Though the need for demand management is emphasized for solving the water supplydemand problems in Japan, there is little detailed information about actual conditions of water use. The objective of the present study is to clarify the residential water demand structure in the Ashida river basin, located in the eastern part of Hiroshima prefecture, with a drainage area of 900 km2 (Fig. 1), by using individual data. Water demands for municipal and industrial sectors have been rapidly increased with urbanization and industrialization since 1960. In 1975, Fukuyama city had a total population of 330, 000 and a total amount of water use of 1.3 ×108m3. A questionnaire survey of 3, 200 households is conducted to obtain information on the factors which have any effects on residential water use. The monthly or bi-monthly water use series of individual households are transcribed from meter books of each municipality for the 1975, 1976 and 1977 fiscal years. In the present study, the following three types of residential water demand models are tested using linear equations. Model A with income variable is used for all of the households which offer the available sets of data (Sample f) . Models B and C, which contain dummy variables on water-saving, and income and price variables, are used for the households utilizing their water only for residential purposes (Samples pi and V). _??_ where Rw is the amount of municipal water use in m3/household/year, Np the number of persons in household, Nb the number of babies (less than two years old) in household, In the income in million yen/household/year, Pr the average price of water in yen/m3, and others are dummy variables except of regression coefficients, ai, bi, and ci (Table 1). Regression coefficients of a3, b2, b3, b4, c2 and c6 are expected to be negative, and all of other regression coefficients with the exception of constant termsa0, b0, and c0 are expected to be positive. The results of a sample survey are summarized as follows ; 1. The mean values of Pw (public water supply), Gw (ground water use), and Tl (flushtoilet use) show the wide regional differences, but there are no remarkable differences in the mean values of other variables (Table 3). The seasonal variations for residential water use show a minimum of 78% to 87% in winter and a maximum of 114% to 130% in summer. 2. The regression analyses using Models A, B and C are listed in Tables 4, 5, and 6. Each equation is the “best” one that yields the expected signs of regression coefficients and no longer explains a significant amount of the remaining unexplained variation by adding another variable, among all equations having every possible combination of explanatory variables. A significant portion of the variation in residential water use can be explained by the variables of the number of persons in household (Np), and the presences or absences of ground water use (Gw), of flush-toilet use (Tl), and of non-domestic water uses around the house (Nr and Ot). The obtained results agree entirely with the previous study of Tachikawa city (Shimmi, 1977). Table 7 shows the household-size, income, and price elasticities evaluated at the mean of respective variables by using individual data. The household-size elasticities are in the ranges of 0.41 to 0.45 for Sample i, of 0.30 to 0.73 for Sample JV, and of 0.38 to 0.51 for Sample V. These ranges are consistent with the previous estimates. The income elasticities range from 0.01 to 0.22 for Sample III, and from 0. 01 to 0. 11 for Sample V.
Read full abstract