Journal of Water Resources Planning and Management

This research develops a mixed integer linear programming (MILP) model that considers simultaneously both the traditional reservoir rule curves and the hedging rules to manage and operate a multipurpose, multireservoir system. During normal periods of operation, when inflows are plentiful, this optimization model efficiently distributes the available stored water from different reservoirs to meet the planned demands imposed by competing users. However, during periods of drought, or when anticipating a drought, the planned demands cannot be fully met, and a water shortage occurs. By considering the hedging rules along with the rule curves, guidelines are provided for reservoir releases. To minimize the impact of drought, the hedging rules effectively reduce the ongoing water supply to balance with the target storage requirement. The MILP model is applied to a multireservoir system in the southern region of Taiwan, where the results obtained demonstrate the applicability and utility of the model.

Nonpoint source management using wet detention ponds to reduce pollutant loading into receiving water bodies is a common practice. Design of these ponds is typically carried out individually to meet a target total suspended solids (TSS) removal level. An improvement to this approach is to generate cost-effective pond configurations that meet system-wide targets for removal of pollutant loadings, including TSS, total nitrogen, and total phosphorus, corresponding to a specific build-out land use plan. The amount of pollutant loading can be managed better through appropriate land use allocation planning conducted simultaneously with the design of detention pond configurations. This paper presents a modeling approach to address this problem and its solution via a genetic algorithm-based search procedure. Application of this approach is demonstrated using an illustrative case study involving the City Lake watershed in North Carolina. The results show cost improvements when considering land management simultaneously with the decisions for pond locations and sizes. Through iterative application of this method, the tradeoff between cost and removal level of each pollutant is also generated.

This study examined hydrological and fiscal effects of residential growth patterns in Virginia. Eleven scenarios that consider a fixed increase in population and vary housing arrangement within a tract (tract form), tract arrangement within the watershed (tract pattern), and shared open space land cover were examined. The analytical tools included a geographic information system, a statistical land value model, tract development budgets, and a hydrological model (HSPF). Low density development has the greatest hydrological impact due to highest per capita impervious area. Varying tract form has more impact on land values and tax receipts than varying tract pattern. Low density development has the highest increase in revenues net of public sewer, water, and education (bus transportation) costs. Higher density settlements reduce hydrological impacts but bear a high cost to local governments in reduced property tax revenues.

Municipal water infrastructure systems are in poor condition and are deteriorating rapidly. A general diagnosis of the actual structural state of these systems is needed, as are tools to assess their rate of deterioration. To make a diagnosis, one must collect and analyze data on the characteristics of water pipes and on their breakage histories. Unfortunately, many municipalities have only been rigorously recording breakage histories for a decade, while their pipes have been in the ground for much longer. A modeling strategy, inspired by survival analysis and using the annual number of water pipe breaks as an indicator of the structural state of a network, was applied to three municipalities characterized by their brief recorded pipe break histories. This paper presents a descriptive analysis of the water pipe and breakage data, and the application of the modeling strategy to these municipalities. Different replacement scenarios were also simulated to assess their impact on the evolution of the annual number of breaks in the three municipalities.

The maintenance of chlorine residual is needed at all the points in the distribution system supplied with chlorine as a disinfectant. The propagation and level of chlorine in a distribution system is affected by both bulk and pipe wall reactions. It is well known that the field determination of wall reaction parameter is difficult. The source strength of chlorine to maintain a specified chlorine residual at a target node is also an important parameter. The inverse model presented in the paper determines these water quality parameters, which are associated with different reaction kinetics, either in single or in groups of pipes. The weighted-least-squares method based on the Gauss–Newton minimization technique is used for the estimation of these parameters. The validation and application of the inverse model is illustrated with an example pipe distribution system under steady state. A generalized procedure to handle noisy and bad (abnormal) data is suggested, which can be used to estimate these parameters more accurately. The developed inverse model is useful for water supply agencies to calibrate their water distribution system and to improve their operational strategies to maintain water quality.

Chlorine concentration is an important parameter used to assess the quality of water supplied by a distribution network. While several water quality simulators are readily available (e.g., EPANET software), water quality optimization decision support tools are currently lacking. A new control-oriented calculation model for chlorine concentrations in distribution networks is presented for general water distribution networks operating under periodic demand conditions. The model is based upon the analysis of the characteristic curves of the concentration transport-reaction equation over each pipe in a network. Numerical results are comparable with EPANET results, but unlike other water quality algorithms, the current method requires neither discretizations along the pipes, nor the choice of initial conditions. Moreover, it is a first step toward a method for controlling concentrations, yielding linear relationships between the concentrations at both ends of each pipe by the explicit inclusion of system dynamics. These relationships can be used as constraints in a linear programming optimization model that minimizes deviations from permissible concentration levels.

Large-scale climatological states [tropical Pacific sea surface temperatures (SST), Pacific-North American (PNA) atmospheric teleconnection and Pacific Decadal Oscillation (PDO)] and local precipitation data are used to predict the April–August Columbia River streamflow at Donald, British Columbia, Canada. Using predictors up to the end of November in the preceding year, forecasts of the April–August streamflow were made by multiple linear regression (MLR) under a jackknife scheme. A correlation skill of 0.52 is attained using PDO, PNA and SST as predictors, with PDO being the strongest and SST the weakest. When local precipitation is added among the predictors, PDO becomes redundant, and MLR with precipitation, PNA and SST as predictors attained a correlation skill of 0.70. Feedforward neural-network models were used for nonlinear regression, but the results were essentially identical to the MLR predictions, implying that the detectable relationships in the short, 49-sample record are linear. A bootstrap process estimates the relative errors of the MLR predictions.