No abstract

No abstract

Understanding population dynamics is critical for meta-population management, especially of endangered species, and also for megaherbivore ecology. We employed complete individual life records to construct census data for a reintroduced black rhinoceros population over 22 years since its founding and investigated its dynamics. Akaike's information criterion applied to scalar models of population growth based on the generalized logistic unambiguously selected an exponential growth model (r=0.102±0.017), indicating a highly successful reintroduction. No evidence of density dependence was detected, and thus, we could not confirm the threshold model of density dependence that has influenced black rhinoceros meta-population management. Our analysis supported previous work contending that the generalized logistic is unreliable when fitted to data that do not sample the entire range of population sizes. A stage-based matrix model of the exponential population dynamics exhibited mild transient behaviour. We found no evidence of environmental stochasticity, consistent with our previous studies of this population that found no influence of rainfall on demographic parameters. Demographic stochasticity was present, principally reflected in annual sex-specific recruitment numbers that differed from deterministic predictions of the matrix model. Demographically driven process noise should be assumed to be a component of megaherbivore population dynamics, as these populations are typically relatively small, and should be accounted for in managed removals and introductions. Increase in age at first reproduction with increasing population size, as manifested in the study population, may provide a warning of possible density feedback prior to detectable slowing of population growth rate for megaherbivores.

Population dynamics is a central component of demography and critical for meta-population management, especially of endangered species. We employed complete individual life records to construct census data for a reintroduced black rhinoceros population over 22 years from its founding and investigated that population's dynamics to inform black rhinoceros meta-population management practice and, more generally, megaherbivore ecology. Akaike's information criterion applied to scalar models of population growth based on the generalized logistic unambiguously selected an exponential growth model (r = 0.102 ± 0.017), indicating a highly successful reintroduction, but yielding no evidence of density dependence. This result is consistent with, but does not confirm, the threshold model of density dependence that has influenced black rhinoceros meta-population management. Our analysis did support previous work contending that the generalized logistic is unreliable when fit to data that do not sample the entire range of possible population sizes. A stage-based matrix model of the exponential population dynamics exhibited mild transient behaviour. We found no evidence of environmental stochasticity, consistent with our previous studies of this population that found no influence of rainfall on demographic parameters. Process noise derived from demographic stochasticity, principally reflected in annual sex-specific recruitment numbers that differed from deterministic predictions of the matrix model. Demographically driven process noise should be assumed to be a component of megaherbivore population dynamics, as these populations are typically relatively small, and should be considered in managed removals and introductions. We suggest that an extended period of exponential growth is common for megaherbivore populations growing from small size and that an increase in age at first reproduction with increasing population size, manifest in the study population, may provide a warning of density feedback prior to detectable slowing of population growth rate for megaherbivores .

The proportion of females calving (PFC) each year has been employed as an indicator of population reproductive performance in ungulates, especially for species that breed annually, because it requires less detailed population data than inter-birthing intervals and age at first reproduction. For asynchronous breeders with inter-birthing intervals longer than a year such as megaherbivores, however, it is unclear how much annual variation in PFC is expected and whether false signals of density feedback or environmental influence might result from analyzing PFC data. We used census data from a well studied, closed, expanding population of black rhinoceros (Diceros bicornis) to study annual variation in PFC over 22 years. Our analysis of PFC data yielded no false signals of density feedback but weak evidence for an unexpected influence of rainfall. The PFC data exhibited considerable variation, which we attribute to autocorrelation in the time series of PFC data, ‘demographic-founding effects’, changes in stage structure, and demographic stochasticity, some of which the modelling of PFC appears to confuse with an influence of rainfall. We expect such variation to be common in introduced populations and to persist for some years, complicating the interpretation of PFC, though moving averages of PFC can help if employed cautiously. While our analysis does not undermine the possible utility of PFC, the analysis and interpretation of PFC values require care.