Roles of maternal condition and predation in survival of juvenile Elk in Oregon

Abstract

Understanding bottom-up, top-down, and abiotic factors along with interactions that may influence additive or compensatory effects of predation on ungulate population growth has become increasingly important as carnivore assemblages, land management policies, and climate variability change across western North America. Recruitment and population trends of elk (Cervus canadensis) have been downward in the last 4 decades across the northern Rocky Mountains and Pacific Northwest, USA. In Oregon, changes in vegetation composition and land use practices occurred, cougar (Puma concolor) populations recovered from near-extirpation, and black bear (Ursus americanus) populations increased. Our goal was to provide managers with insight into the influence of annual climatic variation, and bottom-up and top-down factors affecting recruitment of elk in Oregon. We conducted our research in southwestern (SW; Toketee and Steamboat) and northeastern (NE; Wenaha and Sled Springs) Oregon, which had similar predator assemblages but differed in patterns of juvenile recruitment, climate, cougar densities, and vegetative characteristics.

We obtained monthly temperature and precipitation measures from Parameter-elevation Regressions on Independent Slopes Model (PRISM) and estimates of normalized difference vegetation index (NDVI) for each study area to assess effects of climate and vegetation growth on elk vital rates. To evaluate the nutritional status of elk in each study area, we captured, aged, and radio-collared adult female elk in SW (n = 69) in 2002–2005 and NE (n = 113) in 2001–2007. We repeatedly captured these elk in autumn (n = 232) and spring (n = 404) and measured ingesta-free body fat (IFBF), mass, and pregnancy and lactation status. We fitted pregnant elk with vaginal implant transmitters (VITs) in spring and captured their neonates in SW (n = 46) and NE (n = 100). We placed expandable radio-collars on these plus an additional 110 neonates in SW and 360 neonates in NE captured by hand or net-gunning via helicopter and estimated their age at capture, birth mass from mass at capture, and sex. We monitored their fates and documented causes of mortality until 1 year of age. We estimated density of cougars by population reconstruction of captured (n = 96) and unmarked cougars killed (n = 27) and of black bears from DNA analysis of hair collected from snares.

We found evidence in lactating females of nutritional limitations on all 4 study areas where IFBF_autumn was below 12%, a threshold above which there are few nutritional limitations (9.8% [SE = 0.64%, n = 17] at Toketee, 7.9% [SE = 0.78%, n = 17] at Steamboat, 7.3% [SE = 0.33%, n = 46] at Sled Springs, and 8.9% [SE = 0.51%, n = 23] at Wenaha). In spring, of females known to have been lactating the previous autumn, 48% (SE = 3.3%, n = 56) had IFBF_spring <2%, a level indicating severe nutritional limitations, compared to 20% (SE = 1.7%, n = 91) of those not lactating the previous autumn. These low levels of IFBF_spring of lactating females likely resulted from a carry-over effect of inadequate nutrition during summer and early autumn. We found a positive relationship between summer precipitation and IFBF_autumn in NE, and that IFBF_autumn of pregnant females was inversely related to birth date of their neonates the following spring (F_{1, 52} = 20.37, P < 0.001, R²_adj = 0.27). Mean pregnancy rates of lactating females were below 0.90, a threshold indicating nutritional limitations, at Toketee (0.67, SE = 0.12, n = 15), Wenaha (0.70, SE = 0.10, n = 23), and Sled Springs (0.87, SE = 0.05, n = 47) but not Steamboat (0.93, SE = 0.07, n = 14). Of elk where we sampled femur fat during winter in NE, we saw evidence of imminent starvation in 3 of 21 juveniles (12%) with all 3 killed by cougars, and 2 of 12 adult elk (17%) that both died from non-predation events. Birth mass was <13 kg for 6.5% and 2% of VIT neonates in SW and NE, respectively, a mass associated with reduced probability of survival in previous studies. Birth mass of VIT neonates was greater in Sled Springs ( = 18.3 kg, SD = 2.5, n = 59) than Steamboat ( = 16.3 kg, SD = 2.1, n = 21) or Toketee ( = 16.1 kg, SD = 2.8, n = 24) but not Wenaha ( = 17.1 kg, SD = 2.8, n = 36; F_{3, 132} = 7.63, P < 0.001). Median and mean birth date (29 May) for VIT neonates did not differ between regions (F_{1, 136} = 0.33, P = 0.56), but NE had greater variation around the mean, indicating a longer parturition interval. We documented 293 mortalities of juveniles across study areas and years, and predation was the proximate cause of mortality in 262 cases primarily from cougar (n = 203), black bear (n = 34), and other or unknown predation (n = 25). We also documented causes of mortality as unknown (n = 16), human-caused (n = 8), and disease or starvation (n = 7). We recorded abandonment of 2 (1.4%) and predation mortality of 4 (2.7%) VIT neonates prior to being collared. We found 4-fold differences between regions of subadult female and adult cougar densities (0.90–4.29/100 km²) and 2-fold differences within study areas across years, with cougar density lower in SW than NE. Black bear densities varied from 15–20/100 km² across our study areas.

We estimated survival of neonates to 30 days, 16 weeks, and 12 months using known fates models in Program MARK. Survival of neonates born to females with VITs was associated with cougar density, IFBF_spring, and female mass but not female age or neonate birth date or birth mass. Survival was higher for juveniles born to females with lower IFBF and mass in spring, opposite of what we predicted. In a post hoc analysis, we found females successful in raising their neonate to recruitment were more likely to be successful the following year compared to those not successful the previous year, which may explain this unexpected finding. As cougar density increased, survival of juveniles born to females of known nutritional condition declined.

We conducted separate analyses of survival by region for all neonates captured to evaluate effects of climate, bottom-up (but not maternal condition), and top-down factors. In NE, juvenile survival was little affected by annual variation in climate but decreased as cougar densities increased and as birth date became later. For SW, survival was higher with less April–May precipitation and for later born neonates but less affected by cougar density than observed in NE. Across our 4 study areas, survival varied annually from 0.61 (SE = 0.08) to 1.00 during the first 30 days, 0.41 (SE = 0.11) to 0.90 (SE = 0.09) the first 16 weeks, and 0.18 (SE = 0.06) to 0.57 (SE = 0.11) through 12 months (recruitment) with survival higher in SW than NE. Survival of juvenile elk was inversely related to cougar density through 30 days (F_{1, 18} = 16.59, R²_adj = 0.45, P < 0.001), 16 weeks (F_{1, 18} = 21.07, R²_adj = 0.51, P < 0.001), and 12 months (F_{1, 11} = 18.94, R²_adj = 0.60, P = 0.001). We found that as rates of cougar-specific mortality increased, juvenile survival declined ( = −0.63, 95% CI = −0.84 to −0.42) suggesting cougar predation was partially additive mortality because the estimated regression coefficient was significantly less than 0 but greater than −1. We did not observe a similar relationship with rates of black bear-specific mortality because the estimated regression coefficient overlapped 0, suggesting predation by black bears on juvenile elk was compensatory.

Our results suggest that recruitment in NE but not SW was primarily limited by predation from cougars, which was partially additive mortality. Given that we observed nutritional limitations that influenced juvenile survival in all 4 study areas, we were unable to explicitly quantify how much of the cougar predation was additive mortality. Thus, we caution that a reduction in cougar density may not result in an equivalent increase in recruitment, and maintaining or enhancing summer and winter ranges of elk in our study areas is also vitally important for sustaining populations and distributions. In SW, where cougar densities were lower, maintaining, and enhancing existing elk habitat may be the only management option to improve recruitment. Given the differences we found between regions monitored, basing management on an incomplete understanding of causative factors affecting elk population dynamics may result in ineffective actions to address low recruitment. © 2018 The Authors. Wildlife Monographs Published by Wiley Periodicals, Inc. on behalf of The Wildlife Society.