March
12,
2003
Memorandum
Date
March
12,
2003
To
Memo
to
the
316(
b)
Record
From
Lynne
Tudor,
OW/
EPA;
Elena
Besedin,
Marisa
Mazzotta,
and
Robert
Johnston,
Abt
Associates
Inc.;
and
Elizabeth
Strange,
Stratus
Consulting
Subject
Estimating
Total
and
Non­
Use
Values
for
Fish,
Based
on
Habitat
Values
for
Coastal
Wetlands
and
Eelgrass
(
SAV)

Introduction
Reducing
impingement
and
entrainment
(
I&
E)
losses
of
fish
results
in
both
use
and
non­
use
benefits.
I&
E
losses
to
commercial
and
recreational
fish
that
are
harvested
by
fishermen
can
be
valued
as
direct
use
benefits.
Methodologies
for
estimating
use
values
for
recreational
and
commercial
species
are
well
developed,
and
some
of
these
species
have
been
extensively
studied.
As
a
result,
these
values
are
relatively
easy
to
estimate.
The
portion
of
I&
E
losses
consisting
of
fish
that
are
recreationally
and
commercially
landed,
however,
represented
only
approximately
0.15
percent
of
the
total
age
one
equivalent
I&
E
losses
at
five
estuary,
tidal
river,
and
ocean
case
study
facilities
evaluated
for
the
Phase
II
316(
b)
proposal
(
see
the
proposed
Case
Study
Document,
http://
www.
epa.
gov/
waterscience/
316b/
casestudy/
).
1
The
remaining
I&
E
losses
at
these
five
facilities
are
distributed
as
follows:

<
Unlanded
recreational
and
commercial
fish
represent
0.77
percent
of
the
total
age
one
equivalent
I&
E
losses,

<
Forage
fish
represent
99.08
percent
of
the
total
age
one
equivalent
loss.

Neither
forage
species
nor
the
unlanded
portion
of
recreational
and
commercial
species
have
direct
uses;
therefore,
they
do
not
have
direct
use
values.
The
lack
of
use
values
for
the
unlanded
fisheries
means
that
EPA
did
not
directly
value
approximately
99.85
percent
of
the
total
age
one
equivalent
I&
E
losses
at
the
five
cooling
water
intake
structures
(
CWISs)
discussed
above.
2
Although
individuals
do
not
use
these
resources
directly,
they
may
nevertheless
be
affected
by
changes
in
their
status
or
quality.
Monetary
expression
of
individuals 
preferences
for
these
resources
is
known
as
non­
use
value.
Both
commercial
and
recreational
fishermen,
as
well
as
those
who
do
not
use
the
resource,
may
have
non­
use
values
for
these
species.

Given
that
aquatic
species
without
any
direct
uses
account
for
the
majority
of
CWIS
losses,
it
is
important
to
account
for
non­
use
values
in
the
benefits
analysis.
Stated
preference
methods,
or
benefit
transfers
based
on
stated
preference
studies,
are
the
only
generally
accepted
techniques
for
estimating
non­
use
values.
Stated
preference
methods
rely
on
surveys,
which
ask
people
to
state
their
willingness
to
pay
for
particular
ecological
improvements,
such
as
increased
protection
of
aquatic
species
or
habitats
with
particular
attributes.
Benefit
transfer
involves
adapting
research
conducted
for
another
purpose
in
the
available
literature
to
address
the
policy
questions
in
hand.
Because
benefits
analysis
of
environmental
regulations
rarely
affords
enough
time
to
develop
1
The
estuary/
tidal
river
facilities
incorporated
in
this
estimate
include
Salem,
Big
Bend,
and
Brayton
Point.
The
ocean
facilities
are
Seabrook
and
Pilgrim.

2
Although
the
percentages
vary
by
case
study,
the
same
trend
occurs
in
the
other
case
studies.
For
example,
the
total
percentage
unvalued
in
the
Great
Lake
case
study
(
J.
R.
Whiting
and
Monroe)
was
99.92
percent.

­
1­
March
12,
2003
original
stated
preference
surveys
that
are
specific
to
the
policy
effects,
benefit
transfer
is
often
the
only
option
for
providing
information
to
inform
a
policy
decision.

It
is
not
feasible
to
conduct
a
primary
stated­
preference
study
for
the
316(
b)
rule,
because
of
the
regulatory
schedule
and
the
time
and
significant
resources
necessary
to
properly
perform
such
a
study.
EPA
therefore
analyzed
available
existing
evidence
concerning
total
benefits
(
including
use
and
non­
use
values)
for
the
316(
b)
regulation.
Specifically,
the
Agency
developed
total
values
for
forage
species
and
recreational
and
commercial
species
(
including
the
unlanded
portion
of
commercial
and
recreational
species)
using
a
benefit
transfer
approach.
This
approach
uses
benefit
transfer
of
habitat
values
to
value
the
amount
of
habitat
required
to
offset
I&
E
losses.

The
approach
discussed
here
uses
values
that
survey
respondents
indicated
for
preservation/
restoration
of
eelgrass
(
submerged
aquatic
vegetation,
SAV),
and
wetlands
to
evaluate
losses
of
fishery
resources.
Because
one
of
the
results
of
aquatic
habitat
preservation/
restoration
is
increased
production
of
fish
and
shellfish,
it
may
be
appropriate
to
use
valuation
of
habitat
restoration
as
a
proxy
for
the
value
of
the
fish
and
shellfish
lost
due
to
I&
E.
The
method
used
by
EPA
for
such
indirect
valuation
first
assesses
respondents'
values
for
habitats
that
play
a
significant
role
in
the
production
of
fish
or
shellfish,
and
then
estimates
the
quantity
of
such
habitat
required
to
replace
fish
and
shellfish
lost
to
I&
E.
These
data
are
then
combined
to
yield
an
indirect
estimate
of
household
values
for
fish
and
shellfish.

Researchers,
in
the
original
Peconic
survey
(
Opaluch,
1998)
asked
respondents
to
value
eelgrass
and
wetland
habitat,
which
EPA
then
translated,
using
estimates
of
fish
abundance
in
these
habitats
into
values
for
specific
species
and
quantities
of
fish
or
shellfish.
The
purpose
of
this
memo
describes
the
benefit
transfer
approach
and
its
application
to
estimate
baseline
I&
E
losses
for
the
North
Atlantic
region.
The
methodology
used
in
this
analysis
uses
estimates
of
the
public
value
of
the
fish
habitat
services
from
wetlands
and
eelgrass,
or
sub­
aquatic
vegetation
(
SAV),
to
value
I&
E
damage
to
fishery
resources.

This
benefit
transfer
approach
involves
three
general
steps:

<
(
1)
Estimate
the
amount
of
restored
wetlands
and
eelgrass
needed
to
produce
organisms
at
a
level
necessary
to
offset
I&
E
losses
for
the
subset
of
species
for
which
potential
production
information
is
available;

<
(
2)
Develop
WTP
values
for
ecosystem
services
from
fish
production
by
wetlands
and
eelgrass;
and
<
(
3)
Estimate
the
total
value
of
baseline
I&
E
losses
by
multiplying
the
WTP
values
for
fish
and
shellfish
services
of
restored
wetlands
and
eelgrass
by
the
number
of
acres
of
each
needed
to
offset
I&
E
losses.

The
rest
of
this
memorandum
outlines
the
methodology
and
applies
it
to
the
North
Atlantic
region.
The
benefit
transfer
provides
estimates
of
public
values
for
eelgrass
(
SAV)
and
coastal
wetlands,
both
of
which
provide
fish
and
shellfish
habitat.
Because
wetlands
may
provide
services
other
than
fish
habitat,
the
estimated
wetlands
values
are
reduced
to
reflect
only
the
portion
of
value
that
may
be
attributed
to
fish
habitat.
These
values
are
then
applied
to
residents
living
near
power
generating
facilities
in
the
North
Atlantic
region
to
evaluate
baseline
I&
E
losses
from
these
facilities.

1.
Estimating
the
Amount
of
Different
Habitat
Types
Needed
to
Offset
I&
E
Losses
for
Specific
Species
The
first
step
in
the
analysis
involves
calculating
the
area
of
habitat
needed
to
offset
I&
E
losses
for
the
subset
of
species
for
which
restoration
of
these
habitats
was
identified
by
local
experts
as
the
preferred
restoration
alternative,
and
for
which
production
information
is
available
(
i.
e.,
the
habitat
that
will
produce
the
equivalent
quantity
of
fish
impinged
and
entrained
at
CWIS).
Habitats
that
support
fish
and
shellfish
include
seagrasses,
tidal
wetlands,
coral
reefs,
and
estuarine
soft­
bottom
sediments.
The
analysis
may
also
consider
man­
made
­
2­
March
12,
2003
habitat
enhancements,
such
as
artificial
reefs
or
fish
passageways.
The
most
suitable
habitat
restoration
option
is
selected
for
each
affected
species.
The
proposed
316(
b)
rule
case
study
report
provides
a
detailed
description
of
this
method
(
see
U.
S.
EPA
2002,
Proposed
Section
316(
b)
Phase
II
Existing
Facilities
Rule
Case
Study
Analysis,
http://
www.
epa.
gov/
waterscience/
316b/
casestudy/
).

This
memorandum
provides
estimates
of
the
habitat
required
to
offset
I&
E
losses
of
fish
species
at
estuarine
and
ocean
facilities
in
the
North
Atlantic
region
based
on
the
amount
of
habitat
necessary
for
species
that
could
benefit
from
restoration
of
submerged
aquatic
vegetation
(
SAV)
and
wetlands.
The
regional
analysis
is
based
on
analyses
for
the
Brayton
Point
and
Pilgrim
power
plants
located
in
the
North
Atlantic
region.
A
panel
of
local
resource
experts
was
convened
separately
for
each
facility
to
determine
the
type
of
habitat
restoration
most
likely
to
offset
I&
E
losses
at
the
facility.
Although
a
separate
expert
panel
was
convened
for
each
facility,
the
same
restoration
action
was
identified
for
species
that
were
lost
at
both
facilities,
with
the
exception
of
scup.
Restoration
of
submerged
aquatic
vegetation
(
SAV)
was
identified
as
the
preferred
restoration
method
for
scup
lost
at
Brayton
Point,
while
water
quality
improvement
was
identified
as
the
preferred
restoration
method
for
scup
lost
at
Pilgrim.

Table
1
presents
the
fish
species
in
the
North
Atlantic
region
for
which
SAV
or
tidal
wetland
restoration
was
identified
as
the
preferred
method
for
offsetting
I&
E
losses.
Of
the
37
fish
species
lost
to
I&
E
in
the
region,
experts
determined
that
losses
of
seven
species
could
be
offset
by
tidal
wetland
restoration,
and
losses
of
six
species
could
be
offset
by
SAV
restoration.
Table
1
presents
the
I&
E
losses
and
restoration
action
identified
for
each
of
these
13
species.
Losses
of
these
species
account
for
23,090,067
age
1
equivalent
fish
out
of
the
total
regional
I&
E
loss
of
50,484,839.
This
represents
45.7%
of
the
total
I&
E
loss
in
the
North
Atlantic
region.

Table
1:
Average
Annual
Age
1
Equivalent
I&
E
Losses
and
Preferred
Habitat
Restoration
Alternative
(
SAV
or
Tidal
Wetland)
for
I&
E
Species
in
the
North
Atlantic
Region.
I&
E
Species
Average
Annual
Age
1
Equivalent
Preferred
Habitat
Restoration
American
sand
lance
10,344,711
tidal
wetland
Winter
flounder
8,981,759
tidal
wetland
Grubby
3,505,701
tidal
wetland
Atlantic
silverside
106,159
tidal
wetland
Striped
killifish
8,225
tidal
wetland
Striped
bass
662
tidal
wetland
Bluefish
1,221
tidal
wetland
Threespine
stickleback
81,693
SAV
Atlantic
cod
25,132
SAV
Northern
pipefish
15,224
SAV
Weakfish
12,586
SAV
Pollock
3,380
SAV
Scup
3,614
SAV
(
Brayton)
Other
(
Pilgrim)
TOTAL
I&
E
23,090,067
Table
2
presents
estimated
age
1
equivalent
densities
in
wetland
or
SAV
habitat
for
the
13
fish
species
for
which
restoration
of
these
habitats
was
identified
as
the
preferred
alternative
for
offsetting
I&
E
losses.
3
These
estimates
are
derived
from
abundance
data
for
these
species
in
wetland
and
SAV
habitats.
Abundance
data
were
used
because
estimates
of
production
rates
in
these
habitats
are
not
available
for
the
species
of
interest.
Individuals
were
counted
within
subsampling
areas
of
the
habitats
(
e.
g.,
100
square
meters),
and
the
resulting
counts
were
scaled
up
to
derive
per
acre
density
estimates
by
species.
Usable
data
were
available
for
three
species
for
which
3
Specific
data
sources
for
these
estimates
and
details
of
data
analyses
are
provided
in
chapters
F5
and
G5
of
the
§
316(
b)
Phase
II
Case
Study
Document.

­
3­
March
12,
2003
I&
E
data
were
also
available
that
were
found
in
wetlands
(
winter
flounder,
Atlantic
silverside,
and
striped
killifish)
and
for
three
species
that
were
found
in
SAV
(
threespine
stickleback,
northern
pipefish,
and
scup).

Table
2:
Estimated
Age
1
Equivalent
Densities
in
Tidal
Wetland
and
SAV
for
13
I&
E
Species
Tidal
Wetland
Age
1
Equivalent
Density,
fish/
100
m
2
(
fish/
acre)
a,
b
SAV
Age
1
Equivalent
Densitity,
fish/
100
m
2
(
fish/
acre)
a,
b
Brayton
Point
Estimate
Pilgrim
Estimate
Brayton
Point
Estimate
Pilgrim
Esimate
American
sand
lance
No
abundance
data
No
abundance
data
Winter
flounder
0.05
(
205)
0.09
(
351)
Grubby
No
abundance
data
No
abundance
data
Atlantic
silverside
0.05
(
202)
0.19
(
752)
SAV
not
preferred
restoration
alternative
Striped
killifish
0.19
(
764)
0.17
(
691)
Striped
bass
No
abundance
data
No
abundance
data
Bluefish
No
abundance
data
No
abundance
data
Threespine
stickleback
11.23
(
455)
19.29
(
781)
Atlantic
cod
No
abundance
data
No
abundance
data
Northern
pipefish
Wetland
not
preferred
restoration
alternative
No
I&
E
at
Brayton
Point
2.50
(
101)

Weakfish
No
abundance
data
No
abundance
data
Pollock
No
abundance
data
No
abundance
data
Scup
0.08
(
3)
Not
the
preferred
habitat
restoration
method
for
increased
species
production
a.
Differences
in
the
abundance
estimates
for
a
specific
species
between
Brayton
Point
and
Pilgrim
reflect
incorporation
of
differences
in
site­
specific
life
history.
b.
Abundance
estimates
per
unit
of
habitat
are
rounded
to
the
nearest
fish.

For
any
given
species,
the
number
of
acres
of
restored
habitat
needed
to
offset
I&
E
losses
is
determined
by
dividing
the
species
average
annual
age
1
equivalent
I&
E
loss
by
its
estimated
abundance
per
acre
in
that
habitat.
Table
3
presents
results
of
this
calculation
for
the
species
in
Table
2
with
density
estimates
for
SAV
or
wetland.
The
amount
of
wetlands
acreage
needed
to
restore
I&
E
losses
ranged
from
11­
12
acres
for
killifish
to
25,589­
48,813
acres
for
winter
flounder.
For
SAV,
the
acreage
estimated
ranged
from
105­
180
acres
for
threespine
stickleback,
to
1205
acres
for
scup.

­
4­
March
12,
2003
Table
3:
Quantity
of
Tidal
Wetland
and
SAV
Habitat
Restoration
Required
to
Offset
I&
E
Losses
of
Species
for
Which
Estimates
Could
Be
Developed
Species
Estimated
North
Atlantic
Age
1
Equivalent
I&
E
Loss
Expected
Abundance
per
Restored
Acre
of
Habitat
Total
Acres
of
Habitat
Restoration
Required
to
Offset
I&
E
Loss
Brayton
Point
Estimate
Pilgrim­
based
Estimate
Brayton
Estimate
Pilgrim
Estimate
Tidal
Wetland
Restoration
Winter
flounder
8,981,759
205
351
43,813
25,589
Atlantic
silverside
106,159
202
752
526
141
Striped
killifish
8,225
764
691
11
12
SAV
Restoration
Threespine
stickleback
81,693
455
781
180
105
Northern
pipefish
15,224
No
I&
E
at
Brayton
Point
101
No
I&
E
at
Brayton
Point
151
Scup
3,614
3
Not
the
preferred
restoration
alternative
1,205
Not
the
preferred
restoration
alternative
Table
4
presents
lower
and
upper
bound
estimates
of
the
total
wetland
and
SAV
restoration
required
to
offset
North
Atlantic
I&
E
losses.
These
estimates
reflect
the
acreage
needed
for
the
species
requiring
the
maximum
quantity
of
habitat
restoration
to
offset
its
I&
E
losses
(
see
Table
3).
The
amount
of
tidal
wetland
restoration
in
the
North
Atlantic
region
is
based
on
the
acreage
required
for
winter
flounder.
The
lower
bound
estimate
is
winter
flounder
restoration
estimate
derived
for
Brayton
Point
and
the
upper
bound
estimate
is
the
estimate
for
Pilgrim.

The
acreage
needed
for
winter
flounder
overstates
the
acreage
necessary
for
Atlantic
silverside
and
Striped
killifish.
However,
abundance
data
on
additional
species
identified
by
local
experts
as
benefitting
most
from
tidal
wetland
restoration
was
unavailable
to
determine
the
acreage
necessary
to
replace
other
species
such
as
American
sand
lance,
Grubby,
Striped
Lance,
and
Bluefish
(
see
Table
1),
and
the
Agency
was
unable
to
provide
acreage
estimates
for
these
species.

The
lower
bound
estimate
for
regional
SAV
restoration
is
based
on
the
acreage
needed
for
northern
pipefish
at
Pilgrim
and
the
upper
bound
estimate
is
based
on
the
acreage
needed
for
scup
at
Brayton
Point.

Table
4:
Lower
and
Upper
Bound
Estimates
of
Total
Wetland
and
SAV
Restoration
Required
to
Offset
North
Atlantic
I&
E.
Habitat
Restoration
Category
Lower
Bound
on
Required
number
of
Acres
Upper
Bound
on
Required
number
of
Acres
Tidal
wetland
restoration
25,589
43,813
SAV
restoration
151
1,205
­
5­
March
12,
2003
2.
Developing
WTP
Values
for
Fish
Production
Services
Provided
by
Submerged
Aquatic
Vegetation
and
Wetlands
The
value
for
fish
habitat
can
provide
an
indirect
basis
for
valuing
the
fish
that
are
supported
by
the
habitat.
The
mechanism
for
such
indirect
valuation
first
assesses
public
values
for
ecosystems
that
are
valued
mainly
for
the
provision
of
fish
habitat,
and
then
estimates
the
quantity
of
such
ecosystems
required
to
provide
the
species
and
quantities
of
fish
to
be
valued.
Combining
these
data
yields
an
indirect
estimate
of
the
public s
value
for
the
fish
supported
by
the
habitat.
Survey
respondents
can
be
asked
to
value
acres
of
habitat
(
e.
g.,
eelgrass
or
wetlands),
and
these
values
per
acre
can
then
be
translated,
using
scientific
data
on
the
species
supported,
into
values
for
specific
species
and
quantities
of
fish.
It
is
likely
to
be
easier
for
members
of
the
general
public
to
understand
and
express
their
values
for
certain
ecosystems,
like
wetlands,
with
which
they
are
relatively
familiar,
than
to
express
values
for
specific
fish
species
with
which
they
may
or
may
not
be
familiar.
Existing
wetland
valuation
studies
found
that
members
of
the
general
public
are
aware
of
the
fish
production
services
provided
by
SAV
and
wetlands,
and
that
they
express
support
for
steps
that
include
increasing
SAV
and
wetland
areas
to
restore
reduced
fish
and
shellfish
populations
(
Opaluch
et
al.,
1995;
Mazzotta,
1996).

2.1
Description
of
Studies
Used
for
Benefit
Transfer
EPA
based
the
benefit
transfer
of
both
total
and
non­
use
values
for
fish
habitat
provided
by
eelgrass
and
wetlands
on
a
site­
specific
study
of
the
Peconic
Estuary,
located
on
the
East
End
of
Long
Island,
New
York
(
Johnston
et
al.,
2001a,
2001b;
Opaluch
et
al.,
1995,
1998;
Mazzotta,
1996).
Conducted
in
1995,
the
study
provides
information
for
the
Peconic
Estuary
Program s
Comprehensive
Conservation
and
Management
Plan
(
see
http://
www.
savethepeconicbays.
org/
ccmp/
).
EPA
used
an
additional
wetlands
valuation
study,
Johnston
et.
al.,
2000b,
to
supplement
the
information
provided
by
the
Peconic
Estuary
study.
Section
2.1.2
discusses
this
additional
study.
The
Peconic
Estuary
study
is
discussed
briefly
below.

2.1.1
Measuring
Public
Values
and
Priorities
for
Natural
Resources:
An
Application
to
the
Peconic
Estuary
System
(
M.
Mazzotta,
1996)

EPA
selected
the
Peconic
Estuary
study
for
two
reasons:

<
The
study
elicited
the
public s
total
WTP
values
for
coastal
wetlands
and
eelgrass,
using
the
contingent
choice
method.
Wetland
and
eelgrass
habitats
are
essential
for
supporting
fish
and
shellfish
species.
These
habitat
were
also
the
habitats
most
frequently
identified
by
the
expert
panels
for
offsetting
I&
E
losses
at
the
316(
b)
study
sites.
4
The
survey
described
eelgrass
to
respondents
as
 
fish
and
shellfish
habitat 
but
did
not
specifically
describe
wetlands 
services.

<
The
study
area
is
located
between
the
North
and
Mid
Atlantic
regions,
and
thus
provides
site­
specific
values
for
an
estuary
that
can
be
potentially
representative
of
values
in
both
regions.
5
The
Peconic
Estuary s
eelgrass
and
wetlands
support
aquatic
species
that
are
found
throughout
the
North
Atlantic
region
and
that
are
likely
to
be
affected
by
I&
E
(
e.
g.,
bay
anchovy,
Atlantic
silverside,
scup,
summer
flounder,
winter
flounder,
windowpane
flounder,
weakfish,
and
tautog,
as
well
as
scallops
and
hard
clams).
The
Peconic
Estuary
study
thus
provides
values
for
eelgrass
and
wetlands
that
may
be
representative
of
habitat
needed
to
produce
many
of
the
species
affected
by
I&
E
at
power
plants.

4
The
Agency
held
meetings
on
Habitat
Prioritization
for
Species
Impinged
and
Entrained
by
Brayton
Point
on
9/
10/
2001
in
Fall
River
MA;
and
on
on
Habitat
Prioritization
for
Species
Impinged
and
Entrained
by
Pilgrim
on
9/
12/
2001
in
Lakeville,
MA.
For
a
detailed
list
of
participating
Agencies
and
experts,
See
Chapters
F5
and
G5
in
the
Case
Study
Document
that
accompanied
the
proposed
rule.

5
See
Peconic
Estuary
Program
CCMP,
chapter
4,
www.
savethepeconicbays.
org/
ccmp
for
details.

­
6­
March
12,
2003
For
the
final
316(
b)
rule
analysis,
EPA
will
further
evaluate
the
applicability
of
the
Peconic
study
to
other
study
regions
in
the
Mid­
Atlantic.
EPA
will
also
evaluate
other
aquatic
habitat
valuation
studies
for
their
applicability
to
the
analysis
of
benefits
of
the
316(
b)
rule
in
other
regions.

The
Peconic
study
used
an
original
contingent
choice
survey
to
estimate
the
relative
preferences
of
residents
and
second
homeowners
in
the
study
area
for
preserving
and
restoring
key
natural
and
environmental
resources.
6
The
study
area
is
the
East
End
of
Long
Island,
New
York,
which
includes
the
five
towns
surrounding
the
Peconic
Estuary:
Southold,
Riverhead,
Southampton,
East
Hampton,
and
Shelter
Island.
A
map
in
Appendix
1
depicts
the
geographic
area
used
in
this
analysis.
The
primary
goal
of
the
survey
was
to
learn
about
the
public s
preferences,
priorities,
and
values
for
the
environmental
and
natural
resources
of
the
Peconic
Estuary
that
might
be
affected
by
preservation
and
restoration
actions.

The
contingent
choice
survey
format
asks
respondents
to
choose
between
bundles
of
public
commodities,
which
differ
across
their
physical,
environmental,
aesthetic,
and/
or
monetary
dimensions.
For
example,
respondents
might
compare
two
environmental
policy
proposals,
each
with
a
different
impact
on
coastal
resources
and
a
different
monetary
cost.
By
analyzing
the
choices
that
respondents
make
among
a
variety
of
potential
policies
(
i.
e.,
their
preferences),
it
is
possible
to
estimate
respondents 
relative
values
for
environmental
commodities
(
or
policy
results),
and
their
willingness
to
trade
off
elements
of
policy
packages
(
Cameron,
1988;
Hanemann,
1984).

The
extensive
process
to
develop
the
survey
over
a
six­
month
period,
from
February
to
August
1995,
included
individual
interviews,
focus
groups,
and
pretests
of
preliminary
versions
of
the
survey.
Based
on
concerns
expressed
by
participants
in
focus
groups,
and
natural
resources
identified
as
important
by
the
Technical
Advisory
Committee,
the
survey
addressed
five
natural
resources:
(
1)
farmland,
(
2)
undeveloped
land,
(
3)
wetlands,
(
4)
shell
fishing
areas,
and
(
5)
eelgrass.
The
survey
objective
was
to
determine
respondents 
values
for
improvements
in
natural
resources
above
a
specified
baseline
level.
The
baseline
is
that
level
that
would
exist
in
the
year
2020
if
no
action
occurred
to
preserve
or
restore
the
resource.
The
baseline
was
determined
in
consultation
with
the
Technical
Advisory
Committee,
based
on
historical
declines
and
the
judgment
of
experts,
for
each
resource.

In
the
contingent
choice
questions,
each
resource
was
included
at
three
different
levels:
The
projected
level
for
2020
(
the
 
no
new
action, 
or
baseline,
scenario),
and
two
levels
associated
with
hypothetical
programs
that
would
preserve
or
restore
the
resource.
Eelgrass
was
presented
at
the
current
level
of
9,000
acres;
the
baseline
or
 
no
action 
level
of
8,000
acres;
and
a
high
level,
with
restoration,
of
11,000
acres.
Wetlands
were
presented
at
the
current
level
of
16,000
acres;
the
baseline
or
 
no
action 
level
of
12,000
acres;
and
a
high
level,
with
restoration,
of
17,500
acres.

Each
survey
question
asked
respondents
to
compare
the
 
No
New
Action 
baseline
levels
of
two
of
the
resources
to
two
hypothetical
programs
to
protect
or
restore
these
resources.
7
Respondents
were
asked
to
select
their
preferred
option,
given
the
program
costs.
Figure
1
shows
an
example
question.
A
total
of
60
different
questions
were
developed,
using
Addelman s
fractional
factorial
design,
to
produce
orthogonal
arrays
of
attributes
(
Addelman
1962a,
1962b).
8
Each
booklet
contained
five
contingent
choice
questions.

6
The
values
presented
here
understate
total
benefits
because
they
do
not
include
values
for
visitors
to
the
Peconic
region;
they
do
not
include
values
for
other
individuals
outside
of
the
region
who
may
value
the
resources
of
this
region;
they
do
not
include
producer
surplus
values
to
commercial
users
of
the
resources;
and
they
do
not
include
values
to
users
of
the
resources
in
other
regions
who
may
benefit
from
species
that
are
migratory.

7
In
order
to
avoid
choice
questions
that
are
excessively
complex,
each
question
included
only
two
of
the
five
resources,
plus
the
cost
to
each
household.
The
survey
instructions
led
respondents
to
assume
that
the
three
resources
not
included
in
each
question
would
not
be
affected
by
the
program
being
evaluated,
and
would
thus
be
at
the
baseline
levels.

8
The
orthogonal
design
selected
did
not
allow
for
estimation
of
interactions
among
the
resources.
Such
a
design
would
have
required
a
much
larger
sample
size,
which
was
not
possible
given
the
project
budget
limitations.

­
7­
March
12,
2003
Respondents
completed
a
total
of
968
surveys
and
answered
4,307
contingent
choice
questions.
The
distribution
of
surveys
in
various
locations
within
the
study
area
ensured
response
collection
from
a
wide
cross­
section
of
the
public.
9
The
data
were
analyzed
using
a
conditional
logit
model.
Based
on
standard
economic
consumer
theory,
the
analysis
assumed
that
respondents
choose
the
option
that
maximizes
utility
received
from
attributes
of
the
option,
subject
to
their
budget
constraint.
In
this
study,
the
options
are
described
in
terms
of
a
set
of
natural
resources
and
the
cost
of
the
option.
Thus,
the
utility
function
can
be
specified
as
U(
R),
where
R
is
a
vector
of
natural
resource
levels.
Solving
the
maximization
problem
results
in
the
demand
functions
for
the
natural
resources,
Ri
(
P,
M),
where
P
is
the
vector
of
prices
and
M
is
total
income.
Substituting
the
demand
functions
into
the
utility
function
gives
the
indirect
utility
function,
U(
P,
M),
and
the
conditional
indirect
utility
function,
U(
R,
M).
For
a
detailed
description
of
the
model,
see
Opaluch
et
al.(
1998).

Box­
Cox
tests
were
performed
to
assess
the
functional
form
of
the
model;
these
tests
indicated
that
the
linear
model
best
fit
the
data.
Consequently,
a
linear
form
of
the
valuation
function
is
likely
to
be
appropriate
at
least
in
the
range
of
resource
quantity
changes
considered
in
the
Peconic
study.
10
Based
on
the
model
coefficients,
relative
values
for
the
different
resources
and
dollar
values
for
protecting
an
additional
acre
of
each
resource
can
be
calculated
as
described
by
Hanemann
(
1984).
The
estimated
values
are
marginal
values,
or
WTP
for
an
additional
acre
of
each
resource.

The
study
found
that
the
survey
sample
population
was
better
educated
and
had
higher
incomes
than
the
population
of
the
area.
Thus,
the
estimated
values
were
adjusted
to
be
representative
of
the
general
population
of
the
East
End
in
terms
of
education
and
income.
These
adjustments
were
made
to
the
model
coefficients
prior
to
estimating
welfare
values.
The
study
used
separate
adjustments
for
those
who
live
in
the
area
year­
round
(
about
2/
3
of
the
sample)
and
those
who
are
seasonal
area
residents.

The
original
study
presented
estimates
of
several
statistical
models.
For
the
analysis
presented
below,
EPA
used
the
most
conservative
model,
in
terms
of
estimated
values,
to
calculate
the
per
household
WTP
values
per
acre
of
eelgrass
and
wetlands.
This
model
includes
alternative­
specific
constants,
which
capture
differences
between
taking
action
and
choosing
the
 
no
action 
alternative,
as
well
as
any
unexplained
differences
between
programs
A
and
B.

Table
5
presents
the
Peconic
model
valuation
results
for
eelgrass
and
wetlands.
To
separate
users 
values
from
non­
users 
values
for
purposes
of
this
analysis,
EPA
re­
estimated
the
Peconic
model
with
separate
coefficients
for
users
and
non­
users
of
fishery
resources.
The
Agency
defined
users
as
those
who
stated
that
they
either
fish
or
shellfish,
thus
deriving
indirect
use
values
from
the
fish
habitat
services
of
eelgrass
and
wetlands.
EPA
estimated
non­
use
values
for
those
who
do
not
fish
or
shellfish.
11
For
eelgrass,
the
value
for
non­
users
is
77.7
percent
of
the
total
value
for
users,
and
82.4
percent
of
the
total
value
estimated
for
both
users
and
non­
users;
while
for
wetlands,
the
value
for
non­
users
is
94.4
percent
of
the
total
value
for
users,
and
95.8
percent
of
the
total
value
for
both
users
and
non­
users.

9
In
order
to
obtain
a
wide
cross­
section
of
the
public,
surveys
were
conducted
at
over
37
locations
around
the
East
End.
The
study
collected
451
at
various
grocery
stores
and
shopping
areas;
248
at
public
libraries
and
post
offices;
and
187
at
other
locations,
including
the
Department
of
Motor
Vehicles,
the
ferry
from
New
London
to
Long
Island,
an
aquarium
and
a
vineyard,
and
82
surveys
at
beaches.

10
That
is,
the
WTP
per
acre
of
restored
SAV
or
wetlands
will
not
change
as
the
number
of
acres
required
for
restoration
to
offset
I&
E
losses
increases.

11
Note
that
this
is
not
strictly
true
for
wetlands,
because
other
services
exist
that
allow
for
use
values
such
as
birdwatching.
The
value
of
wetlands
is
adjusted
to
reflect
fish
production
services
only
in
the
section
on
wetlands
below.

­
8­
March
12,
2003
Figure
1
:
Example
Survey
Question
­
9­
March
12,
2003
Table
5:
Estimated
WTP
Values
from
the
Peconic
Study
(
2002$)
12
Wetlandsa
Eelgrass
(
SAV)
$/
HH/
Acre/
Yearb
Non­
Use
Value
%
$/
HH/
Acre/
Yearb
Non­
Use
Value
%
All
Residents
$
0.056
95.80%
$
0.063
82.40%
Users
$
0.057
94.40%
$
0.067
77.70%
Non­
Usersc
$
0.054
100.0%
$
0.052
100.0%
a.
Note
that
wetlands
values
presented
here
are
WTP
for
all
wetland
services,
not
just
fish
habitat
services.
The
adjustment
for
fish
habitat
values
appears
below.
b.
Values
shown
are
WTP
per
household
per
additional
(
i.
e,
marginal)
acre
per
year.
c.
Non­
users
are
defined
as
respondents
who
neither
fish
nor
shellfish.

The
Peconic
survey
described
eelgrass
specifically
as
fish
and
shellfish
habitat.
EPA
is
not
aware
of
other
direct
uses
of
eelgrass.
Based
on
focus
groups
during
survey
development
and
pretesting,
the
Peconic
authors
concluded
that
participants
were
aware
of
eelgrass
and
its
importance
for
fish
and
shellfish
production.
Thus,
EPA
assigned
all
of
the
estimated
WTP
for
SAV
restoration
to
fish
and
shellfish
production
services.
Based
on
these
same
focus
groups
and
pretests
the
authors
also
concluded
that,
individuals
were
aware
of
and
valued
a
number
of
functions
of
wetlands,
including
fish
and
other
wildlife
habitat,
storm
buffering,
and
aesthetics.
Because
coastal
wetlands
provide
several
services
(
e.
g.,
habitat,
water
quality,
storm
buffering,
and
aesthetics),
EPA
assigned
only
a
portion
of
the
estimated
WTP
for
wetlands
restoration
to
fish
habitat
services.

The
survey
data
available
from
the
Peconic
study,
however,
provide
no
direct
means
to
estimate
the
share
of
total
wetland
value
from
fish
and
shellfish
habitat
services
alone.
Wetland
values
presented
in
Table
5
reflect
all
ecological
services
provided
by
the
wetlands,
not
just
fish
and
shellfish
habitat
services.
EPA
therefore
used
a
stated
preference
study
from
Narragansett
Bay,
Rhode
Island
to
adjust
wetland
values.
This
study
was
designed
to
assess
tradeoffs
among
different
services
of
restored
saltwater
wetlands
(
Johnston
et
al.,
2002b).
The
results
from
this
study
allow
estimation
of
the
share
of
saltwater
wetland
restoration
values
associated
with
various
services,
including
fish
and
shellfish
habitat
services.
As
shown
in
section
2.3.1,
below,
EPA
estimated
the
value
of
saltwater
wetlands
associated
with
fish
habitat
services
alone
by
multiplying
this
share
by
the
total
values
in
Table
5.

2.1.2
Description
of
the
Johnston
et
al.
(
2002)
Study
The
survey
instrument
Rhode
Island
Salt
Marsh
Restoration:
2001
Survey
of
Rhode
Island
Residents
was
designed
to
assess
tradeoffs
among
attributes
of
salt
marsh
restoration
plans.
Survey
development
required
more
than
16
months
and
involved
extensive
background
research;
interviews
with
experts
in
salt
marsh
ecology
and
restoration;
and
over
16
focus
groups
with
more
than
100
Rhode
Island
residents.
Numerous
pretests,
including
verbal
protocol
analysis
(
Schkade
and
Payne,
1994)
ensured
that
the
survey
language
and
format
would
be
easily
understood
by
respondents,
and
that
respondents
would
have
a
common
understanding
of
survey
scenarios
(
cf.
Johnston
et
al.,
1995).

Focus
groups
and
pretests
led
to
an
in­
person
survey
approach
that
combined
a
printed
survey
booklet
with
an
eight­
minute
introductory
video.
This
video
introduced
respondents
to
information
about
salt
marshes
and
salt
marsh
restoration;
reminded
respondents
of
tradeoffs
involved
in
salt
marsh
restoration;
reminded
respondents
of
their
budget
constraint
and
the
implications
of
choosing
to
direct
funds
to
restoration
programs;
emphasized
the
importance
of
respondents 
choices;
and
provided
basic
survey
instructions.
The
authors
pretested
the
video
script
and
graphics
extensively
and
revised
both
of
them
and
the
survey
booklet.

12
EPA
made
dollar
value
adjustments
using
the
Consumer
Price
Index
for
all
urban
consumers
for
the
first
half
of
2002.

­
10­
March
12,
2003
Johnston
et
al.(
1995)
chose
attributes
distinguishing
restoration
plans
based
on
background
research,
expert
interviews,
and
focus
groups.
The
authors
tailored
these
attributes
to
reflect
primary
salt
marsh
services
in
the
Northeast
United
States
that
would
be
influenced
by
restoration
activities,
and
characterized
the
size
of
the
marsh,
together
with
effects
of
restoration,
on
(
1)
habitat
for
birds,
(
2)
habitat
for
fish,
(
3)
habitat
for
shellfish,
(
4)
potential
to
control
mosquito
nuisance,
(
5)
recreational
access,
and
(
6)
household
cost.
13
Based
on
the
results
of
focus
groups
and
expert
interviews,
habitat
and
mosquito
control
services
were
presented
from
a
standardized,
statewide
perspective.
For
example,
improvements
to
fish
habitat
were
characterized
as
 
ecological
improvements
to
RI
fish
populations [
resulting
from
a
particular
restoration
project] 
as
judged
by
wetlands
experts,
compared
to
all
other
potential
salt
marsh
restoration
projects
in
Rhode
Island. 

Following
the
general
approach
of
Johnston
et
al.
(
1999),
the
conjoint
(
or
multi­
attribute
choice)
survey
presented
respondents
with
four
sets
of
discrete
choices,
each
involving
two
alternative,
multi­
attribute
restoration
plans.
The
authors
used
fractional
factorial
design
to
construct
a
range
of
survey
questions
with
an
orthogonal
array
of
attribute
levels,
resulting
in
80
contingent
choice
questions
divided
among
20
unique
booklets.
Attributes
distinguishing
plans
were
selected
based
on
background
research,
expert
interviews,
and
focus
groups.
All
attributes
were
free
to
vary
over
their
full
range
for
both
restoration
plans
presented
in
each
question,
with
no
imposed
ordering
of
attribute
levels
between
the
two
plans.
Based
on
these
attributes,
respondents
chose
one
of
the
two
plans,
or
chose
 
Neither
Plan. 

The
survey
was
conducted
from
September
through
December,
2001.
Respondents
were
intercepted
in
person
at
Rhode
Island
Department
of
Motor
Vehicle
offices,
public
libraries,
and
other
survey
sites.
Interviewers
did
not
tell
respondents
that
the
survey
concerned
salt
marsh
restoration.
Rather,
interviewers
asked
respondents
to
participate
in
an
important
survey
regarding
 
environmental
issues
in
Rhode
Island, 
to
reduce
the
potential
for
topic­
related
nonresponse.
In
total,
interviewers
collected
661
completed
surveys,
providing
complete
and
usable
responses
to
2,341
individual
contingent
choice
questions
(
89
percent
of
the
potential
2,644).

Table
6
presents
variables
incorporated
in
the
analysis
of
salt
marsh
restoration
choices.
These
variables
include:
(
1)
a
dummy
variable
identifying
the
 
neither 
option,
(
2)
quadratic
interactions
between
this
dummy
and
certain
demographic
characteristics,
and
(
3)
variables
for
the
restored
salt
marsh
attributes.
Mean
values
for
salt
marsh
attributes
(
Table
6)
indicate
the
mean
values
of
these
attributes
over
all
completed
surveys
included
in
the
analysis.
The
final
column
of
the
table
calculates
these
mean
values
with
"
neither
plan"
data
rows
excluded.
(
As
noted
above,
each
wetland
restoration
choice
included
the
option
of
choosing
neither
plan.
In
the
multinomial
logit
data,
these
options
are
presented
as
a
"
plan"
with
zeros
for
all
wetland
attributes.)

13
Additional,
non­
habitat
services
that
may
be
provided
by
saltwater
wetlands
include,
among
others,
nutrient
transformation,
storm
buffering,
and
coastal
erosion
control.
Interviews
with
experts
on
saltwater
wetland
functions
in
New
England
(
and
Rhode
Island
in
particular)
indicated,
however,
that
wetland
restoration
would
provide
negligible
impacts
on
these
non­
habitat
functions
in
the
majority
of
cases.
They
based
this
assessment
on
the
small
size
of
most
New
England
coastal
wetlands,
and
on
the
fact
that
restoration
may
not
always
increase
substantially
the
ability
of
a
wetland
to
provide
such
functions
as
storm
buffering
or
erosion
control.
Based
on
this
advice,
the
survey
focused
mainly
on
wetland
habitat
functions.

­
11­
March
12,
2003
Table
6:
Definitions
and
Summary
Statistics
for
Model
Variables
Variable
Name
Description
Whole
Sample
Mean
(
Std.
Dev.)
Mean,
Excluding
 
Neither
Plan 
Scenariosa
Neither
Neither=
1
identifies
 
neither
plan 
options
0.3333
(
0.4714)
0.0000
Environ
Dummy
variable
identifying
respondents
with
membership
in
environmental
organizations
0.1900
(
0.3923)
0.1900
Taxgrp
Dummy
variable
identifying
respondents
with
membership
associations
0.0233
(
0.1510)
0.0233
Loincome
Dummy
variable
identifying
respondents
with
household
income
less
than
$
35,000/
year
0.2450
(
0.4301)
0.2450
Hiedu
Dummy
variable
identifying
respondents
with
greater
than
a
four­
year
college
degree
0.1817
(
0.3856)
0.1817
Birds
Ecological
improvement
to
statewide
bird
populations
resulting
from
specified
salt
marsh
restoration
plan,
compared
to
all
other
potential
salt
marsh
restoration
plans
in
Rhode
Island
2.7608
(
2.6072)
4.1413
Fish
Ecological
improvement
to
statewide
fish
populations
resulting
from
specified
salt
marsh
restoration
plan,
compared
to
all
other
potential
salt
marsh
restoration
plans
in
Rhode
Island
2.9075
(
2.6530)
4.3613
Shellfish
Ecological
improvement
to
statewide
shellfish
populations
resulting
from
specified
salt
marsh
restoration
plan,
compared
to
all
other
potential
salt
marsh
restoration
plans
in
Rhode
Island
2.9070
(
2.6518)
4.3619
Mosquito
Increased
ability
to
control
statewide
mosquito
nuisance
resulting
from
specified
salt
marsh
restoration
plan,
compared
to
all
other
potential
salt
marsh
restoration
plans
in
Rhode
Island
2.9077
(
2.6506)
4.3617
Size
Size
of
restored
salt
marsh
(
minimum
3
acres;
maximum
12
acres)
4.8890
(
4.3965)
7.3335
Pro­
access
Dummy
variable
indicating
that
respondent
feels
that
access
to
salt
marshes
should
be
 
somewhat
limited 
or
 
unlimited 
0.8367
(
0.3697)
0.8367
Con­
access
Dummy
variable
indicating
that
respondent
feels
that
access
to
salt
marshes
should
be
 
severely
limited 
or
 
prohibited 
0.2266
(
0.4187)
0.1633
Platform
Dummy
variable
indicating
that
restoration
provides
 
viewing
platforms 
but
no
 
trails 
0.2215
(
0.4153)
0.3400
Both
Dummy
variable
indicating
that
restoration
provides
both
 
viewing
platforms 
and
 
trails 
0.2215
(
0.4153)
0.3323
Cost
Annual
cost
of
restoration
plan
in
increased
taxes
(
minimum
$
0;
maximum
$
200)
63.1694
(
70.7816)
94.7542
in
taxpayer
(
0­
10
scale)

(
0­
10
scale)

(
0­
10
scale)

(
0­
10
scale)

a.
Each
wetland
restoration
choice
included
the
option
of
choosing
neither
plan.
In
the
multinomial
logit
data,
this
option
is
presented
as
a
 
plan 
with
zeros
for
all
wetland
attributes.
The
final
column
of
the
table
calculates
means
with
the
 
neither
plan 
zeros
excluded.

Table
7
presents
results
for
a
conditional
logit
model
of
survey
data.
The
model
is
significant
at
p<
0.0001
(­
2LnL
P2
=
1157.56,
df=
13);
all
individual
parameter
estimates
are
significant
at
p<
0.05,
with
most
significant
at
p<
0.01.

­
12­
March
12,
2003
Table
7:
Conditional
Logit
Results
for
Johnston
et
al.
(
2002)
Wetland
Restoration
Study
Parameter
Estimate
Std.
Error
z
P>|
z|
Neither
1.1568
0.1934
5.98
0.0001
Neither
x
Environ
­
1.1820
0.2232
­
5.30
0.0001
Neither
x
Tax
0.8676
0.3651
2.38
0.0170
Neither
x
Loincome
0.3104
0.1437
2.16
0.0310
Neither
x
Hiedu
0.4147
0.1686
2.46
0.0140
Birds
0.1191
0.0153
7.78
0.0001
Fish
0.1465
0.0157
9.36
0.0001
Shellfish
0.1587
0.0162
9.78
0.0001
Mosquito
0.1611
0.0162
9.95
0.0001
Size
0.0510
0.0098
5.22
0.0001
Pro­
access
x
Platform
0.1678
0.0826
2.03
0.0420
Pro­
access
x
Both
0.4310
0.0844
5.11
0.0001
Cost
­
0.0072
0.0005
­
14.23
0.0001
­
2LnL
P2
1157.56
Prob>
P2
0.0001
The
signs
of
parameter
estimates
correspond
with
prior
expectations
derived
from
focus
groups,
where
prior
expectations
exist.
Respondents
favor
plans
that
restore
larger
salt
marshes;
improve
bird,
fish,
and
shellfish
habitat;
control
mosquitoes;
provide
public
access;
and
result
in
lower
costs
to
the
household.
Comparing
preferences
for
habitat
improvements
and
mosquito
control
(
all
measured
on
a
ten­
point
scale),
respondents
placed
the
greatest
weight
on
mosquito
control,
followed
by
habitat
improvements
for
shellfish,
fish,
and
birds,
respectively.
The
likelihood
of
rejecting
restoration
outright
(
i.
e.,
choosing
neither
plan)
was
smaller
for
members
of
environmental
organizations,
and
larger
for
members
of
taxpayers
organizations,
lower
income
individuals,
and
more
highly
educated
individuals
(
Johnston
et
al.,
2002b).
Given
the
final
model
shown
above,
however,
changes
in
education
and
income
do
not
influence
the
marginal
utility
of
fish
and
shellfish
habitat,
or
that
of
other
wetland
attributes.

2.2
Estimating
Proportion
of
Wetland
Value
Attributable
to
Fish
Habitat
Results
of
the
conjoint
analysis
(
i.
e.,
the
public
survey
results)
presented
by
Johnston
et
al.
(
2002b)
allow
policy
makers
to
rank
restoration
projects
based
on
their
estimated
influence
on
residents 
welfare.
These
results
also
allows
assessment
of
residents 
willingness
to
trade
off
elements
of
wetland
restoration
plans,
or
willingness
to
pay
for
particular
wetland
attributes.
Finally,
for
any
specified
restoration
plan,
provided
that
incremental
gains
or
losses
in
wetland
services
are
known,
it
allows
the
calculation
of
the
proportion
of
the
total
gain
in
social
value
attributable
to
a
particular
service
(
e.
g.,
fish
and
shellfish
habitat).

To
estimate
the
proportion
of
value
associated
with
fish
and
shellfish
habitat,
in
a
representative,
conservative
scenario,
EPA
began
with
the
average
wetland
restoration
scenario
considered
by
the
Rhode
Island
survey
sample.
The
mean
values
of
wetland
attributes
presented
to
survey
respondents
provide
the
most
representative
set
of
results
from
which
value
proportions
may
be
estimated,
and
forecast
the
value
proportions
that
would
result
from
an
average
survey
respondent
confronted
with
an
average
wetland
restoration
scenario,
as
characterized
by
the
Rhode
Island
Salt
Marsh
Restoration
Survey
data.
Excluding
all
 
Neither
Plan 
scenarios,
which
offered
zero
restoration,
Table
6
summarizes
the
mean
values
for
services
considered
by
the
Rhode
Island
sample.

­
13­
March
12,
2003
Although
mean
values
are
used
for
most
attributes
(
i.
e.,
wetland
attributes
or
services
considered
by
survey
respondents
in
choice
scenarios),
changes
in
certain
attributes
are
set
to
zero
to
correspond
more
closely
with
the
policy
scenario
and
with
the
Peconic
Study
described
above
(
because
the
purpose
of
this
analysis
is
to
assess
the
proportion
of
the
Peconic
wetland
values
that
may
reasonably
be
attributed
to
fish
habitat
services).
For
example,
because
the
Peconic
study
survey
did
not
specify
or
discuss
the
provision
of
viewing
platforms
or
trails
at
preserved
wetlands,
EPA
assumed
that
survey
respondents
to
the
Peconic
study
did
not
consider
such
unusual
provisions
when
making
survey
choices.
Accordingly,
in
calculating
value
proportions
in
this
analysis
using
the
Rhode
Island
data,
EPA
assumed
that
viewing
platforms
and
trails
are
not
provided.

EPA
also
assumed
that
any
wetland
created
or
restored
to
provide
fish
habitat
will
likely
not
provide
a
great
degree
of
additional
mosquito
control,
because
a
large
proportion
of
existing
salt
marshes
have
already
been
modified
to
minimize
mosquito
production.
14
For
this
reason,
modern
marsh
restoration
typically
does
not
provide
a
significant
increase
in
mosquito
control.
Rather,
it
often
replaces
older,
more
detrimental
(
to
marsh
function
and
habitat)
forms
of
mosquito
control
with
Open
Marsh
Water
Management
(
OMWM),
in
which
open
water
and
natural
fish
predation
is
used
to
control
mosquito
nuisance
(
Kennish,
2002).
OMWM
has
not
been
an
 
unqualified
success 
at
eliminating
the
mosquito
nuisance
(
New
York
Conservationist,
1997).
Accordingly,
for
many
salt
marshes,
the
positive
net
effect
of
restoration
on
mosquito
nuisance,
if
any,
is
often
minimal.
To
generate
the
most
conservative
estimates,
however,
and
in
recognition
of
the
fact
that
some
salt
marsh
restoration
projects
may
provide
significant
mosquito
control,
EPA
also
estimated
value
proportions
assuming
that
significant
additional
mosquito
control
is
provided.
For
all
other
wetland
attributes
included
in
the
Rhode
Island
survey,
EPA
used
the
mean
values
shown
in
the
final
column
of
Table
6.

Estimation
of
value
proportions
is
based
on
the
estimated
utility
function
v(
A),
which
specifies
the
utility
provided
by
a
wetland
restoration
plan
as
a
function
of
the
attributes
or
services
provided
by
that
plan
(
Johnston
et
al.,
2002b).
That
is,
following
the
standard
random
utility
model
of
Hanemann
(
1984),
the
underlying
model
specifies
respondents 
choices
using
the
conditional
logit
specification,
in
which
the
probability
(
Pi
)
of
choosing
any
wetland
restoration
plan
i
(
plan
A,
plan
B,
or
neither
plan)
over
the
two
remaining
options
(
j
or
k)
is
given
by
exp[
vi()]
 
Pi
=
 
exp[
vi()]
 
 
 +
 
exp[
vj()]
+
 
exp[
vk()]
 
(
1)

where
v(
A)
represents
the
relative
benefits
or
utility
resulting
from
each
restoration
option,
including
the
 
neither
plan 
option.
The
function
v(
A)
is
typically
estimated
as
a
simple
function
of
program
attributes
(
in
this
case
wetland
restoration);
in
practice
linear,
functional
forms
are
often
used
(
Johnston
et
al.,
2002b).

From
the
assumptions
and
model
noted
above,
the
attribute
definitions
given
in
Table
6,
and
the
model
results
of
Table
7,
the
estimated
utility
function
used
to
calculate
value
proportions
is
specified
as
v()
 
(
2)
 =
 
0.1191(
birds)
+
 
0.1465(
fish)
+
 
0.1587(
shellfish)
+
 
0.1611(
mosquito)
+
 
0.0510(
size)

If
mosquito
control
is
not
provided,
then
mosquito=
0.
Given
this
linear
specification,
the
proportion
of
wetland
restoration
value
provided
by
the
gain
in
fish
habitat
services
is
given
by
14
The
mosquito
control
variable
was
included
in
the
survey
in
response
to
the
strong
concern
of
Rhode
Island
residents
over
the
impact
of
restoration
on
mosquitoes
and
related
illnesses
for
which
mosquitoes
are
the
primary
vector.
Wetlands
experts
indicated,
however,
that
salt
marsh
restoration
had
limited
impact
on
mosquito
populations
in
most
cases.

­
14­
March
12,
2003
v()
 
fish
 
 
v()
 
fish
=
 
0
(
3)
  
 
v()
 
fish
 
 
where
v(
A)
fish
represents
the
value
of
v(
A)
with
the
gain
in
fish
habitat
services
set
to
its
mean
value
(
as
described
above),
and
v(
A)
fish=
0
represents
the
value
of
the
function
with
the
gain
in
fish
habitat
services
set
to
zero.

Table
8
shows
the
resulting
value
proportions,
in
which
EPA
calculated
the
proportion
of
wetland
restoration
value
associated
with
different
wetland
services
based
on
mean
values
of
wetland
attributes
presented
to
survey
respondents,
as
discussed
above.
Analogous
methods
were
used
to
assess
value
proportions
associated
with
shellfish
and
other
habitat
services;
Table
8
shows
these
results
for
comparison.
Table
8
also
illustrates
the
results
of
a
sensitivity
analysis
in
which
EPA
calculated
analogous
value
proportions
for
wetland
habitat
services,
but
allow
wetland
size
to
vary.
Wetland
size
was
allowed
to
vary
from
its
minimum
value
in
the
Rhode
Island
survey
data
(
3
acres)
to
its
maximum
value
(
12
acres),
while
holding
habitat
service
changes
constant.
EPA
chose
these
size
values
to
be
representative
of
unrestored
saltwater
wetlands
currently
existing
in
Narragansett
Bay,
which
are
typically
quite
small
(
i.
e.,
less
than
five
acres).
The
three
estimates
of
acreage
are
therefore
likely
closer
to
the
 
average 
Rhode
Island
wetland
than
estimates
based
on
larger
acreages.
(
In
actual
wetlands,
changes
in
restored
acres
are
typically
correlated
with
larger
gains
in
habitat
services
(
Johnston
et
al.,
2002b).
To
illustrate
even
more
conservative
estimates,
however,
Table
8
contains
cases
in
which
restored
wetland
size
increases
from
the
mean,
without
any
resultant
increase
in
habitat
services.)

­
15­
March
12,
2003
Table
8:
Proportions
of
Restored
Wetland
Value
Associated
with
Various
Service
Categoriesa
Restoration
Scenario
Percentage
of
Value
Associated
with
Service:

Fish
Habitat
Bird
Habitat
Shellfish
Habitat
Mosquito
Control
Otherb
1a:
No
additional
mosquito
control;
Mean
values
for
all
other
attributes
0.2906
0.2244
0.3149
0.0000
0.1701
1b:
No
additional
mosquito
control;
Mean
values
for
Habitat
Gains;
Size=
3
acres
0.3231
0.2494
0.3501
0.0000
0.0774
1c:
No
additional
mosquito
control;
Mean
values
for
Habitat
Gains;
Size=
12
acres
0.2622
0.2024
0.2841
0.0000
0.2512
2a:
Mosquito
control
at
mean
value;
Mean
values
for
all
other
attributes
0.2202
0.1700
0.2386
0.2422
0.1289
2b:
Mosquito
control
at
mean
value;
Mean
values
for
habitat
gains;
size=
3
acres
0.2384
0.1840
0.2583
0.2622
0.0571
2c:
Mosquito
control
at
mean
value;
Mean
values
for
habitat
gains;
size=
12
acres
0.2035
0.1571
0.2205
0.2238
0.1950
3a:
Mean
over
all
scenarios
0.2564
0.1979
0.2778
0.1214
0.1466
a.
Results
assume
that
restoration
does
not
provide
viewing
platforms
or
hiking
trails.
b.
Other
services
may
include,
among
others,
nutrient
transformation,
storm
buffering,
and
coastal
erosion
control.

2.2.1
Results:
Proportion
of
Wetland
Restoration
Value
Attributable
to
Fish
and
Shellfish
Habitat
Services
As
shown
by
Table
8,
the
proportion
of
value
associated
with
fish
habitat
ranges
from
0.2035
to
0.3231,
with
a
mean
value
over
all
scenarios
of
0.2564;
and
the
proportion
of
value
associated
with
shellfish
habitat
ranges
from
0.2205
to
0.3501,
with
a
mean
value
over
all
scenarios
of
0.2778.
Scenario
1a
represents
perhaps
the
most
representative
scenario
for
estimating
value
proportions
for
two
reasons:
(
1)
restored
wetlands
are
not
expected
to
provide
additional
mosquito
control
and
(
2)
other
wetland
attributes
are
set
to
their
mean
values
under
this
scenario.
Its
results
are
somewhat
higher
than
those
of
scenario
3a,
which
represents
the
mean
value
over
all
scenarios
presented.
EPA
therefore,
to
be
conservative,
used
proportions
calculated
in
scenario
3a
(
0.2564
for
fish
and
0.2778
for
shellfish)
as
an
estimate
of
the
proportion
of
total
wetland
restoration
value
attributable
to
gains
in
fish
and
shellfish
habitat
services,
given
representative,
mean
values
for
other
wetland
services.
Although
these
numbers
are
not
directly
comparable
to
other
results
found
in
the
literature,
they
appear
to
be
reasonable
and
conservative
compared
to
similar
proportions
generated
for
freshwater
habitats.
For
example,
Schulze
et
al.
(
1995)
estimate
that
between
32.98
percent
and
33.44
percent
of
WTP
for
resource
cleanup
in
the
Clark
Fork
River
Basin
was
associated
with
 
aquatic
resources
and
riparian
habitat 
(
p.
5­
13).

EPA
also
considered
directly
the
parametric
results
of
Table
7
for
further
support
of
the
soundness
of
the
proposed
value
proportions.
Estimates
presented
in
Table
7
indicate
that
the
parametric
weights
are
similar
among
the
dominant
wetland
services
in
Narragansett
Bay
(
i.
e.,
bird
habitat
services,
fish
habitat
services,
shellfish
habitat
services,
and
mosquito
control).
In
other
words,
the
parameter
estimates
are
very
similar
among
these
four
variables.
This
correspondence
suggests
that
restoration
providing
similar
scale
improvements
for
each
of
these
services
should
produce
a
roughly
equivalent
increment
to
utility.
Given
the
four
habitat
services
considered
in
the
survey
(
including
mosquito
control),
each
service
provides
roughly
1/
4
(
or
25
percent)
of
the
total
marginal
utility
associated
with
the
combination
of
habitat
improvements
and
mosquito
control.
For
wetlands
that
do
not
provide
substantial
access
provisions
(
e.
g.,
boardwalks)
and
that
are
of
moderate
or
small
size,
it
would
be
highly
improbable
for
the
proportion
of
value
associated
with
fish
or
shellfish
habitat
to
fall
significantly
below
the
25.64
percent
approximation
estimated
here.

­
16­
March
12,
2003
2.2.2
Applicability
of
Johnston
et
al.
(
2002)
Value
Proportions
to
Peconic
Total
Wetland
Values
As
noted
above,
no
direct
means
is
available
for
assessing
the
exact
proportion
of
Peconic
wetland
values
(
Table
1)
associated
with
fish
and
shellfish
habitat
services,
however,
the
Johnston
et
al.
(
2002b)
value
proportions
(
Table
8)
provide
a
reasonable,
average
approximation,
based
on
a
random­
sample
survey
of
Rhode
Island
residents.
As
with
any
type
of
benefit
transfer,
the
applicability
of
the
Johnston
et
al.
(
2002b)
value
proportions
to
the
Peconic
wetland
values
depends
on
certain
assumptions.
Primary
among
these
are
assumptions
about
the
approximate
constancy
of
value
proportions
with
respect
to
changes
in
policy
scale.

Unlike
the
Peconic
survey,
which
addressed
total
and
marginal
wetland
values
over
large,
long­
term
changes
in
wetland
acreage
(
i.
e.,
up
to
4,000
acres
over
the
entire
Peconic
region),
the
Rhode
Island
study
estimates
restoration
values
over
the
scale
of
a
single
saltwater
wetland
(
i.
e.,
between
three
and
12
acres).
Although
this
difference
in
scale
is
likely
to
influence
the
marginal
WTP
for
wetland
preservation
or
restoration,
EPA
does
not
expect
the
difference
in
scale
to
significantly
influence
the
proportion
of
marginal
WTP
associated
with
fish
habitat
services.
That
is,
the
Agency
assumed
that
if
fish
and
shellfish
habitat
services
each
account
for
approximately
25
percent
of
the
total
value
for
the
tenth
acre
of
restoration
in
a
region,
then
each
service
will
also
account
for
approximately
25
percent
of
value
for
the
hundredth
or
thousandth
acre
in
the
same
region,
even
though
the
total
value
of
each
acre,
on
the
margin,
may
change.
That
is,
the
assumption
of
fixed
value
proportions
associated
with
fish
and
shellfish
habitat
services
concerns
only
the
relative
proportion
of
value
associated
with
fish
and
shellfish
habitat,
which
may
remain
constant
even
as
the
absolute
marginal
value
of
a
wetland
acre
diminishes
with
scale.
15
A
second
key
assumption
of
this
analysis
is
that
residents
of
the
Peconic
region
and
residents
of
Rhode
Island
maintain
similar
relative
values
with
respect
to
the
services
provided
by
saltwater
wetlands.
Although
this
presumption
cannot
be
proven
using
results
from
Johnston
et
al.
(
2002b)
or
Opaluch
et
al.
(
1995,
1998),
and
while
value
proportions
may
differ
to
a
small
degree,
there
is
no
overriding
reason
to
suspect
(
and
the
literature
results
do
not
suggest)
that
relative
value
proportions
would
differ
to
a
significant
degree
across
the
two
regions.

2.3
Wetland
Values
for
Fish
and
Shellfish
Habitat
Services
Table
9
presents
the
values
for
wetlands
from
Table
5,
adjusted
to
reflect
fish
habitat
services
only.
As
discussed
above,
EPA
adjusted
the
values
using
the
average
proportion
of
total
wetland
restoration
value
attributable
to
gains
in
fish
habitat
services,
based
on
a
number
of
reasonable
scenarios,
25.64
percent.
Similarly,
EPA
calculated
the
value
for
wetlands,
adjusted
to
reflect
shellfish
habitat
services.
As
shown
in
Table
8,
the
average
percent
of
total
wetland
restoration
value
attributable
to
shellfish
habitat
services
is
27.78.
Note
that
the
wetland
15
Technical
note
regarding
the
robustness
of
value
proportions:
Following
standard
practice
(
e.
g.,
Adamowicz
et
al.
1998),
the
survey
design
underlying
model
results
(
Table
7)
is
based
on
an
orthogonal
array
of
attribute
levels.
No
imposed
functional
linkage
exists
between
wetland
size
and
habitat
services.
This
independence
is
preserved
by
the
linear
form
of
the
utility
function.
The
model
therefore
allows
one
to
vary
wetland
size
and
habitat
services
independently
when
estimating
public
values,
even
if
such
independence
is
highly
unlikely
in
real
situations.
This
specification
allows
the
researcher
a
large
degree
of
leeway
when
specifying
"
reasonable"
restoration
scenarios.
It
also
allows
the
valuation
of
clearly
unrealistic
scenarios
in
which,
for
example,
the
restoration
of
huge
wetlands
provides
negligible
habitat
gains.
In
such
unrealistic
scenarios,
it
is
possible
to
illustrate
cases
in
which
proportions
of
value
diminish
to
a
significant
degree
as
wetland
size
increases.
However,
if
one
specifies
more
realistic
scenarios
in
which
increases
in
restored
wetland
acreage
and
resulting
increases
in
habitat
services
change
(
approximately)
proportionately,
then
the
proportions
of
wetland
values
associated
with
fish
and
shellfish
habitat
are
robust.
That
is,
assuming
that
the
marginal
gain
in
habitat
(
fish,
shellfish,
etc.)
provided
by
the
tenth
acre
of
restoration
is
equivalent
to
the
gain
provided
by
the
hundredth
acre,
the
proportion
of
value
associated
with
fish
and
shellfish
will
remain
constant
as
one
increases
the
scale
of
restoration.

­
17­
March
12,
2003
values
for
fish
and
shellfish
services
are
not
linearly
additive
within
the
same
acreage,
due
to
the
functional
form
used
in
Johnson
et
al.
(
2002b).
16
Table
9:
Estimated
WTP
Values
per
Household
for
Fish
and
Shellfish
Habitat
Services
of
Wetlands
(
2002$)
from
the
Peconic
Study
$/
HH/
Acre/
Yeara
Fish
Habitat
%
b
$/
HH/
Acre/
Year
for
Fish
Habitatc
Shellfish
Habitat
%
b
$/
HH/
Acre/
Year
for
Shellfish
Habitatd
All
Residents
$
0.056
25.64%
$
0.014
27.78%
$
0.016
Users
$
0.057
25.64%
$
0.015
27.78%
$
0.016
Non­
Usersc
$
0.054
25.64%
$
0.014
27.78%
$
0.015
a.
Values
shown
are
WTP
per
household
per
additional
(
i.
e,
marginal)
acre
per
year.
b.
From
Table
8
above.
c.
Total
value
per
acre
per
year
times
25.64
percent.
d.
Total
value
per
acre
per
year
times
27.78
percent.

2.3.1
Peconic
Estuary
Saltwater
wetlands
provide
a
range
of
valued
services,
including
fish
habitat,
shellfish
habitat,
bird
habitat,
mosquito
control,
recreational
access,
and
various
other
services
that
are
typically
correlated
with
a
wetland s
size.
Therefore,
EPA
assigned
only
a
portion
of
the
estimated
WTP
for
wetlands
restoration
to
fish
habitat
services.

The
exact
proportion
of
wetland
value
associated
with
fish
and
shellfish
habitat
services
will
vary
across
wetlands
with
different
physical
attributes,
because
different
wetlands
will
likely
provide
different
combinations
of
valued
services.
This
analysis
provides
a
conservative
estimate
of
the
proportion
of
wetland
values
(
Table
5)
attributable
to
fish
and
shellfish
habitat
services,
for
average
saltwater
wetlands
of
the
type
considered
by
the
Peconic
survey
respondents.
The
results
of
Johnston
et
al.
(
2002b)
allow
such
proportions
to
be
approximated,
based
on
choice
experiment
responses
of
randomly
selected
Rhode
Island
survey
respondents.

Multiplying
the
value
per
household
from
Table
9
by
the
total
number
of
households
in
the
Peconic
study
area
(
73,423)
yields
the
value
per
acre
of
wetlands
for
the
population
surrounding
the
Peconic
Estuary.
Table
10
shows
these
values.
The
study
defined
the
affected
population
as
the
total
number
of
households
(
both
year­
round
and
seasonal)
in
the
five
towns
bordering
the
Peconic
Estuary.
As
noted
above,
this
definition
of
the
study
area
results
in
conservative
total
values
because
it
does
not
include
the
values
for
people
who
live
on
Long
Island
beyond
these
five
towns,
or
the
values
of
visitors
to
the
area.
For
the
Peconic
Estuary
region,
the
total
annual
value
per
acre
for
fish
habitat
services
of
wetlands
is
$
1,053,
whereas
the
total
non­
use
value
only
is
$
1,009;
and
the
total
annual
value
per
acre
for
shellfish
habitat
services
of
wetlands
is
$
1,141,
whereas
the
total
non­
use
only
value
is
$
1,093.17
These
values
should
not
be
added
together
to
calculate
total
fish
and
shellfish
values
for
the
same
acreage.
For
calculating
a
total
value
for
fish
and
shellfish
habitat,
the
percent
of
WTP
attributable
to
both
habitats
as
a
set
must
be
calculated.

16
Because
the
overall
values
of
Peconic
Estuary
residents
for
eelgrass
and
wetlands
are
similar,
once
adjustments
are
made
to
wetlands
values
to
ascribe
a
portion
to
fish
habitat
services,
the
values
for
fish
and
shellfish
habitat
of
eelgrass
are
estimated
as
four
times
higher
than
those
for
fish
habitat
only
for
wetlands.

17
This
analysis
assumes
that
non­
use
values
are
the
same
for
both
users
and
non­
users
of
the
affected
resources.
It
is
likely
that
users
of
the
resource
have
higher
non­
use
values
than
non­
users,
but
the
data
do
not
allow
us
to
evaluate
non­
use
values
specific
to
users.

­
18­
March
12,
2003
18
For
most
North
Atlantic
facilities,
a
single
county
was
considered
to
define
the
affected
population.
For
those
facilities
located
within
10
miles
of
a
second
county,
the
households
in
both
counties
were
included;
and
for
facilities
on
rivers,
downstream
counties
were
included.

­
19­
Table
10:
Estimated
WTP
Values
for
Fish
and
Shellfish
Habitat
Services
Provided
by
Wetlands
from
the
Peconic
Study
(
2002$)
FishShellfish
$/
HH/
Acre/
YearaTotal
WTP/
Acre/
Yearb$/
HH/
Acre/
YearaTotal
WTP/
Acre/
Yearb
Total
Value$
0.014$
1,053$
0.016$
1,141
Non­
Use
Valuec$
0.014$
1,009$
0.015$
1,093
a.
Values
shown
are
WTP
per
household
per
additional
(
i.
e,
marginal)
acre
per
year.
b.
Total
WTP
per
acre
is
calculated
as
household
WTP
per
acre
times
73,423
total
households
in
the
study
area.
c.
Total
non­
use
value
is
calculated
as
value
per
acre
for
non­
users
only
times
all
households
in
the
region.

2.3.2North
Atlantic
Region
Evaluating
the
total
value
per
acre
of
SAV
for
the
coastal
population
of
the
North
Atlantic
region
required
a
definition
of
the
benefit
population
for
the
region.
The
Peconic
study
defined
the
affected
population
as
the
total
number
of
households
(
both
year­
round
and
seasonal)
in
the
towns
bordering
the
Peconic
Estuary.
Similarly,
EPA
defines
the
affected
population
as
households
residing
in
the
counties
that
abut
the
water
bodies
in
the
North
Atlantic
Region.
These
households
are
likely
to
value
gains
of
fish
or
shellfish
in
the
affected
water
body,
due
to
their
very
close
proximity
to
the
affected
resource.
EPA
estimated
an
average
benefit
population
for
the
region
as
the
average
number
of
households
in
these
nearby
counties.
18
Households
in
counties
that
do
not
directly
abut
the
affected
water
body
will
also
likely
value
the
water
body's
resources.
Analysis
of
data
from
the
Rhode
Island
Salt
Marsh
Restoration
Survey
(
Johnston
et
al.,
2002b)
reveals
that
values
ascribed
to
even
relatively
small­
scale
salt
marsh
restoration
actions
(
i.
e.,
3­
12
acres)
were
stated
by
respondents
from
various
parts
of
the
state.
EPA
thus
assumed
for
the
current
analysis
that
residents
within
a
similar
distance
from
the
coast
as
residents
in
the
Johnston
et
al.
(
2002)
study
would
have
positive
values
for
improving
fish
habitat.
EPA
calculated
the
average
distance
from
Johnston's
studied
locations
to
the
farthest
edges
of
Rhode
Island,
which
totaled
32.43
miles.
The
Agency
then
assumed
that
all
households
living
within
the
same
distance
of
the
affected
resource
as
Rhode
Island
residents
from
the
studied
resource
would
also
value
fish
habitat
improvements
in
the
affected
water
body.

Additionally,
EPA
notes
that
a
study
by
J.
Pate
and
J.
Loomis
(
1997)
found
that
respondents
outside
the
political
jurisdiction
in
which
a
study
site
is
located
were
also
willing
to
ascribe
stated
preference
values
to
the
amenity
being
studied.
The
study
was
designed
to
determine
the
effect
of
distance
on
willingness
to
pay
for
public
goods
with
large
non­
use
values.
Specifically,
the
study
evaluated
environmental
programs
designed
to
improve
wetlands
habitat
and
wildlife
in
the
San
Joaquin
Valley.
It
compared
WTP
values
for
households
residing
in
the
San
Joaquin
Valley,
California,
to
values
for
California
households
outside
the
Valley,
and
to
households
in
Washington
State,
Oregon,
and
Nevada.
The
study
found
that
WTP
values
for
California
residents
outside
the
Valley
were
97.7
percent
of
the
WTP
of
the
Valley
residents;
WTP
values
for
Oregon
residents
were
approximately
27
percent
of
the
WTP
of
the
Valley
residents.

From
the
evidence
cited,
EPA
reasonably
assumed
that
residents
living
outside
the
counties
abutting
affected
water
bodies
would
have
positive
values
for
improving
fish
habitat
in
those
water
bodies.
EPA
therefore
calculated
per
acre
WTP
values
using
two
different
definitions
of
affected
populations:
(
1)
the
average
number
of
households
residing
in
counties
abutting
the
affected
water
body
and
(
2)
the
average
number
of
households
living
within
the
32.4
mile
radius
of
each
affected
water
body
in
the
region.
March
12,
2003
Table
11
presents
this
information.
The
first
definition
of
the
benefit
population
is
the
average
number
of
households
in
the
counties
abutting
affected
water
bodies,
as
described
above.
According
to
the
2000
Census,
this
average
comprises
210,357
households.
As
noted
above,
the
Agency
also
developed
an
alternative
definition
of
the
affected
population
based
on
the
findings
of
the
Johnston
et
al.
(
2002b)
study
of
Rhode
Island
wetland
restoration.
Johnson
et
al.
(
2002b)
found
that
residents
in
all
of
Rhode
Island
ascribed
values
to
even
relatively
small­
scale
wetland
restoration
actions.
EPA
therefore
calculated
the
average
distance
from
the
western
Rhode
Island
border
to
the
shore
of
Mount
Hope
Bay,
a
water
body
affected
by
one
of
the
Phase
II
facilities
located
in
the
North
Atlantic
region.
This
totaled
32.43
miles.
The
Agency
then
assumed
that
all
residents
living
within
a
32.43
mile
radius
of
each
facility
would
also
value
fish
habitat
improvements
in
the
affected
water
body.
EPA
therefore
calculated
the
average
number
of
households,
over
all
North
Atlantic
facilities,
within
a
32.43
mile
radius
of
each
facility.
Calculations
yielded
an
average
affected
population
of
737,711
households.
A
map
in
Appendix
2
depicts
the
geographic
area
used
in
this
analysis.

Table
11:
Affected
Populations
in
North
Atlantic
Region
Households
in
Counties
Abutting
Affected
Water
Bodies
Households
Within
32.4
Mile
Radius
of
Each
FacilityFacility
State
Households
Norwalk
Harbor
Fairfield
CT
324,232
1,437,927
Bridgeport
Harbor
Fairfield
New
Haven
CT
324,232
775,943319,040
Devon
Fairfield
New
Haven
CT
324,232
749,127319,040
County(
ies)

Fairfield
324,232
New
Haven
Harbor
New
Haven
CT
319,040
787,255
South
Meadow
Station
Hartford
CT
335,098
869,513
New
London
99,835
Middlesex
CT
61,341
Millstone
Suffolk
NY
469,299
223,986
Montville
New
London
CT
99,835
277,085
239,936
239,936
Bristol
67,320
Plymouth
168,361
Providence
Bristol
Kent
RI
19,033
67,320
Manchester
Street
Bristol
MA
205,411
920,918
Providence
19,033
Bristol
Kent
Newport
RI
67,320
35,228
Brayton
Point
Bristol
MA
205,411
835,184
Newport
RI
35,228
Somerset
Bristol
MA
205,411
866,945
Canal
Barnstable
MA
94,822
320,343
Pilgrim
Plymouth
MA
168,361
483,020
­
20­
March
12,
2003
Rockingham
NH
104,529
Maine
Energy
Recovery
Cumberland
107,989
Rockingham
NH
Newington
York
ME
74,563
284,093
Schiller
York
ME
74,563
279,969
Co.
York
ME
74,563
202,096
William
F.
Wyman
Cumberland
ME
107,989
203,005
Lincoln
14,158
Mason
Steam
Sagadahoc
ME
14,117
112,236
Average
Affected
Households
210,357
737,711
Source:
Census
2000
Summary
File
3
(
SF
3)
­
Sample
Data
Table
12
presents
the
values
per
acre
per
year
for
the
total
benefit
population
for
the
North
Atlantic
region,
based
on
each
of
the
definitions
discussed
above.
For
the
counties
abutting
affected
water
bodies,
the
total
annual
value
per
acre
for
fish
habitat
services
provided
by
wetlands
is
$
3,017
for
all
households,
whereas
the
total
non­
use
only
value
is
$
2,891.
The
table
also
shows
two
estimates
of
the
values
for
households
living
within
the
32.43
mile
radius
area.
EPA
calculated
these
values
based
on
Pate
and
Loomis s
(
1997)
findings
that
WTP
for
improving
wetland
habitat
in
San
Joaquin
Valley
for
California
residents
outside
the
Valley
comprise
97.7
percent
of
the
WTP
of
the
Valley
residents
and
that
WTP
values
for
Oregon
residents
comprise
approximately
27
percent
of
the
WTP
of
the
Valley
residents.
As
shown
below,
EPA
assigned
the
value
per
household
from
the
Peconic
study
to
all
households
in
the
counties
abutting
affected
water
bodies
(
210,357
households).
For
households
beyond
these
coastal
counties
(
an
additional
527,354
households),
EPA
multiplied
the
Peconic
values
by
97.7
percent
and
27
percent,
to
provide
a
range
of
WTP
values.
Thus,
for
the
32.43
mile
radius
area,
the
total
annual
value
per
acre
for
fish
habitat
services
provided
by
wetlands
ranges
from
$
5,059
to
$
10,390
for
all
households,
while
the
total
non­
use
only
value
ranges
from
$
4,848
to
$
9,958.
Table
11:
Affected
Populations
in
North
Atlantic
Region
Households
in
Counties
Abutting
Affected
Water
Bodies
Households
Within
32.4
Mile
Radius
of
Each
FacilityFacility
State
Households
New
Boston
Essex
Middlesex
Suffolk
Norfolk
Plymouth
MA
275,419
1,582,673
561,220
278,722
248,827
168,361
Mystic
Essex
Middlesex
Suffolk
Norfolk
MA
275,419
1,592,865
561,220
278,722
248,827
Saugus
Resco
Essex
Middlesex
Suffolk
MA
275,419
1,520,553
561,220
278,722
Salem
Harbor
Essex
Middlesex
Suffolk
MA
275,419
1,393,291
561,220
278,722
Seabrook
Essex
Rockingham
MA
NH
275,419
511,616104,529
104,529
County(
ies)

­
21­
March
12,
2003
Table
12
also
shows
the
corresponding
values
if
the
estimated
WTP
share
for
 
shellfish
production
services 
rather
than
the
WTP
for
 
fish
production
services 
is
used.

Table
12:
Estimated
WTP
Values
for
Fish
in
Wetlands
for
the
North
Atlantic
Region
(
2002$)

$/
HH/
Acre/
Yeara
Total
Average
WTP/
Acre/
Year
for
Counties
Abutting
Affected
Water
Bodiesb
Total
WTP/
Acre/
Year
for
Households
in
Proximity
to
Affected
Water
Bodiesc
97.7%
27%
Fish
Total
Value
$
0.014
$
3,017
$
10,390
$
5,059
Non­
Use
Valued
$
0.014
$
2,891
$
9,958
$
4,848
Shellfish
Total
Value
$
0.016
$
3,268
$
11,258
$
5,481
Non­
Use
Valued
$
0.015
$
3,132
$
10,789
$
5,253
a.
Values
shown
are
WTP
per
household
per
additional
(
i.
e,
marginal)
acre
per
year
from
the
Peconic
study.
b.
Total
WTP
per
acre
is
calculated
as
household
WTP
per
acre
times
210,357,
the
average
number
of
households
in
the
counties
abutting
affected
water
bodies.
c.
Total
WTP
per
acre
is
calculated
as
household
WTP
per
acre
times
737,711,
the
average
number
of
households
within
32.43
miles
of
affected
water
bodies.
Values
are
adjusted
with
distance
as
explained
in
the
text.
d.
Total
non­
use
value
is
calculated
as
value
per
acre
for
non­
users
only
times
all
households
in
the
study
area.

2.4
Eelgrass
Values
for
Fish
and
Shellfish
Habitat
Services
2.4.1
Peconic
Estuary
Multiplying
the
value
per
household
from
Table
5
by
the
total
number
of
households
in
the
Peconic
study
area
(
73,423)
yields
the
value
per
acre
of
eelgrass
for
the
population
surrounding
the
Peconic
Estuary.
Table
13
shows
these
values.
The
study
defined
the
benefit
population
as
the
total
number
of
households
(
both
year­
round
and
seasonal)
in
the
five
towns
bordering
the
Peconic
Estuary.
For
the
Peconic
Estuary
region,
the
total
annual
value
per
acre
for
eelgrass
is
$
4,656;
and
the
total
non­
use
only
value
is
$
3,837.19
19
This
analysis
assumes
that
non­
use
values
are
the
same
for
both
users
and
non­
users
of
the
affected
resources.
Users
of
the
resource
likely
have
higher
non­
use
values
than
non­
users,
but
the
data
do
not
allow
us
to
test
this
hypothesis.

­
22­
March
12,
2003
Table
13:
Estimated
WTP
Values
for
Eeelgrass
(
SAV)
from
the
Peconic
Study
(
2002$)
20
$/
HH/
Acre/
Yeara
Total
WTP/
Acre/
Yearb
Total
Value
$
0.063
$
4,656
Non­
Use
Valuec
$
0.052
$
3,837
a.
Values
shown
are
WTP
per
household
per
additional
(
i.
e,
marginal)
acre
per
year.
b.
Total
WTP
per
acre
is
calculated
as
household
WTP
per
acre
times
73,423
total
households
in
the
study
area.
c.
Total
non­
use
value
is
calculated
as
value
per
acre
for
non­
users
only
times
all
households
in
the
study
area.

2.4.2
The
North
Atlantic
Region
Table
14
presents
the
values
per
acre
per
year
for
households
in
counties
abutting
affected
water
bodies,
and
for
households
within
the
larger,
32.43
mile,
radius.
The
total
annual
value
per
acre
for
eelgrass
(
SAV)
for
households
living
in
counties
abutting
affected
water
bodies
is
$
13,341
for
all
residents;
and
the
total
non­
use
only
value
is
$
10,993.
For
households
residing
within
32.43
miles
of
affected
water
bodies,
the
total
annual
value
per
acre
for
fish
and
shellfish
habitat
services
provided
by
SAV
ranges
from
$
22,371
to
$
45,949,
whereas
the
total
non­
use
only
value
ranges
from
$
18,434
to
$
37,863.

Table
14:
Estimated
WTP
Values
for
SAV
for
the
North
Atlantic
Region
(
2002$)
Total
WTP/
Acre/
Year
for
Total
WTP/
Acre/
Year
for
Counties
Abutting
Households
in
Proximity
to
$/
HH/
Acre/
Yeara
Affected
Water
Bodiesb
Affected
Water
Bodiesc
97.7%
27.0%
Total
Value
$
0.063
$
13,341
$
45,949
$
22,371
Non­
Use
Valued
$
0.052
$
10,993
$
37,863
$
18,434
a.
Values
shown
are
WTP
per
household
per
additional
(
i.
e,
marginal)
acre
per
year
from
the
Peconic
study.
b.
Total
WTP
per
acre
is
calculated
as
household
WTP
per
acre
times
210,357,
the
average
number
of
households
in
the
counties
abutting
affected
water
bodies.
c.
Total
WTP
per
acre
is
calculated
as
household
WTP
per
acre
times
737,711,
the
average
number
of
households
within
32.43
miles
of
affected
water
bodies.
Values
are
adjusted
with
distance
as
explained
in
the
text.
d.
Total
non­
use
value
is
calculated
as
value
per
acre
for
non­
users
only
times
all
households
in
the
relevant
study
area.

3.
Estimating
the
Value
of
Baseline
I&
E
Losses
for
the
North
Atlantic
Region
EPA
multiplied
the
estimated
number
of
acres
of
SAV
and
wetlands
needed
to
offset
the
I&
E
losses
for
the
North
Atlantic
region,
as
detailed
in
Section
1
above,
with
the
estimated
per
acre
values
of
SAV
and
wetlands,
to
evaluate
the
benefits
of
I&
E
reduction.
EPA
performed
this
analysis
for
SAV­
and
wetlands­
dependent
species
requiring
the
maximum
restoration
acres
among
these
for
which
productivity
estimates
are
available.

20
EPA
made
dollar
value
adjustments
using
the
Consumer
Price
Index
(
CPI)
for
all
urban
consumers
for
the
first
half
of
2002.

­
23­
March
12,
2003
Table
15
presents
the
estimated
values
for
SAV
restoration
for
the
North
Atlantic.
EPA
estimated
that
between
151
and
1,205
acres
of
revegetated
SAV
(
eelgrass)
is
needed
to
offset
estimated
average
annual
I&
E
losses
to
scup
and
northern
pipefish,
depending
on
whether
Brayton
Point
or
Pilgrim
is
used
for
the
productivity
estimates.
Based
on
the
estimated
value
per
acre
to
residents
of
counties
abutting
affected
water
bodies,
the
total
value
of
restoring
151
acres
of
eelgrass
is
$
2,014,450
per
year.
The
non­
use
only
value
is
$
1,659,243.
The
estimated
value
to
all
households
residing
within
32.43
miles
of
affected
water
bodies
ranges
from
$
3,377,982
to
$
6,938,316
per
year.
The
non­
use
only
value
ranges
from
$
2,783,496
to
$
5,717,253.

Based
on
the
estimated
value
per
acre
to
residents
of
counties
abutting
affected
water
bodies,
the
total
value
of
restoring
1,205
acres
of
eelgrass
is
$
16,075,574
per
year,
with
a
non­
use
only
value
of
$
13,246,458.
The
estimated
value
to
all
households
residing
within
32.43
miles
of
affected
water
bodies
ranges
from
$
26,956,743
to
$
55,368,683
per
year.
The
non­
use
only
value
ranges
from
$
22,212,667
to
$
45,624,433.

Table
15:
WTP
Values
for
SAV
Restoration
of
Fish
to
Offset
the
Baseline
I&
E
Losses
in
the
North
Atlantic
Region
(
2002$)
Species
Benefitting
from
SAV
Restoration
Acres
of
Required
SAV
Restoration
Total
Willingness
to
Pay
per
Acre
per
Year
Counties
Abutting
Affected
Water
Bodies
Scup
Threespine
stickleback
Northern
pipefish
151
acres
Total
Value
$
2,014,450
Non­
Use
Value
$
1,659,243
1,205
acres
Total
Value
$
16,075,574
Non­
Use
Value
$
13,246,458
All
Households
Residing
Within
32.43
Miles
of
Affected
Water
Bodies
(
High
Estimate)

Scup
Threespine
stickleback
Northern
pipefish
151
acres
Total
Value
$
6,938,316
Non­
Use
Value
$
5,717,253
1,205
acres
Total
Value
$
55,368,683
Non­
Use
Value
$
45,624,433
All
Households
Residing
Within
32.43
Miles
of
Affected
Water
Bodies
(
Low
Estimate)
Total
Value
$
3,377,982
Scup
Threespine
stickleback
Northern
pipefish
151
acres
Non­
Use
Value
$
2,783,496
1,205
acres
Total
Value
$
26,956,743
Non­
Use
Value
$
22,212,667
Table
16
presents
the
estimated
values
for
wetlands
restoration
for
the
North
Atlantic
region.
EPA
estimated
that
between
25,589
and
43,813
acres
of
restored
tidal
wetlands
is
needed
to
offset
estimated
average
annual
I&
E
losses
to
winter
flounder.
Based
on
the
estimated
value
per
acre
to
residents
of
counties
abutting
affected
water
bodies,
the
total
value
of
restoring
25,589
acres
of
coastal
wetlands
(
after
adjusting
for
the
estimated
portion
attributable
to
fish
production
services)
is
$
77,194,196
per
year,
with
the
non­
use
only
value
at
$
73,982,015.
For
all
households
within
32.43
miles
of
the
affected
water
bodies,
the
total
value
of
restoring
25,589
acres
of
coastal
wetlands
ranges
from
$
129,445,085
to
$
265,877,962
per
year,
while
the
non­
use
only
value
ranges
from
$
124,058,656
to
$
254,814,331.

­
24­
March
12,
2003
­
25­
Based
on
the
estimated
value
per
acre
to
residents
of
counties
abutting
affected
water
bodies,
the
total
value
of
restoring
43,813
acres
of
coastal
wetlands
is
$
132,170,436
per
year,
while
the
non­
use
only
value
is
$
126,670,601,
adjusted
to
fish
production
services
only.
For
all
households
within
32.43
miles
of
affected
water
bodies,
the
total
value
of
restoring
43,813
acres
of
coastal
wetlands
ranges
from
$
221,633,417
to
$
455,231,200
per
year,
with
the
non­
use
only
value
ranging
from
$
212,410,876
to
$
436,288,260,
adjusted
to
fish
production
services
only.
All
figures
are
2002
dollars.
This
analysis
does
not
include
fish
or
shellfish
losses
caused
by
thermal
discharges
that
are
covered
under
§
316(
a).

Table
16:
WTP
Values
for
Wetlands
Restoration
to
Offset
the
Baseline
I&
E
Losses
of
Fish
in
the
North
Atlantic
Region
(
2002$)

Species
Benefitting
from
Tidal
Wetlands
Restoration
Acres
of
Required
Wetlands
Restoration
Total
Willingness
to
Pay
per
Acre
per
Year
Counties
Abutting
Affected
Water
Bodies
Winter
flounder
Atlantic
silverside
Striped
killifish
25,589
acres
Total
Value$
77,194,196
Non­
Use
Value$
73,982,015
43,813
acres
Total
Value
$
132,170,436
Non­
Use
Value$
126,670,601
All
Households
Within
32.43
Miles
of
Affected
Water
Bodies
(
High
Estimate)

Winter
flounder
Atlantic
silverside
Striped
killifish
25,589
acres
Total
Value$
265,877,962
Non­
Use
Value$
254,814,331
43,813
acres
Total
Value$
455,231,200
Non­
Use
Value$
436,288,260
All
Households
Within
32.43
Miles
of
Affected
Water
Bodies
(
Low
Estimate)

Winter
flounder
Atlantic
silverside
Striped
killifish
25,589
acres
Total
Value$
129,445,085
Non­
Use
Value$
124,058,656
43,813
acres
Total
Value$
221,633,417
Non­
Use
Value$
212,410,876
Summing
the
values
for
wetlands
and
SAV
restoration
(
from
Tables
15
and
16),
the
total
value
of
losses
from
baseline
I&
E
of
the
six
species
identified
above
ranges
from
$
79,208,645
to
$
510,599,883.
The
non­
use
only
value
ranges
from
$
75,641,944
to
$
481,912,693.

EPA
notes
that
the
values
in
Tables
15
and
16
do
not
account
for
all
species
lost
to
I&
E
in
the
North
Atlantic
Region
(
e.
g.,
tautog)
and
include
benefits
for
species
not
affected
by
I&
E.
EPA
continues
to
evaluate
this
approach
as
one
method
for
estimating
comprehensive
non­
use
benefits
associated
with
this
regulation.

4.
Estimating
the
Benefits
of
the
Preferred
Option
for
the
North
Atlantic
Region
Table
17
shows
the
percent
reduction
in
I&
E
losses
for
each
of
the
affected
species
included
in
this
analysis
under
the
preferred
option.
The
preferred
option
is
expected
to
reduce
I&
E
losses
by
18.4
to
23.8
percent,
depending
on
species.
EPA
multiplied
the
percent
reduction
for
the
species
that
determined
the
number
of
acres
of
restoration
required
(
shown
in
Table
3)
by
the
total
number
of
acres
of
the
relevant
habitat
type
to
estimate
the
number
of
March
12,
2003
acres
of
restored
habitat
that
would
achieve
the
same
reduction
in
baseline
I&
E
losses.
For
tidal
wetlands,
winter
flounder
required
the
largest
number
of
acres
of
restoration.
Accordingly,
EPA
used
the
18.73%
reduction
in
I&
E
for
winter
flounder
to
calculate
the
benefits
of
the
preferred
technology.
Similarly,
EPA
used
the
18.97%
reduction
for
northern
pipefish
to
estimate
benefits
of
the
lower
bound
estimate
of
SAV
restoration,
and
the
23.75%
reduction
for
scup
to
estimate
upper
bound
benefits
for
SAV.

Table
17:
Estimated
Reductions
in
I&
E
Losses
under
the
Preferred
Option
Species
Percent
Reduction
Winter
flounder
18.73%
Atlantic
silverside
21.78%
Striped
killifish
18.43%
Threespine
stickleback
31.17%
Northern
pipefish
18.97%
Scup
23.75%

Source:
EPA s
analysis
Table
18
gives
the
range
of
WTP
values
for
the
preferred
option
for
the
North
Atlantic
region.
Summing
the
values
for
wetlands
and
SAV
restoration,
the
total
benefits
of
the
preferred
option
for
the
six
species
identified
above
ranges
from
$
14,840,614
to
$
98,414,866.
The
non­
use
value
only
ranges
from
$
14,171,720
to
$
92,552,594.

Table
18:
WTP
Values
for
Reduction
in
the
Baseline
I&
E
Losses
in
the
North
Atlantic
Region
under
the
Preferred
Option,
Based
on
WTP
for
Wetlands
and
SAV
Restoration
of
Fish
for
(
2002$)
Lower
Bound
Upper
Bound
Counties
Abutting
Affected
Water
Bodies
Total
Value
$
14,840,614
$
28,573,472
Non­
Use
Value
$
14,171,720
$
26,871,437
All
Households
Within
32.43
Miles
of
Affected
Water
Bodies
(
High
Estimate)
Total
Value
$
51,115,141
$
98,414,866
Non­
Use
Value
$
48,811,287
$
92,552,594
All
Households
Within
32.43
Miles
of
Affected
Water
Bodies
(
Low
Estimate)
Total
Value
$
24,885,868
$
47,914,165
Non­
Use
Value
$
23,764,215
$
45,060,065
5.
Per
Household
Values
of
Changes
in
I&
E
Losses
for
the
North
Atlantic
Region
Another
way
of
presenting
these
results
is
to
calculate
the
implied
per
household
WTP
for
households
residing
in
the
two
different
definitions
of
the
study
area.
Table
19
presents
results
of
these
calculations.
A
total
of
3.65
­
26­
March
12,
2003
million
households
live
in
the
counties
abutting
affected
water
bodies
while
4.2
million
households
live
within
a
32.4
mile
radius
of
affected
water
bodies.

5.1
Willingness
to
Pay
to
Eliminate
I&
E
The
results
of
this
analysis
implies
a
total
WTP
to
eliminate
all
I&
E
losses
of
$
21.70
to
$
40.62
and
non­
use
WTP
of
$
20.73
to
$
33.97
per
household
residing
in
the
counties
abutting
affected
water
bodies.
If
a
32.4
mile
radius
is
used
in
these
calculations,
the
implied
WTP
values
to
reduce
all
I&
E
losses
range
from
$
31.62
to
$
121.57
and
non­
use
WTP
range
$
29.92
to
$
113.68
per
household
residing
in
the
32.4
mile­
radius
area.
All
values
are
given
in
2002$.

Table
19:
Values
per
Household
for
Total
Affected
Population
of
the
North
Atlantic,
for
SAV
and
Wetlands
Restoration
Baseline
Losses
Preferred
Option
Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Households
in
Bordering
Counties
Total
Value
(
nonuse
+
use)
$
79,208,646
$
148,246,010
$
14,840,614
$
28,573,472
Total
value/
hh
$
21.70
$
40.62
$
4.07
$
7.83
Total
non­
use
value
$
75,641,944
$
139,917,060
$
14,171,720
$
26,871,437
Non­
use
value/
hh
$
20.73
$
33.97
$
3.44
$
6.52
Households
Within
32.4
Mile
Radius
(
high
estimate)
Total
Value
(
nonuse
+
use)
$
272,816,278
$
510,599,883
$
51,115,141
$
98,414,866
Total
value/
hh
$
64.96
$
121.57
$
12.17
$
23.43
Total
non­
use
value
$
260,531,584
$
481,912,693
$
48,811,287
$
92,552,594
Non­
use
value/
hh
$
61.46
$
113.68
$
11.51
$
21.83
Households
Within
32.4
Mile
Radius
(
low
estimate)
Total
Value
(
nonuse
+
use)
$
132,823,067
$
248,590,160
$
24,885,868
$
47,914,165
Total
value/
hh
$
31.62
$
59.19
$
5.93
$
11.41
Total
non­
use
value
$
126,842,152
$
234,623,543
$
23,764,215
$
45,060,065
Non­
use
value/
hh
$
29.92
$
55.35
$
5.61
$
10.63
5.2
Willingness
to
Pay
to
Reduce
I&
E
The
total
WTP
values
shown
in
Table
18
implies
a
total
WTP
to
reduce
I&
E
losses
of
$
4.07
to
$
7.83
and
non­
use
WTP
of
$
3.44
to
$
6.52
per
household
residing
in
the
counties
abutting
affected
water
bodies.
If
a
32.4
mile
radius
is
used
in
these
calculations,
the
implied
WTP
values
to
reduce
all
I&
E
losses
range
from
$
5.63
to
$
23.43
and
non­
use
WTP
range
from
$
5.61
to
$
21.83
per
household
residing
in
the
32.4
mile­
radius
area.
All
values
are
provided
in
2002$.

­
27­
March
12,
2003
6.
References
Adamowicz,
W.,
P.
Boxall,
M.
Williams,
and
J.
Louviere.
1998.
Stated
Preference
Approaches
for
Measuring
Passive
Use
Values:
Choice
Experiments
and
Contingent
Valuation.
American
Journal
of
Agricultural
Economics
80(
1):
64­
75.

Addelman,
Sidney.
1962a.
 
Orthogonal
Main­
Effect
Plan
for
Asymmetrical
Factorial
Experiments, 
Technometrics
4
No.
1
(
February):
21­
46.

Addelman,
Sidney.
1962b.
 
Symmetrical
and
Asymmetrical
Fractional
Factorial
Plans, 
Technometrics
4
No.
1
(
February):
47­
57.

Cameron,
T.
A.
1988.
A
New
Paradigm
for
Valuing
Non­
market
Goods
Using
Referendum
Data:
Maximum
Likelihood
Estimation
by
Censored
Logistic
Regression,
Journal
of
Environmental
Economics
and
Management
15(
3):
355­
379.

Hanemann,
W.
M.
1984.
Welfare
Evaluations
in
Contingent
Valuation
Experiments
with
Discrete
Responses,
American
Journal
of
Agricultural
Economics
66(
3):
332­
341.

Johnston,
R.
J.,
S.
K.
Swallow
and
D.
M.
Bauer.
2002a.
Spatial
Factors
and
Stated
Preference
Values
for
Public
Goods:
Considerations
for
Rural
Land
Development.
Land
Economics
78(
4):
481­
500.

Johnston,
R.
J.,
G.
Magnusson,
M.
Mazzotta
and
J.
J.
Opaluch.
2002b.
Combining
Economic
and
Ecological
Indicators
to
Prioritize
Salt
Marsh
Restoration
Actions.
American
Journal
of
Agricultural
Economics
84(
5):
1362­
1370.

Johnston,
R.
J.,
S.
K
Swallow
and
T.
F.
Weaver.
1999.
Estimating
Willingness
to
Pay
and
Resource
Trade­
offs
With
Different
Payment
Mechanisms:
An
Evaluation
of
a
Funding
Guarantee
for
Watershed
Management.
Journal
of
Environmental
Economics
and
Management
38(
1):
97­
120.

Johnston,
R.
J.,
T.
F.
Weaver,
L.
A.
Smith
and
S.
K.
Swallow.
1995.
Contingent
Valuation
Focus
Groups:
Insights
From
Ethnographic
Interview
Techniques.
Agricultural
and
Resource
Economics
Review
24(
1):
56­
69.

Johnston,
R.
J.,
T.
A.
Grigalunas,
J.
J.
Opaluch,
J.
Diamantides,
and
M.
Mazzotta.
2001a.
 
Valuing
Estuarine
Resource
Services
Using
Economic
and
Ecological
Models:
The
Peconic
Estuary
System
Study, 
Coastal
Management
30(
1).

Johnston,
R.
J.,
J.
J.
Opaluch,
T.
A.
Grigalunas,
and
M.
J.
Mazzotta.
2001b.
 
Estimating
Amenity
Benefits
of
Coastal
Farmland, 
Growth
and
Change
32:
305­
325.

Kennish,
M.
J.
2002.
Coastal
Salt
Marsh
Systems
in
the
U.
S.:
A
Review
of
Anthropogenic
Impacts.
Journal
of
Coastal
Research
17(
3):
731­
748.

Mazzotta,
Marisa
J.
1996.
Measuring
Public
Values
and
Priorities
for
Natural
Resources:
An
Application
to
the
Peconic
Estuary
System.
Doctoral
Dissertation,
University
of
Rhode
Island.

New
York
Conservationist
(
anonymous).
1997.
Tidal
Wetlands
in
New
York
State
51
(
psupp
4­
5).

Opaluch,
James
J.,
Thomas
Grigalunas,
Marisa
J.
Mazzotta,
Jerry
Diamantides,
and
Robert
Johnston.
1998.
 
Resource
and
Recreational
Economic
Values
for
the
Peconic
Estuary, 
Report
prepared
for
Peconic
Estuary
­
28­
March
12,
2003
Program,
Suffolk
County
Department
of
Health
Services,
Riverhead,
NY,
by
Economic
Analysis,
Inc.,
Peace
Dale,
Rhode
Island.

Opaluch,
James
J.,
Thomas
A.
Grigalunas,
Jerry
Diamantides
and
Marisa
J.
Mazzotta.
1995.
 
Environmental
Economics
in
Estuary
Management, 
Maritimes
Vol.
38,
No.
3
(
Winter):
21­
23.

Peconic
Estuary
Program
Comprehensive
Conservation
and
Management
Plan,
available
at
http://
www.
savethepeconicbays.
org/
ccmp/
.

Pate,
Jennifer
and
John
Loomis.
1997.
The
effect
of
distance
on
willingness
to
pay
values:
a
case
study
of
wetlands
and
salmon
in
California.
Ecological
Economics
Vol.
20
(
1997):
199­
207.

Schkade,
D.
A.
and
J.
W.
Payne.
1994.
 
How
People
Respond
to
Contingent
Valuation
Questions:
A
Verbal
Protocol
Analysis
of
Willingness
to
Pay
for
an
Environmental
Regulation. 
Journal
of
Environmental
Economics
and
Management
26(
1):
88­
109.

Schulze,
W.
D.,
R.
D.
Rowe,
W.
S.
Breffle,
R.
R.
Boyce,
and
G.
H.
McClelland.
1995.
Contingent
Valuation
of
Natural
Resource
Damages
Due
to
Injuries
to
the
Upper
Clark
Fork
River
Basin.
State
of
Montana
Natural
Resource
Damage
Litigation
Program.
Prepared
by
RCG/
Hagler
Bailly,
Boulder,
CO.

U.
S.
EPA.
2002.
Cooling
Water
Intake
Structures
­
Section
316(
b)
Proposed
Section
316(
b)
Phase
II
Existing
Facilities
Rule
Case
Study
Analysis.
U.
S.
EPA
Office
of
Water,
Washington
D.
C.,
EPA­
821­
R­
02­
002,
available
at
http://
www.
epa.
gov/
ost/
316b/.

Wyda,
J.
C.,
L.
A.
Deegan,
J.
E.
Hughes,
and
M.
J.
Weaver.
2002.
The
response
of
fishes
to
submerged
aquatic
vegetation
complexity
in
two
ecoregions
of
the
Mid­
Atlantic
Bight:
Buzzards
Bay
and
Chesapeake
Bay.
Esutaries,
25:
86­
100.

­
29­
March
12,
2003
Appendix
1:
The
Peconic,
Long
Island,
Estuary
System
CT
RI
Long
Island
Atlantic
Ocean
Long
Island
Sound
Block
Island
Sound
Riverhead
Southold
Southam
pton
East
Hampton
Great
Peconic
Bay
Little
Peconic
Bay
Gardiners
Bay
Plum
Is.

Montauk
Point
Shinnecock
Bay
Brook­
haven
Study
Area
Town
Boundary
Legend:

(
map
not
drawn
to
scale)
Shelter
Island
Orient
Point
Figure
1.
Depiction
of
the
Peconic
(
Long
Island)
Estuary
System
­
30­
March
12,
2003
Appendix
2:
The
North
Atlantic
Region
Appendix
3:
Landed
Shares
(
all
results
in
millions
of
Age­
1­
Equivalent
Fish)
C&
R
Landed
Forage
As
Percent
of
Total
Total
Total
C&
R
Landed
C&
R
Unlanded
C&
R
Forage
Salem
Impingement
3.2
1.2
0.114
2
3.6%
33.9%
62.5%
Entrainment
356.3
59
10.5
297
2.9%
13.6%
83.4%
Total
359.5
60.2
10.614
299
3.0%
13.8%
83.2%
Tampa
Impingement
0.42
0.357
0.027
0.062
6.4%
78.6%
14.8%
Entrainment
7710
10.5
1.3
7700
0.0%
0.1%
99.9%
Total
7710.42
10.857
1.327
7700.062
0.0%
0.1%
99.9%
Seabrook
Impingement
0.0131
0.0085
0.0014
0.0046
10.7%
54.2%
35.1%
Entrainment
4.5
0.3
0.0327
4.2
0.7%
5.9%
93.3%
Total
4.5131
0.3085
0.0341
4.2046
0.8%
6.1%
93.2%
Pilgrim
Impingement
0.0528
0.0512
0.0063
0.0016
11.9%
85.0%
3.0%
Entrainment
14.4
2.6
0.121
11.8
0.8%
17.2%
81.9%
Total
14.4528
2.6512
0.1273
11.8016
0.9%
17.5%
81.7%
JR
Whiting
Impingement
21.5
20.7
2.62
0.768
12.2%
84.1%
3.6%
Entrainment
1.8
1.7
0.166
0.144
9.2%
85.2%
8.0%
Total
23.3
22.4
2.786
0.912
12.0%
84.2%
3.9%
Monroe
Impingement
35.8
35.4
4.4
0.37
12.3%
86.6%
1.0%
Entrainment
11.6
11.2
1.2
0.392
10.3%
86.2%
3.4%
Total
47.4
46.6
5.6
0.762
11.8%
86.5%
1.6%
Brayton
Impingement
0.0693
0.029
0.0035
0.0403
5.1%
36.8%
58.2%
Entrainment
3.8
0.606
0.043
3.2
1.1%
14.8%
84.2%
Total
3.8693
0.635
0.0465
3.2403
1.2%
15.2%
83.7%
Aggregate
of
Applicable
Case
Studies
Impingement
61.0552
57.7457
7.1722
3.2465
11.7%
82.8%
5.3%
Entrainment
8102.4
85.906
13.3627
8016.736
0.2%
0.9%
98.9%
Aggregate
8163.4552
143.6517
20.5349
8019.983
0.3%
1.5%
98.2%
Aggregate
of
Case
Studies
(
less
Great
Lakes
facilities
of
Monroe
and
JR
Whiting)
Impingement
3.7552
1.6457
0.1522
Entrainment
8089
73.006
11.9967
Aggregate
8092.7552
74.6517
12.1489
2.1085
4.05%
39.77%
0.75%
0.77%
56.15%
8016.2
0.15%
99.10%
8018.309
0.15%
99.08%
Aggregate
of
Great
Lake
Case
Studies
(
Monroe
&
JR
Whiting)
Impingement
57.3
56.1
7.02
1.138
12.25%
85.65%
1.99%
Entrainment
13.4
12.9
1.366
0.536
10.19%
86.07%
4.00%
Aggregate
8041.4859
27.4167
6.5024
8014.306
0.08%
0.26%
99.66%

­
31­
