This article is an extension and application of Preston and Bedford (1988), especially as relevant to bottomland hardwood (BLH) forests of the southeastern United States. The most important cumulative effects in BLH forests result from incremental forest loss (nibbling) and from synergisms resulting from this nibbling. Present regulatory procedures are ineffective in preventing incremental forest loss because of the focus on permit site evaluation, rather than on large landscapes. Three examples are given to illustrate the need for a landscape focus. This perspective requires preplanning or goal-setting to establish the desired conditions to be maintained in the regulated landscape unit. Spatial and temporal scales are of particular concern for landscape impact assessment. Natural drainage basins of about 106 ha, as identified in U.S. Geological Survey hydrologic units, appear to appropriate spatial units: they have fairly natural boundaries, are of sufficient size to support populations of large, wide-ranging mammals, and are compatible with existing maps and databases. Time scales should be sufficiently long to include recovery of wetland ecosystems from human perturbations. In practice, available data sets limit analysis to no longer than 50 yr. Eight indicators of landscape integrity are identified, based on generally available long-term data sets. Linking technical information concerning cumulativeeffects on landscapes to the evaluation of cumulativeimpacts in regulatory programs (i.e., goal-setting) is a serious issue that can benefit from precedents found in the field of epidemiology, and in the establishment of clean air and clean water standards. We suggest that reference data sets must be developed, relating BLH function to structure (forest area). These can be used to set goals for individual watersheds, based on their present conditions and the magnitude and type of perceived development pressures. Thus the crucial steps in establishing a successful program appear to be (1) establish study unit boundaries, (2) assess the condition of study unit landscape integrity, (3) set goals, and (4) consider the impacts of permit proposals with both goals and the existing condition of the study unit landscape in mind.