POPULATION VIABILITY ANALYSES OF CHAMAECRISTA KEYENSIS: EFFECTS OF FIRE SEASON AND FREQUENCY

Abstract

Conservation of species of concern in fire-dependent ecosystems requires restoring the historical fire regime that has shaped the species' life history. The historical fire regime of the endangered pine rockland ecosystem is controversial. In this study, we attempted to infer the historical fire regime of this ecosystem by studying the effects of contrasting experimental fire regimes on Chamaecrista keyensis, a narrowly endemic species of the pine rocklands of the Lower Florida Keys, USA. We constructed multiple matrix population models for C. keyensis using demographic data from replicated wet- and dry-season burns, 1–30 years since fire, and seed bank dynamics derived from extensive field seed bank experiments. We then carried out deterministic analyses and stochastic simulations. Recently burned sites (1–2 years postfire) had the highest finite population growth rates of all sites. Differences in finite population growth rate between winter and summer fires were more pronounced during the year of burn than in subsequent years. Stochastic simulations show that dry-season (winter) burns generated lower extinction risks and population decline probabilities than a wet-season (summer) fire regime. Fire return intervals of 5–7 years generated the lowest extinction and population decline probabilities for both seasons. C. keyensis may have evolved under a fire regime that consisted of both anthropogenic dry-season fires along with lightning-caused wet-season fires. If phenology is also a key to fire responses in other pine rockland species, a diverse fire regime including a wide range of fire seasons may be critical to maintaining the diversity of the pine rockland ecosystem.