Spotted owls, great horned owls, and forest fragmentation in the Central Oregon Cascades

Nocturnal surveys were conducted in February - May 1989 andJanuary - May 1990 to locate great horned owls (Bubo virginianus) andnorthern spotted owls (Strix occidentalis caurina) throughout the rangeof forest fragmentation levels in the Central Cascades of Oregon.Forest fragmentation levels ranged from landscapes (^ 500 ha in size)containing intact stands of mature/old-growth forest (0% fragmentation)to landscapes containing younger stands with no mature/old-growth forest(100% fragmentation). Six survey visits were made to each of 469calling stations located along 28 roadside survey routes. Total lengthof survey routes was 535.8 road km; relative abundance for great hornedowls and spotted owls was 0.069 and 0.139 owls/road km, respectively.Owl response rates were examined for differences 1) during the night, 2)by moon phase, and 3) by month during the survey period. Great hornedowls responded less than expected before midnight and more than expectedafter midnight, less than expected during full moon and more thanexpected during new moon phases, and less than expected during Januaryand April of the survey period. Spotted owls responded more thanexpected from 1800-1959 hr, more than expected during full moon phases,and generally more than expected during May of the survey period. Thirteen habitat/landscape variables within 500-ha circularlandscape plots surrounding 77 great horned owl, 103 spotted owl, 70no-owl, and 70 random points were assessed. Significant differences.existed between great horned and spotted owl landscapes for 6 variables:great horned owl landscapes contained more shrub/forb and shelterwood,less mature/old-growth and interior habitat, had a higher linear edge-to-mature/old-growth area ratio, and were higher in elevation thanspotted owl landscapes. The amount (Z t SE) of mature/old-growth forestwas 48% t 2% around great horned owls, 60% t 2% around spotted owls, 53%t 3% around no-owl points, and 53% t 2% around random points. Thegreatest number of great horned owl responses were associated withlandscapes containing 10-20% old forest. Great horned owl responsesgenerally declined with increasing amounts of old forest, and few (11%)great horned owls were detected in landscapes containing > 70% oldforest. The majority (62%) of spotted owls were detected withinlandscapes containing > 60% old forest. Spotted owl responses generallydeclined with declining amounts of old forest, and few (7%) spotted owlswere detected within landscapes containing The spatial distribution of old forest stands was compared todispersed (checkerboard) and clumped landscapes: 95% of great hornedowl, 88% of spotted owl, 89% of no-owl, and 86% of random landscapeswere classified as dispersed. Clearly, the forests of the CentralCascades are very highly fragmented. A method for linking owl biologyand landscape level plot size is described.