Journal of Water Resources Planning and Management

Online monitoring of multivariate water quality data is becoming a practical means of improving distribution network management and meeting water security goals. Changes in water quality are often due to changes in the hydraulic operations of the network. These operational changes create patterns of water quality change that are similar, but not exactly the same, from one instance to the next. Classification of multivariate change patterns through trajectory clustering is introduced in this paper to create a pattern library from historical water quality data and as an online process with the goal of reducing false positive water quality event detections. Prior to event declaration, a short sequence of the preceding multivariate data is compared against the pattern library to assess its similarity to a previously observed pattern. A fuzzy clustering algorithm is utilized to assign multivariate pattern memberships for water quality patterns associated with water quality events in both the offline and online modes of operation. The utility of trajectory clustering for multivariate pattern recognition in time-series data is demonstrated with two example applications. The first example uses observed water quality with simulated patterns and events. The pattern matching reduces the number of false positive event detections by 91% relative to the case of not using the pattern matching. The same false positive event reduction is achieved when both patterns and separate water quality events are added and 100% event detection is achieved. The second example uses observed water quality data from a metropolitan distribution system in the United States. The pattern matching approach developed in this paper is able to reduce the false positive event detections by 68%.

Bayesian analysis has application to probabilistic source characterization in water distribution systems. A new implementation of Markov-chain Monte Carlo (MCMC) for this problem is described. The solution addresses the discrete nature of water distribution networks that precludes the application of MCMC methods of general applicability that have been reported elsewhere in the water resources literature. The method is applied to a hypothetical network that has been used by others to test source identification methods. The likelihood function, a key component of Bayes’ rule, is evaluated using a Monte Carlo–based stochastic water-demand model. The results reinforce the need to address the multiple sources of uncertainty in the source characterization, including the stochastic variation of water demand. Further research is needed to make the approach feasible in operational environments. Limitations of the approach and future research directions are discussed.

The true principal factors for the diurnal water-demand pattern of urban water are often difficult to identify using traditional rough-set algorithms because the demand pattern is usually affected by many factors that are uncertain and hard to quantify. An improved attribute-reduction algorithm based on the cumulative weighting coefficient was proposed to solve this problem. The weighting coefficient was determined by the result of the variable precision rough-set algorithm. To discuss the consecutive curves with mathematical tools, an improved fuzzy c-mean (FCM) algorithm was proposed to discretize the diurnal water-demand pattern spatially. The proposed algorithms were then used to analyze the principal factors of the diurnal water-demand pattern in the city of Hangzhou, China. The results show that the improved attribute-reduction algorithm is capable of distinguishing the false attribute from the dynamic reduction sets, and the fuzzy c-mean algorithm is an effective and feasible method of solving the cluster problem for the consecutive curves. The principal factors affecting the diurnal water-demand pattern in Hangzhou are maximum air temperature, minimum air temperature, and weekday or weekend.

Traditional water distribution simulations, known as demand-driven analysis (DDA), are normally analyzed under the assumption that nodal demands are known and satisfied. The DDA is valid under normal conditions and in design requirement. However, in situations such as pipe burst or pump outage, the outflows at nodes affected by low pressures will decrease. Therefore, network simulation under deficient pressure conditions using conventional DDA can cause large deviation from actual situations. The purpose of this paper is to analyze different compositions of nodal outflow, including volume-driven demand, pressure-dependent demand and leakage to simulate deficient network performance more realistically. An extension of the modeling package EPANET, which implements repetitive modifications to nodal outflows based on pressure-dependent demand formulations and leakage models (EPANET-MNO), is first developed. Then, a comparison is made between the respective performances of four different pressure-dependent demand functions with the designated required pressure and minimum service pressure. The EPANET-MNO was verified in two abnormal situations, including fire flow and pipe failure isolation in steady simulation. Finally, pipe failure isolation in extended period simulation was carried out on a real network. This paper demonstrates that good modeling performance and convergence of the EPANET-MNO can be achieved for simulating deficient-networks considering volume-driven demand, pressure-dependent demand and leakage.

The water quality capture volume (WQCV) is a design parameter that is used for sizing urban storm-water treatment facilities. The storage, treatment, overflow, and runoff model (STORM) in NetSTORM was used to translate historical hourly rainfall records into storm event runoff data and to compute the WQCV (set at the 85th percentile capture of the annual runoff volume) of eight cities located in various climatological areas of South Korea. For each city, one-parameter and two-parameter exponential distributions were fitted to the set of storm event runoff volumes that were computed by STORM, and the 85th percentile runoff volume that was determined from the STORM was compared to the 85th percentile runoff volumes that were obtained from the two exponential distributions. The results demonstrate that the more easily applied distribution approach is adequate in South Korea for determining the WQCV. In Seoul, which is a city located in the northwestern part of South Korea, the two-parameter exponential distribution yielded the best estimate of the WQCV computed by STORM. For Chuncheon, Wonju, and Daejeon, which are cities located in the north-central and central parts of South Korea, the WQCVs were accurately estimated using both the one-parameter and two-parameter exponential distributions. Finally, in the cities of Gangneung, Gwangju, Daegu, and Busan, which are located in the northeastern and southern parts of South Korea, the WQCVs were best estimated using the one-parameter exponential distribution. It was found that seasonal variations in rainfall and rainfall event depth are important factors in assessing the WQCV. This study also found that the one-parameter exponential distribution approach, which is the most common approach used in the United States, may not be adequate for WQCV estimation if the standard deviation for average monthly rainfall is significant.

Optimal assessment of the overall state of the degradation affecting sewer systems requires consideration of both the hydraulic and the structural aspects. This work identifies a set of 15 factors—pertaining to both degradation aspects—as primary and constituting the framework for hierarchical structuring of the condition states. The relative importance of these factors was determined by the analytic hierarchy process (AHP). The proposed model was successfully applied to assess the current degradation state of the sewer network in Saint-Hyacinthe, Quebec. The relative significance effect, assessing prevalence of either the hydraulic or the structural criteria in the degradation process, was established through sensitivity analysis. The outcome indicates that hydraulic factors have to be integrated in the pipe-deterioration model as well as the structural factors. With a reference case set at $w_s=0.5$ (i.e., structural and hydraulic criteria having the same significance level), results remained conservative most of the time. Even if the AHP technique was already applied in sewers and water mains as shown in the literature, this paper is unique in that the AHP technique is used to combine a large number (fifteen) of structural and hydraulic criteria and subcriteria. The proposed model will enable researchers and municipal engineers to weigh the structural criterion against the hydraulic criterion to achieve adequate intervention that is carefully planned through a flexible tool outlaying a hierarchy of scores reflecting the infrastructure under consideration.

Decisions concerning dam removal or retention are challenging because they involve tradeoffs, diverse stakeholders, and increasing public safety concerns. A multiobjective portfolio optimization approach, implemented as an integer linear program (ILP), identifies efficient portfolios of dam removals in terms of the objectives of public safety, fish population health, and cost. The ILP integrates judgments by dam safety experts with the results of ecosystem simulations and statistical analysis of empirical data, to explore tradeoffs among the three objectives when choosing a portfolio of dams to be removed in multiple watersheds. This methodology is applied to a case study including 139 dams in 10 watersheds of the Lake Erie basin. Significant tradeoffs are found between maximizing fish population health and minimizing safety risks under a given budget, with different dams recommended for removal in each case. Also, the way dam safety risk is quantified in the ILP affects the selected set, and therefore, deserves further research. Overall, the multiobjective portfolio analysis approach provides a simple, flexible, and useful tool for policy makers to explore the nature and magnitude of tradeoffs to screen potential dam removal projects.

In China, the allocation of water rights in river basins is customarily determined by annual mean volumes of water resources. However, seasonal variability of water availability should be considered for an effective and flexible allocation policy. This paper summarizes a framework for seasonal water allocation. It proposes a modified rule for seasonal water allocation on the basis of the traditional rule used in China and introduces a risk analysis model to assess the performance improvement of the proposed rule when dealing with hydrologic variability in the context of a water rights system. The framework and models were applied to the Shiyang River Basin, Gansu Province, China, and they indicated that seasonally allocated water under the modified rule is more stable and has a lower risk for water use than a rule based on annual mean volumes. This leads to more reliable supplies over the seasons and could provide an effective method for water rights operation in the Shiyang River Basin. It also presents a useful reference for water resource management in other water-deficient regions of China.

This study deals with the historical hydro-climatological watershed characteristics of the Litani basin located in Lebanon, east of the Mediterranean Sea, during 1900–2008. The basin is divided into two distinct sub-basins. The Upper Litani Basin (ULB) is mostly dominated by mountainous geophysical features and the Lower Litani Basin (LLB) is closer to the Mediterranean coast. Monthly and annual temperature and precipitation data were generated from different sources, including global gridded data, satellite data, and local station data. Because long-term runoff records were not available, previously generated monthly synthetic runoff data were employed. Temperature and precipitation trends were investigated using the Mann-Kendall and Sen Slope nonparametric trend tests. The mean annual and monthly runoff were subsequently correlated with temperature and precipitation variations using a multiple linear regression approach. The results show that between 1900 and 2008, the whole Litani basin, including both sub-basins, experienced a drying trend without a significant change in temperature. However, within the period of 1970–2008, the whole basin grew notably warmer in all seasons, without being wetter; this climatic pattern was reflected in the LLB but not in the ULB, which was only slightly warmer, but also slightly wetter in this period. In addition, the results revealed distinctive seasonal and annual correlations between temperature and precipitation changes and the basin’s runoff. However, runoff for both the ULB and the LLB was shown to more frequently correlate with precipitation than with temperature. Positive correlations were found between runoff and precipitation during winter and wet seasons, and a negative correlation was detected during winter between temperature and runoff variations of the ULB. Hence, the continuation of increasing temperature and decreasing precipitation trends may pose a threat to future water resources in the Litani as a whole, and the ULB in particular.

Sediment discharge from erosion, urban runoff, and construction can cause environmental degradation. Governments try to regulate sediment, but the regulatory approach is costly to land owners. In principle, a market-based system could reduce costs, but the associated transaction costs are far too high, because market participants must find trading partners, negotiate, and seek government approval. This paper proposes a smart market design with an associated market clearing model for sediment discharge. Market participants offer and bid for tradable discharge allowances to a central auctioneer. The allowances correspond to kilograms of sediment discharge per year, as estimated by an erosion model. The auctioneer then uses a linear program to price and allocate sediment allowances in a given catchment. Participants do not trade bilaterally, but rather through a central pool, reducing transaction costs. The market design uses the relevant hydrological data, quantifies environmental effects explicitly, and gives price signals based on the environmental features. Additionally, the initial rights could be scaled back or up when catchment is over-allocated or under-allocated to keep a market operating with revenue neutrality. The smart market is simulated for two catchments in Auckland, New Zealand.

An integrated hydrological well field model (WELLNES) that predicts the water level and energy consumption in the production wells of a waterworks is used to optimize the management of a waterworks with the speed of the pumps as decision variables. The two-objective optimization problem of minimizing the risk of contamination from a nearby contaminated site and minimizing the energy consumption of the waterworks is solved by genetic algorithms. In comparison with historical values, significant improvements in both objectives can be obtained. If the existing on/off pumps are changed to new variable-speed pumps, it is possible to save 42% of the specific energy consumption and at the same time improve the risk objective function. The payback period of investing in new variable speed pumps is only 3.1 years, due to the large savings in electricity. The case study illustrates the efficiency and applicability of the developed modeling framework.

In much of the western United States, reducing residential water use is a major source of water conservation, especially as population growth urbanizes agricultural land. Although estimates of the potential of conservation are useful, the experience of Australia provides a realistic target for residential water conservation. Although reliability of urban water use data is often questionable, it is clear that Australians use less water than Californians, with a similar climate, economy, and culture. Per capita usage is compared, and explanations for use differences are offered. If California had the same residential water use rates as Australia, it could have reduced gross urban water use by 2,600 GL (2.1 million acre-feet) in 2009 and potentially saved 1,800 GL (1.5 million acre-feet) for consumptive use by others.

This study develops and applies a simple linear optimization program to identify cost-effective best management practices (BMPs) to reduce phosphorus loading in Echo Reservoir, Utah. The optimization program tests the feasibility of proposed total maximum daily load (TMDL) allocations based on potential BMP options and provides information regarding the spatial redistribution of loads among sub-watersheds. The current version of the TMDL for Echo Reservoir allocates phosphorus loads to existing non-point phosphorus sources in different sub-watersheds to meet a specified total load. Optimization results show that it is feasible to implement BMPs for non-point sources in each sub-watershed to meet reduction targets at a cost of $1.0 million. However, relaxing these targets can achieve the overall target at lower cost. The optimization program and results provide a simple tool to test the feasibility of proposed TMDL allocations based on potential BMP options and can also recommend spatial redistributions of loads among sub-watersheds to lower costs.