2004 Progress Report: A Consistent Framework for Valuation of Wetland Ecosystem Services Using Discrete Choice MethodsEPA Grant Number: R831598
Title: A Consistent Framework for Valuation of Wetland Ecosystem Services Using Discrete Choice Methods
Investigators: Milon, J. Walter , Scrogin, David , Weishampel, John F.
Institution: University of Central Florida
EPA Project Officer: Hahn, Intaek
Project Period: April 1, 2004 through March 31, 2006 (Extended to December 31, 2007)
Project Period Covered by this Report: April 1, 2004 through March 31, 2005
Project Amount: $313,797
RFA: Valuation for Environmental Policy (2003) RFA Text | Recipients Lists
Research Category: Economics and Decision Sciences
The overall goal of this research project is to develop and test a consistent framework to estimate wetland services values. The diverse nature of wetland services negates complete valuation through a single method or data source. Our approach uses a joint modeling strategy to integrate revealed preferences (RP) from a discrete choice model of a housing market and stated preferences (SP) from a choice model for wetland ecosystem services. There are four interrelated objectives to implement this strategy: (1) estimate the demand for proximity to wetlands and other water resources using discrete choice and hedonic pricing models of residential property values; (2) estimate the demand for ecosystem services from different types of wetlands that are not in proximity to residential property using a stated choice survey; (3) develop and test a combined discrete choice model from the RP and SP data to produce a general valuation function for wetland ecosystem services; and (4) estimate the implicit prices of wetland services in wetland mitigation banking markets.
Initial work focused on acquiring the most recent geographic information system (GIS) databases to develop comprehensive land cover/use maps for the study region and integrating the mapping data with residential property parcel data from County Property Appraiser offices in the three Central Florida counties: Orange (Orlando), Polk (Lakeland), and Volusia (Daytona Beach). Digital land cover/use maps were acquired from the two regional water management districts with jurisdiction in the respective counties (St. John’s River Water Management District for Orange and Volusia counties; Southwest Florida Water Management District for Polk county). These maps include data based on medium and low altitude flight imagery collected in 2000 and 2003 at scales of 1:24,000 and 1:6,000, resulting in an image resolution of 1 meter. The flight data were analyzed and interpreted into cover and land use types by the water management districts based on the Florida Land Use and Cover Classification System (FLUCCS). The FLUCCS classifies hundreds of land types and includes over 25 different types of wetlands. These data are more reliable and accurate than the National Wetland Inventory produced by the U.S. Fish and Wildlife Service because the water management districts use the most recent flight imagery and the images are interpreted by local specialists. The initial project schedule was delayed by several months, however, because the most recent digital maps were not available initially.
For this study, land uses were aggregated into eight major categories: developed (residential, commercial, and industrial), agriculture (cropland, pasture, and groves), rangeland (dry prairie and brushland), upland forests (coniferous and hardwood), water (lakes, rivers, and reservoirs), barren land, transportation and utilities, and wetlands. Freshwater wetlands were aggregated into four categories based on the dominant type of vegetation: wetland hardwood forests (loblolly bay, tupelo, and bottomland hardwoods), coniferous forests (cypress, pond pine, and cabbage palm), freshwater marshes (sawgrass, cattail, and other aquatic vegetation), and wet prairies (emergent and sparse vegetation). The first two categories are distinguished by tree cover and crown closure, whereas the latter categories are open habitats with short or no vegetation. These wetland categories are consistent with prior hedonic pricing studies of wetland values. An additional category for named lakes also was included.
For the hedonic housing choice analysis in Objective 1, the GIS land use maps were overlaid with parcel ownership data obtained from the County Property Appraiser’s most recent tax roll and additional GIS maps showing different types of natural amenities, such as parks and golf courses. Using spatial mapping tools available in ArcInfo 9.0, single family residential parcels with sales during the period January 2000 to May 2005 were geo-located in terms of Euclidean distances from the centroid of a parcel to the edge of the nearest wetland by type and to other natural amenities. A distance threshold of 1 kilometer (km) was used as the outer boundary for the effects of wetlands on property values based on prior studies in the literature.
To develop the stated choice survey, the first task was to select a representative sample of wetlands from the Central Florida landscape that could be used to measure willingness to pay to acquire wetlands with different types and levels of ecosystem services. This sampling approach relies upon the natural variability in landscape attributes for the choice set design and differs from previous research on landscape amenities in which artificially generated choice sets are created through factorial design methods. Initial efforts focused on determining the composition of wetlands in the study area based on the water management district FLUCCS data described above.
The first sampling approach considered was based on histograms that displayed wetland patch area and type in each county. This process produced a distribution that was skewed toward smaller wetlands within each type that were 1 hectare or less in size. Visual inspection of the land use maps revealed that these small patches were often located within a few meters of one another but separated by another wetland type or a complementary land cover (e.g., upland forest intermixed with wetland hardwoods). A sample from this distribution of wetlands would tend to show the landscape within each county as highly fragmented and dominated by small wetlands. This type of sample would be difficult to present in a stated choice survey because respondents would be asked for their willingness to pay to acquire mostly small tracts of wetlands and a limited set of wetland services.
An alternative wetland sampling approach was developed that sought to identify larger tracts of land that contained wetland habitats along with contiguous lands that were compatible with the wetland ecosystems. This approach used a random point generator in ArcInfo 9.0 to select 60 sample points within the entire area of each county in the study area. For each random point, a square buffer measuring 1,772 meters on each side was generated for each point. This produced an area equal to that of a circular buffer with a radius of 1 km. The total area represented by the sample square buffer areas reflected the overall distribution of wetlands within each county. But, this approach produced sample plots that varied greatly in land use composition on an individual basis. For example, some sample plots were composed of almost entirely of wetlands, whereas others were dominated by residential land uses. Given the focus of this study, the decision was made to resample the study area, this time focusing on areas of greater wetland density and excluding highly developed areas that contained little or no wetlands.
Through the use of GIS software, a wetland density map was generated for each of the three counties in the study area. These density maps were based on the county land use maps used in previous steps, but areas of high residential, commercial, and industrial land uses were removed along with large lakes. This process produced a refined study area from which a total of 60 random points was selected within each county. Then 1-km square buffer areas were generated again for each point. This produced land use distributions skewed toward areas of higher wetland density and more contiguous wetland areas that would be more amenable to the stated choice survey design. Each of the 60 square buffer sites were then joined with parcel ownership data from the County Property Appraiser’s most recent tax roll to show details within each buffer site for both the ownership data and the land use/wetlands composition. The parcel ownership was then viewed individually within each buffer site to exclude lands that already were developed or within well-known publicly-owned tracts (e.g. state parks and preserves). These data, in conjunction with the aerial photos for 2003, made it possible to determine if newly cleared lands indicated imminent development and therefore were not suitable for possible acquisition as part of the stated choice survey.
The remaining buffer sites then were evaluated using a cluster analysis procedure in Stata 9 to identify sets of similar wetland buffer sites. The analysis used a dissimilarity measure to determine buffer sites that contained similar groups of wetland and surrounding land uses. This analysis, combined with visual inspection of each buffer area using the 2003 aerial photos, made it possible to reduce the number of buffer sites within each county so that the sites would not duplicate attribute combinations in the stated choice survey. This approach is comparable to factorial design methods commonly used in stated choice experiment design to reduce the number of choice alternatives and improve statistical efficiency.
The research team is currently in the process of visiting the sites in the three counties to “ground truth” the FLUCCS and aerial photo information and obtaining property owner’s permission to visit sites where no public access is available. The team also is working with wetlands specialists in regional and state environmental management agencies to evaluate each site and to score the site in terms of four wetland ecosystem services: wildlife habitat, groundwater recharge, hydrologic connectivity, and flood control. These ecosystem services scores for each site will be used as part of the stated choice survey design.
The major focus for the next period will be to conduct focus groups to determine the appropriate visual representation of the GIS and other spatial information to be used in the stated choice survey. The results from the focus groups will guide the development of the final questionnaire to be implemented in the stated choice interviews. The schedule for these focus groups has been delayed from the original project schedule because of the initial delays in acquiring the most recent land use/cover GIS maps and the problems encountered in determining an appropriate wetlands sampling methodology for the stated choice survey. This delay in initiating the focus groups also will delay the field testing and implementation of the stated choice survey. We anticipate that the focus groups will begin in August, which will delay the start of the survey until some time during the fall, depending upon the survey design issues that arise during the focus groups.
Work also will continue on econometric estimation of the housing choice models for the three counties. Both hedonic and discrete choice models will be estimated to evaluate the relative differences in the implicit values for different types of wetlands that result from the alternative modeling approaches. After the completion of the stated choice surveys, these housing choice models will be combined with the stated choice data to develop a joint revealed and stated preference model that will provide a general valuation function for wetland ecosystem services as described in Objective 3.
In addition, contacts have been made with wetland mitigation bank managers in the Central Florida region to obtain data directly describing the characteristics of the wetlands within the mitigation bank and mitigation credit sales prices. Once a full set of data is acquired, the hedonic model for wetland characteristics will be developed as described in Objective 4.