Journal of Water Resources Planning and Management

A water distribution system burst from a sudden pipe failure results in water loss and disruption of customer service. Artificial neural networks, state estimation, and statistical process control (SPC) have been applied to detect bursts. However, system operational condition changes such as the set of operating pumps and valve closures greatly complicates the detection problem. Thus, to date applications have been limited to networks that are supplied by gravity or under consistent operation conditions. This study seeks to overcome these limitations using a nonlinear Kalman filter (NKF) method to identify system condition, estimate system state, and detect bursts.

Technological advances in forecasting the Earth’s climate offer a potentially useful tool to support planning and management decisions in water resources. Previous research has found that the implementation of new ideas and practices are impeded by many challenges such as low forecast skill, institutional obstacles, and political disincentives to innovation. To better understand barriers to forecast use at seasonal-to-interannual, decadal, or longer time scales, this paper evaluates a diffusion of innovations (DoI) framework to assess the adoption of hydroclimatic forecasts by water managers in the Northeast United States. Specifically, this paper seeks to understand how five innovation characteristics, relative advantage, compatibility, complexity, trialability, and observability, influence the rate of adoption. Methods used for this analysis include the distribution of a survey, interviews, and a literature review. Results indicated that while much attention has focused on institutional obstacles, in the Connecticut River Basin obstacles were related to characteristics of the forecasts themselves. Evaluation of the DoI makes clear that the challenges to forecast use are common to all innovations.

Water consumption is perhaps the main process governing water distribution systems. Because of its uncertain nature, water consumption should be modeled as a stochastic process or characterized using statistical tools. This paper presents a description of water consumption using statistics as the mean, variance, and correlation. The analytical equations expressing the dependency of these statistics on the number of served users, observation time, and sampling rate, namely, the scaling laws, are theoretically derived and discussed. Real residential water consumption data are used to assess the validity of these theoretical scaling laws. The results show a good agreement between the scaling laws and scaling behavior of real data statistics. The scaling laws represent an innovative and powerful tool allowing inference of the statistical features of overall water consumption at each node of a network from the process that describes the demand of a user unit without loss of information about its variability and correlation structure. This will further allow the accurate simulation of overall nodal consumptions, reducing the computational time when modeling networks.

Combined sewer overflows (CSOs) have major impacts on receiving waters. It is important to investigate the potential impacts of climate change on CSOs. The main objective of this study was to develop projections on the future evolution of CSO frequency and durations in the context of climate change. First, a model was developed based on available CSO data (date of occurrence and durations over the May to October period for the years 2007 to 2009) from 30 CSO outfalls located in southern Quebec. This model estimates the probability of occurrence and the duration of a CSO given the rainfall total depth. Second, two simulated series from a regional climate model were used to estimate the evolution of CSO frequency and durations between 1961–1990 and 2071–2100 periods. Results suggest that the May to October mean CSO number will remain globally unchanged for the 30 CSO outfalls while mean annual CSO duration will slightly increase ($+4.2\%$). On a monthly basis, CSO frequency and duration will decrease in average during July and August, while they will increase in May and October.

In this study, a two-stage chance-constrained fractional programming (TCFP) method is developed for dealing with water quality management problems associated with stochastic inputs. Two-stage chance-constrained fractional programming is a hybrid of stochastic linear fractional programming (SLFP) and two-stage stochastic programming (TSP) methods. It can not only balance objectives of two aspects through converting a bi-objective problem into a ratio one but can also analyze various policy scenarios when the promised production targets are violated. For demonstrating its advantages, the proposed TCFP method is applied to a case study of water quality management where managers have to consider conflicting objectives between economic development and environmental conservation, as well as stochastic features expressed as probability distributions. The obtained solutions under different significance levels can help managers to identify desired policies under various environmental, economic, and constraint-violation conditions.

Optimal operating policies have been derived for reservoirs in series and in parallel for various purposes, but little formal analysis has been done for flood operation of parallel reservoirs. For flood management, parallel reservoir releases should be managed together to reduce downstream peak flow and minimize flood damage. The optimal allocation of total available flood storage capacity among parallel reservoirs should have each reservoir’s flood storage allocation providing the same incremental reduction to downstream flood flows. This approach is developed mathematically and applied for deterministic and probabilistic inflows. The applicability and effectiveness of these derived flood storage allocation rules are demonstrated by the case study of Oroville Reservoir and New Bullards Bar Reservoir in California’s Sacramento River Basin with a single historical 1997 flood and an uncertain storm.

A fuzzy decision-making framework (DMF) is combined with a hybrid genetic algorithm–linear programming (GA-LP) optimization approach to determine the best booster station network for a water distribution system. The proposed hybrid GA-LP model simultaneously optimizes two conflicting objectives; namely, minimization of total chlorine injection dosage and the number of booster stations. At the same time, residual chlorine concentrations are kept within desired limits. Adjustment of the relative importance of two conflicting objectives results in different optimal solutions. Selection of the best alternative among these optimal solutions is performed through a fuzzy multiobjective DMF. The proposed DMF allows incorporation of the decision makers’ preferences into the booster station network design. In this study, three fuzzy objectives are selected based on economic, operational, and health-related concerns. The hybrid GA-LP model is applied to a case study, and results show that the proposed methodology is effective in maintaining chlorine residuals within desired limits networkwide while minimizing the total chlorine injection, and the fuzzy DMF is a useful tool for incorporating the case specific limitations into the decision process.