Regulatory requirements for mitigation trades require regulators to determine how much restoration or compensating preservation is enough to offset permitted wetland losses. Surveys of wetland mitigation banking practice show that bank program administrators rely on relatively vague, biological function-based compensation ratios. The most common regulatory practice is simply to require an "acre for an acre" of biophysically similar wetland when another is destroyed. Biophysical equivalence is evaluated; however, acre-based or purely functional compensation evaluations fail to account for many of the aspects that determine the social value of a particular ecosystem such as a site?s location in the greater landscape, the importance of local substitutes for and complements to the site, and future risks to the site?s ability to provide services.
Regulators do not often evaluate the economic benefits that arise from lost ecosystem functions, most likely because they are not explicitly required by law. Nothing in the law prevents such analysis, and social benefits analysis is recommended. Trades should preserve what is valuable about ecosystems. Trades will be driven largely by the costs of land acquisition without benefits analysis. This raises the concern that restoration and preservation will occur in relatively remote areas far from direct enjoyment by large numbers of people. This is potentially harmful if wetlands, for example, are providing significant flood protection, water quality improvement, and aesthetic benefits in more densely populated areas. Accordingly, evaluation of ecosystem exchanges should include more than sufficient ecological analysis. Exchanges should include the analysis of services generated by those ecosystems. Unfortunately, regulators are often financially or technically ill-equipped to assess the exchange of the relative social value of environmental assets. For instance, econometric analysis, the economist?s preferred evaluation method, is difficult, costly, and typically does not capture the full range of service benefits at any given site. In practice, econometric analysis is rarely used in wetland-permitting decisions.
Without economic analysis, ecosystem trades may undermine, rather than advance, the achievement of environmental and social welfare objectives. This research project proposes and demonstrates a non-econometric method of ecosystem service site evaluation. The method is a middle ground between no analysis of services and econometric analysis, which is not practical for small-scale permitting applications. The main objective of this research project was to develop an evaluation method, applicable by non-economists, using existing data sources that can identify likely differences in ecosystems? social value. The method is based on geographical information systems (GIS)-based landscape indicators of ecosystem service benefits. Economic precision and the cost and complexity of econometric valuation methods are compromised by a system of simple indicators.
The conceptual foundation for the system is the notion that ecosystems generate socially valuable services. Examples of such services include provision of habitat for rare or endangered species, reduced flood risk, improved drinking water quality, and recreation. This research project began by reviewing federal programs that involve ecosystem trade or compensation. Most regulatory programs rarely or insufficiently evaluate ecosystem service benefits, though there are exceptions. Accordingly, there is a need for better ecosystem benefit estimation tools; these tools need to complement the biophysical assessments that typically take place.
The focus of this research project was the effect of landscape characteristics on ecosystem service benefits. The determinants of ecosystem service benefits are a function of both onsite biophysical characteristics (e.g., the ecological quality of the site) and a function of the landscape context in which sites reside. Both ecological science and economics increasingly stress the importance of spatial analysis. This reflects the understanding that ecosystem functions are related to the surrounding ecological landscape, and that the social enjoyment of ecosystem services is related to the surrounding social and economic landscape. This research project related specific and observable landscape characteristics to the valuation of services.
The centerpiece of this research project was an application of the benefit-indicators approach. Using a wetland mitigation bank in Florida with available GIS data for the region of interest, we evaluated wetland trades associated with the bank. The majority of GIS data used in the case study were acquired from the Florida Geographic Data Library, a repository of more than 200 spatial data layers for Florida. The South Florida Water Management District provided additional data. Our analysis is based on 40 GIS coverages containing demographic, real estate, physical, biological, land use, infrastructure, and planning data. We calculated 66 indicators from these coverages. These indicators were related to the provision of service benefits.
Our analysis focused on wetland services such as the: (1) status of the drinking water supply; (2) prevention of flood damage; (3) improvement of aquatic recreation; and (4) provision of open space, aesthetic, and existence benefits. For each of these services, we sought indicators motivated by the following valuation categories: primary demand, scarcity and substitutability, complementary goods, risk from changed future conditions, income, and equity. In total, we derived 20 different sets of indicators, a set for each of the 5 valuation categories and each of the 4 services. Each of the indicator sets was applied to the bank site and its associated impact (wetland loss) sites to assess the extent to which service benefits were lost or gained by the transfer of wetland acres to the bank.
We concluded that the wetland trades would have benefited from an analysis such as ours. The bank site scores poorly in terms of its ability to provide certain services. In contrast, several of the impact sites remain valuable generators of service benefits. Based on landscape analysis, the bank site's main advantage is its support for species' benefits and recreational benefits related to the support of aquatic species. Those benefits may outweigh the bank's poor performance as a source of other benefits.
In general, the Florida case study demonstrates the practicality of a GIS-based indicator system to characterize sites in terms of ecosystem service benefits. The indicators can be derived from public data suitable for GIS analysis. Although we manipulated data in a variety of ways, the assessment required no original data collection. This was an important aspect of the research plan, because one objective was to propose a relatively affordable, user-friendly evaluation tool. We carefully chronicle the steps necessary to establish an indicator database and apply it to site evaluations. An important conclusion of the study is that public GIS data is rich and varied enough to support benefit-related landscape characterization. Based on the case study, we conclude that GIS landscape indicators can easily be derived and applied to the analysis of ecosystem value. It is an evaluation tool that is practical and has significantly improved, relative to existing regulatory evaluation procedures that tend to ignore service-related benefits.
The research project also included a more critical appraisal of indicator-based evaluation methods. Viewed in isolation, a given indicator is useful information that can almost always inform decisions. Indicators also are useful as inputs to more rigorous valuation methodologies such as benefit-transfer studies or so-called "multi-attribute utility analyses"; however, indicators should not be oversold. Regulators and planners are often under pressure to "generate a number," in this case, a compensation or trading ratio. Landscape indicators are a potentially desirable source of material for such calculations. In addition, they are an organized and empirical approach to evaluation, but indicators are not an ultimate end.
Potentially, an aggregate index could serve as an important input to trading-ratio calculations. However, when a larger set of indicators is generated, numerous questions arise regarding which indicators are most relevant and how the information in different indicators should be weighted. For this reason, indicator tools must address the methods by which decisionmakers weigh noncomparable indicator rankings. Indicators, whether evaluated individually or aggregated into indices, can introduce a false formality that obscures fundamental weaknesses in indicator-based methods.
A particular weakness of indicator-based methods is that they do not easily allow for tradeoff analysis. Without a common metric, such as dollars, it is impossible to determine whether a site scoring highly on one measure is better or worse than a site scoring highly on another measure. The ideal approach for comparing the value of different services is to monetize benefits associated with each service. There are alternative methods that assess tradeoffs and can be used in conjunction with indicator-based methods. One example of an alternative method would be a multi-attribute utility analysis. The project report includes discussion of these issues and relates our indicator method to more conventional econometric assessment techniques, such as benefit-transfer methods.
There also is a set of important questions related to the scaling of indicators. Scaling issues relate to the "shape" of the underlying relationship between an indicator variable and the benefit it describes. Unless the relationship between an indicator and the benefit is linear, indicator interpretation requires subtlety. For example, the difference between indicator values may not be significant if the effect being measured is only noticeable above a threshold. A precise relationship between an indicator and benefits is typically not known, though a general relationship may be well established.
Following the Florida analysis, we undertook an additional case study of wetland trades in Maryland. This case study is similar to the Florida study in that we assessed ecosystem service benefits by mapping landscape characteristics. The Maryland study differs, however, by being a statewide analysis of long-term trends in the spatial pattern of wetland losses and gains. Also, we experimented with a "service zone" approach to evaluation. The objective of service zones is to map conditions that favor the provision of ecosystem services at a broad landscape scale. These maps can be used by planners and ecosystem trade evaluators to quickly assess the desirability of mitigation activities and the social costs of wetland losses. Zones can be mapped statewide, and a specific site can be assessed by determining the zones in which it resides. This is a more economical decision tool, because a statewide analysis can be conducted collectively, rather than site-by-site. The zones simplify the ecosystem service analysis by immediately identifying areas where particular services can be eliminated from consideration (e.g., because they do not meet conditions necessary for the existence of the service). Also, the presence of overlapping zones quickly identifies areas where landscape conditions reinforce the existence and scale of service benefits.
We examined data that included most of the non-tidal wetland impacts that occurred in Maryland between 1990 and 2000, and a subset of the offsite mitigation wetlands developed by permit-seekers to compensate impacts. Wetland impacts are compensated through three types of mechanisms in Maryland. Offsite mitigation occurs in approximately a quarter of all permits, but is used for the largest wetland impacts. With these data, we addressed several questions: (1) Were patterns evident in the demographic setting of sites created and destroyed by inference in the types of services lost? (2) What kinds of changes in wetland quality might be inferred from patterns of loss? and (3) Were trends over time discernable?
The mitigation database contains 46 offsite non-tidal wetland creation or restoration projects, and represents the mitigation sites for which we were able to obtain location data. The acreage mitigated at these sites represents more than 15 percent of the mitigation acreage (onsite and offsite) required by permits in our wetland impact database. Although the two databases are not directly comparable, it is useful to examine the spatial distribution of sites in both datasets.
Our analysis of the wetland impacts showed a spatial correspondence between impacts and high-growth areas. One quarter of the impacts and a third of the impact acreage are within census-defined urban areas, and 30 percent of acreage is within an 800-foot buffer of major roads. The mitigation data show a more limited spatial distribution for mitigations compared to impacts. Most mitigation sites are clustered near or between the major cities, and a third of the sites fall within urban areas. We saw a large difference in median population density between impact and mitigation sites (741 versus 320 persons per square mile) when we examined data by the number of permits. Also, the dominant land use around sites appeared to be residential for impacts and agricultural for offsite mitigation.
In terms of the potential functional quality of wetlands lost and gained, 29 percent of the impacted acreage was within 800 feet of roads, but only 4 percent of the offsite mitigation acreage was similarly situated. This result indicated that offsite mitigation sites may be better situated than impact sites for certain ecological functions, such as the provision of high-quality wildlife habitat, assuming that wetlands develop successfully. An unexpected 32 percent of the mitigation acreage occurred on public land, which reduced risk to functions and correlated with higher quality functional settings. However, these sites were developed largely to mitigate impacts on public land, yielding no net benefit, and perhaps some potential loss of functional quality.
No clear time trends are apparent in either the size of compensation fund impacts or population density of the census block group in which the wetland impact falls. If the impact size of permits using the compensation fund increased through time, this could be an indication that mitigation sites are becoming scarce. Permittees are required to find onsite or offsite mitigation sites when impacts are above a certain size; but instead, they may contribute to the compensation fund if mitigation sites are no longer available.
In the Maryland case study, we also examined specific impact-mitigation site pairs, where data linking an impact site to its mitigation site were available. For these 25 sites, we assessed a specific ecosystem service and the provision of visual amenities, via an application of the service zone method. We also identified five variables thought to be important to the provision of visual amenity benefits. These variables were used to generate a map of zones representing the relative accessibility of areas for aesthetic enjoyment. The state's visual amenity service zones map is created by combining a data layer representing each indicator into a single map, using a GIS overlay operation. Maps of the individual indicators are overlaid and summed to create zones representing the number of value indicators that fall within each zone. Using the zones and other analyses, mitigation-impact pairs were evaluated in terms of differences in demand, access, and scarcity.
When we compared the zone scores for the impact-mitigation site pairs, we found that more than one-half of the "trades" (14 sites) resulted in a lower level of visual access. Seven of the sites had a matching accessibility score, and three mitigation sites had higher accessibility scores. Thus, our results suggest that some visual access was lost as a result of this set of mitigations.
The zones are not intended to provide the same level of detail as the suite of indicators developed in the Florida case study. Rather, the zones are intended to be constructed from a limited set of variables to allow users to screen for the occurrence of particular services and to consider whether the level of service is likely to be high or low. Certain variables or groups of variables are used to represent relative levels of service value. For example, the proximity of densely populated residential areas near a wetland site can indicate easy access for a variety of uses. Screening for this variable alone may be sufficient in some cases to tentatively assign a high or low value to a service.
We estimated that approximately one-fourth of impacted wetland acreage is mitigated offsite. However, our available offsite mitigation data analysis does not support the idea that Maryland's wetland mitigation program results in a major change in landscape setting and service value. In the case of offsite mitigation, some differences in landscape characteristics are evident when localized zones around sites are examined. Also, it is important to note that we did not include an evaluation of compensation fund mitigation sites, which do not require that the mitigation sites be near the impacts. Accordingly, these sites may generate a significant shift in the wetlands' spatial context. Nevertheless, because current regulation of the mitigation program is structured largely to preserve ecological function, the fact that we found only a modest amount of evidence of a systematic change in landscape and social characteristics near wetlands suggests that established controls in Maryland are largely effective at preserving a range of wetland services.
Outline of the Final Project Report: Relating Project Objectives to Outputs. Sections 1 through 4 of the Final Project Report are devoted to the policy background and conceptual motivation for the research project. In particular, we reviewed the role of benefit assessment in federal ecosystem compensation and trade programs. Also, to place our method in a larger context, we have provided an overview of biophysical and economic ecosystem assessment techniques.
Objective 1: This objective corresponds to Tasks 1 and 2 of the project: construction of an inventory of biophysical, geographic, and demographic indicators; and derivation of a system of ecosystem value indicators (EVI).
Output: We derived the inventory of indicators from both a broad review of the science of ecosystem service provision, and from the data available to us in the case studied. Section 5 of the Final Project Report presents a broad overview of ecosystem service provision, and in particular, reviews both biophysical and economic analysis of services. An emphasis is placed in this review on the spatial (landscape) determinants of benefits. Sections 6 and 8 are the Florida and Maryland case studies, respectively. In these cases studies, specific indicators are derived from existing sources of spatial data and as a function of the specific ecosystem services we analyzed. For these indicators to be defensible, they must be related to the science of ecosystem assessment. We are careful to relate specific indicators to their scientific justification in the case studies.
Objective 2: This objective corresponds to Tasks 5 and 6 of the project: integrating the value indicators approach with existing governmental data sources; and field-testing the ranking system via retrospective application of ecosystem benefit indicators to existing trading, banking, and acquisition programs.
Output: We conducted a retrospective evaluation of wetland trades in Lee County, Florida, and in Maryland using the benefit indicators approach. These case studies are the central elements of the research project, and are described in detail in Sections 6 and 8 of the Final Project Report. They demonstrate that the indicator concept is practical, beneficial to ecosystem assessment, and relatively affordable. The indicators rely only on public data, knowledge of GIS techniques, and an appreciation of the underlying economic principles motivating the choice of indicators.
Objective 3: An important task of the project was to assess the weaknesses of ecosystem value indicator systems relative to other valuation methods, and to propose methods by which the application of benefit indicators can be disciplined methodologically. This objective was associated with Tasks 3 and 4 of the project: linking the value indicators approach to established ecosystem valuation methods, such as contingent valuation and revealed preference methods; and evaluating methods for eliciting "relative preferences."
Output: Section 7 of the Final Project Report is devoted to a critique of indicator-based evaluation methods. Monetized benefit estimates are the appropriate aspiration for evaluation. Monetary estimates of benefits are the best means to analyze tradeoffs within programs and across programs because benefits are expressed in dollars, a common metric. This point should not be lost, even while we advocate-for pragmatic reasons-that alternatives to monetization, such as ours, be developed. Also, we devote attention to problems with the interpretation of indicators (nonlinearities and nonconvexities in the underlying function being described) and their inability to resolve fundamental tradeoffs when one site does not strictly dominate another across services. Finally, we address the way in which our indicator system can be used profitably in conjunction with monetization techniques, such as benefit-transfer, and techniques for eliciting and weighting social preferences, such as conjoint analysis and multi-attribute utility analysis.
Objective 4: Do the strengths of indicator-based ecosystem valuation methods outweigh their weaknesses? This objective was pursued in reference to a more specific question, corresponding to Task 7 of the project, which addressed the question: Can an EVI system be effectively applied to determine units of exchange applicable to real-world habitat conservation policies?
Output: This is reflected in the ultimate conclusions of our analysis and throughout the report: as long as the weaknesses of indicator-based methods are appreciated (Section 7), the benefits of the method are significant. The Florida and Maryland case studies show that landscape indicators can provide a more complete understanding of the portfolio of changes associated with an ecosystem bank program. This is desirable for compensation programs and regional planning, where it is also important to understand broad patterns of landscape change. Certainly, the landscape analysis yields a richer description of the sites than functional assessments limited to onsite characteristics. First, the analysis fosters an appreciation of the way in which onsite functions are related to the biophysical characteristics of the larger landscape such as watershed hydrology, floodplain characteristics, and exotic species communities. Second, landscape analysis highlights the human dimension of the surrounding environment. Ecosystem services can be described only if we have data from both the physical and social environments. Third, the analysis is effective in revealing extremely good and extremely poor landscape scenarios.
Landscape-related ecosystem service benefits should matter to regulators and land use planners. If the social value of ecosystems-not just acreage-is to be preserved, then the sites' relative ability to generate benefits must be explored. The argument illustrated in this study is that a visual and numerical presentation of landscape characteristics fosters an intuitive, nontechnical appreciation of these benefits. Accordingly, a better understanding of benefits should be an input to the determination of ecosystem trade and compensation ratios. Determining trading and compensation ratios currently is considered more of an art than a science. Calculation of appropriate ratios remains an art when implementing with a landscape indicator analysis; however, it is an art based on more complete information regarding the sites' ability to provide ecosystem service benefits.