Seeds move but trees stand still: quantifying the importance of spatial structure for plant populations

Background/Question/Methods

 A single tree may produce thousands to millions of seeds during its lifetime, and only one seed needs to recruit to adulthood to maintain the tree population at equilibrium. Consequently, size-structured population models generally reveal that the sensitivity of the population growth rate of long-lived plants to seed survival is low. However, vital rates of individual trees are also defined by location, including conspecific density in the tree’s neighborhood.  If the benefit of being located in a neighborhood with low conspecific density has a consistent effect across tree life stages, seed arrival in low-density locations may have major consequences for population dynamics, even if the sensitivity of the population growth rate to average seed survival is low. We quantified the importance of seed dispersal for a population of Miliusa lineata trees in Thailand, using a long-term dataset consisting of 15 years of spatially-explicit data on thousands of individuals. We parameterized inverse models relating survival and growth to conspecific neighborhood across the entire range of tree size, from seeds to large adults. Next, we built a spatially-explicit, individual-based model (IBM) for tree population dynamics and analyzed the sensitivity of this model to changes in seed dispersal.

Results/Conclusions

Our models show survival and growth benefits for trees of all life stages in neighborhoods with low conspecific density. However, the effect of conspecific density was strongly size-asymmetric, with the largest effect of conspecific neighborhood for small trees with large trees as neighbors. In addition, the strength and spatial scale of negative density dependence varied between life stages. Seed survival was significantly impacted by neighboring trees less than 10 m from the seed, while negative effects of neighboring trees were detectable for saplings up to 20 m away from conspecific neighbors. Including negative density dependent effects of neighborhood tree density in our IBM resulted in very different population dynamics than density-independent models or density-dependent models incorporating mean field density. Finally, our sensitivity analyses revealed that while, on average, larger individuals contribute more to the population growth rate, population dynamics can be very sensitive to the number of seeds dispersed to locations with low conspecific density. These results demonstrate that while most seeds are destined to die before recruiting into adults, seed dispersal can determine survival and growth rates for all subsequent life stages, and thus have major demographic consequences for tree populations.