A wildfire burned over 40,000 ha of conifer and mixed conifer-hardwood forest in the Silver Creek drainage of southwestern Oregon in the fall of 1987 allowing me to assess big game use of a large natural burn. I used fecal pellet group counts to estimate habitat use and effects of forest management activities on Roosevelt elk (Cervus elaphus roosevelti) and black-tailed deer (Odocoileus hemionus columbianus) within the Silver Fire Recovery Project Area (SFRPA) of the Siskiyou National Forest. Pellet decay rate and differences in observers' abilities to detect deer and elk pellet groups (interobserver variability) were estimated to test validity of pellet group counts. Pellet group persistence was estimated during a 10-month period. There were no differences in pellet group persistence between elk and deer (P < 0.05). Observers differed in ability to detect elk (F = 2.7; df = 4, 530; P = 0.03) and deer (F = 10.7; df = 4, 883; P < 0.0001) pellet groups. Interobserver variability related to elk pellet groups was low and was attributed to differences in numbers of transects searched by each observer. Two observers detected greater mean numbers of deer pellet groups than did other observers. I counted 775 elk pellet groups and 3,888 deer pellet groups on four study areas within the SFRPA. I analyzed habitat use for two periods: June to mid-October (summer-fall), and mid- October through May (fall-spring). I used stepwise logistic regression to create models predicting categories of habitat use during each period. Management variables were added to the habitat models to estimate effect of management on predicted categories of habitat use. Total overstory canopy cover was negatively related to deer use during both use periods. Distance to road was the only significant management variable affecting deer habitat use during the fall-spring period (P = 0.03). Slash cover had a negative effect on probability of habitat use by deer during the summer-fall period (P = 0.02). Elk use was negatively affected by steep slopes and hardwood canopies during both periods, while grass seeding positively affected elk use during both summer-fall (P = 0.05) and fall-spring (P = 0.03) use periods. Clearcutting had a negative effect on probability of elk use during the fall-spring period (P = 0.04). / Graduation date: 1995

Forage availability, diets, distribution, and productivity were compared for Roosevelt elk, Cervus elaphus roosevelti, using improved (i.e. seeded, fertilized, and grazed) and untreated areas of the Oregon Coast Range. Seasonal forage availability was substantially different on 1-year-old improved and untreated clearcuts, but the differences diminished by clearcut age 4 and essentially disappeared by clearcut age 7. Elk diets, as evaluated by fecal analysis, were similar in plant species composition and DAPA concentrations on both areas. There was no significant difference in estimated calf birth weights or survival, or in pregnancy and lactation rates, kidney and metatarsal marrow fat concentrations, or breeding dates of adult cows using improved and untreated areas. Summer and winter calf ratios were similar in 7 of 8 seasons sampled. There was no significant difference in home range or core area size, and distances traveled were similar on both areas in all seasons except winter. Elk exposed to summer sheep grazing were displaced an average of 1211 m (SD = 28 m) for an average of 14 weeks (SD = 5 weeks). Elk using improved areas stayed closer to forage areas in spring, but otherwise there was no significant difference in proximity to forage. Elk from both areas demonstrated preferential use of meadows, and used other forage areas in proportion to their availability. The combined results indicated that the forage improvement practices were ineffective in producing any measurable benefits for elk. / Graduation date: 1994