HABITAT SELECTION BY SPOTTED OWLS DURING NATAL DISPERSAL IN WESTERN OREGON

The probability of successful natal dispersal may influence the viabil ity of northern spotted owl (Strix occidentalis caurina) populations w ithin fragmented forests. We tested 6 null hypotheses examining the re lations between habitat selection, forest fragmentation, and the proba bility of mortality for juvenile northern spotted owls during natal di spersal in western Oregon, 1882-85. Older forest (old-growth and matur e stands) was the vegetation type used most frequently during transien ce (35.3%) and colonization (61.2%), and spotted owls selected closed- canopy forests over open-canopy forests during both phases of dispersa l (P < 0.05). The hypotheses that spotted owl habitat selection was in dependent of forest fragmentation were not rejected for either transie nt (P = 0.51) or colonization (P = 0.13) dispersal. Likewise, net disp ersal distance was independent of forest fragmentation (P = 0.92). Use of open sapling stands during transient dispersal decreased the proba bility of mortality (Wald test = 2.21, P = 0.03), whereas use of clear cuts during colonization dispersal increased the probability of mortal ity (Wald test = 1.95, P = 0.03). Net dispersal distance did not affec t the probability of mortality (P = 0.67), yet a negative relation exi sted between dispersal distance and the amount of clearcut used during transient dispersal (P < 0.005). These results provide additional evi dence of the selection of older forests by spotted owls and how their natal dispersal may be affected by harvesting such forests; particular ly, that use of clearcuts may decrease the probability of successful n atal dispersal.