Journal of Water Resources Planning and Management

A goal‐programming model to determine the daily production of a seawater desalination plant and the optimal allocation of desalinated water to various blending stations is developed. At these blending stations, desalinated water is mixed with brackish ground water obtained from several wells with different salinity levels. The model constraints are to meet the target levels of demand for blended water, to define the size of the depletion of the nonrenewable ground water, and to ensure meeting the required salinity level. The model objective is to minimize the goal deviations from five priority levels: (1) demands for blended water, (2) control of the salinity levels, (3) depletion of ground water, (4) demands for brackish water at blending stations, (5) and production of desalinated water. The model has been used to determine an optimal allocation for the distribution and blending of desalinated water in the Dammam metropolitan area of Saudi Arabia.

As a means of examining the performance characteristics of an infiltration trench for storm‐water control, the infiltration rate into the soil surrounding an infiltration trench is simulated with a two‐dimensional saturated‐unsaturated finite‐element model. The model is used to examine infiltration rates with respect to time, depth of water in the trench, distance to the water table, surrounding soil texture, and antecedent moisture condition. The impact of clogging of the bottom of the trench on infiltration rate is examined. The finite‐element‐model results are compared to results obtained using a Darcy's law approach to assess the applicability of the simple model for design purposes. The infiltration rates calculated with the simple model are shown to be conservative.

The theoretical basis of a risk‐management approach to wellhead protection is developed. Contaminant sources are divided into two categories: chronic sources and spills. For chronic sources, such as residential septic tanks, contaminant influx to the ground water, decay rates, and retardation factors are viewed as random variables with defined probability density functions, and risk‐analysis yields the expected risk of an adverse health effect to the population served by the wells. In the case of spills, risk analysis yields the maximum probable concentration (MPC) in the pumped water for a given design spill. From a regulatory viewpoint, chronic sources must have regulated contaminant loadings, and facilities handling significant quantities of hazardous materials must have design spills. Using the risk‐analysis procedure developed in this study, a defined level of protection is associated with land uses within the zone of contribution of a wellfield. This is an improvement over conventional wellhead protection strategies that do not define a level of protection.

This paper presents two multiobjective models for the estuary‐management problem based upon goal‐programming procedures. The multiobjectives are the minimization of freshwater inflow into bays and estuaries and the maximization of commercial‐fish harvest for five fish species. Salinity is expressed as a function of freshwater inflow in a nonlinear‐regression equation, which is used as a constraint. Additional nonlinear constraints are the harvest‐regression equations for the various species that express harvest as a function of the quantity of freshwater inflow. The goal programming model is a nonlinear‐programming problem solved by using the general algebraic modeling system/modular in‐core non‐linear optimization system (GAMS/MINOS) programming code. Because of the uncertainty associated with the regression equations for salinity and harvest, a second formulation was developed in which the constraints are expressed in a chance‐constrained formulation. The chance‐constrained goal‐programming model is solved by using the GRG2 programming code. The models for both formulations are applied to the Lavaca‐Tres Palacios Estuary in Texas.

The channelization project in the Lower Kissimmee River Basin, Florida resulted in severe losses of river floodplain wetlands and waterfowl throughout the river valley. An earth channel, the C‐38 canal, is the main element of this project. Immediately after the project completion in 1971, a strong effort was initiated to restore this basin to its prechannelization status. This project, if and when completed, would set a milestone toward our commitment to the enhancement of ecological harmony because its only major goal is ecological enhancement. The investigation of various alternative restoration plans is described. First, the ecological goals for the restoration plans were formulated. Then, the alternative restoration plans were tested against the ecological goals through a combination of field data collection, physical modeling, and numerical modeling. Finally, the backfilling of certain reaches of C‐38 canal was recommended.

Ground‐water simulation models are increasingly used for the analysis, design, and management of water‐resource problems including ground‐water supply and remediation. The solution to these problems involves some degree of control of ground‐water hydraulics. Combining simulation with linear optimization can provide a useful tool for problem solution. The optimal hydraulic‐control problem for ground‐water systems is defined in terms of controls on the ground‐water system and typical design criteria. Commonly used linear formulations of the control problem are described in detail. Recent applications of the optimal hydraulic‐control approach to field problems are reviewed. The deterministic hydraulic‐control problem is contrasted with more sophisticated techniques which incorporate contaminant transport and uncertainty in aquifer parameters. Conclusions are drawn about the likely future applicability of this method for solving ground‐water management problems.

A hypothetical watershed was constructed to test the efficacy of using the ARC/INFO geographic information system to provide spatially related input for the U.S. Army Corps of Engineers HEC‐1 hydrologic model. The ARC/INFO system performed the tedious and time‐consuming tasks of spatial averaging (basin areas, average runoff curve numbers. etc.) quite well. Difficulties, however, were encountered in using the TIN (triangulated irregular network) module to accurately assess average rainfall intensities. The inaccuracies associated with the computation of average basin rainfall intensities were quantified by using a defined rainfall pattern imposed upon a simple geometric shape (equilateral triangle) and were found to be directly related to the number of prescribed contouring intervals. The highest density of contouring intervals (0.010 nm) resulted in only a 0.55% over‐estimation. The number and location of rain gages relative to watershed boundaries imparted a significant error (approximately ± 10%) to average rainfall estimation.

There are many situations in water‐resources systems in which the statistics of extremes plays a decisive role. This paper investigates the characteristics of extreme values for a strictly monotonic increasing‐loss function, $Y=g(X;)$, where $X$ is an underlying‐random variable whose exact form of probability distribution is unknown. The results are of wide application in water‐planning and management problems. For example, the results are useful in studying the extremal‐flood‐loss function when full knowledge of the flood discharge is not available. A methodology is developed for analyzing the extreme tail of the distribution of loss. Using von Mises' convergence criteria, sufficient conditions are derived to determine the asymptotic distribution of the largest value from realizations of $Y$. Expressions are also given for determining the associated parameters of this distribution. These results are combined to evaluate the expected value and variance of the largest‐sample value of $Y$. Two illustrative examples demonstrate that the new results are useful in the assessment and management of the risk associated with extreme events such as catastrophic floods.