| To explore mechanisms of carbon storage of subtropical forest at individual, community and ecosystem levels. We applied the hierarchical Bayesian model and structural equation models and some other approaches to analysis the sampling tree data and inventory data from the Badagongshan (BDGS) 25 ha plot in Hunan provinces and 107 plots in Hubei provinces. We studied biomass estimate models, biomass distribution and mechanisms, and then we studied carbon storage in tree architecture and relationships between species diversity and forest biomass. The main results are highlighted as follows:(1) We analyzed 147 trees with above-ground part and 23 roots data to develop forest biomass allometric models. We tested the importance of wood density and crown radius in model development, besides, we examined the suitability of foreign models developed in the subtropical forests. Results showed that wood density can improve the model performance, but crown radius cannot, most of foreign models gave a bad prediction to our data. Therefore, we developed a suite of above-and below-ground biomass models in this area.(2) We used these models to estimate forest biomass of BDGS 25 ha plot in Hunan province, Total forest biomass ranged widely from 39.16 to 1024.53Mg/ha with a mean of 252.7 ± 108.7 Mg/ha in the 20 m×20 m quadrats of the BDGS plot. Forest biomass was highly variable in the BDGS plot, however, biomass changes were related to topography; the ridges always had higher biomass and the valleys lowest. Overall, large tree (DBH≥ 25 cm) density accounted for 71% of variation in forest biomass. Variance partitioning of forest biomass showed that 64.8% of the variation was explained by biotic, topographic and spatial factors together. Scatter plots of four significant predictors of biomass showed an increasing trend with increasing convexity, stem density and wood specific gravity, but the decreasing trend as TWI increased.(3) We applied the hierarchical Bayesian model to analyze the tree architecture of 59 subtropical species and detect the variation among intra- and inter-species and functional groups. We also used Kendall rank correlation to examine these architecture relationships changes as stem matures considering the adult stature and light requirement. Tree architecture parameters showed limit convergence at community-wide, and most parameters did not deviated from the community tendency. Large stature species had slender stems, shallow and narrow crowns in small diameter class, but had deep and wide crowns in large diameters. While light demanding species had thick stems in understory but slender stems above canopy; light demanding species had wide crowns at the intermediate size classes.(4) We analyzed data from 107 plots with varying tree species richness, stand ages and soil nutrients in China subtropical forest. Besides, we applied structural equation models to examine the effects of stand ages, topography, soil nutrient variables on rarefied species richness and forest biomass relationships. The positive relationship between species diversity and forest biomass was found and can be reversed by adding the large tree density. Furthermore, we found most of large trees were light-demanding species. Forest biomass and species richness showed positive associations with stand age, altitude via the large tree density, but showed controversial results on the soil nutrient regime. Our results showed the significant associations with environment variables in natural ecosystems. Therefore, the diversity-forest ecosystem function can be driven by resources partitioning via the large trees, the strong light requirement of most large trees indicates sampling effects also shaped the DFB relationships. |
PDF Full Text Request