Variation in biotic interactions mediates the effects of masting and rainfall fluctuations on seedling demography in a subtropical rainforest

Seedling dynamics are critical for determining community composition and future forest structure. Two mechanisms are crucial drivers of seedling vital rates: biotic interactions and environmental filtering. Biotic interactions at the neighbourhood scale, such as the Janzen-Connell effect, are considered among the main mechanisms maintaining species coexistence. These effects can show considerable temporal variation because of changes in neighbourhood density (i.e. seedling recruitment following masting years) and weather conditions. However, there is limited evidence of whether interannual variation in these conditions modulates biotic interactions and its potential cascading effects on seedling vital rates. We used seedling data collected over 17 years in a subtropical rainforest in northern Taiwan. We modelled interannual variation in first-year survival and growth of 12,830 seedlings of 36 species and 47 cohorts using (generalized) linear mixed-effects models with several crowding indices and abiotic conditions (edaphic and topographic factors) as predictors. The first-year seedling survival and growth depended on both weather conditions and neighbourhood density. Seedling survival was greater when recruitment was more abundant and in years with greater rainfall, and this was mostly mediated by the interaction with conspecific and heterospecific seedlings. Seedling growth decreased with the increasing abundance of new recruits and increased with annual rainfall. Growth was negatively affected by the basal area of conspecific and heterospecific trees and by the density of conspecific seedlings when the number of new recruits was greater. The abiotic factors had a limited effect on seedling survival and growth compared to biotic factors. Synthesis. Our results show that variation in recruitment abundance due to masting and annual rainfall significantly affects seedling vital rates by changing biotic interactions. These findings are in accordance with the economy of scale hypothesis, which predicts higher per-capita reproductive success in years of higher seed production. Furthermore, the results support the hypothesis that fluctuations in rainfall can alter biotic interactions in forests. Efforts to predict forest community responses to climate changes should consider the temporal variations in biotic interactions.

Automated summary

Seedling dynamics in relation to biotic interactions and environmental conditions have a significant impact on forest structure and composition. This study showed that interannual variation in recruitment abundance and annual rainfall strongly influenced seedling vital rates, mediated by biotic interactions. These findings support the importance of considering temporal variations in biotic interactions when predicting forest community responses to climate changes.