The main objective of this study is to estimate the welfare values related to sustained water flows in the Zayandeh-Rud River for recreational and cultural amenities in the urban park of Isfahan City in Iran. As is elsewhere the case in arid regions, the drying up of the river due to growing water demand and the increasingly constrained water supply as a result of climate change and more frequent droughts is expected to result in a substantial welfare loss. A double-bounded discrete choice elicitation format is applied in a stated choice survey conducted among local residents and non-residential visitors, focusing on distance-decay and the relationship between income and demand for sustained water flows in publicly provided urban space under climate change. We reject the general finding in the literature that visitors living further away are willing to pay more for unique sites. We show that the recreational services provided by the park can be characterized as a normal economic good for which those living closer by are willing to pay more than those living further away. These results provide an important benchmark for future stated preference research related to welfare valuation of water in urban open space under climate change.

INTRODUCTION

Changes in the hydrological cycle due to climate change can lead to diverse impacts and risks (Jiménez Cisneros et al. 2014). The impacts of climate change on society are felt most severely through the availability and potential future use of water resources, most predominantly in agriculture, but also in other sectors. According to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, water scarcity is expected to be a major challenge for most of the world in the coming decades as a result of increased water demand and lack of good water management (Hijioka et al. 2014). Many of the emerging global climate risks are concentrated in urban areas due to rapid urbanization. The increasing growth of large cities in low- and middle-income countries has been accompanied by the rapid growth of highly vulnerable urban communities at high risk from extreme weather, such as droughts (Revi et al. 2014).

In this paper, we focus on the public amenities provided by urban parks, in particular the role of water in the historical city of Isfahan in Iran, a designated UNESCO World Heritage Site. Conflicting water use as a result of population growth and economic activity, and increasing water scarcity due to climate change, has caused the River Zayandeh-Rud, which runs through the urban park from west to east, to dry up more frequently in the past two decades, ranging from a number of weeks or months to almost an entire year. The historical bridges across the river are one of the characteristic features of the urban park situated along both sides of the river. The drying up of the river is expected to result in a substantial welfare loss.

The main objective of the study presented here is to estimate the welfare value attached to sustained water flows for recreational and cultural amenities in the urban park of the historical city of Isfahan. Visitors to the urban park in Isfahan are asked for their willingness to pay (WTP) to introduce more efficient water saving technologies in all water using sectors, which would secure such a continuous water flow in the future. The valuation method applied in this study is contingent valuation (CV). CV is a survey-based stated preference method where individuals are presented with information about specific environmental changes, the values of which are not accounted for in economic markets or captured through market-based instruments such as an entry fee to the urban park. In the survey, public preferences regarding these changes and their non-market values are elicited. In order to measure the effect of the suggested changes on people's welfare, respondents are typically asked for their hypothetical WTP to support environmental policy and decision-making (Mitchell & Carson 1989; Bateman et al. 2002); in this case study, their WTP to prevent the potential loss of the recreational and cultural amenities associated with sustained urban river flows. Aggregated across those who benefit from the services provided by natural resources, the sum of the WTP amount gives an indicator of the total economic value (TEV) of a change in the level of provision of the services involved, i.e., the water flow-dependent recreational and cultural amenities of the urban park.

Stated preferences (SP) and WTP in particular can be elicited using different formats, including open-ended or closed-ended questions (Bateman et al. 2002). The closed-ended or discrete choice (DC) format presents respondents with given prices (bids) and asks them whether they would be willing to pay a specified price. The simplicity of data collection has made the DC format the most popular technique among CV practitioners (e.g., Alberini et al. 2003; Bateman et al. 2008; Brouwer 2012). In this study, a double-bounded DC elicitation format was developed and applied in a stated choice survey conducted among both local residential and non-residential visitors to the urban park. The double-bounded format implies that respondents, after responding to the first price, are presented with a second price and asked whether they would also be willing to pay that price. Through the introduction of this second WTP question to the follow-up bid, more information on the distribution of WTP is obtained with the same number of interviews. This information helps to further pin down the estimate of WTP (Haab & McConnell 2002). We furthermore investigate two specific methodological challenges in this specific case study: distance-decay and the relationship between income and demand for publicly provided urban space.

The remainder of this paper is organized as a detailed discussion of the study's methodological approach, a short description of the case study area and survey design, the main results, and finally, the conclusion.

METHODOLOGICAL APPROACH

Key methodological issues

There exists an extensive literature on water resources valuation based on SP research in rural areas for recreational and environmental flow purposes, especially in developed countries (e.g., Johnston et al. 2003, 2006; Brouwer et al. 2010, 2015), and urban drinking water supply reliability especially in developing countries (e.g., Whittington et al. 2002; Hensher et al. 2006; Tarfasa & Brouwer 2013). Hardly any SP valuation studies are available focusing on the role of recreational or cultural water flows in urban areas. Water quality levels in the River Tame running through the city of Birmingham in the UK have been valued using SP (Bateman et al. 2006), although more than half of their respondents had never visited the river before. Alternative valuation approaches using economic loss functions for drought-prone urban centers in the USA are presented, for example, in Jenkins et al. (2003).

Although the economic valuation literature of public cultural goods and heritage is steadily growing (e.g., Santagata & Signorello 2000; Noonan 2003), this is, to our knowledge, one of the very few SP studies that aims to estimate the value visitors to an urban park attach to water flows for recreational and cultural amenity purposes. Urban amenities, especially the provision of open space parks, are most commonly valued based on hedonic pricing studies. The few SP studies which value urban open space such as parks focus primarily on urban forest, not water (e.g., Tyrvainen 2001; Kwak et al. 2003; Jim & Chen 2006; Bernath & Roschewitz 2008). Hardly any study combines stated and revealed preference data (Earnhart 2002; Akter et al. 2009). A general equilibrium model of household preferences for urban open space is presented in Walsh (2007).

Revealed preference methods like travel cost and hedonic pricing studies are based on the premise that the value of an environmental good or service decreases as distance increases (e.g., Poor 1999). In the SP literature, such distance-decay is an important, but under-investigated issue (e.g., Loomis 2000; Tyrvainen 2001; Kwak et al. 2003; Jim & Chen 2006; Bernath & Roschewitz 2008), also in the context of cultural heritage (e.g., Tuan & Navrud 2008). In the latter case, it is common practice to simply distinguish between different visitor categories (for example national and international). Typically, foreign visitors are willing to pay more than local visitors, most importantly because of differences in income levels between the two groups (e.g., Mourato et al. 2004).

However, distance-decay would imply that WTP decreases as distance increases. Distance-decay is expected for use values (e.g., Hanley et al. 2003), such as those attached to the recreational and cultural amenity services provided by an urban park. Considerable use values have been found in previous studies for cultural heritage sites (e.g., Alberini et al. 2003). Distance-decay effects have been argued to be less likely to affect WTP for goods that are more widely known and have importance on a larger scale, as these goods may be unique and hence have fewer substitutes (Pate & Loomis 1997). This hypothesis could be extended to cultural heritage, but has never been tested explicitly before as far as we know in the existing literature. The empirical results based on the local versus non-local dichotomy even suggest a reverse distance-decay effect. Hence, our central hypothesis is that distance-decay will not play a significant role in explaining stated WTP in our sample of national visitors due to the wider cultural-historical value of the urban park.

In testing distance-decay, we will control for the influence of income variation between visitor groups. The application of the stated preference approach in a low income, developing country context poses additional challenges (Whittington 2010). Low income levels often result in high shares of non-response and zero bidders (e.g., Akter et al. 2009). In a meta-analysis of 32 CV and hedonic pricing studies, Brander & Koetse (2011) furthermore show that the evidence regarding the relationship between income and WTP for urban space is mixed. Some studies have observed a positive relationship, indicating that urban open space is a normal good. Contrary to theoretical expectations, Brander & Koetse (2011) were unable to detect a significant effect of income on WTP in their meta-analysis. Here, we assume that the urban park is a normal good and income is a significant predictor of WTP, and we will test this explicitly.

WTP elicitation procedure

In view of the incentive compatibility of the DC elicitation format over other CV elicitation formats, and the fact that residential and non-residential visitors are not familiar with being asked to pay to maintain water flow levels in the river running through the urban park (they do not pay an entrance fee or any other fee or charge to gain access to the park), respondents in the survey were asked a DB DC WTP question. Figure 1 illustrates in a flowchart the double- bounded WTP elicitation procedure applied in this study. Five starting bids were used: 0.5, 1, 2, 4, and 8 US dollars (USD), one of which was randomly assigned to every respondent. At the time of the survey in 2010, 10,000 Iranian rials (IRR) equaled approximately 1 USD. Hence, the original starting bids in the survey were IRR 5,000; 10,000; 20,000; 40,000; and 80,000. Depending on the reply to the first bid (yes or no), the second (follow-up) bid was either double or half of the first bid, to which respondents could again answer either yes or no. Hence, if respondents answered no (yes) to the first bid, the follow-up bid was a lower (higher) amount. The follow-up bid intervals were USD 0.25–1; 0.5–2; 1–4; 2–8; 4–16 (IRR 2,500–10,000; 5,000–20,000; 10,000–40,000; 20,000–80,000; 40,000–160,000). The bid levels were based on the results from the survey's pretest, in which an open-ended WTP question was used.
Figure 1

Flowchart of the double-bounded WTP elicitation format applied in the study.

Figure 1

Flowchart of the double-bounded WTP elicitation format applied in the study.

Statistical WTP model

Applying a DB WTP elicitation format, unobserved (latent) individual WTP is given in Equation (1) (Haab & McConnell 2002): 
formula
1
where WTPij represents respondent j's willingness to pay the first or second bid i (i= 1,2). Xij is a vector of respondent characteristics, including household income, use intensity of the urban park and distance living from the urban park, β the corresponding vector of estimated coefficients and ɛ a random error term, assumed to be normally distributed with zero mean and variance σ2: ɛi ∼ N(0, σi2). Based on the DB DC approach, we have information on WTP intervals. Respondents are asked two questions: do you accept the start bid B1 and do you accept the follow-up bid B2. Based on these two questions, four possible intervals can be constructed for WTP (e.g., Alberini et al. 2003):
     
  • WTP B2

    accept both start bid (B1) and follow-up bid (B2)

  •  
  • B1 WTP < B2

    accept the start bid (B1) and reject the follow-up bid (B2)

  •  
  • B2 WTP < B1

    reject the start bid (B1) and accept the follow-up bid (B2)

  •  
  • WTP < B2

    reject both start bid (B1) and follow-up bid (B2)

Deriving the probability of observing each of the possible choice sequences, the jth contribution to the likelihood function can be specified as follows in Equation (2) (Haab & McConnell 2002): 
formula
2
where and are the means for the first and second response and YY = 1 for a yes-yes answer, 0 otherwise, YN = 1 for a yes-no answer, NY = 1 for a no-yes answer, and NN = 1 for a no-no answer, and 0 otherwise. This likelihood function will be estimated using a bivariate probit model. By defining Φɛ1ɛ2 (.) as the standardized bivariate normal cumulative distribution function with zero means (μi), unit variance (σ2), and correlation coefficient ρ, the jth contribution to the bivariate probit likelihood function becomes: 
formula
3
where μ1j = 1 if the response to the first question is yes, and 0 otherwise, μ2j = 1 if the response to the second question is yes, 0 otherwise, d1j = 2μ1j − 1 and d2j = 2μ2j − 1 (Haab & McConnell 2002). Mean and median WTP values are derived by taking the exponential of the negative ratio of the constant term to the bid price coefficient: 
formula
4
 
formula
5
where is a k × 1 row vector of mean values of the explanatory variables including 1 for the constant term, a k-1 × 1 column vector of estimated coefficients, and the estimated variance.

STUDY AREA AND SURVEY DESIGN

Study area

With an average annual precipitation of less than 300 mm, Iran is located in the arid and semi-arid climate zones of the world (Nazemosadat et al. 2006; Kavousi & Meshkani 2007). Available per capita water resources in Iran are merely a quarter of the average world index, and the expected impact of climate change, in particular increasing water scarcity, is becoming more severe in Iran as well as in other parts of the world (Nazemosadat et al. 2006). As a result, mean annual precipitation is decreasing and water resources are becoming increasingly scarce.

The Zayandeh-Rud Basin (see Figure 2) is an important watershed in the central part of Iran. The impacts of climate variability on water flows in the basin are becoming increasingly notable and visible (Massah Bovani & Morid 2005). The basin has a land area of just over 42,000 square kilometers, including 57% of relatively flat lands and 43% of more mountainous landscapes. The average annual rainfall in the basin is only 130 mm, while average monthly temperatures range between 3 °C in winter and 29 °C in summer. The Zayandeh-Rud is the basin's main river. The river flows 400 km eastwards and passes through the desert city of Isfahan, a major historical, cultural, and economic center of Iran with a population of about 1.7 million people (Statistical Center of Iran 2016). According to Hosseini Abari (2003), the city has flourished mainly due to the river. Historic bridges and urban parks located alongside the river provide much needed relief for many from the urban heat and attract many visitors who come to relax and for recreation (Assari & Mahesh 2011). It is the urban park in Isfahan alongside the river that is the area of interest in this study. The freshwater resources, including surface and groundwater, are overexploited in the basin. Water demands continue to grow, while supply has become increasingly constrained due to climate change and the accompanying increase in droughts (Nikouei et al. 2012; Nikouei & Ward 2013). These combined effects have led to the drying up of the river, despite the River Zayandeh-Rud's importance for the city of Isfahan.
Figure 2

Location of the city of Isfahan in the Zayandeh-Rud Basin in Iran.

Figure 2

Location of the city of Isfahan in the Zayandeh-Rud Basin in Iran.

Survey design

The survey was developed and pretested over a time period of three months. A draft version of the survey was critically reviewed and commented upon by 12 Shiraz University employees, followed by two rounds of pretests with a random selection of urban park visitors. Each pretest round consisted of about 35 face-to-face interviews. The pretests were conducted by the same team of six trained interviewers who carried out the main (final) survey. Based on the findings of each pretest round, the survey was modified. The most important change in the second and final pretest round was the conversion of the open-ended WTP values found in the first pretest round into the DB DC elicitation format. No major problems were encountered after the second pretest. After some minor modifications resulting from the second pretest, the final version of the questionnaire was implemented over a time period of two months (May and June 2010). Visitors to Isfahan's urban park were randomly selected on a ‘next to pass’ basis. Given existing socio-cultural conditions, mainly men were interviewed as head of the household.

The final version of the questionnaire consisted of three main parts. The first part referred to the respondent's socio-demographic characteristics, including gender, age, household size and composition, education level, occupation status, income, and place of residence.

The second part included a series of questions related to respondent visitation behavior of the urban park and the recreational activities the respondent undertook when visiting the park (with or without family members). This part also covered issues related to the respondent's experience with and perception of the park's water flow level.

The third and final part of the questionnaire consisted of the CV questions. The discontinuous water flow issue was briefly introduced, followed by a short description of how future water flows could be guaranteed through investments in more efficient water saving technology. These investments would be financed via an increase in the respondent's household municipality bill, which includes all expenses related to water supply management (no separate water bill exists) and is paid by all households in Iran. First respondents were asked whether they would be willing to pay in principle, followed by the DB DC questions. Those respondents who were not willing to pay were asked why not. The data collected with the help of the questionnaire were entered into a database, coded and statistically analyzed using the Stata software.

RESULTS AND DISCUSSION

Sample characteristics

The sample socio-demographic summary statistics are presented in the upper part of Table 1. Most people interviewed were older male heads of household with, on average, 12 years of schooling. A relatively large share of 20% reported having a university degree. The average household size was four. Most respondents were self-employed, about 10% worked for the government, another 10% were unemployed, and just over 10% were retired. Average disposable income was relatively low. More than half of the sample population reported an average monthly income level below the national poverty line of $850 for a household size of four people in urban areas. Half of the sample population earned a disposable monthly household income of $732 (median value). The distribution of income was fairly skewed, and was hence expected to influence stated WTP.

Table 1

Descriptive statistics sample characteristics

Variable Description Statistic St. dev. 
Socio-demographic characteristics 
 Gender Share male (%) 89.8  
 Age Average age 48.2 14.9 
 Household size Average number of people 3.8 1.7 
 Education Average number of school years 12.2 4.5 
Share with university degree (%) 20.0  
 Profession Share government employee (%) 9.1  
Share non-government employee (%) 25.3  
Share self-employed (farmer, shopkeeper, doctor, etc.) (%) 42.6  
Share retired (%) 12.6  
Share unemployed (%) 10.5  
 Income Monthly disposable household income (USD) 927 677 
Share below national poverty line (%) 53.7  
Urban park-related characteristics 
 Resident Share Isfahan City residents (%) 52.8  
 Distance Average distance living from urban park (km) 29.5 91.2 
 Use Average number of days per year visiting the urban park 44.7 55.2 
 Nonuse Share not visiting or much less if river falls dry (%) 60.9  
 Activitya Family picnic (%) 46.5  
Crossing the river's bridges (%) 34.9  
Sports (walking, running, cycling, etc.) (%) 25.8  
Boating or canoeing (%) 8.1  
Enjoying the scenery (%) 3.7  
Earning money (%) 2.3  
Reading and studying (%) 1.6  
Variable Description Statistic St. dev. 
Socio-demographic characteristics 
 Gender Share male (%) 89.8  
 Age Average age 48.2 14.9 
 Household size Average number of people 3.8 1.7 
 Education Average number of school years 12.2 4.5 
Share with university degree (%) 20.0  
 Profession Share government employee (%) 9.1  
Share non-government employee (%) 25.3  
Share self-employed (farmer, shopkeeper, doctor, etc.) (%) 42.6  
Share retired (%) 12.6  
Share unemployed (%) 10.5  
 Income Monthly disposable household income (USD) 927 677 
Share below national poverty line (%) 53.7  
Urban park-related characteristics 
 Resident Share Isfahan City residents (%) 52.8  
 Distance Average distance living from urban park (km) 29.5 91.2 
 Use Average number of days per year visiting the urban park 44.7 55.2 
 Nonuse Share not visiting or much less if river falls dry (%) 60.9  
 Activitya Family picnic (%) 46.5  
Crossing the river's bridges (%) 34.9  
Sports (walking, running, cycling, etc.) (%) 25.8  
Boating or canoeing (%) 8.1  
Enjoying the scenery (%) 3.7  
Earning money (%) 2.3  
Reading and studying (%) 1.6  

aShares do not add up to 100% as respondents are able to undertake and report multiple activities.

Turning to the urban park-related characteristics in the lower part of Table 1, just over half of the sample lived in Isfahan City, and the rest of the sample visited the park from outside the city. As expected, significant differences exist (at the 1% level) between Isfahan City residents and non-residents in terms of average distance traveled to visit the urban park and the average number of days visiting the park per year based on the Mann–Whitney test. Isfahan City residents visit the park on average ten times more often than non-residents, while the latter travel on average about 55 km to visit the park compared to less than 10 km for Isfahan City residents. Non-residents travel up to 500 km to visit the park. Interestingly, Isfahan City residents are less likely to visit the park if the water flow runs dry than non-residents. No significant difference can be detected between the two groups at the 10% level for household income. Also, no big differences exist between the main activities undertaken in the park. Most visitors enjoyed the bridges and had family picnics. Just under 10% hired a boat, and hence used the water flow directly.

Stated WTP

Despite thorough pretesting, about a fifth (22%) of the sample population refused to pay anything at all. Most importantly, because they did not believe that the proposed investments in improved water saving technology could bring about the changes for a continuous flow level in the park (9%), followed by insufficient income (7%), and the belief that the government should pay for the necessary investments (6%). The second reason was theoretically expected. The first and third reason reflect protest bids (e.g., Brouwer & Martín-Ortega 2012). The distribution of WTP responses across the five starting bids is presented in Figure 3. As expected, the relative share of respondents stating a positive WTP reply decreases when going from the lowest to the highest starting bid in the first two columns on the left-hand side of Figure 3 (the columns labeled YY and YN). Vice versa, the relative share of respondents stating a negative WTP reply increases in the last two columns on the right-hand side of Figure 3 (the columns labeled NY and NN).
Figure 3

Distribution of WTP responses across the five starting bids. Note: YY: yes to both bids; YN: yes to first bid, no to follow-up bid; NY: no to first bid, yes to follow-up bid; NN: no to both bids.

Figure 3

Distribution of WTP responses across the five starting bids. Note: YY: yes to both bids; YN: yes to first bid, no to follow-up bid; NY: no to first bid, yes to follow-up bid; NN: no to both bids.

Based on the choice behavior reported in Figure 3, mean and median WTP values were calculated by regressing stated WTP on the bid levels. This was done in Stata 11 using the procedure developed by Jeanty (2007). The Krinsky & Robb (1986) procedure was applied to calculate 95% confidence intervals around mean and median WTP based on 50,000 draws. Bid levels were transformed into their logarithmic form in order to avoid negative WTP. Median WTP values are presented in Table 2, as these appeared to be statistically more efficient than mean WTP values based on the estimated confidence intervals and corresponding standard errors. In Table 2, we only present the difference between the 95% confidence bounds divided by median WTP. The variation coefficients are consistent with the relative efficiency measures presented in Table 2. As expected based on the existing empirical evidence (e.g., Alberini et al. 2003; Bateman et al. 2008; Brouwer 2012), the DB WTP values are consistently and significantly lower than the SB WTP values, albeit at the expense of some measurement precision. Half of the sample population were willing to pay about 40 USD cents per month to secure future flow levels in the urban park. This is 0.05% of disposable household income earned by half of the sample population. Also the average WTP values for local residential and non-residential visitors are presented in Table 2. An important finding is that Isfahan City residents are willing to pay significantly more than non-residents, namely 85 USD cents. This equals just over one tenth of a percent (0.11%) of disposable household income of half of the visitors residing in Isfahan City.

Table 2

Median WTP values in USD per household per month for urban park visitors

  All visitors n = 430 Isfahan City residents n = 227 Non-residents n = 203 
Single bound Median 1.01 (0.81–1.20) 1.62 (1.20–2.13) 0.71 (0.62–0.81) 
CI/median 0.39 0.57 0.26 
Double bound Median 0.38 (0.18–0.58) 0.86 (0.46–1.26) 0.16 (0.05–0.33) 
CI/median 1.06 0.93 1.75 
  All visitors n = 430 Isfahan City residents n = 227 Non-residents n = 203 
Single bound Median 1.01 (0.81–1.20) 1.62 (1.20–2.13) 0.71 (0.62–0.81) 
CI/median 0.39 0.57 0.26 
Double bound Median 0.38 (0.18–0.58) 0.86 (0.46–1.26) 0.16 (0.05–0.33) 
CI/median 1.06 0.93 1.75 

The calculated DB WTP value for all visitors is lower than the more ecosystem services encompassing values found in studies like the one by Ojeda et al. (2008), who estimated the economic value of environmental services in general provided by restored instream flows in the water-scarce Yaqui River Delta in Mexico (USD 4.4 per household per month) or Loomis et al. (2000) who estimated that households would pay on average USD 21 per month for a wide variety of ecosystem services including dilution of wastewater, natural purification of water, erosion control, habitat for fish and wildlife, and recreation along a 45-mile section of the Platte River in the USA.

Explaining stated WTP

Two different bivariate probit models were estimated to test the effect of different visitor categories and distance-decay on stated WTP and examine the robustness of the results at the same time. The first model (Model I in Table 3) includes a simple dummy variable distinguishing between local residential and non-residential visitors to the urban park, while the second model (Model II in Table 3) accounts for possible distance-decay through the inclusion of a more sophisticated continuous distance variable. Both models include otherwise the same theoretically expected variables in the SB and DB WTP function, such as bid price, disposable income, use intensity, engagement in water-related recreational activities, and the impact of not having any water in the river on future use or non-use.

Table 3

Bivariate probit regression results for stated WTP to the initial (single bound) and follow-up bid (double bound)

Explanatory factor Model I
 
Model II
 
Single bound Double bound Single bound Double bound 
coefficient estimate coefficient estimate coefficient estimate coefficient estimate 
Constant  2.382*** (0.685) −1.853 (1.361) 2.410*** (0.615) −1.867 (1.985) 
Bid Natural log IRR × 10−3 −2.334*** (0.273) −1.336*** (0.317) −2.315*** (0.264) −1.372*** (0.333) 
Income Natural log IRR × 10−4 0.213*** (0.065) 0.482*** (0.187) 0.202*** (0.063) 0.529* (0.284) 
Resident Dummy = 1 if respondent is local resident 0.061 (0.308) 0.929*** (0.195) – – 
Distance Kilometers – – 0.0004 (0.001) −0.003*** (0.001) 
Use Days 0.013*** (0.004) 0.001 (0.002) 0.013*** (0.003) 0.005*** (0.002) 
Boating Dummy = 1 if respondent visits park for boating 0.476** (0.237) −0.151 (0.298) 0.465** (0.243) 0.041 (0.291) 
Non-use Dummy = 1 if less likely to visit if there is no water 3.233*** (0.411) 1.539*** (0.433) 3.231*** (0.376) 1.869*** (0.454) 
Model summary statistics 
 ρ  0.110 (0.281)  0.062 (0.277)  
Log likelihood   −256.521  −261.428  
Wald chi-squared  143.470***  116.060***  
 430  430  
Explanatory factor Model I
 
Model II
 
Single bound Double bound Single bound Double bound 
coefficient estimate coefficient estimate coefficient estimate coefficient estimate 
Constant  2.382*** (0.685) −1.853 (1.361) 2.410*** (0.615) −1.867 (1.985) 
Bid Natural log IRR × 10−3 −2.334*** (0.273) −1.336*** (0.317) −2.315*** (0.264) −1.372*** (0.333) 
Income Natural log IRR × 10−4 0.213*** (0.065) 0.482*** (0.187) 0.202*** (0.063) 0.529* (0.284) 
Resident Dummy = 1 if respondent is local resident 0.061 (0.308) 0.929*** (0.195) – – 
Distance Kilometers – – 0.0004 (0.001) −0.003*** (0.001) 
Use Days 0.013*** (0.004) 0.001 (0.002) 0.013*** (0.003) 0.005*** (0.002) 
Boating Dummy = 1 if respondent visits park for boating 0.476** (0.237) −0.151 (0.298) 0.465** (0.243) 0.041 (0.291) 
Non-use Dummy = 1 if less likely to visit if there is no water 3.233*** (0.411) 1.539*** (0.433) 3.231*** (0.376) 1.869*** (0.454) 
Model summary statistics 
 ρ  0.110 (0.281)  0.062 (0.277)  
Log likelihood   −256.521  −261.428  
Wald chi-squared  143.470***  116.060***  
 430  430  

*Significant at 10%.

**Significant at 5%.

***Significant at 1%.

IRR: Iranian rials.

The two models presented in Table 3 are the statistically best-fit models. The Wald test shows that the coefficient estimates in both models are significantly different from zero and no other significant effects on stated WTP could be detected, related to either visitor socio-demographic characteristics (e.g., gender, age, education level, household size), their perception and attitude towards the park, or the recreational activities undertaken in the park. The estimated models account for the panel data structure of the data. The correlation coefficient ρ is not significant in either Model I or II, indicating that the random component of WTP for the initial and follow-up bids is not significantly different. Coefficient estimates were also tested for correlation. No significant correlations could be detected between any of the explanatory variables included in the statistically best-fit models. Hence, similar unobserved factors seem to have played a role when respondents went through the DC sequence. However, the influence of the observed explanatory variables in the first (SB) and second (DB) WTP function is distinctively different in both models. Examining the size and significance of the coefficient estimates across the two estimated models I and II, the models appear robust. Only some differences are observed in the DB WTP functions: the estimated impact of household income is less significant in Model II, while the use intensity variable is highly significant in Model II but not in Model I.

Initially no significant effect can be found for either local residents or non-residents, or distance-decay in the first SB function. This indicates that the recreational and cultural amenities provided by the urban park are valued equally irrespective of where visitors live or come from. This may be partly explained by the fact that local residential and non-residential visitors do not differ in terms of their disposable household income, and partly by the wider cultural heritage status of the urban park. However, when squeezed for their maximum WTP in the follow-up bid function, a significant effect can be observed and residents appear to be willing to pay more than non-residents who live further away. The latter is consistent with the significant distance-decay effect detected in the second DB WTP function. The distance-decay effect is linear. We tested for non-linearity in distance-decay, for instance by also including a squared term for distance or transforming the distance variable into its natural log form, but this did not yield any significant results. So, while controlling for a variety of influencing factors, a significant positive effect is again found for local residents when respondents are asked to reveal their maximum WTP in the second WTP function. This finding is consistent with the expected distance-decay effect for the use values attached to the recreational benefits provided by the urban park, as also reported, for example, in Loomis (2000) and Bernath & Roschewitz (2008), despite its cultural heritage status, rejecting our central hypothesis. Distance plays initially no role when asked to pay extra to conserve future water flows in the park, but when pushed further for their WTP in the follow-up question, visitors living further away are less likely to pay the second bid amount. This finding is in line with Hanley et al. (2003), who found that WTP decreases as distance increases for water quality improvements, while our results show that this relationship also holds for changes in water levels. Visitors who live near to urban parks and use the recreational services provided by sustained urban water flows value these flows more than others. As a result, the former are willing to pay more for sustained urban water flows.

As theoretically expected, the bid levels have a significant negative effect on stated WTP and household income a significant positive effect, as also reported in most other studies (e.g., Carson & Mitchell 1993; Ojeda et al. 2008). The influence of the latter variable is stronger in the follow-up WTP question than the initial WTP question, and hence confirms the other hypothesis that ability to pay is a strong predictor of WTP. Flow levels are considered important by all visitors, but when asked for their financial commitment towards preservation of the water flow in the park, respondents were willing to pay, on average, less than 1% of their disposable monthly household income.

Another important theoretically expected variable is park use. The variable plays a significant positive role in the SB WTP function in both models, but not in the follow-up DB question in Model I. Hence, the more intensively respondents use the park, the more likely they will reply in a positive way to the first WTP question. The same applies to visitors who depend for their enjoyment (partly) on the presence of water, because they go boating. The finding that WTP strongly depends on the kind of recreational services provided by the public goods involved, such as boating, is confirmed in other research (e.g., Bateman et al. 2006; Bernath & Roschewitz 2008). Respondents who stated that they would not visit the park or would visit the park less if there is no water flow in the future are more likely to be willing to pay a positive price, both in the SB and DB WTP function.

CONCLUSIONS

The study described in this paper is one of the first of its kind to use a stated preference approach to assess the non-market value of water in an urban park. The study faced several methodological challenges, of which the low income level of domestic visitors to the urban park in the historical City of Isfahan and the lack of experience in paying for park access, let alone water flow levels in the park, were among the most important ones. A majority of the random selection of visitors (78%) were, nevertheless, willing to pay extra municipality taxes to conserve water flow levels in the park. As a result, the drying up of the river is expected to result in a substantial welfare loss. The protest rate was limited to 15%, which is considered reasonable in view of the above. Less than 10% refused to pay due to income constraints.

Although no benchmark for the non-market value of water in urban parks exists in the stated preference literature, the results in this study seem credible and reliable for use in actual policy and decision-making related to the river's future flow levels. As expected, the mean WTP for the specific urban recreational amenities valued in this study is only a fraction of household income levels and findings reported in other stated preference studies focused on the wider welfare implications of instream flow restoration. This suggests that the applied approach holds promise for application elsewhere in arid regions facing climate change to inform decision-making aiming to reconcile extractive water uses and sustainable urban flow levels. However, in order to make full use of the results presented in this paper in policy and decision-making, additional information is needed regarding visitor numbers to the urban park in order to be able to calculate the TEV of water flow preservation.

The validity of the study was tested based on a priori theoretical expectations. Several model specifications were tested. In general, WTP voting behavior in the choice sequence was influenced by more or less the same variables although the degree of their impact differed between the two WTP bid functions. The central tendency of the WTP values decreased significantly when going from the first to the second WTP question, as evidenced more generally in the CV literature, but the statistical efficiency of the WTP measures did not increase. On the contrary, the relative size of the confidence intervals around the estimated WTP values increased somewhat, suggesting that respondents became slightly more uncertain about their WTP responses.

As expected, income had a significant positive impact on stated WTP, as well as visitation rate and visitors who directly use the water for recreational purposes, while higher price levels and distance living from the park resulted in lower WTP. The latter distance-decay constituted the central hypothesis of this study. No significant distance-decay could be detected initially, only when visitors were pushed to reveal their true WTP in the follow-up question. The estimated distance-decay function is an important step forwards in the existing literature compared to the inclusion of dummy variables for different visitor categories. While controlling for disposable income, we show that non-residential visitors living further away are not necessarily willing to pay more than local residential visitors. On the contrary, we demonstrate that the recreational services provided by a well-known cultural-historical urban park can be characterized as a normal economic good for which those living closer by are willing to pay more than those living further away as predicted by economic theory. This is an important message for municipal policymakers elsewhere who are responsible for the management of cultural heritage sites that depend crucially on local water systems. Under climate change and increasingly frequent drought events, the drying up of urban river systems leads to unevenly distributed welfare losses. Ideally, also non-users of the urban park of Isfahan City would have been included in the sample selection procedure to test possible distance-decay effects for both user and non-user groups, but this would have required a more extensive sampling procedure at different distances from Isfahan City for which we lacked the necessary resources. We consider this an important extension and avenue of future research.

ACKNOWLEDGEMENTS

The authors are grateful for the financial support received from the Isfahan Research and Education Center of Agriculture and Natural Resources, Iran.

REFERENCES

REFERENCES
Akter
S.
Brouwer
R.
Brander
L.
van Beukering
P.
2009
Respondent uncertainty in a contingent market for carbon offsets
.
Ecological Economics
68
,
1858
1863
.
Assari
A.
Mahesh
T. M.
2011
Demographic comparative in heritage texture of Isfahan City
.
Journal of Geography and Regional Planning
4
,
463
470
.
Available at: http://www.academicjournals.org/JGRP (accessed 20 March 2012)
.
Bateman
I. J.
Carson
R. T.
Day
B.
Hanemann
M.
Hanley
N.
Hett
T.
Jones-Lee
M.
Loomes
G.
Mourato
S.
Ozdemiroglu
E.
Pearce
D. W.
Sugden
R.
Swanson
J.
2002
Economic Valuation with Stated Preference Techniques: A Manual
.
Edward Elgar
,
Cheltenham
,
UK
.
Bateman
I. J.
Burgess
D.
Hutchinson
W. G.
Matthews
D. I.
2008
Learning design contingent valuation (lDCV): NOAA guidelines, preference learning and coherent arbitrariness
.
Journal of Environmental Economics and Management
55
,
127
141
.
Brander
L. M.
Koetse
M. J.
2011
The value of urban open space: Meta-analyses of contingent valuation and hedonic pricing results
.
Journal of Environmental Management
,
92
,
2763
2773
.
Brouwer
R.
2012
Constructed preference stability: a test-retest
.
Journal of Environmental Economics and Policy
1
,
70
84
.
Brouwer
R.
Martín-Ortega
J.
Berbel
J.
2010
Spatial preference heterogeneity: a choice experiment
.
Land Economics
86
,
552
568
.
Brouwer
R.
Martinâ-Ortega
J.
Dekker
T.
Sardonini
L.
Andreu
J.
Kontogianni
A.
Skourtos
M.
Raggi
M.
Viaggi
D.
Pulido-Velazquez
M.
2015
Improving value transfer through socio-economic adjustments in a multicountry choice experiment of water conservation alternatives
.
Australian Journal of Agricultural and Resource Economics
59
,
458
478
.
Haab
T. C.
McConnell
K. E.
2002
Valuing Environmental and Natural Resources: The Econometrics of Non-market Valuation
.
Edward Elgar
,
Cheltenham
,
UK
.
Hensher
D. A.
Shore
N.
Train
K.
2006
Water supply security and willingness to pay to avoid drought restrictions
.
The Economic Record
82
,
56
66
.
Hijioka
Y.
Lin
E.
Pereira
J. J.
Corlett
R. T.
Cui
X.
Insarov
G. E.
Lasco
R. D.
Lindgren
E.
Surjan
A.
2014
Asia
. In:
Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part B: Regional Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change
(
Barros
V. R.
Field
C. B.
Dokken
D. J.
Mastrandrea
M. D.
Mach
K. J.
Bilir
T. E.
Chatterjee
M.
Ebi
K. L.
Estrada
Y. O.
Genova
R. C.
Girma
B.
Kissel
E. S.
Levy
A. N.
MacCracken
S.
Mastrandrea
P. R.
White
L. L.
, eds).
Cambridge University Press
,
Cambridge
,
UK
and New York, NY, USA
, pp.
1327
1370
.
Hosseini Abari
S. H.
2003
Zayandeh-Rud and Isfahan
.
Geographical Research
8
,
105
118
(
in Persian
).
Jeanty
P. W.
2007
Constructing Krinsky and Robb Confidence Interval for Mean and Median WTP Using Stata
. In:
North American Stata Users’ Group Meetings 2007
.
Available online at: http://ideas.repec.org/p/boc/asug07/8.html (accessed 8 May 2011)
.
Jenkins
M. W.
Lund
J. R.
Howitt
R. E.
2003
Using economic loss functions to value urban water scarcity in California
.
Journal of American Water Works Association
95
,
58
70
.
Jiménez Cisneros
B. E.
Oki
T.
Arnell
N. W.
Benito
G.
Cogley
J. G.
Döll
P.
Jiang
T.
Mwakalila
S. S.
2014
Freshwater resources
. In:
Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change
(
Field
C. B.
Barros
V. R.
Dokken
D. J.
Mach
K. J.
Mastrandrea
M. D.
Bilir
T. E.
Chatterjee
M.
Ebi
K. L.
Estrada
Y. O.
Genova
R. C.
Girma
B.
Kissel
E. S.
Levy
A. N.
MacCracken
S.
Mastrandrea
P. R.
White
L. L.
, eds).
Cambridge University Press
,
Cambridge
,
UK
and New York, NY, USA
, pp.
229
269
.
Johnston
R. J.
Besedin
E. Y.
Wardwell
R. F.
2003
Modeling relationships between use and nonuse values for surface water quality: a meta-analysis
.
Water Resources Research
39
,
1363
1372
.
Johnston
R. J.
Ranson
M. H.
Besedin
E. Y.
Helm
E. C.
2006
What determines willingness to pay per fish? A meta-analysis of recreational fishing values
.
Marine Resource Economics
21
,
1
32
.
Kavousi
A.
Meshkani
M.
2007
Mapping and spatial analysis of Iran's climate precipitation
.
Environmental Studies
33
,
31
40
(
in Persian
).
Krinsky
I.
Robb
A. L.
1986
On approximating the statistical properties of elasticities
.
Review of Economic and Statistics
68
,
715
719
.
Massah Bovani
A.
Morid
S.
2005
The effects of climate change on water resources and agricultural production: a case study in the Zayandeh-Rud Basin of Isfahan, Iran
.
Water Resources Research
1
,
40
47
(
in Persian
).
Mitchell
R. C.
Carson
R. T.
1989
Using Surveys to Value Public Goods: The Contingent Valuation Method
.
Resources for the Future
,
Washington, DC
,
USA
.
Mourato
S.
Ozdemiroglu
E.
Hett
T.
Atkinson
G.
2004
Pricing cultural heritage: a new approach to managing ancient resources
.
World Economics
5
,
95
113
.
Nazemosadat
S. M. J.
Samani
N.
Bary
A.
Molaei Nikoo
M.
2006
Effective force phenomenon El-Niño –Southern Oscillation (ENSO) on climate change in Iran using analysis of precipitation data
.
Journal of Science and Technology
30
,
555
565
(
in Persian
).
Nikouei
A.
Ward
F. A.
2013
Pricing irrigation water for drought adaptation in Iran
.
Journal of Hydrology
503
,
29
46
.
Poor
P. J.
1999
The value of additional central flyway wetlands: the case of Nebraska's rainwater basin wetlands
.
Journal of Agricultural and Resource Economics
24
,
253
265
.
Revi
A.
Satterthwaite
D. E.
Aragón-Durand
F.
Corfee-Morlot
J.
Kiunsi
R. B. R.
Pelling
M.
Roberts
D. C.
Solecki
W.
2014
Urban areas
. In:
Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change
(
Field
C. B.
Barros
V. R.
Dokken
D. J.
Mach
K. J.
Mastrandrea
M. D.
Bilir
T. E.
Chatterjee
M.
Ebi
K. L.
Estrada
Y. O.
Genova
R. C.
Girma
B.
Kissel
E. S.
Levy
A. N.
MacCracken
S.
Mastrandrea
P. R.
White
L. L.
, eds).
Cambridge University Press
,
Cambridge
,
UK
and New York, NY, USA
, pp.
535
612
.
Statistical Center of Iran
2016
Population estimated for urban and rural areas, edited by Statistical Center of Iran
.
Available at: http://www.amar.org.ir/ (accessed 7 April 2016)
.
Tuan
T. H.
Navrud
S.
2008
Capturing the benefits of preserving cultural heritage
.
Journal of Cultural Heritage
9
,
326
337
.
Tyrvainen
L.
2001
Economic valuation of urban forest benefits in Finland
.
Journal of Environmental Management
62
,
75
92
.
Walsh
R.
2007
Endogenous open space amenities in a locational equilibrium
.
Journal of Urban Economics
61
,
319
344
.
Whittington
D.
Pattanayak
S. K.
Yang
J.
Bal Kumar
K. C.
2002
Household demand for improved piped water services: evidence from Kathmandu, Nepal
.
Water Policy
4
,
531
556
.