Beach nourishment as a dynamic capital accumulation problem

Martin D. Smith
a,
, Jordan M. Slott
b
, Dylan McNamara
c
, A. Brad Murray
d
a
Nicholas School of the Environment, Department of Economics, Box 90328, Duke University, Durham, NC 27708, USA
b
Sun Microsystems Laboratories, Sun Microsystems, Inc. 1 Network Drive, Burlington, MA 01803, USA
c
Department of Physics and Physical Oceanography, University of North CarolinaWilmington, 601 South College Road, Wilmington, NC 28403, USA
d
Nicholas School of the Environment and Center for Nonlinear and Complex Systems, Box 90230, Duke University, Durham, NC 27708, USA
a r t i c l e i n f o
Article history:
Received 9 March 2007
Available online 18 January 2009
JEL classication:
Q24
Q54
Keywords:
Beach nourishment
Optimal rotation
Natural capital
Erosion
Shoreline stabilization
Coastal management
a b s t r a c t
Beach nourishment is a common coastal management strategy used to combat
erosion along sandy coastlines. It involves building out a beach with sand dredged
from another location. This paper develops a positive model of beach nourishment
and generates testable hypotheses about how the frequency of nourishment
responds to property values, project costs, erosion rates, and discounting. By treating
the decision to nourish as a dynamic capital accumulation problem, the model produces
new insights about coupled economic geomorphological systems. In particular,
determining whether the frequency of nourishment increases in response to physical
and economic forces depends on whether the decay rate of nourishment sand exceeds
the discount rate.
& 2009 Elsevier Inc. All rights reserved.
1. Introduction
Most US coastlines are either moderately or severely eroding [14,22]. Both wave-driven alongshore sediment transport
and sea-level rise contribute to coastal erosion, and climate change is likely to intensify both of these forces [30]. At the
same time, more humans are living in the coastal zone [16,31]. Cost estimates of avoiding inundation from a 1m rise in sea
level are between $270 billion and $475 billion [32]. A recent Heinz report suggests that over the next 60 years, erosion will
affect 25% of US structures within 500ft of the coastline, and the lost property value without any further build-out on
vacant lots would be between $3.3 billion and $4.8 billion [14]. These estimates reect diminished property values from
erosion without accounting for any potential damages to structures or lost recreational amenities that are not fully
capitalized into land values.
Trends in the coastal zone point to an inevitable conict between coastal developments and an encroaching shoreline.
As coastal erosion takes place, residential and commercial properties as well as coastal infrastructure are threatened.
Humans can, and do, intervene in the coastal zone to defend against shoreline changes. To address erosion, coastal
managers and engineers can pursue beach nourishment,
1
build hard structures like sea walls, or simply move or abandon
coastal property [17,25].
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/jeem
Journal of
Environmental Economics and Management
ARTICLE IN PRESS
0095-0696/$ - see front matter & 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jeem.2008.07.011

Corresponding author. Fax: +1919684 8741.

E-mail address: marsmith@duke.edu (M.D. Smith).
1
Nourishment is also referred to as re-nourishment or replenishment to reect the need for repeated applications of sand if this strategy is pursued.
Journal of Environmental Economics and Management 58 (2009) 5871
Increasingly, communities are relying on beach nourishment to hold back the sea. Beach nourishment is the practice of
building out a beach with sand that has been dredged from another location [6]. It is a common beach management
strategy in the United States for sandy coastlines along the Atlantic coast and in the Gulf of Mexico [33,34]. At least $2.5
billion ($2002) was spent on nourishment projects between 1950 and 2002 in the US, and the frequency of nourishment
has increased dramatically in recent years with federal appropriations of $787 million from 1995 to 2002 [23]. Beach
nourishment only defends the shoreline temporarily and must be re-done periodically [24]. In this paper, we develop a
model that explains this Sisyphean pattern and suggests what we might expect to see in the future if sea-level continues to
rise and beach erosion accelerates.
We specically develop a capital-theoretic model of a representative community that decides how often to nourish its
beach. We adapt a FaustmannHartman model from the rotational forestry literature to characterize the continuous ow of
amenities from beaches and the periodic nature of beach nourishment interventions [9,13]. As in Faustmann, the choice of
nourishment frequency depends on balancing the marginal net benets (NB) of a single rotation with the interest lost from
delaying all future rotations.
Our model effectively endogenizes the depreciation rate of natural beach capital because nourishment changes the short-
run morphodynamics of the beach. The model generates both expected and surprising predictions. Specically, we nd that
communities will nourish more often if (1) the baseline property values are higher, (2) un-nourished beaches erode faster, (3)
the hedonic price of beach width is higher, (4) xed costs of nourishment are lower, or (5) the discount rate is higher.
However, a higher variable cost of nourishment sand could result in increased or decreased frequency of beach nourishment.
It is thus possible that the input demand for nourishment sand could slope upwards. This result is particularly striking in light
of the scarcity of nourishment-quality sand and the possibility of greater scarcity in the future [10,15,27]. Similarly, a higher
decay rate of nourishment sand could lead to higher or lower rates of nourishment. These latter results lead to new insights
about linked economic and geomorphological models. Signing the change in nourishment frequency hinges on whether the
rate of foregone interest (nancial capital depreciation) exceeds the rate of sand loss (natural capital depreciation).
2. Background
Beaches provide economic value through storm protection and amenity ows, and the economic literature has focused
on quantifying these values. While hedonic models reect storm risks generally [12], some studies specically show that
threats from coastal erosion are capitalized into housing values [7,18]. Amenity ows can also be capitalized into housing
prices; hedonic studies consistently nd a positive price for coastal properties being on the waterfront, in close proximity
to water, or having a view of the water [24,20,25]. In addition, there is empirical support that beach width has a positive
hedonic price [20,21,28], which could reect a combination of amenity and storm protection values. Nevertheless, not all
amenity ows are capitalized directly into beachfront properties. Recreation demand studies nd positive values for beach
trip days using revealed preference [2,3] and stated preference methods [20].
2
Thus, beach communities have an incentive
to preserve beach quality and width to support local tourism.
Several economic studies conduct normative analyses of coastal management strategies. Edwards and Gable [8] compare
amenity values from a hedonic model to costs of periodic beach nourishment and nd support for the efciency of beach
nourishment. Yohe et al. [36] formulate a dynamic model to ask how long coastal communities should defend against sea-
level rise with an application to Charleston, South Carolina. The model incorporates continuous depreciation of capitalboth
land values and structuresas a function of rising sea level, and costs of protection are incurred in continuous time. Parsons
and Powell [25] compare the benets and costs of nourishing beaches versus beach retreat for Delaware and nd that
nourishment provides net benets over a 50-year horizon. Landry et al. [20] combine property benets from a hedonic model
with stated preference data on recreational use to estimate benets of retreat and nourishment strategies on Tybee Island,
Georgia. With an application to Seabrook Island, South Carolina, Woglom [35] compares the costs of nourishment with the
costs of allowing natural shoreline change and moving structures. She nds that nourishment is a preferred strategy when
the interval between nourishments is long (410 years), but the preferred strategy reverses when the interval between
nourishments is short (o5 years). This result leaves open the question of whether nourishment will be a dominant strategy if
communities choose the nourishment interval optimally. Finally, Landry [19] combines empirically driven benet and cost
estimates with the insights from [36] to analyze optimal beach nourishment on Tybee Island using dynamic programming.
Consistent with real-world practices, Landry nds that the optimal nourishment control is periodic. Landrys model also
illustrates how to determine optimal beach width in a dynamic context.
Taken together, these normative analyses and valuation studies highlight factors that will inuence coastal
management. However, they do not generate specic testable hypotheses about coastal management decisions across
communities and across time. Our paper aims to ll this gap in the literature by providing a positive analysis of
nourishment decisions. We propose that choosing whether to nourish a beach and when can be viewed as a dynamic
ARTICLE IN PRESS
2
In a discrete choice recreation demand model, Parsons et al. [26] nd that wider beaches are preferred up to a point. Eventually, beachgoers receive
disutility from wider beaches, suggesting that the recreational value of beach width may have an inverted U-shape. Combined recreational and storm
protection benets of wider beaches are unclear, but it seems reasonable to assume that overall benets are concave in width.
M.D. Smith et al. / Journal of Environmental Economics and Management 58 (2009) 5871 59
capital accumulation problem. We draw on the Faustmann and Hartman models in forest economics to cast nourishment as
an optimal rotation problem [9,13].
In order to explore the capital-theoretic nature of beach management, it is essential to incorporate the geologic response
of a shoreline to beach nourishment in a more physically accurate way than previous treatments in the economic literature. In
the absence of nourishment, beaches erode for any of several reasons, including sea-level rise [5,16], shifts in the shoreline
position on large scales caused by wave-driven sediment transport [1], and coastal structures built by humans [17]. Although
widely believed to cause permanent beach erosion, the effects of storm landfalls on coastal areas are only temporary, and
calm seas restore the wide beaches very soon afterwards [17]. Erosion rates can also differ greatly on beaches within the same
region [30]. To build intuition about how individual communities make nourishment decisions, however, we abstract away
from the spatial particularities of erosion and assume that the background rate of beach erosion is constant.
Beach nourishment places sand on an eroding section of beach, restoring it to some width. Beach nourishment,
therefore, creates an idealized rectangular bump in the plan-view (i.e. birds-eye view) shoreline trend (Fig. 1a, darkened
rectangle). Wave action, assuming most waves approach the shoreline nearly straight-on, tends to spread beach
nourishment sediment laterally, thus smoothing the plan-view bump. This process is known as alongshore sediment
transport. The rate of this smoothing decays exponentially over time, where the time until only half the original beach
nourishment sand volume remains is on the order of several years to one decade [6].
In the absence of beach nourishment, the prole of the nearshore seabed forms a time-averaged, concave up,
equilibrium shape (Fig. 1b, idealized as linear), resulting from a balance between forces tending to cause onshore sediment
transport (from waves) and forces tending to cause offshore sediment transport (chiey gravity outside the zone of
breaking waves). Beach nourishment sand is typically placed only on the dry beach and the region immediately seaward
where waves break and run-up on the beach (the surf and swash zone, respectively). The prole-view wedge created by
beach nourishment disturbs the cross-shore equilibrium prole (Fig. 1b, darkened triangle). Wave action and gravity,
however, tend to redistribute this sediment over the nearshore seabed to restore the equilibrium prole (Fig. 1b, dotted
line) on the time-scale of years. Like the alongshore smoothing of beach nourishment sand, the rate at which sand is
redistributed in the cross-shore also decays exponentially over time [6]. Taken together, the cross-shore and alongshore
nearshore response makes it appear as if nourished beaches erode faster than the background erosion rateU
With potential future increases in erosion rates, coastal property values, and scarcity of sand to nourish beaches, how
often do we expect to see communities nourishing their beaches? By choosing how often to build out beaches through
nourishment projects, coastal managers choose the depreciation rate of beach capital because the net rate of erosion is
effectively endogenous for nourished beaches. With this critical feature in the model, we analyze a representative community
that makes nourishment decisions independently of other communities. Consequently, the testable hypotheses from our
comparative static results apply to nourishment frequencies across different communities and within communities
across time.
3. The model
We explore beach nourishment as a dynamic capital accumulation problem. The model presents a positive analysis of
what we might expect to see if coastal managers follow a dynamically optimal capital accumulation path. We follow the
Hartman [13] approach to amenity ows from a standing stock of forest. In our case, benets accrue as a function of the
stock of beach width. As in a forest rotation problem, this stock changes continuously over time and is reinitialized each
time a control is applied. Cutting the forest returns the stock of timber volume to zero, whereas for beach management,
nourishment implies returning the beach to some initial beach width. Our model differs from the forestry literature in that
there is no harvest benet. All of the benets in our model are ows (amenity and storm protection), but the time-varying
variable cost, essentially negative benets, provide an analog to timber benets in the Hartman model. Assuming the costs
of nourishment are incurred at time zero, net benets of a single nourishment event can be written as a function of the
ARTICLE IN PRESS
Fig. 1. Cross-shore and alongshore response of the coastline to beach nourishment: (a) Beach nourishment creates an idealized rectangular plan-view
bump in the shoreline (darkened rectangle). Wave action spreads beach nourishment sand laterally, smoothing the bump; (b) beach nourishment
creates an idealized triangular prole-view wedge, distributing the cross-shore equilibrium prole (here, idealized as linear). Wave action redistributes
beach nourishment sand to restore the prole its time-averaged equilibrium shape.
M.D. Smith et al. / Journal of Environmental Economics and Management 58 (2009) 5871 60
nourishment interval T:
NBT BT CT, (1)
where C(T) denotes the cost associated with the nourishment project, and B(T) is the benets function. Both of these, in
turn, will depend on the dynamics of beach erosion. Unlike the FaustmannHartman problem, costs are incurred at the
beginning of a rotation because nourishment takes place then, and the benets accrue subsequently. In the
FaustmannHartman model, costs are incurred at the end of a rotation when the forest is cut and re-seeded. Another
crucial difference is that costs in our model are a function of the rotation length. In the FaustmannHartman framework,
there are no variable costs; the costs of cutting and re-seeding are a constant amount per unit of area. In our model,
nourishment sand is a large fraction of the cost. Normalizing to a given alongshore distance, the amount of sand is
proportional the width of beach build-out. This, in turn, is a function of the length of time T that the beach has been
allowed to erode. We develop the analytical framework for costs after introducing the state equation that describes beach
erosion.
3
In order to describe costs of nourishment, it is necessary to specify beach dynamics. Unlike previous economic
literature, our beach erosion dynamics capture both a background erosion rate and the exponentially decaying rate due to
beach nourishment. In the absence of nourishment, the beach erodes at a constant rate of g ft/year (Fig. 1a). Let x(t)
represent the beach width at time, t, and assume beach nourishment restores the beach to some initial width, x
0
, reecting
some realities in the coastal zone, namely xed locations of beachfront houses, utility pipelines and conduits, and
transportation infrastructure. We combine the exponentially decaying, accelerated erosion rate from the separate cross-
shore and alongshore responses into a single term. We, therefore, express the beach width, x(t) as
xt 1 mx
0
me
yt
x
0
gt, (2)
where 0pmp1 is the fraction of the initial beach width, x
0
, which erodes exponentially at rate, yX0.
4
The remainder of the
beach width, (1m), erodes linearly at a rate, gX0. By inspection, x(0) x
0
(Fig. 1a). Differentiating Eq. (2) with respect to
time yields the state equation for beach width:
_
xt mye
yt
x
0
g. (3)
We nowconsider a series of beach nourishments that are done on a periodic basis. Fig. 2 illustrates the beach width over
a 150-year time horizon with a 10-year nourishment interval (rotation length).
5
Initial beach width is 100ft, baseline
erosion is 2ft/year, 35% of the beach decays exponentially due to the nourishment return to equilibrium prole effect, and
the nourishment decay rate is 0.10 (roughly half of the nourishment sand lost in 7 years).
6
Notice that this gure appears
much like a gure that depicts the standing wood volume in a forest rotation problem. However, in the nourishment case,
ARTICLE IN PRESS
0 50 100 150
55
60
65
70
75
80
85
90
95
100
Time (years)
B
e
a
c
h

W
i
d
t
h

(
f
e
e
t
)
Fig. 2. Beach width on a 10-year rotation.
3
As a positive model of beach nourishment, C(T) might only include the engineering, planning, and construction costs of a project. However, to
approach the problem normatively, C(T) would also have to include potential non-market damages to the benthic environment, sea birds, and risks to sea
turtles [11,24]. We note these critical normative dimensions of the beach management problem but leave them as topics for future research.
4
The parameter y captures the half-life of a beach nourishment, typically in the range of 35 years. For empirical approaches, see [6].
5
We use nourishment interval and rotation length interchangeably. When the dynamic problem is time autonomous, the optimal sequence of
nourishment intervals will be a constant interval such that the problem has the rotational feature of the FaustmannHartman model.
6
On long timescales (greater than decades, but on the order of (1m)x
0
/g)), m and y would be functions of absolute time that the beach has been under
nourishment because un-nourished beaches surrounding the community continue to erode at the baseline rate. The positions of these un-nourished
M.D. Smith et al. / Journal of Environmental Economics and Management 58 (2009) 5871 61
the gure is inverted. This makes sense given that the stock of beach width decays, whereas the stock of forest grows. A
rotation resets the forest volume at zero but resets the beach width at its maximum, x
0
.
By changing the rotation length, the manager effectively controls the average beach width and thus the ow of
amenities and storm protection. The associated use of sand depends on the baseline erosion and the exponential decay of
nourishment sand. If the latter effect is zero, then the sand used is the same on average regardless of the rotation length.
The top panel of Fig. 3 depicts this effect. The cumulative sand use paths for a 10-year rotation and a 5-year rotation
overtake each other throughout the 150-year horizon. These overlaps reect only the discreteness of the rotations. The total
amount of sand that is lost from the system is independent of human decisions. In the bottom panel in which nourishment
sand decays exponentially, the two sand use paths diverge. Here, the 5-year rotation uses more sand over time because
more frequent nourishments mean that the beach spends more time in the steeply sloped portion of the state equation (3).
Also note that in both cases, the cumulative sand is higher in the bottom panel compared with the top panel with baseline
erosion only. Thus, human interventions in the geomorphological system increase the amount of sand lost from the
nearshore environment, highlighting what some coastal scientists perceive as the wastefulness of beach nourishment [27].
By introducing benets of beach width, we can weigh these losses of nourishment sand against gains and explore the
circumstances under which, from the communitys perspective, it is optimal to nourish the beach or to allow the shoreline
to retreat naturally.
There are xed and variable components of nourishment project costs [6]. Fixed costs are associated with capital
equipment needed for dredging and spreading sand as well as the costs of planning, obtaining permits, and preparing
environmental impact statements. Variable costs are a function of the amount of nourishment sand required, which is
proportional to the width of beach build-out. Since the shoreline is pinned to x
0
each time nourishment occurs, the amount
of sand is proportional to cumulative erosion. We can thus write the cost function as
CT c fx
0
xT, (4)
where c is the xed cost and f the variable cost of beach sand and includes the engineering conversion from beach width to
sand volume. We assume cX0 and fX0. Substituting for x(T) and simplifying, the cost can be expressed as
CT c fmx
0
1 e
yT
gT. (5)
Because beach widths change continuously but hedonic models capture values of standing capital stocks, it is necessary
to convert stocks to ows for understanding benets of nourishment. Let G(x(t)) capture the value of the stock of beach
width. We convert this to a ow benet using the discount rate. Though only some benets from beaches are capitalized
into home values on the beachfront, while others accrue to beach visitors and to the broader local community, we assume
ARTICLE IN PRESS
0 50 100 150
0
100
200
300
Sand Use with Baseline Erosion Only
C
u
m
u
l
a
t
i
v
e

S
a
n
d

U
s
e
d
5-year Rotation
10-year Rotation
0 50 100 150
0
200
400
600
800
Sand Use with Exponential Decay of Nourishment
Time
C
u
m
u
l
a
t
i
v
e

S
a
n
d

U
s
e
d
5-year Rotation
10-year Rotation
Fig. 3. Sand use patterns with and without nourishment decay.
M.D. Smith et al. / Journal of Environmental Economics and Management 58 (2009) 5871 62
that G(x(t)) encompasses total community values of beach width. Thus,
BT
Z
T
o
e
dt
dGxt dt, (6)
where d is the discount rate, which we assume is strictly positive, and we assume G0 0; @G=@x40; @
2
G=@x
2
o0. The
curvature assumption suggests that as the beach gets wider, the marginal benet of adding more beach width declines, and
there is some empirical support for this assumption [26,28]. Eq. (6) is similar to Hartmans [13] value of the standing stock
of forest. Instantaneous amenities are proportional to the standing stock, which is changing continuously. Computing total
amenities involves integrating the present value ow over the course of a forest rotation. The beach nourishment problem
presents two crucial differences: (1) the beach is eroding over time, not growing (as a forest) and (2) the ow conversion in
the beach problem uses the discount rate, which we assume is the same as the capitalization rate for rent into
housing price; in the forestry problem the ow conversion is a measure of non-timber benets that may not be the same
(or related to) the capitalization rate of the forest.
As in the FaustmannHartman style forestry literature, the problem for communities is to determine how often to
nourish the beach. The choice is a sequence of Ts:
vT
1
; T
2
; T
3
; . . . ; T
n
NBT
1
e
dT
1
NBT
2
e
dT
2
NBT
3
e
dT
n1
NBT
n
, (7)
where v is total present value, and each T
i
represents the absolute time from the beginning of the planning horizon (not the
marginal time from the previous rotation). Assuming that the instantaneous benets function is time autonomous and the
erosion dynamics are stationary, we can write the value of an innite nourishment rotation as an innite geometric series:
vT
X
1
k0
e
dkT
NBT NBT=1 e
dT
, (8)
where T is the length of time since the previous rotation.
The community chooses a T* to solve the following maximization problem:
maxvT BT CT=1 e
dT
. (9)
The rst-order condition is
@v T
@T
B
0
T C
0
T1 e
dT
de
dT
BT CT=1 e
dT

set
0. (10)
Optimal nourishment occurs where T* solves Eq. (10). Two notes are in order. First, by optimal we mean optimal from the
point of view of the coastal manager of a particular location; socially optimal would need to include ecological costs of
nourishment.
7
In a positive sense, one could interpret Eq. (10) to represent narrowly the net benets of nourishment that
are capitalized into property values or values of the surrounding local community. Second, it is possible that no nourishing
is the optimal strategy. To ensure that T* is in fact optimal, one need only check that v4 0 in Eq. (9).
Multiplying Eq. (10) through by (1e
dT
), we see that the optimal nourishment interval occurs where the difference
between marginal benets and marginal costs of a single rotation is exactly offset by the interest payment lost on delaying
all future rotations. This can be seen from the following expression in which the right-hand side is simply the discount rate
multiplied by the present value of all rotations after the rst one:
B
0
T C
0
T dBT CT=e
dT
1. (11)
So far, the intuition generally matches that of the classic Faustmann problem. To clarify interpretation, the Faustmann
rst-order condition can be written (in our notation) as
B
0
T dBT c dBT c=e
dT
1, (12)
where costs in the Faustmann model include only xed costs c. In Faustmann, there are no marginal costs such that the
left-hand side of (12) is different from that of (11), and the right-hand side of (12) has c rather than C(T) as in (11). Another
important difference is that the term d(B(T)c) in (12) does not appear in (11). This reects the fact that Faustmann
generates a lumped benet when the forest is cut, and delaying incurs lost interest on that lumped benet. There are no
such lumped benets in the beach problem; all benets are continuous ows as in Hartmans value of standing forest. The
termd(B(T)c)/(e
dT
1) is interest on future rent and is the opportunity cost of keeping land in forestry [29]. By analogy, for
beaches this term is the opportunity cost of pursuing nourishment as an erosion management strategy.
To generate results from the model, we take derivatives of the benet and cost functions to nd expressions
for the pieces of (10). Applying Liebnizs rule to the benets function in (6), the marginal benet of extending the
ARTICLE IN PRESS
(footnote continued)
beaches ultimately affects the alongshore sediment transport within the nourished region. We treat m and y as xed parameters for analytical tractability
and to develop some insights about the nourishment problem, but we acknowledge this caveat as a limitation of our model.
7
Socially optimal nourishment would also need to include spatial externalities of nourishment (both positive and negative) especially given the
emerging evidence that shoreline perturbations can propagate over large spatial scales [1].
M.D. Smith et al. / Journal of Environmental Economics and Management 58 (2009) 5871 63
rotation is
B
0
T e
dT
dGxT e
dT
dG1 mx
0
me
yT
x
0
gT. (13)
Thus, the marginal benet of extending a single rotation is decreasing. The corresponding marginal cost of delaying a
rotation is
C
0
T fg fmyx
0
e
yT
. (14)
By inspection of (14), the marginal cost of delaying is decreasing. The cost of additional sand is constant, but the
additional sand required is decreasing over time because the total erosion rate decreases and approaches the baseline rate
as the beach returns to equilibrium prole (in both prole- and plan-view senses).
4. Comparative statics of beach nourishment rotations
We cannot analytically sign most of the comparative statics from the model described above because there are two
countervailing forces in the rst-order condition (10). The optimal rotation length (T*) balances the difference between
marginal benets and marginal costs of a single rotation with the interest lost on delaying all future rotations. While this
property characterizes the original Faustmann formula as well, the simple form of the benets function combined with the
assumption of only xed costs in Faustmann leads to cancellations and signable comparative statics.
For the cost parameters, however, it is possible to sign comparative statics by assuming that the second-order condition
holds at the maximum. Because the xed cost c is neither in the benets function nor the marginal cost function, the
comparative static result is relatively simple.
Proposition 1. The optimal rotation length increases if the xed cost of nourishment increases.
Proof.
dT
n
dc

de
dT
n
1 e
dT
n

2
,
@
2
vT
n

@T
2
!
.
The numerator is positive, and the denominator is negative by the second-order condition. Hence, dT
n
=dc40. &
In the next proposition, we explore the importance of the variable cost of nourishment on the optimal rotation length T*.
Here, the sign will neither be strictly positive nor negative, but we can sign the parts and determine the geomorphological
and economic drivers.
Proposition 2. The optimal rotation length can increase or decrease if the variable cost of nourishment increases. The sign
depends on whether the nourishment sand decay rate is higher than the discount rate, the fraction of beach width that decays
exponentially, and the baseline erosion rate g.
Proof. From Proposition 1, we know that signfdT

=dfg signf@
2
vT
n
=@f@Tg. Dene G 1 e
dT
n

2
f@
2
vT
n
=@f@Tg, such
that signfdT

=dfg signfGg. We can then write G as the sum of two terms that have known signs:
G y dmx
0
e
ydT
n
de
dT
n
ye
yT
n
mx
0
|{z}
A
gde
dT
n
T
n
e
dT
n
1
|{z}
B
.
We will show that the sign{A} depends only on whether y4d. First, by inspection, we see that when y d, the A term is
zero. Factoring out mx
0
, A40 if y de
ydT
n
de
dT
n
ye
yT
n
40. We now use second-order Taylor approximations to
sign this expression, noting that e
kT
1 kT k
2
T
2
=2. Thus,
A y y
2
T
n

y
3
2
T
n

2
d d
2
T
n

d
3
2
T
n

2
y y
2
T
n
ydT
n

y
3
2
T
n

yd
2
2
T
n

2
y
2
dT
n

2
d dy d
2
T
n

y
2
d
2
T
n

d
3
2
T
n

2
yd
2
T
n

dy
2
T
n

2
y d .
Thus, sign{A} sign{yd}. To determine the sign{B}, we note that at T* 0, B 0. @B=@T
n
j
T
n
0
gd
2
e
dT
n
o0. Hence, B is
strictly non-positive. Putting together the pieces, whenever yod, both terms are negative, and increasing the variable cost
increases nourishment frequency. When y4d, increasing variable cost increases nourishment frequency if this difference
(between y and d) is small, baseline erosion is high, or the share of beach width that erodes exponentially (m) is small. &
One explanation for Proposition 2 is that the variable cost inuences the relative importance of xed cost of
nourishment in determining T*. When variable costs increase, the relative importance of xed costs decrease, reducing the
incentive to delay. Consider for simplicity that there was only baseline erosion. Then the total variable cost (undiscounted)
would be the same whether nourishment intervals were high or low because the beach is always returned to x
0
. In this
ARTICLE IN PRESS
M.D. Smith et al. / Journal of Environmental Economics and Management 58 (2009) 5871 64
simplied setting, total undiscounted xed costs increase as T* decreases, whereas total undiscounted variable costs do not
change with T*. If there were no xed costs, we would simply nourish continuously and always maintain the beach at x
0
for
maximum benets. When we introduce exponential decay of nourishment sand, more sand is lost quickly, and the variable
costs of nourishing frequently are substantially higher.
The discount rate and exponential decay of nourishment sand play important roles in Proposition 2. In essence, the
manager endogenizes the depreciation rate of beach capital. Depreciation here is a mixture of exponential decay of
nourishment sand and linear sand loss from baseline erosion. Nourishment frequency determines the net rate of
depreciation. Consider rst the effect of discounting in isolation. When the discount rate is high, the weight placed on the
marginal net benets in the short run is high relative to the weight placed on the discounted stream of future rotations. An
increase in the variable cost of sand increases the marginal cost of a rotation that must be offset by increasing the marginal
benets. Since marginal benets are decreasing in T, T* must decrease. When the discount rate is low, the weight on all
future rotations is high. This, in turn, is based on the difference between total benets and total costs of a rotation. With a
low discount rate, total benets of a rotation must increase to compensate for an increase in the total cost of a rotation due
to the variable cost increase. This requires increasing T*.
Nowconsider the exponential decay of sand together with the discount rate. An increase in variable cost of nourishment
will increase the implicit undiscounted cost of maintaining the capital stock. When sand decays rapidly, marginal cost of a
rotation is high but decreasing in T. So, one can lower the marginal cost by increasing T*. If the discount rate is relatively
ARTICLE IN PRESS
0 5 10 15 20 25 30 35
0
5
10
15
20
25
30
35
40
Marginals from a Single Rotation - phi = 1,
gamma = 0.5
Rotation Length
MB
MC
MNB
NB Rent
0 5 10 15 20 25 30 35
0
5
10
15
20
25
30
35
40
Marginals from a Single Rotation - phi = 2,
gamma = 0.5
Rotation Length
MB
MC
MNB
NB Rent
0 5 10 15 20 25 30 35
0
5
10
15
20
25
30
35
40
Marginals from a Single Rotation - phi = 1,
gamma = 2
Rotation Length
MB
MC
MNB
NB Rent
0 5 10 15 20 25 30 35
0
5
10
15
20
25
30
35
40
Marginals from a Single Rotation - phi = 2,
gamma = 2
Rotation Length
MB
MC
MNB
NB Rent
Fig. 4. (a) Illustration of Proposition 2low erosion, low sand cost. (b) Illustration of Proposition 2low erosion, high sand cost. (c) Illustration of
Proposition 2high erosion, low sand cost. (d) Illustration of Proposition 2high erosion, high sand cost.
M.D. Smith et al. / Journal of Environmental Economics and Management 58 (2009) 5871 65
low, we saw above that there is also an incentive to increase T*. When sand decays slowly, marginal cost of a rotation is low
and declining more slowly in T. This dampens the incentive to increase T*, while a high discount rate provides an incentive
to decrease T*.
In the Faustmann model, Eq. (12) shows that optimal rotation balances future rents against marginal net benets of a
single rotation. The beach problem has a similar feature. But in Faustmann, costs are only xed and do not affect the
marginal net benet of the single rotation. As such, a cost increase reduces rents, so marginal net benets must also go
down to compensate. That can only be accomplished by extending the rotation. In our case, a variable cost increase affects
both sides of the rst-order condition (11) such that both terms shift down. When background erosion is high, the shift
down in marginal net benets overshoots the reduction that is required to balance marginal net benets and future rents.
So, the rotation must shrink to increase marginal net benets and balance the rst-order condition.
To clarify Proposition 2 further, consider Fig. 4. The four panels represent two values of baseline erosion (high and low)
and two values of variable cost (high and low). In the rst two panels, background erosion is low (g 0.5). The increase in
variable cost (from f 1 to 2) shifts both the total net benets of future rotations (NB rent) and the marginal net benet
(MNB) down. The shifts are such that the optimal rotation length increases with the increased variable cost (as intuition
suggests). When variable cost increases, it decreases the rent on all future rotations. The rotation increases to reduce
marginal net benets to compensate. This is the same process that governs the Faustmann rotation when xed costs
increase. In the second two panels, background erosion is high (g 2.0). The increase in variable cost still shifts both
curves, but the shifts have different consequences. Here, the shift in marginal net benets with the higher variable cost
overcompensates for the lost rent from the cost increase. Thus, the rotation length must shrink to increase marginal net
benets and compensate for the change in rent. The input demand for sand thus slopes upward in this case. We will see
below that this case is a consequence of xing the extensive margin (the initial beach width x
0
). An analogous
counterintuitive result comes out of the basic Faustmann model. When the extensive margin (i.e. the amount of land in
forestry) is xed in Faustmann, a timber price increase shortens the rotation length; in doing so, the average supply of
timber decreases and output supply slopes downward.
The remaining comparative static results are problematic because they involve integrals that do not have closed-form
expressions. To see this, substitute Eqs. (5), (6), (13), and (14) into Eq. (10):
@v
@T
e
dT
dG1 mx
0
me
yT
x
0
gT fg fmyx
0
e
yT
=1 e
dT

de
dT
Z
T
o
e
dt
dG1 mx
0
me
yt
x
0
gtdt c fmx
0
1 e
yT
gT

1 e
dT

set
0. (15)
Note that the parameters m, x
0
, y, g, and d all enter in both the rst line and the second line, the latter being subtracted
from the former. Moreover, these parameters are all in the integrand of the integral that is being subtracted. By Liebnizs
rule, the integral will appear in all comparative statics for these parameters. We thus resort to numerical simulations to
explore the model further. To this end, we introduce a two-parameter functional form for benets:
BT
Z
T
o
e
dt
daxt
b
dt. (16)
Eq. (16) conforms to our assumptions about the benets function above (strictly positive rst derivative and negative
second derivative). The parameter a can be interpreted as a base value of the beachfront property, while the parameter b
controls the hedonic price of beach width, conditional on having a beachfront property. As before, d converts the capital
value into a ow. A particular advantage of this functional form is that static hedonic pricing models actually estimate the
beach width parameter. For example, Pompe and Rinehart [28] nd b 0.2632 in a constant-elasticity hedonic model.
In the following simulations, we use the same numerical values for parameters that generated Fig. 2. For each set of
simulations, we x all but two parameters and vary the others together. The base parameters are: b 0.25, a 200,
d 0.06, c 10, and f 1. These parameters are illustrative and used to generate testable hypotheses; they do not
represent any particular community. For each combination of parameters, we numerically solve for T* in two different
ways. First, we minimize (v) in Eq. (9) using Matlabs constrained minimization procedure (FMINCON) to obtain
^
T. We use
FMINCON to constrain
^
T to fall between 0 and T
max
, where T
max
is dened implicitly:
xT
max
1 mx
0
me
yT
max
x
0
gT
max
0. (17)
We nest a numerical quadrature (Matlabs QUAD function) within the objective function to evaluate the integral in
Eq. (16). We then evaluate Eq. (9) at
^
T and compare with the value of abandoning the property, which is zero by
construction:
T
n

^
T if v
^
T40;
1 else:
(
(18)
Second, we solve the rst-order condition in Eq. (10) directly to obtain
^
^
T using Matlabs FSOLVE procedure. Again we use
numerical quadrature to evaluate Eq. (16). We compute T* as in Eq. (16), substituting
^
^
T for
^
T and check that the two
numerical solutions for T* are close, i.e. verify that the absolute value of the difference is less than 0.001.
ARTICLE IN PRESS
M.D. Smith et al. / Journal of Environmental Economics and Management 58 (2009) 5871 66
Fig. 5 numerically illustrates Proposition 2 by showing that increasing the variable cost will increase T* when baseline
erosion is low but will decrease T* when baseline erosion is high. Contour labels represent different values of T*. Following
along the horizontal axis, T* increases for low values of g but increases for high values of g. This raises the possibility that
beach nourishment as a coastal management strategy will accelerate in response to climate change-driven sea-level rise
even if the scarcity of nourishment sand drives up the variable cost.
We next present six numerical results that are shown to hold with accompanying contour plots for the optimal
nourishment rotation length. There are four accompanying gures in an online appendix (Figs. A1A4), and Fig. 6 depicts
the less intuitive Numerical Result 6.
ARTICLE IN PRESS
6
7
7
7
8
8
8
8
9
9
9
9
1
0
10
10
11
11
1
1
12
12
1
2
13
13
1
3
14
14
15
15 16
17
Cost of Nourishment Sand
B
a
s
e
l
i
n
e

E
r
o
s
i
o
n

R
a
t
e
Countour Plot of Optimal Rotation Length
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
0.5
1
1.5
2
Fig. 5. Increasing variable cost of sand can accelerate or decelerate nourishment.
4
4
5
5
5
5
6
6
6
6
7
7
7
7
7
8
8
8
8
9
9
9
1
0
1
0
1
0
11
1
1 12
1
2
13
1
3
14
Exponential Decay Rate of Nourishment Sand
D
i
s
c
o
u
n
t

R
a
t
e
Countour Plot of Optimal Rotation Length
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
0.05
0.06
0.07
0.08
0.09
0.1
0.11
0.12
0.13
0.14
Fig. 6. A higher exponential decay of nourishment sand increases (decreases) frequency of nourishment if the decay rate is lower (higher) than the
discount rate.
M.D. Smith et al. / Journal of Environmental Economics and Management 58 (2009) 5871 67
Numerical Result 1. The optimal rotation length decreases as the value of beach width increases.
Nourishment occurs more frequently when the hedonic price of beach width (b) is higher. This result is intuitive because,
ceteris paribus, a high-b community would prefer a wide beach. A higher b increases both the total and marginal benets
of beach width, and more frequent nourishment increases the average beach width.
Numerical Result 2. The optimal rotation length decreases as the base property value increases.
Communities will nourish more frequently when the base property value (a) is higher. As population pressure in the
coastal zone drives up coastal property values, holding other features of the problem constant, our model predicts that
more beach nourishment will occur.
Numerical Result 3. The optimal rotation length decreases as the baseline erosion rate increases.
When the beach erodes faster, communities must nourish more often to keep up.
Numerical Result 4. The optimal rotation length decreases as the share of beach width subject to exponential decay increases.
When more of the beach is subject to exponential decay (m is big), a larger fraction of beach width addition from
nourishment is lost quickly, and communities must nourish more often to keep up. Keeping up with erosion sounds
intuitive, but high-m communities are getting less bang for their buck, and this suggests that Numerical Result 4 is
somewhat counterintuitive. The former effect dominates because all communities will keep up with erosion as long as the
present value of beach nourishment for some T is positive. Communities that lose a lot of nourishment sand in the
adjustment to equilibrium prole essentially become addicted to nourishment as a beach stabilization strategy.
The present value of nourishment exceeds the present value of abandoning the beach for all parameter value combinations,
even though the total present value of the optimal programthe solution to Eq. (9), v (T*)is lower for high-m
communities than for low-m communities. This is no different than suggesting that two otherwise identical properties with
different depreciation rates will be priced differently in the market.
Numerical Result 5. The optimal rotation length decreases as the discount rate increases.
The intuition ows directly from that standard thinking in resource economics; a higher discount rate places more
weight on the near term. In rotation problems, this implies that the marginal net benets of a single rotation receive more
weight than the present value of future rotations. The manager shrinks the rotation length to trade more net benets in the
short run for less net benets in the long run. Note that Figs. A1A4 all illustrate Numerical Result 5.
Numerical Result 6. The optimal rotation length increases (decreases) as the exponential decay rate of nourishment sand
increases if the decay rate is higher (lower) than the discount rate.
Fig. 6 illustrates this result. The thick solid line is the 451 line along which y d. Recall that in the traditional Faustmann
problemand the interpretation due to Samuelson [29]the second term in the rst-order condition is the foregone
interest payment by delaying land rent, where land rent is the discounted present value of all future rotations. A higher
discount rate means that the difference between marginal costs and marginal benets of a single rotation must be greater
to compensate. To the left of the 451 line in Fig. 6, the discount rate is higher than the decay rate of nourishment sand. In
other words, the depreciation of nancial capital (foregone interest) dominates the depreciation of sand capital. To balance
interest lost on land rent with marginal net benets, the rotation length must decrease as the exponential decay rate of
sand increases. The opposite is true to the right of the 451 line where depreciation of sand capital dominates. In essence, if
the discount rate exceeds the decay rate, an increase in decay rate is relatively more costly to the short run because one
places more weight on the short run with a high discount rate, and rotation length is shortened to adjust.
5. Endogenous initial widthallowing the extensive margin to change
We now consider whether our most surprising resultthat input demand for sand can slope upwards or
downwardscontinues to hold if initial beach width (x
0
) is a choice variable. The analytical approach to this extension
is straightforward; we now have an additional choice variable in (9) and a second rst-order condition. However, x
0
enters
both the benet and cost functions, making analytical propositions problematic and requiring numerical analysis.
Numerical Result 7. With endogenous initial beach width, input demand for sand can slope upwards for low variable costs but
turns downwards for high variable costs.
For the same parameter values used above, we nd that with endogenous x
0
an increase in variable cost still decreases
the nourishment interval. However, it also decreases x
0
. Communities are nourishing more often but building the beach out
less when costs increase. When initial beach width is unconstrained, the net result is that sand use decreases when variable
cost of sand increases. That is, input demand slopes downwards. Fig. A5 (available in the online appendix) illustrates this
result for a range of values of baseline erosion. Conditional on baseline erosion (xing a particular level on the y-axis), a
variable cost increase (moving along the x-axis) leads to less sand use (a lower sand use contour).
ARTICLE IN PRESS
M.D. Smith et al. / Journal of Environmental Economics and Management 58 (2009) 5871 68
The generality of downward-sloping sand demand is far from clear because there are physical limits on x
0
, i.e. feasibility
constraints on how far out a beach can be built. When these physical constraints are put in the problem, there is a range over
which managers always choose the maximum x
0
. The desired (unconstrained) x
0
is essentially greater than the physical
maximum. As a consequence, a variable cost increase shrinks the desired x
0
but not necessarily enough to be below the
maximum feasible x
0
. Because rotation length also shrinks with the increased variable cost, the amount of sand used increases.
That is, input demand can slope upwards. Fig. 7(a) illustrates this result for the case of high erosion and lowcost of nourishment
sand. We nd that upward-sloping input demand cannot be sustained at high costs of nourishment sand; eventually sand
demand turns downwards. Fig. 7(b) illustrates this result for the case of high erosion and high cost of nourishment sand.
ARTICLE IN PRESS
4.8494 4.8494
4.8494
4.9907 4.9907 4.9907
5.132 5.132 5.132
5.2733 5.2733
5.2733
5.4146
5.4146
5.4146
5.556
5.556
5.556
5.6973
5.6973
5.6973
5.8386
5.8386
5.8386
5.9799
5.9799
5.9799
6.1212
6.1212
6.1212
Cost of Nourishment Sand
B
a
s
e
l
i
n
e

E
r
o
s
i
o
n

R
a
t
e
Sand Use
0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
1.5
1.6
1.7
1.8
1.9
2
2.1
2.2
2.3
2.4
2.5
3
.
1
5
6
6
3
.
3
7
1
1
3
.
3
7
1
1
3
.
5
8
5
7
3
.
5
8
5
7
3
.
5
8
5
7
3
.
8
0
0
2
3
.
8
0
0
2
3
.
8
0
0
2
4
.
0
1
4
7
4
.
0
1
4
7
4
.
0
1
4
7
4
.
0
1
4
7
4
.
2
2
9
3
4
.
2
2
9
3
4
.
2
2
9
3
4
.
4
4
3
8
4
.
4
4
3
8
4
.
4
4
3
8
4
.
6
5
8
3
4
.
6
5
8
3
4
.
8
7
2
9
5
.
0
8
7
4
Cost of Nourishment Sand
B
a
s
e
l
i
n
e

E
r
o
s
i
o
n

R
a
t
e
Sand Use
1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5
1.5
1.6
1.7
1.8
1.9
2
2.1
2.2
2.3
2.4
2.5
Fig. 7. (a) Sand use with endogenous but physically constrained x
0
high erosion, lowsand cost. (b) Sand use with endogenous but physically constrained
x
0
high erosion, high sand cost.
M.D. Smith et al. / Journal of Environmental Economics and Management 58 (2009) 5871 69
To summarize, when the background erosion rate is high, greater nourishment activity will increase the scarcity of
nourishment-quality sand and likely raise its price. For communities that are already building out their beaches to the
maximum feasible width, this will lead to an acceleration of nourishment (increased sand use) in the short run. If the price
of sand continues to rise, eventually communities will nd it in their interest to reduce the initial beach width as they
reduce the nourishment interval. As in the case with exogenous x
0
, eventually sand becomes so expensive that abandoning
the shoreline defense is optimal.
6. Discussion
Resource economics fundamentally deals with connections between natural capital and nancial capital. Optimal
mineral extraction hinges on whether capital in the bank grows faster than the value of capital in the ground. In sheries
and forestry, optimal management entails balancing growth of the biological resource with growth of the nancial
resource. Viewing a managed beach as a renewable natural resource, one sees again this deep capital-theoretic connection.
Optimal management from a beach communitys perspective depends on the rates of decay of sand and nancial capital.
When nourishment sand decays slower than the rate of foregone interest, an increase in variable cost or increased
nourishment sand decay could accelerate beach nourishment. This result is more nuanced when initial beach width is
made endogenous, but it does not disappear entirely.
By focusing on a representative community, our model produces testable hypotheses about real coastal communities in
general. For empirical work, one would need cross-community variation in erosion rates, base property values and
values of beach width (from hedonic studies), and records of nourishment activities. Such empirical work is especially
important for future research because actions of individual communities could have consequences at the spatial scale of an
entire coastline. That is, the effects of small bumps in an otherwise uniform coastline can propagate over large spatial
scales in economically meaningful time scales [1]. If our positive model describes real behavior, spatially heterogeneous
property values and erosion rates will lead to spatially heterogeneous nourishment interventions. Whether this
process would then lead to a more or less spatially uniform coastline than one that has not been altered by humans is an
open question.
A complementary approach to analyzing beach nourishment would be to use numerical dynamic programming [19].
This approach allows for a more general description of the dynamics such that the state variable evolves as a function of
absolute time, rather than time since the last nourishment event. It also permits the choice of initial beach width for each
rotation. The drawbacks include increased computational burden and the difculty of generalizing comparative static or
dynamic results. In our model, we are able to obtain some analytical results before relying on numerical methods.
Nevertheless, with no previous literature on beach nourishment as a dynamic capital accumulation problem and with
increasing development pressure in the coastal zone, we submit that our Faustmann-style rotation and numerical dynamic
programming are both worthwhile to pursue.
Although most of our model results conform to basic intuition, the possibility that nourishment could increase in
frequency as the variable costs increase is unsettling. As sea level rises or storm patterns change in response to climate
change, an increase in the baseline erosion rate is not unlikely. At the same time, more people are living in the coastal zone,
and property values continue to rise. Nourishment sand that can be recovered through dredging may eventually become
scarce [10,15], and one would expect variable costs to increase. If communities respond by nourishing more often, this
effect could feed back on itself and further accelerate nourishment. Beaches then become increasingly articial, and the
ecological costs of nourishment would grow substantially. In the current policy regime, environmental impacts of
nourishment are cataloged but not counted as costs to weigh against nourishment benets. In a future of accelerated
nourishment activity, policy-makers could begin to require these non-market costs to be counted.
Our model does not address the funding mechanism for beach nourishment. We implicitly assume that a dynamically
optimal nourishment projectfrom the communitys perspectivecan be funded in some manner. Real projects receive
funding from federal, state, and local governments as well as private sources. Both the William J. Clinton and George W.
Bush administrations have supported reducing the federal share of nourishment, though Congress passed the 2002 Water
Appropriations Bill with $47.1 million more for beach nourishment than the $87.6 million that President Bush requested
[23]. An important question for future research is how a change in the availability of federal funds for nourishment will
affect the prevalence and frequency of nourishment. Conventional wisdom suggests that there would simply be less
nourishment, but our capital-theoretic model suggests that outcomes could be more complicated. Whether a reduction in
the federal share will increase or decrease nourishment activity could depend on the allocation of federal funds across xed
and variable costs.
Acknowledgments
This research was funded by the NSF Biocomplexity Program (Grant #DEB0507987). We thank Sathya Gopalakrishnan
for valuable research assistance and Tom Crowley, Ling Huang, Mike Orbach, Joe Ramus, Jim Wilen, Junjie Zhang, and two
anonymous reviewers for helpful comments.
ARTICLE IN PRESS
M.D. Smith et al. / Journal of Environmental Economics and Management 58 (2009) 5871 70
Appendix A. Supplementary data
Supplementary data associated with this article can be found in the online version at doi:10.1016/j.jeem.2008.07.011.
References
[1] A. Ashton, A.B. Murray, O. Arnault, Formation of coastline features by large-scale instabilities induced by high-angle waves, Nature 414 (2001)
296300.
[2] F.W. Bell, V.R. Leeworthy, Recreational demand by tourists for saltwater beach days, Journal of Environmental Economics and Management 18 (1990)
189205.
[3] O. Bin, C.E. Landry, C. Ellis, H. Vogelsong, Some consumer surplus estimates for North Carolina beaches, Marine Resource Economics 20 (2005)
145161.
[4] O. Bin, T. Crawford, J.B. Kruse, C.E. Landry, Viewscapes and ood hazard: coastal housing market response to amenities and risk, Land Economics 84
(2008) 434448.
[5] P. Bruun, Coastal erosion and the development of beach proles, Journal of Waterways Harbors Division American Society Civil Engineers 88 (1962)
117130.
[6] R.G. Dean, Beach Nourishment: Theory and Practice, World Scientic, Hackensack, NJ, 2002.
[7] J.H. Dorfman, A.G. Keeler, W. Kriesel, Valuing risk-reducing interventions with hedonic models: the case of erosion protection, Journal of Agricultural
and Resource Economics 21 (1996) 109119.
[8] S.F. Edwards, F.J. Gable, Estimating the value of beach recreation from property values: an exploration with comparisons to nourishment costs, Ocean
and Shoreline Management 15 (1991) 3755.
[9] M. Faustmann, Berechnung des Wertes welchen Waldboden sowie noch nich haubare Holzbesta nde fu r die Waldwirtschaft besitzen, Allgemeine
Forst- und Jagd-Zeitung 15 (translated into English by W. Linnard and reprinted in Journal of Forest Economics) 1 (1995) 744, (1849).
[10] C.W. Finkl, L. Benedet, J.L. Andrews, B. Suthard, S.D. Locker, Sediment ridges on the west Florida inner continental shelf: sand resources for beach
nourishment, Journal of Coastal Research 23 (2007) 143159.
[11] D.A. Grain, A.B. Bolton, K.A. Bjorndal, Effects of beach renourishment on sea turtles: review and research initiatives, Restoration Ecology 3 (1995)
95104.
[12] D.G. Hallstrom, V.K. Smith, Market responses to hurricanes, Journal of Environmental Economics and Management 50 (2005) 541561.
[13] R. Hartman, The harvesting decision when the standing forest has value, Economic Inquiry 14 (1976) 5258.
[14] H.J. Heinz, Evaluation of Erosion Hazards, The H. John Heinz III Center for Science, Economics and the Environment, Washington, DC, 2000.
[15] C.W. Hoffman, Preliminary assessment of potential sand resource areas offshore of nags head, Kitty Hawk, and Kill Devil Hills, North Carolina, US
Minerals Management Service, State of North Carolina, Sand Resources Task Force, 1998.
[16] IPCC, Climate Change 2001: Impact, Adaptation, and VulnerabilityContribution of Working Group II to the Third Assessment Report of the
Intergovernmental Panel on Climate Change, Cambridge University Press, New York, 2001.
[17] P.D. Komar, Beach Processes and Sedimentation, Simon & Schuster, Upper Saddle River, NJ, 1998.
[18] W. Kriesel, A. Randall, F. Lichtkoppler, Estimating the benets of shore erosion protection in Ohios Lake Erie housing market, Water Resources
Research 29 (1993) 795801.
[19] C.E. Landry, Optimal management of coastal erosion on developed barrier Island Beaches, Working Paper, Department of Economics, East Carolina
University, 2006.
[20] C.E. Landry, A.G. Keeler, W. Kriesel, An economic evaluation of beach erosion management alternatives, Marine Resource Economics 18 (2003)
105127.
[21] W.J. Milon, J. Gressel, D. Mulkey, Hedonic amenity valuation and functional form specication, Land Economics 60 (1984) 378387.
[22] NAST, Climate change impacts on the United States: the potential consequences of climate variability and change, A Report of the National
Assessment Synthesis Team, US Global Change Research Program, 2000.
[23] NOAA, Beach Nourishment: A Guide for Local Government Ofcials, NOAA Coastal Services Center, 2006.
[24] NRC, Beach Nourishment and Protection, National Research Council, National Academy Press, Washington, DC, 1995.
[25] G.R. Parsons, M. Powell, Measuring the cost of beach retreat, Coastal Management 29 (2001) 91103.
[26] G.R. Parsons, D.M. Massey, T. Tomasi, Familiar and favorite sites in a random utility model of beach recreation, Marine Resource Economics 14 (2000)
299315.
[27] O.H. Pilkey, T.D. Clayton, Beach Replenishment: The National Solution? Coastal Zone 87, American Society of Civil Engineers, Seattle, WA, 1987.
[28] J.J. Pompe, J.R. Rinehart, The value of beach nourishment to property owners: storm damage reduction benets, Review of Regional Studies 25 (1995)
271285.
[29] P.A. Samuelson, Economics of forestry in an evolving society, Economic Inquiry 14 (1976) 466492.
[30] J.M. Slott, A.B. Murray, A.D. Ashton, T.J. Crowley, Coastline responses to changing storm patterns, Geophysical Research Letters 33 (2006) 6.
[31] STICS, Spatial Trends in Coastal Socioeconomics: Coastal Zone County Facts, 2006.
[32] J.G. Titus, R.A. Park, S.P. Leatherman, J.R. Weggel, M.S. Greene, P.W. Mausel, S. Brown, C. Gaunt, M. Trehan, G. Yohe, Greenhouse-effect and sea-level
risethe cost of holding back the sea, Coastal Management 19 (1991) 171204.
[33] A.C. Trembanis, O.H. Pilkey, Summary of beach nourishment along the US Gulf of Mexico shoreline, Journal of Coastal Research 14 (1998) 407417.
[34] H.R. Valverde, A.C. Trembanis, O.H. Pilkey, Summary of beach nourishment episodes on the US East Coast barrier islands, Journal of Coastal Research
15 (1999) 11001118.
[35] E. Woglom, The Dynamics of Shoreline Management: An Approach using Renewable Resource Economics, Nicholas School of the Environment and
Earth Sciences, Duke University, 2003.
[36] G. Yohe, J. Neumann, H. Ameden, Assessing the economic costs of greenhouse-induced sea level rise: methods and application in support of a
national survey, Journal of Environmental Economics and Management 29 (1995) S78S97.
ARTICLE IN PRESS
M.D. Smith et al. / Journal of Environmental Economics and Management 58 (2009) 5871 71