| The tropical rain forests have complicated structures, contain rich species and play an important role in conservation of global biodiversity and maintaining universal ecosystem functions. A classification of diverse species in tropical forests based on their different functional traits will greatly contribute to the scientific study and management for tropical forests. In this study, field investigations were conducted in totally 15 ha sample plots of recovery (second-growth) and old-growth forest vegetation in the Bawangling forest region of Hainan Island, South China. The recovery vegetation included different stages of fallows on the shifting cultivation secondary forest stands and logged forests under different logging systems. Based on the field investigations and each functional trait (including life form, potential height, buttress size, wood density, deciduous versus evergreen, seed mass and dispersal agents), plant species in the tropical montane rain forest were classified into different functional groups respectively. NMS (nommetric multidimensional scaling)was used to analyze their change with different environmental factors, disturbance types and recovery time. Then the relationships between different functional traits were analyzed by PCA (principal component analysis). Focusing on all these functional traits, 13 functional groups were identified by TWINSPAN (two-way indicator species analysis), and their change with different environmental factors, disturbance types and recovery time was ordinated by DCCA(detrended canonical correspondence analysis). At last, ecological keystone species in each functional group were identified from five aspects. The main results were as follows: Based on landform, the tropical montane rain forest can be classified into two forest stand types including mountain slope forest stands and ravine forest stands. Based on disturbance types, the tropical montane rain forest can be classified into old-growth forest stands and secondary forest stands (mainly including the logging and shifting cultivation).Trees and shrubs were the most abundant components of tropical montane rain forest, secondarily grasses and lianas, while epiphytes and palms were least in relative abundance. Relative abundance of trees was positively correlated with soil organic matter and the distance to old-growth forest, which had the highest distribution in the secondary forest stands. Relative abundance of shrubs was negatively correlated with stone content, which had high distributions in each forest stands, except for ravine old-growth forest stands. Relative abundance of the life forms such as grasses, palms and epiphytes had positively correlation with soil moisture content, soil stone content, pH, total P and available K, which had more abundance in the ravine old-growth forest stands than in mountain slope old-growth forest stands. Relative abundances of lianas weren't correlated with any environmental factors, which had no difference distribution in different forest stands.1. Shade-tolerant functional groups had the most relative biomass and abundance in each forest stands. Among them, strong shade-tolerant functional groups were mainly distributed in the ravine old-growth forest stands, while less shade-tolerant functional groups were mainly in mountain slope old-growth forest stands. Shade-intolerant functional groups were mainly distributed in secondary forest stands, especially in recovery secondary forest stands after shifting cultivation. With the increasing of recovery time after shifting cultivation, shade-tolerant functional groups increased in relative biomass and abundance, while shade-intolerant functional groups decreased. Relative biomass of stong shade-tolerant functional groups was positively correlated with recovery time, stone content, pH, total P and available K, and negatively with elevation and soil organic matter, while those of shade-intolerant functional groups were negatively correlated with soil available N and total N.2, Functional groups with moderate height (15-25m) had the most relative biomass and abundance in each forest stands, while those with maximal height were mainly distributed in the ravine old-growth forest stands, secondarily in mountain slope old-growth forest stands. With the increasing of recovery time after shifting cultivation; functional groups in the canopy layer of forest stands increased in relative biomass and abundance, while other functional groups had no significant chance. Relative biomass of functional groups in the canopy layer of forest stands were positively correlated with soil stone content, pH, total P, available K and recovery time, and negatively with soil organic matter.3. Functional groups with larger buttress had the most relative biomass and abundace in the canopy layer of forest satnds, especially in the canopy layer of ravine old-growth forest stands, while those with smaller buttress were mainly in secondary forest stands. With the increasing of recovery time after shifting cultivation, functional groups with larger buttress increased in relative biomass and abundance, while smaller buttress decreased. Relative biomass of the functional groups with the largest buttress size was positively correlated with recovery time, stone content, soil pH, total P and available K, and negatively correlated with elevation, soil P and soil organic matter.4. Functional groups with lighter wood density had the most relative biomass and abundace in secondary forest stands, secondarily in ravine old-growth forest stands, while those with heavier wood density were in mountain slope old-growth forest stands. With the increasing of recovery time after shifting cultivation, functional groups with lighter wood density decreased in relative biomass and abundance, while those with heavier wood density increased. Relative biomass of functional groups with heaviest wood density was positively correlated with elevation and soil organic matter, and negatively correlated with logging intensity, recovery time, soil stone content and pH.5. Functional groups with lighter seed mass had the most relative biomass and abundance in shifting cultivation secondary forest stands and ravine old-growth forest stands, while those with heavier seed mass appeared in logged secondary forest stands and mountain slope old-growth forest stands, which showed water difference had more influence on see mass than both disturbance types and recovery time. With the increasing of recovery time after Shifting cultivation, the functional groups with lighter seed mass decreased in relative biomass and abundance, while those with heavier ones increased.6. Functional groups by animal dispersal had the most relative biomass and abundance in each forest stands, especially in old-growth forest stands, while functional groups by auto and wind dispersal had the less. Functional groups by auto dispersal were mainly distributed in ravine old-growth forest stands or recovery secondary forest stands after shifting cultivation, while those by wind dispersal mostly appeared in secondary forest stands. With the increasing of recovery time after shifting cultivation, the functional groups by atuo dispersal increased in relative biomass and abundance, while those by wind dispersal decreased. Relative biomass of functional groups by animal dispersal were negatively correlated with logging intensity and soil available K, while those by auto were positively correlated with soil stone content and total P, and those by wind dispersal were positively correlated with logging intensity.7. Evergreen functional groups had the most relative biomass and abundance in each forest Stands, while deciduous functional groups had the least and mostly appeared in the secondary forest stands. With the increasing of recovery time after shifting cultivation, evergreen functional groups increased in relative biomass and abundance, while the deciduous ones decreased. Relative biomass of evergreen functional groups were positively correlated with elevation, aspect, soil available P and organic matter, and negatively with soil pH and available K, while those of deciduous functional groups were positively correlated with logging intensity.8. Based on the 7 functional traits, major plant species in the tropical montane rain forests could be classified into 13 functional groups. The most important factors (traits) were potential height, buttress size and seed mass and wood density, secondarily shade-intolerant and defoliation/evergreen, while seed dispersal was least in importance. Among them, potential height was positively correlated with seed mass and buttress size (P＜0.05); and seed mass was negatively correlated with shade-intolerant, auto and wind dispersal, and positively correlated with wood density and animal dispersal (P<0.05); and the factors such as shade-intolerant, seed mass, wood density and dispersal agents were all singificantly correlated with defoliation/evergreen (P<0.05).9. The functional groups in ravine old-growth forest stands, partially composed of deciduous species by auto or wind dispersal, had higher potential height and larger buttress than those in mountain slope ones, while those in mountain slope old-growth forest stands, mainly composed of evergreen species by animal dispersal, had heavier seed mass and wood density. With the increasing of recovery time, the potential height, seed mass and wood density of functional groups increased, and deciduous species by wind dispersal changed to evergreen species by animal dispersal.10. Functional groups in the canopy layer of forest stands had higher potential height, larger buttress size, heavier seed mass and wood density than those in the lower layer, which can greatly reflect the gradient change of environmental factors including soilavailable N, available P, available K, pH, organic matter, moisture content and elevation. However, the functional groups in the lower canopy or understorey layer, which had smaller buttress or lighter seed mass, had weak indication to environmental changes.11. To identify the ecological keystone species, five methods were adopted including relative biomass, relative basal area, relative abundance, relative frequency and importance value. The results showed that the ecological keystone species in each functional group could well resemble their repective functional groups' change with environmental and disturbance characteristics, except for a few species that were greatly influenced by seed sources. The ecological keystone species identified in this study can well represent the main characteristics of their functional groups,...
|
PDF Full Text Request