It is generally presumed that intensive survey yields reliable representations of regional archaeological variability. We evaluate this assumption with an analysis of the results of two intensive surveys of the same terrain in the Upper Basin, a heavily forested upland ecosystem located south of Grand Canyon National Park in Kaibab National Forest, northern Arizona. By comparing differences between the results of site-based surveys with those of mapping-unit-based surveys, we demonstrate that units of observation have a profound effect on how archaeological landscapes and their variability are characterized and interpreted. In addition, results of four analyses of survey data show that the archaeological resource inventories created by the application of these two different units of observation cannot be reconciled. We suggest that because some units of observation may be more appropriate for certain problems and for different kinds of surface and near-surface archaeological records, additional studies of the effects of units of observation on characterizing the archaeological content of the same terrain should become a research priority in survey archaeology. Without such studies, the identification of archaeologically sensitive areas, particularly those that necessitate active management and vigilant protection on public lands, will be underdetermined, thereby placing those heritage properties at risk.