Science of the Total Environment 599–600 (2017) 431–441

Contents lists available at ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Effects of inequality of supply hours on consumers' coping strategies and

perceptions of intermittent water supply in Kathmandu Valley, Nepal
B. Guragai a, S. Takizawa a,c,⁎, T. Hashimoto a, K. Oguma b
a
Department of Urban Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan
b
Research Center for Advanced Science and Technology, The University of Tokyo, Japan
c
Department of Environmental Engineering, Faculty of Engineering, Chiang Mai University, Chiang Mai, Thailand

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Lorenz curves and Gini coefficients were

used to measure inequality of piped
supply.
• Inequality was found both between and
within service areas.
• Coping strategies for reduced supply
hours vary between home owners and
tenants.
• Expectations for improved water quality
are higher than those for supply volume.
• Consumers' perceptions of piped water
services worsen as supply hours
shorten.

a r t i c l e i n f o a b s t r a c t

Article history: To investigate the effects of unequal supply hours on consumers' coping strategies and perceptions of the inter-
Received 26 February 2017 mittent water supply (IWS) in the Kathmandu Valley (KV), Nepal we conducted a randomized household survey
Received in revised form 2 April 2017 (n = 369) and on-site water quality tests. Half of the households received piped water for 6 or fewer hours per
Accepted 23 April 2017
week. To augment or cope with the inadequate supply, 28% of the households used highly contaminated and ex-
Available online 5 May 2017
pensive tanker-delivered water. Half of the piped water samples (n = 13) were contaminated with Escherichia
Editor: D. Barcelo coli. Free chlorine concentration in all piped water samples was below the national standards (0.1–0.2 mg/L),
but combined chlorine was detected at an average of 0.24 mg/L, indicating ingression of contaminants in the net-
Keywords: work. Point-of-use devices could increase access to safe water in the KV from 42% to 80%. The use of Lorenz curves
GINI coefficient and Gini coefficients revealed inequality of piped water supply hours per week both between and within service
Lorenz curves areas in the KV, due mainly to a small percentage of households who receive longer supply hours. To cope with
Supply duration reduced supply hours, home owners pay more to get water from alternative sources, while tenants compromise
Supply frequency their water consumption. Under IWS, expectations for improvements in piped water quality and supply regular-
Supply regularity
ity are higher than those for supply volume. Consumers' perceptions of the piped water services worsen with the
Tanker-delivered water
reduction in supply hours, but perceptions of piped water tariff are independent of supply hours.
© 2017 Elsevier B.V. All rights reserved.

1. Introduction
⁎ Corresponding author at: Department of Urban Engineering, Graduate School of
Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan. The supply of piped water on premises, considered the most im-
E-mail address: takizawa@env.t.u-tokyo.ac.jp (S. Takizawa). proved water source (WHO and UNICEF, 2015), does not always take

http://dx.doi.org/10.1016/j.scitotenv.2017.04.182
0048-9697/© 2017 Elsevier B.V. All rights reserved.
432 B. Guragai et al. / Science of the Total Environment 599–600 (2017) 431–441

into consideration the core values of improved water sources: safety, The objectives of this study were to assess the prevalence of IWS in
adequacy, and reliability (Godfrey et al., 2011; Heitzinger et al., 2015; the KV using supply hours and supply frequency as measures of IWS,
Shaheed et al., 2014). Globally, at least 300 million people receive and to graphically express and quantify the supply inequality of piped
piped water under the condition of intermittent water supply (IWS), a water in the KV. The supply inequality was visualized by Lorenz curve
form of water apportioning (Kumpel and Nelson, 2016). Prior studies and quantified by the Gini coefficient in each service area and in the
have identified several causes of IWS that can be broadly grouped into whole KV. Another objective was to evaluate the effects of supply
two categories: (1) purposely reduced supply hours due to water short- hours and supply inequality under IWS on consumers' coping strategies
ages caused by urban population growth, inefficient water supply sys- and their perceptions of the piped water services.
tems, and shrinkage of water resources, and (2) an effect of power
rationing and financial constraints (Galaitsi et al., 2016; Klingel, 2012; 2. Methodology
Kumpel and Nelson, 2016; Rosenberg et al., 2008; Totsuka et al., 2004).
Access to improved water sources among the urban population in 2.1. Questionnaire survey
Nepal is decreasing (from 97% in 1990 to 91% in 2015), unlike in most
developing countries (WHO and UNICEF, 2015). The Kathmandu Valley A questionnaire survey (Approval No. 28-182, the Ethical Review
(KV) is the biggest urban concentration in Nepal, with a population of Board of the University of Tokyo) was conducted in all ten service
2.5 million (CBS, 2011). Its water supply company, Kathmandu areas of KUKL in November and December 2015. The sample size was
Upatyaka Khanepani Limited (KUKL), is continually struggling to pro- determined by Eq. (1) (Daniel, 1999):
vide safe, reliable, and adequate piped water to its consumers. Piped
water was supplied for 6 h a day in the 1970s (World Bank, 1974). Z 2 P ð1−P Þ 1:962  0:5ð1−0:5Þ
Sample size ¼ ¼ ¼ 384 ð1Þ
Since then, because of the widening gap between KUKL's supply volume d
2
0:052
and the water demand of the increasing population, KUKL has regularly
reduced the supply hours. Currently, intermittency prevails throughout where, Z is the Z statistic for the confidence level (1.96 for 95% confi-
the KV (KUKL, 2016), and the supply–demand gap stands in excess of dence level); P is the expected prevalence (for maximum sample size,
200 MLD (million liters per day) (KUKL, 2015). IWS in the KV will con- P = 50% = 0.5); and d is the precision, taken as 5% (Naing et al.,
tinue at least until 2028, when an additional 510 MLD from the 2006). These values gave a sample size of 384, but unforeseeable con-
Melamchi Water Supply Project will eventually become available straints in the survey meant that only 369 samples could be taken. We
(Hazama Ando, 2015). Water utilities, including KUKL, consider physi- assumed that this number was close enough to 384. Proportional strat-
cal water scarcity as the primary cause of IWS and tend to disregard im- ified random sampling (Nardi, 2013) was used to select samples on the
proving other operations, because little is known about how service basis of the number of pipe connections in each service area (Table 1).
levels can be improved under source water volume constraints The surveyed households were spread widely over the KUKL service
(Galaitsi et al., 2016). Consumers are therefore forced to rely on alterna- areas (Fig. 1), and are assumed to be representative of the total
tive water sources to meet their water demands and to use various cop- population.
ing strategies such as storing water in large tanks and using household The questionnaire included both open and closed questions which
water treatment systems to improve water quality (Katuwal and probed respondents' demographic information, water sources, water
Bohara, 2011; Pattanayak et al., 2005). storage and treatment, and perception of the water services. Table 2
KUKL receives an average of 60 complaints a week, most of them re- shows the specific information collected. A 5-point (Likert-type) scale
lated to inadequate supply, leaks, contamination, and inequity in distri- was used to assess perceptions of the service level (quality of piped
bution (KUKL, 2008). A continued water supply–demand gap will water services: very good, good, neutral, poor, and very poor) and the
exacerbate the intermittency and thus the number of complaints will piped water tariff (too high, high, normal, low, and too low). All ques-
increase. Delays or inaction in addressing complaints devalue consumer tionnaire sheets were coded to maintain anonymity before data
trust in utilities (Celik and Muhammetoglu, 2008; García-Valiñas and analysis.
Miquel-Florensa, 2013; Jain et al., 2014; Roche et al., 2013; Vásquez
et al., 2011). Dietrich et al. (2014) consider consumers real-time sen- 2.2. Statistical analyses
sors, whose perceptions are useful in assessing the service levels of
piped water supply. It was reported that consumers' perceptions are in- We investigated five strategies used by the consumers to cope with
fluenced by the comparative levels of piped water services, system reli- IWS: (1) purchasing tanker-delivered water from vendors, (2) storing
ability, and water quality among nearby houses (Vásquez et al., 2011), water in large tanks, (3) treating piped water with household water
and by their concerns for piped water quality (Celik and treatment systems, (4) pumping low-pressure piped water, and
Muhammetoglu, 2008; García-Valiñas and Miquel-Florensa, 2013; Jain (5) spending money on all sorts of water (Katuwal and Bohara, 2011;
et al., 2014; Vásquez et al., 2011). Moreover, previous investigations re- Pattanayak et al., 2005). We also investigated the effects of IWS on con-
vealed that consumers' subjective perceptions (Doria, 2010; Vásquez, sumers' perceptions of four aspects of the piped water services:
2012), as well as their socioeconomic characteristics (such as income,
education level, and household size), influence the selection of coping Table 1
strategies (García-Rubio et al., 2016; García-Valiñas and Service areas and numbers of households surveyed (Hazama Ando, 2015).
Miquel-Florensa, 2013; Roche et al., 2013; Vásquez, 2012; Vásquez Service area Number of pipe Population Households
et al., 2011). Therefore, we hypothesized that the inequality of supply connections served surveyed
hours under IWS, as well as supply frequency per week, can influence Kamaladi 5940 37,677 9
consumers' coping strategies and their perceptions. However, despite Kirtipur 7629 89,180 19
its prevalence, studies of the effects of IWS on consumers' coping strat- Madhyapur Thimi 7962 83,036 28
egies and perception of piped water are sparse. Existing studies consid- Bhaktapur 10,496 100,308 25
Chhetrapati 13,894 115,458 28
ered only the average supply duration as a measure of IWS and did not
Tripureshwor 20,766 196,612 52
elaborate on the inequality of water supply among consumers (Andey Baneshwar 25,796 281,621 50
and Kelkar, 2009; Fan et al., 2014; Zérah, 2000). This study is the first Maharajgunj 31,204 343,191 42
to our knowledge to visualize the inequality in supply hours caused by Mahankalchaur 31,339 424,228 53
IWS and to reveal the effects of such supply inequality on consumers' Lalitpur 39,692 316,403 63
Total households surveyed 369
coping strategies and perceptions of piped water services.
 B. Guragai et al. / Science of the Total Environment 599–600 (2017) 431–441 433

Fig. 1. Locations of households surveyed (shown in red dots); the yellow region represents the KUKL water supply service area. (Household locations may overlap owing to the small scale
of the map.) Service areas are identified in Supplementary material, Fig. S1.

(1) perceived problems of color, odor, taste, microbial quality, and 2.3. Inequality analysis using Lorenz curve and Gini coefficient
pressure; (2) expected improvements in water quality, pressure,
regularity, frequency, and volume; (3) current piped water service The Lorenz curve and the Gini coefficient are widely used to assess
levels; and (4) piped water tariff. Table 3 summarizes the coping inequality in income and wealth distribution and, recently, in health
strategies, perceptions, and assignment of values to each strategy and the environment (Davies et al., 2011; Pullan et al., 2014; van Zon
and perception. et al., 2015). Nonetheless, we believe that ours is the first study to use
The following statistical analyses were conducted: these indices in assessing the inequality of supply hours under IWS. To
construct Lorenz curves, we ordered the weekly supply hours in each
1. Because the continuous numeric data of storing, pumping and
spending did not follow a normal distribution (Shapiro-Wilk test),
comparison across consumer groups with different supply hours Table 3
were carried out by using the non-parametric Kruskal–Wallis test. Assessment of consumers' coping strategies and perceptions.
Dunn's test was performed as a post hoc test. The data are presented
Explanation Value assigned
as boxplots excluding outliers.
Coping strategy
2. For Likert-type data (service level and piped water tariff) and dis- Purchasing Purchasing tanker-delivered Purchase = 1; no = 0
crete dichotomous data (with values “0” or “1”), the Chi-squared water
Storing Size of basement water Exact numerical values (m3)
test (χ2) was performed following Sullivan and Artino (2013). The
storage tank
discrete dichotomous data were collated in a contingency table and Treating Treating piped water by Treat water = 1; no = 0
are presented as percentage bar charts. household water treatment
systems
All tests were conducted in R v.3.2.3 (R Core Team, 2014) and the re- Pumping Costs of pumping may Exact numerical values (NPR)a
sults were considered significant at p b 0.05 level. The tests used in this increase electricity expenses
study were also used in previous analyses of data related to piped water Spending Monthly household water Exact numerical values (NPR)
expenses
quality and consumers' perceptions (Pasakhala et al., 2013; Wedgworth
et al., 2014). Perceptions
Perceived Color Problem perceived = 1; no = 0
problems Odor
Taste
Table 2 Microbial quality
Outline of the survey questions. Pressure
Expected Quality Improvement expected = 1; no
Question group Data obtained
improvements Pressure =0
Demographic information Household size; home ownership; water and Regularity
electricity expenses Frequency
Water sources and perception Major source by volume, frequency, and Volume
of piped water services duration of piped water supply; supply Service level Perception of the piped water Very good = 5, good = 4,
regularity; perceived service levels and piped supply service level neutral = 3, poor = 2, very
water tariff; perceived problems and expected poor = 1
improvements Piped water Perception of the piped water Too low = 5, low = 4, normal
Water storage and treatment Size and material of storage tanks; methods of tariff tariff = 3, high = 2, too high = 1
water treatment a
NPR = Nepalese rupee; USD 1 = NPR 107.69 (Feb. 2017).
434 B. Guragai et al. / Science of the Total Environment 599–600 (2017) 431–441
service area from smallest to largest, then we plotted the cumulative
percentage of the supply hours against the cumulative percentage of
consumers (Fig. S2a).
The Gini coefficient (G) represents the variation in supply hours
across a population and is calculated as the area between the Lorenz
curve and the line of equality divided by the total area below the line
of equality. As different shapes of Lorenz curves can give equal G values,
the shape of a Lorenz curve is described by using the coefficient of asym-
metry (Lasym) (Damgaard and Weiner, 2000). Lasym = 1 indicates perfect
symmetry, Lasym b 1 indicates that inequality is caused by an abundance
of households receiving shorter supply hours, and Lasym N 1 indicates
that inequality is caused by a small number of households who receive
longer supply hours. We also calculated the Hoover index (also known
as Robin Hood index) to quantify the supply hours needed to be
redistributed in order to achieve equality (Hoover, 1941). The Hoover
index is theoretically equivalent to the maximum vertical distance be-
tween the line of equality and the Lorenz curve (Fig. S2b).
2.4. Water quality tests
Water samples were collected from piped waters on premises (n =
8) and on public stand posts (n = 5), tanker deliveries abstracted from
groundwater (n = 5), natural sources including private dug wells (n =
10) and stone spouts (n = 5), household storage tanks (n = 8), pack-
aged 1-L bottled water of different brands (n = 5) and 20-L water jars
(n = 5), and water treated by different brands of point-of-use (POU) de-
vices (n = 5) that use reverse osmosis, triple ionization, ultraviolet and
ozonation processes either singly or in a combination. Parameters for
water quality analysis were selected from those listed in the Nepal
Drinking Water Quality Standards (NDWQS, 2005) on the basis of sig-
nificance in a developing country's situation. Physicochemical parame-
ters, i.e. pH, electrical conductivity (EC), and total dissolved solids
(TDS) were measured with an MP-6 Portable Meter (Hach, CO, USA).
Turbidity was measured with a 2100 Q Portable Turbidimeter (Hach).
The salicylate method was used to measure ammonia nitrogen with
an NH3-N reagent set in Test ‘N Tube Vials that had a measurement
range of 0.2–2.5 mg/L (Hach). Total iron, total chlorine, free chlorine,
and manganese were measured with a DR-890 Portable Colorimeter
(Hach). For microbiological examination, all samples were placed in
37-mm monitor units (Advantec Toyo Kaisha, Ltd., Japan) and m-
ColiBlue24 broth (Hach) was added on site. The monitor units were in-
cubated at 35 °C ± 0.5 for 24 h; blue colonies were counted as
Escherichia coli (E. coli), and the sum of blue and red colonies was count-
ed as total coliforms.
3. Results and discussion
3.1. Household data and water sources
Three-quarters (74%) of respondents were home owners and 26%
were tenants. The mean household size (persons sharing kitchen and
other resources) was 4.34 persons (range: 1–9), similar to that reported
by the National Population and Housing Census (2011). There was a sig-
nificant difference (t-test, p b 0.05) between the mean household size of
owners (4.6 ± 1.3 persons/family; mean ± SD) and of tenants (3.6 ±
1.2 persons/family; mean ± SD). The mean monthly income of the
households was NPR 38,983. The income of home owners (NPR
41,396) was significantly higher than that of tenants (NPR 31,609) (t-
test, p b 0.05). Mean monthly household expenditure on water was
NPR 811, less than on electricity (NPR 1469). Piped water was the
major water source by volume for 73% of respondents. Other major
water sources included dug wells (39%), tanker-delivered water
(28%), public stand posts (6%), and stone spouts (4%). Most respondents
used multiple water sources to meet their water demands.
We compared our results with those of Pasakhala et al. (2013) on the
assumption that the target populations were the same, although the
sampling frame and methods were different. The major water sources
were significantly different (χ2 test, d. f. = 4, p b 0.05) between the
two studies (Fig. 2). The uses of piped water and public stand posts in
our study were slightly lower (by 4% and 3%) than those of Pasakhala
et al. (2013). It is likely that these small differences are the effect of
the sampling frame, but the rapid population growth of 4% per annum
in the KV (World Bank, 2013) might have overwhelmed the extension
of piped supplies by KUKL. The dependence on dug wells was 4%
lower in our study, probably owing to the decline in well yield
(Pandey et al., 2010) and the deterioration of water quality in the shal-
low aquifers of the KV (Haramoto et al., 2011; Shrestha et al., 2015).
Likewise, the use of stone spouts was 12% lower in our study, in agree-
ment with the prediction made by Dixit and Upadhya (2005) that thou-
sands of KV households could be deprived of using stone spouts as the
sources continue to dry out or to be polluted. In contrast, there was a
considerable increase (by 14%) in dependency on tanker-delivered
water, which is substantially more expensive (NPR 4000 for 10 m3 of
water) than piped water (NPR 150 per 10 m3). The comparison of the
results from the two studies reveals that an increasing percentage of
consumers are dependent on tanker-delivered water as an alternative
to piped water. This trend may continue in the future as the population
will further increase in the KV and the conventional sources will shrink.
Thus, the price of tanker-delivered water may increase adding more
financial burden to the consumers who are dependent on such
source.
3.2. Water quality
Between 20% and 60% of tanker-delivered water samples violated
the NDWQS for turbidity, manganese, and total iron (Table 4). Between
7% and 33% of samples from natural source violated the NDWQS for all
parameters except EC. Poor quality of natural water in KV has already
been reported (Haramoto et al., 2011; Khatiwada et al., 2002; Pant,
2011). Notably, 23% of the piped water samples violated the NDWQS
for pH and manganese, 8% for turbidity, and 15% for ammonia nitrogen,
consistent with values previously reported by Warner et al. (2008). Free
chlorine concentration in all piped water samples was below the
NDWQS (0.1–0.2 mg/L). However, combined chlorine was detected at
an average of 0.24 mg/L, indicating reactions of free chlorine with con-
taminants in polluted groundwater that enters the pipes when pres-
sures are low (Kumpel and Nelson, 2014; Shrestha et al., 2013). The
occasional augmentation of piped water with untreated water might
have also led to low concentrations of free chlorine in the piped water
samples (pers. comm., manager of KUKL). About 13% of household stor-
age tank samples violated the NDWQS for turbidity and ammonia nitro-
gen, while 25% violated the NDWQS for manganese and total iron.
However, none of the POU-treated water samples or packaged water
samples violated the NDWQS for any physicochemical parameters.
The NDWQS require that E. coli should not be present in any samples,
and that no N5% of samples should contain total coliforms. Water sam-
ples from tanker-delivered water and natural sources had the highest
levels of microbial contamination among all water sources in the KV
(Fig. 3). Half of all piped water samples contained E. coli and nearly
80% contained total coliforms. Understandably, 50% of household stor-
age tank samples were contaminated with E. coli, possibly because of
the low frequency of cleaning of tanks (Schafer and Mihelcic, 2012).
The concentrations of E. coli in the household storage tanks were higher
than in the piped water samples, perhaps due to mixing of less contam-
inated piped water with more-contaminated tanker-delivered water in
those tanks. Although water treated by POU devices had the lowest level
of microbial contamination, E. coli was still detected in 1 sample (20% of
samples), possibly owing to poor maintenance or inappropriate design
of the POU device (Do et al., 2014). E. coli was also detected in 30% of
the packaged water samples at N 10 CFU/100 mL. Subedi and Aryal
(2010) and Timilshina et al. (2012) also reported the presence of
E. coli in packaged water in the KV. The contamination of packaged
 B. Guragai et al. / Science of the Total Environment 599–600 (2017) 431–441 435

Fig. 2. Major water sources by volume. (The response rate add up to N100 since some respondents used multiple sources.)

water can be attributed to a lack of regular monitoring and regulations
for private water companies (Halage et al., 2015).
Despite poor quality of piped water, consumers in the KV are forced
to use more contaminated water sources such as dug wells and tanker-
delivered water to augment the piped supply. These results highlight
the need of regular monitoring and regulating the water sources serving
large number of consumers, such as stone spouts and tanker-delivered
water which abstract water from ground water (Fox et al., 2016). Con-
sumers should also be encouraged to test their well water which is at
constant risk of exposure to natural and anthropogenic contamination
(Law et al., 2017). In addition, the poor quality of the piped water
samples presented here shows that piped water is not necessarily
safe (Bain et al., 2012; Godfrey et al., 2011; Heitzinger et al., 2015).
Based on the percentage of water samples from various sources that
complied with NDWQS for E. coli (Fig. 3a) and the percentage of con-
sumers dependent on such sources (Fig. 2), the percentage of con-
sumers with access to safe water in the KV is estimated to be 42%
(Tables S1, S2). However, the use of POU devices can increase access
to up to 80% (Fig. 3a). Matsubara et al. (2015) reached a similar conclu-
sion: POU devices increased access to safe drinking water by 13% in
urban areas and by 20% in the suburbs of Hanoi, Vietnam.
abundance of households who receive shorter piped supply hours. In
the remaining eight service areas, Lasym N 1, implying that the inequality
is caused by a small percentage of households who receive piped water
for longer hours. In contrast to our results, Seekell et al. (2011) revealed
that the inequality of water use between nations around the world
was manifested through a large number of low-water-use countries
(Lasym b 1).
Overall, inequality in supply hours under IWS in the KV is relatively
high, as reflected by its Gini coefficient (G = 0.67). Lasym for the KV was
also N 1 (Fig. 5a), indicating that a small percentage of households in the
KV receive piped water for longer hours, contributing to a high inequal-
ity. The Hoover index for the KV was 0.51, which indicates that to
achieve an equality of supply hours, 51% of the supply hours in the KV
needs to be redistributed from the households with longer supply
hours to those with shorter supply hours. The percentage of supply
hours to be redistributed for Baneshwar is the highest among KUKL ser-
vice areas (59%) and that for Madhyapur Thimi is the lowest (13%)
(Fig. 5b). Overall, these indices (G, Hoover index) can be important
guidelines for water utilities, including KUKL, who are embarking
on future supplements to the IWS.

3.3. Inequality under IWS 3.4. Regularity of IWS

The frequency of the piped water supply varied widely (range: 1–7
times/week) within and between all 10 service areas (Fig. S3). Thus,
we considered the total supply hours per week as the basis for assessing
the inequality of supply hours under IWS. The inequality of supply
hours was the lowest in Madhyapur Thimi (G = 0.21) and the highest
in Baneshwar (G = 0.70). These results are reflected by their respective
Lorenz curves: that of Baneshwar (Fig. 4g) is more bowed out than that
of Madhyapur Thimi (Fig. 4c). The Lasym values for Kirtipur (0.97, Fig. 4b)
and Madhyapur Thimi (0.72, Fig. 4c) were b 1, which indicates that the
inequality in supply hours in these service areas is caused by an
It was found that nearly a half (48%) of households received piped
water irregularly; 35% received it in the morning (before 9 a.m.), 4% dur-
ing the day (between 9 a.m. to 3 p.m.), 7% in the afternoon (between
3 p.m. to 6 p.m.), and 5% at night (after 6 p.m.). Fewer than 1% of house-
holds received piped water regularly in two timings. Potential factors
that can affect regularity of water supply include the availability of
water at the source, staff to operate the valves, and power to operate
the pumps. To cope with the IWS efficiently, households in the KV
need to operate their pumps to get water when piped water supply re-
sumes. However, the irregularity of IWS makes this difficult, and thus

Table 4
Percentage of water samples violating the Nepal Drinking Water Quality Standards.

Test parameter NDWQS Tanker-delivered water Natural sources (n Piped water (n Household storage tanks POU treated water Packaged water (n
(n = 5) = 15) = 13) (n = 8) (n = 5) = 10)

pH 6.5–8.5 0 13 23 0 0 0
EC 1500 μs/cm 0 0 0 0 0 0
TDS 1000 mg/L 0 7 0 0 0 0
Turbidity 5 (10) NTU 20 13 8 13 0 0
Manganese 0.2 mg/L 60 33 23 25 0 0
Ammonia–nitrogen 1.5 mg/L 0 33 15 13 0 0
Free chlorine 0.1–0.2 mg/L NA NA 100 NA NA NA
Total iron 0.3 (3) mg/L 40 33 0 25 0 0

μs/cm: microsiemens per cm; NTU: Nephelometric Turbidity Unit; NA: not applicable; NDWQS values in parenthesis ( ) refers the acceptable values only when an alternative is not
available.
436 B. Guragai et al. / Science of the Total Environment 599–600 (2017) 431–441

Fig. 3. Percentages of samples by concentration of (a) E. coli and (b) total coliforms grouped by water source (CFU: colony forming units). (Note that the number of samples for
microbiological tests differs from that for physicochemical tests.)

households have to depend on alternative water sources (Galaitsi et al.,
2016).
3.5. Consumers' perceptions of water services
When asked about the current piped water service level, only 13% of
the respondents considered it to be good or very good, while 39% con-
sidered it to be poor or very poor. These answers reflect the inequality
of the piped water supply, as described above. Nearly a half (48%) of
the respondents considered service levels to be average, probably be-
cause they have become habituated to the IWS over the decades and
have no experience of other water services. About 55% of respondents
reported problems with microbial quality (in agreement with the
rates of E. coli standard violation; Fig. 3a) and inadequate pressure in
the piped supply. In the KV, diarrhea accounts for 30% of all hospital ad-
missions (Shrestha et al., 2015); and frequent news reports and rumors
of waterborne diseases contribute to consumers' perceptions of micro-
bial contamination in the piped water supply (Roche et al., 2013). In
contrast, consumers are less aware of the problems of color, odor, and
taste, except for stains in clothes caused by the presence of iron and
the occasional strong odor of chlorine in the piped water. This may be
because they perceive these aesthetic water quality less significant
than the health-related water quality parameters such as microbial con-
tamination. In addition, consumers in the KV store piped water and
tanker-delivered water in the same tanks and then treat it before drink-
ing, thus concealing the true organoleptic properties of the piped water
(Alameddine et al., 2016). When asked about the current piped water
tariff, only 27% of the respondents considered it to be high or too high,
while most (68%) of them considered the piped water tariff to be
normal.
3.6. Water storage and treatment
Piped supplies in the KV are not only intermittent but also irregular
and inadequate. Under these circumstances, it is common for house-
holds to store water in large tanks. At least 66% of households use base-
ment tanks, consistent with the 71.8% reported by Yoden (2010). These
tanks have an average capacity of 6.7 ± 4.1 (mean ± SD) m3—much

Fig. 4. Lorenz curves of weekly piped water supply hours of the 10 KUKL service areas.
 B. Guragai et al. / Science of the Total Environment 599–600 (2017) 431–441
Fig. 5. (a) Lorenz curve of weekly piped water supply hours in KV. (b) Percentages of weekly piped water supply hours that need to be redistributed to achieve equality.
larger than the 4.0 m3 reported by Yoden (2010)—and 91% of them are
made from reinforced cement concrete. In addition, at least 75% of
households in the KV have rooftop tanks, 99% of them made of high-
density polyethylene. The rooftop tanks have an average capacity of
1.2 ± 0.6 (mean ± SD) m3, the same as reported by Yoden (2010),
but bigger than those in San Lorenzo, Guatemala (0.8 m3; Vásquez,
2016), where consumers store water also in small barrels and buckets.
The larger average capacity of the basement tanks in our study than re-
ported by Yoden (2010) indicates that to cope with the worsening IWS,
householders are building larger basement tanks, although rooftop tank
sizes are unchanged. It also explains that rooftop tank's role as a storage
is secondary as it is primarily used for a gravity-fed supply. The storage
tanks were cleaned 3 times a year on average, ranging from once a
month to less than once a year. Cleaning a large tank is inconvenient,
and under IWS, limitations in the availability of water reduce the
cleaning frequency further. The longest interruption of the piped
water supply varied widely, with 57 respondents reporting N30 days.
In addition, road repair works and the damages caused by the Great
Nepal Earthquake of 25 April 2015 were major reported causes of longer
interruptions to supply.
About 93% of the households in the KV treat piped water before
drinking, by boiling (16%, n = 51), filtration (22%, n = 70), both (40%,
n = 130), solar water disinfection and/or chlorination (3%, n = 8), or
POU devices (4%, n = 12). Most (n = 7) of POU devices used reverse os-
mosis (RO), with an average installation cost of NPR 27,400 and an an-
nual running cost of NPR 2500. Owing to the poor piped water quality
in the KV, consumers will probably invest more in POU devices in the
future.
Fig. 6. Various modes of intermittent water supply in the Kathmandu Valley. (Note that each dot represents a response and the dots are jittered to prevent overlapped plotting.)
437
3.7. Effects of IWS on consumers' coping strategies and perceptions
3.7.1. Modes of IWS and volume of piped water use
Intermittency is commonly reported in terms of average supply du-
ration (h/day or h/week) (Andey and Kelkar, 2009; Fan et al., 2014;
Kumpel and Nelson, 2016). However, this measure alone is not enough
to describe the situation of IWS in the KV. More than half of the con-
sumers in the KV receive piped water 3 or fewer times/week, for 2 or
fewer h/supply, and for 6 or fewer h/week (Fig. 6). This situation is
worse than those in the KV (Pasakhala et al., 2013), in Amman
(Klassert et al., 2015) and in Indian cities (Zérah, 2000). Hereafter, we
use the term “modes of IWS” to refer to the supply frequency (times/
week), supply duration (h/supply), and total supply (h/week) together.
To investigate the linkages between the volume of piped water use,
consumers' coping strategies, and perceptions under different modes of
IWS, we divided the responses into Group A (N 8 h/week), Group B (4–
8 h/week), and Group C (b 4 h/week), maintaining similar numbers of
responses in each group (Shrestha et al., 2015). Responses without fre-
quency and duration were excluded, leaving 304 responses for further
analyses.
Assuming a water consumption rate of 135 L/capita/day (BIS, 2007),
a household of 4.34 persons (Section 3.1) requires 4.1 m3/week of
water, just under the average supply volume of Group B
(5.7 m3/week, 6.2 h/week of piped supply; Table 5). As the consumers
in Group A receive piped water for N8 h/week, most of them would re-
ceive enough water to meet their water demands. The differences be-
tween Groups B and C are a supply frequency of 1.6 times/week, a
supply duration of 4.4 h/week, and a supply volume of 4.4 m3/week
438 B. Guragai et al. / Science of the Total Environment 599–600 (2017) 431–441
Table 5
Piped water supply of consumer groups (n = 304).
Parameters Mean ± SD values by group
A B
Number of responses (n) 86 101
Average supply frequency (times/week) 5.1 ± 1.9 3.2 ± 1.2
Average supply duration (h/supply) 6.4 ± 6.8 2.2 ± 1.0
Average total supply (h/week) 34.1 ± 46.5 6.2 ± 1.4
Average piped water volume (m3/week) 7.6 ± 7.2a 5.7 ± 3.6
a
Households who received water for longer hours were unable to estimate the water
volume.
(Table 5). Individually, an increase in supply frequency of 1 time/week
would increase the supply volume by 0.79 m3, and an increase in supply
duration of 1 h/supply would increase it by 0.32 m3 (Text S1). These re-
sults suggest that supply frequency is more important under IWS than
supply duration. Assuming the independence of supply frequency and
supply duration, the total increase in supply volume due to an increase
in supply frequency of 1.6 times/week and of duration of 4.4 h/supply
would be only 2.7 m3/week. This is less than the increase in supply vol-
ume (4.4 m3/week) when both supply frequency and duration are in-
creased simultaneously. These results indicate that we need to
evaluate both supply frequency and duration under IWS, which is con-
ventionally evaluated only in terms of average supply duration.
Although households in the same groups received piped water for
nearly the same hours per week, their supply frequency (times/week)
and duration (h/supply) varied extensively (Fig. 7). Most Group C
households received piped water 1 to 3 times/week for 1 to 3 h/supply.
Most Group B households received slightly more: 2 to 4 times/week for
1 to 6 h/supply. In contrast, Group A households showed much wider
variations. Several Group A households (top right) received water unin-
terruptedly, as they lie close to the treatment plants. Another group
(bottom right) received piped water briefly but every day. The remain-
ing households (scattered in the middle) showed no distinct character-
istics and differed greatly in supply frequency and duration. These
results suggest that even with the same total supply hours per week,
the supply frequency (times/week) and duration (h/supply) vary
throughout the KV. Kumpel and Nelson (2016) likewise reported that
modes of IWS varied widely in both time and space.
Fig. 7. Frequency versus duration of piped water supply. Blue dashed lines indicate the
average weekly supply hours of each group.
Group A consumers were dominant in districts best serviced by
major water treatment plants (Fig. 8). Group B consumers were domi-
nant in the city center. Much of the Group C consumers live in the sub-
C
urbs, where it is difficult to get reliable piped water.
117
3.7.2. Consumers' coping strategies
1.6 ± 1.0
1.2 ± 0.7 The dependence on tanker-delivered water increased significantly
1.8 ± 0.9 as supply hours decreased: from 8% in Group A to 20% in Group B to
1.3 ± 1.0 51% in Group C (Fig. 9a). Tanker-delivered water is expensive yet conve-
nient for consumers as it is easier and quicker than other sources, and
water can be bought in bulk. Given that 48% of the intermittent supplies
in the KV are irregular and are occasionally interrupted for long periods
(Sections 3.4 and 3.6), the volume of water that can be collected is un-
certain; thus, consumers in all groups depend on tanker-delivered
water. The size of basement tanks showed a unique trend and was sig-
nificantly different among groups (Fig. 9b): Group B consumers, who
are dominant in the city center (Fig. 8), had larger basement tanks
than Groups A and C, probably to meet the demands of a significantly
larger family size (4.9 persons) than Groups A and C (4.3 and 3.7 per-
sons, respectively; Kruskal–Wallis test, Dunn's post hoc, p b 0.05). The
cleaning frequency, however, decreased significantly as supply hours
decreased (Kruskal–Wallis test, p b 0.05). Worsening intermittency,
which is associated with reduced water quality (Shrestha et al., 2013;
Yassin et al., 2006), requires more frequent cleaning of tanks, but it
also reduces water availability; hence, consumers are unable to clean
the tanks as often as required. Thus, N 85% of consumers in all groups
treated piped water (Fig. 9c), which reflects consumers' strong percep-
tions of the poor microbial quality of piped water.
Further, electricity expenses did not differ significantly among the
groups (Fig. 9d), which indicates that differences in supply hours do
not affect pumping costs. In contrast, water expenses differed signifi-
cantly among groups (Fig. 9e): as supply hours decreased from Group
A to Group B to Group C, water expenses increased, probably owing to
the cost of alternative water sources. A breakdown of water expenses
between home owners and tenants reveals a significant increase in
water expenses of home owners from Group A to Group C (Kruskal–
Wallis test, Dunn's post hoc, p b 0.05), but a decreasing trend for tenants
(Fig. 9f). These opposite trends suggest another kind of inequality in
coping strategies: as IWS worsens, home owners pay more for water,
while tenants compromise their water consumption or fetch water
from natural sources. In the KV, home owners and tenants live in the
Fig. 8. Major consumer group in each service area and locations of major water treatment
plants.
 B. Guragai et al. / Science of the Total Environment 599–600 (2017) 431–441 439

Fig. 9. Consumers' strategies to cope with IWS. Box plots (b, d, e, f) compare medians; * #, significantly different at p b 0.05 by Dunn's post hoc test. Bar charts (a, c) compare proportions; *,
significantly different at p b 0.05 by χ2 test. (f) Breakdown of e by home ownership.

same house, and home owners regulate the water in the house by con-
trolling the valves (Pasakhala et al., 2013; Yoden, 2010). This is more
likely to happen when home owners must purchase expensive tanker-
delivered water.
3.7.3. Consumers' perceptions
Significantly fewer Group A consumers reported problems with
color, odor, microbial quality, and pressure than Groups B and C
(Fig. 10a). Despite receiving piped water for longer hours, Group B per-
ceived more problems of color, odor, taste, and pressure than Group C,
probably because Group C consumers use more tanker-delivered
water than piped water. Another reason may be that there is no water
treatment plant in service areas where Group B is dominant (Fig. 8).
Previous studies have shown evidences of increase in contamination
of IWS samples with the increase in distance from the treatment plants
(Kumpel and Nelson, 2013; Lavanya and Ravichandran, 2013; Tokajian
and Hashwa, 2003). Microbial contamination was the most frequently
reported problem, followed by low pressure, followed last by color,
odor, and taste problems (Section 3.5). Significant differences in con-
sumers' perceptions of service levels among the groups indicate that
consumers' perceptions of service levels are based on supply hours.
However, the perception of piped water tariff did not differ significantly
among groups, and so is independent of supply hours. Vásquez (2012)
reported that consumers compare service levels with their neighbors.
This is quite likely to happen in the KV, where piped water supply varies
greatly. However, other factors might influence consumers' perceptions

Fig. 10. Consumers' perceptions of problems and expected improvements in the piped supply. Bars show proportions of each group. *Significant difference at p b 0.05 by χ2 test.
440 B. Guragai et al. / Science of the Total Environment 599–600 (2017) 431–441
of piped water tariff more, such as the costs of other utilities, which are
higher than the water expenses in the KV (Section 3.1). The perceptions
of the piped water tariff (Section 3.5) as normal suggest consumers' la-
tent willingness to pay for any increase in the piped water tariff if KUKL
improved water quality, supply pressure, and regularity.
Improvement in water quality was the most important expectation
for N75% of all consumers (Fig. 10b). Prouty and Zhang (2016) found
similar strong expectations of improvement in water quality in
Uganda, where piped water quality is as poor as that in the KV. The per-
centage of consumers expecting improvements in pressure was much
smaller in Groups B and C than the percentage reporting pressure prob-
lems, probably because most households already use pumps to with-
draw piped water, albeit not always strong enough. However, there
were differences among the consumer groups in expected improve-
ments in supply regularity, frequency, and volume. For instance, over
a third of Group A expected improvements in supply regularity, com-
pared with nearly two-thirds of Groups B and C. Almost twice as
many consumers in Group C (who received piped supply only 1.6
times a week) expected improvements in frequency as in Groups A
and B (3.2 and 5.1 times a week, respectively). Despite receiving piped
water for longer hours than Group C, an almost equal percentage of
consumers in Group B as in Group C (ca. 65%) expected improve-
ments in water volume, compared with fewer than 45% of Group A.
Overall, higher percentages of consumers in all three groups expected
improvements in piped water quality than in any other parameters.
These results suggest that, even under serious IWS, improvements in
water quality and supply regularity may reduce consumers' dissatisfac-
tion. Other factors, such as supply frequency and volume, also influence
consumers' perceptions, although these are difficult to improve while
available water resources remain constrained.
4. Conclusions
We found that the IWS situation in the KV is serious and half of the
consumers receive piped water for 3 or fewer times/week, for 2 or
fewer h/supply, and for 6 or fewer h/week. To augment or cope with
the inadequate piped supply, consumers depend increasingly on alter-
native sources such as tanker-delivered water, which was the major
source for 28% of consumers.
Piped water violated national standards for iron, manganese, turbid-
ity, ammonia and E. coli. However, it was comparatively better than
other water sources such as dug wells and tanker-delivered water.
Treatment of water with point-of-use devices could almost double
access to safe water from the current 42% to 80%.
Inequality in piped water supply hours was prevalent in all KUKL
service areas (G = 0.21–0.70), and over the whole KV (G = 0.67). Over-
all inequality in the KV was due mainly to a small percentage of house-
holds which received piped water for longer hours per week. To achieve
equality under IWS, 51% of the piped supply hours in the KV need to be
redistributed.
To cope with the worsening IWS, home owners have to pay more to
get adequate water, whereas tenants have to compromise their water
consumption. Consumers perceive microbial contamination as the
most serious problem. They expect improvements principally in water
quality, followed by supply regularity, frequency, and volume, but not
so much in supply pressure. Consumers' perceptions of the piped
water service levels gradually worsen as supply hour decrease, but
their perception of the piped water tariff is not influenced by supply
hours.
Although it may take a long time to secure sufficient volumes of
water by developing water resources, the results of this study implicat-
ed that water utilities practicing IWS could improve their service levels
and consumers' perception by two measures: 1) improving water qual-
ity, and 2) maintaining equality in supply hours and supply frequency
among the consumers in their service areas. The GINI coefficient and
the Hoover index are useful parameters to measure inequality in
water supply hours and percentages of supply hours that should be
re-distributed to achieve equality, respectively.
Acknowledgments
This study was supported by a Grant-in-Aid for Scientific Research
(No. 26303013) by the Japan Society for the Promotion of Science
(JSPS). The authors thank Dr. Nawaraj Khatiwada, associate professor
at Kathmandu University, for his suggestions on the conduct of the sur-
vey. The authors gratefully acknowledge the help of Mr. Pratik Karki,
Ms. Sheila Dangol and Mr. Prabhat Joshi in conducting the survey. Spe-
cial thanks go to Mr. Tilak Mohan Bhandari, manager at the Kathmandu
Upatyaka Khanepani Limited (KUKL) and Mr. Kamal Prasad Guragai,
sub-engineer at the District Technical Office, Kathmandu.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.scitotenv.2017.04.182.
References
Alameddine, I., Jawhari, G., El-Fadel, M., 2016. Social perception of public water supply
network and groundwater quality in an urban setting facing saltwater intrusion
and water shortages. Environ. Manag. http://dx.doi.org/10.1007/s00267-016-0803-2.
Andey, S.P., Kelkar, P.S., 2009. Influence of intermittent and continuous modes of water
supply on domestic water consumption. Water Resour. Manag. 23 (12), 2555–2566.
Bain, R.E.S., Gundry, S.W., Wright, J.A., Yang, H., Pedley, S., Bartram, J.K., 2012. Accounting
for water quality in monitoring access to safe drinking-water as part of the Millenni-
um Development Goals: lessons from five countries. Bull. World Health Organ. 90 (3),
228–235A.
BIS, 2007. Bureau of Indian Standards (BIS) Code of Basic Requirements for Water Supply,
Drainage and Sanitation (Fourth Revision). UDC 628.1/.3: 006.7.
CBS (2011), Central Bureau of Statistics, Kathmandu, Nepal.
Celik, E., Muhammetoglu, H., 2008. Improving public perception of tap water in Antalya
city, Turkey. J. Water Supply Res. Technol. AQUA 57 (2), 109–113.
Damgaard, C., Weiner, J., 2000. Describing inequality in plant size or fecundity. Ecology 81
(4), 1139–1142.
Daniel, W., 1999. Biostatistics: A Foundation for Analysis in the Health Sciences. seventh
ed. John Wiley & Sons, New York.
Davies, J.B., Sandström, S., Shorrocks, A.B., Wolff, E.N., 2011. The level and distribution of
global household wealth. Econ. J. 121:223–254. http://dx.doi.org/10.1111/j.1468-
0297.2010.02391.x.
Dietrich, A.M., Phetxumphou, K., Gallagher, D.L., 2014. Systematic tracking, visualizing,
and interpreting of consumer feedback for drinking water quality. Water Res. 66,
63–74.
Dixit, A., Upadhya, M., 2005. Augmenting Groundwater in Kathmandu Valley: challenges and
possibilities. http://r4d.dfid.gov.uk/pdf/outputs/water/r8169-nepalpaperjan05final.pdf
(last accessed 15 Feb 2017).
Do, A.T., Kuroda, K., Hayashi, T., Nga, T.T.V., Oguma, K., Takizawa, S., 2014. Household sur-
vey of installation and treatment efficiency of point-of-use water treatment systems
in Hanoi, Vietnam. J. Water Supply Res. Technol. AQUA 63 (2), 154–161.
Doria, M.F., 2010. Factors influencing public perception of drinking water quality. Water
Policy 12 (1), 1–19.
Fan, L., Wang, F., Liu, G., Yang, X., Qin, W., 2014. Public perception of water consumption
and its effects on water consumption behavior. Water 6 (6), 1771–1784.
Fox, M.A., Nachmanm, K.E., Anderson, B., Lam, J., Resnick, B., 2016. Meeting the public
health challenge of protecting private wells: Proceedings and recommendations
from an expert panel workshop. Sci. Total Environ. 554-555 (2016), 113–118.
Galaitsi, S.E., Russell, R., Bishara, A., Durant, J.L., Bogle, J., Huber-Lee, A., 2016. Intermittent
domestic water supply: a critical review and analysis of causal-consequential
pathways. Water 8 (7), 274–299.
García-Rubio, M.A., Tortajada, C., González-Gómez, F., 2016. Privatising water utilities and
user perception of tap water quality: evidence from Spanish urban water services.
Water Resour. Manag. 30 (1), 315–329.
García-Valiñas, M.Á., Miquel-Florensa, J., 2013. Water services quality in Tanzania: access
and management. Int. J. Water Resour. Dev. 29 (3), 451–471.
Godfrey, S., Labhasetwar, P., Wate, S., Pimpalkar, S., 2011. How safe are the global water
coverage figures? Case study from Madhya Pradesh, India. Environ. Monit. Assess.
176 (1), 561–574.
Halage, A.A., Ssemugabo, C., Ssemwanga, D.K., Musoke, D., Mugambe, R.K., Guwatudde, D.,
Ssempebwa, J.C., 2015. Bacteriological and physical quality of locally packaged drink-
ing water in Kampala, Uganda. J. Environ. Public Health. http://dx.doi.org/10.1155/
2015/942928 (Article ID 942928).
Haramoto, E., Yamada, K., Nishida, K., 2011. Prevalence of protozoa, viruses, coliphages
and indicator bacteria in groundwater and river water in the Kathmandu Valley,
Nepal. Trans. R. Soc. Trop. Med. Hyg. 105 (12), 711–716.
Hazama Ando Corporation, Japan, 2015. Bhaktapur Water Supply Improvement Project
report. http://www.mhlw.go.jp/file/06-Seisakujouhou-10900000-Kenkoukyoku/
0000092634.pdf (last accessed 15 Feb 2017).
Heitzinger, K., Rocha, C.A., Quick, R.E., Montano Jr., S.M., Tilley, D.H., Mock, C.N., Carrasco,
A.J., Cabrera, R.M., Hawes, S.E., 2015. “Improved” but not necessarily safe: an assess-
ment of fecal contamination of household drinking water in rural Peru. American
 B. Guragai et al. / Science of the Total Environment 599–600 (2017) 431–441
Journal of Tropical Medicine and Hygiene]–>Am.J.Trop. Med. Hyg. 93 (3),
501–508.
Hoover, E.M., 1941. Interstate redistribution of population, 1850–1941. J. Econ. Hist. 1 (2),
199–205.
Jain, M., Lim, Y., Arce-Nazario, J.A., Uriarte, M., 2014. Perceptional and socio-demographic
factors associated with household drinking water management strategies in rural
Puerto Rico. PLoS ONE 9 (2), e88059.
Katuwal, H., Bohara, A.K., 2011. Coping with poor water supplies: empirical evidence from
Kathmandu, Nepal. J. Water Health 9 (1), 143–158.
Khatiwada, N.R., Takizawa, S., Tran, T.V., Inoue, M., 2002. Groundwater contamination
assessment for sustainable water supply in Kathmandu Valley, Nepal. Water Sci.
Technol. 46 (9), 147–154.
Klassert, C., Sigel, K., Gawel, E., Klauer, B., 2015. Modeling residential water consumption
in Amman: the role of intermittency, storage and pricing for piped and tanker water.
Water 7 (7), 3643–3670 (2015).
Klingel, P., 2012. Technical causes and impacts of intermittent water distribution. Water
Sci. Technol. Water Supply 12 (4), 504–512.
KUKL, 2008. Kathmandu Upatyaka Khanepani Limited. Annual report, 2065.Available at.
http://www.kathmanduwater.org/reports/annual_report.php (last accessed 15 Feb
2017).
KUKL, 2015. Kathmandu Upatyaka Khanepani Limited. Kathmandu, Nepal. Annual Report,
2072.
KUKL, 2016. Kathmandu Upatyaka Khanepani Limited, Kathmandu, Nepal. Water delivery
schedule.Available:. http://www.kathmanduwater.org/home/water_schedule.php
(last accessed 15 Feb 2017, in Nepali).
Kumpel, E., Nelson, K.L., 2013. Comparing microbial water quality in an intermittent and
continuous piped water supply. Water Res. 47 (14), 5176–5188.
Kumpel, E., Nelson, K.L., 2014. Mechanisms affecting water quality in an intermittent
water supply. Environ. Sci. Technol. 48 (5), 2766–2775.
Kumpel, E., Nelson, K.L., 2016. Intermittent water supply: prevalence, practice, and micro-
bial quality. Environ. Sci. Technol. 50 (2), 542–553.
Lavanya, V., Ravichandran, S., 2013. Microbial contamination of drinking water at the
source and household storage level in the peri-urban area of southern Chennai and
its implication on health, India. J. Public Health 21:481. http://dx.doi.org/10.1007/
s10389-013-0573-8.
Law, R.K., Murphy, M.W., Choudhary, E., 2017. Private well groundwater quality in West
Virginia, USA-2010. Sci. Total Environ. 586 (2017), 559–565.
Matsubara, K., Do, A.T., Kuroda, K., Nga, T., Takizawa, S., 2015. Estimation of the access to
safe drinking water sources and improvement by household water treatment in
Hanoi city, Vietnam. J. Jpn. Soc. Civ. Eng., Ser. G (Environ. Res.) 71 (7), III_69–III_78
(in Japanese).
Naing, L., Winn, T., Rusli, B.N., 2006. Practical issues in calculating the sample size for
prevalence studies. Arch. Orofac. Sci. 2006 (1), 9–14.
Nardi, P., 2013. Doing Survey Research. A Guide to Quantitative Methods. third ed.
Paradigm, Boulder, CO.
National Population and Housing Census, 2011. Central Bureau of Statistics, Nepal. Available
at http://cbs.gov.np/image/data/population/national%20report/national%20report.pdf
(last accessed 15 Feb 2017).
NDWQS. Nepal Drinking Water Quality Standards, 2005. Government of Nepal.Available
at. http://mowss.gov.np/assets/uploads/files/NDWQS_2005_Nepal.pdf (last accessed
15 Feb 2017).
Pandey, V.P., Chapagain, S.K., Kazama, F., 2010. Evaluation of groundwater environment of
Kathmandu Valley. Environ. Earth Sci. 60 (6), 1329–1342.
Pant, B.R., 2011. Groundwater quality in the Kathmandu valley of Nepal. Environ. Monit.
Assess. 178 (1–4), 477–485.
Pasakhala, B., Harada, H., Fujii, S., Tanaka, S., Shivakoti, B.R., Shrestha, S., 2013. Household
coping measures with water scarcity: a case study in Kathmandu, Nepal. J. Jpn. Soc.
Civ. Eng. 69 (7), III_73–III_81.
Pattanayak, S., Yang, J., Whittington, D., B.K.K.C., 2005. Coping with unreliable public
water supplies: averting expenditures by households in Kathmandu, Nepal. Water
Resour. Res. 41:W02012. http://dx.doi.org/10.1029/2003WR002443.
Prouty, C., Zhang, Q., 2016. How do people's perceptions of water quality influence the life
cycle environmental impacts of drinking water in Uganda? Resour. Conserv. Recycl.
109 (2016), 24–33.
Pullan, R.L., Freeman, M.C., Gething, P.W., Brooker, S.J., 2014. Geographical inequalities in
use of improved drinking water supply and sanitation across Sub-Saharan Africa:
mapping and spatial analysis of cross-sectional survey data. PLoS Med. 11 (4):
e1001626. http://dx.doi.org/10.1371/journal.pmed.1001626.
441
R Core Team, 2014. R: A Language and Environment for Statistical Computing. R Founda-
tion for Statistical Computing, Vienna Available from. http://www.r-project.org/.
Roche, S.M., Jones-Bitton, A., Majowicz, S.E., Pintar, K.D.M., Allison, D., 2013. Investigating
public perceptions and knowledge translation priorities to improve water safety for
residents with private water supplies: a cross-sectional study in Newfoundland and
Labrador. BMC Public Health 13, 1225 (2013).
Rosenberg, D.E., Talozi, S., Lund, J., 2008. Intermittent water supplies: challenges and
opportunities for residential water users in Jordan. Water Int. 33 (4), 488–504.
Schafer, C.A., Mihelcic, J.R., 2012. Effect of storage tank material and maintenance on
household water quality. American Water Works Association]–>J. Am. Water
Works Assoc. 104 (9), E521–E529.
Seekell, D.A., D'Odorico, P., Pace, M.L., 2011. Virtual water transfers unlikely to redress
inequality in global water use. Environ. Res. Lett. 6 (2), 024017.
Shaheed, A., Orgill, J., Montgomery, M.A., Jeuland, M.A., Brown, J., 2014. Why “improved”
water sources are not always safe. Bull. World Health Organ. 92 (4), 283–289.
Shrestha, S., Malla, S.S., Aihara, Y., Kondo, N., Nishida, K., 2013. Water quality at supply
source and point of use in Kathmandu Valley. J. Water Environ. Tech. 11 (4), 331–340.
Shrestha, S., Haramoto, E., Malla, R., Nishida, K., 2015. Risk of diarrhea from shallow
groundwater contaminated with entero-pathogens in the Kathmandu Valley,
Nepal. J. Water Health 13 (1), 259–269.
Subedi, M., Aryal, M., 2010. Public perception about drinking jar water and its bacteriological
analysis. Nepal Med. Coll. J. 12 (2), 110–114.
Sullivan, G.M., Artino Jr., A.R., 2013. Analyzing and interpreting data from Likert-type
scales. J. Grad. Med. Educ. http://dx.doi.org/10.4300/JGME-5-4-18.
Timilshina, M., Dahal, I., Thapa, B., 2012. Microbial assessment of bottled drinking water of
Kathmandu valley. Int. J. Infect. Microbiol. 1 (2), 84–86.
Tokajian, S., Hashwa, F., 2003. Water quality problems associated with intermittent water
supply. Water Sci. Technol. 47 (3), 229–234.
Totsuka, N., Trifunovic, N., Vairavamoorthy, K., 2004. Intermittent urban water supply
under water starving situations. 30th WEDC International Conference, Vientiane,
Lao PDR, 2004.
Vásquez, W.F., 2012. Reliability perceptions and water storage expenditures: evidence
from Nicaragua. Water Resour. Res. 48 (10), W10547.
Vásquez, W.F., 2016. An empirical analysis of household choices among water storage
devices. Water Resour. Rural Dev. 8 (2016), 12–24.
Vásquez, W.F., Trudeau, J., Franceschi, D., 2011. Can user perception influence the quality
of water services? Evidences from León, Nicaragua. Int. Rev. Adm. Sci. 77 (3),
481–503.
Warner, N.R., Levy, J., Harpp, K., Farruggia, F., 2008. Drinking water quality in Nepal's
Kathmandu Valley: a survey and assessment of selected controlling site characteris-
tics. Hydrogeol. J. 16 (2), 321–334.
Wedgworth, J.C., Brown, J., Johnson, P., Olson, J.B., Elliott, M., Forehand, R., Stauber, C.E.,
2014. Associations between perceptions of drinking water service delivery and mea-
sured drinking water quality in rural Alabama. Int. J. Environ. Res. Public Health 11
(7):7376–7392. http://dx.doi.org/10.3390/ijerph110707376.
WHO, UNICEF, 2015. 25 years Progress on Water Supply and Sanitation and Drinking
Water 2015 Update and MDG Assessment Report. WHO/UNICEF, Geneva,
Switzerland.
World Bank, 1974. Water Supply and Sewerage Project, Nepal. World Bank, Washington,
DC http://documents.worldbank.org/curated/en/960401468291690797/pdf/multi-
page.pdf (last accessed 15 Feb 2017).
World Bank, 2013. Feature story: Managing Nepal's urban transition.Available at. http://
www.worldbank.org/en/news/feature/2013/04/01/managing-nepals-urban-
transition (last accessed 15 Feb 2017).
Yassin, M.M., Amr, S.S., Al-Najar, H.M., 2006. Assessment of microbiological water quality
and its relation to human health in Gaza Governorate, Gaza Strip. Public Health 120
(12), 1177–1187.
Yoden, K., 2010. Analysis of Domestic Water Use: A Case Study of Kathmandu, Nepal.
University of Tokyo, Japan (Ph.D. Thesis).
Zérah, M.H., 2000. Household strategies for coping with unreliable water supplies: the
case of Delhi. Habitat Int. 24 (3), 295–307.
van Zon, S.K.R., Bültmann, U., Mendes de Leon, C.F., Reijneveld, S.A., 2015. Absolute and
relative socioeconomic health inequalities across age groups. PLoS ONE 10 (12):
e0145947. http://dx.doi.org/10.1371/journal.pone.0145947.