Journal of Water Resources Planning and Management

Many coastal ecosystems have been adversely impacted by increased storm-water drainage due to expanding urbanization. The ecosystem of the St. Lucie Estuary (SLE), located on the east coast of south Florida, has been greatly influenced by development of an intricate network of storm-water drainage canals in the tributary watershed. A suite of models dealing with watershed hydrology, reservoir optimization, and estuary salinity and ecology are applied for optimal sizing and operation of storm-water reservoirs. The multipurpose storm-water control facilities provide for hydrologic restoration to predrained or natural hydrologic conditions for recovery of salinity-sensitive biota in the SLE, as well as supplemental irrigation water and pollution control through connected storm-water treatment areas. The optimization is challenging since the ecological goal is for mean monthly storm-water discharges to the SLE to coincide with the desired natural frequency distribution, rather than simply attempting to control individual extreme events. The OPTI6 optimization model applies a genetic algorithm, coupled with a daily simulation model of the storm-water drainage network, to optimize the sizing and fuzzy operating rules of reservoirs for controlling storm-water discharges to the SLE. Results indicate that the desired frequency distribution is closely matched and the level of service for the supplemental irrigation demand is met under reduced storage requirements.

This study addresses the dynamic issue of optimum pumping and optimum target ending aquifer elevation. A stochastic dynamic programming (DP) approach is used to address this optimal elevation question. The analytical approach relies on the generation of stochastic state transition information from a stochastic nonlinear, static programming model of the aquifer. Methodologically, the model linkage allowed us to handle a large number of potential state variables by collapsing them into a simple ending elevation state variable in the DP.

Water resources optimization models often use spatial numerical models to approximate the physics of natural systems. The discretization of the numerical grids can affect their search for optimal solutions, in terms of both solution reliability and computational costs. Computational costs are particularly significant for population-based optimization techniques such as genetic algorithms (GAs), which are being applied to water resources optimization. To overcome these bottlenecks, this paper proposes multiscale strategies for GAs that evaluate designs on different spatial grids at different stages of the algorithm. The strategies are initially tested on a hypothetical groundwater remediation problem, and then the best approach is used to solve a field-scale groundwater application at the Umatilla Chemical Depot in Oregon. For the Umatilla case, the multiscale GA was able to save as much as 80% of the computational costs (relative to the GA that used only the fine grid) with no loss of accuracy, thus exhibiting significant promise for improving performance of GA-based optimization methodologies for water resources applications.

A series of uncertainty analysis methods was applied to investigate the propagation of parameter uncertainty within a coupled model system and to evaluate the effects of uncertainty on model outputs and decision-making processes. First-order error analysis showed that among a large number of model parameters, only a few significantly affected the variation in pollutant loads at the watershed outlets and concentrations in the receiving body of water, and the variation in pollutant concentrations is greater than the variation in pollutant loads. The uncertainty analysis regarding the loads and concentrations showed different patterns, underscoring the importance of a complete uncertainty analysis and the need for an explicit quantification of the errors associated with the predicted loads. Monte Carlo simulation showed that best management practice scenarios considered as a safe scheme based on a deterministic model could actually lead to a significant risk of violating the water quality standards when model uncertainty is considered. With a modeling framework that considers uncertainty, feasible alternatives can be evaluated and ranked based on their risks of exceeding the target water quality criteria.

New microscale techniques have become available to assist urban designers in better water management. Urban water management has focused on two different areas: storm water and water supply. The focus of storm-water management is shifting toward low-impact development, which emphasizes better management of urban storm water through reductions in postdevelopment runoff by increasing on-site infiltration, while water supply planning has been enhanced by the emergence of end-use demand management, especially outdoor irrigation. Implementation of these two objectives requires examination of processes at smaller scales in order to evaluate changes being contemplated at a parcel level. A modeling approach is presented that incorporates decentralized options for management of both storm water and urban water supply. Management options that can be evaluated with this approach include restrictive irrigation policies and rainwater harvesting. A simpler model based upon Soil Conservation Service hydrology is then calibrated to the more complex model using a commercially available nonlinear optimizer. A method for comparison of costs and benefits from a consumer perspective is presented.

A genetic algorithm (GA)-based method for the least-cost design of looped pipe networks for various levels of redundancy is presented in this paper. Redundancy constraints are introduced in the optimization model by considering the number of pipes assumed to be out of service at any one time. Using this approach, trade-off relationships between cost and redundancy are developed. The GA-based approach is computationally intensive, and implementations on a custom fault-tolerant distributed computing framework, called Vitri, are used to satisfy the computational requirements. The design methodology is applied to two water distribution networks of different sizes, and a comparison of the performance of the distributed GAs for the design problems is also presented. We conclude that a GA-based approach to obtaining cost-effective, redundant solutions for the least-cost design of looped pipe networks can be effectively used on a heterogeneous network of nondedicated workstations.

The frequency of water conflicts in riparian areas has increased in recent years, forcing water agencies to consider regulating water withdrawals from streams. In this paper, we evaluate permitting programs that would be part of such regulations. The research focuses on fixed- and fractional-flow-based permit systems, the former of which allows a constant volume of withdrawal similar to the practice in appropriative areas; the latter of which allows water abstractions proportional to streamflow. The evaluation is undertaken with respect to three performance criteria: Low flow frequency, users’ net benefit, and withdrawal interruption frequency. It is concluded that the fixed-flow-based permit program is able to provide a high net benefit but causes frequent minimum streamflow violations and withdrawal interruption due to its relatively simple withdrawal formula. The fractional-flow-based permit program is found to achieve a better tradeoff in preserving the natural flow regime, the net economic benefit, and convenience to water users and administrators.

This technical note develops a simple drought storage allocation rule to minimize evaporative and seepage water losses from a system of reservoirs. Such a rule might have value during a prolonged drought, when the value of lost water is likely to be particularly high. Typically, concentrating water storage in one or a few reservoirs reduces overall water losses compared to “balancing” storage among reservoirs. Paradoxically, concentrating storage during a drought tends to minimize reservoir surface area available for recreation, increasing recreation losses even as it minimizes water losses.