ÁÔÆæÖØ¿Ú

We investigated recent forest mortality dynamics by revisiting a network of 1617 trees in unlogged roadless areas in the Blue Mountains of eastern Oregon for which we have precise age and growth data. There was extensive mortality of the oldest trees in stands—a quarter of trees ≥300 years of age died over a ten-year period. Trees with below average  increment and older trees were more likely to die in stands that have not experienced wildfire in the last 130 years. Smaller trees were more likely to die in stands that experienced wildfire during the ten-year observation period. Douglas-fir and western juniper experienced a lower rate of mortality than ponderosa , western larch, or grand fir in sites that have not experienced wildfire. There was little evidence for a difference in mortality between species in sites that experienced wildfire. We describe a novel simulation model which demonstrated that succession of young trees is unlikely to replace recent losses of old trees. The confluence of a rare insect defoliator outbreak, drought, and past management decisions that increased competitive stress and fuel connectivity is the most likely explanation for the elevated mortality we observed. The  where we collected data are not achieving old tree conservation objectives. Active management to reduce forest competition is the most realistic pathway for old tree conservation in dry forests.