The California spotted owl (Strix occidentalis occidentalis) has been at the center of political and administrative debate due to its association with commercially valuable forest. Several studies have compared the forest cover types used by California spotted owls with the cover types that are generally available, establishing the association between spotted owls and old/large tree components of forests at the landscape scale. We sought a deeper understanding of spotted owl habitat associations in areas in which owls had already selected territories. We mapped and classified vegetation within circular plots (radius 2.4 km) around 67 spotted owl sites in northeastern California, USA. We evaluated the relationships between habitat composition within the different owl sites and variation in (1) nest success (1990–2000) and (2) site occupancy, apparent survival probability, and reproductive output (1993–1998). All analyses included data representing 2 spatial scales: core area (814 ha) and nest area (203 ha). Site occupancy was positively associated with the amount of the nest area dominated by large trees with high canopy cover within the nest area. It was negatively associated with the amount of nonhabitat (nonforested areas and forest cover types not used for nesting or foraging) and with medium-sized trees with high canopy cover. Site occupancy also decreased with time and elevation. Apparent survival probability varied annually and was positively related to the area of each habitat class multiplied by the quotient proportion used/proportion available for each type, at both the nest and core scales. Reproductive output was negatively related to elevation and nonhabitat within the nest area. Nest success was positively associated with the presence of large remnant trees within the nest stand.