Watering cities: spatial analysis of urban water use in the Southeastern United States

Abstract

This study explores the spatial pattern of urban water use in the southeastern US using Moran's I and Local Moran statistics, and identifies the factors influencing urban water use quantity using a regression model. The findings suggest that counties with a large quantity of water use are spatially clustered. The regression analysis reveals that the quantity of water use decreases as the water price increases in cities in the light water use group and when a city government manages water in cities in the heavy use group. Effective water management policies in the southeastern region of the US need to be established in consideration of these effects.

Keywords:

• urban water use
• spatial analysis
• southeastern United States
• water price
• water management

Notes

1. ‘Hot spot’ refers to a statistically significant spatial cluster of high values (large water use in this case).

2. The spatial analysis in the first part is conducted of the counties of six states, excluding Tennessee, as information pertaining to domestic water use in Tennessee is not included in the USGS data (Estimated Use of Water in the United States in 2000) which provided the information on urban water use. As a result, the amount of urban water use in this state cannot be compiled.

3. When the data on climate were collected, the priority was on whether the source is either a US federal government agency or a nationwide public institution, as the sources of all of the other variables are. The July mean temperature and the relative humidity were obtained from ‘Natural Amenities Scale’ provided by the USDA. However, as annual precipitation data are not provided from this source, an additional search was made in an effort to find the data published by a federal source. This was not successful, and PRISM data offered by Oregon State University was used instead.

4. For example, Smith (1956) notes that the fragmentation of the problem led by public districts (e.g. water districts) makes effective resource management more difficult.

5. The correlation is also checked for climate, socio-economic and price variables across the three groups. Among the 192 pairs of variables, 32 pairs (approximately 16.7%) show a significant correlation at the 95% significance level. However, all have a mild correlation (smaller than 0.4) except for three cases (PORT-VR (0.473), INCOME-PR3K (0.419) and BAGE-PI1.5M (0.416)). In addition, because the multicollinearity condition number from the regression analysis with the factors obtained from the factor analysis is as low as 2.821, it is confirmed that this does not cause any statistical problem.

6. 613.70 million gallons per day by Nashville and 271.89 million gallons per day by Miami are extraordinarily large water use quantities compared to the 93.97 million gallons per day by the third largest city in terms of water use.

7. Regional water authorities established in 1973 in the UK may be an example of regional scale governance. These authorities were able to respond more effectively to the drought in 1975–1976 than the previous local management authorities (Castro et al. 2003).

8. Public districts such as water supply districts are organised to deal with specific public problems (Smith 1956). Therefore, the policy options that they have are usually limited to those specific issues. However, local governments are able to take a more comprehensive and integrated approach by not necessarily limiting their policy attention to a specific problem itself but considering various policy options from related fields, which in turn have an influence on any specific problem of interest.