Understanding and predicting changes in the abundance of natural populations is a central goal of ecology. These changes are influenced by a variety of exogenous processes (weather, floods, fire); variation in these processes leads to variation in vital rates (survival, fecundity) that may be positively or negatively correlated across the life cycle. We used 20 years of data and a hierarchical Bayesian model to estimate vital rates and their covariation in an endangered plant, Dicerandra frutescens ssp. frutescens (Lamiaceae), as a function of time since fire and random year effects. Germination and the number of flowering branches declined with time since fire, and all plants were increasingly likely to become nonreproductive with time since fire. Time since fire had negative effects on survival of seedlings, vegetative plants, and small flowering plants, and positive effects on survival of medium and large flowering plants. Model comparison strongly supported inclusion of time-since-fire effects and weakly supported inclusion of year effects influencing all vital rates (“model-wide” year effects). We used samples from the joint posterior distribution of model parameters to simulate population dynamics as a function of fire regime and year-to-year environmental variation. These simulations suggest that populations of Dicerandra frutescens ssp. frutescens are least likely to go extinct if the average time between fires is ∼24–30 years. The design of the simulations allowed us to distinguish variation in stochastic population growth associated with process variability (fire, year effects, and demographic stochasticity) from variation associated with parameter uncertainty (finite amounts of data). Even with 20 years of data, half or more of the uncertainty in population growth rates was due to parameter uncertainty. This hierarchical Bayesian population viability analysis illustrates a general analytical framework for (1) estimating vital rates as a function of an exogenous environmental factor, (2) accounting for covariation among vital rates, and (3) simulating population dynamics as a function of stochastic environmental processes while taking into account uncertainty about their effects. We discuss future areas of development for this approach.