Kuwait J. Sci. Eng.

30(1) 2003

Household demand for water: A case study in Kuwait

JASEM M. AL-HUMOUD AND IBRAHIM AL-GHUSAIN

Kuwait University, Department of Civil Engineering
P.O.Box 5969, Safat 13060, Kuwait, Tel: (+965) 967-4825; Fax: (+965) 481-7524
E-mail: jasem@civil.kuniv.edu.kw

ABSTRACT
This paper reports the results of a research study undertaken to determine water
consumption in Kuwait. The consumption and other related data were collected
randomly from di€erent households within Kuwait. In addition, the paper gives
forecasts for water consumption (demand) in Kuwait. The government of Kuwait
heavily subsidizes water production. These subsidies add a heavy burden to the current
®scal de®cit. Results of the analyses indicated that there is considerable waste of fresh
water by the average Kuwaiti household. Predictions of future water demand included in
this study may help the government in its e€orts to reform its subsidy policy. The paper
concludes with useful remarks to both the water authority (Ministry of Electricity and
Water) and the citizens of Kuwait on water usage rationalization.

Keywords: Forecasting; Households; Kuwait; Water Consumption.

INTRODUCTION
Signs of water scarcity worldwide are numerous. Water tables are falling, lakes
are shrinking, rivers are becoming more polluted and wetlands are disappearing.
Engineering solutions to water shortages include schemes such as building river
diversions (e.g. in California, U.S.A), and the construction of dams and other
expensive projects, some with questionable environmental consequences (Postel
1992).
Since 1950, global water demand has more than tripled to an estimated
annual level of 4,340 km3 in 1993 (Shiklomanov 1990). Because of improved
standards of living, the per capita water use worldwide has also increased
dramatically to 800 m3 per year. This amount is 50% more than what it was in
1950, and it continues to grow. To meet this rising demand, responsible policy-
makers have mainly emphasized the construction of "water development"
projects, particularly dams and river diversions. Today, more than 36,000 large
dams have been built around the world to control ¯oods and provide energy,
irrigation, industrial supplies, and drinking water to a growing global
population (Falkenmark 1989 and PRB 1993).
198 Jasem M. Al-Humoud and Ibrahim Al-Ghusain

The problem of securing reliable water supplies has been an issue of great
importance to the people of Kuwait since the earliest days of settlement in the
region. Today, as the population and industrial developments grow, the water
supply issue promises to become more acute.

Covering an area of 17,818 km2, Kuwait lies in the northwestern corner of the
Arabian Gulf. The Kuwait mainland extends between latitudes 288C 300 and 308
060 north, longitudes 468 300 and 488 300 east, and measures about 200 km from
the most northern point to the most southern points, and about 170 km from
east to west between the longitude extremes. The weather of the country is
typical of a desert (arid) region and receives only about 100 mm (3.94 in) of rain
annually. This precipitation is highly variable from year to year (40 - 240 mm)
and ground water is scarce, complicating water development e€orts (MOP,
2001). In the past the people of Kuwait relied on a scant number of wells to
satisfy their water needs. Those wells, supplemented with fresh water
transported by boats from Basra, Iraq, were the main source of water supply to
the people. Transporting water by boats continued for some time, and in 1939 a
company was established to manage the ¯eet of water carries from Iraq at the
same time constructing three reservoirs on the shore for storage. The ®rst major
breakthrough in the supply & water came in 1951 when Kuwait Oil Company
(K.O.C) built a small sea water desalination plant with a capacity of 80,000
gallon per day at the port of Al-Ahmadi (Mina Al-Ahmadi), and distributed
part of the water to the town of Kuwait. The ®rst major desalination plant was
built in 1953 with a capacity of 1 million gallons per day (mgpd). In 1978,
another desalination plant was build in Doha with a capacity of 42 mgpd
(MEW 2001 and MOP 2000).

The country was and is very anxious to exploit all available groundwater,
both freshwater for drinking and brackish water for irrigation. Fresh
groundwater is considered a matter of prime importance. Fresh groundwater
was discovered in limited quantities at both Al-Rawdhatain and Umm Al-Aish
®elds. Pumping operations commenced in 1962, with the estimated natural
reserve of bothy ®elds about 40,000 million gallons. In 1980, the Rawdhatain
Water Production and Bottling Project started to produce 1800 m3/year of
mineral water. The Umm Al-Aish ®eld is currently producing 8000 m3/year of
water (MEW 20001, Omar et al. 1996).

Along with fresh water, the country makes use of its large supply of brackish
groundwater (1200 - 7000 ppm). The Ministry of Electricity and Water (MEW)
distributes this water to the consumers through separate networks parallel to the
fresh water network. The brackish water is intended to be used for various
purposes such as blending with distilled water, irrigation, livestock watering and
construction works. Fresh and brackish water are stored by two di€erent means
 Household demand for water 199

in Kuwait; one is by ground reservoirs, and the other is by elevated tanks

(towers). The present storage capacities of the di€erent reservoirs for both fresh
and brackish water are presented in Table 1. The consumption of fresh and
brackish water has been increasing with the increase of population.

Table 1. Types and capacities of storage reservoirs for fresh and brackish water

Type of Ground Reservoirs Elevated Tanks (Towers)

Water Number Capacity (MIG) Number Capacity (MIG)
Fresh 64 2143 39 25
Brackish 26 498.8 15 9.6

As can be noted in Fig. 1, the average consumption of water has increased

gradually, with the exception of the years 1990, 1991 and 1992. This lack of
increase is explained by the invasion of Iraq and the start of the Gulf War.
During that period, water production was interrupted tremendously. In
addition, the population was at it lowest levels, due to the large exodus of
people from the country. Figure 2 shows the per capita consumption is
increasing yearly as well, with a slight drop during 1998. The data for the years
1990 and 1991 is not available because of the Gulf War. By the end of the year
2000, the total length of the entire water network was 131,45 Km. It is worth
mentioning that the number of consumers connected to the fresh water network
totaled 111,581 by the year 2000, while consumers connected to the brackish
water network totaled 67,657 by the end of the same year (MEW, 2001).

Fig. 1. Fresh Water Consumption (1970-2000)

200 Jasem M. Al-Humoud and Ibrahim Al-Ghusain

Fig. 2. Per capita consumption of fresh water (1975-2000)

Another point to add is that in the a‚uent State of Kuwait, nearly 98 percent
of the population reside in Metropolitan Kuwait. A unique characteristic of the
State of Kuwait and other nations of the Gulf Cooperation Council (GCC) is
the constant change in the expatriate population from year to year. Each year, a
signi®cant percentage of Kuwait's expatriate population (62% of Kuwait's
population are non-Kuwaitis) leave the State upon termination of their
contracts. These individuals are replaced by incoming expatriate workers,
mostly from Egypt, India, Bangladesh, and Sri Lanka.

THE STUDY
Following a comprehensive review of the related literature and the
establishment of organizational structure for the implementation of the
research, a survey plan was designed. The computation of the sample size, the
development of a questionnaire, the determination of sampling technique and
sample population were all addressed in the survey design. The computation of
the sample size was made in accordance with the commonly utilized statistical
equation (Walpole and Myers, 1985). A con®dence level of 95%, and an error
level of +5% were considered to be appropriate by the research team. By
utilizing a recommended value of 0.5 for the standard deviation (for maximum
possible standard error of the mean), the minimum required sample size may be
computed from the following equation:
p
N ˆ ‰…z1 =2  S†=eŠ …1†
 Household demand for water 201

where:
N = the minimum required sample size.
Z = the number of units of the standard deviation in a standard normal
distribution curve (Z = 1.96%,  = 5%).
 = signi®cance level.
S = standard deviation (0.5 when the true population standard deviation is
not known).
e = acceptable error (+5%).
Equation 1 and the above input values result in a minimum sample size of 385.
A simple yet structured questionnaire was designed and pre-tested to obtain
information on households' socio-economic and residence characteristics, as
well as the daily water-consuming activities of sample households. The
questionnaire, after being pre-tested and modi®ed, was distributed to a
systematic random sample of 3000 households residing in the ®ve governorates
of Metropolitan Kuwait. Special care was taken to ensure the representativeness
of various socio-economic groups in the sample. The random sample of 3000
households were interviewed in person. The objective of the study was to
examine the water supply and consumption situation in Kuwait. The interviews
were conducted by senior year Kuwait university undergraduates. The objective
of the study was to examine the water supply and consumption in Kuwait.

Living conditions
The ®rst aspect of the study was concerned with the living conditions, house sizes,
average monthly income, etc. It was found that the majority of the families, or
24%, consists of four people, and that 68% of the population live in two-story
housing. Most of the houses surveyed (86%) had a garden or a yard that they
water. All of the 3000 households in the survey had at least one car, 10% had six
cars, 48% had ®ve cars, 17% had three cars, 8% had two and only 2% had one car
per household. These results are shown in Table 2. About 27% of the heads of the
households that were surveyed made between 3,343-5,010 US dollars per month,
22% made more than 6,680 dollars, 21% make between 2,341-3,340 dollars, 14%
make between 1,670-2,338 dollars, 13% make between 5,010-6,680 dollars, and
only 3% make less than 1,670 dollars per month. Each head of household was also
asked about the level of education that he or she had.
The average monthly water bill varied from house to house; 39% had a bill
between 104-167 dollars per month, 30% had greater than 234 dollars, 17%
between 171-234 dollars, and 14% had a bill between 67-101 dollars per month.
202 Jasem M. Al-Humoud and Ibrahim Al-Ghusain

Table 2. Number of cars per household

Number of Cars Percentage (%)
1 2
2 8
3 15
4 17
5 48
6 10
Total 100

Water use
Since people in Kuwait use both fresh and brackish water daily, the respondents
were each asked to give an idea of what kind of water they use in their di€erent
daily activities.
Half of the households, or 50%, use bottled water, while 40% use tap water,
and 10% use them both for drinking. It was also discovered that 44% of the
people water their garden with fresh water, 29% with brackish water and 14%
use both types of water for gardening (see Table 3). It was also found that 44%
of the households with gardens water them at least three times per week, 12%
twice, and 2% once a week (see Table 4).

Table 3. Type of water people prefer to drink

Type of Water Percentage (%)
Bottled water 50
Tap water 40
Bottled water & tape water 10
Total 100

Table 4. Number of times garden is watered each week

Number of times per week Percentage (%)
1 2
2 12
3 43
4 32
5 10
6-7 1
Total 100
 Household demand for water 203

An interesting element in the survey was a question regarding the number of

times clothes are washed each week. It was found that 44% of the respondents
do their wash seven days a week, 23% twice, 13% four times, 12% three times,
5% once, 2% ®ve times and 1% six times a week (Table 5). Washing the entire
house (¯oor scrubbing) consumes a lot of water; 46% of the people wash their
houses once a week, 31% twice, 13% three times, 4% four times, and 6% at
least ®ve times per week (Table 6).

Table 5. Number of times clothes are washed per week

Number of Washes per Week Percentage (%)

1 5
2 23
3 12
4 13
5 2
6 1
7 44
Total 100

Table 6. Number of times ¯oors are scrubbed each week

Number of Washes per Week Percentage (%)

1 46
2 31
3 13
4 4
5-7 6

Total 100

The use of a shower or bath is another cause of high water consumption. Our
results show that every member of the household takes a shower or bath once a
day or once every other day.

Finally, an attempt has been made to ®nd out what bothers people most about
the water. It was found that the color of the water is the most bothersome item
to them. Forty-four percent of the people are bothered by color alone, 40% are
204 Jasem M. Al-Humoud and Ibrahim Al-Ghusain

bothered about the color and other related factors, such as taste, impurity, etc.,
and 16% are bothered by other factors than the color of the water.

Forecasting water consumption (demand)

Forecasting water demand in Kuwait carries major policy implications for both
the Kuwaiti government and the water industry. Kuwait's government heavily
subsidizes water consumption. While 1000 Imperial gallons (4,545 m3) of water
costs MEW KD 2.7 to produce, the customer is charged KD 0.8 for it. These
subsidies add a heavy burden to the already large ®scal de®cit. Predictions of
future water consumption (demand) would help the government in its e€orts to
reform its subsidy policy and ®nd alternative methods of reducing its de®cit.
Also, the water sector invests continuously in expanding its production capacity.
It would be useful indeed to form an idea about the future demand for this
public utility. This research uses time series methodology in forecasting water
consumption, which has never been used in forecasting consumption and/or
demand for public utilities in Kuwait.

In times series, the series is ®rst examined by computing the autocorrelation

function (ACF) and the partial autocorrelation function (PACF) for the original
series (Granger & Newbold 1986; Mills 1990) or by formal unit root analysis
(Dicky et al. 1986). The correlograms associated with ACF and PACF are often
good visual diagnostic tools (Hanke & Reitsch 1995). The tentative model is
then estimated. A number of alternative parameter estimation procedures are
commonly used. These various procedures typically yield quite similar estimates
when the sample size is large. However, for shorter series, there can be larger
di€erences, particularly if the model involves substantial moving average terms.
The full-maximum likelihood approach is usually preferred in those cases for
which di€erent estimation procedures yield signi®cantly di€erent results
(Newbold & Bos 1994). The residuals from the tentative model are examined to
determine whether or not they are white noise. The adequacy of the model is
also indicated by the Box-Ljung statistic (Ljung & Box 1978). Model
inadequacy is indicated by large absolute values for the residual
autocorrelations, and consequently large values for the Box-Ljung statistic. If
the residuals are white noise, the tentative model is probably a good
approximation of the underlying stochastic process. If they are not, then the
process is started all over again.

The model ®nally selected can then be used for forecasting. Forecasts can be
made for a single period or several periods in the future. Con®dence intervals
can also be constructed about these estimates. In general, the further into the
 Household demand for water 205

forecast, the larger the con®dence intervals. Therefore, as more data become
available, the same model can be used to revise the forecasts by choosing
di€erent time origins (Box & Jenkins 1970, Ho€ 1983, Chat®eld 1989).
As in our case, a series of 30 observations, representing total annual water
consumption in Kuwait during the period from 1970 to 2000, was collected. As
seen before, Fig.1 gives a plot of the original series. A close inspection of the
®gure suggests that water consumption in Kuwait experienced exponential
growth over the sample period with the exception of the period 1990-1992, as
was explained earlier in the paper. The series suggests a constant proportional
growth rate of approximately 8% per annum. This was estimated using the
regression:

Ln Ct ˆ b0 ‡ b1 T ‡ ut …2†

where Ct is the total annual water consumption in period t, T is time, b0 and b1

are constants and u is the error term. The coecient b1 represents the (constant)
proportional rate of growth (r), given by:

r ˆ …dC=dT†=C: …3†

The ®nal model was estimated using the SAS program (SAS/ETS Software
1991) and it can be written as follows:

Yt ‡ 96:6 ˆ 0:11…Yt 1 ‡ 96:6† ‡ et : …4†

The model shown above in Fig.3 and in Table 7 appears to provide a good ®t to
the water consumption. None of the Q statistics values are signi®cant and none
of the residual autocorrelations are very large in the autocorrelation check of

Fig.3. Actual and forecasted water consumption in Kuwait

206 Jasem M. Al-Humoud and Ibrahim Al-Ghusain

residuals, indicating that the model provides a good ®t to the data. The seasonal
moving average parameter estimate of 0.11 is signi®cant. In addition, the
Akaike's Information Criterion (AIC) (Akaike 1974, Harvey 1981) and
Schwartz's Bayesian Criterion (SBC) (Schwartz 1978) which measure the
goodness of ®t are small, indicating good ®t of the model. Moreover, more
elaborate models did not produce superior results. Therefore, this model
adequately describes the behavior of the water consumption time series. Figure
3 shows how close the forecasted values given by the model are to the actual
values; the results suggest a very good ®t. In addition, Table 8 lists the forecasts
with their associated forecast standard errors (Std Error) and 95 percent
con®dence limits. Note that the forecast standard errors increase and the 95
percent con®dence limits widen for each future period.

Table 7. Statistical Analysis Results

Parameter Estimate Std. Error t-value Approx. Pr>jtj Lag

MU -96.60151 338.49424 -0.29 0.7753 0
MA1 0.11182 0.40863 0.27 0.7844 12

Constant Estimate -96.6015

Variance Estimate 2530673
Std. Error Estimate 1590.703
AIC 406.8082
SBC 408.1643
Number of Residuals 24

To Chi- Pr >
Lag square DF Chi-sq Autocorrelations
6 11.93 5 0.9035 0.013 -0.588 -0.070 0.135 0.083 0.032

12 12.97 11 0.9370 -0.046 -0.058 -0.018 0.086 0.092 -0.048

18 13.56 17 0.9757 -0.079 -0.021 0.001 0.031 0.039 0.007

24 13.63 23 0.9854 -0.019 -0.010 0.004 0.001 -0.002 -0.002

 Household demand for water 207

Table 8. Lower and upper 95% levels of water consumption.

Actual Water
Year consumption Forecast Std. L95 U95 Residual
1970 6638 - - - - -
1971 7688 - - - - -
1972 8572 - - - - -
1973 9303 9525.40 1696.21 6200.89 12849.91 -222.40
1974 10023 10090.40 1696.21 6765.89 13414.91 -67.40
1975 11602 10627.40 1696.21 7332.89 13981.91 944.60
1976 14328 12225.40 1696.21 8900.89 15549.91 2102.60
1977 17312 15810.40 1696.21 12485.89 19134.91 1501.60
1978 20699 19941.40 1696.21 16616.89 23265.91 757.60
1979 23067 23586.40 1696.21 20261.89 26910.91 -519.40
1980 23443 26357.40 1696.21 23032.89 29681.91 -2914.40
1981 24917 25714.40 1696.21 22389.89 29038.91 -797.40
1982 28181 25196.40 1696.21 21871.89 28520.91 2984.60
1983 31470 29558.40 1696.21 26233.89 32882.91 1991.60
1984 34522 34637.40 1696.21 31312.89 37961.91 -115.40
1985 36904 37738.96 1685.83 34434.79 41043.13 -834.96
1986 38469 39866.84 1685.83 36562.67 43171.02 -1397.84
1987 40306 40650.08 1685.83 37345.90 43954.25 -344.08
1988 43422 41542.18 1685.83 38238.01 44846.36 1879.82
1989 47605 44996.56 1685.83 41692.39 48300.73 2608.44
1990 47546 50540.73 1685.83 47236.56 53844.90 -2994.73
1991 30814 51689.76 1685.83 48385.59 54993.93 -20875.76
1992 42641 51856.03 2384.04 47183.40 56528.66 -9215.03
1993 49456 46776.23 1685.83 43472.05 50080.40 2979.77
1994 57165 49496.04 1685.83 46191.87 52800.22 7668.96
1995 61577 63972.28 1685.83 60668.11 67276.45 -2395.28
1996 67464 68902.14 1685.83 65597.97 72206.32 -1438.14
1997 73284 71872.75 1685.71 68568.83 75176.67 1411.25
1998 78496 79230.68 1685.71 75926.76 82534.61 -734.68
1999 84070 84257.87 1685.71 80953.95 87561.79 -187.87
2000 88452 88975.23 1685.71 85671.30 92279.15 -523.23
2001 - 93637.76 1685.71 90333.84 96941.68 -
2002 - 98257.99 2383.95 93585.54 102930.43 -
2003 - 103347.15 4129.11 95254.23 111440.06 -
2004 - 107870.77 5330.66 97422.86 118318.68 -
2005 - 112863.33 7347.81 98461.88 127264.77 -
208 Jasem M. Al-Humoud and Ibrahim Al-Ghusain

CONCLUSION
Forecasting water demand in Kuwait carries major policy implications for both
the Kuwaiti government and the water industry. Kuwait's government heavily
subsidizes water consumption, which adds a heavy burden to the already large
®scal de®cit. Also, the water sector invests continuously in expanding its
production capacity. It would be very useful indeed to form an idea about the
future demand for this public utility.
Total annual water consumption in Kuwait is predicted to reach 1.13 billion
Imperial gallons (5.14 million cubic meters) in the year 2005. However, this
®gure could be as high as 1.3 billion Imperial gallons (5.91 million cubic meters).
This represents an increase of 17.5% to 25% over the current consumption
rates. Such an increase would de®nitely add constraints to the budget.
It is apparent that people in Kuwait are bothered most about the color of the
water that reaches their houses. That is why the majority of the people surveyed
use bottled water for drinking. It is noticed as well that people in Kuwait waste
a lot of water. The water authority and the Government should start a public
campaign in advising and guiding people on consuming water. One way to
encourage the people to reduce water consumption is by increasing the charge
rates for water use. In view of the rapidly increasing cost of water supply
schemes, in¯ation, and the ®nancial crises the country is going through
(especially after the Gulf War), the above rates seem to be unreasonably low.
The water authority should advise the consumers to use more of the brackish
water for their daily chores, such as irrigation, landscaping, household purposes,
live-stock watering and construction works. They could also make a separate
network of the brackish water to the toilets. Perhaps the water authority should
have two di€erent water rates: one for fresh water and a lower one for brackish
water. By the same token, the water authority should study the reduction in the
volume of water consumption for personal washing and bathing, by replacing
the conventional sprays with ones that could reduce water ¯ow.
The implications for the budget suggest that a more e€ective rationalization
of water consumption may be needed. This may require use of partial
subsidization and multiple tari€ pricing system. If conditions continue as they
are, the above results suggest that Kuwait may need to consider expanding the
current water supply by at least 20% to cope with the increase in demand.
The Ministry of Electricity and water should seriously think of using
wastewater treated e‚uents. In many places around the world, secondary
treated wastewater is used for irrigation. In fact, tertiary treated e‚uents (BOD
5 - 10 mg/l) are produced in Kuwait. Such high quality e‚uents can be used for
irrigation. Alternatively, they could be discharged into arti®cally constructed
 Household demand for water 209

wetlands, which may represents a viable option. In addition, the cost of

producing tertiary e‚uents is low compared to producing fresh water in
Kuwait. It costs less than KD 0.6 to produce 1000 Imperial gallons (4.545m3) of
tertiary treated e‚uent compared to KD 2.7 per 1000 Imperial gallons
(4.545m3) of fresh water. For example, the fresh water production from the
Ministry of Electricity and Water (MEW) in the year 2000 was 89436106
Imperial gallons at a cost of KD 245,938,000.000. On the other hand, the
treated wastewater from the treatment plants in the country was 33,1666106
Imperial gallons at a cost of KD 19,899,600.00. That is a di€erence of KD
226,038,400.00.

ACKNOWLEDGEMENTS
The authors are very thankful to the distingished reviewers of the Journal. The
authors also wish to thank the Ministry of Electricity and Water (MEW) for
their help and support.

REFERENCES
Akaike, H. 1974. A new look at statistical model identi®cation. IEEE Transaction on Automatic
Control, AC 19: 716-723.
Box, G. & Jenkins, G. 1976. Time Series Analysis: Forecasting and Control. Holden-Day, San
Francisco.
Chat®eld, C. 1989. The Analysis of Time Series: An Introduction (4th edition) Chapman and Hall,
London.
Dickey, D.A., Bell, W.R & Miller, R.B. 1986. Unit roots in time series models: Tests and
applications. The American Statistician, 40: 12-26.
Falkenmark, M. 1989. The massive water scarcity in Africa: Why isn't it being addressed? Ambio
18(2).
Hankee, J.E. & A.G., 1995 Business Forecasting. (5th edition) Prentice Hall International, Inc,
Englewood Cli€s, New Jersey.
Harvey, A.C. 1981. The Economic Analysis of Time Series. John Wiley & Sons, Inc. New York.
Ho€, J.C. 1983. A Practical Guide to Box-Jenkins Forecasting. Lifetime Learning Publication,
Belmont, CA.
Ljung, G.M. & Box, G.E.P. 1978. On a measured of lack of ®t in time series models. Biometrica, 65
279-303.
Postel, S. 1992. Last Oasis: Facing Water Scarcity. W.W. Norton & Co., New York, USA.
MEW, Ministry of Electricity & Water, 2001. Statistical Year Book (Water). Edition 26 "B".
MOP, Ministry of Planning, 2000. Annual Statistical Abstract, Edition 36.
Newbold, P. & Bos, T. 1994. Introductory Business and Economic Forecasting. South-Western
Publishing Co., Cincinnati, Ohio.
Omar, S.S., Alyqub, & Senay, Y. 1996. Geology and Groundwater Hydrology of the State of
Kuwait, Gulf Arabian Penin. Study 1: 9-51.
PRB 1993. "World Population Data Sheet", Water Resources Institute, World Resources.
210 Jasem M. Al-Humoud and Ibrahim Al-Ghusain

SAS/ETS Software: Applications Guide 1. 1991. Time Series Modeling and Forecasting, Financial
Reporting, and Loan Analysis. Version 6, 1st Edition. Cary, N.C. USA.
Schwartz, G. 1978. Estimating the dimensions of a model. Annals of Statistics 6: 461-464.
Shiklomanov, I.A. 1990. Global Water Resources, Nature and Resources 26(3).
Walpole, R.E., & Myers, R.J. 1985. Probability and Statistics for Engineers and Scientists (3rd
Edition) Macmillan Publishing Co., New York.
WRI, World Resources Institute 1992. World Resources 1992-1993. Oxford University Press, New
York.

Submitted : 18/4/2001
Revised : 1/6/2001
Accepted : 10/6/2002
 Household demand for water 211

?+yR"}yG g9+}yG J9F9+AIG

B*(wyG ?yhO ,p ?+z}f ?SGQO

#+ZjyG ~+$GQ<E h O(}JyG O}J| ~S9F

B*(wyG ?g|9F - ?+!O}yG ?SO"%yG ~Ts
13060 I9qZyG 5969 .H.U

? Y ;L

.B*(wyG ?yhO ,p g9+}yG c;%ASG ?SGQOy EJ< ?G+A! ,$ ?+}zgyG ?sQ(yG gP$ fE
k
?yhO ,p ?qzAM| dR9"| #| 9+FG(Wf Bg}F A9}yG c;%ASG #f J9|(zg|h J9!9+<
,p g9+}yG )L9A!E( c;%AS: ,z=tAT}yG Q(ZAyG ,ag* EJ=yG GP$ f{p ?p9\ 9< .B*(wyG
k
Q+=v {wW< ?|(wJyG {=s #| 9+y9I e(fO| B*(wyG ?yhO ,p g9+}yG L9A!E .B*(wyG ?yhO
k k
?+}wy ,z=tAT}yG |="AyG f{p xyPy .?+!GR+}yG )zf 98=f {wW* K=YC &!C E+J< ,GOF
HF9A! .L9A!5y ~fOyG ?i9+Y IO9f ?J+JZyG IQ(ZyG ?|(wJyG ,ag* g9+}yG c;%ASG
?*9%! ,p .,A*(wyG dR"}yG {=s #| QO%@ g9+}yG #| IQ+=v J9+}v c9"$ fz< #+=@ ?SGQOyG
O+VQAy ,A*(wyG #_G(}zyh A9}yGh A9<Q%wyG IQGR(y ?}%| J9dI;| c9"$ ?SGQOyG
.g9+}yG c;%ASG