- This study compared effects of human and natural wildfire disturbance on age class distribution and associated ecosystem properties of forests in a 15,670 km2 area of the western Cascades of Oregon. The study site is characterized by three forest use types: low elevation, intensively harvested private industrial lands; mid elevation partially harvested public forest lands; and high elevation wilderness lands. Hypothetical landscapes were constructed, representing wildfire conditions, the current (1995) condition, and forest management strategies. The spatial pattern of forests in each landscape was depicted by the distribution of four forest age classes (0-30,31-80,81-200, and 201+ years). Wildfire-affected forests were simulated for 3000-year periods using an existing model with six alternate parameterizations obtained from fire history studies. Forest patterns on the 1995 landscape were obtained from classified remotely sensed imagery. Three landscapes representing hypothetical forest management strategies were constructed: 1) riparian buffers, 2) reserve allocations, and 3) multi-age rotation harvests. Forest age class amounts, spatial distribution, and arrangement were compared between wildfire-affected, 1995, and managed landscapes. Wood production, carbon storage, water yield, and species diversity of each landscape were evaluated using simple measures from empirical studies relating forest structural patterns to ecosystem properties. Comparison of the 1995 landscape with simulated wildfire-affected landscapes indicates that private industrial lands in 1995 had more young forest, whereas wilderness lands had more old forest, than most, or all, simulated wildfire conditions. Because private industrial and wilderness areas deviate from the range of natural (wildfire-affected) variability in opposite directions, the study area as a whole is within the probable range of natural variability over the past few millennia. Forest ages classes on public non-wilderness lands were within the range of simulated wildfire-affected landscapes. Ecosystem properties of the current and managed landscapes were within the range of variability of those properties in the simulated wildfire-affected landscapes. Given the caveats imposed by the simplistic assumptions, some ecosystem properties appear to be sensitive to the arrangement of forest age classes. Future studies would benefit from the use of structurally based, rather than age-based, forest classes, for evaluating forest pattern effects on ecosystem properties.