Population dynamics in Digitalis purpurea:: the interaction of disturbance and seed bank dynamics

1. Plant ecologists have long since realized that the persistence of many facultative biennial plants depends upon disturbance. However, we still have a limited knowledge of the population-level effects of disturbance, and the connection between adult and seed bank dynamics. 2. Using data from a 3-year demographic study combined with experimental gap-opening in a large population of Digitalis purpurea, we parameterized stochastic transition matrix models in 'disturbed' vs. 'undisturbed' areas. We simulated different gap sizes (fraction of population that was disturbed) and temporal disturbance patterns (constant, random, regular and irregular return intervals) and evaluated the effects on population growth rate and seed bank dynamics. To explore seed bank importance we used two alternatives for seed bank survival rate (0.75/0.35) and three alternatives for seed bank recruitment fraction (0.9/0.5/0.1). 3. Observed background recruitment levels were insufficient to ensure a positive population growth rate. Increased amounts of gap-opening led to higher growth rates, and population persistence was predicted at moderate disturbance levels if seed bank survival was high (0.75). 4. Temporal disturbance pattern affected model results; random and interval scenarios resulted in lower population growth rates and higher extinction risks than constant scenarios of the same average disturbance level. Small and frequent disturbances led to considerably higher growth rates than large and rare disturbances. 5. Stochastic elasticity analyses identified the seed bank as the most important life cycle stage with respect to population growth and persistence in most scenarios, and its relative impact was positively related to seed bank survival rate and negatively related to disturbance level. Variation in the recruitment fraction from seed bank vs. seed rain affected both population growth rate and elasticity patterns, indicating the large impact of spatial variation in seed bank density. 6. Synthesis: Despite the existence of a large seed bank, our data suggest that recruitment may be locally seed-limited due to a patchy seed bank structure. Local population development may consequently differ widely from gap to gap. These results illustrate how spatial structures in both seed bank, adult population and gap formation interact to shape plant population dynamics, as well as the occurrence of microsite- vs. seed-limitation.