Journal of Water Resources Planning and Management

Conflicts in water resources exist because river-reservoir systems are managed to optimize traditional benefits (e.g., hydropower and flood control), which are historically quantified in economic terms, whereas natural and environmental resources, including in-stream and riparian resources, are more difficult or impossible to quantify in economic terms. Multicriteria decision analysis provides a quantitative approach to evaluate resources subject to river basin management alternatives. This objective quantification method includes inputs from special interest groups, the general public, and concerned individuals, as well as professionals for each resource considered in a trade-off analysis. Multicriteria decision analysis is applied to resources and flow alternatives presented in the environmental impact statement for Glen Canyon Dam on the Colorado River. A numeric rating and priority-weighting scheme is used to evaluate 29 specific natural resource attributes, grouped into seven main resource objectives, for nine flow alternatives enumerated in the environmental impact statement.

Multiobjective systems analysis is used to size reservoir storage and identify noninferior system operating rules that mitigate the impacts of consumptive appropriations. The marginal impacts of consumptive use are offset by adding reservoir storage to the system, balancing technical efficiency, economic efficiency, and equity. The “price” to receive a consumptive appropriation permit (augmentation storage) is effectively equal to the marginal cost of the new withdrawal. In this context, prices and costs are measured not in dollars, but in units of storage, days at minimum instream flow, and the other direct operating impacts of consumptive use. In return for “paying” the efficient marginal price to join the system, new appropriators become equal participants in regionally coordinated operation, with equal reliability of meeting unrestricted demands. Parametric operating rules to size augmentation storage are developed as a multiobjective extension of firm yield analysis, applied to forecast-based operation of a multireservoir system. Examples drawn from Maryland's Potomac River consumptive use regulation illustrate how operational definitions of equity and reliability offer a normative framework to manage risk-based appropriation within a permitted riparian regulatory system.

An economic-engineering modeling approach is presented for integrating urban water supply reliability analysis with shortage management options such as dry year option and spot market water transfers, water reuse, and long- and short-term water conservation. The integrated model uses a probability plotting position formula to link supply-side yield simulation to probabilistic shortage management optimization. The approach can help joint planning of supply and demand management and estimate the economic value of expanding infrastructure and changing institutional constraints. Results from a case study show that jointly managing supply capacity and operating rule decisions with shortage management decisions within a risk analysis framework can create significant economic savings. The importance of hydrologic and institutional uncertainties is explored. Tuning supply system and reservoir operating policies to match the seasonal and cost characteristics of available shortage management measures is shown to be essential to least-cost planning.

This paper describes the development and implementation of a modeling approach using both nonlinear optimization and finite-element simulation to improve the control of river reach water levels for navigation and hydroelectric energy production in run-of-river power plants. These two modeling methods have been incorporated into a decision support system (DSS) for use by the energy company RWE Energie AG for the real-time control of a section of the River Mosel in Germany. The use of this system has eliminated the need to spend several million EURO in dredging to guarantee the navigation depths desired on the river. This paper briefly describes the models and the experiences we had calibrating and implementing them. It points out the importance of involving people in the implementation process who have the power and authority to provide necessary data and equipment. The paper concludes with a discussion of the application of the algorithm to a system of three reservoirs when energy prices vary over the day.

Fifty years ago, Jeddah, on the Red Sea Coast of Saudi Arabia, was a very small city. In recent decades it has grown rapidly but the sewerage system has not. As a result, there is extensive on-site wastewater disposal, which raises the ground-water table, and as this reaches the surface, it damages property and health. The cost falls on private and public sectors. The costs, damages, and losses from the use of cesspools are estimated and compared with the costs of proper sewerage over the next 50 years. The present value cost of the cesspool system exceeds that of sanitary sewerage by an estimated factor of 4.5 in the developed area and 1.1 in the undeveloped. Sanitary sewerage for the whole city should have the highest priority; delay will increase the final cost.

Predictive equations for the harmonic mean flow; the 7-day, 2-year low-flow (7Q2); and the 7-day, 10-year low-flow (7Q10) of ungauged Texas streams are developed using multiple regression analysis based on meteorological and physiographic data from 63 gauged streams. Harmonic mean, 7Q2, and 7Q10 values from the data range from 0.0001 to 1.6448 m³/s (0.004 to 58.12 cfs), 0 to 0.566 m³/s, (0 to 20 cfs), and 0 to 0.425 m³/s (0 to 15 cfs), respectively. Gamma-produced 7Q2 and 7Q10 estimates yielded more statistically reliable equations than either the log-Pearson Type III or mixed-log-Pearson Type III produced estimates. The basin parameters considered include contributing drainage area, channel length and slope, basin shape factor, mean annual precipitation, predominant hydrologic soil group, and the 2-year 24-h precipitation. The regression equation for the harmonic mean included channel slope, mean annual precipitation, and drainage area and had a higher adjusted R² than the 7Q2 and 7Q10 equations. A regional regression analysis yielded higher adjusted R² for the harmonic mean than the statewide equation, implying that regional equations are more appropriate than statewide equations for Texas.

Overland pollutant transport via surface runoff and flooding is a primary concern in the management of agricultural land resources in the Delaware River Basin. The Catskills is home to multiple water reservoirs that supply the drinking water for New York City. Contamination of this water by pollutants emanating from agricultural sources located in floodplain areas necessitates risk quantification for these locations. This study was performed to assess the risk of rivers topping their banks, or flooding, during various time periods of the year. Streamflow data for seven stations were analyzed to produce monthly partial duration and maximum series. A log-Pearson Type II analysis was performed on the maximum and partial duration series to produce monthly probabilities of flooding. Based on the maximum series, on average, March had the highest probability of flood occurrence at 28%, and April and December had the next highest, at 18 and 12%, respectively. Larger flood probabilities were calculated using the partial duration series, and this result was attributed to the tendency of some months to have multiple floods. Calculated yearly flood probabilities from monthly results agreed well with theoretical values. This study has important ramifications for decision making based on hydrologic risk of pollutant transport in floodplain soils (e.g., this information is needed to determine surface water pollution risks from manure-spread agricultural fields and develop schedules to reduce this risk).

Reservoir system operation requires a variety of hydrologic information, of which streamflow data are most essential. Typically, available gauged streamflows are processed to represent flows at the location of a reservoir. Rainfall-runoff models provide alternatives to estimate streamflows when gauged data are unavailable or inadequate. Reservoir inflows may also be estimated, based on water budget computation if storage levels and gauged releases are available. This study presents a series of computational procedures to estimate flows into a multiple reservoir system where storage levels and gauged releases are available at irregular intervals. The inflows are estimated by water budget computation. Then the weekly and monthly inflows are derived from cumulative mass curves of inflows at irregular intervals. The inflow to a reservoir includes releases from upstream reservoirs, which are controllable, and also the flow from the immediate watershed, which is not controllable. In order to preserve, for further analysis, the stochastic characteristics of the unregulated inflow, the flow contributed by an immediate watershed is estimated separately from the total inflow to a reservoir.