Meeting an objective of returning a riparian site to "original conditions," or a close approximation thereof, may be difficult because those conditions may not be known due to the site's long history of human impacts (Cairns et al. 1979). Collection of historical data on the site can greatly aid in development of a restoration site plan and success criteria. Several studies have used historical regional lists to determine desired plant or animal diversity of the completed site (e.g., Anderson and Ohmart 1985; Hey and Philippi 1985a). Gross and Brown (1987) stressed that the long-term success of mine reclamation projects should be measured by how well the landscape functions, not by how well the design mimics the landscape that existed prior to mining.

Unfortunately, little information is available regarding success of riparian ecosystem creation/restoration projects (Platts et al. 1987b; Ischinger and Schneller-McDonald 1988). Data on survival and growth of planted vegetation have been the most commonly used parameters to support the success of these projects. Typically, these variables are measured for only the first few growing seasons. However, it may take several years beyond that time for the revegetated site to achieve desired species diversity. This is particularly true for the relatively slower growth of riparian hardwoods. A reestablished bottomland forest can take 40-60 years to become self-regenerating, and to produce the full value to many wildlife species (Haynes and Moore 1988).

The Washington State Department of Transportation sets standards for success of wetland mitigation projects as measured by the survival and growth of plants (Miller 1988). Projects are monitored annually using standard vegetation sampling techniques to measure coverage and species composition. Sites also are evaluated for wetland values. At the end of the first year of monitoring, the project should have 50% survival of species indicated on the revegetation plan. At the end of the fifth year of monitoring, a wetland project is considered successful if areal coverage by wetland species is 90% of adjacent natural wetland areas.

Ideally, success of a creation/restoration project should be based on a number of variables. One project that is attempting to measure a variety of parameters is the Des Plaines River Wetlands Demonstration Project in northeastern Illinois (Hey and Philippi 1985a; Hey 1988). To document changes in physical structure and vegetation with change in faunal composition over a 5 year period following habitat creation and restoration, baseline and subsequent data collecting will include water quality and chemistry, hydrology, soils, microorganisms, invertebrates, fish, amphibians, reptiles, birds, mammals, and human use, along with vegetation and productivity studies. The Fish and Wildlife Service Habitat Evaluation Procedures and Habitat Suitability Indices will be used for three species of mammals, nine nongame birds, and six fishes.

Quantitative comparisons with control sites lends support for documenting success of a creation/restoration effort. On the Lower Colorado River Project, a 4-year preliminary study of control sites was used to determine which species and components of vegetation attracted birds and nocturnal rodents during various seasons (Anderson et al. 1979). The information was used to design the revegetation project and to quantify expected avian and mammalian use of the proposed site. Success could then be quantified by comparing resulting use of the site with expected use by chi-square analysis.

Platts et al. (1987a) stressed inclusion of control areas that do not receive management treatments in evaluating success of riparian habitat improvement efforts. Treatment and control sites must have the same premanagement characteristics and the same potential for response to management to document changes actually attributable to management. Resources are frequently not monitored long enough to permit management responses to occur. Unfortunately, there has been a tendency for neglecting advance consideration of statistical tests, often resulting in data that cannot be used to quantitatively support determination of project success.

Success of stream relocation projects involving riparian habitat improvements in mined areas has been determined by comparing aquatic invertebrate or fish populations with natural sites (e.g., Thompson 1984; Robertson et al. 1987). The Florida Department of Environmental Regulation often includes an assessment of aquatic macroinvertebrate colonization as one of its measures of success. Diversity must be similar to that in a reference wetland, either the identical wetland monitored before mining occurred, or in a nearby wetland with similar characteristics (Robertson et al. 1987).

Evaluation of the success of the Agrico Swamp Restoration Project in Florida involved comparisons between natural and reclaimed sites over 4 years, once sites had stabilized (Erwin 1986). Species richness, percent cover, survival, and growth of vegetation were measured. Diversity and abundance of macroinvertebrates of reclaimed sites were compared to natural communities. Improvement of water quality was termed successful as confirmed by State water quality standards. Monitoring of hydrology, tree seedling survival and growth, and wildlife and fish use of the reclaimed habitat will continue for 3 additional years.

Comparing success of creation/restoration projects conducted by different agencies or organizations can be difficult due to the wide variation in monitoring and evaluation techniques used. Nelson et al. (1978) addressed this problem in their documentation of successful and potentially successful habitat and population improvement measures accompanying water resource development projects (e.g., construction of dams and reservoirs) recommended by the Fish and Wildlife Service for reservoirs and streams of the western U.S. They classified the improvement as successful if it apparently accomplished a major part of its intended purpose, marginally successful if it accomplished a moderate part, and unsuccessful if it accomplished only a minor part, in terms of habitat and population categories. Confidence in the estimate of success as a function of the reliability of reports on actual biological effects was rated high, medium, or low. This rating was determined by whether the reported effects of a habitat or population improvement measure were derived quantitatively from field measurements or qualitatively from direct field observations or indirect reports of anglers or hunters.

Continued success of creation/restoration projects may depend on management, particularly in grazing-impacted areas. For example, Clay (1984) documented success of a high mountain meadow restoration project (as indicated by stabilized banks, raised water table and channel bottom, and filled eroded channels) in Modoc County, California, after 3 years of fencing. He stated that fencing could be removed, but only if grazing was managed. The U.S. General Accounting Office (1988) reviewed 22 riparian areas restored by the Bureau of Land Management or the Forest Service in the western U.S. and noted that the overriding factor in achieving success was improving the management of livestock to give the native vegetation more opportunity to grow. In some cases, fences were built to keep the livestock out of the area, either permanently or until the vegetation had recovered and streambanks were stabilized. In others, livestock continued to graze in the area, but their use was restricted by herding or fences, or a combination of both, to a shorter period of time, specific season, or only part of the area.