We address the problem of the relative importance of the intrinsic chaos and the external noise in determining the complexity of population dynamics. We use a recently proposed method for studying the complexity of nonlinear random dynamical systems. The new measure of complexity is defined in terms of the average number of bits per time-unit necessary to specify the sequence generated by the system. This measure coincides with the rate of divergence of nearby trajectories under two different realizations of the noise. In particular, we show that the complexity of a nonlinear time-series model constructed from sheep populations comes completely from the environmental variations. However, in other situations, intrinsic chaos can be the crucial factor. This method can be applied to many other systems in biology and physics.