Journal of Water Resources Planning and Management

In cold-region environments, ice-jam floods (IJFs) can result in high water levels in rivers to overtop levees, leading to devastating floods. Since climatic conditions play an important role in ice-jam flooding, there is a growing concern among property developers, insurance companies, government agencies, and communities on future IJF probabilities, especially in the context of changing climate. This study presents a stochastic framework for future IJF risk assessment and hazard mapping. Future hydrological conditions derived from a physically based hydrological model (forced with meteorological inputs from the Canadian regional climate model driven by two atmospheric-ocean general circulation climate models) were coupled to a fully dynamic hydraulic river-ice model to evaluate ice-jam scenarios and subsequent backwater-level profiles for the 2041–2070 period along the Athabasca River at Fort McMurray in Alberta, Canada. The modeling results show that future IJF risk will be lower and flood inundations less severe due to projected changes in climate, hydrology, and ice phenology in the 2041–2070 period compared to the baseline period of 1971–2000. However, extreme IJF events are still probable and can flood a considerable area of the town.

The detection of contaminant intrusion into a water-distribution network (WDN) is a difficult issue due to uncertainty related to the type of injected contaminant, source location, and intrusion time. The placement of water quality sensors has received increasing interest in the last years, and it still represents an open problem and a great challenge for researchers and utilities. Efficient numerical techniques are needed to support any contamination warning system (CWS) design. These require a well-calibrated hydraulic model of the WDN and a great deal of information, both of which are often unavailable to water utilities. In addition, as the size of the WDN increases, the choice of effective sensor placement becomes a computationally intractable problem. This paper introduces a methodology to support water utilities in the design of an effective CWS without any use of hydraulic information, but just exploiting the knowledge of the topology of the WDN. To ensure a complete coverage of the network, the method relies on a priori clustering of the WDN and on the installation of quality sensors at the most central nodes of each cluster, selected according to different topological centrality metrics. The procedure is tested on a benchmark network and on a real WDN serving a town close to Naples, Italy. The solutions obtained with topological criteria are effective in terms of detection time, detection likelihood, redundancy, and population exposed through ingestion.

Water loss reduction is important in sustainable water resource management. As one of the main water loss control methods, early detection of hydraulic accidents in district metering areas (DMAs) has emerged as a research focus. This study presents a data-driven method for burst detection which consists of three stages: prediction, classification and correction. A prediction stage is used to improve accuracy of flow prediction, a classification stage utilizes multiple thresholds to make the method robust to time variation, and an outlier feedback correction stage allows consecutive detection of outliers. The proposed method was capable of triggering burst alarms with 99.80% detection accuracy (DA), 85.71% true-positive rate (TPR), and 0.14% false-positive rate (FPR) in simulated experiments, and 99.77% DA, 94.82% TPR and 0.21% FPR in synthetic experiments over a 10-min detection time in a real-life DMA. The identifiable minimum burst rate was as low as 2.79% of average DMA inflow. The proposed method outperformed the single threshold-based method, window size–based method, and clustering-based method. It provides a sensitive and effective solution for burst detection in water distribution systems.

Water management and infrastructure design depend on quantifying thresholds in minimum flows. Decreasing trends in low flows have been observed at many stream gauges in the Southeast US; however, a comprehensive quantitative assessment of regional trends and shifts in flow minima is lacking. This study examines trends and abrupt shifts in the annual minimum 7-day mean streamflow in the Southeastern US for four distinct time periods over the last century. A type II error analysis is conducted to evaluate the probability of erroneously declaring that a trend does not exist. A decreasing trend in low-flow magnitude is identified in 80% of the streamflow records. An abrupt shift in low-flow magnitude was identified in 50% of the gauge records, occurring predominantly around 1975–1985 and 1995–2005. Trend slopes indicate an accelerated rate of decline in low-flow magnitude over recent decades compared to the last 50–75 years. Where statistically significant trends are not identified, short record lengths ($<50\mathrm{years}$) and high variability in flow records result in a high probability of a type II error.

Motivated by the emerging need for long-term and sustainable reservoir management, this paper developed a multicriteria decision-making (MCDM) model of reservoir operation with consideration of balanced applicability in both the past and the future. The proposed method consists of four modules: (1) the multiobjective optimization model of reservoir operation is built based on historical streamflow and then solved by the evolutionary multiobjective direct policy search (EMOPDS) with the nondominated sorting genetic algorithm-II (NSGA-II); (2) an evaluation criteria matrix is made considering the benefits and risks of the same objectives under future climate change; (3) objective interrelationships are qualitatively identified by visual analytics and quantitatively described by the structural equation model (SEM); and (4) an objective-based space coordinate system (SCS) is established to calculate the evaluation criteria weights based on SEM results, from which the optimum Pareto solution is determined by a decision-making function. The Three Gorges Reservoir (TGR) was selected as a case study, which simultaneously provides hydropower generation (pow), ecology (eco), and water storage (stor). Results indicate that eco negatively interacts with both pow and stor; however, pow is positively correlated with stor. Moreover, the design of the SCS-MCDM model can significantly improve benefits in both pow and stor and slightly sacrifice benefit in eco, compared with fuzzy optimum selection (FOS)-1 and FOS-2. Generally, the SCS-MCDM model outperforms FOS-1 and FOS-2 in terms of benefits for past and future and robustness in operation process under climate change. Therefore, the SCS-MCDM model facilitates decision makers to make a sustainable and balanced decision for reservoir management under uncertain climate change.

With increasing pressure on water resources availability and dependability and constraints due to environmental concerns, the traditional approaches for defining reservoir management rules are often inadequate. In particular, in multireservoir systems, when multiple input variables (e.g., the storage of other reservoirs in the system, water demand in different districts) must be taken into account, it is almost impossible to figure out which shape the operating rule(s) could have. For these reasons, neural network (NN) based rules have been increasingly adopted in the last decade. NN-based rules are well known as universal approximators that can help determine the most interesting input variables, their mutual relations, and how they contribute to the definition of the optimal releases. Two approaches to the identification of neural management rules are discussed in the paper. The first solves a deterministic open-loop (i.e., with known inflows) problem and then identifies neural closed-loop policies using the classical regression method, so that the rules approximate as much as possible the solution found in the first step. The second approach, direct policy search, assumes that the operating rule is represented by an NN, the parameters of which are optimized directly by solving the optimal closed-loop problem. This work applies the two approaches to the case of the downstream portion of the Nile River basin system, which contains some large reservoirs, and for which several years of synthetic streamflows are available. The comparison of the two approaches highlights intrinsic differences, showing the benefits and disadvantages of each. In the specific case of the Nile, the first approach performs better in terms of global agricultural deficit and hydropower production.

Modern drip-irrigation technologies improve water-use efficiency while simultaneously transforming areas that are not otherwise irrigable in practice (too distant or too high to be reached by surface waters). Although drip irrigation is expanding rapidly in India, adoption remains low in neighboring Pakistan. To gain deeper insight into the factors constraining adoption of drip irrigation in Pakistan, a discrete choice experiment framed around the hypothetical subsidized purchase of a drip-irrigation system in four districts within Pakistan’s Punjab Province was used. The results show higher valuation of drip systems among new users, which suggests that limited technical support and upstream maintenance facilities are not posing significant barriers to drip-irrigation adoption. It was observed that aspirations for cropping systems under drip were better predictors of farmers’ valuation for drip systems than current cropping patterns, implying that a different agricultural landscape might reasonably emerge under more widespread adoption of drip. Both aspirations were observed for high-value crops such as fruits, as well as lower-value crops such as wheat, under drip systems, suggesting a number of ways through which drip irrigation may transform Pakistan’s agricultural landscape.

Seawater intrusion is a common problem in coastal aquifers and can be intensified by various factors, such as unplanned groundwater extraction. Hence, making appropriate decisions for coastal groundwater management is essential to prevent excessive seawater intrusion, although the uncertainties of this procedure can undermine the reliability of obtained results and consequently affect proposed decisions. To address such challenges, this study presents an uncertainty-based decision model based on a combined simulation-optimization framework and parallel processing technique. In this model, the SUTRA numerical simulation model is combined with an evolutionary algorithm of continuous ant colony optimization, while Monte Carlo simulation is used to consider the uncertainties in the permeability of porous media. Two sets of probabilistic management problems are formulated for maximizing groundwater extraction and controlling the seawater intrusion in a real case coastal aquifer of Ajabshir, Iran. The results indicate that a 46% increase in groundwater extraction is possible for a low-exploitation-stress period of the year. Moreover, using a parallelization technique with 15 threads, the computational time of the decision model is cut in half.

With the development of the power market in China, the operation of hydropower stations will be changed significantly. In a competitive environment, each hydropower station or station group tries to improve its own net benefits, compared to maximizing a power grid-level objective with centralized scheduling. To analyze the impact of market reformation on hydropower operations, both noncooperative and cooperative sampling stochastic dynamic programing game (SSDP-game) models are proposed in this paper to obtain monthly operation policies under a competitive environment. The models are tested using data for the hydropower stations of Longtan, Xiaowan, and Goupitan and a hypothetical demand curve. The results show the potential increase in net benefits and potential energy loss of each station, along with the influential factors of the gaming approach. The seasonal regulation demands, typically addressed through firm power constraints, are better achieved through market competition, with 1.0%–5.0% energy losses at individual stations.

Incorporating streamflow forecasts into reservoir operations can improve water resources management efficiency, yet the forecast value in multipurpose reservoir systems is rarely investigated, let alone the relationship between forecast accuracy and value in multiobjective reservoir operation. Here, we propose a forecast-informed framework to derive multiobjective operating rules based on radial basis functions and the Pareto archived dynamically dimensioned search optimization algorithm and subsequently develop indicators reflective of Pareto fronts with and without forecast information to characterize forecast value. Based on a case study of the Hanjiang cascade of reservoirs in the Yangtze River Basin, China, the optimal inclusion of streamflow forecasts notably improves the performance of multiobjective reservoir operations, mainly by significantly increasing the hydropower generation. The relationship between forecast accuracy and value is explored by comparing four accuracy indicators (Nash–Sutcliffe efficiency, mutual information, correlation coefficient, and Kullback–Leibler distance) and forecast value. The correlation coefficient is found to be the most suitable forecast indicator given its high correlation with forecast value and stability in the regression. For multiobjective forecast-informed reservoir systems, it is critical to understand and define the relationship between forecast accuracy and forecast value; if improvements in accuracy lead to steep gains in value, investing in further forecast model development may be warranted.

Water distribution network (WDN) resilience depends on both the underlying flow dynamics and the network topology. Although existing literature has examined a variety of complex network centrality and spectral properties, very little attention has been given to understand multiscale flow-based feedback loops and their impact on overall stability. Here, high-dimensional resilience arising from multiscale feedback loops is inferred using a compressed one-dimensional proxy measure called trophic coherence. This is hypothesized to have a strong impact on both the pressure deficit and the water age. The results show that trophic coherence is positively correlated with the time to recovery (0.62–0.52), but it is negatively correlated with the diffusion of a disruption ($-0.66$ to $-0.52$). Finally, random forest analysis is used to combine resilience measures, showing that the new resilience ensemble provides a more accurate measure for networked resilience.

Condition assessment of water pipelines using fluid transient waves is a noninvasive technique that has been investigated for the past 25 years. Approaches to identify different anomalies and to identify elements of the topology of a pipeline have been proposed but often require detailed modeling and knowledge of the system. On the other hand, artificial neural networks (ANN) have become a useful tool in a range of different fields by enabling a computer to solve a problem without being explicitly programmed to do so, but rather by learning from a series of known examples. This paper presents a new methodology that uses ANNs to predict the presence of features in a pipeline. First, the location and characteristics of a junction have been predicted as a way to identify elements of the topology of a pipeline followed by identification of the location and sizing of a leak. The ANN characteristics and training approaches have been determined for both the junction and the leak example. Results show that the ANN that has been designed for this research is able to accurately predict the location of a junction with an error in this estimation of 2.32 m (out of a 1,000 m long pipeline) or less in 95% of the tested cases. The prediction of the two different diameters on either side of the junction was extremely accurate with only one misidentification of one of the diameters in the 5,000 tested examples. When the ANN was trained and tested to locate and size a leak, the results were also successful. A total of 95% of the tested examples located the leak with an error equal or less than 3.0 m (out of a 1,000 m pipe length) and the leak size was predicted with an average absolute error of only 0.31 mm. The results presented in this paper demonstrate the potential of combining the use of both fluid transient pressure waves and ANNs for the detection of features in pipelines.

Over the past few decades, the development of municipal societies along flood-prone areas such as waterfronts and riverbanks has increased rapidly, causing flood-induced disasters more often than ever before. Flood defense systems have defaulted to gray infrastructure for many centuries. However, more recent storms such as Superstorm Sandy revealed the shortcomings of gray solutions, and therefore the notion of adaptability has been directed to the use of green infrastructure. In flood risk management, nature-based strategies have already proven considerable flood-reducing potential while effectively conserving the ecosystem. However, while nature-based approaches are extremely beneficial in preserving the ecological characteristics of the region, they cannot be solely relied on when it comes to rarely occurring events such as extreme floods. Best management practices (BMPs) are one of the most commonly used safeguards in which natural elements can be integrated into traditional infrastructure designs. In this paper, different strategies are proposed to highlight the immense effects of combining green and gray approaches on both flood mitigation and prevention cost. These strategies focus on flow through vegetation (dense plants), which could be patterned as flow through porous media; and were applied to Westchester Creek, New York. Suggested conceptual designs were then simulated in a hydrologic model. Flood inundation maps generated through the modeling process were then compared in five different scenarios including before and after BMPs. The results of this study showed how the combination of levee and vegetation could significantly lower surge effects while proving to be cost-effective. The methodology outlined in this paper can be used for other urban coastal settings to plan for flood preparedness.

Municipal water utilities commonly manage drought by encouraging conservation. Efforts to temporarily reduce water consumption can cost significantly less than investments in infrastructure to increase supply. However, reduced water sales during an extended drought can hinder a utility’s ability to recover its (mostly fixed) costs, potentially leading to lower credit ratings and higher bond financing rates. Drought surcharges have emerged as an increasingly common solution to this problem. Compared to typical restriction-based policies (i.e., residential irrigation mandates), pricing incentives increase the economic efficiency of conservation while also mitigating some or all lost utility revenue. However, drought surcharges can also provide utilities with financial disincentives for pursuing cost-effective alternatives to conservation, like water transfers or wholesale (treated water) purchase agreements, when they are available. Wholesale water purchases between neighboring water systems are playing a more prominent role within urban water management plans, particularly in growing metropolitan regions of the southeastern United States. By reducing the need for conservation during drought, wholesale purchases can help utilities avoid welfare losses associated with consumption reductions. Using simulations of changing consumption patterns obtained from the detailed billing data of a mid-sized water utility, we found that the use of surcharges can cause conservation to appear more financially attractive to utilities than wholesale purchase agreements, even when the consumer welfare losses from conservation are larger than purchase costs. Results illustrate how the pursuit of efficiency with respect to one policy option (i.e., conservation) can alter the incentives to integrate that policy with a wider range of potentially beneficial alternatives.

As a large open lake that connects to the lower Yangtze River, the Poyang Lake, China, has persisted dramatic water level decline since 2003. For better management of lake water resources and ecosystem, this study quantitatively examined the contributions and spatiotemporal differences of compounding influencing factors to this hydrological change. Attempts were achieved through the reconstruction of lake water level series by two combined neural network models. The results indicate that with reference to the period 1980–1999, the average contributions to lake decline during 2003–2014 of lake-bottom topography change, the operation of the Three Gorges Dam (TGD), and climate change and other human activities over the Yangtze River basin were 50%, 18%, and 32%, respectively. The response of the lake water level to the three factors shows obvious spatiotemporal differences due to different influencing mechanisms. It is worth noting that the effect of lake-bottom topography change is still increasing across the lake. In addition, the reduction of precipitation in the Yangtze River basin should be highly correlated to the water level decline of the lake. From the perspective of maintaining a certain water level of the lake and preventing continuous decline and shrinkage, constructing the proposed water control structure would be an effective solution.

Concerns over drought-related financial risk have been growing for water utilities as adaptive measures, such as conservation (which reduces revenues) and water transfers (which increase costs), play larger roles in drought management. Recent research suggests that both utility-specific and generalized forms of index insurance can play an effective role in managing a water utility’s drought-related financial risk. However, any entity offering these contracts to the water utility sector, whether a third-party insurer or a “mutual” formed by multiple utilities, will need to manage the “tail” risks of severe or widespread drought impacting many utilities simultaneously, which has the potential to drive high aggregate financial losses. When managing these risks, this entity must balance the opportunity costs of capital associated with higher reserve funds against the costs of mechanisms for transferring this tail risk (e.g., reinsurance). This research characterizes the financial risks of drought for a mutual, composed of surface water utilities distributed across all 344 climate divisions of the contiguous United States, that insures utilities against drought-related financial losses using contracts based on the Palmer Hydrologic Drought Index (PHDI). Tradeoffs associated with the mutual’s risk management strategy involve varying levels of reserves and reinsurance. Results show that risk pooling via a mutual combined with reinsurance coverage can reduce the average annual net cost of risk management for member utilities by 7.5% while significantly reducing the average total net cost over 10-year periods by 17%.

Public water utilities face uncertain decisions every day in their efforts to meet drinking water needs of customers. Real-time decision-support tools (DST) are often used by water managers to solve a variety of water management challenges, including meeting customers’ demands, forecasting floods, and developing reservoir operating rules. The incorporation of seasonal forecasting can improve operational decision making by explicitly including uncertainties that affect these near-term decisions. This study presents an application of DST that incorporate rainfall/streamflow uncertainties, seasonal demand forecasts, and system operational constraints to assist utility decision-makers. Large-scale climate information is used in monthly precipitation forecasts using a hidden Markov-chain model. An ad hoc seasonal demand forecasting model considers weather conditions explicitly and socioeconomic factors implicitly. The seasonal system operation is modeled as a mixed-integer optimization problem that aims at minimizing operational costs. It embeds the flexibility of incorporating operational rules at different components, e.g., surface water treatment plants, desalination facilities, and groundwater pumping stations. The proposed framework is illustrated for a water supply agency in the southeastern United States, Tampa Bay Water. The use of the tool is demonstrated in providing operational guidance for taking a large storage reservoir offline for a two-week period to conduct a required inspection. The results provided insights for the best time to take the reservoir offline and yet meet an operational objective of filling the reservoir by October 1. Although this application is illustrated for Tampa Bay Water, it demonstrates the use of DST for regional water management in other areas.

EPANET 2 software has been widely used among researchers and practitioners from its inception for hydraulic analysis. This technical note presents a method to simulate pressure-dependent volume-based demands in a single extended-period simulation run of EPANET 2. In the proposed method, artificial strings made up of a pipe with a minor loss, pressure-sustaining valve, pipe of negligible resistance, and tank are added to the demand nodes before running the hydraulic analysis. The time required to satisfy a volume-based demand is determined as the time required to fill the artificial tank connected to the demand node for normal and pressure-deficient conditions. All the simulations were carried out using the available functions of the graphical user interface of EPANET 2. The proposed method was applied to a two-source looped network available from the literature and additionally compared with an existing method available in the literature using two example networks.

This study discusses practical implications of considering unit head loss in different pipe sizes and in different locations of water distribution systems (WDSs) with regard to operation and maintenance. By visualizing unit head loss (using the Hazen-Williams relationship) in pipes obtained from 18 WDSs in North America, changes in unit head loss are put into perspective in different pipe sizes and different WDS locations. The results suggest that the importance of diameter is greater than that of the Hazen-Williams roughness factor, that flow rate plays a more important role than diameter in determination of head loss in pipes closer to water sources, and that diameter seems to be more important than flow rate in pipes at the periphery. Moreover, aging, tuberculation, and subsequently reduction in effective diameter can have a more critical effect on head loss in smaller pipes at the periphery of a system. Finally, effects of water conservation and pump scheduling in different locations of the network, as far as head loss is concerned, can potentially be more evident on larger pipes closer to the water source and in some cases on smaller pipes at the periphery. Therefore, it is suggested that network-level energy management decisions can have different effects on different pipe sizes in different locations.

Hydraulic and quality models are widely used in water distribution networks (WDNs) for assisting network analysis, design, and operation. In a calibrated model, the difference between the simulated and measured data reaches an acceptable error level. In this study, to calibrate the hydraulic and quality model of the network, the pressure and residual chlorine levels were measured initially during 2-h steps in the network. Furthermore, the bulk chlorine decay values were measured using the laboratory tests. Then, hydraulic and quality calibrations were carried out simultaneously using particle swarm optimization (PSO) as the optimization tool, Hazen-Williams roughness ($C_{H-W}$) and fitting coefficients as the decision variables, and mean absolute percentage error (MAPE) as the objective function. According to the results, the PSO algorithm was highly accurate in determining the $C_{H-W}$ and could minimize the difference between the observed and simulated values.

Water distribution system models have long been widely used for design and planning purposes. Their application for supporting real-time operational decisions has been also gaining increasing interest over the past decade. Accurate end-user nodal demands are critical to the reliability of hydraulic simulations for real-time decision support. Conventionally, nodal demands are set to a handful of periodically updated ensembles of demand patterns, which cannot represent the vast heterogeneity and volatility of demands. With advances in metering technology, consumption data with unprecedentedly high temporal and spatial resolutions are available to water utilities on a real-time basis. A framework is developed here to create a dynamic demand assignment hydraulic model, in which consumption data are assigned to nodes to update the water network model with the streaming data from the data center and without interruption of the hydraulic simulation run. The developed framework is cloud-based and scalable, making it suitable for water distribution systems of all sizes. The framework modifies the core EPANET engine to directly assign updated demands in order to overcome current software limitations. The model is applied and demonstrated using a real-world case study in the US. The results show the importance of the real-time demand assignment for the reliability of hydraulic models for making real-time operational decisions and the realization of digital twins of water infrastructure systems.