Climate and fragment area jointly affect the annual dynamics of seedlings in different functional groups in the Thousand Island Lake

Abstract

Habitat fragmentation and climate change are the two main threats to global biodiversity. Understanding their combined impact on plant community regeneration is vital for predicting future forest structure and conserving biodiversity. This study monitored the seed production, seedling recruitment and mortality of woody plants in the Thousand Island Lake, a highly fragmented anthropogenic archipelago, for 5 years. We analyzed the seed-seedling transition, seedling recruitment and mortality of different functional groups in the fragmented forests and conducted correlation analyses involving climatic factors, island area, and plant community abundance. Our results showed that: 1) shade-tolerant and evergreen species had higher seed-seedling transition, seedling recruitment and survival rate than shade-intolerant and deciduous species in time and space, and these advantages increased with the island area. 2) Seedlings in different functional groups responded differently to island area, temperature and precipitation. 3) Increasing active accumulated temperature (the sum of the mean daily temperature above 0 °C) significantly increased seedling recruitment and survival, and warming climate favored the regeneration of evergreen species. 4) The seedling mortality rate of all plant functional groups increased with the increase of island area, but the increasing strength weakened significantly with the increase of the annual maximum temperature. These results suggested that the dynamics of woody plant seedlings varied among functional groups, and can be regulated separately and jointly by fragmentation and climate.

Keywords: fragmentation; habitat filtering; regeneration niche difference; seed-to-seedling transition; seedling recruitment; seedling survival; species coexistence.

Figure 1

Seedling dynamics [seed-seedling transition (A, D), seedling recruitment (B, E) and seedling mortality (C, F)] from 2018 to 2022.

Figure 2

Seedling dynamics [seed-seedling transition (A, D, G), seedling recruitment (B, E, H) and seedling mortality (C, F, I)] of different functional groups on islands with different area from 2018 to 2022. “s”, “m”, and “l” represent small, medium and large islands, respectively. Seedling dynamic of all species (grey), evergreen (EG, green) and deciduous (DC, yellow) species, Shade-intolerance (SI, red) and shade-tolerance (ST, blue) species were shown. Different letters between functional groups or island types indicated significant differences.

Figure 3

The relative influence of potential factors [the community abundance (Abundance), island area, annual maximum temperature (T_max ), annual precipitation (P_total ) and active accumulated temperature (T_active )] on the transition (A, D, G), recruitment (B, E, H) and mortality (C, F, I) rate of seedlings of evergreen(EG, green), deciduous(DC, yellow), shade-intolerat(SI, red), shade-tolerant(ST, blue), and all (black) species. Solid points with asterisk marks indicate a significant effect (significant codes for p-value: p= 0~ 0.001 “***”, p= 0.001~ 0.01 “**”, p= 0.01~ 0.05 “*”). If the coefficient value of the explanatory variable is<0, it means a significant negative correlation. The parameter estimate<0 indicated a significant negative effect. The parameter estimate intersects with 0 indicated an insignificant effect. The parameter estimate >0 indicates a significant positive effect.

Figure 4

Linear regression correlation between z-value and annual maximum temperature (T_max : (A–E), annual precipitation (P_total : (F–J), and the annual active accumulated temperature (T_active : (K–O) for all species (A, F, K), evergreen species (B, G, L), deciduous species (C, H, M), shade-tolerant species (D, I, N), and shade-intolerant species (E, J, O). The five points in the panels represent the z-index values calculated based on year-level climatic factors during 5 years. To be noted, the p-value in this figure was calculated based on the results of LM with time offset.