Journal of Water Resources Planning and Management

The development and validation of a methodological approach to predicting runoff over large regions is considered. The Soil Conservation Service (SCS) runoff curve number procedure is modified to allow for areal variation of land use and land cover, soils, and rainfall over a regular‐gridded network. Ten rainfall events, from the period 1973–1980, are selected to estimate runoff for the Mahantango Creek Watershed $\left(421{\mathrm{km}}^2\right)$ in east‐central Pennsylvania. The method may be useful in generating a geographic information system database for distributed‐components hydrologic models and for water resources management.

Results of interview surveys for potential flood damages are used to develop depth—damage curves for structural and content groupings and alternative residence types. Statistical significance tests are used to collapse 28 structural categories into seven basic groups. The exterior of the residence is demonstrated to not necessarily be a good indicator of the potential flood damage to the contents. Use of average curves are recommended except in the very obvious circumstances of very low and/or very high valued residences.

Estimates are provided for quantifying adjustments to depth‐damage tables as a result of flooding events of different types on residential dwellings. The adjustments reflect estimated changes to flood damages due to the availability of flood warning, long durations of flooding, and ice and high velocity flow. Regression analyses are described that examine the degree to which a range of variables, including a resident's experience with flooding, householder income, floor area of residence, and assessed value of property, can explain the variability of flood damage curves.

Seattle, Washington, suffered its most extreme drought on record during the summer and fall of 1987. Severe and continuing water use restrictions were required to limit the drought's impact on municipal water supplies, fish populations, and navigation. This paper describes a decision support system used to aid in drought decisions. Its components include an expert system, a linear programming model, database management tools, and computer graphics. The expert system incorporates operator experience and intuition using a rule base developed through interviews with management personnel from the Seattle Water Department. The expert system also integrates the other programming techniques into a single system. A linear programming model determines system yield and optimal operating policies for past hydrologic regimes. Database management and graphics software store and allow the display of over two thousand operating policies to decision‐makers. The system provides user‐friendly support to help decision‐makers explore a wide range of management alternatives.

Underground disposal of liquid wastes poses a danger of groundwater contamination. The prediction of their impacts is subject to many uncertainties. This paper develops and applies a risk analysis model that addresses several of these uncertainties. The model inputs are hydrogeologic parameters, the reliability of waste disposal, and demographic information. The outputs of the model are probability distributions of the number of rural home wells or municipal groundwater well fields that may be contaminated. Contamination is defined as occurring when the concentration for any disposed chemical exceeds its water quality standard. The model is used to analyze the risks posed by annular disposal of oil and gas brines in Ohio. It is found that the economic costs associated with those risks are an order of magnitude smaller than the benefits of the practice.

One means to increase effective water supply yield is to have pumpedwater intertie of reservoirs. The management of Hoover Reservoir and its associated pumped‐water intertie with Alum Creek Reservoir in the water supply system of Columbus, Ohio, is examined using the method of implicit stochastic optimization. Optimal monthly pumping policies are derived in a multiobjective environment involving trade‐offs of pumping and water supply shortage costs, target draft rates, and reliability levels. Policies heuristically derived from a dynamic programming formulation are tested for statistical performance using simulation, and are parametrically varied to determine the most cost‐effective policies for a given reliability level and draft rate. Multiobjective trade‐off analysis can then be carried out between pumpage cost, target draft rate, and reliability. In the case study, significant net savings result from the use of pumped water rather than alternative water supply sources due to lower softening costs, and a savings maximization procedure is developed. The advantages and disadvantages of implicit stochastic optimization are highlighted.