Journal of Water Resources Planning and Management

The assessment of flood risk is a difficult task and usually requires solution of a flood routing problem as a part of the assessment. Due to the large number of scenarios that have to be developed and analyzed, simplified numerical models are used for the computation of flooded areas in each scenario. More sophisticated models are often too complex to manage or, due to their design generality, not well suited to deal with the specific needs of flood routing problems. A comparison among three different models, with varying degrees of simplification and abstraction, is presented and discussed. Some considerations on the effectiveness of simpler models are drawn, focusing on the prediction of flooded areas.

A model package including a watershed model, an atmospheric loading model, a hydrodynamic model, and a eutrophication model are used to evaluate the benefit of nutrient and solids load controls on the Virginia tributaries to the Chesapeake Bay. Quantities examined include nutrients, solids, chlorophyll, anoxic volume, mesozooplankton, benthos, light attenuation, and submerged aquatic vegetation. Nutrient load controls are beneficial in reducing chlorophyll concentration and anoxic volume but produce no major benefits for zooplankton and benthos. Load controls benefit aquatic vegetation biomass, but more extensive solids controls are required to restore widespread SAV distribution.

A prototype spatial decision support system (SDSS) is presented for integrated, real-time river basin flood control in a multipurpose, multireservoir system. The SDSS integrates a geographic information system with a database management subsystem, a real-time meteorological and hydrological data monitoring system, a model-base subsystem for system simulation and optimization, and a graphical dialog interface allowing effective use by system operators. The model-base subsystem employs an artificial neural network in a real-time flood forecasting module providing spatially distributed forecasted flows that are updated as the flood event progresses. Forecasted, basinwide discharges are input into a dynamic programming module providing optimal gate-control strategies, which are also updated in real-time. The SDSS for flood control is applied to the Han River Basin in Korea and demonstrated through simulated application to a severe 1995 flood event. Results of the case study indicate that integrated operational strategies generated by the SDSS for flood control substantially reduce downstream flood impacts, while maintaining sufficient conservation storage for water use subsequent to the flood season.

Economical integration of permanent and emergency flood control options is a long-standing problem in water resources planning and management. A two-stage linear programming formulation of this problem is proposed and demonstrated, which provides an explicit economic basis for developing integrated floodplain management plans. The approach minimizes the expected value of flood damages and costs, given a flow or stage frequency distribution. A variety of permanent and emergency floodplain management options can be examined in the method, and interactive effects of options on flood damage reduction can be represented. The approach is demonstrated and discussed for a hypothetical example. Limitations of the method in terms of forecast uncertainty and concave additive damage function forms are discussed also, along with extensions for addressing these more difficult situations.

A storm-water quality control system consists of trickle channels and water quality control basins (WQCB). To be economical, a WQCB is often designed to capture rainfall events smaller than extreme events. In current practice, the design rainfall statistics and runoff capture rates for sizing a WQCB have to be derived from the local long-term continuous rainfall record. Such a rainfall-runoff simulation process is lengthy and time-consuming. In this study, the analyses of 30- to 40-year continuous rainfall data recorded from seven metropolitan areas in the United States indicate that the frequency distribution of rainfall event-depths can be described by an exponential decay function. Applying the exponential distribution to a complete rainfall data series, the normalized runoff capture curve is derived in this study to describe the nonexceedance probability distribution of runoff depths. Similarly, the normalized runoff delivery curve is also developed to describe the nonexceedance probability distribution of runoff rates. These two curves provide necessary and important design information by which both the trickle channel and the WQCB can be sized on a consistent basis of overflow risk.

The possible impacts of climate change induced by global warming on a highly permeable, unconfined aquifer located in the humid northeastern U.S. is analyzed for the years 2030 and 2100. The groundwater model MODFLOW is calibrated and verified for the area and used to analyze the impacts of several mean and drought climate change scenarios. Pumping conditions and land use are assumed to remain the same as at present. The climate change scenarios result in either slightly higher, no different, or significantly less annual recharge and groundwater elevations, producing a variety of impacts on wetlands, water supply potential, and low flows. Impacts are most severe under some drought scenarios. The policy response to the possible impacts as recommended by the leader of a watershed advocacy group is that wise management of the aquifer should be advocated with a particular focus on limiting the expansion of water supply from the aquifer and increasing the present amount of groundwater recharge.

This paper presents a multiscale derivative method for solving a successive approximation linear quadratic regulator model for optimal in situ bioremediation design. An efficient one-sided forward divided difference numerical derivatives calculation was implemented as the first stage of the method, which only required assembling the right-hand-side vector of the linear systems of equations of the simulation model and performing backward substitution. The derivative calculation was reduced from $O\left(N^3\right)$ to nearly $O\left(N^2\right)\text{,}$ where N is the number of non-Dirichlet state variables. A V-cycle multiscale derivatives approximation was implemented as the second stage, which used coarser mesh derivatives to interpolate finer mesh derivatives. Implementing the numerical derivatives method in a case study with over 1,600 state variables caused a reduction of more than two-thirds in computing time over the previous analytical derivatives method without loss of accuracy. Using the V-cycle multiscale derivatives approximation further reduced computing time by 29%, resulting in an overall 77% reduction compared to the previous analytical derivatives method. The reduction will be even greater for applications with more state variables, enabling the solution of much larger-scale problems than was previously possible.