impacts are landscape-level phenomena that result from decisions at many individual permit sites, as well as activities that are not regulated under Section 404 (Gosselink and Lee, 1989).

Conversion of a wetland forest to agricultural use results in a typical cumulative impact (Gosselink et al., 1990b). Historically, the incremental clearing of 10 ha to as much as 2,000 ha in an individual permit has been perceived to have no significant ecological impact on a total forest system of several million hectares, and the cumulative effect of many such permitted activities has been ignored (Gosselink et al., 1990b). Currently no methodology for cumulative impact assessment is generally accepted by scientists and managers (Gosselink et al., 1990a). Furthermore, wetland restoration has generally been undertaken on an ad hoc basis, and the potential role of restoration in a cumulative context or on a landscape scale is not usually considered.

Gosselink and Lee (1989) described a methodology for cumulative impact assessment and management that incorporates a process of ecological characterization, goal setting, and planning. The method also focuses attention on the landscape level and bases planning on landscape ecology principles. Researchers used the Tensas River basin, an area of approximately 1 million hectares in northeastern Louisiana, as a case study to test this general approach to cumulative impact assessment and management (Gosselink et al., 1990a).

Most ecosystems will be affected by global warming over the next century, and planning for restoration of self-sustaining ecosystems must consider the potential impact of climate change. The problem is predicting how the climate will change at specific locations (will it get wetter, drier, warmer, or cooler?) and what measures might offset any negative impacts of changing weather and hydrology. For biodiversity reserves, Peters (1988) has suggested that a number of management actions may be needed to prevent species extinctions as climate changes. For example, much larger reserves may be necessary to perpetuate populations, so that heterogeneous topography and soils are included, thus protecting a wider range of habitat types (and presumably also ecotypes). We may need to plan to control environmental conditions in reserves (e.g., through irrigation or drainage). Control of predators and exotic species may become more critical. Peters (1988) also suggests that creating reserves outside the normal climatic regions of an ecosystem type may be needed. Finally, he mentions the possibility of moving reserves as the climate changes,