The ecological significance of canopy seed storage in fire-prone environments: a model for resprouting shrubs

1. Based on a parameter set describing mean life-history attributes for the resprouting shrub Banksia attenuata in south-western Australia, the relationship between fire interval and extent of canopy seed storage (serotiny) was analysed. Deterministic and stochastic computer models were used to try to identify the circumstances under which serotiny is favoured, and what degree of serotiny maximizes the estimated finite rate of natural increase (lambda) for species which, in addition to serotiny, have the capacity to resprout vegetatively. 2. The deterministic model indicated a maximum finite rate of natural increase (lambda = 1.015) when the fire interval was 13 years and all seeds were retained on the plant until fire occurred (i.e. complete serotiny). Populations were predicted to decline if the fire interval exceeded about 40 years. 3. Changes to biological attributes associated with timing of reproduction shifted the optimum fire interval but did not change the overall pattern in relation to serotiny. Reduction in the probability of adults resprouting after fire also shifted the optimum, so that as probability of resprouting declined, optimum fire interval increased. 4. Increasing stochasticity around the mean fire interval (CV > 50%) under regimes of frequent fire (fire interval < 20 years) led to strong, but not complete, serotiny as the optimum strategy (lambda = 1.009 +/- 0.001). If fire was less frequent, and the probability of inter-fire recruitment was similar to that of post-fire recruitment, then the optimum strategy was zero serotiny. 5. The probability of seedling recruitment was low after most fires due to low seed availability and high seedling mortality over the first summer. Stochastic weather (variation in summer rainfall) resulted in a higher rate of population growth (lambda = 1.019) than under constant average weather conditions, emphasizing the importance of rare recruitment events in the population behaviour of resprouting species. 6. Resprouters were able to cope with more frequent fire than non-sprouters, but there was no evidence to support the contention that they are also favoured under regimes of infrequent fire.