Dynamic elasticities of Soay sheep in stochastic environments

Perturbation analysis aims to identify the demographic rate of pivotal importance for population growth.  The population growth rate is a function of deterministic and stochastic processes, and even relatively small amounts of stochasticity may appreciably alter any conclusions drawn.  Despite this influence, such analyses often fail to address the frequently violated assumptions of density-independent growth and a stable age-distribution in one, constant environment.  The population dynamics of the Soay sheep population on St Kilda is characterized by over-compensatory density-dependence.  What consequences does analysis in a dynamic rather than static environment have on the relative importance of demographic rates on population growth rate?  Stochastic (E^S) and deterministic (E^D) elasticities did not differ qualitatively.  E^S calculated in Markovian environments, which are flexible ways of modelling environmental change, can be partitioned to assess the importance of different moments of demographic rate distributions and different habitat states.  Elasticities of mean rates were more influential on E^S than elasticities of variation.  The overall mean is a function of the means in each habitat state, and the relative importance of age-classes and demographic rates was inconstant between states.  This has the potential to impact the probability of population persistence.  Deterministic analysis, whilst superficially acceptable, may reduce predictive power in a changing and uncertain world.