The URL can be used to link to this page

Home My WebLink About04 April 2012 packetHEATHY RIRS AND STRES CITIZENS ADVISORY BOAR Church at Redstone Redstone, CO AprIl1Q. 2012 - 4 p.m. I 4:00 Public Comment 4:05 Board Comment 4:15 Approval of Minutes February 16, 2012 meeting 4:20 Financial Assistance for On site Carla Ostberg- Wastewater Treatment System Pit Co Env Health Replacement Dept Update of Watershed Plan and Water Sharon Clarke - Conservation Report Roaring Fork Update on Watershed Plan Brand Conservancy Mark Fuller - RWPA New Watershed Gauging Project Tim Flynn Chelsea Congdon Brundige Request for Proposals on study Virginia Newton describing the relation of the real estate market in Pitkin County to the Roaring Fork River City of Aspen's Storm Water Program April Long City of Aspen Storm Water Manger Upcoming 2012 reaular meeting dates May 17 June 21 July 19 HEATHY RIRS AND STRS CITIZENS ADVISORY BOAR Meeting Minutes February 16, 2012 Aspen, CO Board members present:Ruthie Brown, Greg Poschman, Lisa Tasker, Bil Jochems, and Steve Hunter Board members absent:Rick Neiley and Andre Wile Public Comment - None John Ely, Lisa MacDonald, Tim Flynn, Greg Espegren and Lee Rozaklis Others present: Board Comment - None Approval of Minutes Ms. Tasker moved to approve the minutes motion. The motion passed 5 to o. . Hunter seconded the Pitki COUl Stream Health Meth Rozaklis, Hydrologist AMEC Environmen Mr. Espegren and Mr. Rozaklis presented Discussions ensued and several embers of Executive Session Ms. Brown moved to e the purpose of discussin instreamjlows. ask pursu t to C.R.S. § 24-6-402 (4)(b)for .ect and discussion of drought relieffrom . Motion passed 5 to o. CORE Lette the Colorado Wa The Board receive (CORE) to the ColoradMr. Hunter moved t motion. The motion p rd tter 0 port from the Community Offce for Resource Effciency n Board for a water conservation education and outreach grant request. letter of support to CORE. Mr. Jochems seconded the Adiourn The meeting adjourned at approximately at 6:45 pm. Approved:Attest: Greg Poschman - Chairman Healthy Rivers and Streams Board Lisa MacDonald Minutes -Healthy Rivers and Streams Citizens Advisory Board February 16, 2012 Page 1 AGENDA ITEM SUMMAY AprIl19, 2012 TO:River Board FROM:Carla Ostberg, Pitkin County Environmental Health Department SUBJECT: Financial Assistance for Onsite Wastewater Treatment System (OWTS) Replacement Introduction: This summary is intended to generally inform the Board of the effect of failng OWTSs, commonly referred to as septic systems, on the environment and public health, as well as to alert the Board of a current situation having the potential to negatively impact the environment and public health in the Capitol Creek drainage. "Wastewater", also known as sewage, is defined in the Pitkin County OWTS Regulation as a combination of liquid wastes that may include chemicals, household wastes, human excreta, animal excreta, other animal or vegetable matter in suspension or solution, and that is discharged from any building, dwellng, or other establishment. Pitkin County has approximately 3000 parcels that fall outside of a sanitation district, relying on OWTSs to treat their wastewater. Malfunctioning OWTSs contribute to non- point source pollution and negatively impact water quality. Non-point source pollution differs from point source pollution in that the pollution is not coming from a single source, such as a factory. Non-point source pollution comes from the cumulative effect of day to day activities including road maintenance, fertilzer application, and use of OWTSs. The impact of non-point source pollution is often diffcult quantify in the same way pollution from a point source can be identified and measured. In the last several years, the Pitkin County Environmental Health Department has seen an increase in OWTS permits issued to address failures. Aging systems, improper maintenance, and increased water use are all factors that are likely contributing to OWTS failure rates in Pitkin County. However, failing OWTSs are not just a problem in Pitkin County. Nationwide, contamination from septic systems and sewage overfows is an issue of great concern. American Rivers, an organization founded in 1973 to protect and restore the nation's rivers and streams, posts the following information regarding the implications of sewage contamination on the environment and public health: "Sewage pollutes our waters with pathogens, excess nutrients, heavy metals, and other toxins. It kils aquatic life and creates algal blooms that can suffocate fisheries. Even worse, sewage carries pathogens that can end up in our drinking water supplies and swimming areas. These disease-causing microorganisms cause diarrhea, vomiting, respiratory, and other infections, hepatitis, dysentery, and other diseases. Common ilnesses r: April 19, 2012 LJ Request for owrs funding caused by swimming in and drinking untreated or partially treated sewage include gastroenteritis, but sewage is also linked to long term, chronic ilnesses such as cancer, heart disease, and arthritis." (ww.americanrivers.org) In February, I was contacted by the owners of a residence located at 7030 E. Sopris Creek Road. The OWTS servng the residence on the propert, permitted and installed in 1999, had failed and wastewater was coming to the ground surface. Capitol Creek is located approximately 125 feet downhil from the lowest portion of the failed absorption area. There are also several private drinking water wells in close proximity to the OWTS. The cause of the failure is diffcult to pinpoint as the system is relatively new, water use is conservative, and maintenance of the system has been consistent; however, the receiving soils are clay, meaning that it takes longer for any water or wastewater introduced to the soil to absorb. The owners immediately addressed the problem by replacing the failed system with appropriate technologies for the site and soil conditions. I understand that maintaining and improving water quality and quantity within the Roaring Fork watershed is one of several objectives of the Healthy Rivers and Streams fund was established to attain. I believe that providing loans to homeowners needing to replace and repair malfunctioning OWTSs aligns well with that objective. There are a number of other states and counties that have established loan programs specifically to address OWTS repair and replacement. A loan program of this tye would benefit Pitkin County residents while facilitating greater protection of our water resources. Project Cost: $25,000 Requested Board Action: Approve funding a loan, based on the County Attorney's approval of the loan agreement, for the property owners of 7030 E. Sopris Creek Road to offset the cost of replacing the OWTS servng this residence. Exhibit A: Google Earth location of propert Exhibit B: Photographs of failure after excavation Attachments: Exhibit B This photograph shows excavation of the existing absorption ofthe subject property. The house shown is located on an adjacent property. Capitol Creek run behind the house shown here, approximately 125' from the failed absorption area. This photograph illustrates the saturated absorption area, no longer accepting the effluent, or wastewater, causing it to surface. In this case, effluent first surfaced on the downhill side of the absorption area, toward the neighboring house, pictured above, and Capitol Creek. To: Pitkin County Rivers Board From: Mark Fuller, Ruedi Water and Power Authority and Sharon Clarke, Roaring Fork Conservancy Date: April 12, 2012 The April 12 meeting of the Roaring Fork Watershed Collaborative featured the formal unveiling of the Roaring Fork Watershed Plan, the Water C'onservation Report, and a Watershed Plan Brand. Pitkin County Rivers Board contributed to both the Plan and Water Conservation Report. On April 19th we want to thank you for your support, present an overview of the documents, and provide an opportunity for you to ask questions. Both of these documents are attached and can be found at www.roaringfork.org/watershedplan. The Plan is the culmination of a five-year process that began with the compilation of the Roaring Fork State of the Watershed Report in 2008 and included the development of several supplementary documents, including two Phase ii Guidance Documents, an Implementation Guide, a Front-Range Water Supply Update and the Water Conservation Report. The Plan is the product of dozens of meetings and the efforts of hundreds of local citizens and government and agency representatives who worked together to develop the Plan through a series of public meetings and workshops. The Ruedi Water and Power Authority is the sponsor of the Plan; Roaring Fork Conservancy, the lead consultant; and Rose Ann Sullivan, Kootenay Resources, LLC assisted with work on all aspects of the Plan. The Plan contains over 200 recommendations for managing the watershed in the future. The recommendations are sortable by topic, by geographic area, and by responsible parties. The Plan is subdivided into five major topic areas. These include Groundwater, Surface Water, Riparian and Instream Habitat, Regional Water Management, and Water Quality. The recommendations are further subdivided into projects, legislation and regulation, and studies and range from specific and urgent (i.e. Plan and Implement Key Riparian Restoration and Protection Projects) to more general and long-term (i.e. Review and Revise Master Plans to Address the Impacts of Climate Change). The Plan's recommendations are intended to provide a guide and a series of goals for local governments, water management agencies, land managers and non-governmental organizations that can be implemented as opportunities arise. This is a comprehensive and holistic document that should help everyone involved in water, from large government agencies to individual users, to address the needs of the watershed in a systematic way. We recognize that implementation of this Plan will be an incremental process and RWAPA and the Roaring Fork Conservancy are committed to fostering that implementation. In the future, we will be working to identify appropriate projects, sources of funding, partnerships and government support as opportunities arise. The Watershed Plan was funded by contributions from local governments and two $40,000 grants from the Colorado Water Conservation Board. The total cost of the Planning Process, including the State of the Watershed Report and the supplemental documents, was around $250,000. The Plan has been 1 presented in draft form to the City and County governing boards in the Valley as well as the Colorado Division of Parks and Wildlife and several local Water and Sanitation District Boards. The Authority and the Conservancy, with the help of the Schwener Design Group, also completed a process to identify a "brand" for the Roaring Fork Watershed Plan. The purpose of branding is to build awareness by giving the Watershed Plan its own graphic identity that can be used on everything from letterhead to project signage. This brand will allow us to label implementation actions as Watershed Plan projects and not solely as projects of the Roaring Fork Conservancy, the Ruedi Water and Power Authority, or other agencies. Puttng a "face" on the Plan will provide a recognizable symbol and an image that can be the focus of activities ranging from political action to fundraising to communications. The Water Conservation Report entitled Opportunities for Water Conservation: Realizina the Streamflow Benefits from Local Water Conservation Efforts was completed by G. Moss Driscoll, Elk Mountain Consulting. The report investigates the potential strategies for employing water conservation to benefit streamflows and offers ten recommendations for a Water Conservation Campaign in the Roaring Fork Watershed. The research and recommendations presented in this report arose from the desire among local interests in the Roaring Fork Watershed to understand how water conservation efforts could be used to improve local streamflows. The report has its origins in a commonly heard local question: "Why hasn't Roaring Fork Conservancy engaged in a water conservation campaign to improve streamflows?" Generally, the reason has been because in Colorado water conservation raises a host of complicated legal issues, such as abandonment, waste, and potential injury to other water users. Yet such legal barriers do not change the fact that conservation efforts are likely to prove essential to ensuring adequate streamflows in the Roaring Fork Watershed. 2 TO: The Board of the Healthy Rivers and Streams Fund; John Ely From: Virginia Newton Date: 4/16/12 RE: RFP Draft for Real Estate Economics and the Roaring Fork River Dear All, Thank you for allowing me to get involved with the question "What does the river have to do with our real estate market?" Real estate is not the first thing that comes to mind when protecting and conserving water. River flow rates, aquatic and riparian flora and fauna, kayaking, fishing and rafting have all been a part of the conversation in river protection communities. They are important and will continue to be. However, a more complete picture of the impact of healthy rivers and streams to our area will also include the real estate market. Pitkin County is unique in many ways, and people pay extraordinary prices to own property here. In 2011, (a down year) real estate sales totaled 1.27 billion dollars. Is there a relationship between our rivers and streams and this remarkable economic engine? If so, to what degree? If the relationship is clear and significant, how might that information be used to benefit the goals ofthe Healthy Rivers and Streams Board? At the April 19, 2012 board meeting, we can discuss the viability of an RFP, a draft which is attached. Many thanks, Virginia Newton vnewtonC§sopris.net Reauest for Proposals To provide for a study describing the relation of the real estate market in Pitkin County to the Roaring Fork River. Introduction With approximately twenty-six billion in actual real estate values, and over one billion a year in real estate transactions, Pitkin County's real estate market drives much of the economic success of the Roaring Fork Valley and related communities. While local studies have quantified the economic benefits related primarily to tourism and fishing, little work has been done to quantify and qualify the relation the Roaring Fork River has to real estate values and sales. Currently, forty percent of the natural river flow of the Roaring Fork River is diverted at the headwaters to the Eastern Slope of Colorado for municipal and agricultural use. Several studies, including the 2010 Interbasin Compact Committee report to the Governor of Colorado, indicate that increased needs for water for the Front Range communities of Colorado will require the consideration of further diversions from headwater counties such as Pitkin County. Climate change factors that cause average winter snowpack levels to vary would further impact natural stream flow volumes of the Roaring Fork River. Scope of Work Pitkin County is seeking qualified individuals or firms to conduct an analysis of the relationship between real estate values and the Roaring Fork River. The study will include research to support the relationship between healthy river systems and real estate values such as the following: . Analyze the real estate sector in Pitkin County and the economic impacts of the over one billion dollars in annual transactions that occur in the transfer of real estate, as well as the economic impacts of the approximately twenty-six billion in real estate value residing in Pitkin County. . Review the literature of previous economic impact studies of rivers in Pitkin County and describe the gap in these studies related to the real estate sector. . Develop a statistically valid method for quantitatively illustrating the relationship between the Roaring Fork River and real estate values. . Develop a persuasive method for qualitatively showing the relationship between rivers, river health and real estate values. . Describe potential future diversions of native river flow from the Roaring Fork River headwaters. . Describe the impacts that future diversions would have on riparian and aquatic health management in the Roaring Fork River. . Develop a conclusion to ilustrate that reduced flow rate and compromised riparian and aquatic health of the Roaring Fork River would have a negative impact on the real estate economy, including dollar value estimates for this impact. Other communities with denuded or restored river environments may be examined and contrasted to the rivers in Pitkin County. eo J 996 by 5 E.L. & A~~dat1"5 Estimating the Benefits of Urban Stream Restoration Using the Hedonic Price Method Carol F. Streiner John B. Loomis Dt'p,~r: Itt'rit of Agricu/tunil ami Resuiira EiinamriC$ Colorado State Llnh1ersity Fort Collll::, Colorado 80523 USA ABSTRACT: The hedonic prict" m~thod \\'i1S used to estimate residents' wil- ingness to pa,' for improvementit in urban streilms. This stud". examined C\li- fomia's Department of Water Resources Urban Stream Rt't'tor~tion Progrilm to determine the economic value of stream restoration measures such as reducing flood damage and improving fish habitat. Seven projects from three counti~s- Con~ra Costa, Santa Cruz, and Solano-were pooled for analysis. Property prices in areas \'\'ith restored strt'am~ were found to increase by 54,500 to $19.000 due to sti\bilzing stl'eambanks and acquiring land for educåtion trails. This repre- sents rrom 3 to 13'i at" the mean property prici' in th(~ study. Recommt'ndiltions fur facilitating further analysis are made and implication~ e¡) quantifying tht' benefits of similar programs in other slates are pro\'ided. KE)' WOllDS: ~onmarkèl \'illudtion. property v",lues. water quality. willing. ness to pay. INTRODUCTION Urban streams can be assets to a com- Hons such as concrete channels w('re typo muiiity. yet flooding, erosion, ically constructed to stabilize streambanks streambank instabilities, and ba5ic envi- and protect properties. mnmental degradation along urban creek Currently, (he trend is toward restora- channels pose significant problems to tion of the n.1tural attributes of the stream landowners adjacent to these creeks. Prob- as \.vell as structural improvements. Some lems with urban streams arise from a structures that incorporate re\'~getation ar"" change in n,1tural stream corridor charac- wood crib walls, live fascines (which are teristics. often brought about by urbaniza- bundles of live cuttings anchored into the Hon itself. Some properties and buildings streambdnk) and other li\'e cuttings, and are endangered by such processes, and de- brush matting or brush layering. All are spite individuals' efforts to remedy the ef. designed to slow the velocity of the stream feets. problems can recur orcan cause dam- and prevent bank instability. These meth- age farther downstream from the site. ods as \\o'e11 as others, such as check walls, Flooding, erosion. and streambank in- ruck ur stune walls, and wood plank walls, stability also deteriorate the natural values are recommended (or aesthetic and func- of streams. Historically, structural solu- tion,11 purposes alike. Many characteristic~ Riv~rs · Volume 5, Number 4 Pages 267 -27R 267 I~..... of a natural ~tream are restored using such methods. but benefits provided by these more natural methods h,we not been mea- sured. URBAN STREA~i RESTORATION PROGRAMS For some communities in California, lo- cal and state agencies have realized the need for a program to mitigate the adverse affects of flooding, erosion. and bank in. stability. California's Department of Water Resuurces (DWR) initiated the Urban Stream Restoration Program in J 985 to as- sist communities in reducing damages from Hooding .1nd stll~arnbank instability as well as to restore environmental and aesthetic values of urban stream channels. By en- couraging the involvement of local ~gen- cìc!' and citizens' groups. both paid and volunt('cr, thc DWR hopes 10 promote stewardship and main (cnM\(' of strl'ams bv communit\' members.. .An urban stream restoration program can result in a varietv of bcnef,ls. both 10 in. dividual property owners c'nd to commu- nities as a ,vhole. These can be grouped into two broad categories: reduction in proptCrty damages. and restor.ition of the natural values of the stream itseli. Damages to adjacent properties can be reduced by mitigating the effects of fluod- i ng. ('rosion, and stream ban k i nst.ibilty through various restoration me,lsures. The benefits accruing to property owners are intact yards, minimal damages to trees. structures. and Icindsç.iping, and healthier streamside parks. Benefits of returning a stream to its more natural state are more- stable stream banks, rl?stor.ition of tht"' riffe.pool sequences en- hancing fish habitat and other aquatic hab. itats. and a more aesthetically pleasing eco- system with riparian \,("geiation restored and wildlife habitats protected. These ben- efits are identifiable yet diffcult to quan- tify in dollar terms comparable to the cost. ~ieasuring homeowners' value of stJch benefits is important in documenting the economic contribution of an urban stream restoration program and allowing compar- ison to the costs of such r1 program (L1.s. Water Resources Council 1983). Estimating residents' willingness to p,iy (\\TP) for improvements in urban streams can be' clccomplished by using the hedonic price method (HPlv1). Hedonic pricing uses residential property value differl'ntials to measure changes in \VTP fur environmen- tal amenities in two stages. The slupe of the first stage hedonic equc1tion measures how the value of the property changes ,...ith a small change in the level of an attribute. The second stage hedonic measures the de- mand for larger changes in the levels of that attribute. allowing calculation of the residents' WIP for that change. In this studv, we used the first stage hedonic func- tion -to meêls.ure a discrete' value resulting from performing a speciA( stream resto- ration measure. We hypothesized that properties along streams that have been restored will exhibit nigher property \'al- ues than areas that h.we not been restored. if buyers perceive restored streams as an amenity. Price differentials between the two areas can be stcltistically related to stream improvements via multiple regres- sion analysis. BASICS OF THE HEDONIC PRICE ~lETHOD (HPM) The theory behind the HPM, as ex- plained by Palmquist (19lJ 1), lies in diftl:r- entiated consumer products_ Houses that are single commodities difier in envii'ona mental attributes at their location. Con- sumers select properties fur the number clnd quality of charılcterisiics that c1re pres- ent at the site. Housing price differentials, therefure, reflect differences in housing characteristics (Freeman 1993). Rivers. Volume 5, \lumber 4 October J 995 Ii~-i 268 The basic hedonic property model is pre. sented by i:reeman (1993) as the price ot' a property as a function ot' its structural. neighborhood. and environmental char- acteristics, or: P, = f (5,. \l,. (,,)( 1) "there P, 5, price of propert~. i :; structural characteristics of i neighborhood chardctl'ristics of i environmental qualitv char- acteristics of i - Both Freeman (1993) and Palmquist (1991) agree that the above equation should be Jlonlinear based on the fact that in the housing market, "repackaging" of prop- crty charactt:ristics is unlikely. That is to say, individual characteristics of each property cannot be varied independenllv. In a particul.- property's "bundle," a (on- sumer cannot trade t\'...o rooms for better air qualiy. Thus, consumers must choose a "bundle" that best meets their needs. The marginal implicil price of a characteristic is the 'part~al deriv,itive of the first stage hedoJlc price function in equation (1) with res~e~t to a marginal change in Qi or "the additlOnal amount that must be paid by any household to move to a bundll' with a higher level of thaI characteristic, all oth- er things being equal" (Freeman 1993). A nonlinear hedonic function vields a mar- ginal implicit price for a char~cteristjç that depends on the level of that particular at- tribute and on the level of other charac- teri~lics a~ well. ln the second stage he- donic, estimates of residents' \.VTP for dif. ferent le\'e15 of each attribute can be ob- tained. :-, Qj Specifying the Hedonic Pricing l..fodel The dependent variable of a hedonic price function is the full price of house and land, which is regressed .igainst the ex- ~ected determinants of property price. Freeman (1993) maintains that data on ac- tual market transactions are preferred, but if professional assessors' values are used, care !l\U5t be taken to assure that they ap- proximate actual sales value. Dat.i for the explanatory variables can be obt.iined through county tax-assessors' records ilnd census data. As noted above, both freem.1n (1993) and Palmquist (l99t) suggest that a nonlinear functional form is appropriate. Freem.1l and Palmquist agree that a Box-Cox Trans- formation works '..ell in selecting ihe ap- propriate functional form. In general, the Box-Cox Transformation takes the form (Johnston 1984): Y\H, = a" + ßX(~2) + u (2) Depending on the estill1altd values for ,\ and À~, the best functional form can be de. termined. For example, if ". = Ài = 0, tlw form would b(' a log-log modeL. Hedon.ic Price Analysis for Urban WatN Issues Early studies using the hedonic price ap- proach have focused on air quality iSSU~5, yet the method is equ.il1y applicable to wa- ler3~ality issues: Dornbusch and ßarrager (19/.)) used multiple regression analysis toestimate the benefit of \',;ater poiiution abatement.. They concluded that property values of single family residences on wa- terfront lots increased from 8 to 2S'¡r with \VClter pollution control. In 1980, Feen berg and Mills also wn- ducted a study to measuri~ the benefits of \."raler pollu!ion abatement using property data. They lound that the determinMlts of demand were water quality at the nea1''st beach and number of people per dwelling. The authors conduded th.:t the ,...illing- ne5S to pay for slightly cleaner water rises rapidly as water becomes dirtier. To determine the impact of degraded water quality on the value of seasonal res- idential properties, Young and Teti (1984) studied the shordine properties of St. Al~ bans' Bay in Vermont, Th('ir main obiec- tÍ.ves WNe 10 use a hedonic mode) to pro- vide a mt:asure of water quality's inllu- ences OIl property values and to estimatl' the benefits from water quality improve- ments. Young and Teti concluded that the largest impact of waler pollution in Sl. 1'\ 1- bans' Bay affected residents and recrea- tionists. The benefits of wat('l quality im- provements, therefore, would be higher property values and enhanced recreation. as well as improved wildlife habitat aru! ~n~ironmental aesthetics. Gnl' significant Ulsight from this study is thêlt property val- ue data reflect only those benefits to prop- erty owners, When evaluating the bendits of '~,;ater quality improvements, it is critical to include other potential benefits as '..ell. A more recent paper dealing with urban ,-vater management problems is a study by Kriesel et a1. (1993) of the benefits of shur.to erosion protection in Ohio's Lake Erie housing mıll'ket. The purpuse of their study "..'as to measurt: the discount of erosiOJi- prone lakeshore properties using hedonic C. F. Streiner and J. B. Loomis 269 II~._ price analysis. They point out that deter- mining the benefits of erosion protection is ùiffcult because private and social ben- efits differ and market information is lack- ing. Their objectives were to determine how erosion and protection devices affect property prices and to calculate the bene- .£ts ùf such measures. The authors con- cluded that an average erosion control de- vice lasting 8 years would raise property val ue bv $5,500 from an initial time of 20 "ears to setback (\'ears until shoreline property is eroded ~p to the house). while a device lasting 20 years would add $1 i ,000 to propert)' value. These benefits .iccruc to private property owners. Other social ben- efits are not mentioned; thus. total benefits of erosion protection may be understated. DATA ANALYSTS cated within 1.100 ft of the creek and 45 properties greater than 1.100 ft from the creek. Total sample size included 521 prop- erties for the funded projects and 478 for the unfunded projects. Property data were obtained from the respective county asses- sor's offce. These data list the lot number, type of residence, sale date, sale .imount, assessed value. lot size. improvement size, number of rooms, bedrooms. and baths. and existence of a garage. An)! information missing from the reports was researched by the DWR. Groups of propertics were traced to the census maps to determine the census tract and the census block. This matched each propert)' ,vith its respective demographic characteristics. Organi:;ations that received funding from the O\.VR submitted a survev to the department detailing the complet~d proj- ect. Groups that failed to receive funding. and consequently did not carry out the proposed project, were interviewed by phone to determine their planned objec- tives, goals. and projected costs. All data came from property sales be- tween 1983 and 1993. Sales prices and as- sessed values were adjusted by fixed ,veights from the U.s. residential price in- dex found in the Survey of Currcnt Busi- ness (U.S. Dept. of Commerce 1995). Ide- allv, a California or San l=rancisco Bav area sp~cilÏC price index would have bee~ used but we could not locate one. Percent changes in the price iiidex were used to convert propert)' values to a base year, 1982. for comparison over time without the in- fluence of inflation or a char.icteristic in- crease in property prices in California's housing market. To correctly account for changes in stream attributes, properties that were sold before the restoration projects began were coded with zeroes for the restoration mea- Creating the Data Set Data for estimating the economic bene- fits of California's Urban Stream Restora- tion Program consist of property transac- tions. property characteristics, stream proj- ect characteristics, and demographics of the residents living in the area. The D\'VR com- piled the data, beginning with the pairing of unfunded and funùed projects accord- ing to similar locations, demographics. and project characteristics. A total of 7 project pairs were pooled for analysis. Initially 12 projects were selected to reflect a geo- graphic mix throughout California, and to represent urban. suburban, and rural stream restoration projects. Funded proj- ects similar in location were paired with unfunded projects in an attempt to control for location specific elements that might be diffcult to quantify in a regression. This \vas done because we were not certain if \\o'e could pool the sample across projects in different locations due to the possibility that they might have different regression coeffcients. Unfortunately, several proj- ects had. to be dropped because data on sales transactions or characteristics of the unfunded project were not .waila.ble. Two pairs were from Santa Cruz County, four from Contra Costa (near the San Fran- cisco Bav area) and one from Solano Coun- tv in no'rthern California. The streams in- ':olved in the funded projects averaged a flow of about 500 cis during storms and ranged from 2.000 to 3.000 ds during peak winter flO\'\s. Each pair of projects con- tained an average of 80 properties adjacent to a funded project and 70 properties near an unfunded proposed project. The major- ity of the properties v".ere single family res- idences (see Streiner 1995 for more details). Each project. whether funded or un- funded. contained at least 50 properties 10- I~J 270 October 1995 IRivers. Volume 5. Number 4- TA8LE 1 C orrt'alioii mat rIXt5 for reshmltroPl pad:agt's. fish Acquire Ed ucatio ri h.ibitat l(lnd trail Restoration package A Fish habitat 1.0000 0.8273 0.8273 Acquire land 0.527.3 1.0000 1.0000 Education trail 0.S273 1.0000 1.0000 Restoration package B Stilbiliz~Redfldam Cle inup Clrobs Rt!veg Aesthetics Stabilize 1.0000 0.6383 0.6362 0.ï481 0.6996 0.6026 Reduce flood damage 0.6383 1.0000 0.6ï43 0.8332 0.6805 0.5019 Clean up 0.6562 0.6ï-l3 1. 0000 0.Si71 0.6996 0.7665 Clear ob5tnl(tion~O.i 481 0.8532 0.8i71 1.0000 0.7977 0.6591 Rt'vegl:tah'0.6996 0.6805 0.6996 0.79iï 1. 0000 0.9) 28 Al'i-thetics 0.6026 0.50 I 9 0.7665 0.6591 0.9128 1.0000 sures. Of the seven funded projects in the sample, 0111 but on(' were initiated in sum- mer 1989. On average, restoration projects were completed in 1~.i years. Estimating the Hedunic Equation As shown in equation (1), a hedonic equation is specified as a function of struc- tural. neighborhood, and environmental variables. We chose variables that repre- sented l'ach of the three c,11egories. In our data set, many of these variables within each category ,,,'ere currelated with each other. Therefore, the first step in variable selection is to conduct an analysis for mul- ticollinearity among the candidate e),plan- awry variables. A correlation matrix ,.vas calculated using Econometric Views (Lil- ien et al. 1994). Many of the stream char- acteristic variables are highly correlated with each other, having correlation cuef- fici('nts greater than 0.80. In addition. many of the property and demographic charac- teristics are highly correlated. To avoid high sampling variances and low J-statis- tks. variables must be chosen to minimize the effect of multicollinearity. The second step was to conduct regres- sion analyses on groups of independent variables to calculate partial R2's. Results from this test should indicate ,vhich vari- ables in each of the three groups (property, stream, and demographic characteristics) are most influenced bv the others. Inde- pendent variables ,\'íth low partial R2's within the thrt'e groups of projects are pre- ferred to minimize multicollnearity. The stream meaSures were grouped into "po1ckages'" according to each variable's correlation with the others. The packages and currelations are shown in T.1ble I. A correlation coeffcient of one indicated that the two ...ariables ,'.rere ahl.ra)'s carried uiit in the restoration projects together and, therefore, were exactly the sam(' in the re- gression, and, in fact, influence each oth('r. For example, because both "acquire land" and "establish an education trail" variables were needed together; the correlation be- tween those measures is one. Based on the review of the literature and the partial R~'s, a subset of the variables in each packag(' was selected as candidate variables that were most likcIy to deter- mine the value owners place on a property. Omitting a stream restoration variable that is correlated with an included restoratiun variable may lead to mis-specifícation and bias in the included ..ariables. We recog- nize this problem and consider the esti- mat('d coeffcient on the included stream restoration variable to renect the joint ef- fect of aU the stream restoration variables that it is correlated with in its package. The robustness of using different stream res- toration \o'ariables from the same package in the hedonic regression is then test('d by comparing the resulting sign, size, and sta- tistical significance of the cueffcients in Tabl.: 2 and the resulting marginal values in Table 3. I c. F. Strciner and J. B. Loomis 271 II~ -r~l- IV",.,TABLE 2 R(lx.Cin IIOH/iilllir TI'.'lI''"..Ùm /lodti.. I\"""n of Ass"",.d Prop..rty Prki' ~ SI4.l,08S Joint Modd Edtrai!StabiliLi'Acaland Fi5h,ib Redßd,im ~i Vdri,lhlt'.1 Codf.1-st,it Cot'L '-st,ll Coefi.I-stoll Codi.'-5Ial CoefL I-,I.il Coefi.1-,IOiL lmp~iz..0,0104 6.9.1 0.0106 .. 18 0,0104 4.33 0.0114 4.21 0.0112 4.24 0.0104 4..11 ;¡.lotsiii'0.0002 7.22 0.0002 3.57 0.0002 3.69 0.0002 :1.59 0.0002 :1.62 0.OOLL2 3.67~Yl!ëJr 0.2504 1.57 0.4434 2.53 0.11l55 1.26 0.3902 2.22 0.3782 2.22 0.2527 1.72:;G.lf.i¡;,'2.0.147 2.80 2.1222 2.51 2.1556 2.70 2.0530 2.32 2.1594 2.47 2.61166 3.04 Ckdist -0.On06 -2.0.1 - 0.0007 - 1.92 -0.0005 -1.56 -LL.OO06 - 1.6t1 0.0006 1.70 O.OOOS 1.6:1iPrindl90.0010 8.iii 0.0010 4.14 0.0009 4.27 0.0011 4.14 0.001 i 4.J8 0.00 io .1.28 i:lra"i.ii -0.4419 -6.24 0.4671 -3.51\0.5246 3.87 0.5367 3.7 0.5323 3.71 0.5052 3.79 3 Mi'an,iiti'0.2280 3.64 0.2544 2.96 0.2014 2.74 0.2595 2.87 0.2509 2.18 0.2264 2.'l1 It Un..mpri 0.2138 3.22 -0.2621 -3..11 -0.2242 -.1.11 -0.2707 -.1 22 -0.2591 -.1.20 -02133 - 2.'l1, ,'"Edir.iiJ 7.1364 2.17 'l5'l90 2.79 7.SI.ibiliLl'3.018.1 ,iOO 2,(,617 2.59i:Acaland 10.596 2.1l330-Fish.ill 8.1l092 V!H It R..dOdam 4.6474 2.48...i Con..t.1lH -';.1'401 -0.40 - 21.1183 -1.42 3.1560 0.24 15.062 n.'l7 - 14.307 0.%5.0554 0.:'9 1.1 inbd,i 0.7169 IS.H I 0.7230 15.86 0.7047 16.07 0.7375 16.27 0.7309 1/t.27 0.70'17 In.IS ,\' =1.055 1.055 1,077 1,077 1.077 1,077 1\2 _0.553 0.54'1 0.534 0.540 0.539 0.537 o,. 5'0-'l.. · wh,'rt' Imp,iz.. IS hOllo;.. siii', ""Isi"i' is pril','rtr siii', Yi'ar is ihi' vi'ar in which It". pni¡wrt\' ""is s"ld, Caraiii' is pri'Sl'iicl' of a ¡tara!tl', l1..d"i is di,I.1OCI' fr"m Crt...k. Pcindl'l is llLr ,.ipil~ in.-"m,' iii I 'lR'l. TrOiv,'11 is tr,i",') tin", In ",,,rk in mil1uti's, I\lt'olnage iio 11.. 1l1',ii ,Iiti' of the popul"!'''" in th.. sam I'll', Un,'mprt is iI", ul1,'mpIOVII\l111 r.-h', St.ibili".. is ,iabili"" stl''.iiii hani... Edltail ¡, ",I.ibli.h ,10 ~ducnlÍlIl Ir"iI. ,\iql.ind is ,lCquir,' I.ind.Fish.ll, i. fish hoibit.it. R,'dßd.i1i i.. ,,'uun' n""d d,imil!tl' ,ind Lnmbd.i is Ih,' ""pon"ni to whiih ih.. di'pend..nt vari"bl.. b r.ibed -0-0,"' The depe ndc n 1 vtliab Ie in 1 he regi'l's- sions is the a~~es~ed valult of the properties at the time of sale. It is not significantl~. diffcr~nt from the actual sale \'alut'. This is not surprising bec,lusi' with ('.lliforni.i's Pri1positic,n i 3, a~sessed propi.'rty vahiC's ,i ri' updated to the new Ildrkt.t prict. at ih(' time of sale. A Box-em, tr.instormatlL1n was employed to dctl.'rminc the functional fnrm that fits thi' data and the \!.iiabli's that Of'st explained the viHiation in the Jepltndl'IH variable (the assessed value in real terms). To conduce multiple ll~glt'ssion analysis. the question of functional form \\lS ,1d- dressed using the Box-Cox transformation method (Greene 1992). This method allo\..'~ tlw tr.3nsformatinn nf both the dependent and ini.kpciid('lH v,iri,1bl('s. Th~ Box.Cox transiormation indi"lll'd a nonlini',ll iunco tioned iiirm to tit the dat¡¡. This Ilh"'del was used in the multiple re~fl'ssi(1n ¡:lalysis. The Box-Cox transformation method al- !ü\\s Iransform,llÎons of both tlw Idt.hand- side and right-hand-sidt:' variables in ,1 n~- gression equation. depending on thl. na- lUl'(" of thtl data and the int~raction of the independent and depi'ndent vilri.:bles. Preliminary regressions were dtt~mpted using all four pussible BI)X-COX specifica- tions: imnsformation of the dependent variable by I.imbda, tl,, indepi'ndt~nt ,..an- abies by lambd", both independl.'nt i1J1d de- pende~t vari.ibles by lambda, and trans- fiirm i ng t hl.' depi~nden t v.uiabli: by i hl..i and the indepenJ~nt vdlÌilbles by ldmbdd. Using LlMDEP (Greem.' 19n), the model chosen for this data set, tl'ansforms cinly the dependt'nt v,iriabli': assi'sst'd propl"ty v,ilui'. Basl"d on Ill(' minimum v,ilul.' oi thl' lng-likelihood vclhiL." prelimin.iry l('gl'es- ~il.ins indicated thi:- transfl.irniatiiiJl Wd~ b~st. lambd.i represents the exptment hi \vhich tlw di~pendent variable is raised. TIlt modd l~ nimlinl.ilr, a~ sugge~ii.d by 111l hedonic price theory reviewed pn!\'iously. \Vhite's general test for heteruskedastic- ity was perrormi'd on the fin.,! n'glt'ssiOll modeL. l\ posiii\'~ result indicatl.s l,rgi' v.iri"ncl!s. \Vt'ightt!d It!at.t :-quMeb (ould hI. used to curri'ct for this, but there w~re sev- eral possiblt' variables that might bE' ap- propriaii' as ihi' w('ighl. HOWl"Vi'l', only 011( vari,lble can be USE'd a:= the wt'ighl. rhus, to avoid this pnibIL'm, we corn'cted for het- eroskedasticity by using \yhitt"s Hett'ros- ked.istidty-Consislt'nt St¡indard Errors ,md CovaJ'anc(' (Lilii'n et a!. 1(94). which r('- duces the stant.liru error:: .md n1cikes thl.' t-statistics mort' accurate. Tabh' 2 lt'ports the I.statistics afrt' the stand,ird i'rrors ,1ll' corrl'cti'd. R~'5 fur l'(ich oj the rl'gression~ were ca1cubted by ein urdiii.lY least squares cstim.it1on procedure using the transfor- mation t)f the independent vari.ibli' b~. i ,1m hd.1. SI.itistic,11 Results STATISTICAL AND PROPl:::ny VALUE RESULTS Tht' final regression equation cont;iins three property v.iriables that have minimal cOlTi'lations among l'.lch oill('r, ihri'l' neighborhood or lotatioIldl \'dridDJes, thri'e demographic variables, .md t\\o stream im- provement variables, one from l'ilch r~s- toration "pa(k,lgc." This can be loul1i. in the first column of Table 2. Stream rt'sto- ratiun measures in Tables i and 2 are sim- ply dummy variables for whether or not th,1t mc.1SIJr(' WilS pi'rformi'd in ilw proW'1. ror all the regression results in Table 2, improvement ~iLe. lot siLl'. ~aragl', per cap- ita income, travl' time to work, mean age. unemployment rate, and lambda are sig- nifkal1eat k',lSI.lt ihe 5';; Ii.'\'el. Codf;Ch'nts on the nonstrl?am vari.ibh.'s rL'main sl,ibk I c. F. Strt'iner ,lOd J. B. Loom¡; ~ i'vi'n whl?n di(fi'rl'ni stn'l1m v.iri,ibli.':' ,iri' used in the r('grl'ssion. Cri!(.i- distance i:- signific.int at least at tht: (0'.; le\'eL The stahilitv or restoration nwasures is indi- (,llld hv 11"- col.ficÍCnls on cduc¡ilion trail. stahihæ 5trl'ambanks, fish habitat, acquiJ't.' land, and reduL't' flood deirnagt', \',..hidi are each significant at tht.' 5'.., level in their respect ivf' n~g rt'SSÍt111S. \ViWrt~.lS the rear of sale is nol signj fie.ml ,It t hl' I 0':7, i('v~l, i I has a f-statistit" uf grt'Hter them une ilnd i~ included in the model to ,iccount for po~- ~ibii' i nCJ"~,ises in real propi~rty prices over thi. time frcime of ihe study. r"blt'l shi,\\~ rl'gression cOt'tticienls fOJ th~ moueb with two mea::ure!: that MC groupt.'d togl'th('r iii il restoration pilckagi'. Both "reduce flood damag('" and "stabiliLe strl'i1nibanks" ,lrt' by thl'msl.'lv('s positive i7J -i I~_I ~4.. and significant. As shown iii Table 2., "sta- bilize stream banks" and "education trail" are both significant when induded togeth- er. Both "reduce flood damage" and "ed- ucation traiL" however, were not signifi- çant when induded together in the modeL. Property Value Results The individual values for the stream im- provement measures can be calculated from the coeffcients estimated by the modeL. The method evaluates the change in propE.ty price due to the measure itself, holding all other variables constant at zero. The fol- 100...ing equation from Greene (1992) for the Box-Cox nonlinear regression model was used to calculate the individual values. y = l(X..8X) + iil"~ (3) As an example of this calculation. the value of the fish habitat variable is set equ.il to one. This indicates the complt:tion of thai measure. If the other variables are held constant at zero, the equation becomes: Change in property price = (0.730948.8092 + i¡i ,01:100 The result is 15.571, or $15,571 change per property in the arc,) where the restoration project improved fish habitilt. When both an education trail and bank stabilization are carried out together in the projects, property values increase $19,078 over properties without these measures. These restoration measures include other aspects that were generally carried out iii the sample projects, such as clearing ob- structions, revegetating, acquiring land, and making other improvements. Vêll ucs for the restoration i'packages" that contain stream restoration measures that can be grouped together according to each variable's correlation with each other art~ presented in Table 3. From Packilge A holding all other variables at zero, estab- lishing an education trail is perceived by buyers to contribute $17,560 to the price of a property. When compared to the mean property value of the sample, this makes up 12% of the property value. Maintaining fish habitat was perceived to be worth 515,571, and acquiring land was \Talued at 519,123 per property. Similar values (each I~,. .. 274~"- TABU: 3 Va/iit of resloraliori p(irkages Absolute amount r'er- cent of pnip- erty value Restoration packagü A" Fish habitat Acquire land Education trail Restoration package Ba Stabilize Reduce flood dam¡igc loint model Package A (Edtrail) Package B (Stabilize) Total $15,571 II $19,123 13 $17,560 12 $ 4.48H 3 $ 7.804 5 Si19.07R 13 a ~OTE: The individual measures c,mnot be added together becau&e they are simply alter- n,itive measures of the joint effect of all tht!sc variables in the package. is within S2,000 to $4,000 of each other) and their high correlations with each other indicate that these measures are not in- dependent of each other but, in fact, ap- pear to measure the joint effect. Establish- ing an education trail and acquiring land are perfectly correlated. This follows be. cause, to establish a trail, land is needed along the riparian area. The end result is that the value for an education trail cal- culated irom the s.imple of projects also includes the value of acquiring land and maintaining fish habitat. Both "stabilize" and "reduce flood dam- age" enter into the regression equation positively and their values are also listed in Table 3. Stabilize has a value of $4,1188 per property and reduce flood damage con- tributes $ï,804 or 5% of property value. Individually. it appears as though stabiliz- ing streambanks does not add as much to property value as does reducing flood dam- age and the measures in Package A. The values retlect buyers' perceptions at the time of purchase; perhilps education trails and less damaged banks are more visually important to home buyers than the tech- nical details of bank stabilization. Reducing flood damage is highly cor- related with other strcilll measures, such Rivers . Volume 5, Number 4 October 1995 ~ I ~'1 ~. ~'1..~Co'; T MILE 4 Liii'a. '1','1 'ii ;1 n',/I) 'I'M,lm' ItllI r",,,r,.~,,,m" ~ I ê.1'r ~It'an of ,'\~~t'~~l.d l'ropèrty lriù..I 44,OSS Educ.lIion tr.iiJ !'sh habit.it V..i..hlt.CPt' if i-Raiio ~1l~dn n( \eiiett.r-R~iiii M..an 01 X---_.- Imp"....0010,..I.IIN 1..1517 :i O.OIIJ 4.31 10455" I.'lsiil.00002 3.5;'11 7A713 0.0002 3. H 7.518.7 Yt'.ir 0...1).1 i.53tl 8887 0,3810 2.220 88,'11 C.iri.~L':!.1:!2:!2.512 O.ti:!ï 2.131,1 2.-21\0.1115 C~di,t -O.OOO~- i.'I20 i.SI\.1 -O.OOOb -1.70 i. S l. ('cineS'!0.00111 4.145 1~.3i!!0.0011 4.173 17A31 Tr.1\l't (lA117 3.5S1 :!g.27H 0.532 3.6%2li.2ì Mi..in~gi'0.2~.¡.2.%~3312'1 0,25..11 2.tI!!/.33.1,,2 L:Il..inprt -0.2,,21 -:UI7 51)'1 0.2,,(l!l 3.IS"5.'12 Tr,lilfi'i'l 0038):!.7'JIi b.li- Fi,hh','i 003112 2,iQQ 7315 D.1in.iHt'lt'l'l Ciin~t.int 2 i.Stl 1..2 -1456 -0.%5 L.imbd.i 0.7230 ¡5.Sbl 0.7314 1/i.2!111 ;' =i ,055 LOii R~ =0.5.1'1 0,5.in--- I j' i I I i 1 I !yj- ~~I ~edue,' ßood d.im.'gi' e""if. i-Ratio J\e.in ,ii .\ 0,0 ii( 00002 0.3027 2A21ot. - 0.000/1 0.001 i 0.512 0.247'1 0.221l5 00)1\1 -9.19 0.nii3 1.077 0.541 4.2Ml 3.M3 1.8'10 2.1., - 1.633 4.2111 3.6'10 2.'101 2.952 3.113 -o.Me; Ib.3J6 1.455,'1 7,518,7 !l8."1 0".3'; 1.1'1,3 170431 :!~.37 33.162 5."2 11...4 as stabilizing streambanks, clearing ob- structions, and revegetating the riparian area. The value obtained for reduced flood damo1ge, therefore, cannot be entirely at- tributed to that individual measure. It is important to remember that the val- ues for each measure do nut take intu ac- count the costs of performing that mea- sure. The benefit differential among mea- sures may be offset by a differential in the costs of restoration. In other words. decid- ing to acquire land solely because it pro- duces the greatest gross benefit is not an adequate reason. The higher cosl of ac- quiring land may far outweigh the benefits of the measures. thus reducing the net ben- efit (benefit-cost) of the measure. Maximizing net benefits, therefore, would be a better objective in deciding which restoration measures to carry out in a project. The benefits must be weighed against the costs. For instance, the follow- ing hypothetical example ilustrates the proper analysis; The benefit of acquiring land has been determined to be S19,123 per property. The benefit of stabilizing stream banks is $4.488. If costs of acquiring land, however, are S18,000 per parcel of land in a project, and it costs only S400 to stabilize the same area of streambank, the net benefit of acquiring land is onlv $1,123 while the net benefit of stabilizi~g is 54,088. Performing bank stabilization and estab- lishing an education trail together yield a total value of 519,078, or 13% of mean prop- erty value. It must be kept in mind that because each variable in a package contains the influence of the other variables, only one variable from each package can be cho- sen to calculate total value. In addition, individual measures from A cannot be add- ed to individual measures from B to de- termine the total value because of the non- linear functional form of the hedonic equa- tion. Continuous Measures of Rcstoration Acti 'lities Where the previous values have been discrete measures, values for certain res- toration activities can be represented in an alternative fashion, in terms of linear feet restored. Regression results are reported in Table 4. Changes in the value of property with different amounts of restoration, mea- sured in linear feet, are determined for es- tablishing an education trail, improving £ish habitat, and reducing flood damage. The values for a change in property value as the linear feet of an education trail along a creek are expanded are valid for resto- ration between one and 250 linear ft. These values are based on the: r.-nge of our data and are approximately 51.000 to 517,560 per property. Note this value of 517,560 is the: same value determined using the dum- my variable approach for establishing an education traiL. The property value changes with in- creasing linear feet of fish habitat main- tained are relevant from one to 250 linear ft of restoration. The valucs for improving fish habitat range from $1,000 for one lin- ear ft to 515,000 for the mean amount oi restoration (225 linear ft). Again, these val- UE'S are based on the range of the data in the sample. Changes in property values from $1,000 for one linear it to $11,350 for 175 linear ft of reduced flood damage are determined from the modeL. These values represent a continuous measure of the val- ue of restoration activities. CONCLUSIOl" The hedonic pricing method proved ap- plicable to measuring the benefits of se- lected urban stream measures. The Box-Cox nonlinear regression model pro\'ided an equation for which the coeffcients of stream restoration variables could be esti- mated. From the regression coeffcients, property value changes were calculated and the value of restoration measures deter- mined. These increases in property values were attributed to specific stream restora- tion measures, yet the high correlation among measures indicates that generally more th.-n one measure is reflected in the value of anyone individu.i1 measure. Fur measures such as establishing an ed- ucation traiL. maintaining fish habitat, and acquiring land and/or easements alung a stream, the one time increase in property value ranges from about 51 ;,570 to $19.120 i~ 276 October 1995Rívers · Volume 5. r\umber 4 per single filmily rcsidenc('. ror stabilizing streilmbanks (which includes de.uting ob- structions, revegetating streamban k!i, and cleaning up the stream) and reducing flood damage, property volJU('S increase about S4,480 to Sï,800 per single family resi- dt'l1(('. The-51. val ues are specific tu our sam- ple of projects. which reflects the San Fran- cisco Bay area and Santa Cruz. It may not be appropriate to gener.:lizl,' these ..-alut:s to other geograph Ie .:rl'as. Howl'ver, the bilSk method would be applicable. To alleviate the problem of multicollin- earity ëlmong project measures, it would be useful to have information ll1 iht: amount of restoration dOlii'. such as tne number of Iin('ar f~et. Three measures uf this type- linear feet of fish habitat maintained, ed- ucation trail established, and flood dameigl' reduced-produc('d changes in property prin~s fnmi 51.000 to S1i,560, depending on ihe number of linear feet restored. It is important for analysis to havl,' all projects repurt the amount of rl,'SIOl;itii'H ci)mplet- ed for valid Cèilcuhition of ..alul:s. Another recommendation for alll.'vicit- ing the correlation among slr('am \'ari.ibles is to defini' me.isures more specifically. If one objectiv(' can bl' accomplbhed by per- forming another, thtm the objective is too broadly defined and increases the difficuJ- ty of computing a !ieparati' \'ciJut ior each ml:asure. The benefit of these increases in prup- l'ft\' values also bentdits communites as a wl;ole.ln Califurnia, using the Proposition J 3 tax r.lti' of i .25': of propt'riy value, an increa!ie in property value of 51 Y.Oì8 would provide about 52-10 pl:r house in additional property ta.' to thl: cummunity annually. \Vhen .iddet.i up over the large number i)f single family homes in the fund~d areas, the pn~senl valut.' of the added tax money OVl'r the life of the restoration praji'ct is likely to contribute far morl' !"evenu(' than the program costs (which in our study has a rnedi,in veilui' of $),t920). Based on our r~search, the basic hedonic property approach appears tn be usdul for evaluating the benefits of a \\'idi. variety of urban stream l"t'i'tofaiion prúgrain~. ."cknowJ~dgments Tlianks an.' in order to Sara Den.lIcr a!id Earle Cumming,; cif tht' (.ilifornia Dep.lrtllt.l\t of W,iicr Rt.súurrcs for the ..xt..n~i\'l. dai,i collection ¡ind guidi.llCt... I'~ofl!~s.or Stt'vE' Davil.S was instrumental in assisting in tht., implt'mf'mation of I he llo\-Cox (unclion.iJ form. Funding fDr chi!. !ôtudy was pro- \'ided. in pari, by the Colorado Agricultural Ex- pt'rime-ni St,llion .1nl! by tht.. Cilifiiriia Dt'part- mt'nt of Watt'r Hl.sources. R EfEREXÇE5 Dornbu!'ch, (J 1\1., .md 5 M B,lfragc~. 1973 Benefit ()f wdh..r pollution c(in~roi 01\ propt.'rty valut!l'. Co;itract 1'0. 6R-Ol-0753 Project UIAAb-07. Washington. DC: U.S. E¡wiruninL'ntilJ I'rotcr~jiin Agt'rlcy. OffCE' nf Rt'~t',1TCh ilnd ~.ionilOrjng fe~iib~rg. D.. ilJld E. S. ~HIIs. 19RO. :\ property villue 5ludv. Ml'rlSIIlJI,~' tiii' ~tlidib ot Wrllt'r ¡'aIlIH;.'!: Abilll't)lt.I/. Nt.'w Yurk, ~Y: Acadi'mic PrE'~~. Frel.nl,ln, A. M. III. 1993. Propt.'rty value miidl'!.. Pagt's 3(,7-420 ir T!lt ,'.i,iisiirnt¡nr/ ii,t ëii T'm1mtJL"I/al ,¡tlli RI':''IIr(I' l'rilu,'s. fultiiiorc. MD: Johns H,'pkins Univl.r';ity I'rl'sli. GrE't'nt'. ~\". H 1 't'n L1MDEf' V~l5iOii 6.0. L"S(..r'S. Mill1lill ,ind Rdcrt'llCI! Guide. BE'lIpnrt. ~Y: ECI)foml'tri, SofIW,lfl'. Johnston. J. 19í\4. EI.,tt,I/lit'lr,. Akl!liiiiS, 3rd editiun ~c'\. Y,iT\. McCr.iw-Hill. Krj~sci. W.. A. RandalL, Llid F. LIC"lit koppl;?r i 'j'J3. F~..timaling the bl'ndits (\f ~liore crosiün pwt~i:tion iii Ohju's Lilkl' Erie housing m¡irk..t. l\.it,'1 J.,.S,.IIH,CS 1\t5tliti-ri ;:9(4):795-801. Lilien. l) ~1. R. E. HdJl, ,ind uthl'rs. 199~. ITin.,,, U~f''.~ GWdl', \'t'r"ion 1.0 In'in;?, Ct\; Qii.\nILt,ltivl' Micro SOIlW,lTl'. Pal mqiiisl, R B i 99 i. Hcdunjc inclllOds. P,igt..s. 77-120 m J. ll. ßradt'n and C. D. Kolstad, pdiwrfo Ali"1511Iili.i lilL' ¡.l('tlii..d (¡" flll'lnl!/tllt'I/lll Qlwlil.lI. N(irth-I rulland; Elsi'vit.'r Snt'nce PublishE'!\. Stri!irllr. C. S. 199';. Fscim.iiiig rht' ht'n""tiis 01 the urb.1l1 stream fl'stordtiún prLlgrilm. Mmih.'r's tht.'sis.. Fiirt C(liiin~: Cotor,ldn Statl' Unive-rSÎt\', De¡:.utnlt'nt of A~ricuilural ,ind ¡~t.i;ource Ei:oiwrnics. L:5. Deparlmt.nt oí COmmL'rcL'. 1995. Surv,'y of C"urrt'ni business. \..ashinp,ton, DC. Econiimics .ind ~t,lljslics Administr¡itioii. BUrL',lU oí Economlt. Amilysis.. 75(5). C. F. Streiner and 1. ß. Loomis 277 II~._ L'.5. WatE'r Resources CounciL. 1983. Economic and environmental principles for water and related land rE'50UrCE:5 implementation sludies. \-Vashington, DC Young. C. E.. and F. A. TetL 1984. The influence of water quality on the value of recreational properties adjacent to Sr. Albans Bay. VE:rmonc. Staff Reporl No. AGES 831116. Washing- lon, DC: V.S.D.A. Economic Rl!search ServicE:. Natural Resurce Economics Division. Rt(tÌi.'t4; 31 O(to.litr 1995 ,4.cc!'pt!'d: 18 Jiinuary 1996 i~ 278 Rivers · Volume 5, Number 4 October 1995