Optimal resource allocation in a serotinous non‐resprouting plant species under different fire regimes

Summary Serotiny, the retention of mature seeds in closed fruits within the canopy for over a year, is a common trait in fire‐prone environments. When competition with adult plants prevents seedling establishment between fire events and in the absence of post‐release soil seed dormancy, strong serotiny, i.e. the retention of all seeds until the next fire, appears as the best strategy. Despite the low levels of inter‐fire seed recruitment for several species in both Australian and South African fire‐prone environments, considerable variation in the duration of fruit retention is nevertheless observed among species. Our aim is to predict optimal age‐specific reproductive schedules in a perennial, serotinous species, when cone maintenance is costly. We focus on species where adults are killed by fire, without a soil seed‐bank. We explicitly consider a trade‐off between growth (which determines plant survival), seed production and seed maintenance. In our model recruitment relies upon fire events. We use dynamic programming to determine, for given fire regimes, the optimal pattern of resource allocation. We further study the effect of changes in fire regime on the viability of populations adapted to some historical fire regime. We find that, whenever maximal plant survival probability is low, the optimal strategy consists in reducing resource allocation to seed maintenance while increasing resource allocation to annual seed production. This illustrates a trade‐off between current and future reproduction. A low rather than a strong level of serotiny should evolve whenever the variance of fire intervals is large and the mean fire interval is low. Low levels of serotiny could constitute a bet‐hedging strategy with decreasing predictability of the arrival of fire. Once adapted to some historical fire regime, serotinous populations are highly sensitive to a change in mean fire frequency and to an increase in the variance of fire intervals. Populations adapted to a historically high level of variance in fire return are more robust to changes in fire regime. Synthesis: Life‐history trade‐offs and low predictability of fire intervals may favour low rather than strong levels of serotiny even when recruitment essentially occurs just after fire events.