Correcting mortality rates of asocial animals with prolonged periods of maternal dependence in demographic modeling

Background/Question/Methods

Demographic modeling of population growth, such as that done with Leslie matrices, is a commonly employed tool in wildlife management and conservation that has been used to regulate harvest of bobcats, assess the recovery of gray wolves, and compare different strategies for the conservation of cheetahs. Although population modeling is inherently based on a number of simplyfying assumptions (e.g., timing of reproduction and mortality, transitions between age-classes, etc.), some of these simplifications can result in substantial deviation from reality if not properly re-evaluated in certain situations. For example, demographic models of asocial animals that exhibit a prolonged period of maternal dependence (e.g., leopards, moose, and bears) often disregard the strong correlation between the survival of the mother and that of the young. In such populations where juvenile survival depends almost entirely on the survival of the mother, the death of a female would naturally lead to the death of any of her dependent young. Failure to account for this dynamic can lead to signficant overestimation of population growth. Here we propose an analytically derived correction factor that adjusts the apparent mortality of the juvenile age-classes in matrix modeling to include deaths of maternally dependent individuals that result from mother mortality.

Results/Conclusions

Failure to account for deaths of maternally dependent individuals resulting from mother mortality in matrix models resulted in a significant over-estimation of population growth over a 20-year period in all theoretical modeled scenarios (p=0.86-0.98). Use of the correction factor significantly improved the accuracy of the matrix model in recreating population growth rates across a wide range of mortality and fecundity rates, numbers of age-classes, and duration of maternal dependence (p=0.019-0.038). This analytically-derived correction factor provides an easy and accurate way to implicitly account for this added mortality in Leslie matrices and other matrix models without having to resort to more complex individual-fate models that can explicitly account for this dynamic. Because these models are often used to aid managers in setting harvest limits and regulations, overestimation of population growth can severely influence a population’s persistence and risk of extinction. We believe that the robustness and simplicity of this correction factor, combined with the risk of overestimating population growth in its absence, should encourage wildlife managers and biologists to incorporate it into employed matrix models and use it to fine-tune population projections, evaluate the effects of harvest timing on population growth, and better inform current and proposed management strategies and regulations.