Random environmental influences, such as snow cover, are widely regarded as an integral feature of caribou population dynamics. We conducted computer simulations to explore the ramifications of such stochastic variability for caribou demography. We devised 4 models with increasing levels of complexity: Model 1, density-independence under different levels of stochasticity and r; Model 2, non-linear effect of snow cover on r; Model 3, non-linear effect of snow cover on r and stochacticity as a function of population size; and Model 4, non-linear effect of snow cover on r; stochasticity as a function of population size, and density-dependence according to the logistic equation. The results of Model 1 indicated that nearly all caribou populations subject only to environmental vagaries experienced either extinction or irruption. Model 2 revealed that non-linear effect of snow cover depressed the realized r as a function of population size. Finally, Model 4 suggested long-term population as previously reported in literature, but with reduced chance of overshooting K under moderate to high environmental variability.