Small-scale, spatially distributed water management practices: Implications for research in the hydrologic sciences
 Traditional water resources management in the United States relies heavily on the use of centralized facilities, such as regional wastewater treatment plants and flood control reservoirs. Increasing concern for human impacts on aquatic systems and diminished federal support for large water management projects are motivating the increased use of small-scale, widely distributed practices, such as treatment wetlands and infiltration practices. These practices, which exploit or enhance natural systems and processes, can be used alone or in conjunction with traditional practices to enhance performance and reduce environmental impacts. The use of smaller, more distributed water management practices challenges the hydrological science community to improve its capacity for assessing and predicting hydrologic conditions and to make this capacity accessible to water resource practitioners.
 Traditional water resources management in the United States relies heavily on the use of centralized facilities, such as regional wastewater treatment plants and flood control reservoirs. Centralized facilities achieve efficiency through economies of scale. For a variety of reasons, however, such facilities are falling into disfavor. They often cause undesirable environmental impacts, an increasingly important issue given population growth and anticipated climate change. In some cases, such as flood control, centralized facilities have failed to fully meet their objectives. Large water resource projects typically depend on federal subsidies, which have declined in recent years. Small, spatially distributed practices, such as the use of constructed wetlands for wastewater and storm water treatment and infiltration practices for reducing runoff and increasing groundwater recharge, are emerging to augment or even replace traditional facilities. These practices are designed to exploit or enhance natural systems and processes and typically cause less environmental impact. They can be used alone, or in conjunction with traditional practices to enhance performance and reduce environmental impacts. This increased use of smaller, more distributed water management practices challenges the hydrological science community to improve its capacity for assessing and predicting hydrologic conditions and to make this capacity accessible to water resource practitioners.
2. Flood Control
 Traditional flood control has relied heavily on the use of extensive systems of levees, dams, and diversions to control flood flows. This emphasis on flow control has been very effective, contributing significantly to economic growth. However, it has also caused unintended and often unaccounted for impacts, most notably the alteration of natural flow patterns and the associated degradation of aquatic ecosystems [Poff et al., 1997]. In recent years increased attention has been directed to spatially distributed solutions, such as avoidance of flood-prone areas, use of land management practices that minimize increases in storm runoff, and restoration of the natural flood storage capacity of wetlands and floodplains [Interagency Floodplain Management Review Committee, 1994].
3. Urban Water Management
 In urban areas, water supply, wastewater treatment, and storm water management are all centralized. Although efficient, this approach has some notable failings. The problems associated with combined sewers (sanitary and storm) have long been recognized, although not fully addressed. More recent is the recognition that centralized water supply and wastewater treatment result in large diversions of water from the original flow systems, contributing to degradation of local aquatic systems that are often signature landscape features [Alley et al., 2002]. Local planned water reuse for urban irrigation and other nonpotable purposes offers a spatially distributed strategy that conserves water and reduces environmental impacts, and is increasingly being considered as part of an overall strategy for urban sustainability.
 Detention ponds, the ubiquitous solution to the problems associated with increased storm runoff peaks, do nothing to control increases in storm runoff volumes and decreases in groundwater recharge [Potter, 2003]. Decreased groundwater recharge exacerbates the impacts of groundwater pumping. Increased runoff volumes increase the magnitude and frequency of downstream flood peaks, even when local runoff peaks are controlled. Small-scale, distributed infiltration practices, such as rain gardens and bioretention facilities, offer promise for such control and have the potential to mitigate groundwater depletion [Dussaillant et al., 2004]. Such practices are also proving to be effective at improving the quality of urban runoff [Pitt and Clark, 2003].
4. Research Challenges
 A shift to small, spatially distributed practices for managing water resources requires an improved capacity to assess present and future hydrologic conditions at a range of spatial scales. This challenges the hydrologic research community to develop better methods for collecting relevant data and better operational hydrologic models for evaluating alternative future scenarios. In addition, the community must improve its capacity to deliver these products to practitioners.
 Centralized facilities rely on engineered solutions that can be designed and constructed to specification. Local information may be required, but only at a relatively small number of locations. Spatially distributed practices rely on local conditions that are often subject to limited control; their design requires detailed assessment of these conditions. The use of the flood storage capacity of wetlands and floodplains requires accurate topographic information. LIDAR and other remote sensing techniques offer great potential for providing such information, but they must be demonstrated over a wide range of landscape conditions, such as heavily vegetated wetlands. The design of distributed infiltration practices to reduce runoff volumes and increase groundwater recharge requires spatially distributed assessment of the depth to groundwater and the hydraulic parameters of both shallow subsurface layers that may cause surface ponding and deeper layers that may limit percolation to critical aquifers. Protection of specific ecosystems that depend on groundwater discharge requires detailed understanding of the stratigraphic controls on groundwater flow.
 The design and assessment of small, spatially distributed facilities also require a suite of vetted operational models that can predict with known and acceptable levels of accuracy the performance of water resource management practices, both individually and collectively. Such models would be useful for a range of design and management questions. For the case of urban storm water management, the following questions come to mind: How will an individual infiltration practice affect the runoff/recharge regime of a given parcel of land? How will a collection of infiltration practices in conjunction with other storm water management practices affect the runoff/recharge regime of a large land development? How will alternative scenarios of land development and water resource management (water supply, storm water, and wastewater) affect highly valued ecosystems? How should water management regulations be designed to achieve specific goals regarding peak discharges, flow regimes, and groundwater sustainability? Models that are widely accepted and supported by accurate data are needed to explore these questions.
 In addition to developing new methods and models for supporting distributed water management, the hydrologic science community must greatly improve the effectiveness of its communication with practitioners. Over the past several decades, researchers have developed numerous "decision support systems" for specific applications. Yet most water resource decisions do not exploit modeling innovations associated with these systems. For example, event-based hydrologic analysis is still dominant, even though it has been well documented that continuous methods yield better designs for many of today's problems. What are the reasons for apparent gaps between theory and practice? Are the gaps simply due to failures in technology transfer? Or are they due to inadequacies in research that result from poor communication between scientists and practitioners? The hydrologic science community should participate in organizing and implementing major efforts to better understand and close the gap between research and practice.
 Small, spatially distributed water management practices are being increasingly used to augment or even replace large water management projects. It is incumbent on the hydrologic science community to conduct the research and develop and make available the measurement and modeling tools necessary to support this fundamental change in water management.