Survival and Breeding Transitions for a Reintroduced Bison Population: a Multistate Approach

Abstract

The iconic plains bison (Bison bison) have been reintroduced to many places in their former range, but there are few scientific data evaluating the success of these reintroductions or guiding the continued management of these populations. Relying on mark-recapture data, we used a multistate model to estimate bison survival and breeding transition probabilities while controlling for the recapture process. We tested hypotheses in these demographic parameters associated with age, sex, reproductive state, and environmental variables. We also estimated biological process variation in survival and breeding transition probabilities by factoring out sampling variation. The recapture rate of females and calves was high (0.78 6 0.15 [SE]) and much lower for males (0.41 6 0.23), especially older males (0.17 6 0.15). We found that overall bison survival was high (.0.8) and that males (0.80 6 0.13) survived at lower rates than females (0.94 6 0.04), but as females aged survival declined (0.89 6 0.05 for F L 15 yr old). Lactating and non-lactating females survived at similar rates. We found that females can conceive early (approx. 1.5 yr of age) and had a high probability (approx. 0.8) of breeding in consecutive years, until age 13.5 years, when females that were non-lactating tended to stay in that state. Our results suggest senescence in reproduction and survival for females. We found little support for the effect of climatic covariates on demographic rates, perhaps because the park's current population management goals were predicated from drought-year conditions. This reintroduction has been successful, but continued culling actions will need to be employed and an adaptive management approach is warranted. Our demographic approach can be applied to other heavily managed large-ungulate systems with few or no natural predators.

Pregnancy rates in central Yellowstone bison

“…The age of primiparity among ungulate populations is highly variable and strongly influenced by population density, forage availability, and weather conditions (Eberhardt , Eberhardt ; Gaillard et al , Gaillard et al ). Our detection of a single pregnant 1‐year‐old bison is consistent with reported low levels of breeding by yearling plains bison (Meagher , Green , Berger and Cunningham ) and wood bison (Fuller ); however, breeding by yearling female plains bison has not been detected in other populations (Haugen , Pyne et al ). Female American bison of both subspecies more commonly breed for the first time between 2 and 4 years (Fuller , Haugen , Wolfe et al ).…”

Plausible combinations: An improved method to evaluate the covariate structure of Cormack-Jolly-Seber mark-recapture models

“…Even if such methods correctly adjust for the number of times a covariate is used, our view is that the use of model weights to assess the importance of individual covariates is overemphasized. Covariate combinations define model structure and covariates are rarely uncorrelated, hence model structure is more effectively assessed by examining cumulative weights associated with unique covariate combinations [38]. Note that the subset of models evaluated under the PC strategy is naturally balanced for covariate combinations.…”

Bison Weights From National Parks in the Northern Great Plains

“…The park manages for a population of 700 bison, a level based on drought conditions. 6 The bison probably consume about 12% of plant productivity in a normal-precipitation year.…”

“…For purposes of presenting the average weight of a bison within a herd I had to account for the sex and age composition of the herd. I used fecundity values from Millspaugh et al 9 and the mid-point of the survival rates from Millspaugh et al 9 and Pyne et al 6 to parameterize a demographic model; those studies used the Badlands and Wind Cave NP bison roundup databases. The modeled sex and age composition of a herd under an annual cull of 16% of each age and sex cohort is presented in Table 1.…”