Journal Search Engine
Search Advanced Search Adode Reader(link)
Download PDF Export Citaion korean bibliography PMC previewer
ISSN : 1225-8504(Print)
ISSN : 2287-8165(Online)
Journal of the Korean Society of International Agricultue Vol.31 No.3 pp.207-213

Valuing Preferences for Water Quality Improvement in Urban Area: a Case Study in To Lich River, Hanoi, Vietnam

Nguyen Thi Hai Yen*, Hio-Jung Shin*, Saem Lee**†
*Dept. of Agricultural & Resource Economics, Kangwon National Univ., Chunchon, 24341, Korea
**Research Institute for Gangwon, Chunchon, 24265, Korea
Corresponding author (Phone) +82-33-250-1319 (E-mail)
June 3, 2019 September 17, 2019 September 20, 2019


The purpose of this study is to estimate residents’ Willingness to Pay (WTP) for the water quality improvement along To Lich river. A case study was implemented on six Hanoi’s districts locate along the river. The empirical results from a double-bounded dichotomous choice contingnet valuation method reveal that the expected annual mean WTP for improving the water quality was VND 202,560 per person (US $8.6). In bivariate probit model, we find that the variable health concern and income were statistically significant with positive sign. Meanwhile, the variable age appears to have negative effect on the WTP. This means that the respondents with health concerns and higher income would have more interests in the WTP and younger respondents tend to have higher willing to pay for the water quality improvement. This paper can provide useful information for sustainable water resource management in Vietnam.

베트남 To Lich River 수질 개선에 대한 소비자 행태 분석에 관한 연구

응 웬 티 하이 엔*, 신 효중*, 이 샘**†
*강원대학교 농업생명과학대학 농업자원경제학과



    The nexus of water-food-energy is vital to sustainable development and this concept can be used to support the water management policy (Waughray, 2011; Stijn et al., 2017). The rapid economic growth, urbanization and global population boom have put a huge pressure on the demand for the freshwater resources, energy and food production and supply (FAO, 2014). To ensure sustainable development, where all the three purposes of water and food security, sustainable agriculture, and energy conservation are achieved, a suitably integrated approach is required. The linkage between the three domains is strong so that the negative situation of one aspect will affect the two aspects. Especially for the agricultural economy relies mainly on wet-rice production like Vietnam, the management of water quality and quantity is vital for food security and the economic development.

    After the economic revolution in the 1980s, rapid economic development and population boom in Hanoi have put a massive demand on freshwater all around the city. However, the water supplying and purifying infrastructure is not improved as fast as the economics and population increase. According to the Ministry of Natural Resources and Environment of Vietnam (MONRE), most of the rivers in Hanoi are heavily polluted due to wasted water discharged from industrial and domestic activities (MONRE, 2010). The water pollution is a serious problem of recent decades, distracting daily life and economics practices of residents live in the river basins. However, because most of the inner Hanoi city’s domestic wastewater is not treated properly, the polluted water can affect both the residents who live near the river and the bigger society through consumption of commodities that produce by the water directly and indirectly.

    To Lich River (TLR) is important for Hanoi city, as it provides irrigation water for the downstream area (MONRE, 2010;Nguyen et al., 2008). However, the river is severely contaminated because it receives nearly twothirds of the wastewater discharged from the inner-city area. Since 1998, Hanoi had implemented several efforts to improve the water quality, for example, they built an embankment along the river to protect the river from solid waste and non-point wasted water. However, due to continuously growing population, urbanization, and the increasing in the number of factories, in Hanoi, the polluted and inundation problem have happened every year and the phenomenon is getting worse. Nguyen et al. (2008) indicated the river severely polluted, the water is not sufficient for irrigation purpose. The water quality improvement is essential for Hanoi city. However, there is a lack of researches that investigate citizen’s willingness to pay for improving rivers’ water quality using a contingent valuation method (CVM).

    The CVM is a method broadly used for estimating nonmarket value and non-used value or both of environmental resources (Bishop and Romano, 1998;Mitchell and Carson, 1989). Despite some strong controversial arguments around the technique, CVM method is broadly used as it provides helpful information to policy decision-makers (FAO, 2000). Many of contingent valuation studies have been done in developed countries such as US and Europe (Connelly et al., 2002;Julián and Salvador, 2012;Halkos and Matsiori, 2014). However, in developing country, in particular, in Vietnam, recent studies using CVM are in its infancy stage. There is still a lack of CVM literatures in relation with water quality protection in Vietnam.

    Thus, the purpose of this study is to: (i) estimate the willingness to pay (WTP) for the water quality improvement along the To Lich river in Vietnam using CVM, (ii) investigate important determinants affecting the WTP for the water quality improvement. This study can contribute to future policy establishment for sustainable water management as suggested in integrated water management.


    1. Study area

    The capital of Vietnam, Hanoi is one of the most densely populated cities in Vietnam. The GDP per capita of Hanoi city was around US $ 3,425 in 2017 (HSO, 2018). According to Hanoi Statistics Office (HSO), the total population of Hanoi was 7,739.4 thousand, and the population density was 2,304 people/km2 in 2017, increased from 1,929 people/ km2 in 2009 (HSO, 2009). As stated by MONRE, an urban citizen tends to discharge waste into a river at average 2-3 times, compares with one people lives in rural area (MONRE, 2010).

    To Lich river (TLR) is a small river which originates from West Lake, flows through six central districts of Hanoi city. In its way draining into Nhue river at downstream, it receives water from Kim Nguu river, Lu river and Set river. TLR meets Nhue river at Thanh Liet dam in rainy season. The river flows into Nhue river and it is one of the Nhue river’s main pollution sources. However, when the TLR water level decreases below the dam in dry season or when government closes the dam to protect the agricultural practices in the downstream area, the water is directed towards Yen So lake and is pumped into Red River.

    To Lich river (TLR) is a small river, with a length of approximately 14 km, originates from West Lake and flows through 6 districts of Hanoi. The six central districts along TLR are Cau giay, Ba dinh, Dong da, Thanh xuan, Hoang mai, and Thanh tri districts. These districts population density is nearly 13,100 people per km2, approximately 5.7 times the Hanoi’s average (HSO, 2018). While the total area accounts for only 4.2% of Hanoi city, their total population is 24.4% of Hanoi’s population. Only TLR flows through these districts and receives almost all the untreated industrial and domestic wastewater in the area, equal to two-thirds of the Hanoi’s inner wasted water (Nguyen et al., 2013).

    TLR is regarded as one of the most polluted rivers in Hanoi and highly polluted in surface water and sediment by heavy metal and other substances (Nguyen et al., 2015;Nguyen et al., 2013;Fuhrimann et al., 2016). The wastewater of TLR is used for vegetable cultivation and aquaculture practices without any cleaning treatment (MONRE, 2006;Vuong et al., 2007). These agricultural products, then be sole in Ha Noi and other regions all year round. This is important considering the need of improving the TLR (Nguyen et al., 2008) because the residue of harmful substances and microorganism in the vegetable and fish, which are cultivated using insufficient wastewater, will be absorbed into the human body (MONRE, 2006;Nguyen et al., 2015). In rainy season, because the water level increases, the polluted water from TLR flows through Thanh Liet dam into Nhue river, decreasing the Nhue river’s quality and affect the daily life of residents live along Nhue river and economic activities in Nhue river’s basin area. During our survey, residents living along the river have claimed that odor and mosquito spread out from the River effects their daily activities heavily.

    2. Survey and data

    CVM survey is based on a hypothetical scenario which must be designed in a way as much real as possible (Hanemann, 1994). NOAA (1993) has released guidelines in which, they provided some recommendations on the CVM survey design and administrations. This study’s survey was designed and implemented carefully based on the NOAA’s recommendation.

    We implemented a face-to-face interview in Vietnam in July and August 2018. In addition, before conducting the final survey, the overall questionnaire was tested by a pilot survey. The survey is then implemented in six districts locate along TLR in Hanoi, Vietnam. Compensation was given to the surveyed respondents as an incentive. Total 375 samples were collected through the survey. However, due to non-responses, 53 samples were deleted.

    The questionnaires are composed of three parts. The first part is about how vital the environment in general and water environment to them by asking them. In this part, the interviewers asked respondents how the TLR quality affects their family’s daily life and health. Secondly, the questionnaire provided respondents with information about current status of TLR water and told them that there is a program for improving the water quality and supposed that the project will require every citizen to pay an amount of tax. We provided four types of initial bid amounts (VND 5,000, VND 10,000, VND 15,000 and VND 20,000). Then, the second bids were offered by increase or decrease in the first bid by 20%. Finally, the third part includes socio-economic background, such as age, education level, and income.

    3. WTP elicitation formats

    The double bounded dichotomous choice (DBDC) method was utilized for estimating the WTP (Hanemann et al., 1991). The double-bounded method is more statistically effective and minimizes bias compared with a singlebounded method (Haab and McConnell, 2002).

    The DBDC model is a close-ended format and has a binary answer, a yes or no response to the first values (B1) and follow-up WTP values (B2). The follow-up values rely on the respondents’ response to the proposed initial WTP value; if the initial value is taken, the follow-up value is double, whereas if the first value is not accepted, the follow-up value is half of the value as much. This DBDC method can give an economic measure of individual welfares relevant to environmental change. In the format, the probability of WTP is equal to or greater than offered bids (В) can be indicated as:

    Pr ( yes ) = Pr ( W T P B ) 1 F c ( B ) ,

    where Fc(В) indicates the cumulative distribution function of WTP. In the indirect utility function, all components are unobservable. In the random utility model, error terms means random variables, the probability of the “yes” answer can be written as:

    Pr ( yes ) = Pr {(C(Q 0 , Q 1 , Y, P, Z, ε B ) Pr{V(Q 1 , Y-B,P, Z,ε) V(Q 0 , Y, P, Z,ε)} 1 F c ( B )

    where (Q) demonstrates the scalar for values of environmental quality change, (P) is the price vector for the market goods, (Z) means the socio-economic information, and (Y) is the income of respondents. The utility function is considered as V (Q0, Y, P, Z, ε). If water quality change happens to the given scenario, the utility function is shifted to V(Q1, Y– B, P, Z, ε).

    In this respect, the compensation variation (C) means the expected mean WTP for the water quality change. It makes the maximum WTP of respondents for the water quality change from the first situation, status quo (Q0) to changed situation (Q1). If μWTP is equal to E[WTP(Q0, Q1,Y, P, Z, ε)], δ WTP 2 is equal to Var[WTP(Q0, Q1, Y, P, Z, ε)] and F(•) denotes the cumulative distribution function of the standard normal variate ω =  ( WTP- μ W T P ) / δ W T P , the probability function can be written as:

    Pr ( y e s ) = 1 F [ B μ W T P δ W T P ] 1 F ( α + β B ) ,

    where α = μ W T P / δ W T P and β = 1/ δ W T P . In this dichotomous choice method, the model for WTP estimates is determined by cumulative distribution function of WTP (C), Fc(В) and distribution of the random component.

    In our study, a respondent (j) is given the first and the second bid amount for improving the water quality. There are four types of format as follows in DBDC format: Yes- Yes (YY) answers (WTPj ≥ B2), Yes-No (YN) answers (B1 ≤ WTPj < B2), No-Yes (NY) answers (B1 ≥ WTPj > B2), and No-No (NN) answers (WTPj < B2).

    The Yes-Yes(YY) answers are if the respondent answers “yes” for both the first bid and the second bid which mean WTPj ≥ B1 and WTPj ≥ B2. The Yes-No (YN) answers mean if the respondent answers “yes” for the first bid and “no” for the second bid, WTPj ≥ B1 and WTPj < B2 , the highest willingness to pay is between WTPjL and WTPjH. No- Yes, if the respondent answers “no” for the first bid and “yes” for the second bid, WTPj < B1 and WTPj > B2. The No-No answers mean “no” for both the first bid and the second bid, WTPj < B1 and WTPj < B2, the highest willingness to pay is between 0 and WTPjL.

    Thus, the probability of the responses is offered by

    Pr(Yes  Yes)P yy = Pr (WTP j 1 > B 1 , WTP j 2 > B 2 )
    Pr(No  No)P nn = Pr (WTP j 1 <B 1 , WTP j 2 <B 2 )
    Pr(Yes No)P yn = Pr (WTP j 1 > B 1 , WTP j 2 < B 2 )
    Pr(No Yes)P ny = Pr (WTP j 1 <B 1 , WTP j 2 >B 2 )

    The basic econometric model for the DBDC method can be defined as below.

    WTP b i d i = μ b i d + ε b i d i

    WTPbidi represents the ith individual’s WTP, and bid means the first and second bid answers. The μ b i d is the mean value for WTP

    L j ( μ / B ) = Pr ( μ b i d 1 + ε b i d 1 i B 1 , μ b i d 2 + ε b i d 2 i < B 2 ) Y N
    Pr ( μ b i d 1 + ε b i d 1 i > B 1 , μ b i d 2 + ε b i d 2 i B 2 ) Y Y
    Pr ( μ b i d 1 + ε b i d 1 i < B 1 , μ b i d 2 + ε b i d 2 i > B 2 ) N Y
    Pr ( μ b i d 1 + ε b i d 1 i < B 1 , μ b i d 2 + ε b i d 2 i < B 2 ) N N

    Where YY = 1 for a Yes-Yes answer, 0 otherwise, NY = 1 for a No-Yes answer, 0 otherwise, YN = 1 for a Yes-No answer, 0 otherwise and NN = 1 for a no-no answer, 0 otherwise. This formulation is defined to as the discrete choice model.

    The mean WTP from bivariate probit model were calculated using the formula (Habb and Mconnell, 2002). Mean WTP = μ =  α β , where α is the constant or intercept term, β is the coefficient of the given bid value to the respondents. To produce the likelihood function, it can derive the probability of observing each of the possible two-bid response (Yes-Yes, Yes-No, No-Yes, No-No).

    In the DBDC method, a bivariate probit model was used to calculate the expected annual mean WTP for the water quality improvement. The offered bids for the WTP varied to avoid initial bid biases. The estimates of the probit model are gained through maximum likelihood techniques. A quantitative interpretation of the model coefficients can be implemented in marginal effect. All statistical analysis was implemented using Stata version 14.


    1. Socio-economic characteristics

    Table 1 shows the descriptive statistics for the total samples. Of 322 respondents, 81.9% of respondents were in the age from 20 to 59 years old, in which respondents who are thirties account for about 28% of total responses. There are 74 interviewees in the age of 40s, account for 23% of total respondents. The numbers of respondents in the age of 60s and above 60s are equal, each of them accounts for 16.5%. As the median age of Hanoi population were 30.5 years in 2018, Hanoi residents aged 25-54 years occupied 45.6% which was similar with the rate of the respondents.

    About 42% respondents have under-graduated diploma, indicates that respondents in this survey have high education level. 33.5% of respondents were high school education level. 58 respondents (18%) said that they have secondary school degree and only 6.8% respondents have primary and below primary education level.

    The majority of respondents (44.7%) answered that their family total income ranges from VND 12 to 17.9 million per month. 32.3% of respondents answered that their income is more than VND 18 million per month. Only 74 out of total 322 respondents (23%) said that their family total monthly income is less than VND 12 million.

    2. Mean WTP

    Table 2 shows the expected mean and aggregate WTPs for the water quality improvement. The mean WTP per person is VND 16,880 (US $ 0.72) per month, the aggregate WTPs for improving the TLR are VND 130.6 billion, approximately USD 5.6 million per month. Specifically, the expected annual WTP is VND 202,560, about 0.24% of Hanoi’s average income per capita. Multiplied with Hanoi’s total population of 7,739.4 thousand people, the aggregate annual WTPs for the TLR are VND 1,579.68 billion (US $ 67.8 million) for improving TLR quality.

    3. Key determinants of the WTP

    Table 3 shows the estimation results from bivariate probit model. The statistically significant factors are health concerns, age and income in the model. The variable age was found to be a significant contributor to paying the WTP for the water quality improvement, with a 28% decrease. In addition, our finding indicates that the predicted probability of the WTP for the water quality improvement increases by 21% as health concerns of the respondent’s increases. This implies that if the respondents think polluted water has bad effect on them and their family’s health, their payment tendencies will increase. If the respondent considers the effect is terrible, their willingness to pay for water improvement increases and vice versa. Higher income respondents have a higher probability of the willingness to pay for the water quality improvement by 8%.

    The results find that age is a negative determinant and income is a positive determinant of WTP’s probabilities. Our results are consistent with the results by Khai et al. (2014), Salvador et al. (2012) who reported the sociodemographic characteristics such as age, income have an important influence on individual’ WTP probabilities for water quality improvement. However, in this case, there is no significant relationship between education and the paying intensions of the WTPs (Halkos and Matsiori, 2014; Pham et al., 2018). In addition, our finding is in line with the result of previous researches for tap-water system improvement (Beaumais et al., 2010;Jianjun et al., 2016). Those researches showed that citizens who have high perception of health risk caused by low tap-water quality might express higher willingness to pay.

    Further research endeavors should avoid methodological limitations. This research tried to make sophisticated survey design in CVM. However, it is still the case using a typical CVM interview, although this study provided the application of the CVM to the issue of the water quality in Vietnam. Further studies are required to which can address potential bias such as starting point bias and yea-saying bias by CVM. Additional works should be done to show the effect of endogeneity on the WTPs using CVM.

    Moreover, while the total benefit calculated in our study is involved in only one stakeholder, evaluating economic benefits and costs by emphasizing on dual roles in between upstream (supplier and polluter) and downstream (consumers and sufferers) is recommended. In the context of water-food-energy nexus, our analysis obtained by empirical case study could be extended with spatial modeling studies such as inVEST for integrated management approaches. For an integrated approach, the empirical studies applied cognitive psychology in future CVM survey could be recommendable to determine the key cognitive process of respondents in information process.


    This study aims to estimate the willingness to pay (WTP) for the water quality improvement along the To Lich River (TLR) in Vietnam and to identify key determinants affecting the WTP for the water quality improvement. This result indicates that the expected annual mean WTP is VND 202,560 (US $8.6). Our results reveal that respondents who think polluted water could influence badly on their family’s health and themselves have higher tendencies to participate in the program than who do not. Meanwhile, younger people express the higher probability of participation in the program for water quality improvement than older and poorer people. The result implied that higher income residents are more likely to pay for the program for the water quality improvement.

    Institutional and financial supports from the government are required to improve the public health and sustainable water resource management. It might be helpful to raise the people’s environmental concerns, in particular, in relation to water quality pollution. Further researches should be needed targeting to the downstream residents along the TLR basins. This study can contribute to future policy establishment for sustainable water management as suggested in integrated water quality management.

    적 요

    본 연구는 비시장평가법인 조건부가치평가법을 통해 수질개 선을 위한 지불용의액을 추정 하였다. 연구 지역은 심각한 수 질오염으로 직·간접적인 피해를 입고 있는 베트남의 하노이 To Lich 강 유역이며, To Lich 강 주변의 거주민들을 대상으 로 설문조사를 실시하였다. 이중양분선택형 질문을 통해 설문 내용을 디자인하였고, 설문조사를 통해 얻어진 표본으로 이변 량프로빗모형을 사용하였다. 그 결과, 베트남 하노이지역의 거 주민이 수질개선을 위하여 연간 지불용의액은 약 VND 202,560원으로 나타났다. 또한 지불용의액에 영향을 주는 중 요한 변수는 사회경제적 변수로 나이, 소득으로 나타났고, 거 주민의 건강에 대한 관심도 통계적으로 유의한 변수로 나타났 다. 본 연구는 하노이의 To Lich 강에 조사 지역을 한정하여 진행하였기 때문에, 연구 지역 범위를 확대시켜 향후 연구가 진행 되어야 할 것이다. 본 연구의 결과는 지속 가능한 수질 관리의 중요성을 강조하며, 추후 베트남 수질 개선 정책에 환 경적 편익 정보를 제공하는데 도움을 줄 것으로 기대된다.


    본 논문은 제1저자 응웬 티 하이 엔(Nguyen Thi Hai Yen) 의 석사학위 논문을 수정 · 보완한 것임.



    A summary of socio-economic characteristics of sample

    Annual mean and aggregate willingness to pay values

    Estimation results of the bivariate probit model.


    1. Arrow, K. , Solow, R. , R. Portney, P., E. Leamer, E. , Radner, R. ,& Schuman, H. (1993). Report of the NOAA Panel on contingent valuation. Washington D.C: Federal Register.
    2. Beaumais, O. , Briand, A. , Millock, K. , & Nauges, C. (2010). What are Households Willing to Pay for Better Tap Water Quality? A Cross-Country Valuation Study. Université Paris1 Panthéon-Sorbonne (Post-Print and Working Papers).
    3. Bishop, R. , & Romano, D. (1998). Environmental Resource Valuation: Application of the Contingent Valuation Method in Italy. Kluwer Academic Publisher.
    4. Connelly, N. , Knuth, B. , & Kay, D. (2002). Public Support for Ecosystem Restoration in the Hudson River Valley, USA. Environmental management, 467-476.
    5. FAO. (2000). Application of Contingent Valuation Method in Developing Countries: A Survey. Rome: Economic and Social Development Paper 146.
    6. FAO. (2014). The Water-Energy-Food Nexus: A new approach in support of food security and sustainable agriculture. Rome: Food and Agriculture Organization of the United Nations.
    7. Fuhrimann, S. , Nauta, M. , Nguyen, T. T. , Pham, D. P. , Nguyen, V. H. , Utzinger, J. , . . . Vuong, T. (2016). Microbial contamination along the main open wastewater and storm water channel of Hanoi, Vietnam, and potential health risks forurban farmers. Science of the Total Environment 566-567, 1014-1022.
    8. Fuhrimann, S. , Pham, D. P. , Utzinger, J. , Cissé, G. , S. Winkler, M. , Do, T. , & Schindler, C. (2016). Intestinal parasite infections and associated risk factors in communities exposed to wastewater in urban and peri-urban transition zones in Hanoi, Vietnam. Parasites & Vectors, 9:537.
    9. Haab, T. , & McConnell, K. (2002). Valuing Environmental and Natural Resources: The Econometrics of Non-Market Valuation. Edward Elgar Publishing, Inc.
    10. Halkos, G. , & Matsiori, S. (2014). Exploring social attitude and willingness to pay for water resources conservation. Journal of Behavioral and Experimental Economics 49, 54-62.
    11. Hanemann, M. , Loomis, J. , & Kanninen, B. (1991, November). Statistical Efficiency of Double-Bounded Dichotomous Choice Contingent Valuation. American Journal of Agricultural Economics, 73(4), 1255-1263.
    12. Hanemann, W. (1994). Valuing the environment through contingent valuation. Joumal of Economic Perspectives, 19-43.
    13. HSO. (2009). Hanoi statistical yearbook 2008. Hanoi: Hanoi StaValuing Statistical Office.
    14. HSO. (2018). Hanoi statistical yearbook 2017. Hanoi: Hanoi Statistical Office.
    15. Jianjun, J. , Wenyu, W. , Ying, F. , & Xiaomin, W. (2016). Measuring the willingness to pay for drinking waterquality improvements: results of a contingent valuationsurvey in Songzi, China. Journal of Water and Health 14(3).
    16. Mitchell, R. C. , & Carson, R. T. (1989). Using surveys to value public goods: the contingent valuation method. Resources for the Future.
    17. MONRE. (2006). Environment report of Vietnam 2006 The state of water environment in 3 river basins of Cau, Nhue - Day and Dong Nai river system”. The Ministry of Natural Resources and Environment of Viet Nam.
    18. MONRE. (2010). 2010 National State of Environmental Report – An overview of Vietnam’s environment. Hanoi: The Ministry of Natural Resources and Environment of Viet Nam.
    19. Nguyen, T. , Yoneda, M. , & Matsui, Y. (2013). Does embankment improve quality of a river? A case study in To Lich River inner city Hanoi, with special reference to heavy metals. Journal of Environmental Protection 4, 4, 361-370.
    20. Nguyen, T. , Ohtsubo, M. , Li, L. , Higashi, T. , & Kanayama, M. (2008). Assessment of the water quality of two rivers in Hanoi City and its suitability for irrigation water. Paddy and water environment 6, 257-262.
    21. Nguyen, T. , Yoneda, M. , Shimada, Y. , & Matsui, Y. (2015). Assessment of trace metal contamination and exchange between water and sediment systems in the To Lich River in inner Hanoi, Vietnam. Environmental Earth Sciences 73, 3925-3936.
    22. Salvador, d.-S. , & Julián, R.-H. (2012). Estimating the Non-market Benefits of Water Quality Improvement for a Case Study in Spain: A Contingent Valuation Approach. Environmental science and policy 22, 47-59.
    23. Stijn, R. , Jan, V. , Wouter, W. , & Ruerd, R. (2017). Water-foodenergy nexus. Wageningen: Wageningen Economic Research.
    24. Vuong, T. , van der Hoek, W. , Kjær Ersbøll, A. , Nguyen, V. , Nguyen, D. , Phung, D. , & Dalsgaard, A. (2007). Dermatitis among farmers engaged in peri-urban aquatic food production in Hanoi, Vietnam. Tropical Medicine and International Health 12, 59-65.
    25. Waughray, D. (2011). Water security: The water-energy-food-climate nexus. Washington and London: Island Press.