SPATIAL PATTERNS OF TREE MORTALITY IN AN OLD-GROWTH ABIES-PSEUDOTSUGASTAND

The processes that generate spatial patterns of organisms are often in ferred from analysis of the patterns themselves. To test the reliabili ty of such inferences, we examined spatial patterns of tree mortality over 16 years in a permanent 1-ha study area where causes of mortality are reasonably well known. The area is an old-growth stand in the Cas cade Mountains of Oregon, dominated by Abies amabilis (60% of stems bu t only 12% of basal area), Abies procera (52% of basal area), and Pseu dotsuga menziesii (22% of basal area). Mortality was dominated by the two Abies species. Dying A. amabilis occupied lower canopy strata and often had been suppressed or damaged by falling limbs or trees. Dying A, procera occupied upper canopy strata and often had been attacked by pathogens. The positions of dying trees and causes of mortality sugge sted that dying A. amabilis should be found near canopy trees and that dying A. procera should be aggregated. These hypotheses were tested b y quadrat- and distance-based methods. On average, dying A. amabilis w ere not significantly closer to canopy trees than expected for a rando m distribution or closer than random selections of trees from the enti re population of A. amabilis. Dying A. procera were aggregated at a sc ale of about 20 m; however, the degree of aggregation was nor signific ant compared with the distribution of the entire population. The poten tial for large falling trees to strike other trees at some distance fr om their bases, and the underlying non-random distribution of some tre e populations, may limit the extent to which causes of tree mortality can be inferred from spatial patterns.