EVIDENCE FOR VARIABLE FIRE SEVERITY AND NON-EQUILIBRIUM DYNAMICS IN PRE-MANAGEMENT ERA DRY FORESTS OF THE INLAND NORTHWEST, USA (POSTER)

HESSBURG, P.F., SALTER, R.B. AND JAMES, K.M.

USDA Forest Service, Pacific Northwest Research Station, Forestry Sciences Laboratory, 1133 North Western Avenue, Wenatchee, WA 98801-1229

Pre-management era (ca. 1900) dry forests of the Inland Northwest are represented in the fire literature as types that were tightly coupled with low severity fires. To evaluate the validity of this premise, we identified three Ecological Subregions in the eastern Oregon and Washington Cascades, USA that contained extensive dry forests. We randomly sampled ~10% of the area and subwatersheds of each Subregion, and photo-interpreted the vegetation attributes of each patch in each subwatershed from the oldest available continuous stereo coverage of aerial photos (1930s to 1950s). Attributes included total crown cover; crown cover, species composition, and size classes of both the understory and overstory; and number of canopy layers. To remove the effects of early selection cutting, we statistically reconstructed the vegetation attributes of all patches showing any evidence of harvesting. We then classified patches as being last affected by low, mixed, or high severity fires using published percent canopy mortality values (≤ 20%, 20.1-69.9%, ≥ 70%, respectively), attributes of the reconstructed forest structure, and in some cases, the cover type. We found that highly variable mixed severity fires were the prevailing type in forests of all three Subregions, and more common than expected in the dry forests. Patterns of pre-management era dry forest structure were apparently formed by a mix of low, mixed, and high severity fires. This suggested that variable fire severity and non-equilibrium dynamics rather than low fire severity and equilibrium fire dynamics were at work. We found that dry forests and their cool-moist analogues were more similar than different in their distribution of pre-management era fire severity and that the potential vegetation type poorly separated patches by their expected fire severity. We concluded that one reason that low severity fires have been so strongly coupled with dry forests in the past is that estimates of historical fire severity have been largely based on point rather than patch or area observations. Point-based observations of fire severity across a landscape record fires in individual recorder trees, but there is a bias for allocating samples to topographic and physiographic settings that have a history of low impact fires. Patch scale observations, on the other hand, can account for the spatial extent of all three fire severities directly. A worthwhile addition to existing sampling designs would be to include both point and patch scale estimates of fire severity when studying landscapes. The point observations would register the events for which recorder trees remain and distribute them spatially. In addition, understory cohorts could be sampled evenly across the landscape and aged to determine whether they were initiated in response to events registered on the set of surviving recorder trees in the near vicinity, or in response to other events not represented by the recorders.