Sporadic seed production in mast seeding trees

Seed production, germination and establishment are critical processes determining structure of plant populations and composition of terrestrial ecosystems. In many long-lived plants, seed production shows strong variation between years with a gradient ranging from years without seed production to years with heavy seed crops (masting). In a number of plant populations, this intermittent production of very large seed crops is synchronized over large areas and sometimes even across species. Mast seeding events are common in boreal, temperate and tropical forest ecosystems and have spurred considerable scientific interest. For the causes of this phenomenon and its evolutionary advantages, a few hypothesis have been brought forward: the satiation hypothesis assumes that in mast years the food availability for seed eaters all of a sudden rises high leading to an over-satiation. This would leave seeds for the tree population. A second hypothesis emphasizes the higher pollination success of wind-pollinated plants in years where many plants have many flowers. The basis for seed production on a plant level is a good enough resource status a factor which can trigger mast years as well. As the number of long-term studies of the phenomenon grows, a new picture of temporal variation in tree seed production is emerging. In particular, many studies have documented both the irregular length of intervals between very large seed crops, and the more frequent production of small or moderate seed crops between the years of peak seed production. To date, there has been little examination of the consequences of these inter-mast seeding events for the long-term population dynamics of tree species. The general goal of the proposed research is thus to understand and predict the consequences of spatial and temporal variation in tree seed rain for long-term population dynamics of tree species. We hypothesize that inter-mast seed production is a bet-hedging strategy for capturing windows of opportunity for successful regeneration after stochastic disturbances for species with longer masting cycles, and that the long-term average abundance of species is increased as a result of inter-mast seed production. We further hypothesize that, for species with palatable seeds, inter-mast seed production is particularly successful when seed predators are driven into hunger cycles by mast years, and subsequent crashes in seed predator populations result in lower than average population sizes during inter-mast years. The hypotheses will be tested using a combination of empirical research, e.g. on seed fate, combined with the parameterization of a spatially explicit individual based tree population model (SORTIE-ND). Seed fate will be documented by tracking of marked seeds. Small mammal population dynamics as a driver of seed survival and seedling input into the tree population will be modeled by the MARK package.

Seed production by forest tree populations shows large changes between years - in some years, very few or no seeds are produced, but in the next year, the forest floor is covered with hundreds of seeds per square meter. This behavior, called mast seeding, often covers large areas and has important consequences for seed-dependent animal populations. One of the big unresolved questions about masting is whether it is only in years of large seed crops that young trees establish into the forest, or whether intermediate years also contribute to the age structure of the population. This is important because to date, scientists assumed that only high seed years lead to high survival of seeds and seedlings and eventual recruitment into the adult population because in other years, not enough seeds are pollinated or they are consumed by rodents. However, our results from 14 years of detailed monitoring of seed production and population dynamics in an old-growth montane forest dominated by beech, spruce and fir reveal that the effect of annual variation in seed production on seedling densities declines rapidly, suggesting that high levels of seedling mortality counteract the yearly effect of masting. In contrast, however, the spatial signature of seed rain does affect recruitment of new trees - areas with high seed density and light show the highest probability of seedling establishment. Using live-trapping, we also conducted yearly monitoring of small rodent populations to investigate how quickly they can respond to changes in seed production. We found that forest mice and bank voles responded differently to seed crops, depending on the type of forest they occupied. While forest mouse populations showed similar fluctuations with seed crops across all habitats, bank voles only settled into open habitat following years of very large crops, thus changing their habitat preference based on population density. This finding suggest that open areas next to stands that just produced intermediate seed crops may provide a safer location for seeds than elsewhere. Using seeds marked with small radio-tags, we investigated the role of small rodents as seed dispersers and predators in more detail, and found that in years of high seed density, rodents quickly remove most of the seeds and that transported seeds, which could potentially be dispersed and establish, are taken too deep into the soil to survive. Contrary to common assumptions, seed transporting by small rodents plays a negligible role in terms of seed dispersal. To study the combined effects of masting on rodent populations, seed predation, and the competition among trees through which seedlings grow into adults, we are currently using high-resolution simulations based on our study plots, to determine the conditions under which masting actually results in recruitment of adult trees.