Journal of Water Resources Planning and Management

Hourly water‐demand data is forecasted with a model based on a combination of pattern recognition and time‐series analysis. Three repeating segments are observed in the daily demand pattern: “rising,” “oscillating,” “falling,” then “rising” again the following day. These are called “states” of the demand curve, and are defined as successive states of a Markov process. The transition probabilities between states are “learned,” and low‐order auto‐regressive integrated moving average (ARIMA) models fitted to each segment, using a modest amount of historical data. The model is then used to forecast hourly demands for a period of one to several days ahead. The forecast can be performed in real time, on a personal computer, with low computational requirements, at any time the system state deviates from the planned, or when new data become available. The process of model development, application, and evaluation is demonstrated on a water system in Israel.

In an era when subsidized water development is not available to municipalities, development costs are fully borne by the rate‐paying clientele. Proposed water projects therefore impact consumers both through increased rates and increased water supply. Economic evaluation requires attention to the interdependent effects of both changes. Under these circumstances it is inappropriate to calculate the benefits of a project due to the enhancement of supply and then obtain net benefits by deducting project costs. A framework for the proper analysis of locally financed water development is proposed and then employed to evaluate the highly controversial Applewhite Reservoir in San Antonio, Tex. The study indicates that cost recovery will take substantially longer than proposed, net present value is negative under baseline conditions, and the project was initiated prematurely from the perspective of the welfare of local water consumers.

A model is developed for optimal operation of a multiquality water‐supply system, under steady‐state conditions. The system contains: sources of different qualities, treatment facilities, pipes, and pumping stations. The objective is to minimize total cost, while delivering to all consumers the required quantities, at acceptable qualities and pressures. A special approximation of the equation for water quality in pipes is used, which enables the model to select the flow directions in pipes as part of the optimization. The steady‐state operation of an example system is optimized: it supplies six consumers from three sources, two of them with treatment plants, and has three pumping stations and 10 pipes. The optimization is carried out with GAMS/ MINOS, which employs a projected augmented Lagrangian algorithm. The example system has been analyzed through a base run and four additional runs, aimed at studying the effects of modifications in key data. The optimal solutions of the five cases demonstrate the response to changes in economic and operational conditions.

A model is developed for optimal operation of a multiquality water‐supply network under unsteady conditions, for a time horizon that is divided into a number of time periods. The objective is to minimize total cost, which includes the cost of water at the sources, of treatment, and of the energy to operate the system. The constraints include equations that describe the change in flow and quality over time throughout the system, the physical laws of flow and concentrations, and the requirements for level of service. The equations that describe concentrations in pipes are of a form that allows the flow direction to reverse during the iterative solution process. The model is solved with GAMS/MINOS. An example system is optimized, with two sources, one with a treatment plant, two reservoirs, 6 consumers and 11 pipes, operated over five time periods. The system has been analyzed through a base run and three additional runs.

A Monte Carlo simulation experiment was conducted to explore the relationships between land use, climate, and water supply yields. A watershed streamflow and sediment yield model was coupled with a reservoir model and used to estimate storage‐yield relationships for otherwise identical 5,000‐ha forested, agricultural, and urban watersheds using weather data from four different sections of the eastern U.S. (Georgia, Indiana, New York, and Texas). Simulation results indicated that although land use very significantly affects water supply yields, the magnitudes of the effects are quite dependent on climate. Thus the differences in yields among the three watershed types are much greater in areas such as Georgia and Texas that have high evapotranspiration than they are in Indiana and New York. For example, at the Texas and Georgia sites, urban watersheds produced more than twice the water supply yield of forests at high levels of flow regulation (large reservoir capacities) while in Indiana and New York, increases of 25–50% were obtained. Agricultural watersheds also generally provided greater yields than forests, although at smaller reservoir capacities, the loss of reservoir capacity due to sedimentation in agricultural watersheds sharply reduced yields.

The ground‐water table in four districts of the city of Jeddah, in Saudi Arabia, is to be lowered by installing a subsurface drainage system. In connection with the drainage studies, the pollution potential of the drainage water for discharge to the coastal waters has to be determined. Quality of the subsurface waters in the areas to be drained was characterized by their physical, chemical, and bacteriological parameters. Samples were collected from 31 shallow (3–5 m) and from four deep (40–86 m) observation wells. The study showed that the quality of the drainage water satisfied the 1989 standards of the Meteorology and Environmental Protection Agency (MEPA) for discharge into the Red Sea.