Partial life-cycle models: how good are they?

Partial life-cycle models are simplified, matrix-based population models that allow demographic analyses using incomplete demographic data. Number of parameters in most partial life-cycle models are fixed, and this permits direct interspecific comparison of population dynamics and life histories. Because of these desirable properties, partial life-cycle models have recently received substantial applications. However, adequacy of partial life-cycle models has not been investigated using empirical data. I tested the adequacy of a post-breeding census partial life-cycle model by examining whether and to what extent dynamical properties of the age-structured model are retained in the post-breeding census partial life-cycle model. I applied the age-structured Leslie matrix model and a post-breeding census partial life-cycle model to demographic data for 142 populations of mammals, and compared the projected population growth rate (lambda), and elasticity of lambda to changes in ages at first and last reproduction, juvenile survival, adult survival, and fertility estimated from the two models. Population growth rate estimated from the two models were practically indistinguishable, and elasticities estimated from the two models were very similar. These results suggest that dynamical properties of the age-structured model are generally captured in the partial life-cycle model, and that application of the post-breeding census partial life-cycle model to age-structured or partial demographic data will not substantially compromise precision of analyses nor conclusions of an investigation. These results have substantial implications in basic and applied population ecology. because age-specific demographic data are seldom available for many species of conservation concern, and also because partial life-cycle models permit perturbation analyses involving ages at first and last reproduction using standard analytical techniques. (C) 2003 Elsevier B.V. All rights reserved.