Scholarly article on topic 'Impact Fees Coupled With Conservation Payments to Sustain Ecosystem Structure: A Conceptual and Numerical Application at the Urban-Rural Fringe'

Impact Fees Coupled With Conservation Payments to Sustain Ecosystem Structure: A Conceptual and Numerical Application at the Urban-Rural Fringe Academic research paper on "Earth and related environmental sciences"

CC BY
0
0
Share paper
Academic journal
Ecological Economics
Keywords
{"Land use regulation" / "Land preservation" / "Impact fees" / Ecosystems / "Public finance" / "Urban sprawl" / Wetlands / "Development rights" / Spatial / Metapopulation}

Abstract of research paper on Earth and related environmental sciences, author of scientific article — Yong Jiang, Stephen K. Swallow

Abstract Communities in exurban areas increasingly rely on land preservation as a strategy to balance sprawling land development with maintaining environmental amenities. Based on a review of existing approaches for preserving land, we consider a conceptual model of environmental impact fees (EIFs) coupled with conservation payments for managing private land of ecosystem value. In this framework, conservation payments are intended to cost-effectively target fair market value compensation for heterogeneous land for preservation that sustains ecosystem health. EIFs serve as a financial instrument to augment conservation payments and to allow flexibility for landowners with private information to pursue development opportunities while accounting for environmental impacts. Using a bioeconomic model of nature-reserve design, we develop an empirical illustration of how to estimate the EIF of development damage to critical habitat in southern Rhode Island in an effort to preserve land as an environmental infrastructure that maintains ecosystem health.

Academic research paper on topic "Impact Fees Coupled With Conservation Payments to Sustain Ecosystem Structure: A Conceptual and Numerical Application at the Urban-Rural Fringe"

Contents lists available at ScienceDirect

Ecological Economics

journal homepage: www.elsevier.com/locate/ecolecon

Impact Fees Coupled With Conservation Payments to Sustain Ecosystem Structure: A Conceptual and Numerical Application at the Urban-Rural Fringe

Yong Jiang a,b,*f Stephen K. Swallowc

a UNESCO-IHE Institute for Water Education, Westvest 7,2611AX Delft, The Netherlands

b Department of Public Management, Faculty of Humanities and Social Sciences, Dalian University ofTechnology, Dalian 116024, China c Department of Agricultural and Resource Economics and Center for Environmental Sciences and Engineering, University of Connecticut, Storrs, CT 06269, USA

ARTICLE INFO

ABSTRACT

Article history:

Received 1 June 2016

Received in revised form 4 February 2017

Accepted 6 February 2017

Available online xxxx

Keywords: Land use regulation Land preservation Impact fees Ecosystems Public finance Urban sprawl Wetlands

Development rights Spatial

Metapopulation

Communities in exurban areas increasingly rely on land preservation as a strategy to balance sprawling land development with maintaining environmental amenities. Based on a review of existing approaches for preserving land, we consider a conceptual model of environmental impact fees (EIFs) coupled with conservation payments for managing private land of ecosystem value. In this framework, conservation payments are intended to cost-effectively target fair market value compensation for heterogeneous land for preservation that sustains ecosystem health. EIFs serve as a financial instrument to augment conservation payments and to allow flexibility for landowners with private information to pursue development opportunities while accounting for environmental impacts. Using a bioeconomic model of nature-reserve design, we develop an empirical illustration of how to estimate the EIF of development damage to critical habitat in southern Rhode Island in an effort to preserve land as an environmental infrastructure that maintains ecosystem health.

© 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://

creativecommons.org/licenses/by/4.0/).

1. Introduction

Maintaining the quality of public services, particularly environmental amenities, is of policy interest to communities at the urban-rural fringe. Over recent decades, amenity-driven migration has strongly influenced the evolution of American life, characterized by pursuit of low density residential development in exurban and rural areas (see Marcouiller et al., 2002; Irwin et al., 2009). Sprawling land development, while imposing pressure on provision of public services, threatens local ecosystems and environmental amenities, as well as agricultural and rural landscapes (Johnson, 2001; Daniels and Daniels, 2003; Odell etal., 2003; Mcdonald et al., 2009). Concern over environmental degradation motivates local policy initiatives to regulate growth and to protect the environment (Myers and Puentes, 2001; Marcouiller et al., 2002; Bengston et al., 2004).

Recognizing the potential impact of urban sprawl, communities increasingly rely on land preservation as a strategy to balance residential development and maintain environmental amenities (Daniels and Lapping, 2005; Jiang and Swallow, 2015). The rationale is that

* Corresponding author. E-mail addresses: y.jiang@unesco-ihe.org (Y.Jiang), stephen.swallow@uconn.edu (S.IK Swallow).

environmental amenities can be sustained by protecting from development certain private undeveloped land that is environmentally valuable. The efficiency, effectiveness and political feasibility of land preservation in pursuit of intended goals, however, depends upon the extent to which land acquisition takes into account the economic cost of land (Ando et al., 1998; Polasky et al., 2001), the development rights of owners (Innes, 1995,1997), incentives for different land uses including conservation (Innes and Frisvold, 2009), as well as the role of land and its use in the remaining ecosystem (Swallow, 1996a, 1996b).

In this study, we propose an impact fee framework coupled with conservation payments to manage private land of environmental value in an incentive-based system. This approach is motivated by a review of existing land acquisition approaches such as planning and incentive-based programs, which are found insufficient to achieve land development while sustaining valued environmental resources at the urban-rural fringe, particularly in a heterogeneous landscape with ecologically interdependent land parcels, constitutional protection of property rights, and local conservation financing challenges (see section 2 for further detail). Our intent is to provide an intuitive framework to motivate and implement impact fees as a tool to mitigate environmental impacts of land development while being ecologically effective and financially self-sufficient. This intent follows from the recommendations of Portney (2004)

http://dx.doi.org/10.1016/j.ecolecon.2017.02.007

0921-8009/© 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

who calls economists to offer environmental policy alternatives of potential practical value for improving social welfare, even when complete benefit-cost analysis may be impossible.

We consider a novel policy, environment impact fees (EIFs), which are inspired by the practice of development impact fees in the public sector. Development impact fees are used to finance local infrastructure, such as schools and sewage systems, and to control overdevelopment and urban sprawl (Brueckner, 1997; Burge et al., 2007). Similarly, EIFs can serve as a tool to finance conservation, to create "green infrastructure" providing ecosystem services such as pollution purification, flood mitigation, and a green space or wildlife habitat network. For this paper, such a network is analogous to the infrastructure supporting conventional community services such as schools or public safety services, and the EIF approach strives to reduce the cost for a community to achieve a conservation network designed to sustain a targeted level of ecological health. Simultaneously, EIFs can serve as a Pigovian instrument to internalize any negative impact of land development on the local environment (Clinch and O'Neill, 2010a, 2010b). Indeed, previous studies have suggested using impact fees for environmental purposes such as protecting open space (Nicholas and Juergensmeyer, 2003) and encouraging "green" buildings (Kingsley, 2008).

In our conceptual model, the EIFs are intended to augment conservation payments or other incentive-based conservation programs ofa community, and to address the potential conflict between landowner discretion and public interest in environmental amenities. In a heterogeneous landscape, conservation programs cost-effectively target private land of critical environmental value by offering payments for enrollment in conservation. Such payments comprise just compensation for taking private development rights for the public purpose of establishing a conservation reserve network (or green infrastructure). EIFs explicitly account for the possible external impact of an individual development decision involving private land that has been identified and targeted for a conservation network. That is, EIFs directly link the assessment of development damage imposed on a planned conservation network to the additional financial expenditures a community would face to compensate for such damage and allow the community to achieve level of environmental amenities expected from the original plan.

To empirically demonstrate the EIFs, we apply a spatially explicit bioeconomic model ofa nature reserve design to guide the conservation program in cost-effectively targeting land for preservation, thereby establishing green infrastructure - e.g., a wildlife habitat network - that sustains a healthy ecosystem. We estimate EIFs for each undeveloped land parcel that has been incorporated within a community's plan to create, cost effectively, a conservation network that will sustain the local ecosystem. The EIFs are derived from minimization of the cost of conservation payments needed to acquire land that is essential to sustaining a socially (municipally) chosen ecosystem health or environmental quality index. Our case study illustrates empirical estimation of spatially-sensitive EIFs for different levels of development damage to the community's plan to establish a conservation reserve network.

This study is relevant to land use policy and public decision-making. There is an increasing literature that reveals the linkage between landscape elements and ecosystem structure and process (e.g., Araujo et al., 2002; Bauer et al., 2010; Bennett et al., 2006; Ernoult et al., 2006; Pearson and Dawson, 2005; Polasky et al., 2001; Polasky et al., 2005; Swallow et al., 1997). The present study attempts to develop and illustrate a conceptual, integrated framework that links the conservation literature to community land use management in a socio-bioeconomic framework. It addresses an important land use issue, characteristic of communities at the urban-rural fringe. Our case study demonstrates the potential and importance of research integration for a more comprehensive analysis of land use management aimed to improve economic efficiency, environmental quality, and political acceptance (Plantinga, 2015).

In the next section, we review existing approaches. Section 3 describes the conceptual model of EIFs coupled with conservation

payments designed to protect the local landscape's capacity to sustain a minimal level of ecosystem health while accommodating land development. Section 4 presents an example, estimating EIFs for different levels of development damage to a network of land that, if preserved at an optimal level, could cost-effectively sustain a target level of ecosystem health. Section 5 draws conclusions.

2. Literature Review

A traditional approach to protecting land relies on zoning and land use planning, such as low density zoning (Fischel, 2000) and conservation subdivision (Arendt, 1999). The planning-based approaches can increase the amount of undeveloped land by regulating development density or intensity, but may not effectively address the conservation needs of local ecosystems (Kretser et al., 2008; Carter, 2009). Moreover, the welfare consequences in terms of efficiency and equity of those planning approaches are often of concern as those instruments may be rigid and may create land rent and windfall gains differentially affecting landowners while producing significant transaction costs reducing the efficiency of land use management (Thorson, 1996, Heikkila, 2000, cf. Lewis et al., 2009).

The pitfalls of the planning-based approach may partially explain the increasing popularity of market-based instruments to promote desirable land use, such as fee-simple purchases of land, conservation easements, or transferable development rights (e.g., Rushman, 2000; Bengston et al., 2004; Watzold and Drechsler, 2005; McConnell et al., 2006; Carter, 2009). Market-based instruments typically acknowledge landowners' development rights and create a market setting enabling retirement of landowners' development rights and, thus, land preservation through the more flexible market mechanism. Their major advantage over the traditional planning-based approaches lies in the potential to improve equity and efficiency moderated through the market mechanism.

To improve land use patterns, particularly in the context of protecting biodiversity and ecosystem function, economists have also examined incentive-based mechanisms in designing payments or programs to acquire land (Parkhurst and Shogren, 2003; Lewis et al., 2011). Examples include contracts with landowners who protect endangered species (Smith and Shogren, 2002), "agglomeration bonus" to encourage preservation of large tracts of land (Parkhurst et al., 2002; Drechsler et al., 2010), spatially uniform versus heterogeneous compensation payments (Watzold and Drechsler, 2005; Lewis and Plantinga, 2007), direct land acquisition for preservation versus indirect approaches to affect relative returns to various land uses (Langpap and Wu, 2008), and incentives for reducing habitat fragmentation (Lewis et al., 2009). Exhibiting varying advantages with desirable welfare implications, these programs all can reduce the negative impact of land use on local ecosystems.

In a heterogeneous landscape with ecologically interdependent land parcels, pure incentive-based programs or market-based instruments alone may still be insufficient to address local ecosystem needs (e.g., Anderson and King, 2004). Ecosystem health relies on spatially heterogeneous land uses and attributes which establish a structure through which parcels may contribute an unequal share to ecosystem process and function (Swallow, 1996b; Swallow et al., 1997; Wiens et al., 2006). This reality implies that the spatial configuration of preserved land is an important element in addition to the total acreage that may be preserved. Yet, incentive or market-based conservation programs often do not explicitly link land preservation to ecosystem process and function (Jiang etal.,2007; Lewis et al., 2011), and by relying on voluntary participation and landowners' discretion, decentralized land use decisions can leave uncertain the resulting pattern of land use and outcomes for ecosystem function (Lewis et al., 2011). From the ecological and political perspective, a more effective mechanism is needed to protect landscape elements and the structure consistent with ecosystem

health while retaining the desirable voluntary nature of incentive-based programs.

In addition, pure incentive-based conservation programs can raise a fiscal issue for a local government budget. Proactive land use planning incorporates land preservation as an integral part (Beatley, 2000; Daniels and Lapping, 2005). To finance land preservation, local governments often issue bonds or dedicate portions of their annual budgets to conservation funds for land acquisition. Yet, a land parcel of critical ecosystem value may not be fully available due to some level of development taken by the landowner between the planning and implementation stages. In an extreme case, a landowner's development decision may even preempt the parcel from preservation. In such cases, the original conservation budget may be no longer sufficient to achieve the targeted level of ecosystem protection due to differences in costs of land parcels that could be chosen to fill the functional role of a parcel lost to development. To achieve conservation goals, local communities thus need a reliable funding source to take into account the uncertain outcome of voluntary conservation programs and to finance any possible increase in cost to achieve the targeted ecosystem outcome. The funding mechanism to cover public conservation cost requires serious consideration, particularly given that government authority to collect money by taxation has been limited even if it is for public interest (see for example Silicon Valley Taxpayers Association, Inc. v. Santa Clara County Open Space Authority (2008)).

3. A Conceptual Model

In this section, we introduce a conceptual framework to identify how EIFs may play a role in assisting a community to improve efficiency of land allocation. Our purpose is not to develop a definitive theoretical model, but rather to establish the intuitive foundations for how EIFs could serve in an efficient land allocation scheme. Here, our framework provides a sufficient, intuitive basis for understanding the broader theoretical challenge while, similar to the implementation of impact fees for conventional public services, our empirical approach will maintain a focus on the cost-effective support of the policy target considered here, that of sustaining a community-selected level of ecosystem health by establishing a conservation reserve network.

3.1. The Economic Foundation of Conservation intervention

Consider a community at the urban-rural fringe where rapid residential development threatens local ecosystem health by destroying and fragmenting wildlife habitat. Denote Q(Ap) as a unit-free measure of ecosystem health for the community, which depends on the total amount of land allocated to preservation Ap. Denote R(Q) as the Ricardian rent of community land parcels, which indicates the monetary value per unit of land and which is a function of ecosystem health Q(Ap). Assume land preservation can improve ecosystem health Q(Ap) that in turn may raise the Ricardian rent R(Q) of community land. Therefore, dQ/dAp > 0 and dR/dQ> 0, thereby defining dR/dAp = (dR/dQ) ■ (dQJ dAp) > 0.

Suppose the community has a total amount of land A. If the total value of land in the community is R(Q) ■ (A — Ap), then the socially optimal amount of land, Ap*, allocated to preservation in order to maximize the total value of community land must satisfy the following necessary condition:

Consider a landowner i's decision on allocating his own land between development and preservation to maximize his wealth. At a social optimum, the amount of undeveloped land of each landowner in the community will jointly provide the optimal amount of preservation Ap*. Denote A'd as the total amount of developed and developable land held by landowner i. If the landowner allocates a unit of his undeveloped land to preservation, he would enjoy a benefit of (9R/dAp)A'd while incurring an opportunity cost of the forgone rent R. Since the amount of developed and developable land of landowner i is only a portion of all such land in the community of many landowners, such that Add < A — A* then

In other words, the landowner suffers a loss in his wealth if, when the community is at the optimum, he continues to hold a marginal unit ofundeveloped land in preservation rather than in development. This incentive leads landowner i to develop more land than is socially optimal. These private land use decisions, in aggregate, eventually lead to a reduction in the amount of undeveloped land below the amount that would maximize the total value of community land. These considerations outline the intuitive economic argument for policy intervention in land use management to protect against environmental degradation, which motivates our conceptual model of EIFs.

As indicated by condition (2), policy interventions may be needed to create incentives to align the private cost with the private benefit for the socially optimal level of preservation. For the landowner i, those incentives may be: 1) collecting payments b from beneficiaries other than the landowner i to generate revenues which are used to reward private preservation such that (9R/dAp)Ad + b = R,2) imposing a charge t on private land development by landowner i to have (dR/dAp)A'd = R — t,or3) a combination of 1) and 2) to satisfy (9R/dAp)Ad + b = R — t depending on political preference for allocating the property right to environmental quality (Stavins, 2003). The above incentive framework for conservation management constitutes the intuitive foundation for our EIFs model.

3.2. Environmental Impact Fees Model

For our EIF model, we assume that a land manager (or a benevolent community planner) strives to protect both the public interest in the environment, including ecosystem structure critical to ecological functions, and the landowners' development rights.

3.2.1. Incentive-based Conservation Programs

The land manager establishes in his annual budget a conservation program that purchases undeveloped land in certain locations to be preserved as a green infrastructure, establishing a network supporting landscape ecosystem capacity. Economically, the conservation program attempts to implement the following incentive mechanism for each landowner, say, i:

w/d+m- r

dR(A—Ap) - R

where R(R) represents a payment rate per unit of land to the landowner who agrees to enroll a particular parcel in preservation, with the payment rate R(R) designed to balance the advantages of preservation1

That is, increasing land preservation at the margin raises the value of all developable (or developed) land parcels in the community, which must be balanced against the opportunity cost of the preserved marginal parcel, represented on the right-hand-side by the foregone value of development R.

1 The structure in Eq. (3) allows the policy to recognize that the landowner may benefit from marginal preservation that induces an increase in the value of parcels she retains for land not preserved. In practice there will be cases when this factor plays a small role and other cases where it is more significant at least for some larger landowners. In the former case, the incentive payment may equal or nearly equal the development value of the parcel preserved.

with the value R the landowner would gain on a per unit basis from leaving the land available for developed uses.

As community land is heterogeneous in terms of its ecological value and development opportunity cost (Swallow, 1996a; Ando et al., 1998), a practical policy goal for targeting land for preservation is to protect a targeted level of ecosystem health, Q0, at minimum cost of land withdrawn from availability for developed uses. Denote A'p as the amount of landowner i's land targeted for preservation and Ri (Ri) as the conservation payment rate per unit of enrolled land for landowner i who faces an opportunity cost R'. By indexing this notation to landowner i, we allow heterogeneity of parcels across landowners, so that our spatially explicit model below can accommodate rents and incentive payments at different rates for different landowners. The total cost of planned conservation payments across planned payouts (if accomplished instantaneously) to all landowners can be represented by E = ^ AlpR'(R') .2

With a conservation program cost-effectively targeting land for preservation, the total cost of conservation payment E represents the minimum cost for achieving the chosen level of environmental quality, Q0.

3.2.2. Mixed Mechanism of Conservation Payments and EIFs

In the incentive-based conservation programs, the land manager makes an offer at the rate R'(R') for purchasing a portion or whole parcel of undeveloped land privately owned by a landowner i for preservation. The payment rate R'(R' ) could be derived from a fair market value for the land, and may be based on a real estate appraisal of the development value R' of the land, such as the one employed for real estate taxation. In a practical land-use setting with private property rights, some landowners may make development decisions rather than accepting payment for preservation, particularly if the compensation payment rate

R'(R') is privately considered insufficient to cover the foregone benefit of land development. In such case, we need to distinguish between the landowner i's estimate of her opportunity cost for a parcel, which

we now denote R', and the planner's appraisal or estimate of that opportunity cost, which we denote with the symbol R'(R') as used earlier. Then, in this case, the landowner may be viewed as having an opportunity such that

Ad) + RYRM <F~

where R~i represents the landowner i's opportunity from taking a development option that is valued more highly than the manager's appraisal, R'r', of land rent for the parcel.

To achieve the conservation goal of Q0 at the planned cost E, an EIF may be introduced by the land manager to augment the incentive mechanism (4) for retiring land from development. That is, if the landowner i declines the offer at a rate R'(R') for preservation and pursues development, then he is subject to assessment of an EIF to cover the potential increment in the community's conservation cost triggered by developing his land; this increment arises from the developer's decision to remove or modify a parcel already identified as necessary to minimize the cost of Q0. If the landowner i's opportunity is sufficiently lucrative such that

OR-M) + R'iR1) <~' -EIF

he will likely choose to pay the EIF and continue with development;

otherwise he would accept the offer based on the rate of R'(R'). As our framework and empirical illustration show below, this EIF is calculated to capture the additional cost to the planner if the landowner chooses to develop rather than accept the conservation payment. This additional cost reflects the higher cost of land needed to maintain the targeted environmental quality Q0 if a parcel identified for the network is lost to development. In short, the EIF assessment internalizes the impact, if any, of land development on the conservation program's budget (cost). The revenue collected is dedicated to covering the costs of maintaining Q0 with an altered set of land to be preserved.

The EIF, combined with government offered conservation payments at rate R'^), is expected to induce cost-effective land preservation while allowing landowner discretion on land use within existing development rights. Without EIFs, only those land parcels for which the planner's appraisal of opportunity cost is less than the owner's appraisal, R' < R'(R'), would be enrolled in the conservation program in pursuit of environmental quality Q0 (see Lewis et al., 2011). With EIFs, additional land subject to r'(R')<R' <Ri(Ri) +EIF would be enrolled for preservation, and land with high development value R'>r'(R') + EIF would be allocated to development. The approach could move the community toward an optimal landscape supporting the ecosystem health Q0 while maximizing the value of community land and recognizing limits on a community's capacity to finance conservation.

3.3. Model Formulation for Estimating Environmental Impact Fees

We now formulate a spatially explicit model for empirically estimating EIFs in the context of establishing a community's green infrastructure, a conservation network. We begin by dividing the community landscape into standard grid cells. Denote the preservation status of each grid cell by y, which ranges from 0 to 1 indicating the percentage of preservation of the grid cell. The preservation status of the community landscape can then be described by a matrix of ys denoted as r, which determines the level of local ecosystem health Q(r) of the landscape (so Q(r) takes the place of Q(Ap) in the conceptual model above).

Denote C as the matrix of conservation payment rates R'(R'), which now are location-specific and represent the land cost per unit of grid cells in the landscape matrix r that comprise the conservation program. Denote Q0 as the socially desirable level of ecosystem health that the land manager has identified and attempts to achieve. Mathematically, the conservation problem can be formulated as minimizing the cost of land acquisition necessary to establish conservation reserve network as green infrastructure that will sustain Q0:

Min E = l'C®n r

st.Q (T) > Q0

2 In anticipation of Eq. (4), we allow heterogeneity of parcels across landowners, so rents and incentive payments may be at different rates for different landowners, based on their parcel's role in the ecosystem and value in real estate markets.

where I denotes a vector of unity and Eq. (6) uses Kronecker multiplication ®. The solution to the problem prescribes the proportion of each land parcel to be acquired for preservation that could achieve the conservation goal Q0 at minimum cost E0. Note that the cost E0 represents the level of budget needed to achieve the goal Q0, assuming all landowners cooperate with the plan and assuming (or conditioned on) the planner's appraisal of the opportunity cost of each parcel is correct.

Suppose the land manager has identified, based on a set of fair

compensation rates R'(R'), an optimal solution to the conservation program - a set of preservation ratios r0 = {ym„:ymn £[0,1]} for land parcels in r under the conservation goal Q0, where m and n are row and column indices for parcels (grid cells) in matrix r. Denote ym„ as the identified optimal preservation ratio for a land parcel Lmn. With a voluntary program, the owner of the parcel Lmn may select a

development level (1 —Ymn') that is beyond the level allowed by ym„ such that the landowner is only willing to enroll less than the optimal proportion of the parcel in the conservation program. In such case, the land manager needs to identify any adjustment in the set r0 = {ym„:ymn £[0,1]} necessary to account for the insufficient preservation on parcel Lmn in achieving the goal Q0. The conservation spending to achieve the desired level Q0 accounting for the landowner's decision can be estimated as follows:

Min E1 = I'C®n (8)

s.t.Q (r) > Q°(r°) (9)

Jmn'<J*mn (10)

where E1 is the new minimum cost to reach the goal Q0 by preserving undeveloped land with less than the ideal, cost minimizing proportion (from Eqs. (6)-(7)) of parcel Lmn available for preservation.

From Eqs. (8)-(10), denote the alternative set of preservation ratios as r1. By the mathematical condition of optimization (minimization), E1 = E(r1) is no less than E0 = E(r0). Consequently, the limited availability of parcel Lmn due to overdevelopment might negatively affect the community's conservation effort by necessitating an increased cost to achieve the goal Q0. In our EIF framework, this cost increase E1 — E0 would be imposed as the EIF on the development of parcel Lmn for its damage to the cost of the conservation goal Q0. In general, if the parcel is proposed in r0 for something less than full preservation (ym^ < 1) and the landowner proposes to develop more than this optimal proportion but not necessarily entirely, then the EIF would be scaled by the additional intensity of development, ym„ — ymn'.

4. An Empirical Illustration

This section demonstrates, in a case study, the empirical estimation of ElFs for the impact of land development on critical habitat. This example uses a spatially explicit model to capture the functional connections between habitat patches in nature reserve networks that can sustain sensitive wetland species as an indicator of a healthy ecosystem.

4.1. Study Region and Ecosystem

We focus on an area in southern Rhode Island, which involves several townships including West Greenwich, East Greenwich, Richmond, Exeter, and North Kingstown (see Fig. 1). This region is geologically notable for an abundance of glacial kettles and other shallow depressions (Flint, 1971), which support temporary wetlands providing a suitable environment for many wetland-dependent species, including birds and amphibians (Rhode Island Department of Environmental Management, 2005). Yet, concentration of large cities within commuting distances from southern Rhode lsland makes the focal area an ideal place for permanent residence or summer houses with coastal access. Rapid land conversion associated with residential development represents a serious challenge to local ecosystems (Skidds et al., 2007) and land use management in the local communities3 (Ryan, 2002, 2006).

A main ecological feature of the study area is the presence of species that may be characterized by metapopulation structure. ln

3 In Rhode Island, all land is incorporated and these communities are independent. However, the towns collaborate through a private non-profit watershed association and an intergovernmental land conservation alliance. For the purposes of the present discussion, we set aside governance details and treat the watershed as a single community planning for ecosystem health within the watershed.

metapopulation ecology, the landscape is regarded as a physical "matrix" (or template) in which scattered habitat patches are embedded; each patch supports a local population of a species, with extinction of the population or recolonization of the patch dependent upon the species ability to disperse across the landscape matrix from surrounding occupied patches known as source patches (Hanski and Gilpin, 1997; Hanski, 1998). One important feature of species characterized by a metapopulation structure is that local extinction may not necessarily cause regional species extinction, and species long-term persistence depends on the dynamic balance of successful recolonization from source patches. By this model, land development may negatively affect species persistence by disturbing the recolonization process through restricting or destroying either the dispersal corridors or patches which could provide a source of individuals for dispersal.

While many ecosystems could be used to illustrate the application of ElFs in land use management, we chose to focus on a metapopulation structure for two reasons. First, while some species may be adapted to habitats that naturally exist in a patchy pattern across the landscape, historic land development in a community may have generated a landscape with remnant habitats such that native species exist in a metapopulation structure shaped by human land-use. The human community may then judge that future, unacceptable losses of ecosystem health may occur with further landscape fragmentation following additional development. This intrinsic concern for ecosystem health may lead a community to explore tradeoffs between environmental quality and development, ultimately choosing a target level of environmental quality Q and using a collectively chosen indicator species' population as an index for Q. The metapopulation, or patch occupancy model, for species persistence may then constitute a useful framework for connecting human concerns with natural systems in a planning model (Bauer etal.,2010).

Second, addressing ecosystem health (environmental quality goals) through a metapopulation structure provides a convenient case with sufficient ecological sophistication to illustrate ElFs as a policy instrument suitable for managing critical land that plays a spatial role of ecosystem value while ElFs also facilitate financing conservation spending. Conservation practices have increasingly relied on designation of spatially located nature reserves as a cornerstone to sustain a healthy ecosystem or to provide environmental goods (Noss and Cooperrider, 1994; Swallow, 1996b). While the spatial relationship of preserved land parcels can be critical to the ecological process underlying population dynamics and species survival, developing one land parcel in the reserve network could compromise the ability of the whole network to sustain ecological health. This heterogeneity and interdependency of preserved land exemplifies a typical challenge for assessing ElFs as compared to using development impact fees for financing conventional types of local infrastructure, such as schools or roads.

4.2. Bioeconomic Model

To quantify the linkage between land preservation and landscape capacity of a metapopulation, we integrate two bioeconomic models developed, respectively, by Jiang et al. (2007) and Bauer et al. (2010) for designing nature reserve networks for ecosystems supporting a species in a metapopulation structure. Specifically, we adopt the model structure of Jiang et al. (2007) but re-formulate the model as a discrete non-linear programming problem to allow a working landscape with development interacting with habitat patches and corridors (Polasky et al., 2005,2008). Our model considers the aggregate population of migrating individuals, instead of the extinction probability modeled in Jiang et al. (2007), to measure the level of ecosystem health. This approach assumes that dispersal of more individuals can lead to a lower probability of overall species extinction and thus a healthier ecosystem. We incorporate the metapopulation-explicit landscape element of Bauer et al. (2010) to re-define habitat patches and corridors to include

Fig. 1. Demonstration of habitat patches in the study area. The area is located in the Wood-Pawcatuck River Watershed in southern Rhode Island, USA. Each circle with a unique ID number represents a habitat patch.

core habitat plus upland buffer zones and dispersal matrix while allowing for heterogeneous land costs.

Eq. (11) describes the ecosystem health of the community landscape that can be achieved by preservation of selected land parcels as habitat patches and corridors for dispersal (see Appendix A for derivation of Eq. (11)):

Q (r) = EE (1-e~^Xi) (1 -e-liX> )mji1—e~KZi')P'j (11)

' j£Ni

where Q denotes the aggregate population of migrating individuals, reflecting the degree of ecosystem health; mj measures the size of the local population of amphibians in patch j; Pj represents the effective rate of dispersal from habitat patch j to patch i on an undisturbed landscape with complete connectivity; (3 and k are biologically relevant parameters linking population size and dispersal rates to the ecosystem health Q; and X' (or Xj) and Zj are variables that indicate, respectively, the proportions of habitat patches (Xs) and corridors (Zs) to be preserved that link the land use of grid cells or land parcels in landscape matrix r to the re-colonization process of the species. The adoption of patch and corridor variables, X' and Zj, has the advantage of being tied directly to the ecologically-based, metapopulation model reviewed and developed by Bauer et al. (2010). That model links each parcel to a spatially delineated role in the ecosystem structure. Our model then chooses the proportion of each patch or corridor that would be

preserved, thereby setting the value of y for each grid cell within the patch or corridor.

This production function (Eq. 11) implies that: i) the ecosystem value of a neighboring patch j in contributing to dispersal of individuals that colonize a patch i depends on the proportions of both patches being preserved, given by X' and Xj, and ii) the dispersal capacity of species may be enhanced by preserving a larger portion of the dispersal corridor between patches j and i, as measured by Zj. Specifically, the contributions to ecosystem health of habitat patches and corridors increases exponentially with the proportions of land preserved within the patches and corridors.

The patch variable X' may vary between 0 and 1, indicating the portion of a biologically identified patch i to be preserved. Patches are biologically identified to include both core wetland habitats and their surrounding upland habitats (e.g., Semlitsch and Bodie, 2003). We define habitat corridors as land parcels creating a 100-meter wide segment connecting pairs of those biologically identified patches. The corridor variable Zj may vary between 0 and 1, assuming that further expanding the 100 m wide segment of land parcels connecting patches i and j does not functionally enhance species dispersal. The cost minimization identifies the land to be preserved by using Eqs. (6)-(7), with Eq. (11) implementing Q; we then calculate EIFs by using Eqs. (8)-(10), with Eq. (11). In the cost minimization of Eq. (11), the decision variables are the Xs and Z's that directly and correspondingly link the decisions for ecologically-relevant land units to the landscape matrix r. With the above setting, we can conduct a series ofempirical exercises to estimate

EIFs for development damage to land of ecosystem importance (see Appendix B for the specific procedure).

4.3. Data

4.3.1. Biological Data

We use pond-breeding amphibians as indicator species to identify landscape elements that are critical to sustaining the ecosystem characterized by species exhibiting a metapopulation structure (Green, 2003; Semlitsch, 2007). Pond-breeding amphibians have complex life cycles that involve varying types of habitat at different development stages that expose them to high risk from habitat loss and fragmentation in both wetlands and adjacent uplands (Baldwin and Demaynadier, 2009). Indeed, habitat loss, degradation, and fragmentation caused by urbanization currently threatens over one-third of the world's known amphibian species (Hamer and McDonnell, 2008), which require preservation of not only terrestrial and aquatic habitat-patches but also landscape connectivity in corridors to facilitate dispersal (Cushman, 2006; Hamer and McDonnell, 2008). Amphibians provide an opportunity to identify critical habitat and to assess the impact of rural land use change on ecosystem health in a spatial framework (Homan et al., 2004; Porej et al., 2004; Skidds et al., 2007; Egan and Paton, 2008). While Rhode Island has five species of amphibians that breed only in temporary ponds, our empirical demonstration focuses on wood frogs (Rana sylvatica) for which biological data were available (see Crouch and Paton, 2000; Skidds and Golet, 2005; Skidds et al., 2007).

The dispersal behavior of amphibians is critical to modeling landscape connectivity, which can affect the selection of conservation strategies and land to be preserved (Jiang et al., 2007). Ideally, field studies would calibrate a dispersal parameter, a, which determines the effective dispersal rate Pj in Eq. (11) and which characterizes the migration behavior of the indicator species within landscapes involving different land use combinations. Parameter a represents a species-specific effect of distance dj between patches i and j on effective dispersal rate. Unfortunately, available field data are limited so we adopt the only available dispersal function, Pj = (0.4392)10-Qdjja=0.000560 in Eq. (11), estimated by Berven and Grudzien (1990) for wood frogs and we use available literature to calibrate the a for our study area following Jiang et al. (2007).

4.3.2. Land Cost Data

We use tax-assessed land value data maintained by local governments to measure the acquisition cost (or the government offered compensation for preservation) of land (i.e., r'(R')). In the study area, land values are appraised every three years and the tax-value appraisals are regarded as a good proxy of the market value of land parcels.4 Tax-assessed land value may best represent the cost information available during a community effort to plan for preservation of nature reserves as a community's green infrastructure.

The land parcels for which the tax-assessed value data has been collected are ofvarying size and shape, and therefore a procedure is needed to convert the tax-assessed value ofland parcels to the economic cost of our modeling unit, i.e., habitat patches and corridors. In our model, the landscape is divided into a matrix of square grid cells of 1 ha. Since the spatial information of the land parcels is unavailable, except for a corresponding postal address, we are not able to proportionally split the tax-assessed land value into component grid cells. Instead, we employ a spatial regression technique to estimate the natural logarithm of per unit land value, V, as a function of land attributes, and then apply this spatial hedonic function to estimate the value of each grid

cell based on its own attributes (see Appendix C for estimating the land value function). With the estimated land costs of grid cells, the land cost of each patch and corridor is derived based on their constituent grid cells.

4.4. Results

In this empirical illustration, the set r0 = {ym„:ymn £[0,1]} prescribes the proportions of one-hectare cells that comprise patches and corridors to be preserved through outright purchases in a conservation plan. Tables 1 and 2 present the optimal preservation ratio for the patches and corridors identified and estimated EIFs for different levels of development damage to those patches and corridors. In both Fig. 1 and the tables, each patch and corridor is labeled by a unique identification number.

As indicated by the preservation ratios in Tables 1 and 2, preserving the entirety of the patches and corridors identified is not necessary to sustain the target level of ecosystem health with the available budget of $3 million.5 Indeed, the majority of the identified patches and corridors have an optimal ratio <0.5, suggesting that only a proportion of the patch and corridor needs to be preserved such that some development is allowed on those land parcels. Table 1 also shows that there are some patches (e.g., patches 69, 84 and 85 with preservation ratios > 0.5) that require not only full preservation of core habitat but also preserved surrounding upland buffer zones.

In this empirical illustration, development damage is measured by the relative percentage loss of the proportion (i.e., the preservation ratio) of each patch or corridor that needs to be preserved for the conservation program to cost-effectively achieve its goal within the allocated budget. For example, patch 67 has an optimal preservation ratio prescribed at 0.48, which indicates 48% of the identified patch needs to be preserved. For this patch, a 10% development damage is modeled as a 10% reduction (due to development) in its preservation ratio such that only around 43% of the patch is available at most for preservation although a 48% preservation is optimal (cost-effective). Similarly, 100% development damage refers to the complete loss of the portion (48%) of the patch required to be preserved for the target level of ecosystem health.

As demonstrated by Tables 1 and 2, the estimated EIFs exhibit two characteristics that deserve mentioning. First, Tables 1 and 2 show that for the same level of development damage, the EIF varies across the patches. This result can be attributed to the heterogeneity of individual land parcels in sustaining ecosystem health relative to their costs, their role in the ecosystem health function Q, and the costs and role of the replacement land. A low EIF implies that the related patch is less critical to maintaining the target level of ecosystem health with the available budget, and it is relatively easier to compensate the damage by minor adjustments in the conservation program without incurring a significant cost increase. In contrast, a high impact fee indicates an ecologically valuable patch such that any development damage would require expensive adjustment in the conservation program, including acquiring more expensive land to enhance other identified patches and corridors available for preservation or to target new patches and corridors. It is these factors, of linkage to the ecological model and relative costs, that establish that EIFs may meet the judicial tests required for acceptability under the U.S. Constitution's prohibition against takings of private property for public purposes (see Innes, 1997; Fenster, 2007). These tests include: 1) rough proportionality between the EIF and its effect on the community facility, in this case the land

4 Rhode Island law requires assessments based on professional market appraisal methods every 3 years.

5 This budget is established based on tax-assessed values for raw land averaged with vacant land available for buildable lots. Purchase of land that had already completed an engineering assessment and building and zoning review process may be substantially more costly; the purpose here is to provide an illustrative example.

Demonstration of estimated impact fees for different levels of development damage to habitat patches.

Patch Preserve. ratio Impact fee, $/acre Patch Preserve. ratio Impact fee, $/acre

10% damage 50% damage 100% damage 10% damage 50% damage 100% damage

67 0.24 88.26 2628.95 5706.88 191 0.51 46.44 2224.55 15,590.30

69 0.61 66.19 3695.85 36,386.16 193 0.50 43.91 2045.99 12,640.97

70 0.25 132.37 4233.56 13,277.74 205 0.33 80.12 2661.11 5645.02

71 0.40 400.07 16,386.35 74,767.17 872 0.43 37.10 1592.89 11,755.83

82 0.55 69.83 3513.03 24,961.59 873 0.44 36.70 1583.02 11,755.83

83 0.52 68.51 3260.81 17,361.02 922 0.30 164.05 5676.50 21,412.28

84 0.63 110.59 6442.16 81,438.47 924 0.30 151.61 5192.06 18,765.61

85 0.64 103.03 6165.68 82,655.27 925 0.29 148.41 5057.36 17,985.92

86 0.54 83.03 4124.22 32,360.90 927 0.28 149.44 4514.24 8895.12

88 0.50 53.57 2492.45 17,860.28 928 0.38 308.99 11,832.07 40,806.73

89 0.48 45.15 2032.94 11,555.43 929 0.32 221.86 8026.97 36,091.62

91 0.35 127.86 4592.73 7438.67 935 0.21 116.22 3514.02 10,108.90

92 0.48 53.85 2387.50 3107.51 941 0.31 191.89 6272.53 15,622.46

93 0.33 119.39 3700.59 3107.51 942 0.45 172.17 7280.33 35,989.27

94 0.25 78.72 2228.13 2293.37 943 0.37 128.31 4737.45 22,784.57

95 0.45 143.33 5982.88 37,390.22 945 0.24 124.66 4027.05 15,495.36

96 0.45 151.52 6557.64 35,097.40 946 0.59 57.90 3120.23 16,655.28

162 0.31 91.05 3213.88 11,665.41 951 0.36 82.56 2979.44 10,190.67

163 0.31 91.16 3215.03 11,665.41 952 0.60 108.58 5992.70 53,275.94

164 0.31 167.50 5945.08 24,201.07 953 0.49 169.97 7687.47 44,486.00

165 0.32 149.49 5205.27 19,616.79 954 0.42 237.16 9769.94 57,042.50

173 0.42 31.34 1178.73 3265.70 955 0.40 208.31 8189.25 48,787.65

174 0.49 41.77 1238.28 8865.53 959 0.27 78.54 2515.34 6614.47

175 0.42 29.86 1258.51 9470.19 960 0.47 42.35 1851.08 7675.97

176 0.30 11.03 257.95 388.47 1127 0.35 143.88 5422.67 20,624.94

177 0.36 21.20 799.49 3643.31 1215 0.36 142.40 5397.94 20,624.94

190 0.54 37.19 1849.11 13,748.84

conservation network, treating similarly situated lands similarly conservation purpose. The magnitude of the ElFs has important im-based on their role within the ecological model; and 2) essential plications for cost-effective targeting of preservation land when not nexus between the treatment of the lands and the community's all land of ecosystem value can be preserved instantaneously.

Table 2

Demonstration of estimated impact fees for different levels of development damage to corridors between habitat patches.

Corridor Preserve ratio Impact fee, $/acre Corridor Preserve ratio Impact fee, $/acre

i j 10% damage 50% damage 100% damage i j 10% damage 50% damage 100% damage

67 69 0.16 2.75 74.51 328.80 88 89 0.48 13.62 450.79 2693.60

67 70 0.02 0.06 1.57 6.36 91 959 0.37 25.61 755.71 2091.70

67 71 0.38 17.99 539.19 1333.10 91 960 0.27 19.37 554.83 2173.80

69 70 0.14 27.81 748.56 3289.20 92 93 0.31 18.16 535.28 2577.30

69 71 0.51 317.33 10,747.00 66,705.00 92 94 0.30 5.85 171.61 842.56

70 71 0.46 436.66 13,968.00 46,326.00 93 94 0.22 12.65 334.80 586.85

70 164 0.01 0.03 0.78 3.16 93 95 0.03 0.43 10.92 44.39

71 164 0.33 35.92 1075.90 5384.40 95 96 0.55 55.97 1946.10 12,230.00

71 165 0.23 25.67 723.35 3330.70 95 205 0.16 4.77 128.86 563.36

82 83 0.31 5.78 172.84 883.71 96 205 0.28 13.55 391.43 1819.70

82 84 0.39 7.69 240.85 1322.60 162 163 0.75 141.87 5515.20 21,414.00

82 85 0.41 8.36 266.02 1494.90 164 165 0.64 170.48 6215.30 41,987.00

82 86 0.13 1.21 32.41 142.20 173 174 0.10 3.45 89.95 376.33

82 88 0.01 0.00 0.10 0.41 173 177 0.46 8.35 268.13 1158.00

83 84 0.26 5.19 150.67 738.14 174 175 0.76 25.20 1007.50 7144.90

83 85 0.19 2.73 75.89 348.53 174 176 0.12 3.83 88.88 162.37

83 86 0.31 7.02 209.19 1064.70 190 191 0.45 8.65 281.61 1628.70

84 85 1.00 70.40 2613.50 29,171.00 190 193 0.36 4.60 141.06 745.20

84 86 0.59 22.40 798.75 5403.00 191 193 0.39 7.76 241.67 1302.90

84 88 0.37 9.42 291.79 1572.40 872 873 0.94 44.40 2021.90 16,162.00

84 89 0.27 5.64 164.23 806.82 922 924 0.49 58.25 1907.30 10,583.00

85 86 0.44 13.74 444.30 2563.40 922 925 0.48 48.95 1592.60 8779.90

85 88 0.54 13.06 450.88 2882.00 924 925 0.41 165.73 5177.00 27,216.00

85 89 0.40 9.01 283.31 1555.80 927 928 0.31 217.98 6405.10 19,363.00

86 88 0.06 0.27 7.03 29.39 927 929 0.20 21.29 586.63 2639.60

927 941 0.07 1.38 35.61 148.52 951 952 0.39 13.06 407.31 2132.50

928 929 0.58 129.38 4549.90 24,189.00 951 953 0.16 3.53 95.89 424.99

928 941 0.20 17.27 481.83 2213.90 952 953 0.51 34.68 1174.80 7313.60

928 942 0.03 0.42 10.75 43.98 952 954 0.15 6.55 178.75 802.34

935 945 0.49 11.40 372.38 2069.50 952 955 0.08 1.61 42.09 178.50

935 946 0.49 11.70 379.97 1585.40 953 954 0.19 8.58 239.28 1106.80

941 942 0.33 37.77 1136.40 5743.90 953 955 0.15 4.33 117.66 524.46

941 943 0.06 1.19 30.62 127.15 954 955 0.72 201.15 7821.70 59,732.00

942 943 0.47 54.88 1803.10 8417.80 1127 1215 0.55 161.56 5499.40 19,810.00

945 946 0.61 14.34 512.08 3130.90

Second, as expected, EIFs can increase quickly with the extent of development damage to patches and corridors. Table 1 shows that for a 10% development damage, the estimated EIF on a per acre basis ranges from $11.03 for patch 176 to $400.07 for patch 71; in contrast, when the development damage reaches 100%, the majority of the estimated EIFs are on the order of $10,000 per acre with the highest at over $82,000 per acre for patch 85. This result implies that: 1) conservation cost increases non-linearly, at a rising rate with development damage, suggesting it is relatively cheaper and easier to establish a conservation reserve in the early stage of community development, and 2) minor development damage could be mitigated at low costs and it can be very expensive to compensate severe development damage to land of ecosystem value. Note that extremely high EIFs seem likely to prevent complete development to the corresponding patches and corridors of critical ecosystem value and yet these high EIFs remain near or below typical market (tax) value of buildable lots in the study area.

5. Conclusion with Discussion

While traditional land use planning has been criticized for being mainly focused on development, the so-called "smart growth" symbolizes a tendency in modern land use regulation to promote an environment-balanced community growth (e.g., Ye et al., 2005). Yet, maintaining desirable environmental amenities while accommodating land development is challenging, particularly at the urban fringe where conservation unfavorably competes with development for private land in free markets. On one hand, residential development represents a profitable use for undeveloped land while generating a negative impact on the environment and rural quality of life. On the other hand, the environmental amenities provided by preserved land that are valued by the public may not fully and exclusively accrue to their providers. Because of the market's failure to internalize the externalities associated with land use, more land than is socially optimal may be converted to developed use in a laissez-faire model of community growth. Policy intervention is needed to protect land of environmental importance, and thereby sustain environmental amenities and foster efficient land use in communities at the urban-rural fringe.

In this study, we propose an impact fee framework coupled with conservation payments to implement cost-effective land preservation consistent with local ecosystem characteristics, while considering both landowner development rights and public interest in the environment. The EIF framework is based on the understanding of both the relative advantages of different land manage approaches and the need for a combined use of multiple policy instruments that reinforce and complement each other to increase effectiveness and minimize unintended consequences (Bengston et al., 2004). In the proposed framework, conservation payments are devised to compensate enrollment of undeveloped land in preservation while acknowledging landowner development rights; meanwhile, EIFs may be imposed on land development to internalize the potential damage of development on the public's environmental interest and to finance mitigation efforts to maintain environmental quality. This study is intended to illustrate the conceptual model of EIFs applied to managing private land of ecosystem value and to complement conservation payments in protecting local ecosystems.

Our conceptual model of EIFs can be a useful tool for community land use management. It combines the "beneficiary pays principle" from public finance and the "polluter pays principle" from environmental management to address the potential conflict between decentralized land development and public environmental interest. It represents a direct extension of development impact fees in the public sector to managing private land of environmental value and to financing green infrastructure and environmental goods and services. It internalizes the negative impact of private land development on the cost-effectiveness of maintaining an environmental quality target identified for the public interest, and the EIF could reduce the gap between the values

of development and preservation and thus mitigate possible overdevelopment of community land, with the potential to lead toward more efficient land use patterns.

EIFs offer an alternative approach complementing existing market-based instruments for protecting local environmental amenity. Transferable development rights (TDRs) are a tool that has long been discussed and practiced for land conservation for diverse purposes (Pruetz and Standridge, 2008; Kaplowitz et al., 2008; McConnell and Walls, 2009). Essentially a tradable permit scheme or a quantity-based approach, TDRs as compared to our EIFs have various advantages and weaknesses. Clinch and O'Neill (2010a) extended the classic understanding of pricing versus quantity approach (i.e., Weitzman, 1974) to land use management and planning, identifying theoretical conditions under which a pricing approach such as our EIF outperforms a TDR approach in terms ofsocial welfare outcome. More importantly, a common understanding of empirical literature is that TDRs require an active market with sufficient demand and supply, but a "thin" market is often found to limit the operation of a TDR approach where there is insufficient demand or supply in a community or a specific area to allow frequent market trading of development rights under competitive (efficient) conditions (McConnell and Walls, 2009). In this regard, our EIFs exhibit relative advantages as they do not require market transactions. EIFs can target individual land development with government enforcement that accommodates land development while internalizing the external cost impact of development on cost effective conservation of a community's environmental or ecosystem-based amenity at a desired level.

The EIF framework also addresses important political or institutional concerns. In a social context that protects private property rights, the concept of EIFs like any land exactions inevitably may interact in the legal arena with the "takings" of private property (Innes, 1995,1997, Innes et al., 1998, Curtin and Gowder, 2003, Fenster, 2007). While the actual application of the takings clause of the U.S. Constitution is case-specific, previous jurisprudence and case law have established "essential nexus" and "rough proportionality" as two common judicial tests for claims of takings (Nollan v. the California Coastal Commission, 1987; Dolan v. City of Tigard, 1994). We view our conceptual model of EIFs as consistent with the legal framework by tying EIFs of land development quantitatively and scientifically to development's marginal impact on the public cost of environmental conservation (or the replacement cost of damaged habitat for maintaining environmental quality through a conservation network representing a public infrastructure or facility). By placing land parcels in a quantitative model capturing their ecological role in an environmental quality index, and by using professional appraisal data for estimating land costs and planning a community's green infrastructure, the EIFs place land situated to play a similar ecological role in the community's greenway plan on equal footing. Our empirical example illustrates the application of the conservation science in land use regulation consistent with such a legal framework, establishing the EIFs in a manner that is tied to (or, judicially, roughly proportional to) the cost a development decision would impose on the community's effort to maintain a minimum of ecosystem structure and function. While the implementation of our approach requires some ecological and economic modeling expertise, this expertise resides within the capabilities of many public agencies (at least at a state level), perhaps in collaboration with academia. While advancements in science may improve modeling and estimation, it should be noted that the legal tests to allow an EIF approach explicitly seeks a "rough" proportionality, allowing a practical perspective not unlike Portney's (2004) admonition.

Funding

This research was partially supported by the US EPA STAR Grants R829384 and R828629; the USDA/CSREES/NRI Grant 2002-3540111657, the Rhode Island Agricultural Experiment Station, and the

University of Connecticut Agricultural Experiment Station project CONS00971.

Appendix A. Formulation of the Ecosystem Health Measure for Community Landscape in Relation to Land Conservation

This appendix provides the linkage between past work cited and our Eq. (11). For our bioeconomic model, the objective of model formulation is to describe and measure the ecosystem health of the community landscape in relation to preservation of certain land parcels as habitat patches and corridors connecting patches allowing species dispersal. We approach this objective by following the modeling structure of Jiang et al. (2007), which links preservation of land parcels as habitat patches and corridors (via the probability of recolonization (dispersal) of migrating species) to the probability of species extinction which implicitly assumes, based on the metapopulation theory, that more dispersal of individuals can lead to a lower probability of overall species extinction. Our model, however, differs from that ofJiang et al. (2007) by directly modeling the aggregate population of migrating individuals instead of the extinction probability. Our approach is mainly motivated by the need to account for a working landscape and to incorporate the metapopula-tion-explicit landscape element of Bauer et al. (2010), thus allowing for the preservation of core habitat areas plus upland buffer zones and portions of dispersal matrix.

Consider a habitat patch i and a neighboring patch j, which both are undisturbed remaining in their natural condition. Denote nOj as the number of individuals migrating from patch j to i, which can be calculated as:

n0 = mP

where mj measures the size of the local population of species in the neighboring patch j, and Pj represents the effective rate of dispersal from habitat patch j to patch i on an undisturbed landscape with complete connectivity.

To capture the impact of land availability or conservation status of habitat patches and dispersal corridors in a working landscape, we introduce the decision variables Xj and Zj to indicate, respectively, the proportions of pre-identified habitat patches (Xs) and corridors (Zs) to be preserved that link the land use of grid cells or land parcels in the landscape matrix r to the recolonization process of the species such that the effective number nij of individuals migrating from patch j to i can be calculated as:

.... = C1 -e-3Xi

nij = 0

mj( 1-e-KjPj

where ¡3 and k are biologically relevant parameters. Note that here we introduce two multiplying factors in front of mj and Pj, respectively, to effectively capture the impact of the conservation status (i.e., the percentage of conservation land) of habitat patches and corridors on migrating individuals and dispersal capability. As implied by the equation, if habitat patch j is fully developed, then Xj becomes zero and thus the contribution of migrating individuals from patch j is zero to reflect species loss due to destroyed habitat; similarly, if habitat corridor Zj is fully developed, then the species dispersal cannot happen. We use the exponential formulation here as it captures many biological and ecological relationships. Of course, the biological parameters can be validated or calibrated with experiment or field data; for this study, parameters were drawn from literature as discussed in the main text.

Therefore, the total number ni of species migrating to patch i from all neighboring patches can be calculated as:

ni = E (l~e-3Xj

)mj(1-e-

Zi')Pa

Similarly, the aggregate number of migrating individuals for the community landscape can be calculated as:

Q (0 = E (1-e

-e-fiXi)

(l-e^) E (1-e-3X>)mj( 1-e-KZ«)Pij

which leads to the objective function or the ecological benefit function of land conservation, i.e., Eq. (11) in our empirical application:

Q (r) = EE (1 -e-№) (1 -e-3X )mj (1 -e-KZj )P„(A5) i jm

Appendix B. Empirical Procedure for Estimating EIFs

The estimation procedure involves two steps. First, a fixed amount of conservation budget E0 is assigned, which allows estimation of the maximum level of ecosystem conservation Q0 and the optimal set of preservation ratios r0 = {ym„:ymn £[0,1]}for selected patches and corridors.6 Define the maximum level of ecosystem conservation Q0 achievable for the budget E0 to be the socially efficient level of conservation, as selected by the community through a public process. By duality, the budget E0 represents the minimum cost required for the conservation goal Q0, and thus may be used as the bench mark to calculate the EIF for any development beyond the allowed level represented by preservation ratios r0 = {ymn:ymn £[0,1]} for each patch or corridor.

The second step involves re-solving the cost-minimizing conservation model for a series of different model constraints on the optimal set of preservation ratios. The model is re-solved for each parcel involved in the initial conservation network wherein ymm is non-zero. Those new constraints represent the proportions of patches and corridors that are available depending on each landowner decision, as represented by Eq. (10) transferred to constraints on the Xs and Z's. The minimized cost E1 for the constrained model represents the new conservation spending needed to account for a landowner decision. The difference between the minimized cost E1 and the budget E0 is the EIF for the development damage as represented by the model constraint (10) on the preservation ratio. For demonstration purposes, we consider a single development damage to a patch or corridor at one time, assuming availability of other land of ecosystem importance.7 Repeating the second step allows estimation of ElFs for different levels of development damage to each patch and corridor via the parcels Lmn that each patch or corridor incorporates.

Appendix C. Estimating Land Value Function for Generating Landscape Cost Matrix

To estimate the hedonic land value function, we specified a general spatial econometric model that included both spatial lag and spatial error to account for possible spatial processes (Anselin, 1988):

V = pW1V + Y8 + 8

8 = \W28 + u

<N (0, ct2

(A6) (A7)

where V is a vector of n observations of the natural logarithm of per unit land value, Y is an n x k matrix of independent variables, s is an nx 1 vector of spatially correlated errors, u is annx 1 vector of independently normally distributed disturbance with mean zero and variance o2, W1 and W2 are weight matrices describing spatial relations among

6 The conservation budget E0 may be justified by aggregate willingness to pay of the public for ecosystem conservation.

7 It is straightforward to extend the modeling framework to multiple damages to patches and corridors, which we did not consider in this example.

each observation, and 6, p, and \ are coefficient parameters to be estimated. A matrix manipulation of the spatial model yields

V =(I-pW1)"1Y6 + (I-pW1)"1(I-\W2)"1u u ~ N(0,ct2), (A8)

where I represents the identity matrix.

We first used the consumer price index to standardize all values from either 2003 or 2004 to the 2005 level. We adopted, with modifications, a public Matlab toolbox for spatial econometrics developed by LeSage (1998) to estimate the spatial model with alternative weight matrices, as described by Jiang et al. (2007). We applied the estimated hedonic land value function to assign a cost to each grid cell in the standard landscape matrix r based on their attributes, i.e., the independent variables Y. However, the spatial scale of the grid cell was much smaller than that of the actual land parcels associated with the land value data such that a direct application of Eq. (A8) caused a serious computing problem in calculating (I — pWi)—1 due to the large spatial weight matrix W1. Consequently, we used an iteration approach to approximate the land value for each grid cell:

V = (pW1 + p2W1 + ... + pnWn) Y8 + pn+1 W?+1V (A9)

Since W! is a distance-based matrix whose elements are the inverses of the distances between land parcels, pn+1Wn+1V would converge to zero for a sufficiently large n. In this example, we take n equal to 100.

References

Anderson, C.M., King, J.R., 2004. Equilibrium behavior in the conservation easement game. Land Econ. 80, 355-374.

Ando, A., Camm, J., Polasky, S., Solow, A., 1998. Species distributions, land values, and efficient conservation. Science 279,2126-2128.

Anselin, L., 1988. Spatial Econometrics: Methods and Models. Kluwer Academic Publishers, Dordrecht.

Araujo, M.B., Williams, P.H., Fuller, R.J., 2002. Dynamics of extinction and the selection of nature reserves. Proc. R. Soc. Lond. B 269,1971-1980.

Arendt, R.G., 1999. Growing Greener: Putting Conservation into Local Plans and Ordinances. Island Press: Washington, DC.

Baldwin, R.F., Demaynadier, P.G., 2009. Assessing threats to pool-breeding amphibian habitat in an urbanizing landscape. Biol. Conserv. 142 (8), 1628-1638.

Bauer, D.M., Swallow, S.K., Paton, P.W.C., 2010. Cost-effective species conservation in exurban communities: a spatial analysis. Resour. Energy Econ. 32 (2), 180-202.

Beatley, T., 2000. Preserving biodiversity - challenges for planners. J. Am. Plan. Assoc. 66, 5-20.

Bengston, D.N., Fletcher, J.O., Nelson, K.C., 2004. Public policies for managing urban growth and protecting open space: policy instruments and lessons learned in the United States. Landsc. Urban Plan. 69 (2004), 271-286.

Bennett, A.F., Radford, J.Q., Haslem, A., 2006. Properties of land mosaics: implications for nature conservation in agricultural environments. Biol. Conserv. 133, 250-264.

Berven, K.A., Grudzien, T.A., 1990. Dispersal in wood frog Rana sylvatica: implications for genetic population structure. Evolution 44, 2047-2056.

Brueckner, J.K., 1997. Infrastructure financing and urban development: the economics of impact fees. J. Public Econ. 66 (3), 383-407.

Burge, G.S., Nelson, A.C., Matthews, J., 2007. Effects of proportionate-share impact fees. Housing Policy Debate 18 (4), 679-710.

Carter, T., 2009. Developing conservation subdivisions: ecological constraints, regulatory barriers, and market incentives. Landsc. Urban Plan. 92 (2), 117-124.

Clinch, J.P., O'Neill, E., 2010a. Assessing the relative merits of development charges and transferable development rights in an uncertain world. Urban Stud. 47 (4), 891-911.

Clinch, J.P., O'Neill, E., 2010b. Designing development planning charges: settlement patterns, cost recovery and public facilities. Urban Stud. 47 (10), 2149-2171.

Crouch, W.B., Paton, P.W.C., 2000. Using egg mass counts to monitor wood frog populations. Wildl. Soc. Bull. 28, 895-901.

Curtin, D.J., Gowder, W.A., 2003. Exactions updates: when and how do the Dolan/Nollan rules apply? Urban Lawyer 35, 729-738.

Cushman, S.A., 2006. Effects of habitat loss and fragmentation on amphibians: a review and prospectus. Biol. Conserv. 128 (2), 231-240.

Daniels, T., Daniels, K., 2003. The Environmental Planning Handbook for Sustainable Communities and Regions. Planners Press-American Planning Association, Chicago.

Daniels, T., Lapping, M., 2005. Land preservation: an essential ingredient in smart growth. J. Plan. Lit. 19, 316-329.

Dolan v. City of Tigard, 1994. (S. Ct. No. 93-518).

Drechsler, M., Watzold, F., Johst, K., Shogren, J.F., 2010. An agglomeration payment for cost-effective biodiversity conservation in spatially structural landscapes. Resour. Energy Econ. 32 (2), 261-275.

Egan, R.S., Paton, P.W.C., 2008. Multiple scale habitat characteristics of pond-breeding amphibians across a rural-urban gradient. In: Jung, R.E., Mitchell, J.C. (Eds.), Urban Her-petology. Society for the Study of Amphibians and Reptiles, Salt Lake City, Utah.

Ernoult, A., Tremauville, Y., Cellier, D., Margerie, P., Langlois, E., Alard, D., 2006. Potential landscape drivers of biodiversity components in a flood plain: past or present patterns? Biol. Conserv. 127 (1), 1-17.

Fenster, M., 2007. Regulating land use in a constitutional shadow: the institutional contexts of exactions. Hastings Law Journal 58 (4), 729-72 +.

Fischel, W., 2000. Zoning and land use regulation. In: Boudewijn, B., De Geest, G., (edi), Encyclopedia of Law and Economics, Volume II: Civil Law and Economics, Edward Elgar, Cheltenham, pp. 403-423.

Flint, R.F., 1971. Glacial and Quaternary Geology. John Wiley and Sons, New York, NY, USA.

Green, D.M., 2003. The ecology of extinction: population fluctuation and decline in amphibians. Biol. Conserv. 111, 331-343.

Hamer, A.J., McDonnell, M.J., 2008. Amphibian ecology and conservation in the urbanizing world: a review. Biol. Conserv. 141 (10), 2432-2449.

Hanski, I., 1998. Metapopulation dynamics. Nature 396,41-49.

Hanski, I., Gilpin, M.E., 1997. Metapopulation Biology. Academic Press, San Diego, California.

Heikkila, E.J., 2000. Economics of Planning. CUPR Press, Rutgers University.

Homan, R.N., Windmiller, B.S., Reed, J.M., 2004. Critical thresholds associated with habitat loss for two vernal pool-breeding amphibians. Ecol. Appl. 14,1547-1553.

Innes, R., 1995. An essay on takings, choices: the magazine of food. Farm & Resource Issues 10 (1), 4-7.

Innes, R., 1997. Takings, compensation, and equal treatment for owners of developed and undeveloped property. J. Law Econ. XL, 403-432.

Innes, R., Frisvold, G., 2009. The economics of endangered species. Ann. Rev. Resour. Econ. 1,485-512.

Innes, R., Polasky, S., Tschirhart, J., 1998. Takings, compensation and endangered species protection on private lands. J. Econ. Perspect. 12,35-52.

Irwin, E.G., Bell, K.P., Bockstael, N.E., Newburn, D.A., Partridge, M.D., Wu, J.J., 2009. The economics of urban-rural space. Ann. Rev. Resour. Econ. 1,435-462.

Jiang, Y., Swallow, S.K., 2015. Providing an ecologically sound community landscape at the urban-rural fringe: a conceptual, integrated model. J Land Use Sci. 10 (3), 323-341.

Jiang, Y., Swallow, S.K., Paton, P.W.C., 2007. Designing a spatially-explicit nature reserve network based on ecological functions: an integer programming approach. Biol. Conserv. 140, 236-249.

Johnson, M.P., 2001. Environmental impacts of urban sprawl: survey of the literature and proposed research agenda. Environ. Plan. A 33, 717-735.

Kaplowitz, M.D., Machemer, P., Pruetz, R., 2008. Planners' experiences in managing growth using transferable development rights (TDR) in the United States. Land Use Policy 25, 378-387.

Kingsley, B.S., 2008. Making it easy to be green: using impact fees to encourage green building. New York University Law Review 83 (2), 532-567.

Kretser, H.E., Sullivan, P.J., Knuth, B.A., 2008. Housing density as an indicator of spatial patterns of reported human-wildlife interactions in Northern New York. Landsc. Urban Plan. 84 (3-4), 282-292.

Langpap, C., Wu, J.J., 2008. Predicting the effect of land-use policies on wildlife habitat abundance. Can. J. Agric. Econ. 56 (2), 195-217.

LeSage, J.P., 1998). Spatial Econometrics. accessed. http://citeseerx.ist.psu.edu/viewdoc/ download?doi=10.1.1.143.3703&rep=rep1&type=pdf.

Lewis, D.J., Plantinga, A.J., 2007. Policies for habitat fragmentation: combining econometrics with GIS-based landscape simulations. Land Econ. 83 (2), 109-127.

Lewis, D.J., Plantinga, A.J., Wu, J., 2009. Targeting incentives for habitat fragmentation. Am. J. Agric. Econ. 91 (4), 1080-1096.

Lewis, D.J., Plantinga, A.J., Nelson, E., Polasky, S., 2011. The efficiency of voluntary incentive policies for preventing biodiversity loss. Resour. Energy Econ. 33,192-211.

Marcouiller, D.W., Clendenning, J.G., Kedzior, R., 2002. Natural amenity-led development and rural planning. J. Plan. Lit. 16,515-542.

McConnell, V., Walls, M., 2009. U.S. experience with transferable development rights. Rev. Environ. Econ. Policy 3 (2), 288-303.

McConnell, V., Kopits, E., Walls, M., 2006. Using markets for land preservation: results of a TDR program. J. Environ. Plan. Manag. 49 (5), 631-651.

Mcdonald, R.I., Forman, R.T.T., Kareiva, P., Neugarten, R., Salzer, D., Fisher, J., 2009. Urban effects, distance, and protected areas in an urbanizing world. Landsc. Urban Plan. 93 (1), 63-75.

Myers, P., Puentes, R., 2001. Growth at the Ballot Box: Electing the Shape of Communities in November 2000, Discussion Paper, Center on Urban and Metropolitan Policy. The Brookings Institution, Washington, D.C.

Nicholas, J.C., Juergensmeyer, J.C., 2003. Market based approaches to environmental preservation: environmental mitigation fees and beyond. Nat. Resour. J. 43, 837-863.

Nollan v. California Coastal Commission, 1987.483 U.S. 825.

Noss, R.F., Cooperrider, A.Y., 1994. Saving Nature's Legacy: Protecting and Restoring Biodiversity. Island Press, Washington, D.C.

Odell, E.A., Theobald, D.M., Knight, R.L., 2003. A songbird's case for clustered housing developments. J. Am. Plan. Assoc. 69,1 -15.

Parkhurst, G.M., Shogren, J.F., 2003. Evaluating incentive mechanisms for conserving habitat. Nat. Resour. J. 43,1093.

Parkhurst, G.M., Shogren, J., Bastian, C., Kivi, P., Donner, J., Smith, R.B.W., 2002. Agglomeration bonus: an incentive mechanism to reunite fragmented habitat for biodiversity conservation. Ecol. Econ. 41 (2), 305-328.

Pearson, R.G., Dawson, T.P., 2005. Long-distance plant dispersal and habitat fragmentation: identifying conservation targets for spatial landscape planning under climate change. Biol. Conserv. 123 (3), 389-401.

Plantinga, A., 2015. Integrating economic land-use and biophysical models. Ann. Rev. Resour. Econ. 7, 233-249.

Polasky, S., Camm, J.D., Garber-Yonts, B., 2001. Selecting biological reserves cost-effectively: an application to terrestrial vertebrate conservation in Oregon. Land Econ. 77 (1), 68-78.

Polasky, S., Nelson, E., Lonsdorf, E., Fackler, P., Starfield, A., 2005. Conservation species in a working landscape: land use with biological and economic objectives. Ecol. Appl. 15 (4), 1387-1401.

Polasky, S., Nelson, E., Camm, J., Csuti, B., Fackler, P., Lonsdorf, E., White, D., Arthur, J., Garber-Yonts, B., Haight, R., Kagan, J., Montgomery, C., Starfield, A., Tobalske, C., 2008. Where to put things? Spatial land management to sustain biodiversity and economic production. Biol. Conserv. 141 (6), 1505-1524.

Porej, D., Micacchion, M., Hetherington, T.E., 2004. Core terrestrial habitat for conservation of local populations of salamanders and wood frogs in agricultural landscapes. Biol. Conserv. 120, 399-409.

Portney, Paul, 2004. Obligations of a policy economist. J. Agric. Resour. Econ. 33 (2), 159-161.

Pruetz, R., Standridge, N., 2008. What makes transfer of development rights work?: success factors from research and practice. J. Am. Plan. Assoc. 75 (1), 78-87.

Rhode Island Department of Environmental Management, 2005. Rhode Island watershed partnership [online].Available from. http://www.dem.ri.gov/programs/bpoladm/ stratpp/partnership/pawcatuc/index.html (assessed August 2005).

Rushman, M.J., 2000. Protecting the land: conservation easements past, present, and future. Landsc. Urban Plan. 52, 63-64.

Ryan, R.L., 2002. Preserving rural character in New England: local residents' perceptions of alternative residential development. Landsc. Urban Plan. 61,19-35.

Ryan, R.L., 2006. Comparing the attitudes of local residents, planners, and developers about preserving rural character in New England. Landsc. Urban Plan. 75, 5-22.

Semlitsch, R.D., 2007. Differentiating migration and dispersal processes for pond-breeding amphibians. J. Wildl. Manag. 72, 260-267.

Semlitsch, R.D., Bodie, J.R., 2003. Biological criteria for buffer zones around wetlands and riparian habitats for amphibians and reptiles. Conserv. Biol. 17,1219-1228.

Silicon Valley Taxpayers Association, Inc. v. Santa Clara County Open Space Authority, 2008. No. S 136468.

Skidds, D.E., Golet, F.C., 2005. Estimating hydroperiod suitability for breeding amphibians in southern Rhode Island seasonal forest ponds. Wetl. Ecol. Manag. 13, 349-366.

Skidds, D.E., Golet, F.C., Paton, P.W.C., Mitchell, J.C., 2007. Habitat correlates of reproductive effort in wood frogs and spotted salamanders in an urbanizing watershed. J. Herpetol. 41 (3), 439-450.

Smith, R.B.W., Shogren, J.F., 2002. Voluntary incentive design for endangered species protection. J. Environ. Econ. Manag. 43 (2), 169-187.

Stavins, R., 2003. Chapter 9 experience with market-based environmental policy instruments. In: Maler, K.G., Vincent, J.R (Eds.), Handbook of Environmental Economics. Vol. 1. Elsevier, Amsterdam, pp. 355-435.

Swallow, S.K., 1996a. Resource capital theory and ecosystem economics: developing non-renewable habitats with heterogeneous quality. South. Econ. J. 63 (1), 106-123.

Swallow, S.K., 1996b. Economic issues in ecosystem management: an introduction and overview. J. Agric. Resour. Econ. 25 (2), 83-100.

Swallow, S.K., Talukdar, P., Wear, D.N., 1997. Spatial and temporal specialization in forest ecosystem management under sole ownership. Am. J. Agric. Econ. 79 (2), 311-326.

Thorson, J.A., 1996. An examination of the monopoly zoning hypothesis. Land Econ. 72 (1), 43-55.

Watzold, F., Drechsler, M., 2005. Spatially uniform versus spatially heterogeneous compensation payments for biodiversity-enhancing land-use measures. Environ. Resour. Econ. 31 (1), 73-93.

Weitzman, M.L., 1974. Prices vs. Quantities. Rev. Econ. Stud. 41 (4), 477-491.

Wiens, J.A., Moss, M.R., Turner, M.G., Mladenoff, D.J., 2006. Foundation Papers in Landscape Ecology. Columbia University Press, New York.

Ye, L., Mandpe, S., Meyer, P., 2005. What is "smart growth?" really? J. Plan. Lit. 19(3), 301-315.