Journal of Water Resources Planning and Management

Considerable research has been carried out on the optimization of water distribution systems (WDSs) over the last three decades. In previous research, attention has mainly focused on the minimization of cost, due to the high expenditure associated with the construction and maintenance of such systems. However, the impacts of WDSs on the environment usually have not been considered adequately. The recent increasing awareness of sustainability and climate change, especially global warming, has led to research where greenhouse gas (GHG) emissions are considered. In the study described in this paper a multiobjective genetic algorithm for WDS optimization has been used as an explorative tool to investigate the trade-offs between the traditional economic objective of minimizing costs and an additional environmental objective of minimizing GHG emissions. The impacts of minimizing GHG emissions on the results of WDS optimization have been explored for a case study in this paper. The results indicate that the inclusion of GHG emission minimization as one of the objectives results in significant trade-offs between the economic and environmental objectives. Furthermore, a sensitivity analysis has been conducted by using different discount rates in a present value analysis for computing both ongoing costs and GHG emissions. The results obtained show that the Pareto-optimal front is very sensitive to the discount rates used. As a result, the selection of discount rates has a significant impact on final decision making.

The upper Blue Nile Basin in Ethiopia harbors considerable untapped potential for irrigation and large-scale hydropower development and expansion. Numerous water resources system models have been developed to evaluate these resources, yet often fail to adequately address critical aspects, including the transient (e.g., filling) stages of reservoirs, relevant streamflow retention policies and downstream consequences, construction staggering, and the implications of stochastic modeling of variable climate and climate change. This omission has clear economic impacts on benefits and costs and could be pivotal in national policy and decision making. The Investment Model for Planning Ethiopian Nile Development dynamic water resources system model is outlined and applied to address these aspects. For the hydropower and irrigation development projects specified, model results disregarding transient and construction stagger aspects demonstrate overestimations of $6 billion in benefits and 170% in downstream flows compared to model results accounting for these aspects. Benefit-cost ratios for models accounting for transient conditions and climate variability are found to range from 1.2–1.8 under historical climate regimes for the streamflow retention policies evaluated. Climate change scenarios, represented either by changes in the frequency of El Niño and La Niña events or by the Special Report on Emissions Scenarios projections, indicate potential for small benefit-cost increases, but also reflect the potential for noteworthy decreases, relative to the historical climate conditions. In particular, stochastic modeling of scenarios representing a doubling of the historical frequency of El Niño events indicates benefit-cost ratios as low as 1.0, even under perfect foresight optimization modeling, due to a lack of timely water. However, even at this ratio, Ethiopia, at current growth rates, may still be unable to absorb all the potential energy developed, reinforcing the need for significant economic planning and the necessity of securing energy trade contracts prior to extensive expansion.

The South Florida Water Management District (SFWMD) has a large and expanding surface water level-monitoring network for various water bodies including lakes and streams. SFWMD seeks to optimize the existing network of monitoring stations to collect the data from the least number of monitoring stations in the best possible locations without compromising information. Tabu combinatorial search algorithm is used to optimize water level-monitoring stations in lakes and streams within the SFWMD. Given a network of $n$ stations, the methodology involves searching for the best spatial combination of $r$ (where $r<n$ ) monitoring stations within the existing network to estimate water surface levels in lakes and streams within a given tolerance. Separate techniques are used to compute water levels in lakes and streams to compare possible solutions from Tabu search with observed daily data. Results from application of proposed techniques to Kissimmee River basin show that implementation of Tabu search within geographic information system (GIS) provides a computationally efficient way of optimizing a water level-monitoring network. It is also found that factors such as length of data records, slope of hydraulic profiles, and data at control structures play a significant role in the overall optimization process. The proposed methodology can also be used in the design of new monitoring networks and other water resources applications that involve GIS, computational modeling, and optimization.

The Lower Arkansas River (LAR) Basin in Colorado, like many intensively irrigated river basins in the Western United States, faces a variety of problems associated with inefficient irrigation, seepage from earthen canals, and inadequate drainage facilities. Upward flows from high water tables have salinized and waterlogged agricultural soils of the Valley, contributing to reduced crop yields and nonbeneficial water consumption on adjacent uncultivated lands. River water quality has also suffered since intensive irrigation of alluvial soils results in evaporative concentration and the accelerated dissolution of inherent salts and other mineral pollutants into the underlying aquifer, appearing as return flows that threaten the ecological health of the river. A geographic information system-based river basin decision-support system (River GeoDSS) has been developed and applied to the LAR Basin for assessing basinwide strategies for improving agricultural productivity, salvaging water from nonbeneficial consumptive use, and reducing solute concentrations while maintaining compliance with decreed water rights and the Arkansas River Compact between Colorado and Kansas. Development and calibration of River GeoDSS, based on extensive field data collection programs in the LAR Basin, is described herein, with a companion paper detailing its application to evaluation of a wide range of agroenvironmentally focused water management strategies.

Research conducted at the field and regional scales in the Lower Arkansas River (LAR) Valley of Colorado has identified water management alternatives with potential for enhancing agroenvironmental conditions in the basin by reducing waterlogging and soil salinity, salt loadings to the river, and nonbeneficial evapotranspiration in the irrigated stream-aquifer system. The LAR geospatial decision support system (GeoDSS), presented in a companion paper as a customized version of the generalized River GeoDSS, is applied to the evaluation of the feasibility and performance of water management strategies at the basin scale. The LAR GeoDSS allows comparative evaluation of management options for improving irrigation efficiency, minimizing water shortages, and improving water quality at selected control points by augmenting groundwater return flows through dynamic regulation of reservoir releases to abide by legal and administrative constraints on river operations. Results show that conditions favorable to increased agricultural productivity and water conservation can be accommodated, along with the benefits of improved river water quality through reduction of excess irrigation recharge and canal seepage and augmented subsurface drainage, without violating existing water rights and the Colorado-Kansas Interstate Compact Agreement.

Actual projections provided by climate models suggest that the probability of occurrence of intense rainfall will increase in a future climate due to increasing concentrations of greenhouse gases. Considering that the design of urban drainage systems is based on statistical analysis of past events, an increase in the intensity and frequency of extreme rainfall events will most probably result in more frequent flooding. The design criteria must therefore be revised to take into consideration possible changes induced by climate change. A procedure is proposed to revise the design criteria of urban drainage infrastructures. This procedure integrates information about (1) climate projections for extreme rainfall over the region under consideration; (2) expected level of performance (or acceptable level of risk); and (3) expected lifetime of the infrastructure/system. The resulting design criterion ensures that the service level remains above the selected “acceptable” level over a predefined portion of the infrastructure lifetime. It is argued that the definition of new design criteria should be part of a global adaptation strategy combining various measures to maintain an acceptable level of service in a long-term perspective. Defining this level of service is however a challenge in a context where uncertainties on projected changes in intense rainfall are still important.

In the present work, we locate sensors in water distribution networks and make inferences on the presence of contamination events based on sensor signals. We fully consider the imperfection of sensors, which means that sensors do provide false positive and false negative signals, and we propose a two-stage model by combining a facility location model with Bayesian networks to (1) identify optimal sensors locations and (2) infer the probability of the occurrence of a contamination event and the possible contamination source based on sensor signals, the probability of a contamination event being detected by the sensors given that there is a contamination event, and the probability of detecting a contamination event given that there is actually no such event (overall false positive rate). This two-stage model can also be used to construct the trade-offs between the number of sensors and the power (the false negative and false positive rates) of individual sensors while guaranteeing the performance (the probability of detecting random contamination events) of the sensor network system (all the sensors). The method can be generalized to address similar problems in deploying sensors in harsh environments.

Reservoir flood control operation (RFCO) is a large scale multiobjective problem with complex constraints that require powerful algorithms to solve it. As a new metaheuristic evolutionary algorithm, shuffled frog leaping algorithm (SFLA) has the potential ability to solve multiobjective optimization problems because of its group evolution characteristic. In this paper, we present a novel multiobjective shuffled frog leaping algorithm (MOSFLA), which incorporates an archiving strategy based on self-adaptive niche method to maintain the nondominated solutions, and improves the memetic evolution process of SFLA to adapt to the multiobjective optimization problem. The numerical experiments of five Zitzler-Deb-Thiele functions indicate that MOSFLA yields better-spread solutions and converges closer to the true Pareto frontier than non-denominated sorting genetic algorithm (NGSA)-II and SPEA2. Furthermore, MOSFLA is applied to solve RFCO of the Three Gorges Project, and the results demonstrate that this algorithm can generate a solution set with uniform spread and good convergence for the problems with two conflicting objectives, including minimizing the highest reservoir water level and minimizing the peak flood discharge. Additionally, if compared with dynamic programming and NGSA-II, MOSFLA is verified to be more efficient and competitive, and thus can be provided as a new effective alternative for solving the complex reservoir operation problems.

In the last three decades the optimal design of water distribution systems problem has been studied by a great many researchers, and this has resulted in the development of a large number of models and the application of optimization techniques. The design of these infrastructures is based on future predefined and perfectly known working conditions for the water distribution networks, a premise that may direct the optimization process to solutions which, although optimal for the imposed scenario, may perform badly if reality turns out to be significantly different. In fact the working conditions can be disrupted by accidents such as broken pipes or reservoirs, technical failures, change in demand, etc. In the context of a proactive attitude toward risk, it is important to consider these aspects at the design phase. This paper presents a robust optimization-based approach for designing a water distribution network aimed at obtaining solutions that can cope with the uncertainty of the network’s working conditions. Robust optimization is a scenario based technique, and in the present case its goal is to provide significant savings in comparison with the worst case scenario solution, while incurring only minor suboptimality for the rest of the scenarios, and being almost feasible for all the scenarios. The solution of the robust optimization problem is obtained with a methodology that uses the link between a simulated annealing algorithm, the optimizer, and a hydraulic simulator, here used to solve the problem’s hydraulic constraints. The application of the proposed methodology is illustrated with an example of a water distribution network.

Time domain continuous simulation models using simple or detailed process models are routinely used to analyze storm-water control options. The interest in decentralized controls has generated the need to develop tools that can evaluate performance when response times are very short and may even be shorter than the available precipitation data. The effect of storage is much less in infiltration based systems so simpler frequency domain approaches can be useful. A simple, but reasonably reliable, model has been developed on the basis of system response time and frequency analysis using long-term precipitation data. In the model, the performance of a storm-water management facility is related to the loading rate of storm water onto the facility. Performance is estimated using long-term precipitation data, catchment information, and storm-water management facility specification. The developed frequency model provides a simple and relatively accurate way to estimate the overall performance of infiltration best management practices (BMPs) over a range of storm events. It can be calibrated using local field data and a more detailed time domain process simulator and then used to provide a macrolevel evaluation model that can be used to find the overall optimal control strategy.

The current paper aims at the assessment of the effectiveness of a popular transient-based technique—inverse transient analysis (ITA)—for leak detection in water pipe systems using physical data. Data were collected from two pipe systems made of polyethylene: a laboratory facility at Imperial College London and a quasi-field system at Thames Water Utilities. A hydraulic transient solver especially developed for viscoelastic pipes (typically made of polyethylene) and an optimization algorithm for the simultaneous creep calibration and leak location have been developed for the application of this technique. The evaluation of the presence, location, and size of leaks was carried out using collected data. ITA allowed for the identification of the approximate location of leaks as long as the physical characteristics of the pipeline are well known, the transient is generated by a fast change of flow conditions, the leak has a “reasonable size,” and the transient solver is accurate enough to describe the transient event.

Factor analysis is used to develop an integrated indicator system that evaluates the state of watersheds from three perspectives (subindexes): water-use management, flood management, and environment and ecosystem management. The system, which has a total of nine indicators, is applied to evaluate the state of watershed management in the Han River basin, South Korea. For all the subindexes, preliminary tests, namely, the Kaiser-Meyer-Olkin test and the Bartlett test of sphericity, are conducted to evaluate the suitability of the data collected for factor analysis. The variables related to each subindex are then grouped into three factors (indicators) with the variables’ statistical characteristics taken into account. The results of the application of the system to the case study show that the indicator system reasonably depicts the state of the watersheds in the Han River basin and can be effectively applied to resolve concerns in watershed management; the system is one of the inevitable elements in the adaptive watershed management process. Examining the correlations between the indicators proves that the selection of the indicators and their variables is appropriate from a statistical perspective. In addition to providing evaluation information about the state of watersheds from multiple perspectives, the system can be used as a preprocessing step for watershed management modeling regarding the choice of watershed variables within the context of river basins.

The continuous growth of urban areas and the increasing public awareness of the environmental impacts of storm water have raised interest on the quality of the receiving water bodies. In the past two decades, many efforts have been directed at improving urban drainage systems by introducing mitigation measures to limit the negative environmental impacts of storm water. These mitigation measures are generally called best management practices (BMPs), sustainable urban drainage systems, or low impact developments, and they include practices such as infiltration and storage tanks that reduce the peak flow and retain some of the polluting materials. Choosing the best mitigation measure is still a controversial topic. To gain insight on the best technique, this study compares different distributed and centralized urban storm-water management techniques, including infiltration and storage facilities. The main objective of this study is to use modeling to assess the effects of the different urban drainage techniques. To this end, a homemade model that was developed in previous studies is applied. This model enables us to simulate both combined sewer systems and ancillary structures such as storm tanks or infiltration trenches to determine water quantity and quality characteristics. A long-term simulation is employed to account for the effects of sediments in BMPs, which generally reduce the hydraulic capacity. The results allow us to draw some conclusions on the peculiarities of BMP techniques, on the possibility of integrating different techniques for improving efficiency, and on BMP maintenance planning.

The paper describes the development of a DENET computer model that involves the application of an evolutionary optimization technique, differential evolution, linked to the hydraulic simulation solver, EPANET, for optimal design of water distribution networks. A model is formulated with the objective of minimizing cost and this formulation is applied to two benchmark water distribution system optimization problems—New York water supply system and Hanoi water distribution network. The study yielded promising results as compared with earlier studies in the literature and encouraged to reformulate the model for a new objective of maximizing network resilience. The results of the analysis demonstrate that DENET can be considered as a potential alternative tool for economical and reliable water distribution network planning and management.

The suitability of genetic algorithms (GAs) for the optimization of water distribution systems (WDSs) has been demonstrated extensively. However, despite many years of application in many different fields, the selection of the GA parameters remains a difficult and time consuming task. In this paper, two methodologies that do not require trial-and-error GA parameter calibration have been tested on a WDS optimization problem to determine their suitability for application in the water resources field and to assess their ability in locating near-optimal solutions. The results indicate that both approaches located solutions that were significantly better than a GA using typical parameter values, while the methodology based on convergence of the GA population located the best solutions overall. This method can be easily applied to assist GA users in identifying suitable GA parameters without requiring a time consuming trial-and-error approach.