Journal of Water Resources Planning and Management

A monthly planning and operational model has been developed for the Adelaide headworks system in South Australia, Australia. The model uses linear goal programming to aid in the identification of optimum operating policies for the system. The planning and operational policies obtained are aimed at achieving maximum yield for a given level of reliability. In this particular system, where a significant fraction of the supply is pumped from a distant river, a further objective is to minimize pumping cost. The results obtained for the Adelaide system indicate that savings of between 5% and 10% of total pumping costs can be achieved through the application of the model developed. Potential also exists for further reductions in pumping costs through improved forecasting of up to 20% of total pumping costs. Examples are also given of the use of the model to assist with operational decision making. This model has subsequently been implemented and the predicted savings in operating costs are being achieved by the local water authority, the Engineering and Water Supply Department.

A practical way of assessing the impact of various pipe failure conditions on water distribution networks is described. The method avoids the need to undertake a full network analysis or simulation for each particular component failure. The method assesses the vulnerability of the network to the loss of any particular pipe element, and provides a quantitative estimate of the impact on each nodal demand, and the postfailure utilization of nodal sources and pipe elements. Implementation of the method requires only knowledge of the network configuration and a set of typical operating conditions, such as might already be available from a routine network analysis of the intact distribution network. Detailed non‐linear network analysis covering all possible failure conditions is thus avoided. In cases of networks supplied by a number of sources of differing quality, the same overall technique can also be used to determine the chemical/biochemical quality of supply at any point in the network. The results of the method can be combined with pipe failure probabilities to provide measures of network reliability.

A hydrometerological study of the Yockanookany watershed in Mississippi was conducted to demonstrate the use of National Weather Service (NWS) radar and rain gage data for flood forecasting. Weather radar scenes were recorded at approximately 10 min intervals using a Kavouras Triton X computer for three storms in 1986 by the U.S. Army Corps of Engineers. The NWS recorded precipitation hourly at five gaging stations in or near the watershed. A U.S. Geological Survey (USGS) stream gage was located at the watershed outlet. Calibrated weather radar is shown to be effective for flood forecasting. Both the modified Brandes and Kriging calibration techniques gave good results. The most consistent results were achieved when the rainfall intensity was computed from six radar scenes per hour. Storm routing can be effective in forecasting precipitation several hours in advance, if conducted when the storm intensity is relatively constant. The proposed storm translation method is effective for stratiform‐type storms with motion trajectories and intensities almost constant, and should not be used in forecasting severe convective‐type storms.

Flood hydrograph determination under arid climate for countries like Saudi Arabia is essential for the assessment of water resource potential and for the design of various hydraulic structures. To meet such objectives, hydrological and morphological data from different regions of Saudi Arabia are evaluated for their suitability and tested for model application. The suggested model used in the present study, geomorphologic instantaneous unit hydrograph (GIUH), is applied by the researchers for its application in the selected thirteen representative basins. The calibrated model for gaged wadis can be applied to other ungaged wadis by following the proposed procedure. The mathematically and graphically derived relations are found to be the key elements in the prediction of flood hydrographs for ungaged wadis.

This paper describes the initial phase of an investigation to develop a methodology for estimating the total costs of a dam failure. This methodology takes into account the fundamental issues of accounting stance, damage to fixed assets, loss of income, intangible losses, and loss of life and limb. A review and analysis of the English‐language literature on historical floods and dam failures was used as the starting point for developing a framework for evaluating dam‐failure costs. Cost‐related information is portrayed by type and collection priority in a data‐collection matrix to guide an on‐site investigation. This matrix was tested by evaluating five major dam failures that occurred in the United States during the past 20 years, and significant limitations in the current data base were identified. Once validated by trial data collections involving future events, the methodology might be used to assist in conducting a probabilistic safety assessment of a dam, or to develop more rational public policies aimed at risk mitigation.

Quantification of the amount of redundancy in a looped water‐distribution network is necessary in order to compare different designs and to use an optimization approach in the design of the system. In this work, redundancy is quantified using the expected shortage due to failure of individual pipes as a surrogate measure of reliability that permits incorporation of some considerations of frequency, duration, and severity of damage. Based on this surrogate, a gradient‐modified linear‐programming model is developed for minimum‐cost design subject to reliability constraints. The model constrains the shortage at each node in the network to be less than or equal to some specified fraction of demand. A solution approach is proposed to overcome computational complexity, and is shown to bring practical‐sized network solutions within reach. The model is used to explore the trade‐off between cost and reliability. An extension of the algorithm considers storage tanks within the system.