell based air quality data. That is, some grid cells will have many monitors in them, some
will have just one, and some will have none. BenMAP uses the interpolation methods to
generate representative air quality metric values for each grid cell from monitor data for all
of these cases.

128

September 2008

-------
Air Quality Grid Creation

w Monitor Direct Settings

~0B

Grid Type:

Interpolation Method

C Closest Monitor

Pollutant:

* Voronoi Neighborhood Averaging

C Fixed Radius (km.) I

Library | Database, Columns | Database. Rows | Text File

Monitor DataSet:

Monitor Library Year:

I
I

Advanced

Cancel

Map

Go!

BenMAP includes three interpolation methods. The Closest Monitor method simply uses
the monitor closest to a grid cell's center as its representative value. The Voronoi
Neighbor Averaging method takes an inverse-distance weighted average of a set of the
monitors surrounding a grid cell's center as its representative value. The Fixed Radius
method averages all of the monitors within a fixed radius. Each method is described
below. For more detail, also see the appendix on Air Pollution Exposure Estimation
Algorithms.

In the middle of the Monitor Direct Settings window, you can choose the rest of the
options to create your air quality grid. You need to specify the Grid Type and the
Pollutant. The choices for Grid Type include only those that you have previously
defined for the particular Setup in which you are working. Similarly, the choices for
Pollutant include only those that you have already defined in your Setup. And you need
to specify a source for your monitor data. In the Library tab you may select data that you
have already loaded into BenMAP.

You can also input your own monitor data file rather than use monitor data already
existing in a dataset, so long as it is in the format that BenMAP recognizes (see the Load
Data chapter for details on the format for monitor datasets). Click on the tab that matches
one of the three possible formats for your data: (1) in a database file, with monitor
definition information and monitor values in a single row (Database, Rows), (2) in a text
file, with monitor definition information and monitor values in a single line (Text File),
and (3) in two database files/tables, with monitor definition information in rows in one
file/table and monitor values in a single column in the other file/table (Database,
Columns). There is no preferred approach; you should use the approach with which you

129

September 2008

-------
Air Quality Grid Creation

are most comfortable.

5.2.1 Closest Monitor for Monitor Direct

If you choose the Closest Monitor option, BenMAP identifies the monitor closest to each
grid cell's center, and then assigns that monitor's data to the grid cell. Closest Monitor
interpolation can be modified by clicking on the Advanced button at the bottom of the
window and typing in & Maximum Neighbor Distance (in km).

$ Advanced Options fj[n][x]

Maximum Neighbor Distance (in km):

- No Maximum Distance -

Maximum Relative Neighbor Distance:

- No Maximum Relative Distance -

Weighting Approach
(* ilnverse Distance!

C Inverse Distance Squared

|~ Get Closest if None within Radius

Custom Monitor Filtering
Cancel OK

The Maximum Neighbor Distance specifies the maximum distance (measured in kilometers) that a
monitor may be from a grid cell (distances are calculated using the grid cell centroid). Cells
without any monitors within this distance will not be included in the resultant air quality grid. The
default setting is infinite (i.e. no limit to the distance between a monitor and a grid cell).

Note: Both the Maximum Relative Neighbor Distance and the Weighting Approach options are
irrelevant, since BenMAP is only choosing a single monitor to assign to any given grid cell.

5.2.2 Voronoi Neighbor Averaging (VNA) for Monitor Direct

If you choose the VNA option, BenMAP first identifies the set of monitors that "surround"
each grid cell's center (these monitors are referred to as the grid cell's neighbors), and
then BenMAP calculates an inverse-distance weighted average of these neighboring
monitors. In this section, we provide some examples of the different ways that BenMAP
calculates the average of the neighbor monitors. See the appendix on Air Pollution
Exposure Estimation Algorithms, for an expanded discussion of VNA, including how the
VNA algorithm actually chooses the neighbor monitors, as well as the different ways that
it may be used.

VNA interpolation has three advanced interpolation options, which can be modified by
clicking on the Advanced button at the bottom of the window:

130 September 2008

-------
Air Quality Grid Creation

> Maximum Neighbor Distance (in km) specifies the maximum distance that a monitor may be
from a grid cell, and still be included in the set of neighbor monitors used to calculate air
pollution exposure at a particular grid cell. The default setting is infinite (i.e., no limit to the
distance between a monitor and the grid cell).

> Maximum Relative Neighbor Distance specifies the maximum ratio for the distance of each
monitor to the distance of the closest monitor. The default setting is infinite.

> Weighting Approach specifies whether BenMAP should use inverse-distance weighting for the
monitors, or inverse-distance-squared weighting of the monitors. The default setting is
inverse-distance weighting.

The following examples illustrate how varying these options affects the final average

concentration estimate.

Example 1: Monitor Direct VNA method
Default options

Consider the following example at an hypothetical rural grid cell, where there are
relatively few monitors, and where the distance from a monitor to the grid cell can be
fairly large. With VNA, BenMAP first identifies the set of "neighbor" monitors for each
grid cell. The number of neighbors is usually in the range of about three to eight. In this
case, assume that there are five monitors at distances of 25, 50, 100, 200, and 400 miles
from the grid cell, with annual PM2.5 levels of 8, 13, 12, 18, and 15 |ag/m3, respectively.
BenMAP would calculate an inverse-distance weighted average of the monitor values as
follows:

11111

•8+13+ 12+ • 18+ 15

25 50 100 200 400

PM25 average^ = 10.68

25+ 50+ 100+ 200+ 400

Example 2: Monitor Direct VNA method

Maximum Neighbor Distance = 75

Using the same example that we used above, let us say you have specified a Maximum
Neighbor Distance of 75 miles, and left unchanged the default options (infinite value) for
Maximum Relative Neighbor Distance. BenMAP would only consider the first two
monitors, and would calculate an inverse-distance weighted average of the monitor values
as follows:

131

September 2008

-------
Air Quality Grid Creation

1 1
•8+13
r>j ^ 25 50 n

i M2 5 average^ - - = 9.67

25+ 50

Example 3: Monitor Direct VNA method

Maximum Relative Distance = 10

Alternatively, if you have specified that the Maximum Neighbor Distance is infinite, but
the Maximum Relative Neighbor Distance should have a value of, say, 10, then
BenMAP would calculate the ratio of the distance for each monitor to distance of the
closest monitor. In this case, the ratios would be 1 (=25/25), 2 (=50/25), 4 (=100/25), 8
(=200/25), and 16 (=400/25), and BenMAP would drop the monitor with a ratio of 16.
BenMAP would then calculate an inverse-distance weighted average of the monitor values
as follows:

111 1

•8+13+ 12+ 18

25 50 100 200

rM25 average^ - - - - = 10.53

25+ 50+ 100+ 200

Example 4: Monitor Direct VNA method

Inverse-distance squared neighbor scaling

In addition, you can specify the an inverse-distance-squared weighting of the monitors.
Let us say that you have left unchanged the defaults (infinite values) for Maximum
Neighbor Distance and Maximum Relative Neighbor Distance, and specified that the
Weighting Approach is inverse-distance-squared. BenMAP would then calculate an
inverse-distance-squared weighted average of the monitor values as follows:

1111 1

-8+ 13+ 12+ 18+ 15

625 2,500 10,000 40,000 160,000
PM2 5 average^ - - - - - = 9.26

625 + 2,500 + 10,000 + 40,000 + 160,000

Example 5: Monitor Direct VNA method

Maximum Neighbor Distance = 75
Maximum Relative Neighbor Distance = 10
Inverse-distance squared Weighting Approach

132

September 2008

-------
Air Quality Grid Creation

Finally, you could specify changes to all three options: a Maximum Neighbor Distance
of 75 miles, a Maximum Relative Neighbor Distance of 10, and a Weighting Approach

of inverse-distance-squared weighting. BenMAP would then calculate the following
average:

1 1
•8+ 13

625 2,500

PM25 average^ = 9.00

- +

625 2,500

5.2.3 Fixed Radius for Monitor Direct

The Fixed Radius method averages all of the monitors within a fixed radius (measured in
kilometers) that you specify. The the way that the monitors are averaged depends on the
Weighting Approach that you choose after clicking the Advanced button. You can
choose either Inverse Distance or Inverse Distance Squared weighting.

Note that the default option with the Fixed Radius approach is that BenMAP will not
calculate an average for a grid cell if there are no monitors within the fixed radius
(distance) that you specify. In the Advanced Options window, if you click Get Closest if
None within Radius, then BenMAP will calculate an average for those grid cells without
any monitors within the fixed radius, by reaching out and choosing the nearest monitor
regardless of distance.

5.2.4 Custom Monitor Filtering for EPA Standard Monitors

The Custom Filtering option only applies to EPA Standard Monitors library. This option
allows you to filter, map, and export your monitor data. You can reach the Custom
Filtering option by first choosing your pollutant, data source (e.g., monitor library) and
year.

133

September 2008

-------
Air Quality Grid Creation

w Monitor Direct Settings

~(1®

Grid Type:

Interpolation Method

|CAMX

C~ Closest Monitor

Pollutant:

(* Voronoi Neighborhood Averaging

PM2.5

» Fixed Radius (km.) I

Library | Database, Columns | Database, Rows ] Text File

Monitor DataSet:

EPA Standard Monitors

Monitor Library Year:

2000

Advanced

Cancel

Map

Go!

Click the Advanced button. This will take you to the Advanced Options window.

v Advanced Options

Maximum Neighbor Distance (in km):

Maximum Relative Neighbor Distance:
- No Maximum Relative Distance -

r Use original (slow) VNA algorithm

P Get Closest if None within Radius

Weighting Approach
* Inverse Distance
r Inverse Distance Squared
C Area Reduction
C Voronoi Edge Length
C Mean Value Coordinate

Custom Monitor Filtering

Cancel

~ K

Click the Custom Monitor Filtering button. This will take you to the Filter Monitors

window. Note that the first six options are essentially the same for each pollutant, and the
seventh option depends on the pollutant.

r 1. Include specific monitors. Here you can specify particular monitor IDs that you
want to include in your analysis. If this is left blank, then Ben MAP will include all
monitors that meet the rest of the selection criteria.

134

September 2008

-------
Air Quality Grid Creation

> 2. Exclude specific monitors. Here you can exclude any particular monitor IDs from
your analysis. Here again, if this option is left blank then BenMAP will include all
monitors that meet the rest of the selection criteria.

> 3. Restrict to particular states and/or latitude/longitude. You can choose monitors to
include from particular states, by listing the two-character state abbreviation (e.g., CA =
California). You can also choose monitors within particular latitude and longitude
ranges. The default values for latitude (20 to 50) and longitude (-130 to -65) completely
includes the continental U.S. Here again, if this option is left blank then BenMAP will
include all monitors that meet the rest of the selection criteria.

> 4. POC code. The Maximum POC specifies the highest POC value allowed in the data.
The default is a value of 4. And to choose one monitor when more than one monitor is
in the same location, the POC Preference Order specifies the preferred ordering of POC
codes.

> 5. Methods. The Method codes listed depend on the pollutant. In the case of 03 and
PM10, all methods are checked by default. And in the case of PM2.5, only so-called
federal-reference methods (FRM) are chosen by default — specifically numbers 116
through 120.

> 6. Monitor Objectives. The default is to choose all monitors regardless of monitor
objective.

> 7. Parameters Specific to the Pollutant. The default options vary by pollutant. Below,
we have listed the windows that appear with PM2.5, PM10, and ozone.

PM10 Monitor Filter

The Number of Valid Observations Required Per Quarter specifies the number of days
of data needed. The default is to require 11 observation per quarter. The Data Types to
Use option specifies whether to use Local data (parameter code 85101), Standard data
(parameter code 81102), or Both. The default for PM10 is to use both local and standard
data. When standard and local data are available at the same monitor location, the
Preferred Type allows you to choose which to use — the default is Local. The Output
Type option is designed to allow you to make the data reasonably consistent when both
local and standard data are used. The default is to use the local output type, so standard
data will converted to local.

135

September 2008

-------
Air Quality Grid Creation

v Filter Monitors

1. If you wish to include individual monitors by ID, enter the IDs
here., separated by commas. If you do not enter any
monitors here, all monitors which meet the rest of the
selection criteria will be included:

2. If you wish to exclude individual monitors by ID, enter the IDs
here, separated by commas:

If you wish to restrict monitors to certain states or areas, enter
comma separated state abbreviations, and minimum and/or
maximum latitudes and longitudes:

States:

Minimum Longitude: -130 Minimum Latitude: 20
Maximum Longitude: [-65 Maximum Latitude: [55

4. Select maximum PQC code to include and the POC
preference:

Maximum POC: 4 % POC Preference Order: 1,2,3,4

5. Select the Methods you wish to include:

0 062
0 063
v 064
0 065
0 071

073
076
079
081
098

0 125
0 126
v 127
0 130
~ 785

G. Select the Monitor Objectives you wish to include:

v GENERAL/BACKGROUND
0 HIGHEST CONCENTRATION
0 MAX OZONE CONCENTRATION

7. Select the parameters specific to the pollutant:

Number of Valid Observations Required Per Quarter:

FT \

Data Types To Use:

C Local C Standard (• Both

Preferred Type:

(J" Local C Standard

Output Type:

(* Local C Standard

Export

Map

Go!

Cancel OK

PM2.5 Monitor Filter

The default for PM2.5 is to also require 11 observations per quarter. A key difference with
PM10, is that only local PM2.5 data is used by default.

136

September 2008

-------
Air Quality Grid Creation

w Filter Monitors

1. ! f you wish to include individual monitors by ID, enter the! D s
here., separated by commas. If you do not enter any
monitors here, all monitors which meet the rest of the
selection criteria will be included:

2. If you wish to exclude individual monitors by ID, enter the IDs
here, separated by commas:

If you wish to restrict monitors to certain states or areas, enter
comma separated state abbreviations, and minimum and/or
maximum latitudes and longitudes:

States:

Minimum Longitude: -130 Minimum Latitude: [20
Maximum Longitude: -65 Maximum Latitude: [55

4. Select maximum PQC code to include and the PQC
preference:

Maximum POD [4 %] POC Preference Order: |1,2,3,4

5. Select the Methods you wish to include:

0 116

0 119

~ 702

0 117

0 120

~ 704

0 118

701

~ 711

ijX|

6. Select the Monitor Objectives you wish to include:

0 EXTREME DOWNWIND

GENERAL/BACKGROUND

0 HIGHEST CONCENTRATION

m ir ~-7l)

7. Select the parameters specific to the pollutant:

Number of Valid Observations Required Per Quarter:

Data Types To Use:

(• Local C Standard Both

Preferred Type:

(* Local C Standard

Output Type:

(* Local r Standard

Export

Map

Go!

Cancel

Ozone Monitor Filter

The ozone specific options differ from PM10 and PM2.5 in large part because ozone is an
hourly pollutant. The Number of Valid Hours specifies the number of hours needed for a
particular day of monitor to be considered "valid." BenMAP counts the number of non-
missing hourly values from the Start Hour through the End Hour and compare this
number with that specified in the Number of Valid Hours.

The Percent of Valid Days specifies the percent of days between the Start Date and the
End Date that need to be valid for the monitor itself to be considered valid. The default is
50 percent of the days between May 1 and September 30.

137

September 2008

-------
Air Quality Grid Creation

w Filter Monitors

2. If you wish to exclude individual monitors by ID, enter the IDs
here, separated by commas:

If you wish to restrict monitors to certain states or areas, enter
comma separated state abbreviations, and minimum and/or
maximum latitudes and longitudes:

States:

Minimum Longitude: -130 Minimum Latitude: [20
Maximum Longitude: -65 Maximum Latitude: [55

4. Select maximum PQC code to include and the PQC
preference:

Maximum POD [4 POC Preference Order: |1,2,3,4

5. Select the Methods you wish to include:

6. Select the Monitor Objectives you wish to include:

v EXTREME DOWNWIND

~ GENERAL/BACKGROUND

0 HIGHEST CONCENTRATION

< mi -M

k®

7. Select the parameters specific to the pollutant:

Start Hour:

^4 N umber of Valid H ours:

12 9 H

Percent of Valid Days:
End Date: [September 30 f 50 %

End Hour: 19
Start Date:

May 01 -H

Export

Map

Go!

Cancel

5.3 Monitor and Model Relative

Th & Monitor and Model Relative option lets you scale interpolated monitor values with
model data. As with the Monitor Direct option, you can choose Closest Monitor or
Voronoi Neighbor Averaging (VNA). In addition, you can scale the monitor values with
three different approaches: Spatial Only, Temporal Only, and Spatial and Temporal. As
discussed below, these approaches let you combine the advantage of the actual monitor
observations with the information provided by the models.

Monitor and Model Relative air quality grid creation is exactly the same as Monitor Direct
air quality grid creation with the exception of scaling. The concept of scaling is to use
modeling data to improve interpolation and/or forecast future air quality trends.

138

September 2008

-------
Air Quality Grid Creation

$ Monitor Model Relative Settings

Grid Type:

Interpolation Method
C Closest Monitor

Pollutant:

(* Voronoi Neighborhood Averaging!

r Fixed Radius (km.) T

Select Scaling Type:
C Spatial Only

C T emporal 0 nly

(* Spatial and Temporal

Library | Database, Columns | Database, Rows | Text File
Monitor DataSet:

Monitor Library Year:

Generic Model Data File |
Base Year Model File:

Browse

Text (Comma delimited). Excel Access or Dbf file.

The column headings are Column, Row, Metric, Seasonal Metric, Statistic, Values.

Generic Model Data File |

Future Year Model File:

Browse

Text (Comma delimited). Excel, Access or Dbf file.

The column headings are Column, Flow, Metric, Seasonal Metric, Statistic, Values.

Advanced

Map

Cancel

Go!

Identify and load your modeling data into BenMAP using the Base Year Model File box.
If temporal scaling is used, load your modeling forecasts using the Future Year Model
File box. The base year file should be created with modeling data which closely reflects
the historical conditions of the monitor data to be used. Typically multiple future year
files will be used to create multiple air quality grids for use in an analysis. That is, one
future year file might represent a future projection of current trends, while another might
represent the results of implementing a regulatory program.

5.3.1 Spatial Scaling

Spatial scaling involves only a Base Year Model File, and scales the concentrations of

139

September 2008

-------
Air Quality Grid Creation

each neighboring monitor by the ratio of the modeled concentration at the grid cell to the
modeled concentration at the grid cell containing the monitor. This approach takes into
account what the air quality modeling reveals about spatial heterogeneity in pollution
levels. For example, if the monitors are in relatively polluted urban areas, and the grid cell
is in a relatively unpolluted rural area, then the scaling will result in multiplying the
monitor values with ratios less than one, and thus produce lower values at the rural grid
cell than would be estimated with interpolation of the unsealed monitor data.

Spatial scaling, then, is useful because, while monitors provide invaluable information
about historical conditions, there are only a limited number of monitors. Many areas,
particularly rural areas, do not have close monitors. Model data can provide additional
information that improves the interpolated concentration estimates, and provides a more
accurate picture of air quality.

5.3.2 Temporal Scaling

Temporal scaling involves both a Base Year Model File and & Future Year Model File, and
scales the concentrations of each neighboring monitor by the ratio of the modeled
concentration at the grid cell containing the monitor in the future year to the modeled
concentration at the grid cell containing the monitor in the base year. This approach takes
into account what the air quality modeling reveals about the changes in pollution levels
over time at the monitor sites. For example, if the modeling forecasts that in the future,
pollution levels will decrease, then the scaling will result in multiplying the monitor values
with a ratio less than one, and thus produce lower forecasts at the grid cell than would be
obtained with the unsealed monitor data.

Temporal scaling, then, is useful because monitors cannot provide any information about
future conditions. Model data can provide this information, which can then be used to
project future monitor concentrations.

5.3.3 Spatial and Temporal Scaling

Using both spatial and temporal scaling involves both a Base Year Model File and a
Future Year Model File, and is simply a combination of spatial scaling and temporal
scaling, except that the future year data is used for the spatial scaling. That is, it scales the
concentrations of each neighboring monitor first by the ratio of the modeled concentration
at the grid cell in the future year to the modeled concentration at the grid cell containing
the monitor in the future year (spatial scaling), and then by the ratio of the modeled
concentration at the grid cell containing the monitor in the future year to the modeled
concentration at the grid cell containing the monitor in the base year (temporal scaling).
Notice, however, that the two future year concentrations at the grid cell containing the
monitor cancel out, allowing the ratio used to be simply the modeled concentration at the
grid cell in the future year to the modeled concentration at the grid cell containing the
monitor in the base year.

Using both spatial and temporal scaling gives the benefits of both approaches - it both
improves the interpolated estimates of air quality, and provides a future forecast of air
quality.

140

September 2008

-------
Air Quality Grid Creation

5.3.4 Examples

The first example presented is a. Monitor Direct example, which will provide a foundation
for the following Monitor and Model Relative examples. Recall that Monitor and Model
Relative air quality grid creation is exactly the same as Monitor Direct grid creation with
the exception of scaling. For additional examples with more detail, see Appendix C.

Example 1: Monitor Direct VNA Method
Default options

Consider the example at an hypothetical rural grid cell, from Section 5.2.2, where there are
five monitors at a distance of 25, 50, 100, 200, and 400 miles from the grid cell. Further,
let us say that the monitors have annual PM2.5 levels of 8, 13, 12, 18, and 15 (ig/m3.
Without any model-based scaling, BenMAP would calculate an inverse-distance weighted
average of the monitor values as follows:

llll 1

— S + 13 + 12+ 1S + 15

_ 25 50 100 200 400
PM^ average - j j j j j 10.6S

25 + 50 + 100 + 200 + 400

Example 2: Monitor and Model Relative VNA Method
Default options
Spatial Only scaling

Assume the same monitors and monitor concentrations as above. Additionally, assume
that the grid cell model value in the base-year is 6 |ig/m3, and the model values at the
monitors in the base-year are: 10, 14, 11, 17, and 15 |ig/m3. (This modeling suggests that
the grid cell is a less-polluted area than the area around the monitors.) Incorporating
Spatial Only scaling, BenMAP would calculate an inverse-distance weighted average of
the monitor values using the same approach as before, with the difference being that the
monitor values are scaled with the modeling values:

- fs -)~ — (l3 -]~ — fl2 -U — [is -]~ — fl5 -1
21 ^ 10J 51)1 Uj 100 ^ llj 200 ^ 17J 400 ^ 15j
PM^ average - j j j j

25+50 + 100 + 200+400

Example 3: Monitor and Model Relative Closest Monitor Method

141

September 2008

-------
Air Quality Grid Creation

Default options
Spatial Only scaling

Assume the same monitors and monitor concentrations as above. Additionally, assume the
same model values as above. The Closest Monitor interpolation of these same values
using Spatial Only scaling would be calculated as follows:

PM 25 average - S • — - 4.5

Example 4: Monitor and Model Relative VNA Method

Default options

Temporal Only scaling

Again, assume the same monitors and monitor concentrations as above. Additionally,
assume that the model values at the monitors in the future-year are: 8, 11, 9, 14, and 11 |i
g/m3, and that base-year model values at the monitors are: 10, 14, 11, 17, and 15 |ig/m3.
(This modeling suggests that the future-year model values are generally lower.)
Incorporating the Temporal Only scaling, BenMAP would calculate an inverse-distance
weighted average of the monitor values as follows:

— (8 —]~ — (,3 —]» — (l2 — (iS —]~ — (l5 —]

25 \ 10J 50 (u) 100 ^ llj 200 ^ 17 400 15 .
PM^ average - j j Sj2

25+ 50 + 100 + 200 + 400

Example 5: Monitor and Model Relative VNA Method

Default options

Spatial and Temporal scaling

Again, assume the same monitors and monitor concentrations as above. Additionally,
assume that the future-year model value at the grid cell is 4 |ig/m3, and that base-year
model values at the monitors are: 10, 14, 11, 17, and 15 |ig/m3. (This modeling suggests
that the future-year model value at the grid cell is significantly lower than the base-year
model values at the monitor.) Incorporating the Spatial and Temporal scaling, BenMAP
would calculate an inverse-distance weighted average of the monitor values as follows:

142

September 2008

-------
Air Quality Grid Creation

— fs —| + f 13 1 + (l2 —1 + [lS —1 + f15 —]

25 [ 10J 50 I U) 100 [ llj 200 [ V) 400 { 15} „„
PMz: average - - 3JS

25 + 50 + 100 + 200 + 400

5.4 Monitor Rollback

The Monitor Rollback option allows you quickly test what the benefits would be from
reducing historical monitor levels. BenMAP has three methods to reduce, or "roll back,"
monitor data: percentage rollback, incremental rollback, or rollback to a standard. Each of
these methods are depiocted below. Note that with each of these methods you can use the
same two interpolation algorithms (closest monitor, and VNA) as you can use with
Monitor Direct and Monitor and Model Relative.

Percentage rollback reduces all monitor observations by the same percentage.

Mg/m:-

33% (percentage)

DAY

Incremental rollback reduces all observations by the same increment.

Mg/m:-

~j— 10 [jg/m3

(incremental)

DAY

Rollback to a standard lets you choose a standard, and then reduces monitor observations
so that they just meet the standard.

143

September 2008

-------
Air Quality Grid Creation

(jg/rn:-

60 |jg/m3 daily standard
(to standard)

DAY

To apply a monitor rollback, first click the Create Air Quality Grids button. On the Air
Quality Grid Creation Method window, choosq Monitor Rollback.

s Monitor Rollback Settings: (... [T][n]| x]

Pollutant:

I 3

Library | Database, Columns | Database, Rows | Text File |

Monitor DataSet:

I 3

Monitor Library Year:

I 3

Advanced

Rollback Grid Type:

Cancel

There are three steps to the Monitor Rollback method.

1. Monitor Rollback Settings: (1) Select Monitors. Choose the Pollutant for the
monitor data. If you use data from an existing dataset, then choose the Library tab, and
from the drop-down list choose the Monitor DataSet and the Monitor Library Year.

If you want to use your own data, then choose the format tab that matches the data that you
want to load: Database, Columns; Database, Rows; or Text File. The file should have
the monitor data format specified in Load Data chapter.

Then choose the Rollback Grid Type from the drop-down list. This allows you to
determine how detailed the rollback scenario may be. If the whole region (e.g., United

144

September 2008

-------
Air Quality Grid Creation

States) will have the same type of rollback then you may simply choose a grid outlining
the area of interest. If you are interested in different rollbacks within a region, then you
should choose a more finely detailed grid definition (e.g., states).

When you have finished making your choices, click Next.

3 Monitor Rollback Settings: (2) Select Rollback Regions and Settings (-J[ry|xJ

Rollback Regions

• GUQjQ

Add Region Delete Region |- Region to Delete - ~»] V Export After Rollback

Select All

Deselect All

Back

2. Monitor Rollback Settings: Select Rollback Regions and Settings. In this section,
you can specify the type of the rollback method(s) that you would like to use, as well as
the location of the monitors that you want to rollback.

Choosing the Add Region button brings up the three rollback methods: Percentage
Rollback, Incremental Rollback, and Rollback to a Standard.

145

September 2008

-------
Air Quality Grid Creation

$ Select Region R... |T][n][x]

Rollback Type

'¦*" iPercentage Rollback]
C Incremental Rollback
Rollback to a Standard

Cancel

After choosing the rollback type, you then need to specify the amount of the rollback and
the region to which you want to apply it. In this example, we specified a 10 percent
reduction, a background of 0 ppb, and applied it to all monitors in all states by clicking
the Select All. So, BenMAP will reduce all observations for all monitors in all states by
10 percent.

Note that just above the map of the states there are four buttons, typically seen in mapping
programs, that allow you to zoom in and zoom out, and to focus on the particular groups of
grid cells that interest you.

At any time you can change the grid cells that you have selected. This particular example
is quite simple, so we will use a more complicated example below. After choosing the

146

September 2008

-------
Air Quality Grid Creation

counties where you want to rollback monitors, click on the Next button. BenMAP will
then perform the rollback you specified on the monitors in the grid cells that you have
chosen.

3. Monitor Rollback Settings: Additional Grid Settings.

The third stage is similar to the Monitor Direct grid creation method. As in Monitor
Direct, you need to specify the Interpolation Method {Closest Monitor or VNA) and the
Grid Type. You may select a Scaling Method (None or Spatial Only). If you choose
spatial scaling, it works exactly as described in the Monitor and Model Relative section,
where the modeling data is used to provide information in those areas that are
unmonitored.

$ Monitor Rollback Settings: (... (T](n][xJ

Select Interpolation Method—

C Closest Monitor

(• Voronoi Neighborhood Averaging

C Fixed Radius (km.) \

Grid Type: 0555

Adjustment File: | Browse |

Make Baseline Grid (in addition to Control Grid).

Advanced | Map
Back Go!

Select Scaling Method
(* None

C Spatial Only

By checking the Make Baseline Grid (in addition to Control Grid) you may create a
baseline grid at the same time as the control grid. The baseline grid uses the same
parameters as the control grid, with the exception of the rollback. That is, the baseline
uses the same monitor data, interpolation method, scaling (if any), and the same grid type.
The two resultant grids can then serve as both baseline and control scenarios in later stages
of an analysis.

Note that there is an Advanced button that lets you select the Maximum Neighbor
Distance, Maximum Relative Neighbor Distance, and Weighting Approach. The specific
availability of advanced features depends on the interpolation method that you choose.
You may also click the Map button to view the inputs to the rollback grids that you are
creating, as well as to view the grids themselves. The chapter on GIS/Mapping discusses
the mapping of grids in more detail.

147

September 2008

-------
Air Quality Grid Creation

5.4.1 Example: Combining Three Rollback Approaches in Different Regions

BenMAP allows you to have different rollback approaches in different regions. In this
example, we combine the three rollback types: Percentage Rollback, Incremental Rollback
, and Rollback to a Standard. Start by clicking on the Create Air Quality Grids button,
and choosing Monitor Rollback. On the Monitor Rollback Settings: (1) Select Monitors
window, select your pollutant (Ozone) and year (2000), and click Next. On the next
window, click the Add Region button and enter 10 for the Percent. In the previous
example, we used the Select All button to include all states in the rollback region. In this
example we want to create three regions instead, so just click on the three western-most
states to add them to the region. The states you have added to the region will turn yellow,
as in the picture below.

s Monitor Rollback Settings: (2) Select Rollback Regions and Settings [T](n][x]

Rollback Regions

~ Region 1
Rollback Parameters-

Percent: 10

Background: 0.00

»|Qx|Qs|q|

Add Region Delete Region |~ Region to Delete -- V Export After Rollback

Select All

Deselect All

Back

To add states with a second type of rollback, click on the Add Region button, choose the
rollback type, and then click on the counties to include in this second region, which
BenMAP denotes as Region 2. In this example, we have chosen an Incremental Rollback
of 5 and a background of 20, and applied it to the 14 next western-most states.

The map now depicts two rollback regions. We can toggle back and forth between each
region by clicking on the legend on the left side of the map. Any counties that have not
yet been included in a region may be added to an existing region, or we may create one or
more regions for these remaining counties. Note that once counties have been included in
a rollback region, they cannot be included in a different rollback region. In our example,

148

September 2008

-------
Air Quality Grid Creation

the three western-most states are highlighted in gray.

3 Monitor Rollback Settings: (2) Select Rollback Regions and Settings {-

Rollback Regions

I I Region 2
Rollback Parameters-

Increment 5

Background: 20

Region 1
rRollback. Parameters-

Percent: 10

Background: 0.00

Add Region Delete Region |- Region to Delete - ~»] V Export After Rollback

Select All

Deselect All

Back

To add a third rollback type covering the rest of the counties, click again on the Add
Region button, and then choose the rollback type. However, instead of choosing
individually the states, simply click the Select All button. This will select all of the states
that are not yet included in a region, and these remaining counties will now become
Region 3.

In this third region, we have chosen a Rollback to a Standard, which involves two groups
of parameters - those associated with the Attainment Test, which determines whether a
monitor is in attainment (meets the standard), and those associated with the Rollback
Methods, which are used to bring out-of-attainment monitors into attainment.

The Attainment Test parameters are Metric, Seasonal Metric, Metric Statistic,
Ordinality, and Standard. A monitor is considered in attainment if the nth highest value
of a daily metric specified by Metric is at or below the value specified by Standard,
where n is the value specified by Ordinality. For example, if Metric is D8HourMax,
Ordinality is 4, and Standard is 85, a monitor will be considered in attainment if the
fourth highest value of the eight-hour daily maximum is at or below 85 ppb. In this step
BenMAP calculates the metric to be used to determine whether a monitor's values must be
rolled back and, if so, how much (e.g., if Metric is D8HourMax, BenMAP calculates the
8-hour daily maximum for each day at each monitor).

149

September 2008

-------
Air Quality Grid Creation

$ Monitor Rollback Settings: (2) Select Rollback Regions and Settings [-J|nJ[xJ

Rollback Regions

| Region 3

"Attainment Test-
Daily Metric:

|D8HourMax
Seasonal Metric:

Annual Statistic:

Ordinality:

Standard:

4th Highest value of DSHourMax <= 85

rRollback Methodsr-

Interday Rollback Method, Background:
| Percentage ^ |0.00

Intraday Rollback Method, Background:
|Percentage ^ |0.00

'I k"

] ¦

"H—

1 f

¦1 ^

L-A—5-

[ ¦1 1

/ ^ _ j?

Add Region Delete Region |-- Region to Delete -- ~[] I- Export After Rollback

Select All : Deselect All

Back

The Attainment Test can also be used for seasonal metrics (by choosing previously
defined seasonal metrics from the drop-down list below Seasonal Metric), as well as for
annual metrics (by using the drop-down list below Metric Statistic). For example, if you
want the annual mean ozone level to go stay below 60 ppb, then you would choose the
daily 24 hour mean (D24HourMecm) from the drop-down list below Metric, choose Mean
from the drop-down list below Metric Statistic, and set the Standard to 60. (Note that in
this case Ordinality cannot be chosen because there is only a single annual value.)

The Rollback Method parameters are Interday Rollback Method, Interday
Background Level, Intraday Rollback Method, and Intraday Background Level.

These four parameters determine the rollback procedures used to simulate
out-of-attainment monitors coming into attainment. The Interday Rollback Method and
Background Level are used to generate target values for the metric specified by the
Attainment Test. The Intraday Rollback Method and Background Level are used to
adjust hourly observations to meet the target metric values generated in the previous step.

BenMAP provides several types of Interday Rollback Methods (Percentage,
Incremental, Quadratic, and Peak Shaving) and several types of Intraday Rollback
Methods (.Percentage, Incremental, and Quadratic). The methods involved for each can
be somewhat complicated, so we have included a section in the Monitor Rollback
Algorithms appendix goes through several examples.

150

September 2008

-------
Air Quality Grid Creation

5.5 Questions Regarding Creating Air Quality Grids

>How can I generate a map of an air quality grid and export it?

You need to choose the GIS/Mapping option from the Tools drop-down menu available in the
upper left-hand corner of the main BenMAP window. The GIS/Mapping option, as described in
the GIS/Mapping chapter, provides the option to map and then save air quality grids.

r For the Rollback to a Standard option, why are there Interday and Intraday rollback
options?

The Interday Rollback Method option identifies the approach (e.g., Percentage) to reduce
daily air pollution levels, in order to meet the specified standard. (In other words, there is more
than one way to reduce daily pollution levels so as to meet the standard you have chosen, and
BenMAP lets you choose from among several approaches. The Background level associated
with the Interday Rollback specifies the bound, below which, BenMAP will not make
adjustments to daily levels.

-Rollback Methods

Interday Rollback Method, Background:

Percentage

0.00

Intraday Rollback Method, Background:

Percentage ~ |

0.00

The Intraday Rollback Method option is only relevant for hourly pollutants, like ozone. This
option specifies the approach (e.g., Percentage) used by BenMAP to reduce hourly air pollution
levels to reach is only relevant for hourly air pollution data. That is, once you have chosen the
approach to reduce daily air pollution levels, on any given day there is more than one way to
reduce the hourly air pollution values to meet the targeted pollution level for that day. The
Background level associated with the Interday Rollback species the bound, below which,
BenMAP will not make adjustments to hourly levels

The Interday and Intraday options are complicated. The appendix on Monitor Rollback
Algorithms explains these options in more detail and gives some numerical examples.

>Can I use air quality grids based on different Grid Types in the baseline and control
scenarios?

No. In any given analysis, you need to use the same Grid Type in the baseline and
control scenarios.

>Can I use air quality grids of the same Grid Type but based on different Grid Creation
Methods?

Yes. In any given analysis, you may use air quality grids made with different methods. Air
quality grids made with Model Direct, Monitor Direct, and Monitor and Model Relative may be
used interchangeably, if desired. Similarly, air quality grids made with different interpolation
methods may be compared. However, it generally is not recommended to create grids with
different methods and use them in the same analysis.

151

September 2008

-------
Air Quality Grid Creation

>Can I do I an analysis with multiple pollutants?

You can only estimate impacts one pollutant at a time, however, BenMAP allows you to
aggregate the results of more than one pollutant. This is discussed in the chapter on
Aggregation, Pooling, and Valuation.

> Why does it take so long to generate an Ozone Air Quality grid if there are a of grid cells?

It can take a long time to create an air quality grid because the file being generated can be quite
large. In some cases, air quality grids can be several hundred megabytes in size. (One reason
the ozone files are large is that the definition of ozone has, by default, four metrics. If you do
not need all of the default metrics for the health impact functions in your database, then delete
the unneeded metrics and BenMAP will run faster and will generate smaller air quality grids.
This is an advanced step, so do not do it if you are unsure.)

>How do I access data in an Air Quality grid?

You can access the data in an air quality grid by going to the Tools drop-down menu
and choosing the Export Air Quality Grid option. (The Tools drop-down menu is
available in the upper left-hand corner of the main BenMAP window.) You then locate
the air quality grid from which you want to export air quality data and then give a name
to your exported file. BenMAP will generate a text file that you can then examine.

This is discussed in detail in the Tools Menu chapter.

>How do I create Air Quality grids for a One-Step Analysis?

Follow the same steps outlined

152

September 2008

-------
Incidence Estimation

CHAPTER 6

Incidence
Estimation

In this chapter...

>Get an overview of health incidence estimation.

>Use the Create and Run Configuration button to specify
various options for calculating incidence results.

>Learn about baseline and control scenarios.

>Learn the difference between Point Mode and the Latin
Hypercube option.

>Select health impact functions.

>Run, save and reopen a configuration.

153

September 2008

-------
Incidence Estimation

A configuration is a record of the choices you make in estimating the change in adverse
health effects between a baseline and control scenario. The choices include the following:

>The air quality grids for the baseline and control scenarios;

>The year for the analysis;

>The threshold for the analysis;

> Whether the analysis will focus on a single "point" estimate (Point Mode), or a range of results
that mirror the variability in the inputs to the health impact functions (Latin Hypercube Points);
and,

>The health impact functions to be used in estimating adverse health effects.

Once these choices are made, they can be saved in a configuration file for future reuse.
BenMAP gives you flexibility in the creation, editing, and saving of configuration files.
You can open an already existing configuration, and proceed directly to the estimation of
incidence. Or, you can create a new one, and proceed with the incidence estimation. In
addition, you may save any edits made to existing or new configuration files. BenMAP
saves configuration files with a ".cfg" extension. After calculating the change in adverse
health effects, BenMAP saves the results in a "configuration results" file with a ".cfgr"
extension.

After clicking the Create and Run Configuration button, you will be asked if you want
to create a new configuration, or open an existing one, if you have already created one that
you would like to use again. Select the desired option and click Go!. Below, we discuss
the subsequent steps for each option.

$ Configuration Creati...

C Create New Configuration
C Open Existing Configuration

Go!

Cancel

6.1 Introduction to Estimating Health Incidence

Health impact functions relate a change in the concentration of a pollutant to a change in
the incidence of a health endpoint (i.e., premature mortality or work loss days). It is
typically derived from the estimated relationship between the concentration of a pollutant
and the adverse health effects suffered by a given population. This relationship is called
the concentration-response (C-R) function. For example, the concentration of the
pollutant may be particulate matter (PM10), and the population response may be daily
premature deaths. For the purposes of estimating benefits, we are not interested in the C-R
function itself, but rather the relationship between the change in concentration of the
pollutant, and the corresponding change in the population health response. We want to
know, for example, if the concentration of PM10 is reduced by 10 micrograms per meter

154

September 2008

-------
Incidence Estimation

cubed, how many premature deaths will be avoided?

To produce health incidence results, the first step is to calculate the change in pollution
concentrations that would be produced by the application of a given set of emissions
controls. The concentration change in a pollutant is the increment between the control
scenario and the baseline scenario. This increment and a gridded population dataset are
then used in health impact functions to calculate the change in health incidence that would
result from this change in pollution. These functions are based on epidemiological studies
and can be selected by the user. Typically, these health incidence results show the
decrease in health incidence (e.g. the decrease in asthma, bronchitis, mortality, etc) due to
a decrease in pollution.

In BenMAP, the selected health impact functions are stored in configurations, which can
be re-used over and over again.

6.2 Create Health Impact Configuration

If you choose to create a new configuration, there are two steps. First, fill out the
Configuration Settings form with the following options: the air quality grids for the
Baseline File and the Control File, the Population DataSet, Population Year,
Threshold, and whether BenMAP will run in Point Mode or use the Latin Hypercube
Points option. In the second step, specify the health impact functions that you want to use
in your analysis.

6.2.1 Configuration Settings

The Configuration Settings window opens after you select Create New Configuration
and click Go!. In this window, you first specify the air quality grids for the Baseline File
and Control File. The baseline file contains the air quality metrics for the scenario
assumed to occur without any change in policy. The control file specifies the air quality
metrics assuming that some type of policy or change has been implemented. The air
quality grids should be of the same pollutant, and should also be based on the same
grid-type. If you choose a particular grid type (e.g., County) for the baseline file, then the
same grid type must be used in the control file. Conversely, it would not possible to use
County grid-type in the baseline and a Tract grid type in the control file.

You may choose existing air quality grids, by clicking the Open button, and selecting an
air quality grid, which is designated with an ".aqg" extension. You may also create a new
air quality grid by clicking the New button. This will take you to window for Air Quality
Grid Creation Method, where you follow the same steps outlined in the chapter on air
quality grid creation.

The Pollutant specified in the air quality grids determines the suite of health impact
functions available for the configuration. Only functions associated with the specified
pollutant will be available for the configuration. Furthermore, if only certain Metrics
associated with the pollutant are present in one (or both) of the air quality grids (this see
the information on monitor and model data formats above for more information on how

155

September 2008

-------
Incidence Estimation

this can occur) only those health impact functions associated with those metrics will be
available.

In choosing the Population DataSet and Population Year, you specify the population
data that will be used in the health impact function. The values in the menu for the
Population Year depend on the range of values in your Population DataSet.

« Configuration Settings ['- I t I x

Cancel Previous Next

The Threshold indicates the minimum value that may be used in either the baseline or
control air quality metrics. That is, air quality metrics below the threshold will be replaced
with the threshold value. With a threshold of zero, there is no impact on the estimates
generated by the health impact functions. However, as the threshold increases, then it will
have a progressively larger impact on the incidence estimation. For most analyses, a
threshold of zero is appropriate, as there is little evidence suggesting the existence of a
threshold. Nevertheless, the Threshold option allows you to explore the impact of any
given threshold on the incidence estimation. This is also useful for scenarios where you
might want to know the incidence associated with changes in air quality occurring above a
standard.

You can also explore thresholds with health impact functions explicitly specified in the
functional form. For example, some PM2.5 functions in the default health impact function
database have thresholds built into them. You can identify these functions by checking the

156

September 2008

-------
Incidence Estimation

"Qualifier" variable field, where you will see the notation "Adjusted Coefficient." The
adjustment process is discussed in the Appendix describing the Particulate Matter Health
Impact Functions.

The Point Mode and Latin Hypercube Points options allow you to generate an average
incidence estimate, or a range of results that mirror the variability in the inputs to the
health impact functions. With the Point Mode option, BenMAP uses the mean values of
the inputs to the health impact functions, and generates a single "point estimate" of the
change in adverse health effects.

With the Latin Hypercube Points option, you can generate a number of estimates that
mirror the variability in the inputs to the health impact functions. The Latin Hypercube
Points option allows you to generate specific percentiles along the estimated incidence
distribution. For example, if you specify 20 points, then BenMAP will generate estimates
of the 2.5th percentile, 7.5th percentile, and so on, up through the 97.5th percentile. The
number of points suggested in the drop down menu varies between 10 and 100. In
addition, you can simply type in the desired number of points. The greater the number of
chosen points, the greater the time needed by BenMAP to process the results. The
relationship between the number of points and time needed is essentially linear, so a
doubling of the number of points would double the processing time.

If Point Mode is chosen, the number of Latin Hypercube Points cannot be modified and
will be ignored (treated as zero). However, with the Latin Hypercube Points option, a
point estimate will still be generated. As discussed in the chapter on Aggregation,
Pooling, and Valuation, by choosing the Point Mode, you have fewer pooling options.
You cannot conduct fixed/random effects pooling, nor any other procedure that depends
on knowing the distribution, or the range of variability of the incidence estimates.

6.2.2 Select Health Impact Functions

The second step in creating a configuration is to select the health impact functions. After
filling in both the Select Air Quality Grids and Settings of the Configuration Settings

window, click Next to select the health impact functions.

157

September 2008

-------
Incidence Estimation

$ Configuration Settings

Available CR Functions:

Tree

Data

DataSet

Endpoint Group

Endpoint

Metric

Seasonal Metric

Metric Statistic

Aui

0 BenMAP Health Impact Functions

_il

Selected CR Functions:

Function Identification

Function Parameters

DataSet

Endpoint

Race

Ethnicity

Gender

Start Age

End Age

Incidence DataSet

Prevalence Data...

Variable DataSet

-------
Incidence Estimation

$ Configuration Settings

Available CR Functions:

Tree

Data

DataSet

Endpoint Group

Endpoint

Metric

Seasonal Metric

Metric Statistic

B-

BenMAP Health Impact Functions

Mortality

Asthma Exacerbatio

Asthma Exacerbatior

—

D24HourMean

None

D24HourMean

None

D24HourMean

None

D24HourMean

None

Asthma Exacerbatior

Selected CR Functions:

Function Identification

Function Parameters

DataSet

Endpoint

Race

Ethnicity

Gender

Start Age

End Age

Incidence DataSet

Prevalence Data...

Variable DataSet

BenMAP He

Asthma E

Cancel

Run

If you want to delete some of the health impact functions that you added to your
configuration, just highlight the studies of interest and hit the Delete key on your
keyboard.

When you drag over a study, BenMAP displays the study's information in two broad
categories: Function Identification and Function Parameters. The Function
Identification includes information such as the Endpoint Group, Endpoint, Metric,
Location, and other variables. This identification information is useful when
distinguishing between multiple health impact functions. The Function Parameters
include those variables that you may directly edit: Race, Ethnicity, Gender, Start Age,
End Age, Incidence DataSet, Prevalence DataSet, and Variable DataSet..

You can drag over the same study multiple times, and then make edits, in order to be able
to calculate the impact of changes in these variables. To edit Start Age and End Age, just
highlight the appropriate cell and type in the desired age values. Keep in mind that these
age represent inclusive age bounds, so if you type in 10 and 12 this will include all
children ages ten, eleven, and twelve years old. If you want just a single age year, then
type the same year in both the Start Age and the End Age. Note that the accuracy of the
populations calculated for these age ranges will depend on the specificity of the population
data present in your selected population dataset.

159

September 2008

-------
Incidence Estimation

3 Configuration Settings

Available CR Functions:

T ree

Data

DataSet

Endpoint Group

Endpoint

Metric

Seasonal Metric

Metric Statistic

BenMAP Health Impact Functions

Mortality

Asthma Exacerbatio

Asthma Exacerbatior

—

D24HourMean

None

D24HourMean

None

D24HourMean

None

D24HourMean

None

lii-

Asthma Exacerbatior

Selected CR Functions:

Function Identification

1 Function Parameters

DataSet Endpoint

Race Ethnicity Gender Start Age

End Age

Incidence DataSet Prevalence Data... Variable DataSet

BenMAP He Asthma E

i >i

Cancel

Run

At the same time you can edit the Race, Ethnicity, and Gender variables, by clicking on
the appropriate cell, and then scrolling through the drop-down menu. Similarly, if you
have multiple datasets to choose from, you can change the source of the incidence and
prevalence data that is used for each function. Use the drop-down lists in the Incidence
DataSet, Prevalence DataSet, and Variable DataSet fields.

6.3 Run Health Impact Configuration

To begin the calculation of incidence for the health impact functions in the configuration,
you click the Run button on the bottom right-hand corner of the Configuration Window
that has the list of chosen health impact functions. After clicking Run, BenMAP allows
you to save the configuration, or begin the calculation. If you wish to save the
configuration for future use, click Save and specify a file with a ".cfg" extension.

160

September 2008

-------
Incidence Estimation

$ Save Configuration |T](n](x]

Ready to run configuration. If you wish to save
this configuration, click the Save button. When
ready, click OK. If you are not ready to run this
configuration, click Cancel.

\ Save

Cancel

When ready to generate incidence estimates, click on the OK button. BenMAP then
requires that you specify a file in which to save the results, with a ".cfgr" extension.

6.4 Open & Modify Existing Health Impact Configuration

To use the same settings as a previous BenMAP run, you can choose to open an existing
configuration. After clicking the Create and Run Configuration button, you simply
choose Open Configuration and click the Go! button. Once an existing configuration is
open,, you can do all the things with it that you would do with a newly created
configuration - modify settings, add and delete health impact functions, save it to a
configuration file, generate results, etc.

6.5 Questions Regarding Health Impact Configurations

> I got an error message the "specified Baseline file does not exist." What does that mean
and how do I fix it?

The path and/or the file that you have specified do not exist. Go back and ensure that
the path and file are correctly identified.

Error fx"

The specifiei

d Baseline file i

[ OK

does not exist.

> How do I know which health impact functions to use? Which functions does EPA use?

161

September 2008

-------
Incidence Estimation

One option regarding the choice of health impact functions is to work with someone,
say another BenMAP user, who is familiar with the epidemiological literature and
develop your own set of health impact functions. Reviewing the epidemiological
literature can be time-consuming, though in some situations, this might be the best
option, such as if you want to estimate the health impacts of carbon monoxide exposure,
for which BenMAP does not have pre-installed health impact functions.

Another option is to use the ozone and PM2.5 configurations used by EPA that come
installed with BenMAP. You will find them in the Configurations folder. If desired you
can edit this configuration and then save it under a different file name — it is always a
good idea to keep the original version, so you can go back to it if needed!

> How do I edit or add other health impact functions?

To edit or add health impact functions you need to go to Modify Setup option in the
Tools drop-down menu available in the upper left-hand corner of the main BenMAP
window. See the health impact function section in the chapter on Loading Data for
details on how to do this.

> How do I learn more about the population data in BenMAP?

The appendix on Population Data for U.S. Setup describes the population data in detail.

> Why did I not get results for a given geographic area that I wanted in my analysis?

Check to see if your air quality grids mapped properly.

> Why does BenMAP give me a range check error when I try to run the configuration step?

If you get a "range check" error, it usually means that you are using files that are
incompatible with your current version of BenMAP. Make sure that the files you are
currently using were generated with the version of BenMAP that you have installed. If
the problem persists, contact: benmap@epa.gov.

162

September 2008

-------
Aggregation, Pooling, and Valuation

CHAPTER 7

Aggregation,
Pooling, and
Valuation

In this chapter...

>Getan overview of valuation, discounting, and pooling.

>Use the Aggregation, Pooling, and Valuation (APV) button
to create an APV configuration.

>Sort and pool incidence results.

>Learn the differences between the pooling methods.

> Assign economic values to incidence results.

>Aggregate incidence results and valuations.

>Run, save and re-open an APV configuration.

163

September 2008

-------
Aggregation, Pooling, and Valuation

Once you have created a configuration results file with incidence results based on your
two air quality grids (using the Create and Run Configuration button), you can use the
Aggregation, Pooling, and Valuation button to combine the incidence results and place
an economic value on the combined results. You have two options.

> Create a New Configuration for Aggregation, Pooling, and Valuation. You can create a new
type of configuration, termed an Aggregation, Pooling, and Valuation (APV) Configuration.

This allows you to specify whether to aggregate incidence results at the county, state or national
level, or whether to leave them at the grid cell level. In addition, you can specify how you might
want to combine or "pool" the incidence results, using a variety of pooling options. Given the
aggregated and pooled incidence results, you then can specify how you might want to value
them - typically there are multiple valuations. These valuation results can then be further
aggregated, and these aggregated valuation results can be pooled. Having made all of your
selections, you may save this APV Configuration file ("apv") for future use, and then proceed to
calculating the results, which are stored in an APV Results file ("apvr).

>Open Existing Configuration for Aggregation, Pooling, and Valuation. You can load an
existing APV Configuration file, edit the configuration, save it with the same or a different
name, and then proceed to calculating the results.

7.1 Introduction to Valuation, Discounting, and Pooling

This section presents an introduction to valuation, discounting, and pooling. Valuation
generally refers to placing a dollar value on estimated health incidence. In BenMAP, you
might also think of valuation as including the assignment of quality adjust life year
(QALY) weights, in the sense that the placement of dollar values and QALY weights both
occur in the APV configuration. We discuss both dollar values and QALY weights here.
In addition, we provide a brief introduction to discounting, which has to do with placing
less weight on things occurring in the future than we do for things today. Finally, we
discuss pooling, which has to do with combining comparable results.

7.1.1 Overview of Valuation

Improvements in ambient air quality generally lower the risk of developing an adverse
health effect by a fairly small amount for a large population. A lower risk for everyone
means that fewer cases of the adverse health effect are expected, although we cannot
predict which people would be spared. Therefore, the health benefits conferred on
individuals by a reduction in pollution are actually reductions in the risk of having to
endure certain health problems. Monetizing the benefits of a reduction in air pollution
involves estimating society's willingness to pay (WTP) for these reductions in risk and is
typically referred to as valuation. BenMAP uses valuation functions to estimate the
benefits of reducing air pollution.

These benefits (reductions in risk) may not be the same for all individuals (and could be
zero for some individuals). Likewise, the WTP for a given benefit is likely to vary from
one individual to another. In theory, the total social value associated with the decrease in

164

September 2008

-------
Aggregation, Pooling, and Valuation

risk of a given health problem resulting from a given reduction in pollution concentrations
is generally taken to be the sum of everyone's WTP for the benefits they receive.

Epidemiological studies allow us to estimate the number of cases of an adverse health
effect that would be avoided by a given reduction in pollutant concentrations. If we have
an estimate of the average individual's WTP for the risk reduction conferred upon him, we
can derive from that an estimate of the value of a statistical case avoided. Suppose, for
example, that a given reduction in pollutant concentrations results in a decrease in
mortality risk of 1/10,000. Then for every 10,000 individuals, one individual would be
expected to die in the absence of the reduction in pollutant concentrations (who would not
be expected to die in the presence of the reduction in pollutant concentrations). If the
average individual's WTP for this 1/10,000 decrease in mortality risk is $100, then the
value of a statistical life is 10,000 x $100, or $1 million. In general, the ex ante WTP for a
risk reduction of x can be converted into an ex post value of a statistical case avoided by
dividing the average individual's WTP for the risk reduction of x by x (e.g. $100/0.0001 =
$1,000,000). The same type of calculation can produce values for statistical incidences of
other health endpoints.

Sometimes those values come from contingent valuation studies, in which study
participants are queried about their WTP to avoid a specific adverse health effect. When
WTP estimates are not available, it can be approximated by other measures, most notably
cost of illness measures.

An individual's WTP to avoid an adverse health effect will include, at a minimum, the
amount of money he would have to pay for medical expenses associated with the illness.
Because medical expenditures are to a significant extent shared by society, via medical
insurance, Medicare, etc., the medical expenditures actually incurred by the individual are
likely to be less than the total medical cost to society. The total value to society of an
individual's avoidance of an adverse health effect, then, might be thought of as having two
components: (1) the cost of the illness (COI) to society, including the total value of the
medical resources used (some portion of which will be paid by the individual), plus the
value of the lost productivity, as well as (2) the WTP of the individual, as well as that of
others, to avoid the pain and suffering resulting from the illness.

The COI approach attempts to estimate the total value of the medical resources used up as
well as the value of the individual's time lost as a result of the illness. Because this
method does not include the value of avoiding the pain and suffering resulting from the
illness (a potentially large component), it is generally believed to underestimate the total
value of avoiding the illness, perhaps substantially.

The contingent valuation method attempts to elicit from people what they would be willing
to pay to avoid the illness. Because of the distortion in the market for medical goods and
services, whereby individuals generally do not pay the full value of the medical care, this
method too is likely to understate the total value of avoiding the illness.

Although the COI and contingent valuation are the two most common methods, other
methods have been used in certain circumstances. The method the benefit analyst chooses
to value a particular health endpoint will depend in part on what is available. The unit

165

September 2008

-------
Aggregation, Pooling, and Valuation

values currently available for use in BenMAP are data or estimates that have been
collected or generated by researchers and can be readily obtained in publicly available
databases or in the open literature. When reviewing the economic literature to determine
the appropriate valuation functions to use, it is important to have an economist assist.

7.1.1.1 Valuing Reductions in Premature Mortality

The economics literature discussing the value of changes in fatality risks is extensive and
provides a basis for monetizing benefits when the number of deaths avoided as a result of
a regulatory action can be calculated, but the literature on certain issues regarding the
appropriate method for valuing reductions in premature mortality risk is still developing.
Issues such as the appropriate discount rate and whether there are factors, such as age or
the quality of life, that should be taken into consideration when estimating the value of
avoided premature mortality are still under discussion. BenMAP currently offers a variety
of options reflecting the uncertainty surrounding the unit value for premature mortality.
See the appendix on Economic Value of Health Effects for more detail on the valuation
functions available in BenMAP.

Monetary estimates of changes in fatality risk are often expressed in terms of the Value of
a Statistical Life (VSL). This term is easily misinterpreted and should be carefully
described when used in benefit analysis. In particular, VSL refers to the WTP for
reductions in the risk of premature death aggregated over the population experiencing the
risk reduction; that is, VSL refers to the sum of many small reductions in fatality risks.
The basic assumption underlying the VSL approach is that equal increments in fatality
risks are valued equally. For similar reasons, the VSL approach is only appropriate for
marginal changes in the risk of death and should not be used to value more significant
changes. Because changes in individual fatality risks resulting from environmental
regulation are typically very small, the VSL approach is usually acceptable for these types
of benefit analyses.

7.1.2 Overview of Discounting
What is discounting?

In general, people prefer current consumption to future consumption. In other words, a $1
today is worth more today than a $1 tomorrow is worth today, and that dollar continues to
decrease in value as you go further out into the future. (This concept is also referred to as
the social rate of time preference or the time value of money. This is a different concept
than inflation, which is a general increase in the price level of goods and services.)
Discounting is the process of converting that future dollar into a value that can be
compared to the value of a dollar today. The discount rate expresses this process in
quantitative terms. The higher the discount rate, the faster value decreases over time. For
example, $1 twenty years from now is worth $0.55 today at a 3% annual discount rate, but
worth only $0.26 at a 7% annual discount rate.

A basic discounting function is as follows:

Present Value = Future Value / (l+r)1

166

September 2008

-------
Aggregation, Pooling, and Valuation

where r is the discount rate and t is the time period (usually years).

Example: $1 twenty years from now at a 3% annual discount rate is worth $0.55 today

Present Value = $1.00/(1 + 0.03)20 = 1 /(1.03)20= 1 / 1.806111 = 0.553676 = $0.55

Why do we discount benefits?

The benefits of reductions in air pollution may need to be discounted for several reasons.

1. Today's society values benefits that occur today more than benefits in the future.
Therefore, we must discount in order to compare those future benefits with current
benefits.

2. For a cost-benefit analysis, benefits estimates need to be comparable to the cost
estimates. Discounting can be used to compare the future streams of benefits and costs.

3. BenMAP only estimates changes in adverse health effects based on changes in air
quality for one specified analysis year, even though certain health benefits may occur after
the analysis year. Discounting can be used to compare the future benefits with benefits
occurring during the analysis year.

Under which scenarios would I need to discount benefits?

Health benefits may occur three different ways after the analysis year specified in
BenMAP.

1. Certain health endpoints accrue medical expenses or lost earnings for multiple years.
The future medical expenses would need to be discounted to compare with expenses
occurring in the analysis year.

2. Pollution exposure and the resulting health effects do not occur within the same year
(a.k.a. a cession lag). The monetized benefits of future health effects would need to be
discounted to compare with the benefits of health effects that occur during the analysis
year.

3. In some analyses, you may want to estimate a stream of benefits occurring over
multiple analysis years instead of just one analysis year. In this scenario, you would need
to discount the future benefits occurring in each year analyzed back to the present year in
order to present the cumulative total estimate of benefits (i.e., the net present value of a
stream of benefits).

When would we not discount benefits?

In many instances, it is not necessary to discount the benefits estimates generated by
BenMAP. If the health effect and the monetized value of all the medical expenses, lost

167

September 2008

-------
Aggregation, Pooling, and Valuation

earnings, and suffering occur entirely in the analysis year, then you may not need to
discount your benefits. For example, school loss days occur within the analysis year, and
all monetized expenses occur within the analysis year. It is important that you understand
the assumptions within the health and valuation functions before you decide whether you
need to discount. (If your analysis year for your benefits estimates does not match the
analysis year for your costs estimates, you may need to discount in order to compare your
benefits with your costs even if you meet the criteria listed above.)

Which discount rate should I choose?

Selecting a discount rate is challenging and is one of the most contentious methodological
issues encountered in economic analyses of environmental policies. Because
environmental regulations frequently have differing streams of costs and benefits over
time, the selected discount rate may determine whether the benefits of a regulatory action
exceed the costs. In addition, selecting a higher discount rate may result in a smaller
benefits estimate because the future benefits are worth much less than they would be if a
lower discount rate was selected. For benefits that occur well into the future, the issue of
intergenerational equity further complicates the selection of the discount rate. (In the
context of environmental policy, intergenerational equity refers to the fairness of the
distribution of the costs and benefits of a long-lived policy when those costs and benefits
are borne by different generations. Most criteria pollutants are not considered to have
intergenerational equity issues, but the issue frequently arises in analyses of climate and
mercury.)

There are various economic arguments in support of and in opposition to various discount
rates. (For more details see EPA [1999; 2000] listed in the section on Sources for More
Information.) To comply with OMB and EPA's recommendations, EPA currently uses
discount rates of 3% and 7% for benefit analyses.

Which health endpoints accrue medical expenses or lost earnings for multiple years,
and how do I discount them?

BenMAP includes health and valuation functions for several chronic health effects,
including PM2.5-related chronic bronchitis and non-fatal acute myocardial infarctions
(AMIs, or heart attacks). Technically, AMIs are discrete, acute events, not chronic
conditions. However, heart attacks cause chronic follow-up health effects that accrue
medical expenses over time, similar to chronic conditions. You can discount the economic
value of these chronic effects through the valuation function in BenMAP. Chronic
bronchitis is assumed to last from the initial onset of the illness throughout the rest of the
individual's life. BenMAP currently includes one WTP function as well as two COI
functions representing the two discount rates for chronic bronchitis. AMIs are assumed to
accrue costs over five years. Although WTP functions for AMIs are not available,
BenMAP currently includes several COI functions that incorporate the direct medical
costs and the opportunity cost (lost earnings) for specific age groups at two discount rates.
See Appendix H for details on the discounting assumptions within the valuation functions.

168

September 2008

-------
Aggregation, Pooling, and Valuation

Should I discount the health incidence as well as the valuation?

You should not discount the health incidence for any of the scenarios mentioned above.
Changes in the lag assumptions do not change the total number of estimated deaths, for
example, but rather the timing of those deaths. If you discounted the health incidence
along with valuation, you would essentially be discounting twice. In contrast, the process
for conducting cost-effectiveness analyses and calculating QALYs involves discounting
life years. See Appendix G of the PM RIA for more information on cost-effectiveness
analyses (http://www.epa.gov/ttn/ecas/regdata/RIAs/Appendix%20G—Health%20Based%
20Cost%20Effectiveness%20Analysis.pdf).

Which health endpoints do not occur in the same year as exposure?

In many cases, the health effect from exposure to air pollution occurs shortly after
exposure, but there can be a significant lag between exposure and the health effect. The
cession lag can be a matter of hours or days in duration, but some health effects may lag
exposure by much longer. If exposure and the health effect do not occur within the same
year, it is necessary to discount those benefits back to the analysis year. The only health
function currently in BenMAP that falls into this category is PM2.5-related premature
mortality. Discounting PM-related premature mortality is controversial because the lag
structure is unknown, but scientific literature on similar adverse health effects and new
intervention studies suggest that premature mortality probably would not occur in the same
year as the exposure. (See: Roosli M, Kunzli N, Braun-Fahrlander C, Egger M. 2005.
"Years of life lost attributable to air pollution in Switzerland: dynamic exposure-response
model." International Journal of Epidemiology 34[5]: 1029-35.)

EPA's Science Advisory Board recommends future research to support the development of
defensible lag structures and provided a lag structure that could be assumed until
additional research has been completed. See Chapter 5 of the PM RIA for more detail on
assumed lag structures for PM2.5-related premature mortality (http://www.epa.gov/ttn/
ecas/regdata/RIAs/Chapter%205—Benefits.pdf). Some example lag structures from the
PM RIA are shown in Figures 6-1 and 6-2 below. Currently, BenMAP does not have the
capability to do this type of discounting, so you must discount outside of BenMAP.

Note: Discounting is not necessary for ozone-related premature mortality because it
occurs within the analysis year.

Figure 6-1. Graphical representation of assumed lag structures analyzed in EPA's

PM RIA as sensitivity analyses

169

September 2008

-------
Aggregation, Pooling, and Valuation

Assumed Lag Structures for PM2 5 Premature Mortality

SAB segmented lag
OMB segmented lag
Exponential decay model
5-year distributed lag

1st 2nd 3rd 4th 5th 6th 7th 8th 9th 10th 11th 12th 13th 14th 15th 16th 17th 18th 19th 20th
year year year year year year year year year year year year year year year year year year year year

Year Following Reduction in PM2 5

Figure 6-2. Graphical representation of cumulative assumed lag structures analyzed

in EPA's PM RIA as sensitivity analyses

Assumed Lag Structures for PM2 5 Premature Mortality (Cumulative)

SAB segmented lag
OMB segmented lag
Exponential decay model
5-year distributed lag

1st 2nd 3rd 4th 5th 6th 7th 8th 9th 10th 11th 12th 13th 14th 15th 16th 17th 18th 19th 20th
year year year year year year year year year year year year year year year year year year year year

Year Following Reduction in PM2

170

September 2008

-------
Aggregation, Pooling, and Valuation

7.1.3 Overview of Pooling

For many of the health endpoints (e.g., mortality), BenMAP contains many different
functions from different studies that you could choose to include in your configuration.
Combining data from several comparable studies in order to analyze them together is often
referred to as meta-analysis. For a number of reasons, it is often impractical or impossible
to combine the original data sets. Combining the results of studies provides a second-best
way to synthesize information. This is referred to as pooling.

BenMAP allows the pooling of incidence changes predicted by several studies for the
same pollutant-health endpoint group combination. It also allows the pooling of the
corresponding study-specific estimates of monetary benefits. See the appendix on
Uncertainty & Pooling for more detail on the pooling options available in BenMAP.

7.2 Create Aggregation, Pooling, and Valuation (APV) Configuration

To start you need to choose a configuration results file that contains incidence estimates at
the grid cell level (created by clicking the Create and Run Configuration button, see
Chapter 6). These incidence results are in files with a *.cfgr extension, and typically
stored in the Configuration Results folder.

171

September 2008

-------
Aggregation, Pooling, and Valuation

r i

Open

Look in:

^ Configuration Results ^ 11T

§03 2000 County ClosestMonitor lOPctRollback.cfgr

^]PM2.5 2000 County ClosestMoriitor lOPctRollback.cfgr

My Recent

Documents

FrZL

Desktop

My Documents

My Computer

My Network
Places

File name: _^J

Upen

Files of type: Configuration Results (".cfgr'l

Cancel

Once you open this file, you can begin creating your APV configuration. You will start
with selecting and pooling your incidence results, then move on to valuation. These
processes are described in detail below. Advanced functions, including aggregation, are
described below.

7.2.1 Pooling Incidence Results

After opening the configuration results file, you will find a list of Available Incidence
Results on the left-hand side of the window. The results are represented by the health
impact functions from which they were created, and are displayed in a tree-structure with
three levels, similar to the tree-structure found in the Configuration Settings Window.
Endpoint Groups occupy the top-most level, followed by Endpoints, and then individual
health impact Function identifiers. You will also see a pooling window at the right, where
you can select pooling options.

There are several steps to pooling your incidence results:

>Step 1. Add incidence results to the pooling window

In the Incidence and Pooling window, you will see all the incidence results generated

172

September 2008

-------
Aggregation, Pooling, and Valuation

from your configuration in the left hand column. You can drag individual incidence
results, or groups of results at any level of the tree and drag them over to the pooling
window. Note that once you drag a result or group of results into the pooling window it
will still be displayed on the left side. You do not have to drag all of your incidence
results over into the pooling window, but note that only those results showing in the
pooling window will be included in the pooled incidence or valuation results.

Incidence results are displayed in the pooling window in a tree structure determined by (1)
the order of the columns, and (2) the values of the identifying variables of the health
impact Functions from which the incidence results were generated (Endpoint Group,
Endpoint, etc.).

3 Incidence Pooling and Aggregation

Available Incidence Results

Select Pooling Methods

B- PM2.5
EE Mortality
B Chronic Bronchitis
B Acute Bronchitis
B Acute Myocardial Infarc
B Hospital Admissions, Re

Pooling Window N ame: | Pooling Window 1

Endpoint Group

Pooling Method

Hospital Admissions.

HA, Chronic Lung D

None

Moolgavkar

Adjusted Coefficient

Ito

Adjusted Coefficient

|j" HA, Chronic Lung D
0 - HA, Pneumonia
0- HA, Asthma
B Hospital Admissions, Ca
B Emergency Room Visits.
B Acute Respiratory Symp
B Lower Respiratory Symp
B Asthma Exacerbation
B Work Loss Days

B Upper Respiratory Symp

J- Window to Delete - Delete

Add J

Target Grid Type: |County

Configuration Results File Name(s): |C:\Program Files\BenMAP 3.0\Configuration Results\PM2.5 2000 County Clo ~ |

Browse

Advanced |

Cancel

Newt

Each line in the pooling window represents a node in the tree structure, with each node
representing either an individual incidence result or a collection of incidence results which
have common values for their leftmost identifying variables. The tree structure is
generated by comparing the leftmost values of the incidence result's identifying variables.
High level nodes in the tree are formed when results have common values for identifying
variables, and branches in the tree occur when the values differ.

173

September 2008

-------
Aggregation, Pooling, and Valuation

$ Incidence Pooling and Aggregation

Available Incidence Results

Select Pooling Methods

B- PM2.5
EE Mortality
0 Chronic Bronchitis
EE Acute Bronchitis
0 Acute Myocardial Infarc
EE Hospital Admissions, Re
0- Hospital Admissions, Ca
EE Emergency Room Visits.
0- Acute Respiratory Symp
EE Lower Respiratory Symp
0 Asthma Exacerbation
0 Work Loss Days
0 Upper Respiratory Symp

£|. UU_ il £l

Pooling Window N ame: | Pooling Window 1

Endpoint Group

Endpoint Author | Start A... End Ag\

Pooling Method

Asthma Exacerbatio

None

Asthma Exacerbation, Cough

None

Vedal et al.

Dstro et al.

Asthma Exacerbation, Wheeze

Ostro et al.

Asthma Exacerbation, Shortness of Breath

~stro et al.

J- Window to Delete --

Delete

| Add |

Target Grid Type: |County

Configuration Results File Name(s): |C:\Program Files\BenMAP 3.0\Configuration Results\PM2.5 2000 County Clo ~ |

Browse

Advanced

Cancel

In the above example, four incidence results have been dragged into the pooling window.
Each of the four health impact Functions has Endpoint Group Asthma Exacerbation.
Thus, the top line, or root of the tree structure, represents all four incidence results. A
branch then occurs in the tree structure, because two studies have Endpoint Asthma
Exacerbation, Cough, while two others have Endpoint Asthma Exacerbation, Wheeze
and Endpoint Asthma Exacerbation, Shortness of Breath. A further branch occurs within
Endpoint Asthma Exacerbation, Cough when Author of the two incidence results differs.
Once a node has only a single incidence result, no further branching can occur.

>Step 2. Sort results

After dragging incidence results into the Pooling Window, you can rearrange the order of
the columns (variables), and thus change the tree structure. To do this, click on a column
and hold the button down as you drag it to its new location. Note that Endpoint Group is
always the first column, and Pooling Method is always the last column. All the other
columns can be moved. To see how the order of the columns in the pooling window
affects the tree structure, consider the following example:

174

September 2008

-------
Aggregation, Pooling, and Valuation

$ Incidence Pooling and Aggregation

Available Incidence Results

Select Pooling Methods

B- PM2.5
EE Mortality

Pooling Window N ame: | Pooling Window 1

0 Chronic Bronchitis

Endpoint Group

Author

Endpoint

Start A...

End Agi

Pooling Method

EE Acute Bronchitis

Asthma Exacerbatio

None

E Acute Myocardial Infarc
EE Hospital Admissions, Re
E] Hospital Admissions, Ca
EE Emergency Room Visits.
0 Acute Respiratory Symp
EE Lower Respiratory Symp
0 Asthma Exacerbation
EE Work Loss Days
S Upper Respiratory Symp

Vedal et al.

Asthma Exacerbation, Cough

Dstro et al.

None

Asthma Exacerbation, Cough

Asthma Exacerbation, Wheeze

Asthma Exacerbation, Shortness of Breath

-j) ,*j

¦¦Window to Delete ¦¦ t| Delete

Add

i i

Target Grid Type: |Coi

1 1

Configuration Results File Narme(s): |l:''Program Files\BenMAP 3.0\Configuration Flesults\PM2.5 2000 County Clo

Browse

1 1

Advanced

Cancel

This example uses exactly the same incidence results as the previous example, but with
the Author column (variable) immediately after the Endpoint Group column.

> Step 3. Select pooling methods

Once the tree structure is set up in the Pooling Window, you are ready to select your
pooling methods. Essentially each pooling method involves a different method of
combining input incidence results to generate new incidence results. Results can be
pooled any time a branch occurs in the tree structure - that is, any time two or more results
share common values for their leftmost variables. BenMAP helps you to identify these
spots by inserting a value of None in the Pooling Method column at each spot where
pooling is possible.

Table 7-1 summarizes the different types of pooling approaches, and the Appendix on
Uncertainty and Pooling provides a detailed discussion of the approaches. Note that some
pooling methods are only available in Latin Hypercube mode. This is because these
pooling methods attempt to combine distributions of results into new distributions, and no
distributional information is available in Point Mode. The Pooling Method column will
thus have different values in its drop down list depending on the mode used to generate the
incidence results being pooled.

175

September 2008

-------
Aggregation, Pooling, and Valuation

Table 7-1. Pooling Approaches for Incidence and Valuation Results

Pooling
Approach

Description of Pooling Approach3

Availability

Point
Mode

Latin
Hypercube

None No pooling performed.

Sum Results are summed assuming they are perfectly correlated. In
(Dependent) Point Mode, this is just a simple sum. In Latin Hypercube
mode, BenMAP chooses the first point from each result in the
pooling and does a simple sum to generate the first point in the
pooled result, and so on for all of the points in the distribution
of results.

Yes
Yes

Sum Results are summed assuming that they are independent. A

(Independent) Monte Carlo simulation is used. At each iteration, a random
point is chosen from the Latin Hypercube of each result, and
the sum of these values is put in a holding container. After
some number of iterations, the holding container is sorted low
to high and binned down to the appropriate number of Latin
Hypercube points.

Yes

Subtraction Results are subtracted assuming they are perfectly correlated.

(Dependent) All subsequent results are subtracted from the first result (the
highest result in the display - to reorder results, simply click
and hold a result and then drag it to its new position). In Point
Mode, this is a simple subtraction. In Latin Hypercube mode,
BenMAP chooses the first point from each result in the pooling
and does a simple subtraction to generate the first point in the
pooled result, and so on for all of the points in the distribution
of results.

Yes

Subtraction Results are subtracted assuming that they are independent. A
(Independent) Monte Carlo simulation is used. At each iteration, a random
point is chosen from the Latin Hypercube of the first result, and

then random points are chosen from the Latin Hypercube of
each subsequent result and subtracted from the first. The result
is put into a holding container. After some number of
iterations, the holding container is sorted low to high and
binned down to the appropriate number of Latin Hypercube
points.

Yes

Subjective Weights are specified by the user (see Step 5, below). In Point
Weights Mode, the new result is generated by a simple weighted sum of
the input results. In Latin Hypercube mode, the results are
combined using the user specified weights with the "Round
Weights to Two Digits" Advanced Pooling Method. Note that
the weights you enter need not add up to one - BenMAP will
normalize them internally. Also note that BenMAP initializes
all the weights to 1/n, where n is the number of results being

Yes

176

September 2008

-------
Aggregation, Pooling, and Valuation

pooled.

Fixed Effects Pooling weights are generated automatically based on the
inverse variance of each input result, with the weights
normalized to sum to one. Results with a larger absolute
variance get smaller weights. Results are then combined
according to the chosen Advanced Pooling Method.

Yes

Random / Fixed BenMAP first tests if random weights should be used. If not,
Effects BenMAP uses fixed effects weights. If yes, the weights take
into account both the variance within each set of results and the
variance between sets of results. Results are then combined
according to the chosen Advanced Pooling Method.

Yes

>Step 4. Create additional pooling windows if needed

Within a given pooling window, you can have only one ordering of the columns
(variables). As we have seen, however, the ordering of the columns determines the
structure of the tree used to pool results. It may thus sometimes be necessary for analyses
to have multiple tree structures to handle the various pooling trees they require. To
facilitate this, BenMAP allows additional pooling windows to be added and deleted. To
open a new pooling window, simply click on the Add button. You may do this as many
times as needed to accommodate different sort orders. You can add the same incidence
results to as many different pooling windows as you like.

As needed you can also delete a pooling window by using the Window to Delete —"
drop-down menu to identify the pooling window, and then hitting the Delete button.

177

September 2008

-------
Aggregation, Pooling, and Valuation

$ Incidence Pooling and Aggregation

Available Incidence Results
E PM2.5

Select Pooling Methods

Pooling Window Name: IBasic Functions

Endpoint Group Start Age

Mortality

Author

Endpoint

Laden et al.

Pope et al.

Mortality, All Cau:

Pooling Method

None

Pooling Window Name: AMI IX Calcs

Endpoint Group

Acute Myocardial In

Endpoint

Author

Qualifier

Acute Myocardial Inl Peters et al.

Adjusted Coeffici

Pooling Method

None

Delete

IBasic Functions

Add

Target Grid Type: nunty

Configuration Results File Name(s): C:\BenMAP_Developrnent\Standar

AMI 7X Calcs

AMI Summed Incidence
Expert Functions
Mort Alt Thresholds
HA Summed Incidence

1 \County_t

Advanced

Cancel

7.2.1.1 Example: Simple Sorting & Pooling

If you add a single incidence result to the right-hand window, you will see just one line,
and therefore no opportunities to pool. This shown in the example below.

178

September 2008

-------
Aggregation, Pooling, and Valuation

« Incidence Pooling and Aggregation [_J|taJ(x]

Available Incidence Results

belect Pooling Methods

B PM2.5
EE Mortality
[jl- Chronic Bronchitis
[Jj Acute Bronchitis
EE Acute Myocardial Infarc
B Hospital Admissions, Re
El- HA, Chronic Lung D
EE HA, Chronic Lung D
EE HA, Pneumonia
0 HA, Asthma
EE Hospital Admissions, Ca
|j Emergency Room Visits.
EE Acute Respiratory Symp
[jl Lower Respiratory Symp
[j Asthma Exacerbation
EE Work Loss Days
EE Upper Respiratory Syrup

Pooling Window N ame: | Pooling Window 1

Endpoint Group

Endpoint Author | Start Age End Age

Pooling Method

Hospital Admissions, Respiratory

HA, Chronic Lung Disease l Moolgavkar 65 99

J-- Window to Delete -- ~ | Delete

Add

<_i ^ I |8l

Target Grid Type: [County

Configuration Results File Name(s): |C:\Program Files\BenMAP 3.0\Corifiguration Results\PM2.5 2000 County Clo Browse

Advanced |

Cancel

If you add a second incidence result to the window whose health impact Function has the
same Endpoint Group, but a different Author, you will then have a tree with two items in
it. The tree branches at the point where the two health impact Functions vary - at the
Author column.

Endpoint Group

Endpoint

Author

Start Age

E nd Age

Pooling Method

Hospital Admissions, Respiratory

HA.. Chronic Lung Disease

N one

Moolgavkar

I to

-------
Aggregation, Pooling, and Valuation

If you now add four more incidence results to the window whose health impact Functions
have the same Endpoint Group, you will see the following:

Endpoint Group

Endpoint Author | Start Age | End Age

Pooling Method

Hospital Admissions, Respiratory

None

HA, Chronic Lung Disease

None

Moolgavkar

I to

HA, Pneumonia

I to

HA, Asthma

Sheppard

-------
Aggregation, Pooling, and Valuation

Endpoint Group

Endpoint Author | Start Age | End Age

Pooling Method

Hospital Admissions, Respiratory

Subjective Weights

HA, Chronic Lung Disease

None

Moolgavkar

I to

HA, Pneumonia

I to

HA, Asthma

Sheppard

First, the HA, Chronic Lung Disease results are pooled to a give a single HA, Chronic
Lung Disease result.

>Next, the three separate Endpoint results are pooled to give a single Hospital
Admissions, Respiratory result.

Endpoint Group

E ndpoint

Author

Start Age

End Age

Pooling Method

Hospital Admissions, Respiratory

Subjective Weights

HA, Chronic Lung Disease

Fixed Effects

Moolgavkar

I to

HA, Pneumonia

I to

HA, Asthma

Sheppard

-------
Aggregation, Pooling, and Valuation

For example, you might want to pool all results of health impact Functions by a particular
author, rather than pooling all results of health impact functions of a particular endpoint.
The examples below show the same set of incidence results, first sorted by Author, then
sorted by Endpoint. As you can see, the pooling options are very different.

Endpoint Group

Author

Endpoint

Start Age

End Age

Pooling Method

Hospital Admissions.

None

Moolgavkar

None

HA, Chronic Lung Disease

HA, Chronic Lung Disease (less Asthma)

I to

None

HA, Chronic Lung Disease

HA, Pneumonia

Sheppard

HA, Asthma

-------
Aggregation, Pooling, and Valuation

databases available, and various pooling windows on the right side representing the
selected valuations and pooling options.

There will be one pooling window in the Select Valuation Methods, Pooling, and
Aggregation screen for each pooling window Incidence Pooling and Aggregation

screen. In each pooling window, there will be one result present for each incidence result
left over after all incidence pooling has occurred. Each of these results will be represented
by a Select —" value in the Valuation Method column.

The columns present in the Select Valuation Methods, Pooling, and Aggregation

window are determined by the incidence results left after all incidence pooling has
occurred. There will be exactly enough columns in each pooling window to represent the
"least" pooled incidence result. That is, the columns will be in the same order they were
in the Incidence Pooling and Aggregation window, but the only columns present will be
those up to the level of the pooled incidence result with the most columns left over after all
pooling has occurred. Here is an example:

« Select Valuation Methods, Pooling, and Aggregation Q[n](x]

Valuation Methods

Variable DataSet:

Pooling Window Name: Pooling Window 1

Endpoint Group

Hospital Admissions

E ndpoint

HA, Chronic Lung Disease

HA, Chronic Lung Disease (less Asthma)

HA, Pneumonia

HA, Asthma

Valuation Method Pooling Method

¦Select-

--Select--

-Select-

None

Advanced p Skip QALY Valuation

Cancel

There are several steps to take in the Select Valuation Methods, Pooling, and
Aggregation window:

>Step 1. Select your valuation methods

183

September 2008

-------
Aggregation, Pooling, and Valuation

Valuation methods are specific to endpoint groups, and sometimes to endpoints as well.
The only valuation methods which appear in the left side tree views are those which have
the same endpoint group values as the pooled incidence results which are available to be
valued. To select a valuation method, select it in the left side tree views and drag and drop
it onto the appropriate incidence result in the pooling window. Note that BenMAP will
only allow you to drop valuation methods onto incidence results which have the same
endpoint group value. For example, BenMAP will not allow you to drop a Mortality
valuation on & Hospital Admissions incidence result. Note also that you can only drag and
drop individual valuation methods, not entire groups of them. For explanations of the
various valuations, see the appendix on the Economic Value of Health Effects.

If you have added any of your own valuation methods, as described in the Valuation Data
section of the Loading Data chapter, you can drag and drop them in the same way as the
EPA Standard valuations.

« Select Valuation Methods, Pooling, and Aggregation

~SDH

Valuation Methods

Variable DataSet:

- EPA Standard Valuation Functions
- H ospital Admissions, R espiratory
+]¦ HA, All Respiratory
+] HA, Asthma

B HA, Chronic Lung Disease

-1

Pooling Window Name: Pooling Window 1

Endpoint Group

Endpoint

Valuation Method

Pooling Method

I^^^^^^^HCOI: med costs + waae loss ll

Hospital Admissions.

None

CO I: med costs + wage loss I
COI: med costs + wage loss I
+] HA, Chronic Lung Disease (less A
+] HA, Pneumonia

HA, Chronic Lung Disease

COI: med costs + w<

HA, Chronic Lung Disease (less

-Select-

HA, Pneumonia

-Select-

HA, Asthma

-Select-

—=4 ¦

Advanced p- Skip QALY Valuation

Cancel | Previous

When BenMAP runs the APV Configuration, it will generate a valuation result for each
valuation method you select by running the valuation method's valuation functions on the
incidence results for which they were selected. You do not need to select valuation
methods for every incidence result - incidence results without any valuation methods will
simply be ignored when valuation results are generated, aggregated, and pooled.

Because valuation functions have uncertainty associated with them, generating valuation
results is fairly complicated. The procedure used depends on whether the incidence results
being used were generated in Point Mode or with Latin Hypercube Points (see the chapter
on Incidence Estimation for details on Point Mode and Latin Hypercube Points).

In Point Mode, BenMAP simply runs the valuation functions once using the point

184

September 2008

-------
Aggregation, Pooling, and Valuation

estimate of the incidence result and the mean of the valuation function as inputs.

With Latin Hypercube Points, on the other hand, BenMAP generates one hundred
percentile points (from the 0.5th percentile to the 99.5th percentile) to represent the
distribution of the inputs to the valuation function. To get the value of the health
incidence, BenMAP multiplies each combination of values from the incidence result, Latin
Hypercube, with each of the hundred valuation points, and puts the results into a holding
container. (For example, if the incidence result has 10 Latin Hypercube Points and there
are 100 valuation points, then the holding container will have 1,000 values.) Finally, the
holding container is sorted low to high and binned back down to 100 Latin Hypercube
points (representing the 0.5th percentile to the 99.5th percentile of the economic value of
the incidence).

>Step 2. Sort results

Depending on how your incidence results were pooled, the columns in the valuation
pooling windows can be resorted in the same way as the incidence pooling window
columns. This resorting will have the same sort of impact on the tree structure of
valuation results that it had on the tree structure of incidence results. (See Step 2 in the
section on Pooling Incidence Results.)

> Step 3. Select pooling methods

The same pooling methods are available for valuation results which were available for
incidence results. (See Step 2 in the section on Pooling Incidence Results.) You should
note that when more than one valuation method is selected for a particular pooled
incidence result, it is possible to pool the generated valuation results.

185

September 2008

-------
Aggregation, Pooling, and Valuation

$ Select Valuation Methods, Pooling,

and Aggregation

00®

¦ i

Valuation Methods

Variable DataSet:

E EPA Standard Valuation Functions
- H ospital Admissions, R espiratory
+] HA, All Respiratory
+] HA, Asthma
-] HA, Chronic Lung Disease

COI: med costs + wage loss I
COI: med costs + wage loss I
COI: med costs + wage loss I
EE] HA, Chronic Lung Disease (less A
EE HA, Pneumonia

Pooling Window Name: Pooling Window 1

Endpoint Group

E ndpoint

Valuation Method

Pooling Method

Hospital Admissions.

None

HA, Chronic Lung Disease

•r

COI: med costs + wi

None

COI: rned costs + w<

Sum (Dependent)
Sum (Independent)
Subtraction (Depende
Subtraction (Indepenc

HA, Chronic Lung Disease (less

¦¦Select-

HA, Pneumonia

-Select-

HA, Asthma

-Select-

Random / Fixed Effec
Fined Effects

d§l

Advanced | p Skip QALY Valuation

Cancel | Previous

>Step 4. Select Variable DataSet

In order to proceed to the next step, you must select a Variable DataSet from the
drop-down menu. The files in the Variable DataSet can include a variety of data, such as
income and poverty data that might be used in health or valuation functions. For the
default EPA health and valuation functions, you just need to select the EPA Standard
Variables.

If you have developed your own setup, then you need to make sure that you also load a
Variable DataSet. This is necessary even if you do not need the extra variables that can
included in this dataset.

186

September 2008

-------
Aggregation, Pooling, and Valuation

$ Select Valuation Methods, Pooling,

and Aggregation

i ¦

Valuation Methods

Variable DataSet:

E EPA Standard Valuation Functions
- H ospital Admissions, R espiratory
+] HA, All Respiratory
+] HA, Asthma
-] HA, Chronic Lung Disease

COI: med costs + wage loss I
COI: med costs + wage loss I
COI: med costs + wage loss I
EE] HA, Chronic Lung Disease (less ^
El HA, Pneumonia

I Poverty

Endpoint Group

E ndpoint

Valuation Method

Pooling Method

Hospital Admissions.

None

HA, Chronic Lung Disease

Sum (Dependent]

COI: med costs + wi

COI: med costs + w<

HA, Chronic Lung Disease (less

¦¦Select-

HA, Pneumonia

-Select-

HA, Asthma

-Select-

Mi mi I Si

Advanced p Skip QALY Valuation

Cancel Previous Next

7.2.3 Using QALYs Weights

After you have specified your incidence and valuation pooling options, you can select
QALYs and pooling options. By default, the Skip QALY Weights box is checked in the
Select Valuation Methods, Pooling, and Aggregation window. This is done to save
time, as it takes BenMAP a few moments to load the QALY datafiles. If you want to use
QALY weights, then uncheck this box, and then click Next.

Advanced

f~ IS kip QALY Valuations

Cancel | Previous

The Select QALY Methods, Pooling and Aggregation window appears after you click
Next on the Select Valuation Methods, Pooling and Aggregation window. This
window should look quite similar to the Incidence Pooling and Aggregation and Select
Valuation Methods, Pooling and Aggregation windows, with tree views on the left side
representing the valuation databases available, and various pooling windows on the right
side representing the selected valuations and pooling options.

There will be one pooling window in the Select QALY Methods, Pooling, and
Aggregation screen for each pooling window Incidence Pooling and Aggregation
screen. In each pooling window, there will be one result present for each incidence result
left over after all incidence pooling has occurred. Each of these results will be represented
by a Select —" value in the QALY Method column.

As with the Select Valuation Methods, Pooling, and Aggregation window, the columns

187

September 2008

-------
Aggregation, Pooling, and Valuation

present in the Select QALY Methods, Pooling, and Aggregation window are determined
by the incidence results left after all incidence pooling has occurred. There will be exactly
enough columns in each pooling window to represent the "least" pooled incidence result.
That is, the columns will be in the same order they were in the Incidence Pooling and
Aggregation window, but the only columns present will be those up to the level of the
pooled incidence result with the most columns left over after all pooling has occurred.

Here is an example:

« Select QALY Methods, Pooling and Aggregation

QALY Valuation Methods

Variable DataSet:

E PA S tandard Variables

Pooling Window Name: |Pooling Window 1

Endpoint Group

Endpoint

QALY Valuation Pooling M ethod

Hospital Admissions.

None

HA, Chronic Lung Di

-¦Select-

HA, Chronic Lung Di

-Select-

HA, Pneumonia

-Select-

HA, Asthma

-Select-

Advanced

Cancel | Previous Run

There are several steps to take in the Select QALY Methods, Pooling, and Aggregation

window:

>Step 1. Select your QALY methods

The QALY datasets that come installed with the U.S. setup have only a limited number of
QALY methods — mortality, myocardial infarction (heart attack) and chronic bronchitis.
And, as with valuation methods, the QALY methods are specific to endpoint groups. The
only QALY methods which appear in the left side tree views are those which have the
same endpoint group values as the pooled incidence results which are available to be
valued. For example, in the standard APV Configuration for PM2.5 used by EPA, the
available QALY methods appear in the left side tree.

188

September 2008

-------
Aggregation, Pooling, and Valuation

On the other hand, in the case of the Endpoint Group Hospital Admissions, Respiratory,
there are no QALY methods from which to choose in the data that comes loaded with
Ben MAP, as seen in the following screenshot:

« Select QALY Methods, Pooling and Aggregation

QALY Valuation Methods

Variable DataSet:

| E PA S tandard Variables

Pooling Window Name: jPooling Window 1 w 1

Endpoint Group

Endpoint

I QALY Valuation Pooling Method

Hospital Admissions, Respiratory

None

HA, Chronic Lung Disease

-Select-

HA, Chronic Lung Disease (less Asthma)

--Select-

HA, Pneumonia

-Select-

HA, Asthma

--S elect-

Advanced |

Cancel I Previous

Run

To select a QALY method, select it in the left side tree views and drag and drop it onto the
appropriate incidence result in the pooling window. Note that BenMAP will only allow

189

September 2008

-------
Aggregation, Pooling, and Valuation

you to drop QALY methods onto incidence results which have the same endpoint group
value. For example, BenMAP will not allow you to drop a Mortality QALY method on a
Acute Myocardial Infarction incidence result. Note also that you can only drag and drop
individual QALY methods, not entire groups of them.

If you have added any of your own QALY methods, as described in the QALY Data
section of the Loading Data chapter, you can drag and drop them in the same way as the
EPA Standard valuations.

When BenMAP runs the APV Configuration, it will generate a QALY result for each
QALY method you select. You do not need to select QALY methods for every incidence
result - incidence results without any QALY methods will simply be ignored when QALY
results are generated, aggregated, and pooled.

Because QALY method databases have uncertainty associated with them, generating
QALY results is fairly complicated. The procedure used depends on whether the
incidence results being used were generated in Point Mode or with Latin Hypercube Points
(see the chapter on Incidence Estimation for details on Point Mode and Latin Hypercube
Points).

In Point Mode, BenMAP simply runs the QALY functions once using the point estimate
of the incidence result and the mean of the valuation function as inputs.

With Latin Hypercube Points, on the other hand, BenMAP uses a 5,000 draw Monte
Carlo approach to estimate QALYs. That is, BenMAP draws once from the QALY
database and once from the incidence Latin Hypercube, and multiplies the two draws
together, repeating this process 5,000 times and putting the results into a holding
container. Finally, the holding container is sorted low to high and binned back down to
100 Latin Hypercube points (representing the 0.5th percentile to the 99.5th percentile of
the estimated QALY distribution).

>Step 2. Sort results

Depending on how your incidence results were pooled, the columns in the QALY pooling
windows can be resorted in the same way as the incidence and valuation pooling window
columns. This resorting will have the same sort of impact on the tree structure of
valuation results that it had on the tree structure of incidence results. (See Step 2 in the
section on Pooling Incidence Results.)

> Step 3. Select pooling methods

The same pooling methods are available for QALY results which were available for
incidence and valuation results. (See Step 2 in the section on Pooling Incidence Results.)
You should note that when more than one QALY method is selected for a particular
pooled incidence result, it is possible to pool the generated QALY results.

190

September 2008

-------
Aggregation, Pooling, and Valuation

$ Select QALY Methods, Pooling and Aggregation

QALY Valuation Methods

Mortality

Mortali

Mortality

Mortali

Mortality

Mortali

Mortal

Mortali

Mortality

Mortali

Mortal

Mortali

Mortality

DR05^

DR o;

DR 05
DR 35
DR 35
DR 35
DR 35
DR 35
DR 35
DR 35
DR 35
DR 75
DR 75
DR 75
DR 75
DR 75
DR 75
DR 75
DR 75
DR 75 v

In-

variable DataSet:

EF'A Standard Variables

Pooling Window Name: Basic Functions

Endpoint Group

Mortality

:art Age

Author

QALY Valuation Pooling Method

None

1. / I ti

Mortality DR OX I 25

Mortality DFl IX I 25

-Select-

Pooling Window Name: AMI IX, Calcs

None
Sum (Dependent)
Sum (Independent)
Subtraction (Dependent)
Subtraction (independent)
Subjective Weights
Random / Fixed Effects

Endpoint Group

Acute Myocardial In

Endpoint

Author

Acute Myocardial Inl Peters et al.

QALY Valuation Pooling Method

-Select-

-S elect-

None

Advanced

Cancel

Run

7.2.4 APV Configuration Advanced Settings

At any point when specifying the incidence and valuation pooling options, you may click
on the Advanced button on the bottom-left of either the window for Incidence Pooling
and Aggregation, Select Valuation Methods, Pooling and Aggregation, or Select
QALY Methods, Pooling and Aggregation. This button will open the APV
Configuration Advanced Settings window.

The APV Configuration Advanced Settings window has three advanced option tabs for
the following:

p Aggregation and Pooling. This lets you choose the level of aggregation for the
incidence, valuation, and QALY results. You can also specify details on the pooling
procedure, such as whether to use a Monte Carlo or Latin Hypercube approach.

5^ Currency and Income. You can choose the currency year, how to adjust for inflation,
and how to adjust for income growth.

191

September 2008

-------
Aggregation, Pooling, and Valuation

$ APV Configuration Advanced Setting? |-J|nj|><_

Aggregation and Pooling | Currency and Income |

I ncidence Aggregation:

Report_Regions

Valuation Aggregation:

Report_Regions

QALY Aggregation:

Default Advanced Pooling Method:

Round weights to two

digi_J

Default Monte Carlo Iterations:

15000

Random Seed:

Random Integer

[Sort Incidence Result? 17

Cancel OK

7.2.4.1 Aggregation & Pooling

The Aggegation & Pooling tab lets you choose the level of aggregation for the incidence,
valuation, and QALY results. You can also specify details on the pooling procedure, such
as whether to use a Monte Carlo or Latin Hypercube approach.

192

September 2008

-------
Aggregation, Pooling, and Valuation

Incidence Aggregation: I

Valuation Aggregation: |l

QALY Aggregation: |

Default Advanced Pooling Method: lRound wei9hts to two

digi -

Default Monte Carlo Iterations: p000

Random Seed: |Random Integer

Sort Incidence Results W

Default Advanced Pooling Method

The relative contribution of any one study in the pooling process depends on the weight
assigned to that study. A key component of the pooling process, then, is the determination
of the weight given to each study. BenMAP lets users assign "subjective" weights and it
assign weights using a fixed effects or a random effects approach. There are three options
for using weights available in the Default Advanced Pooling Method drop-down menu:

Round weights to two digits. BenMAP rounds each weight to two digits (e.g. 0.73), and then
multiplies these weights by 100 to get two digit integers. Each entire distribution (set of Latin
Hypercube points) is then put into a holding container an integral number of times, according
to its integral weight. This holding container is then sorted low to high and binned down to
the appropriate number of Latin Hypercube points.

Round weights to three digits. BenMAP rounds each weight to three digits (e.g. 0.732), and
then multiplies these weights by 1000 to get three digit integers. Each entire distribution (set
of Latin Hypercube points) is then put into a holding container an integral number of times,
according to its integral weight. This holding container is then sorted low to high and binned
down to the appropriate number of Latin Hypercube points.

Use exact weights for Monte Carlo. BenMAP uses exact weights and a Monte Carlo
simulation. On each iteration of the procedure, a particular result is chosen with a probability
equal to its weight. Once a result is chosen, one of its Latin Hypercube points is chosen at
random and put into a holding container. This is done some number of times (see Monte
Carlo Iterations, below), and the holding container is then sorted low to high and binned

193

September 2008

-------
Aggregation, Pooling, and Valuation

down to the appropriate number of Latin Hypercube points.

Default Monte Carlo Iterations

This drop down list is only enabled when Use exact weights for Monte Carlo is selected as
the Advanced Pooling Method. It specifies the number of iterations the Monte Carlo
simulation should be run (see above). Its initial value is set by the Default Monte Carlo
Iterations value from the APV Advanced Settings window (see Step 1, above).

Random Seed

The APV Configuration Advanced Settings window allows the specification of a
Random Seed. As mentioned at the beginning of this chapter (see Open an APV Results
File, above) many of the pooling methods require the generation of sequences of random
numbers (e.g. choosing a random Latin Hypercube point during a Monte Carlo
simulation). Providing a specific Random Seed value allows you to ensure that the same
sequence of random numbers is generated as in a previous analysis, thus allowing exact
results to be reproduced.

If you do not set the Random Seed for a particular run, one will be generated
automatically from the system clock (the number generated will depend on the date and
time, and should change every minute). Normally, you should not set the Random Seed
value. If you need to reproduce a specific set of results, however, the random seed used to
generate previous APV Configuration Results can be obtained from an APV
Configuration Result file (*.apvr) Audit Trail Report.

Sort Incidence Results

The Sort Incidence Results should generally be always checked. This setting ensures that
the incidence Latin Hypercube results are sorted low to high.

7.2.4.2 Currency & Income

The Currency & Income tab allows you to specify an Inflation DataSet and a Currency
Year, which in combination allow you to change the year of the valuation dollars. The
Income Growth Adjustment panel allows you to adjust the valuation estimates to
account for the growth in income over time.

194

September 2008

-------
Aggregation, Pooling, and Valuation

Inflation Adjustment"

Inflation Dataset: \
Currency Year: 2000

Income Growth Adjustment-

Income Growth
Adjustment Dataset:

Year:

Endpoint Groups:

Inflation Adjustment

The inflation adjustment needs to be carefully considered in relation to the valuation
database that you are using. (This is discussed in detail in the section on loading inflation
data.) The default valuation database in the U.S. Setup has a currency year of 2000, so the
inflation dataset has a value of 1 for the year 2000.

Income Growth Adjustment

Willingness to pay (WTP) estimates are believed to be tied to the income of individuals.
As income rises over time, WTP estimates are likely to increase as well. The income
growth adjustment is designed to take this phenomenon into account, allowing you to
account for income growth between the time when WTP estimates were calculated and the
year of your analysis.

As with the inflation adjustment, the income growth adjustment has a close connection to
the valuation estimates. For example, the valuation estimates in the default version of
BenMAP are assumed to be based on income levels from 1990, so the income growth
adjustment database has a value of 1 for the year 1990. (This is discussed in detail in the
section on loading income growth data.)

To use the income growth adjustment, you need to choose a dataset and then choose the
income year that you want to use. It is common to set the "Year" variable to the year of

195

September 2008

-------
Aggregation, Pooling, and Valuation

the population forecast in your analysis. Of course, you can only choose from the
available data. If the income growth adjustment data only goes to 2024 and the population
data in your analysis are for 2030, then there will be some, unavoidable mismatch.

7.3 Run APV Configuration

After having specified the various aggregation, pooling, and valuation/QALY options, you
have the opportunity to save your APV Configuration for future use. The file that you
save has an "apv" extension. The configuration that you have specified is similar in idea
to the configuration that you developed for choosing health impact functions. (That
configuration has a "cfg" extension.) Both files allow you to save choices that you have
made, and re-run them at a later time.

You can save your APV configuration when you have finished making your valuation
pooling choices. Click the Run button, and then choose Save.

« Save Aggregation, Pooling, an... £j|nj|xj

Ready to run Aggregation, Pooling, and Valuation
Configuration. If you wish to save this configuration,
click the Save button. When ready, click OK. If you
are not ready to run this configuration, click Cancel.

Save

Cancel

You then need to name your configuration (*.apv) file. We suggest that you save this in
the Configurations folder. When ready to generate APV Configuration results, click the
OK button. BenMAP then requires that you specify a file in which to save the results,
with an ".apvr" extension.

7.4 Open & Modify Existing APV Configuration

If you have an existing configuration (*.apv) file, you can open, and then edit it.

196

September 2008

-------
Aggregation, Pooling, and Valuation

$ APV Configuration Creation Method (T](n](>^

C Create New Configuration for Aggregation, Pooling, and Valuation.

(• pDpen Existing Configuration file for Aggregation, Pooling, and Valuation (Kapv file).;

Cancel Go!

If you have only a few changes to make to an existing configuration, it is typically much
quicker to open the previous configuration, rather than entering all of your choices again.
Note that the various parts of an APV Configuration are quite interdependent, so
modifying part of the configuration may cause other parts to be reset. For example,
modifying the tree structure for incidence pooling will cause the valuation method
selection and valuation pooling tree structure to be cleared and reset. Changing the
Configuration Results Filename in the Incidence Pooling and Aggregation window will
not reset the incidence or valuation pooling trees as long as the new file contains incidence
results generated from the same health impact Functions as the old file. This can be quite
helpful for generating new APV Configuration Results from several different
Configuration Results files which were generated from different baseline / control
scenarios, but with the same set of health impact functions.

7.5 Questions Regarding APV Configurations

> I am at the BenMAP valuation window and cannot proceed. What should I do?

In order to proceed to the next step, you must select a Variable DataSet from the
drop-down menu in the Select Valuation Methods, Pooling, and Aggregation

window. The files in the Variable DataSet can include a variety of data, such as
income and poverty data that might be used in health or valuation functions. For the
default EPA health and valuation functions, you just need to select the EPA Standard
Variables.

>How do I edit or add other valuation functions?

To edit or add valuation functions you need to go to Modify Setup option in the Tools
drop-down menu available in the upper left-hand corner of the main BenMAP window.
See the valuation function section in the chapter on Loading Data for details on how to
do this.

197 September 2008

-------
Aggregation, Pooling, and Valuation

> How do I get to the QALY selection window?

By default, the Skip QALY Weights box is checked in the Select Valuation Methods,

Pooling, and Aggregation window. This is done to save time, as it takes BenMAP a
few moments to load the QALY datafiles. If you want to use QALY weights, then
uncheck this box, and then click Next.

Advanced

I- IS kip QALY Valuations

Cancel

Previous |

The Select QALY Methods, Pooling and Aggregation window appears after you
click Next on the Select Valuation Methods, Pooling and Aggregation window.

>How do I know what year dollars were used?

You can find the answer in the Audit trail for the APVR file that you generated.

>Do the currency year and year of the population data have to match?

No. The currency year and the year of the population data do not need to match.

198

September 2008

-------
Reports

CHAPTER 8
Reports

In this chapter...

>Use the Create Reports button to create incidence,
valuation, and QALY reports.

>Find the file formats, including variable definitions and
program compatibility, for each report type.

>Find out about using an Audit Trail report to keep track of
the options and assumptions underlying each analysis.

199

September 2008

-------
Reports

If you are interested in merely exporting results or merely viewing them, there are three
types of reports that you can access by clicking on the Create Reports button. You will
be asked which type of report you wish to create:

>Incidence, Valuation, & QALY Results: Raw, Aggregated, and Pooled use an Aggregation,
Pooling, and Valuation Results file (with the "".apvr" extension) to create report for incidence,
aggregated incidence, pooled incidence, valuation, aggregated valuation, pooled valuation,
QALY, aggregated QALY, or pooled QALY results. These reports are comma separated values
(CSV) files (*.csv) which can be read into various spreadsheet and database programs, such as
Microsoft Excel.

>Raw Incidence Results use a Configuration Results file (with the "".cfgr" extension) to create
reports for incidence results. These reports are CSV files.

> Audit Trail Reports provide a summary of the assumptions underlying each of five types of
files generated by BenMAP: Air Quality Grids (with the "".aqg" extension), Incidence
Configurations (with the "".cfg" extension), Configuration Results (with the ".cfgr"
extension), Aggregation, Pooling, and Valuation Configurations (with the "".apv" extension),
and Aggregation, Pooling, and Valuation Results (with the ".apvf' extension). These reports
can be viewed within BenMAP in an expandable tree structure, or can be exported to
tab-delimited text files.

8.1 One-Step Reports

200

September 2008

-------
Reports

One Step Results

Incidence Results

Table of Mortality
Incidence

Table of Morbidity
Incidence

Box Plot of
Mortality Incidence

Valuation Results

Table of Mortality
Valuation

Box Plot of
Mortality Valuation

Table of Morbidity
Valuation

Bar Chart of
Valuation

¦Audit Trail

Audit Trail Report

ReSU|ts pj|e- uration R esults\PM 2.5 Wizard, apvi|

Cumulative
Distribution
Functions

Load Apvr

8.1.1 One-Step Incidence Reports

There are three One-Step incidence reports:

Table of Mortality Incidence. Available for both the nation and report-region grid
definitions.

Table of Morbidity Incidence. Available for both the nation and report-region grid
definitions.

Box Plot of Mortality Incidence. Available only for the national grid definition.

8.1.1.1 Table of Mortality Incidence

This option can be used for both regional and national results. An example of regional
results is as follows:

201

September 2008

-------
Reports

« Report Preview

[h[xj

I File Page Zoom

|ta B & M < ~m Page 1 of 1 % % [P, 1 Zoom ^0.0 %

¦ I

Estimated Reduction in Premature Mortality
90th Percentile Confidence Intervals Provided in Parentheses

Western U. S.

Eastern U.S. Excluding CA California National Total

/V
V

Motility Impact Functions Derived from Epidemiology Lrtei ature

Harvard Six-City
Study

39,000
(21,000-56,000)

1,400
(760-2,000)

5,600
(3,100-8,100)

46,000
(25,000-66,000)

ACS Study

17,000
(6,700-27,000)

610
(240-990)

2,500
(970-4,000)

20,000
(7,900-32,000)

Woodruffetal.,
1997 (infant
mortality!

38
(19-58)

3.1
(1.5-4.7)

5.6
(2.7-8.4)

(23-71)

Mortality Impact Functions Derived from Expert Elicitation

ExpertA

46,000
(8,400-82,000)

4,000
(740-7,200)

6,600
(1,200-12,000)

56,000
(10,000-100,000)

Expert B

37,000
(5,400-74,000)

3,100
(350-6,500)

5,400

(800-11,000)

46,000
(6,600-91,000)

Expert C

38,000
(6,300-74,000)

3,300
(550-6,500)

5,500
(920-11,000)

47,000
(7,800-91,000)

Expert D

25,000
(5,100-40,000)

2,200
(440-3,500)

3,600
(740-5,800)

31,000
(6,300-49,000)

Expert E

57,000
(29,000-85,000)

5,000
(2,500-7,500)

8,300
(4,200-12,000)

71,000
(35,000-110,000)

Expert F

33,000
(22,000-45,000)

2,700
(1,800-3,700)

4,900
(3,200-6,600)

41,000
(27,000-56,000)

Expert G

20,000
(0-37,000)

1,800
(0-3,300)

2,900
(0-5,400)

25,000
(0-46,000)

Expert H

25,000
(95-59,000)

2,200
(8.3-5,200)

3,600
(14-8,600)

31,000
(120-73,000)

Expert I

1 A i-ifi

t- onn

1 A

8.1.1.2 Table of Morbidity Incidence

This option can be used for both regional and national results. An example of regional
results is as follows:

202

September 2008

-------
Reports

s Report Preview

~I XJ

1 File Page Zoom

|'£i ~ J Page |l

of 1 ^

> ^ l&i IB Zoom

^o.o % g

Estimated Reduction in Morbidity
90th Percentile Confidence Intervals Provided in Parentheses

Western U. S.

Eastern U.S. Excluding CA California National Total

Moil»idity Functions Dei ived f 10111 Epidemiology Liteiatme

Chronic bronchitis
(age»25and over)

11,000
(2,000-20,000)

820
(150-1,500)

1,900
(350-3,400)

14,000

(2,600-25,000)

Emergency room
visits for asthma
(age<19)

15,000
(8,900-21,000)

310
(180-430)

760
(460-1,100)

16,000

(9,600-23,000)

Acutebronchitis
(age 8-12)

29,000
(-1,000-58,000)

2,300
(-80-4,600)

4,800
(-170-9,500)

36,000
(-1,300-72,000)

Lowerespiratory
symptoms (age
7-14)

250,000
(120,000-380,000)

16,000
(8,000-25,000)

41,000

(20,000-62,000)

310,000
(150,000-470,000)

Upperrespiratory
symptoms
(asthmatic
children age 9-18)

190,000
(59,000-320,000)

12,000
(3,800-20,000)

31,000
(9,700-52,000)

230,000
(73,000-390,000)

Minor
restricted-activity
days(age18-65)

9,700,000
(8,200,000-11,000,000)

600,000
(500,000-690,000)

1,700,000
(1,500,000-2,000,000)

12,000,000
(10,000,000-14,000,000)

Work loss days
(age18-65)

1,600,000
(1,400,000-1,800,000)

100,000
(89,000-120,000)

290,000
(250,000-330,000)

2,000,000
(1,800,000-2,300,000)

Asthma
Exacerbation

230,000
r?R non-fi70 nnm

15,000
f1 R00-49 ooni

38,000
f 4 3no.i 1 n nom

280,000
n? nnn.R?n oooi

8.1.1.3 Box Plot of Mortality Incidence

This option can be used only for national results. An example of national results is as
follows:

203

September 2008

-------
Reports

* Box Plot

Results of Application of Expert Elicitation:
Annual Reductions in Premature Mortality in 2020

110,000
105,000
100,000
95,000
90,000
85,000
80,000
75,000
70,000
65,000
60,000
55,000
50,000
45,000
40,000
35,000
30,000
25,000
20,000
15,000
10,000
5,000
0

Distributions
Derived from
Epidemiology
Studies

Distributions Based on Expert Elicitation

-O-

Laden Expert Expert Expert Expert

et al Pope A Expert C Expert E Expert G Expert
(2002) et al B D F H

8.1.2 One-Step Valuation Reports

There are five One-Step valuation reports:

Table of Mortality Valuation. Available for both the nation and report-region grid
definitions.

Table of Morbidity Valuation. Available for both the nation and report-region grid
definitions.

Box Plot of Mortality Valuation. Available only for the national grid definition.
Cumulative Distribution Functions. Available only for the national grid definition.
Bar Chart of Valuation. Available only for the report-region grid definition.

8.1.2.1 Table of Mortality Valuation

This option can be used for both regional and national results. An example of regional
results is as follows:

204

September 2008

-------
Reports

s Report Preview

n x

I Fi'8

Page Zoom

Page

of 1 | % % pi

IB Zoom

80.0 %

Estimated Reduction in Premature Mortality
90th Percentile Confidence Intervals Provided in Parentheses

Eastern LI. S.

Western U. S.
Excluding CA

California

National Total

Mortality Impact Functions Derived from Epidemiology Literature

Harvard Six-City
Study

$210,000
($59,000-3:410,000)

$7,700
($2,100-115,000)

$31,000
($8,500-$59,000)

$250,000
($70,000-$480,000)

ACS Study

$94,000
($22,000-$190,000)

$3,400
($800-$6,900)

$14,000
($3,200-$28,000)

$110,000
($26,000-$230,000)

Woodruff eta!.,
1997 (infant
mortalityl

$210
($55-$410)

$17
($5-$34)

$31
($8-$60)

$260
($68-$500)

Mortality Impact Functions Derived from Expert Elicitation

Expert A

$250,000
($36,000-$560,000)

$22,000
($3,100-$49,000)

$36,000
($5,200-$81,000)

$310,000
($44,000-$690,000)

Expert B

$210,000
($22,000-$520,000)

$17,000
($1,400-$46,000)

$30,000
($3,300-$76,000)

$250,000
($27,000-$650,000)

Expert C

$210,000
($27,000-$490,000)

$18,000
($2,300-$43,000)

$30,000
($3,900-$71,000)

$260,000
($33,000-$600,000)

Expert D

$140,000
($22,000-$280,000)

$12,000
($1,900-$25,000)

$20,000
($3,200-$41,000)

$170,000
($27,000-$350,000)

Expert E

$310,000
($84,000-$610,000)

$28,000
($7,300-$54,000)

$46,000
($12,000-$89,000)

$390,000
($100,000-$750,000)

Expert F

$180,000
($53,000-$340,000)

$15,000
($4,400-128,000)

$27,000
($7,700-$49,000)

$220,000
($65,000-$420,000)

Expert G

$110,000
($0-$270,000)

$9,700
($0-$24,000)

$16,000
($0-$39,000)

$140,000
($0-$330,000)

Expert H

$140,000
($470-$380,000)

$12,000
($41 -$33,000)

$20,000
($69-$55,000)

$170,000
($580-$470,000)

;V|

8.1.2.2 Table of Morbidity Valuation

This option can be used for both regional and national results. An example of regional
results is as follows:

205

September 2008

-------
Reports

« Report Preview

- x.

| File Page Zoom

|H ~ # I M * ~ M Page

of 1 5

^ ^ IP, H Zoom

^0.0 % |

Estimated Reduction in Morbidity
90th Percentile Confidence Intervals Provided in Parentheses

Western U. S.

Eastern U.S. Excluding CA California National Total

Morbidity Functions Derived from Epidemiology Literature

Chronic bronchitis
(age>25andover)

$3,700
($310-$13,000)

$280
($23-$930)

$640
($53-$2,200)

$4,600
($390-$16,000)

Nonfatal
myocardial
infarction
(age>17)

$2,600
($800-$5,300)

$130
($41-$270)

$370
($120-$740)

$3,100
($960-$6,300)

Emergency room
visits for asthma
(age<19)

($2-$6)

$0
($0-$0)

$5
($3-$7)

Acutebronchitis
(age 8-12)

($0-$4)

$0
($0-$0)

$0
($0-11)

$2
($0-$5)

Lowerespiratory
symptoms (age
7-14)

C$2-$7)

$0
($0-$0)

$1
($0-$1)

$5
($2-$9)

Upperrespiratory
symptoms
(asthmatic
children age 9-18)

$5
($1-$10)

$0
($0-$1)

$1
($0-$2)

$6
($2-$13)

Minor
restricted-activity
days(age18-65)

$490
($290-$710)

$30
($18-$43)

$87
($51-$130)

$610
($360-$880)

Work loss days
(age 18-65)

$170
r si sn-®i am

$11
rsa-Jii

$36
mi -$411

$210
ran an-Si?4m

8.1.2.3 Box Plot of Mortality Valuation

This option can be used only for national results. An example of national results is as
follows:

206

September 2008

-------
Reports

* Box Plot

H0B

Results of Probabalistic Uncertainty Analysis:
Dollar Value of Health and Welfare Impacts

$1,050
$1,000
$950
$900

$750
$700
$650
$600
$550
$500
$450
$400
$350
$300
$250
$200
$150
$100
$50

000
,001
,001
00|,

oosji,

000
,000
,000
000
000

poo

,000
,000
000
,000
,000

poo

000
000
,000

poo

>ti ilnitions
< lived from
itlemiolcKjy
idles

-$50,000

Distributions Based on Expert Elicitation

-O--

Laden Expert Expert Expert Expert Expert Expert

et al Pope A Expert C Expert E Expert G Expert I Expert K Expert
(2002) et al B D F H J L

8.1.2.4 Cumulative Distribution Functions

This option can be used only for national results. An example of national results is as
follows:

207

September 2008

-------
Reports

« Cumulative Probabilty

Results of Probabilistic Uncertainty Analysis:

Cumulative Distributions of Dollar Value of Health and Welfare Impacts

0.9

0.7

0.6

0.5

0.3

0.2

0.8

$200,000

$400,000 $600,000

Million ($1999)

$800,000

$1 ,ooo,oor

— Laden et al

— Pope et al

Expert A

— Expert B

Expert C

— Expert D

— Expert E — Expert F

— Expert G

— Expert H

Expert I

— Expert J

— Expert K

Expert L

8.1.2.5 Bar Chart of Valuation

This option can be used only for regional results. An example of regional results is as
follows:

208

September 2008

-------
Reports

« RegionsBarForm

Q@®

Total Monetized Benefits
(Millions of 1999$)

~ (3) CA

~ (2) East

~ (1) West

| $210,000 |

I $94,000 | | $92,000 | | $93,000 |

I $82,000 I

Laden Pope Expert Expert Expert Expert

et al et al L K J I

8.1.3 One-Step Audit Trail

As described in the section on Audit Trails, the Audit Trail Reports provide a summary
of the assumptions underlying the various parts of the analysis. The report for the
One-Step analysis provides information regarding the Aggregation, Pooling, and
Valuation Results file (with the ".apvr" extension). Below is an example of the Audit
Trail Report:

209

September 2008

-------
Reports

« Audit Trail Report

CAPrograrn Files^BenMAP 3.0\Configuration ResultsMJser Manual Configuration Results\PM2.5 Wizard.apvr

-]¦¦ Configuration Results: CAPrograrn Files\BenMAP 3.0\Configuration ResultsMJser Manual Configuration Results\PM2.5 Wizardcfgr
Latin Hypercube Points: 10
Population Dataset: United States Census - County
Year: 2020
Threshold: 0
+ Grid Definition
+ Selected Studies

+ Baseline Air Quality Grid: CAPrograrn FilesVBenMAP 3.0\Air Quality GridsMJser Manual Air Quality GridsVBaseline PM25 2000 Closestf
+ Control Air Quality Grid: CAPrograrn Files\BenMAP 3.OVAir Quality GridsMJser Manual Air Quality Grids\Control PM25 2000 ClosestMoi
E Advanced

B Incidence Pooling Windows

+ Incidence Pooling Window: Basic Functions
El Incidence Pooling Window: AMI 1% Calcs
El Incidence Pooling Window: AMI Summed Incidence
El Incidence Pooling Window: Expert Functions
+ Incidence Pooling Window: Mort Alt Thresholds
El Incidence Pooling Window: HA Summed Incidence
El Valuation Pooling Windows
El QALY Pooling Windows

Export

8.1.4 Load One-Step Results to Generate Reports

Alternatively, you can load in a data file and create reports at the One Step Analysis
window.

210

September 2008

-------
Reports

$ One Step Analysis

~OKI

Run Name:
Output Directory:

1. Air Quality Li rids
Baseline File:

Control File:

Select

Open Create |
Open | Create

Map Deltas

2. Incidence

At what level would you like to aggregate your incidence results?

Nation

r3. Valuation

At what level would you like to aggregate your valuation results?

Valuation Aggregation must be at a level the same as or higher than the Incidence aggregation.

Nation

Open Results

Cancel

Click on the Open Results button. This will bring up the One Step Results window.

211

September 2008

-------
Reports

One Step Results

Incidence Results

Table of Mortality
Incidence

Table of Morbidity
Incidence

Box Plot of
Mortality Incidence

Valuation Results

Table of Mortality

Box Plot of

Valuation

Mortality Valuation

Table of Morbidity

Bar Chart of

Valuation

Cumulative
Distribution
Functions

Audit T rail

Audit Trail Report

Results File:

Load Apvr

Click on the Load APVR button, and browse for the APVR file that you want to load.
Make sure that the file you load has the results needed for the reports that you will be
generating. If you find that the tables, box plot, or bar charts look incomplete,
misformatted, or otherwise odd, check that the CFG and APV files are reasonably similar
to the default files that come with BenMAP. To the extent that there are divergences, then
it is likely that these pre-defined reports (which depend on the output looking a certain
way) will not look right.

8.1.5 Export One-Step Reports

You have the option to export tables, box plots, and graphs by clicking the Save button at
the top of the window containing the figure of interest, or go to the File menu option at the
top of the window and choose Save As. Name the file in the File name box, and from the
Save as type drop-down menu choose the file type that you want to save.

To export an Audit Trail, click the Export button. BenMAP will generate a text file with
the information contained in the Audit Trail.

272

September 2008

-------
Reports

8.2 Incidence, Valuation, & QALY - from APVR file

Using the results in the Aggregation, Pooling, and Valuation Results file (".apvr"
extension), you can create nine types of reports: raw incidence, aggregated incidence,
pooled incidence, valuation (of the pooled incidence), aggregated valuation, and pooled
valuation. After you click on the Create Reports button, you need to specify the APV
Configuration Results File that you want to use.

After choosing your APVR file and clicking the Open button, you then need to choose a
Result Type.

213

September 2008

-------
Reports

$ Choose a Res... 0[~](><]

Result Type
C Incidence Results
C Aggregated Incidence Results
r Pooled Incidence Results
C Valuation Results
C Aggregated Valuation Results
r Pooled Valuation Results
r QALY Results
C Aggregated QALY Results
C Pooled QALY Results

Cancel

Table 8-1 describes the results contained in each type of report.

Table 8-1. Summary of the Reports Generated from APVR Files

Result Type Description

Incidence Results Incidence results for each health impact function at the grid-cell level or

aggregated at the county, state, or national level.

Aggregated Incidence Incidence results for each health impact function aggregated to the level you
Results specified in the Aggregation, Pooling, and Valuation configuration file.

Pooled Incidence Results Incidence results aggregated and pooled as you specified in the Aggregation,

Pooling, and Valuation configuration file.

Valuation Results Valuation results for the pooled and aggregated incidence results.

Aggregated Valuation Valuation results aggregated to the level you specified in the Aggregation,
Results Pooling, and Valuation configuration file.

Pooled Valuation Results Valuation results aggregated and pooled as you specified in the Aggregation,

Pooling, and Valuation configuration file.

QALY Results QALY results for the pooled and aggregated incidence results.

Aggregated QALY QALY results aggregated to the level you specified in the Aggregation,

Results Pooling, and Valuation configuration file.

214

September 2008

-------
Reports

Pooled QALY Results QALY results aggregated and pooled as you specified in the Aggregation,

Pooling, and Valuation configuration file.

8.2.1 Incidence Results

The Incidence Results report gives you the opportunity to examine the results of each
health impact function at the grid-cell level, or aggregate them to, say, the state or
national level. Simply select the options that you desire from the four main sections of
the Configuration Results Report window: Column Selection (these include Grid
Fields, C-R Function Fields, and Result Fields), Grouping Options, Display Options,
and Advanced Options. As you modify your choices, the Preview section will be updated
accordingly.

The Column Selection section allows you to choose the field names (and values) which
will appear in the report. The Grid Fields section allows the inclusion of Col and Row
fields, which can be helpful in identifying the grid-cell of a particular line in the report.
These will not always be necessary, however - for example, when results have been
aggregated to the national level. The C-R Function Fields section allows the inclusion of
various fields which can be helpful in identifying the health impact function of a particular
line in the report. Almost all of the field names have appeared previously in the
preparation of a Aggregation, Pooling, and Valuation Results file. Finally, the Result
Fields section allows the inclusion of various types of results.

215

September 2008

-------
Reports

In the Grouping Options section, you can change the sorting of the results, by clicking
the radio buttons Group by Gridcell, then by C-R function and Group by C-R function,
then by Gridcell.

In the Display Options section, you may set the number of digits that appear after the
decimal point, and you can set the number of rows that appear in the preview window.

In the Advanced Options, you can set the level of aggregation at the grid cell (none),
county, state, and national levels. You can also choose to generate Population Weighted
Deltas, which BenMAP calculates at the national level for each health impact function by
weighting the change in the pollution metric at each grid cell with the population of the
grid cell. For example, if there are large changes in highly populated urban grid cells and
relatively small changes in lightly population rural grid cells, then the population-weighted
change would reflect the large urban changes and be relatively large.

As noted above, most of these fields have been seen previously, so Table 8-2 provides a
summary of the fields that are new to this report format. (See the section on Health Impact
Functions data format in the Loading Data chapter for a description of the C-R Function
Fields.)

216

September 2008

-------
Reports

Table 8-2. Selected Variables in the Reports Based on the APVR file

Variable

Description

Col The column of the grid cell of the result. For grid cell level results, this is the column

of the grid cell. For county and state level results, this is the state FIPS code. For
national results, this is always 1.

Row The row of the grid cell of the result. For grid cell level results, this is the row of the

grid cell. For county results, this is the county FIPS code. For state and national

results, this is always 1.

DataSet Specifies the dataset from which a health impact function was chosen.

Population Population provides the number of persons used in the health impact function

calculation.

Delta The difference between the baseline and control scenarios for the metric used in the

health impact function. Calculated by subtracting the metric value in the control
scenario from the metric value in the baseline scenario.

Point Estimate The point estimate for the result from the health impact function. The point estimate
is generally based on the mean estimate of the "Beta" from the health impact function.

Mean Mean of the points in the Latin Hypercube for this result. The mean is set to missing

if the Point Mode option is chosen.

Baseline Estimate based on the baseline function, which typically estimates health impacts due

to all causes (not just air pollution-related causes).

Percent of Baseline Estimates the percentage change in health impacts (e.g., hospital admissions) due to
the change in air quality from the baseline to the control scenario. Calculated by
dividing the Point Estimate by the Baseline.

Std Dev Standard deviation calculated based on the points in the Latin Hypercube for this

result.

Latin Hypercube The number of percentiles depends on the number of points in the Latin Hypercube
Points for this result.

8.2.2 Aggregated Incidence Results

The Aggregated Incidence Results report presents the incidence results at the
aggregation level that you have previously specified in the Aggregation, Pooling, and
Valuation Configuration file. If you want to change the level of aggregation, you need to
revise your choices in the APV Configuration Advanced Settings window (see the
section on advanced APV options).

217

September 2008

-------
Reports

The Column Selection section looks largely the same as in the Incidence Results Report.
The Grouping Options section and Display Options section are exactly the same, as is
the Advanced Options section.

« Configuration Results Report

File

Column Selection
Grid Fields:

C-R Function Fields:

DataSet
v Endpoirit Group
Endpoint

~ Pollutant
Metric

~ Seasonal Metric

~ Metric Statistic

Year
Location

~ Other Pollutants
Qualifier

Reference ~ P2Beta

~ Race ~ A

~ Ethnicity ~ NameA

~ Gender ~ B

~ Start Age ~ NarneB

~ End Age ~ C

~ Function ~ NameC
I ] Incidence DataSet

E Prevalence DataSet
Beta
DistBeta
PIBeta

~ID®

Result Fields:

@ Point Estimate
0 Population
@ Delta
@ Mean
0 Baseline
jy9 Percent of Baseline
0 Standard Deviation
@ Variance

0 Latin Hypercube Points

Grouping Options

(• Group by Gridcell, then by C-R function.
Group by C-FI function, then by Gridcell.

Display Options
Digits After Decimal Point: 4

Advanced Options

Population Weighted Deltas: f"

Elements in Preview: 25

Aggregation Level: -1

Preview

Column

Row

Endpoint Group

Author

Point Estimate

Population

Delta

Mean

Baseline *

Mortality

Laden et al.

5,510.7412

28,673,236.0000

1.4390

5,506.3750

301,383.0313 ~

e. joe nco nnnn

1 *07n

O HOrt QQ.1 A

TJQ C .1 Q riL' \'R —

Done

8.2.3 Pooled Incidence Results

The Pooled Incidence Results report provides results aggregated and pooled to the level
that you previously specified in the Aggregation, Pooling, and Valuation Configuration
file. This report has fewer health impact Function Fields than the Aggregated Incidence
Results Report, and values for others will be blank. This is because after pooling, only
enough fields are retained to uniquely identify individual results. In addition, the
Advanced Options section (available for Incidence Results and Aggregated Incidence
Results reports) no longer exists, as the options in it do not apply to aggregated incidence
results.

218

September 2008

-------
Reports

$ APV Configuration Results Report

ED®®

File

Column Selection
Grid Fields:

Pooled C-R Function Fields:

Result Fields:

0 Column
0 Row

y! Endpoint Group

~ Start Aqe

~ Endpoint
Qualifier
Location
End Age
Year

~ Other Pollutants
Reference
Race
Gender

Add Sum:

Function
Pollutant
Metric

~ Seasonal Metric
C I Metric Statistic

DataSet
Version

~ Pooling Window

0 Point Estimate
0 Mean

~ Standard Deviation

~iSBHHi

v1 Latin Hypercube Points

Grouping Options

(• Group by Gridcell, then by Pooled C-F! Function.
C Group by Pooled C-R Function, then by Gridcell.

Display Options -
Digits After Decimal Point:

E lements in Preview: 25

"3
1

Preview

Column

Row

Endpoint Group

Author

Point Estimate

Mean

Percentile 5

Percentile 15

Percentile Jl

Mortality

5,511

5,506

3,011

3,938

4,488 —'

Mortality

2,428

2,426

952

1,489

1,824

-------
Reports

$ APV Configuration Results Report

File

Column Selection

Grid Fields:

Valuation Method Fields:

0 Endpoint Group

~ Author

~ Endpoint

~ Qualifier

~ Location

~ End Age

~ Year

~ Other Pollutants

~ Reference

~ Race

~ Gender

Add Sunns

~ Function
ID Pollutant

~ Metric

~ Seasonal Metric

~ Metric Statistic
C DataSet

~ Version

~ Pooling Window

~ ValuationMethod

Point Estimate

Result Fields:

\qr

I Mean

ID Standard Deviation
~ Variance
I Latin Hypercube Points

Grouping Options

C Group by Gridcell, then by Valuation Method.
(* Group by Valuation Method, then by Gridcell.

Display Options
Digits After Decimal Point:

Elements in Preview: 25

"3
1

Preview

Endpoint Group Start Age Column Row Mean

Percentile 0.5

Percentile 1.5

Percentile 2.5

Mortality

6,207,985,664

-467,934,976

584,980,929

1,001,921,344

Mortality

3,447,534,336

¦259,800,606

324,785,280

555,720,064

Mortality

3,086,246,144

-232,595,184

290,774,912

497,714,208

Mortality
-

30,715,885,568

-2,315,094,764

2,894,176,256

4,955,653,120

Done

8.2.4.1 Add Sums Button

To calculate the total value of different groups of effects, click on the Add Sums button in
the APV Configuration Results Report window. In the next window you can specify the
effects you want to sum and how you want to sum them. In the Include in Total field,
simply check the effects that you want to sum together, and then name the sum in the
Valuation Sum Identifier field. You also need to specify whether you are going to use
the Dependent or Independent sum approach. (See Chapter 7 for a discussion of these two
summing approaches.) In the Summation Type use the drop-down menu, choose the
approach that you want to use. If you choose the Independent sum, then you also need to
specify the number of Monte Carlo draws that you want BenMAP to use. The default is
5,000. You may repeat this procedure to generate as many totals as you like.

Note that the Add Sums button is only enabled for reports involving monetary valuations,
not those involving incidence estimates. Typically, incidence estimates should not be
summed across endpoint groups (for example, Mortality and Hospital Admissions,
Respiratory). Within endpoint groups, incidence estimates can be summed - you may do
this in Aggregation, Pooling and Valuation Configurations. Once results are in monetary

220

September 2008

-------
Reports

values, however, summing across endpoint groups can be useful in calculating aggregate
benefits.

$ Add Valuation Sum (T]| nj(xj

Pooling Window

Endpoint Group

Start Age

ValuationMethod

Include in Total

-j

Mortality

VSL, based on range from $1 to $10 million, nor

Mortality

VSL, based on range from $1 to $10 million, nor

Mortality

VSL, based on range from $1 to $10 million, nor

Mortality

VSL, based on range from $1 to $10 million, nor

Chronic Bronchitis

WTP: average severity I 30-99

Acute Myocardial Inl

CO I

5 yrs med, 5 yrs wages, 3% DR, Russell (1!

Acute Myocardial Inl

CO I

5 yrs med, 5 yrs wages, 3% DR, Wittels (1J

Acute Myocardial Inl

CO I

5 yrs med, 5 yrs wages, 3% DR, Russell (1!

Acute Myocardial Inl

CO I

5 yrs med, 5 yrs wages, 3% DR, Wittels (1J

Acute Myocardial Inl

CO I

5 yrs med, 5 yrs wages, 3% DR, Russell (1!

Acute Myocardial Inl

CO I

5 yrs med, 5 yrs wages, 3% DR, Wittels (1J

Acute Myocardial Inl

CO I

5 yrs med, 5 yrs wages, 3Z DR, Russell (1!

a l _ l j i: _ i i.. r

rni. n i c nn \. m rl

I 1 .

M I ~

Valuation Sunn Identifier: Summation Type:

|Mortality & Chron Bronc |Dependent ^ | I5000 Cancel

8.2.5 Aggregated Valuation Results

The Aggregated Valuation Results report presents valuation results aggregated to the
level you specified at the Advanced button when creating the APV configuration file.

Note that the aggregation level specified for the valuation does not need to be same as that
for the incidence aggregation. In the example below, the valuation results are aggregated
to the three so-called "report regions," so in the Row field, you can see the value changing
from 1 (East), 2 (West minus California), and 3 (California). (The underlying incidence
data in this example were aggregated to the state level.).

As with the Valuation Results Report, you can use the Add Sums button to create totals
with various valuation results.

227

September 2008

-------
Reports

« APV Configuration Results Report

0®®

File

Column Selection

Grid Fields:

Valuation Method Fields:

0 Column
SI Row

V Endpoint Group

BIZ. _

~ Author

~ Endpoint

~ Qualifier
Location
End Age
Year

~ Other Pollutants

~ Reference
Race
Gender

AddS ums

Function
!~ Pollutant

Metric
~ Seasonal Metric
C I Metric Statistic
DataSet
Version
D Pooling Window
D ValuationMethod

Point Estimate

Result Fields:

fqr

Mean
Standard Deviation
~ Variance

w* Latin Hypercube Points

Grouping Options

C Group by Gridcell, then by Valuation Method,
f* Group by Valuation Method, then by Gridcell.

Display Options —

Digits After Decimal Point:

E lements in Preview: 25

Preview

Endpoint Group

Start Age

Column

Row

Mean

Percentile 0.5

Percentile 1.5

Percentile 2.5 *"

Mortality

30,681,671,680

-2,312,515,840

2,890,852,448

4,950,132,736

Mortality

212,938,0SE,640

-16,048,456,704

20,062,701,568

34,344,161,280

Mortality

7,725,341,184

-582,146,368

727,760,384

1,245,023,744

Mortality

13,519,849,472

-1,018,885,120

1,153,763,456

1,900,353,152

Mortality

93,907,435,520

-7,076,745,728

8,013,245.952

13,196,695,552

Mortality

3,393,062,400

-255,669,120

289,475,936

476,564,096

Mortality
4

47.530.772

-60.356.316

-51.246.468

-45.839.916 TJ

Done

8.2.6 Pooled Valuation Results

The Pooled Valuation Results report presents valuation results aggregated and pooled to
the level you specified at the Advanced button when creating the APV configuration file.
As with the Pooled Incidence Results Report, fewer Pooled Valuation Method Fields are
available, because only enough fields are retained to uniquely identify individual results.
As with the Valuation Results Report, you can use the Add Sums button to create totals
with various valuation results.

222

September 2008

-------
Reports

« APV Configuration Results Report

~EDS

File

Column Selection
Grid Fields:

Pooled Valuation M ethod Fields:

y! Column
yi Row

151 Endpoint Group

Start Age

Author

~ Endpoint

~ ValuatioriM ethod

C Pooling Window

Qualifier

Location

Point Estimate

Result Fields:

rnr

M ean
Standard Deviation
~ Variance

Latin Hypercube Points

Add S urns

Grouping Options

• Group by Gridcell, then by Pooled Valuation Method.
C Group by Pooled Valuation Method, then by Gridcell.

Display Options —

Digits After Decimal Point:

E lements in Preview: 25

"3
1

Preview

Column

Row

Endpoint Group

Mean

Percentile 0.5

Percentile 1.5

Percentile 2.5

Percentile 3.5 *

Mortality

30,681,671,680

-2,312,515,840

2,890,952,448

4,950,132,736

6,654,758,912—'

Mortality

13,519,849,472

-1,018,885,120

1,153,763,456

1,900,353,152

2,503,057,864

-------
Reports

« APV Configuration Results Report

~us

File

Column Selection
Grid Fields:

0 Column
0 Row

QALY Weights Fields:

Endpoint
Author
Qualifier
Location
0 Start Age
E rid Age
Year
D Other Pollutants
Reference
Race
~ Ethnicity

AddS urns

Gender
Function
Pollutant
Metric
C Seasonal Metric
Metric Statistic
DataSet
Version
L Pooling Window
& QALY Weights

Point Estimate

Result Fields:

Mean

~ Standard Deviation
Variance

Latin Hypercube Points

Grouping Options

C Group by Gridcell, then by QALY Weights.
IGroup by QACy Weights, then by Gridceiij

Display Options
Digits After Decimal Point:

Elements in Preview: 25

Preview

Endpoint Group

Start Age

QALY Weights

Column

Row

M ean

Percentile 0.5 Percentile 1.5 Percentile i •"

Mortality

Mortality DR 3X130-34

108

Mortality

Mortality DR 3% 130-34

105

Mortality

Mortality DR 3% I 30-34

Mortality

Mortality DR 3% I 30-34

553

211

214

216

Mortality

Mortality DR 3% I 30-34

Mortality

±1 I

Mortality DR 3% I 30-34

Done

8.2.8 Aggregated QALY Results

The Aggregated QALY Results report presents QALY results aggregated to the level you
specified at the Advanced button when creating the APV configuration file. (Recall that
you can aggregate incidence, valuation, and QALY results to different levels.) In the
example below, the QALY results are aggregated at the national level, so the values in the
Column and Row fields are set to 1. As with the Valuation Results Report, you can use
the Add Sums button to create totals with various QALY results.

224

September 2008

-------
Reports

« APV Configuration Results Report

~HIS

File

Column Selection
Grid Fields:

Aggregated QALY Weights Fields:

Result Fields:

0 Column
@ Row

@ Endpoint Group

Gender

Endpoint

Function

Author

Pollutant

~ Qualifier

Metric

C Location

Seasonal Metric

Efl Start Age

Metric Statistic

End Age

DataSet

Year

~ Version

Other Pollutants

i Pooling Window

Reference

~ QALY Weights

Race

~ Ethnicity

~ Point Estimate
IZ Mean

~ Standard Deviation

ls£ Latin Hypercube Points

Add S urns

Grouping Options

(• jGroup by Gridcell, then by Aggregated QALY Weights.;
C Group by Aggregated QALY Weights, then by Gridcell.

Display Options ~

Digits After Decimal Point:

Elements in Preview: 25

Preview

Column

Row Endpoint Group Start Age Mean Percentile 0.5 Percentile 1.5 Percentile 2.5 Percentile 3.5

Mortality

4,432

1,676

1,698

1,711

1,724

Mortality

12,321

4,SS6

4,349

4,331

5,025

Mortality

23,105

8,740

3,343

3,324

3,336

Mortality

47,684

18,023

13,265

18,425

18,543

Mortality

61,375

23,404

23,631

23,332

24,057

Done

8.2.9 Pooled QALY Results

The Pooled QALY Results report presents QALY results aggregated and pooled to the
level you specified in the Aggregation, Pooling, and Valuation configuration file. As with
the other valuation reports, you can use the Add Sums button to create totals with various
valuation results.

225

September 2008

-------
Reports

« APV Configuration Results Report

U@E

File

Column Selection

Grid Fields:

* Column
0 Row

Pooled QALY Weights Fields:

0 Endpoint Group
E ndpoint
Author

~ Qualifier
Location

~ Start Age

~ QALY Weiqhts

Pooling Window

Result Fields:

~ Point Estimate
0 Mean

~ Standard Deviation

0 Latin Hypercube Points

Add Sums

Grouping Options

'~ IGroup by Gridceii, then by Pooled QALY Weights.;
f" Group by Pooled QALY Weights, then by Gridceii.

Display Options-
Digits After Decimal Point:

E lements in Preview: 25

Preview

Column

Row

Endpoint Group

Pooling Window

M ean

Percentile 0.5

Percentile 1.5

Percentile 2.5

F'erci

Mortality

3 percent DR

220,911

83,539

84,596

85,312

85,85

Chronic Bronchitis

3 percent DR

65,456

5,363

7,640

9,130

10,3'

Acute Myocardial Infarction

3 percent DR

29,023

3,459

5,110

6,142

7,04<

Mortality

7 percent DR

175,356

68,786

68,7!

Chronic Bronchitis

7 percent DR

42,570

3,445

4,883

5,920

6,62^

Acute Myocardial Infarction

7 percent DR

23,806

3,021

4,361

5,178

5,87'

-------
Reports

« APV Configuration Results Report

JnJ(:x

Save Ctrl+S

Print Ctrl+P

Column

W
y Row

Dooled C-R Function Fields:

Author

~ Endpoint
Qualifier
Location
End Age
Year

~ Other Pollutants
Reference
Race
Gender

Add Sums

Function
3 Pollutant

Metric
~ Seasonal Metric
C Metric Statistic
DataSet
Version
D Pooling Window

Result Fields:

~ Point Estimate
0 Mean

0 Standard Deviation
0 Latin Hypercube Points

Grouping Options

(• Group by Gridcell, then by Pooled C-Fl Function.
C Group by Pooled C-R Function, then by Gridcell.

Display Options —

Digits After Decimal Point:

E lements in Preview: 25

Preview

Column Row Endpoint Group Start Age Mean

Standard Deviation Percentile 5 Percentile 15 Percentile 25 Percentile 35

Mortality

1,129

291

613

808

921

1,011

Mortality

498

172

195

308

374

428

Mortality

Done

8.3 Raw Incidence - from CFGR file

The Raw Incidence Results report gives you the opportunity to examine the results of each
health impact function at the grid-cell level, county, state, or national level. It is based on
the Configuration Results file (with the *.cfgr extension), and is otherwise identical to the
Incidence Results Report generated from the Aggregation, Pooling, and Valuation results
file (with the *.apvr extension). BenMAP includes both versions to increase the reporting
flexibility for users. (See the section on Incidence Results reports for a description of the
options available for this report type.)

227

September 2008

-------
Reports

« Configuration Results Report

0BB

File

Column Selection
Grid Fields:

C-R Function Fields:

Result Fields:

0 Column
0 Row

DataSet
|!
-------
Reports

$ Configuration Results Report

File

Save Ctrl+S

Print Ctrl+P

v Column
0 Row

!:-R Function Fields:

~ DataSet

0 Endpoint Group

~ Endpoint

~ Pollutant

a ¦

~ Seasonal Metric

~ Metric Statistic
0 Author

D Year

~ Location

~ Other Pollutants
D Qualifier

~ Reference ~ P2Beta

~ Race ~ A

~ Ethnicity ~ NarneA

~ Gender ~ B

ID Start Age ~ NarneB

~ End Age ~ C

~ Function ~ NanneC
C Incidence DataSet

D Prevalence DataSet

~ Beta

~ DistBeta

~ PIBeta

Result Fields:

& Population
0 Delta
0 Mean
0 Baseline
0 Percent of Baseline
~ Standard Deviation
D Variance

0 Latin Hypercube Points

Grouping Options

(* Group by Gridcell, then by C-R function.
Group by C-R function, then by Gridcell.

Display Options
Digits After Decimal Point: 1

Advanced Options

Population Weighted Deltas: |~

Elements in Preview: |25

Aggregation Level: |Nation

Preview

-3

olumn Row Endpoint Group Author

Population

Delta

Mean

Baseline

Mortality

Laden et al.

228,054,695.9

1.3

45,723.3

2,725,115.9

Mortality

Pope et al.

204,949,580.8

1.3

20,242.3

2,702,189.8

Mortality

Pope et al.

204,949,580.S

1.3

20,160.7

2,702,169.8

Done

8.4 Audit Trail Reports - from all BenMAP files

The Audit Trail Reports provide a summary of the assumptions underlying the various
parts of the analysis. You may generate an audit trail with any of the file types used in
BenMAP: Air Quality Grids (with the ".aqg" extension), Incidence Configurations (with
the ".cfg" extension), Configuration Results (with the ".cfgr" extension), Aggregation,
Pooling, and Valuation Configurations (with the ".apv" extension), and Aggregation,
Pooling, and Valuation Results (with the ".apvr" extension). The report itself has a tree
structure that lets you easily find the information that you are seeking. Below is an
example of an Audit Trail Report for a Aggregation, Pooling, and Valuation Results file.

Note that each successive step in an analysis contains a summary of its assumptions, and
those of each previous step in the analysis. For example, in the below report the
assumptions of the Configuration Results file used to generate the APV Results are present
in the Configuration Results node. Similarly, the assumptions of both the baseline and
control air quality grids are present under the Configuration Results node.

229

September 2008

-------
Reports

« Audit Trail Report

^CAPrograrn FilesVBenMAP 3.0\Configuration ResultsMJser Manual Configuration Results\PM2.5 Wizard.apvrB

El Configuration Results: C:\Prograrn Files\BenMAP 3.0\Configuration ResultsMJser Manual Configuration Results\PM2.5 Wizard.cfgr

Latin Hypercube Points: 10
Population Dataset: United States Census - County
Year: 2020
Threshold: 0
0 Grid Definition
El Selected Studies

El Baseline Air Quality Grid: C:\Prograrn Files\BenMAP 3.0V\ir Quality GridsMJser Manual Air Quality Grids\Baseline PM25 2000 Closi
E Control Air Quality Grid: C:\Progrann Files\BenMAP 3.0\Air Quality GridsMJser Manual Air Quality Grids\Control PM25 2000 Closestl
El Advanced

ED I ncidence Pooling Windows

El Incidence Pooling Window: Basic Functions
El Incidence Pooling Window: AMI IX Calcs
El Incidence Pooling Window: AMI Summed Incidence
E) Incidence Pooling Window: Expert Functions
El Incidence Pooling Window: Mort Alt Thresholds
El Incidence Pooling Window: HA Summed Incidence

B- Hospital Admissions, Respiratory [Pooling Method: Sum (Dependent)]

0, Sheppard, HA, Asthma, Adjusted Coefficient With 10 ug Threshold, Seattle, WA, 64, 2003,, Sheppard, L. Ambient Air F
18, Moolgavkar, HA, Chronic Lung Disease (less Asthma), Adjusted Coefficient With 10 ug Threshold, Los Angeles, CA, B
- 65 [Pooling Method: Random I Fined Effects] [Advanced Pooling Method: Round Weights to Two Digits]

Moolgavkar, HA, Chronic Lung Disease, Adjusted Coefficient With 10 ug Threshold, Los Angeles, CA, S3, 2003,, Mo>
+ I to [Weight: 0.01, Mean: 6,457.16, StdDev: 3,686.65] [Pooling Method: Sum (Dependent)]

*¦' " -—— mU —Hj|'" ' " r' " jjjj-

Export I OK

8.4.1 Export Audit Trail

Note that you can export Audit Trail as a text file . Each branch in the tree structure will
be converted to a tab in the exported file, allowing for easy viewing in Excel, WordPad,
and a variety of other programs. Simply click on the Export button, name the file, and
click Save.

Note that by default BenMAP will export the file to the Reports folder. Carefully name the
file that you are generating so that you will recognize it in the future!

230

September 2008

-------
Reports

Save As

Save in: \B Reports

"3 «- ®

My Recent
Documents

Desktop

/^Sl

My Documents

My Computer

My Network
Places

File name:
S ave as type:

PM2.5 2000 "lOPctRollback. - APVR audit trail

[7 ext Files (Ktxt)

H
~3

Save

Cancel

8.5 Questions Regarding Creating Reports

>When creating reports from *.cfgr and *.apvr files, why do some of the variables that I
have checked appear as blanks?

When results are pooled together, some of the identifying information for individual health
impact functions gets lost. For example, when pooling together endpoints within the same
endpoint group, such as "HA, Pneumonia" and "HA, Chronic Lung Disease" (both within
"Hospital Admissions, Respiratory"), there is no longer a unique endpoint name for the pooled
result. So, BenMAP would leave the endpoint name blank.

>How do I export my results?

Identify the type of report that you want to create, then refer to the sections in this
chapter on exporting reports.

>How do I determine what the Column and Row refer to?

The Column and Row are variables designed to uniquely identify each grid cell in the
grid definition. In the case of the County grid definition, the Column refers to the state

231

September 2008

-------
Reports

FIPS code and the row refers to the county FIPS code. One way to get a good sense of
the Column and Row variables is to create a map (discussed in the next chapter) and
then view whether particular Column and Row variables occur in the map.

>In the report step, why is it that when I click the Done button the report is not
saved?

You have to go to File menu and choose Save.

232

September 2008

-------
GIS/Mapping

CHAPTER 9
GIS/Mapping

In this chapter...

>Learn about BenMAP's mapping functions.
>Map configuration inputs like air quality grids.
>Map incidence and valuation results.

>Map different variables and modify the map display.

233

September 2008

-------
GIS/Mapping

Mapping can be quite useful to check the quality of your work, and, of course, mapping
provides a useful platform for presenting your results. BenMAP features powerful,
integrated mapping capabilities which you can access at several points in the model. The
main Mapping/ GIS tool, available via the Tools drop-down menu in the main window,
allows you to map various types of files and data associated with an analysis, including
air quality grid files, monitor data, population data, and valuation results.

In addition, at several points from within the program, you can view data being used in an
analysis. You can view maps when filtering monitor data, creating air quality grids, and
when creating configuration files (with the *.cfg extension). In particular, when creating a
configuration file, you have the option to map your baseline and control air quality grids,
as well as to map the difference between the two (the "delta"), which is used in health
impact functions. The "delta" is typically a key input to your calculations, and this is the
only way to map it — mapping of the delta is not available from the Tools menu.

9.1 Overview of Mapping

You can access the some of the mapping capabilities in BenMAP by going to the Tools
drop-down menu, and choosing the GIS/Mapping.

234

September 2008

-------
GIS/Mapping

3 ©IX

Help

Air Quality Grid Aggregation
Model File Concatenated
Database Export
Database Import
Export Air Quality Grid

GIS/Mapping

Modify Setup
Neighbor File Creator
One-Step Setup

Use BenMAP: Which Analysis Meets your Needs?

uality data
ol to apply
te a report.

Custom Analysis

Step 1 — Import ail quality data

Air Quality Grid Creation

Step 2 - Set custom parameters

Incidence Estimation

Step 3 — Use results from Step 2
to set custom parameters

Pooling. Aggregation
and Valuation

Step 4 — Run report

Report

Active Setup:

United States

In addition to the Tools menu, BenMAP provides several places where you can map the
input data that you are using for a particular analysis. Below, we provide examples of
both types of mapping. First, we di scuss the display options and taskbar buttons which
can be used in both cases.

9.1.1 Display Options

After choosing the file or data type that you want to map, you must select the variable you
want to map, and how you want it displayed. Double-click on the layer name on the
left-hand side of the window. A small box will appear with the Display Options for the
layer.

235

September 2008

-------
GIS/Mapping

$ Display Options [T]fb]fx~]

¦ i

Variable: ^

Start Color: | |

Min Value: 0.00

End Color: [ |

Max Value: 10.00

Default Color: j |

Grid Outline:

Decimal Digits: 12

Cancel OK

This box allows you to set the following options:

Variable. The drop-down menu lists all of the variables contained in the layer you chose.
Select the variable that you want to view. Note that each layer can only show one variable
at a time.

Start Color and End Color. These are the colors that represent the gradations in the
selected variable. BenMAP uses 10 equal-sized increments for the variable, with a
gradual transition between the Start Color and the End Color. To change either color,
click on the colored square and select a new color.

Default Color. This is the color is used for values that fall outside of the range of the Min
Value and the Max Value.

Min Value and the Max Value. These options define the range of the selected variable
and are automatically set to the minimum and maximum of the variable. However, you
may wish to enter other minimum and maximum values, such as in the case where there is
one outlier that is dominating the color scale. For example, some urban grid-cells such as
those in Los Angeles have extremely large values. Since BenMAP creates 10 equal-sized
increments between the minimum and the maximum, it is not uncommon to have most of
the grid values in just a few of the increments. If you change the minimum and/or
maximum values to exclude outliers from the color scale, the outliers will then appear in
the Default color.

Decimal Digits. This specifies the number of digits used in displaying the results. The
default is two decimal points. However, for variables with values much smaller than one,
you will want to increase the number of digits that you display.

Grid Outline. This appears as a fine white line around the border of each grid cell. For
some grid types, this outline can make the map too complicated, and you may want to
uncheck this option. An alternative is to use the Reference Layer drop-down window
(described in the Taskbar Buttons section), which lets you add a blue outline for the
nation, states, or counties.

236

September 2008

-------
GIS/Mapping

Note that if you have specified point data, such as monitors, the Display Options window
includes the ability to set the size of the points on the map. The Start Size, End Size, and
Default Size have a default value of 75,150, and 120, respectively, which you can then
edit, if desired. For example, if you want smaller values to be all the same size, you might
set all of the values to 100.

$ Display Options (]^|(n](xj

Variable:

QuarterlyMea

Start Size:

Start Color:

Min Value: 2.75

End Size:

150

¦v

End Color:

Max Value: 28.33

Default Size:

|l 20

Default Color:

Decimal Digits: |2

Cancel

9.1.2 Taskbar Buttons

There are a number of standard buttons used in most map viewing programs which you
can use to navigate and customize the map view.

Open a file. Use this to open maps for viewing.

y Save active layer to file. Creates a shape file for use in other map-viewing
programs. The active layer is the topmost visible layer.

Zoom to full extent. Allows you to view the whole map that you are
viewing.

Increase zoom and Decrease zoom. Allows you to zoom in and
out.

Select a region for zooming. Allows you to select a region to view.

237

September 2008

-------
GIS/Mapping

Drag mode. Allows you to manually move the map by clicking and
dragging.

© Click to display info for the cell under the mouse. Allows you to display
information (all the variable values) for individual cells or points by clicking
on them.

Q Build Query. Allows you to view grid-cells that satisfy certain criteria.
Hitting the Execute button will produce a map of the cells that meet the
criteria that you have specified. Below is an example of how you might use
this function.

£ Layer Statistics. Provides information about the active layer. In the Fields
section simply choose the variable of interest, and BenMAP will display
statistics and sample values for that variable.

238

September 2008

-------
GIS/Mapping

9 Layer Statistics

Fields:

Statistics:

COL

ROW

iBronch

POPO

DELTAO

Myo1824

P0P1

DELTA1

, *1 ' jj

Mean:
Min:
Max:
StdDev:
Sunn:

22S.63

131.81

484.84

144.78

1148.14

Count: 5

PM2.5 UityOne County 25 Pet Rollback 2003 t

Sample Values:

165.61
131.81
164.36
201.53
484.84

~ K

The drop-down menu provides alternative projections for displaying the data, with the
default set to Albers Equal Area Conic.

9.2 Mapping from the Tools Menu

To access the main mapping capabilities within BenMAP, go to the Tools drop-down
menu, and choose the GIS/Mapping option. A blank window will appear, with buttons at
the top for managing files and navigating the map. To see the name of each button, simply
hold the cursor over it.

Use the Open a file button in the top-left corner to choose the file (or other type of data)
that you want to view. All of the options are straightforward, though some require a few
more steps than others. Below we give some step-by-step examples. See Table 9-1 for a
brief description of each type.

239

September 2008

-------
GIS/Mapping

« BenMAP GIS

0©B

Air Quality Grid (*.aqg)

APV Configuration Results (*.apvr) ~

Configuration Results (*.cfgr)

Modeling Data

Monitors

Population

Table 9-1. Description of Mapping File Types

File Type Description Source File Used

Air Quality Grid Annual summary values (e.g., daily average, daily Air quality grids created by
maximum, or other metric where available) within BenMAP (*.aqg)

each grid cell.

APV Presents the full range of incidence, valuation and Aggregation, Pooling and

Configuration QALY results. The results are available at the level Valuation Results files (*.apvr)
Results of aggregation of the air quality grids. In addition

aggregated results are available, as well as both
aggregated and pooled results.

Configuration Presents the incidence estimate, change in air Configuration Results files

Results pollution, and population used in the health impact (* cfgr)

calculation.

Modeling Data
Monitors

Annual summary values (average, median or other
metric where available) for each monitor.

Select monitors from the
BenMAP library or use your
own by loading a file.

240

September 2008

-------
GIS/Mapping

File Type Description Source File Used

Population Total population, and cross-tabulations of age and Internal BenMAP population
gender with race/hispanic origin within each grid files and projections, based on
cell (for the Year and GridType you specify). Census block-level data.

Each map that you select and load into the GIS viewer will appear on the left-hand side of
the window as a layer. All of the options to create maps are straightforward, though some
require a few more steps than others. Below we give some step-by-step examples of each
type.

9.2.1 Air Quality Grids

Air quality grid maps present annual summary values of air pollution metrics (e.g., daily
average, daily maximum, or other metric where available) within each grid cell. After
choosing the air quality grid that you want to map, double-click on the layer name on the
left-hand side of the window. A small box will appear with the Display Options for the
layer. Choose the metric that you want to map. When mapping an air quality grid with
county-level PM2.5 values, the available default metrics are D24HourMean and
OuarterlyMecm.

* Display Options [^lfn~l|x|

Variable:

CUL
ROW

D24HourMean

Start Color:

QuarterlyMean

| j i_i.

End Color:

| Max Value: Cl.OO

Default Color:

Grid Outline: |7 Decimal
Cancel

Digits: [2 ~^j
OK |

If you keep the default display options, the maps can often look a little complicated or
busy. For example, when mapping county-level level PM2.5 levels with the default
display options you might get a map looking like the following:

241

September 2008

-------
GIS/Mapping

« BenMAP GIS

B0B

&
Layers

£ | Albers Equal Area Conic 3 |BJjyy.ujfc[yi TJ

V Baseline PM25 200C
2.75-4.8G
4.88-6.98
6.98-9.09
9.03-11.20
11.20-13.31
13.31-15.43
15.43-17.54
17.54-19.65
19.65-21.77
21.77-23.88

Unchecking the Grid Outline box, reducing the Decimal Digits from 2 to 0, and using a
State Reference Layer can simplify the map, as follows:

242

September 2008

-------
GIS/Mapping

« BenMAP GIS

"Layers—

t? Baseline PM25 200C

2-4

4-6

6-9

9-11

11-13

13-15

15-17

17-20

20-22

22-24

#KK|g|eH®H

Albers Equal Area Conic

9.2.2 APV Configuration Results

When mapping APV Configuration results, you can generate nine different types of maps:
(1) Incidence, (2) Aggregated Incidence, (3) Pooled Incidence, (4) Valuation, (5)
Aggregated Valuation, (6) Pooled Valuation, and (7) OALYResults, (8) Aggregated OALY
Results, and (9) Pooled OALY Results. All of the options use the information contained in
the Aggregation, Pooling and Valuation Results files (*.apvr), which you can create
using the Aggregation, Pooling, and Valuation button on the main window.

243

September 2008

-------
GIS/Mapping

« BenMAP GIS

0@B

Air Quality Grid (*.aqg)

APV Configuration Results (*.apvr) >

Incidence Results

Configuration Results (*.cfgr)

Modeling Data

Monitors

Population

Aggregated Incidence Results
Pooled Incidence Results
Valuation Results
Aggregated Valuation Results
Pooled Valuation Results
QALY Results
Aggregated QALY Results
Pooled QALY Results

After you have chosen the results you want to map, a window will appear with identifiers
for each of the variables. On the right-hand side is the GIS Field Name. Mapping
programs typically have a default length of 10 characters for variable names, so it is
necessary to specify your own names or to use the default name given in the GIS Field
Name column.

244

September 2008

-------
GIS/Mapping

Edit GIS Field Names

DataSet

Endpoint Group

Endpoint

Pollutant

M etric

Seasonal Met

Gis Field Name

—

BenMAP Health Imp

Mortality

Mortality, All Cause

PM2.5

D24HourMean

QuarterlyMear

FiesultO

BenMAP Health Imp

Mortality

Mortality, All Cause

PM2.5

D24HourMean

QuarterlyMear

Resultl

BenMAP Health Imp

Mortality

Mortality, All Cause

PM2.5

D24HourMean

QuarterlyMear

Result2

BenMAP Health Imp

Mortality

Mortality, All Cause

PM2.5

D24HourMean

QuarterlyMear

Result3

BenMAP Health Imp

Mortality

Mortality, All Cause

PM2.5

D24HourMean

QuarterlyMear

Result4

BenMAP Health Imp

Mortality

Mortality, All Cause

PM2.5

D24HourMean

QuarterlyMear

Result5

BenMAP Health Imp

Mortality

Mortality, All Cause

PM2.5

D24HourMean

QuarterlyMear

ResultG

BenMAP Health Imp

Mortality

Mortality, All Cause

PM2.5

D24HourMean

QuarterlyMear

Result7

BenMAP Health Imp

Chronic Bronchitis

PM2.5

D24HourMean

QuarterlyMear

Results

BenMAP Health Imp

Acute Bronchitis

PM2.5

D24HourMean

QuarterlyMear

Results

BenMAP Health Imp

Acute Myocardial In

PM2.5

D24HourMean

Resultl 0

BenMAP Health Imp

Acute Myocardial In

PM2.5

D24HourMean

Resultl 1

n _..i j a n ii _ _iii. i

a L-jk J j:-¦ i..i

f> .l_ i j i:_i i..i

ni j n c

r. -"i a i i i j

n iii ^

Note that if you used more than one pooling window (described in Aggregation, Pooling,
and Valuation chapter), then BenMAP will create one layer for each pooling window.

Also, note that the "aggregated" results are aggregated to the level specified in the APVR
file. If results were aggregated to the national level, the US map will only show one color.

For each of the three valuation options (Valuation, Aggregated Valuation, and Pooled
Valuation) and for each of the three QALY options (OALY, Aggregated OALY, and Pooled
OALY), a second window appears that allows you to specify variables that you want to add
together. (If you have selected Incidence, Pooled Incidence, or Aggregated Incidence,
BenMAP will go directly to loading the new layer.)

245

September 2008

-------
GIS/Mapping

$ Valuation Sums Layer Q(n](x^

Click on the Add Sum button. Then in the far-right column, Include in Total, check the
variables that you want to include in the total. In the lower-left corner, give a name (no
longer than 10 characters) for the total in the Valuation Sum Identifier, and then choose
whether to use a Dependent or Independent sum. If you choose the latter, then you also
need to choose the number of Monte Carlo draws. To finish, click OK. You may repeat
this step as many times as desired.

246

September 2008

-------
GIS/Mapping

$ Add Valuation Sum U nj(x]

Pooling Window

Endpoint Group

Start Age

Author

Endpoint

Include in Total

-J

Mortality

Woodruff et al.

Mortality

Woodruff et al.

Chronic Bronchitis

Acute Myocardial Inl

-------
GIS/Mapping

s Display Options

Variable:

COL

ROW

Ladenl0

Start Color:

POPO
DELTAO

End Color:

Pope75
P0P1

Default Color:

DELTA1

Grid Outline: W Decimal Digits: 12

Cancel OK

$ BenMAP GIS

After choosing the rest of your di splay options, you can then view your map. Note that it
is fairly common for maps of Configuration Results to be dominated by a few regions,
such as southern California, with many other parts of the country indistinguishable from
each other.

12?

Layers
[7 PM 2.5 Wizard - Nati
0-193
198-395
395-593
593-790
790-988
988-1186
1186-1383
1383-1581
1581-1779
1779-1976

By changing the Max Value, End Color, and Default Color, as follows, you can develop
a map that provides more information regarding impacts in other parts of the country,
aside from those with the largest impacts.

248

September 2008

-------
GIS/Mapping

$ Display Options Q| nJIK

1 1

Variable: IBSEIfllil

Start Color: Q

Min Value:

End Color: |
Default Color: J

Max Value:

300

Grid Outline: f~

Decimal Digits: 2

Cancel

In this example, the areas with the largest impacts appear in bright red (i.e., southern
California, Phoenix, Houston, Chicago, and a handful of other locations). The areas with
the next largest impacts appear in dark green.

* BenMAP GIS

~0®

Albers Equal Area Conic »

v PM2.5 Wizard - Nati
0-30
30-G0
60-90
90-120
120-150
150-180
180-210
210-240
240-270
270-300

You might also use the Create a Query button to focus on your results of interest. For
example, if wanted to focus on those areas (in this instance counties) with at least one
hundred deaths as calculated with our "Laden" function, then we would first click the
Create a Query button and build the following query.

249

September 2008

-------
GIS/Mapping

$ Build Query [-J

Fields:

COL
ROW
Ladenl0
POF'O
DELTAO

n tc

Sample Values:

Not

And

13
50
G
5

Ladenl 0 >= 1 UC|

Cancel

Execute

Clicking OK will complete the query.

* BenMAP GIS

&\a\&\ »KKIo,|e|o|ol g |

Layers

Albers Equal Area Conic »

PM2.5 Wizard - Nati
0-30
30-G0
60-90
90-120
120-150
150-1 SO
180-210
210-240
240-270
270-300

Running the query again on this hypothetical, we might to focus on just those areas with at
least 200 deaths. The exact type of map that you want generate will of course depend on
your particular set of results and analytical goals. The point is that mapping capabilities in
BenMAP can provide some interesting, easy-to-use options.

250 September 2008

-------
GIS/Mapping

9.2.4 Modeling Data

You can map any modeling data that you planning to use to create an air quality grid (*.
aqg file), as long as you have a grid definition and a pollutant definition that matches your
data. Choose Modeling Data from the Open a file drop-down menu on the BenMAP GIS
window. This will bring up the Select Model Data to Map window. Click the Browse
button to find the file you want to map, and then choose the appropriate Grid Type and
Pollutant from the drop-down menus.

$ Select Model Data to Map |-_j[n]| x_|

Model File: er Manual\Baseline Modeled PM.xls

Browse

Grid Type: |CityOne Grid t

Pollutant: | PM 2.5

Cancel

Click OK, and then you will be ready to map. Double-click on the layer in the left panel,
and this will bring up the Display Options window, so you can choose the variable you
want to map and how you want it to look. In this example, the Grid Outline has been
unchecked and the Decimal Digits has been set to 0.

s Display Options [T][~"][><"]

Variable: |QuarterlyMean

-I

Start Color: | |

Min Value: 10.42

End Color: | |

Max Value: 17.16

Default Color: | |

Grid Outline: \~

Decimal Digits: |0

Cancel OK

To add context to the map (because the Grid Outline is unchecked), you will likely want to
use a Reference Layer from the drop-down list.

251

September 2008

-------
GIS/Mapping

Model Data (Elaselin

11-12

12-12

12-13

13-14

14-14

14-15

15-16

16-16

16-17

9.2.5 Monitoring Data

You can map monitor locations and concentration levels using the Monitors option. Click
the Open a file button and select Monitors from the drop-down menu. The Select
Monitors window will appear in which you select the Pollutant and then load monitor
data for that pollutant from either an existing library (in which case you choose the
Library tab) or from a dataset outside of BenMAP (in which case you would use either
the Database Columns, Database Rows, or Text File tabs, depending on the file format).
The same monitor data can be used many times, so it is generally a good idea to create a
library — see the Loading Data chapter for how to do this.

252

September 2008

-------
GIS/Mapping

$ Select Monitors

EBB

Pollutant:

| Ozone

Library Database, Columns | Database, Rows |

Text File

Monitor DataSet:

I EPA Standard Monitors^^^^^^^^^^^H

Monitor Library Year:

12000

Advanced

Cancel

If you map data from the BenMAP library, you need to specify the Monitor DataSet and
the Monitor Library Year. In addition, you can change the default monitor filter options.
Generally speaking, this is not necessary. However, if desired, click the Advanced button.
This will bring up Filter Monitors window, where you can make whatever adjustments
are deemed necessary to the default options. (Note that the filtering is hardcoded into
BenMAP, so whatever filter options you choose will not persist to any future mapping or
use of monitor data.)

253

September 2008

-------
GIS/Mapping

« Filter Monitors

M@(K

1. If you wish to include individual monitors by ID, enter the IDs
here, separated by commas. If you do not enter any
monitors here, all monitors which meet the rest of the
selection criteria will be included:

2. If you wish to exclude individual monitors by ID, enter the IDs
here, separated by commas:

3. If you wish to restrict monitors to certain states or areas, enter
comma separated state abbreviations, and minimum and/or
maximum latitudes and longitudes:

States:

Minimum Longitude: -130 Minimum Latitude: [20
Maximum Longitude: -65 Maximum Latitude: 55

4. Select maximum POC code to include and the PGC
preference:

Maximum POC: [4 POC Preference Order: ] 1,2,3,4

5. Select the Methods you wish to include:

G Select the parameters specific to the pollutant:

Go!

~ 003
V! 011
0 014
0 019

0 020

v 047
0 053
0 056
0 078

0 087

0 091
0 103
0 112

Start Hour:
End Hour:

Start Date:

Number of Valid Hours:

End Date: September 30 -rj [50

Percent of Valid Days:

Cancel OK

Click Go! after you have made your selections. This will bring you back to the Select
Monitors window. Click OK again, and BenMAP will generate a monitor layer that you
can view.

254

September 2008

-------
GIS/Mapping

« BenMAP GIS

00®

& H\&\ 9

Q 0 O E Albers Equal Area Conic

Layers

. : ¦ i

r.
¦J

17 Ozone, 2000

In this map, each red square is a monitor location. To see the monitor values displayed
with colors varying by the level of the monitor, double-click on the layer on the left side of
the map, and follow the steps outlined above for setting the display options. Values shown
using the Monitors mapping option are annual average values (|ig/m3 for PM and ppb for
ozone).

255

September 2008

-------
GIS/Mapping

s BenMAP GIS

~HIS

Layers

P Ozone.
23.03-
28.11
33.20-
38.28-
43.36-
43.45-
53.53-
58.61-
63.63-1
68.78-

2000
28.11
33.20
38.28
43.36
48.45
53.53
58.61
63.69
68.78
73.86

9.2.6 Population Data

You can map the population data that you have already loaded into BenMAP. Choose
Population from the Open a file drop-down menu on the BenMAP GIS window. This
will bring up the Select Population Data window. Choose the Population DataSet that
you want to map and the particular year from the Population Year from the drop-down
menus.

$ Select Populate

Population DataSet:

|CityOne Tract Population

Population Year:

12005)

Cancel

256

September 2008

-------
GIS/Mapping

been unchecked and the Decimal Digits has been set to 0.

« Display Options |-J| nj|xj

Variable: SKmimillTRI

Start Color: Q

Min Value:

E nd Color: J

Max Value:

241

Default Color: J

Grid Outline:

Decimal Digits: |2 ^

Cancel

* BenMAP GIS

To add context to the map (because the Grid Outline is unchecked), you will likely want to
use a Reference Layer from the drop-down list.

&\a\&\ #KlQ.[qre|o[o

Layers

; Equal Area Conic I It I

ounties

1^ Population T raots 2C
0-24
24-48
43-72
72-96
96-121
121-145
145-169
169-193
193-217
| 217-241

9.2.7 Mapping Multiple Layers of Data

It is possible to map multiple layers simultaneously. The layer that you have opened most

257

September 2008

-------
GIS/Mapping

recently appears on the top of the list, and in the map its values lie on top of the other
layers. By right-clicking on any given layer, you can move its position within the list
(select Move Up or Move Down). For example in the following screenshot, we move the
Control Grid layer to the top by right-clicking on the Control Grid layer and choose
Move Up (or you can right-click on the Delta layer and choose Move Down).

Move Up

* BenMAP GIS

Move Down

Albers Equal Area Conic

|y|#|

Layers

P Delta
0.00-0.24
0.24-0.48
0.48-0.72
0.72-0.96

L0.9S-1.20
1.20-1.43
1.43-1.67
1.67-1.91
1.91-2.15
2.15-2.39

_ A

State

Control Gric'
2-4
4-6

By checking the box to the left of the layer name, you can turn a layer's visibility off and
on. For instance, if you want to see the second layer in the list, simply un-check the box
next to the first layer in the list. The second layer will then be on top and be visible, as in
the example below.

258

September 2008

-------
GIS/Mapping

« BenMAP GIS

„ *1

Layers

Delta

0.00-0.24

0.24-0.48

0.48-0.72

0.72-0.96

0.86-1.20

1.20-1.43

1.43-1.67

1.67-1.91

1 91-2.15

2.15-2.39

1? Control Grid

2-4

4-6

6-8

8-10

10-12

12-14

Q|n[0

Albers Equal Area Conic

9.3 Viewing Maps in a BenMAP Analysis

In addition to the Tools menu, BenMAP provides several places where you can map the

input data that you are using for a particular analysis. We provide examples of these

mapping options below.

• Air Quality Deltas. When creating your Configuration (*.cfg file), you can map the
baseline minus the control scenario to get a sense of where the major impacts are
occurring in your analysis.

• Air Quality Grids. When creating an air quality grid (*.aqg file), you can map what the
air quality grid will look like. And in the case of air quality grids based on monitor data,
you can map the monitors as well.

• Air Quality Monitors from the Advanced Monitor Filter. When filtering monitors, you
can map the monitors to see the impact of the filtering.

• Monitor Rollback Inputs & Outputs. When creating a monitor rollback air quality grid
(*.aqg file), you can map both the baseline and control air quality grids and the monitor
data used.

259

September 2008

-------
GIS/Mapping

9.3.1 Mapping Air Quality Deltas

When developing a Configuration file (with the *.cfg extension), you can map the baseline
and control air quality grids, as well as the difference between the two. In the
Configuration Settings window, after choosing the grids that you want to use, click on the
Map Grids button.

$ Configuration Settings

Select Air Quality Grids
Baseline File:

|les\BenMAP 3.0\Air Quality GridsMJser Manual Air Quality Grids\Baseline PM25 2000 ClosestMonitor County.aqg Open

Create

Control File:

|.0\Air Quality GridsMJser Manual Air Quality Grids\Control PM25 2000 ClosestMonitor County 10pctRollback.aqg| Open | Create

Map Grids

"Settings—

Latin Hypercube Points:

|10

Population DataSet:

|United States Census
Run In Point Mode:

Threshold:

Population Year:
] [2520

Cancel

This will generate layers with the delta (baseline minus control), baseline, and control
values for each grid cell. You can then double-click on the layers in the left panel to set
your display settings.

260

September 2008

-------
GIS/Mapping

* BenMAP GIS

QBE

y ^

I "Layers—

©Jf? o

T Albers Equal Area Conic jJ

w* Delta

W Control Grid
1^ Baseline Grid

Close !

9.3.2 Mapping Monitor Direct Air Quality Grids

When generating air quality grids there is an option to map the grid and the data used to
create it. For example, when generating a Monitor Direct grid, click on the Map button on
the Monitor Direct Settings page.

261

September 2008

-------
GIS/Mapping

$ Monitor Direct Settings Q|n^|(x^

Grid Type:

Interpolation Method

| County

C Closest Monitor

Pollutant:

(* Voronoi Neighborhood Averaging

|PM2.5

( Fixed Radius (km.) I

Library 1 Database, Columns

Database, Rows | Text File |

Monitor DataSet:

EPA Standard Monitors

Monitor Library Year:

[2000

Advanced

Map

Cancel

Go!

BenMAP will then generate the grid, generate a map with both the monitors and the
monitor data interpolated to the grid cells. You can then use the display options to
generate the map you desire.

262

September 2008

-------
GIS/Mapping

s BenMAP GIS

\a\* e|o|o| ^

Layers

SUB

17 Monitors

17 Air quality grid

Albers Equal Area Conic -

9.3.3 Mapping Air Quality Monitors from the Advanced Monitor Filter

You can access mapping through the advanced monitor filter when generating air quality
grids. Click on the Create Air Quality Grids button and choose either Monitor Direct,
Monitor and Model Relative, or Monitor Rollback. On the Settings page, click on the
Advanced button, which brings up the Advanced Options page.

263

September 2008

-------
GIS/Mapping

$ Advanced Options |

Maximum Neighbor Distance (in km):
[-No Maximum Distance -

Maximum Relative Neighbor Distance:
-No Maximum Relative Distance -

Weighting Approach
Inverse Distance

Inverse Distance Squared

Get Closest if None within Radius

Custom Monitor Filtering
Cancel OK

Then click on the Custom Monitor Filtering button, which will bring up the Filter
Monitors window.

$ Filter Monitors

If you wish to include individual monitors by ID, enter the IDs
here, separated by commas. If you do not enter any
monitors here, all monitors which meet the rest of the
selection criteria will be included:

2. If you wish to exclude individual monitors by ID, enter the IDs
here, separated by commas:

If you wish to restrict monitors to certain states or areas, enter
comma separated state abbreviations, and minimum arid/or
maximum latitudes arid longitudes:

States:
Minimum Longitude: -130
Maximum Longitude: [-65

Minimum Latitude: 120
Maximum Latitude: 55

Select maximum PQC code to include and the ROC
preference:

Maximum ROC: |4 ROC Preference Order: 11,2,3,4

5. Select the Methods you wish to include:

0 116

0 120

711

0 117

701

~ 712

0 118

702

~ 713

0119

704

~ 722

< i

6 Select the parameters specific to the pollutant:

Number of Valid Observations Required Per Quarter:

I11 23

Data Types To Use:

(• Local <"* Standard C Both

Preferred Type:

(• Local C Standard

Output Type:

(• Local C Standard

Export

Map

Go!

Cancel OK

Choose your filtering options and click Go! Then click the Map button. An initial map
appears with each monitor location identified by a red square. To immediately provide

264 September 2008

-------
GIS/Mapping

some context, you may want to choose, say, the States reference layer.

* BenMAP GIS

Albers Equal Area Conic * I TI

Layers

Double-click on the layer to the left of the map. Choose the display options that you
would like, and then click OK.

« Display Options [J_n ][x J

Variable:

QuarterlvMeanl

Start Size:

100

Start Color:

Min Value:

2.0

End Size:

100

End Color:

Max Value:

30.0

Default Size:

100

Default Color:

Decimal Digits:

I1 U

Cancel

A map with your display options will then appear. You may then use the various mapping
options available, such as zooming in to an area of interest, getting information on
particular monitors, querying the map to display monitors with certain characteristics, and
saving the map as a shapefile and viewing it in another map viewer.

265

September 2008

-------
GIS/Mapping

« BenMAP GIS

¦ |h|#I •KK|q,|oIoIo|s

Layers—

[7 Filtered Monitors
2.0-4.S
4.8-7.6
7.G-10.4
10.4-13.2
13.2-1 G.O
16.0-13.8
18.8-21.6
21.6-24.4
24.4-27.2
27.2-30.0

Albers Equal Area Conic -

Jw 1,pJ^i
1% ¦

(m m m

r. .

7 V"

¦ ¦

¦1

¦"l " ¦

¦ ¦
1 J

) m 1

S rl"j|

¦lL

¦
¦

¦ _

— 1 —Vl

LVT ¦ 1

¦ -j

9.3.4 Mapping Monitor Rollback Inputs & Outputs

When using the Monitor Rollback option, you can generate two types of maps. First, you
can map the inputs to the rollback - the monitor data and any spatial adjustment file that
you may have used. Second, you can map the inputs and at the same time map the grid
based on these inputs. Start to create a monitor rollback grid, and when you get to the
Monitor Rollback Settings: (3) Additional Grid Settings, make your selections for the
Select Interpolation Method, Select Scaling Method, Grid Type, and whether you want to
Make Baseline Grid (in addition to Cotnrol Grid).

266

September 2008

-------
GIS/Mapping

« Monitor Rollback Settings: (3) Additional Grid Settings (-J|H](x^

Select Interpolation Method

Select Scaling Method—

C Closest Monitor

(* None

(* Voronoi Neighborhood Averaging

C Fixed Radius (knn.) \

C Spatial Only

Grid Type: [bounty "*]

Adjustment File: | Browse

|7 Make Baseline Grid (in addition to Control Grid).

Advanced

Map

Back

Go!

In this example, BenMAP will produce a map with both the inputs, as well as the baseline
and control grids. (The baseline grid gets created because we checked the box for Make
Baseline Grid (in addition to Control Grid). Click on the Map button. This will bring up
the BenMAP GIS window.

267

September 2008

-------
GIS/Mapping

« BenMAP GIS

g* e|o|o| ^

Layers

Albers Equal Area Conic "

w* Rollback Grid

17 Baseline Grid

17 Rolled back Monitor:

7 Baseline Monitors

As before, you may use the various display options to generate the map that you desire.
Recall that the topmost layer lies on top of the other mapped layers. To change the
ordering of the layers, simply right-click on the layer that you want to move (up or down).

9.4 Questions Regarding Mapping

>Why is the Open a File menu button disabled?

This happened because you did not use the Tools button and choose Mapping GIS. When
viewing maps while generating air quality grids, filtering monitor data, and other activities
within BenMAP you do not have access to other types of maps. This is to avoid too many
competing activities.

>A11 of the mapped values have the same color. How do I avoid this?

This can happen when the values are extremely small and you have not specified a sufficient
number of decimal points. Go to the Display Options window, and change the Decimal Points.
This can also happen when one of the grid cells is an outlier, with either a very low or very
small value. You can go to the Display Options window, and change either the Min Value or
the Max Value. Finally, this can happen if you have mapped national data. In this instance, you
should expect all areas to have the same color, since there is only a single national number for
display, such as when mapping national results, or mapping incidence rates that do not vary by
region (e.g., MRAD incidence rate).

268

September 2008

-------
GIS/Mapping

>Can I map air quality for individual days?

No. BenMAP only maps annual averages. In the case of hourly metrics, such as the one-hour
daily maximum for ozone, BenMAP will map the average of the metric for the available days.

>Can I print maps in BenMAP?

No. BenMAP does not currently allow printing directly from the program. However, you can
export shapefiles, and then read these shapefiles into a program that does support printing, such
as ArcView.

>Why is my map all grey or all black?

You may need to select a variable or uncheck the grid outlines checkbox. You can do this by
double-clicking on the layer name.

>How do I clear a query?

When using a query, you limit the areas that get displayed. That is, after using a query, some
areas will be white and other areas will have colors that you have specified. If you want to be
able to display all areas simply run a query that includes all areas. For example, if you run a
query such as choose all areas with population greater than or equal to zero, then you will get a
complete set of results.

>What are the units?

The units that get displayed depend on the particular map that you are creating. For example, if
you are mapping PM2.5 levels, then the units will be in micrograms per meter cubed. If you are
mapping ozone, the units will be in parts per billion. If you are mapping mortality or some
other type of health effect, then the units will be numbers of cases. If you are mapping
valuation results, then the results will be in dollars.

>How do I change which layer is on top?

When mapping multiple layers, the top-most layer is the active layer. To change the position of
a layer, simply right-click on it, and you can move it up or down. This is discussed in the
section Mapping Multiple Layers of Data.

269

September 2008

-------
Tools Menu

CHAPTER 10
Tools Menu

In this chapter...

>-Learn about the options in the Tools menu.

270

September 2008

-------
Tools Menu

The Tools menu, available on the main BenMAP screen, provides access to a number of
specialized applications. We list them below and give a brief description. Note that other
sections of the manual have already covered several of them, such as GIS/Mapping, so we
merely list them here and point you to the appropriate section.

Help

Air Quality Grid Aggregation
Model File Concatenated
Database Export
Database Import
Export Air Quality Grid
GIS/Mapping
Modify Setup
Neighbor File Creator
One-Step Setup

Use BenMAP: Which Analysis Meets your Needs?

uality data
ol to apply
ite a report.

ffirutramyTsrld Crsalion

\ar

JUL

Preloaded
EPA parameters

Report

Active Setup:

Custom Analysis

Step 1 — Import air quality data

Air Quality Grid Creaiion

* ' *" i

Step 2 — Set custom parameters

Incidence Estimation

Step 3 — Use results from Step 2
to set custom parameters

Pooling, Aggregation
and Valuation

Step 4 — Run report

Report

.11

Air Quality Grid Aggregator. You can change an air quality grid based on one grid
definition to another grid definition, using a simple spatially-weighted average approach.
This option is described below.

Model File Concatenator. This is a highly specialized option that allows you to combine
air quality modeling data text files and produce a smaller, more efficient, binary file. This
option is described below.

271

September 2008

-------
Tools Menu

Database Export. Export entire databases (all Setups), individual Setups, and parts of
individual Setups (e.g. all GridDefinitions, or individual health impact Function DataSets).
This functionality can be used to share your data with other users, as well as to allow you
to move databases between computers. This is described in the Export Setups section of
the Loading Data chapter.

Database Import. Import entire Setups or parts of individual Setups. This option is
described in the Import Setups section of the Loading Data chapter.

Export Air Quality Grid. Generate a data file with all of the data in an air quality grid.
This option is described below.

GIS/Mapping. Map various types of files and data associated with an analysis, including
air quality grid files, monitor data, population data, and valuation results. See the
GIS/Mapping chapter for details.

Modify Setup. BenMAP encapsulates in a single dataset, called a "Setup," all of the data
needed to run analyses for a particular geographic area. Using the Modify Setup option,
you add to, modify, and delete setups. See the Loading Data chapter for details.

Neighbor File Creator. Identify the monitors and weights used in the interpolation
process when creating air quality grids. This option is described below.

One-Step Setup. Allows you to set the CFG and APV files, as well as currency year, for
the One-Step analysis. This option is described below.

10.1 Air Quality Grid Aggregator

Using the Air Quality Grid Aggregator you can change an air quality grid based on one
grid definition to another grid definition, using a simple spatially-weighted average
approach.

To start, choose Air Quality Grid Aggregator from the Tools drop-down menu. This will
bring up the Air Quality Grid Aggregator window. Click the Browse button to find the
air quality grid that you want to change and then use the drop-down menu in the
Aggregation Grid panel to identify the new grid definition that you want to use. (For
example, you might want to aggregate a county-level air quality gird to state-level.) Click
OK when done.

$ Air Quality Grid Aggregator Q[n][x^

Air Quality Grid:

|ds\Baseline PM25 2000 ClosestMonitor County, aqg Browse
Aggregation Grid:

[Hi t]

Cancel I OK

272

September 2008

-------
Tools Menu

* BenMAP GIS

EBB

This will bring up the Save Aggregated Air Quality Grid window, where you specify the
name of the file you are creating and its location. After the file is created, BenMAP will
take you back to the main BenMAP screen. You can then use the new file just as you
would any other air quality grid.

Below is an example of going from a county-level PM2.5 air quality grid to a state-level
PM2.5 air quality grid. The county-level file looks as follows:

| y|q*| #J

Layers 1

k Baseline PM25 200C

2-4

4-6

6-9

9-11

11-13

13-15

15-17

17-20

20-22

22-24

T Baseline PM25 200C

2-4

4-6

6-9

9-11

11-13

13-15

15.17

©s|jq e[o|o s

Albers Equal Area Conic f State

The state-level file looks as follows:

273

September 2008

-------
Tools Menu

« BenMAP GIS

Layers

17 Baseline PM25 200[
2-4
4-6
6-9
9-11
I 11-13
13-15
15-17
17-20
¦ 20-22
| 22-24

r Baseline PM25 200C
2-4
4-6
6-9
9-11
I 11-13
13-15

¦ 1 R.I 7

® e f? © £

Albers Equal Area Conic -

State

The scale in the left panel is set to the same values, so you can compare the two maps
more easily. One key thing to note is that there is more variation in the county-level file.
For example, in California the county-level file has colors ranging from yellow to dark
red, while in the state-level file, California is a solid orange color. This pattern is
expected, because BenMAP is just taking a simple spatially-weighted average of the data.

10.2 Model File Concatenator

The Model File Concatenator is a specialized tool that allows you to combine air quality
modeling data text files and produce a smaller, more efficient, binary file. For example,
ozone modeling files often are generated one day at a time and split between eastern and
western modeling domains, so the input modeling files to BenMAP might consist of
dozens of files. The Model File Concatenator combines these files into a single file that
can then be easily used in BenMAP.

You can use this tool if you have data for both a western and eastern domain. In case you
want to use this tool for a single modeling domain, then just fill in the options for the
western domain, and ignore the eastern domain.

To start, choose Model File Concatenator from the Tools drop-down menu, the Model
File Concatenator window will appear. There are a number of options that need to be
specified. We describe each below.

274

September 2008

-------
Tools Menu

Model File Concatenator

Input Files

(* jRow/Colunnni ' Column/Row

Column Delimiter
(* Whitespace C Comma

Convert model values from parts per million to parts per billion

Lower Upper
Bound Bound

Select Western Domain files

Columns |~
Rows: |~

Select Eastern Domain files:

Lower
B ourid

Columns |~
Rows: |"

Upper
Bound

Done

Go!

The Input Files options are Row Column and Column Row. The choice depends on how
the data files are formatted. Below is an example of what a typical datafile looks like. In
this example, the first column specifies the Row and the second column specified the
Column. (In other cases, this might be reversed.)

275

September 2008

-------
Tools Menu

e 2030bn

_J4a_12km_us12b.o3_hr_shift_LST.0616.txt -

WordPad

1 File Edit View Insert

Format Help

d & y

m r

| Courier New

j 110 v | | Western

"^| B /

inn

lii
SI

1 1

26.5623

25.8350

26.8762

27.3641

27.4579

27.6055

27.6513

27.4290

28-

1 2

26.8234

2 6.4892

27.2904

27.7655

28.0426

28.4414

28.4601

27.7336

1 3

26.6103

26.2705

27.1781

28.0081

28.5588

28.9647

28.6379

27.7269

1 4

27.6259

28.1685

28.8431

29.1286

29.4561

29.7432

29.53 63

28.6144

1 5

27.7258

28.3655

29.0252

29.2747

29.5688

29.8001

29.6565

28.7194

1 6

27.3433

28.4953

29.1180

29.3925

29.5763

29.8120

29.7781

28.8796

1 7

2 6.6847

28.1185

28.9395

29.4130

29.7871

30.0345

29.8349

29.1045

1 8

26.8281

26.8445

28.5588

29.4126

29.8215

30.0141

29.7688

29.0178

1 9

27.6705

28.2606

28.8573

29.2847

29.7062

30.0008

29.7188

28.9144

1 10

2 6.4482

27.0996

28.3808

29.2019

29.8162

29.8011

29.2660

28.4902

1 11

25.9659

27.1938

28.3777

29.2658

29.7222

29.6352

29.0991

28.3078

1 12

25.3104

26.2073

27.5506

28.5092

28.9848

28.9250

28.2751

27.6080

1 13

2 6.5879

27.5554

28.8133

29.5936

30.0683

30.1861

29.7461

28.9540

1 14

27.2379

28.3031

29.4225

30.1436

30.4931

30.5879

30.3055

29.4991

1 15

27.1276

28.4479

29.5017

30.0784

29.9426

29.5764

29.6402

29.8531

1 16

26.6852

28.2787

29.3640

29.8157

29.4465

29.0940

29.3511

29.9123

1 17

2 6.5073

27.6797

28.7189

28.9697

28.9559

28.9189

29. 1223

29.5941

1 IS

24.5983

26.2094

27.4657

27.8673

28.3008

28.5781

28.8050

29.3971

1 19

22 .8515

23 .7349

25.2685

26. 1621

27.2968

27.9337

28.2830

28.8351

1 20

19.6247

20.5592

22.7508

24.4561

26.1729

26.9226

27.1932

28.1863

1 21

17.5897

18.4304

19.2873

21.5901

22.4980

23 .2718

24.3704

25.4730

1 22

13.9149

13.8449

14.4835

15.4996

17.1143

18.7123

18.8846

20.2761

24V

\M\

|For Help, press F1

NUM

The Column Delimiter options are White Space and Comma. In the example above,
White Space is the correct choice.

The BenMAP health impact functions for ozone generally assume that the data is in parts
per billion (ppb). If your model data happens to be in parts per million (ppm) then check
the box Convert model values from parts per million to parts per billion.

Click on the Select Western Domain Files button and choose the files that you want to
load. Note that each file represents a different day of modeling, and there can be any
number of days. After selecting the files for the western domain, then specify the number
of Columns and Rows in each file by filling in the Lower Bound and Upper Bound
boxes. For example, if the file has 122 columns and 188 rows in the western domain, and
213 columns and 192 rows in the eastern domain, then boxes would be filled out as
follows:

276

September 2008

-------
Tools Menu

Model File Concatenator

Input Files

~ Row,'Column C Column/Row

Column Delimiter

(*" Whitespace Comma

r Convert model values from parts per million to parts per billion

Lower U pper
Bound Bound

Columns |1

Select Western Domain files

Flows: 1

|l 22

|l 88

i\2030bn_12_thoga\2030bn_J4a_12km_us12b. o3_hr_shift_LS T. 0G1 G.txt
A2030bn_12_thoga\2030bn_J4a_12km_u$12b.o3_hr_shift_LST.0617.txt
A2030bn_12_thoga\2030bn_J4a_12km_us12b. o3_hr_shif t_LS T. 0618.txt
A2030bn_12_thoga\2030bn_J4a_12km_us12b.o3_hr_shift_LST.0S19.txt
i'\2030bn_12_thoga\2030bn_J4a_12km_us12b.o3_hr_shift_LST.0620.txt v
< I

Lower U pper
Bound Bound

Select Eastern Domain files:

Columns 1 1213

Rows: [i~~ fl92~

i'.,2030bn_12_thoga'..2030bn_J4a_12km_us12b.o3_hr_shift_LST.061 G.txt a
A2030bn_12_thoga\2030bn_J4a_12km_us12b. o3_hr_shif t_LS T. 0617.txt
A2030bn_12_thoga\2030bn_.J4a_12km_us12b.o3_hr_shift_LST.0618.txt
A2030bn_12_thoga\2030bn_J4a_12km_us12b. o3_hr_shif t_LS T. 0619.txt
i'\2030bn_12_thoga\2030bn_J4a_12km_us12b.o3_hr_shift_LST.0620.txt v

#1 k ar iiar

Done Go!

In this example, BenMAP will renumber the eastern domain so that it follows the
numbering in the western domain. There will be 335 columns and 380 rows numbered
consecutively.

When you have completed filling in the options, then click the Go! button. BenMAP will
begin processing the files. When the processing is complete, the Specify an Output File
window will appear. Provide a name for the file in the File Name box. Note that
BenMAP automatically uses a *.model extension for this type of file.

277

September 2008

-------
Tools Menu

After the file is saved, BenMAP brings you back to the Model File Concatenator window.
You can reuse the tool by choosing a new set of western and eastern domain model files
and making the other appropriate selections. When you are done using this tool, you can
just click the Done button.

10.3 Export Air Quality Grid

The Export Air Quality Grid tool generates a data file (*.csv) with all of the data in an
air quality grid. After choosing Export Air Quality Grid from the Tools drop-down menu,
the Export Air Quality Grid window will appear. Click the Browse button to choose the
air quality grid that you want to examine.

278

September 2008

-------
Tools Menu

$ Export Air Quality Grid

Qfn^T

Air Quality Grid: |aseline 03 2000 VNA CMAQ36Nation0verlap.aqg

Browse

Cancel

Click OK after you have selected your file. This will bring up the Reports window. Use
the Save in drop-down menu to choose the director where you want to save your file. And
in the File name box, type in the name of the file.

To help keep track of what you are doing, you might want to name the file the same thing
as your air quality grid, or something very similar. (If you name it the same as your air
quality grid, you can always distinguish the two files with the extension. An air quality
grid has a *.aqg extension and the file you are generating here has a *.csv extension.)

When done click the Save button. You can view the files you have created with any
database viewer. For each Metric and Seasonal Metric, you can see the actual values. In
addition, you can see the Statistics calculated for each. In the example below for a

279

September 2008

-------
Tools Menu

county-level PM2.5 air quality grid, you can see in the first county (Column = 1, Row =1)
that the Mean of the D24HourMean metric is 17.24, with & Median of 15.1, & Maximum of
52.1, Minimum of 0.4, and finally the Sum of the D24HourMean values is 6292.

£ Microsoft Excel - Baseline PM25

ClosestMonitor County.csv

® File Edit View Insert Format Tools Date Window Help

I Q H^yjl # Gk v' ? 3*1 H4 * •"? 10 * 4 2

'£1 11

T _ 11 X 1

{{| ioo% £?) di - "1

E2 - f*

E! C

Column

Row

Metric

Seasonal Metric

Statistic

Values

D24HourMean

4.80000019073486,4.80000019073486,4.8000

D24HourMean

Mean

17.24

1
1

D24HourMean

Median

15.1

D24HourMean

Max

52.1

1
1

D24HourMean

Min

0.4

D24HourMean

Sum

6292

D24HourMean

QuarterlyMean

14.2877779006958,14.4659337997437,19.955

1
1
1

D24HourMean

QuarterlyMean

Mean

17.21

D24HourMean

QuarterlyMean

Median

17.21

D24HourMean

QuarterlyMean

Max

20.15

D24HourMean

QuarterlyMean

Min

14.29

D24HourMean

QuarterlyMean

Sum

68.86

D24HourMean

5.90000009536743,5.90000009536743,5.9000

D24HourMean

Mean

14.57

1
1

D24HourMean

Median

13.9
40.3

D24HourMean

Max

D24HourMean

Min

5319.19

D24HourMean

Sum

D24HourMean

QuarterlyMean

11.8411111831665,13.2747249603271,18.798

D24HourMean

QuarterlyMean

Mean

14.55

D24HourMean

QuarterlyMean

Median

13.79

D24HourMean

QuarterlyMean

Max

18.8

D24HourMean

QuarterlyMean

Min

11.84

D24HourMean

QuarterlyMean

Sum

58.22

Hi.'

m .1 U-., .„h ,1 ^

-j innnnnncioicio n mnnnnncnoi cnj> n -inryj T 1
NUM

h ^ ~ h [\ Baseline PM25 2000 CloseslMonit / | •

Ready

Note that the files may well be quite large, with more lines than can be completely read by
a program like Excel, which can read only the first 65,000, or so, lines of data. With large
files, you might want to use a database program like SAS. Alternatively, these files can
also be read by simple text editors, which do not have the 65,000 line limitation of Excel.

10.4 Neighbor File Creator

The Neighbor File Creator tool generates a file that specifies the monitors used to
estimate air quality in each grid cell of a Monitor Direct, Monitor and Model Relative, or
Monitor Rollback air quality grid.

To start, choose Neighbor File Creator from the Tools drop-down menu. This will bring

280

September 2008

-------
Tools Menu

up the Create Neighbor File window. Click on the Browse button, and find the air quality
you want to analyze.

$ Create Neighbors File

n_|x_

I Air Quality Grid: aseline D3 2000 VNA CMAQ36Nation0verlap.aqg|

Browse

Cancel

After finding the file, click the OK button. BenMAP will then ask you to provide a name
for the file you want to create. BenMAP will then save the file as a text file that can be
easily viewed in a variety of programs, including WordPad. An example is as follows:

VNA CMAQ36NationOverlap... tJ(nJ(x]

a Baseline 03 2

File Edit View Insert Format Help

iEt M E

Courier New

v 10 v Western

110

390490081442011

0.207786

52.711

110

390890005442011

0.207349

52.822

110

390970007442011

0. 177358

61.754

110

390271002442011

0.122828

89.170

110

541071002442011

0.098704

110.964

110

390870011442011

0.095200

115.048

110

210890007442011

0.090777

120.654

110

390890005442011

0.294657

25.365

110

390490081442011

0.293583

25.457

110

390830002442011

0. 189328

39.476

110

390970007442011

0.110911

67.386

110

541071002442011

0.061357

121.810

11 n

390870011442011

0.050164

148.988

For Help., press F1

NUM

The first two columns specify the Column and Row variables for each grid cell. In the
example above, you will see that Column =110 and Row = 60 are repeated seven times,
indicating that seven different monitors were used to estimate air quality at this grid cell.
The third column gives the monitor identifier. The fourth column the weight used in the
air quality calculation (e.g., Voronoi Neighbor Averaging). And the fifth column gives the
distance (in kilometers) from the monitor to the center of the grid cell.

Note that if an air quality grid was created using the Closest Monitor option, then only a
single monitor is used for any given grid cell. As a result the resulting neighbor file only
has three columns, the Column and Row identifier and the monitor used.

281

September 2008

-------
Tools Menu

r Baseline PM25 2000 ClosestMonitor Co... QQ(x]

File

Edit Format

View Help

P-

011010007831011

010030010881011

010 69 0 0 02881011

010731009881011

01073 5002881011

011010007881011

010 530002881011

010 5 50010881011

010270001881011

13115000 5881011

011170006881011

011190002881011

010270001881011

010690002881011

1 1

10.5 One-Step Setup

The One-Step Setup tool lets you specify the CFG and APV files (and currency year) for
the One-Step Analysis.

To start, choose One-Step Setup from the Tools drop-down menu. This brings up the
One-Step Setup Parameters window. From the Pollutant drop-down menu, choose one of
the pollutants that you have already specified. The U.S. setup that comes with BenMAP
already has defaults for PM2.5 and Ozone.

282

September 2008

-------
Tools Menu

$ One Step Setup Parameters

For each pollutant specify the standard cfg and apv files you wish used for the One-Step Analyses.

Pollutant: lPMZ5

APV File Name:
Currency Year:

0\Configurations\User Manual Configurations\PM2.5 Wizard APV.apv

Browse |

2000

Cancel

If desired, you can change the defaults. However, the default CFG and APV files for
PM2.5 are specifically tied to the various reporting options available in the One-Step
Analysis, so if you use different CFG and APV files, the reporting options may not look as
intended.

283

September 2008

-------