To explore the regeneration dynamics of major tree species in an evergreen broad-leaved forest in Gutianshan National Nature Reserve, a 5-ha dynamic plot was established in 2002. All woody plants with DBH (diameter at breast height) ≥1 cm were tagged, mapped, measured and identified to species. Based on data from one census each in 2002 and 2007, we analyzed the characteristics of dead and recruited individuals to estimate mortality and recruitment rates, and population change for 43 tree species. The results showed the following: (1) A total of 4,758 stems were recruited and 1,384 died during the five years, and the stem increment ratio was consequently 26.05%. (2) The amounts of recruits within three growth forms (tree, subtree or shrub, and sub-shrub) all declined as DBH size class increased; a pattern different from that observed in dead individuals. (3) The average annual mortality and recruitment rate of these 43 tree species was 2.26% and 5.90%, respectively. (4) Comparing mortality and recruitment rates in each DBH class, we found that 19 subtree or shrub species consistently had higher recruitment than mortality rates in small DBH class, and almost no difference in both middle and large DBH classes. (5) 31 out of 43 species showed population growth, while 12 species’ populations declined slightly. Populations of 10 species increased or decreased at an annual rate of >5%, indicating that they were experiencing fast turnover. Among these, Chimonanthus salicifolius (45.13%) and Itea oblonga (16.35%) had the highest increment rate in population size, whereas Loropetalum chinense (-3.10%) and Quercus serrata var. brevipetiolata (-1.86%) had the highest rate of decline. In conclusion, both recruitment and mortality rates varied among species and among DBH size classes, pointing to the unique regeneration dynamics of the 43 tree species. These differences might be associated with patterns of density dependence in addition to habitat fluctuations.

We determined the best-fit model for, and explored the mechanisms shaping species-abundance distributions (SADs) by fitting five widely-used SAD distribution models at several scales. We used data collected in 2005 from a 24-ha dynamic plot in an evergreen broad-leaved forest in the Gutianshan National Nature Reserve. We estimated SAD at different sampling scales from the mean value of SADs taken from 100 randomly-selected subplots within the 600 m×400 m Gutianshan plot. We subsequently used the SADs to test the fit of different models, including the broken stick, lognormal distribution, niche preemption, Zipf, Zipf- Mandelbrot, and neutral models. We employed AIC and χ²values to test goodness-of-fit for these models. All computations were conducted using the Vegan package in R 2.7.1. At smaller scales (10 m×10 m and 20 m×20 m), the broken stick, lognormal distribution, niche preemption, Zipf, and Zipf-Mandelbrot models all fit well to the observed species-abundance distribution. The Zipf-Mandelbrot was the best model at the 20 m×20 m scale. The Lognormal was the best-fit model at the 40 m×40 m scale, and the Zipf-Mandelbrot model was the only suitable one in explaining the observed SAD at scales of 60 m×60 m and 80 m×80 m. None of these models performed well at a scale of 100 m×100 m, but the neutral model was better at explaining patterns of SADs at larger scales (40 m×40 m to 100 m×100 m) than smaller scales and it is suitable in explaining patterns of SADs at all scales. Patterns in SAD were scale-dependent, suggesting that SADs at different scales are likely structured by different ecological processes.

We inventoried tree damage within a 24-ha plot in Gutianshan National Nature Reserve immediately after the ice storm that affected south China in 2008. This ice storm caused severe damage to one third of trees and relatively slight damage to another third of trees in the plot. Results from our multinomial logistic regression analysis showed that all of the four examined factors, diameter at breast height (DBH), habitat type, plant life form and leaf habit (evergreen/deciduous), had close relationship with the damage types of forest trees. The results of multinomial logistic regression model indicated that evergreen canopy trees in low valleys with larger DBH suffered more severe damage. Detrended correspondence analysis was employed to compare variation in forest community composition before and after the ice storm. Variation in community composition before and after the ice storm differed significantly from stochastic variation modeled with null models at scales of 40 m×40 m. This result indicates that some species are more prone to ice-storm damage than others, which may result in directional changes in community composition.

To study litter production, composition, temporal dynamics, and the effects of an ice storm on litter production in a 24-ha evergreen broad-leaved forest dynamic plot in Gutianshan National Nature Reserve, Zhejiang, we set up 169 seed traps, and collected litterfall weekly from October 2006 to December 2009. Total annual litter production in 2007 and 2009 was 532.05 g/m² and 375.17 g/m², respectively. We attribute the remarkable drop in production due to an ice storm in February 2008. Leaves, twigs, bark and miscellaneous materials accounted for 79.0%, 14.7%, 3.3% and 3.0% of mass, respectively. The annual leaf litter production of four dominant tree species (Castanopsis eyrei, Schima superba, Pinus massoniana and Quercus serrata var. brevipetiolata) contributed about 71.36% of the total leaf litter production in 2007. There were two litter fall peaks each year, one in spring (April) and the other in autumn to early winter (between late October and early December). Total litter production, leaf production and twig production decreased significantly after the ice storm (P<0.05), as did the annual leaf production of C. eyrei and Distylium myricoides (P<0.05) and S. superba (P<0.1). The production of miscellaneous materials mainly composed of insect dung increased (P<0.01) in April immediately following the ice storm, indicating that a compensating growth of branches and leaves was occurring in the damaged community.

All free-standing trees within a 6-ha plot with diameter at breast height (DBH)≥1 cm were tagged, mapped, measured and identified to species. The spatial distribution patterns of four dominant canopy tree species in the plot were analyzed using a point pattern analysis Ripley’s L-function. A total of 12,131 free-standing individuals were recorded, including 68 species belonging to 49 genera and 25 families. Lithocarpus hancei had the highest basal area and the largest importance value. The second was Castanopsis wattii. Camellia forrestii, an understory tree species, ranked third in terms of importance value, although it showed the highest abundance (1,712 individuals). And this evergreen broad-leaved forest did not contain an obviously dominant species. The four dominant canopy species had a large number of seedlings and saplings and tended to be patchily distributed. We also examined the spatial distribution of 26 tree species with ≥40 individuals at each growth period (DBH ≥1 cm). Thirty-seven of 53 life history stages showed aggregated distribution pattern, suggesting that habitat heterogeneity may be more important than density dependence for regulating the population spatial structure of most tree species in the subtropical forest studied here.

Community structure is one of the key features in the process of vegetation succession. Warm temperate mixed deciduous broad-leaved secondary forest is the main forest vegetation type in China’s warm temperate zone. In order to better study the mechanisms of succession and species coexistence in this forest type, we established the Donglingshan 20-ha forest plot (DLS plot) in 2009 and 2010 using the same protocol as the well-established plot on Barro Colorado Island in Panama. In the plot, all free-standing woody plants with DBH (diameter at breast height) ≥1 cm were mapped, tagged, and identified to species. Here, we address preliminary results on floristic characteristics, community composition, and size-class, vertical, and spatial structure of the DLS plot. These datasets will serve as baseline information accessible to a wide range of future studies. We tagged a total of 52,136 genotype individuals (103,284 individuals including branches of genotype individuals), belonging to 58 species, 33 genera and 18 families. All of these tree species were deciduous. Floristic characteristics of the community suggested a temperate flora, including some subtropical and tropical elements. There were very obvious dominant species in the plot. Five species comprised 61% of all individuals, and 20 species comprised 92% of all individuals, while the other 38 species comprised only 8% of all individuals. Vertical structure was composed of an overstory layer (19 species), midstory layer (18 species), and a shrub layer (21 species). The DBH size-class structure of all species in the plot generally fitted a “reverse J” distribution, indicating good regeneration across the community. The size-class structure of the main species in the overstory layer showed a bimodal or nearly normal distribution, while the most abundant species in the midstory and shrub layers showed “reverse J” or even “L” distributions. Spatial distribution patterns of the dominant species varied with size-class and scale and shifted from closer aggregation to looser aggregation from small to adult or old trees. The size-class spatial distribution patterns of the dominant species showed the different diameter levels of their own individuals occupied different spatial positions in the plot.

Ailao Mountain National Nature Reserve covers 504 km² and is one of the largest tracts of evergreen broad-leaved forests in China. A 6-ha plot was established in the reserve in 2008 by Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences, for the purpose of monitoring long-term dynamics of the forest. Tree seedlings were sampled in this plot to understand their composition and spatio-temporal distribution. Five of the top 10 seedling species in terms of importance values were the same as 5 of the top 10 adult tree species with the highest importance values. Both abundance and species richness of tree seedlings dropped between the end of the rainy season and the end of dry season, likely due to drought stress. Seedlings in canopy gaps were richer in species, more abundant in terms of density, and experienced lower mortality compared with those under the forest canopy, suggesting that forest gaps facilitate the recruitment of seedlings in the forest.

The following are typical communities in the secondary deciduous broad-leaved forests in the Dongling Mountain area: Quercus wutaishanica forest (QWF), Betula dahurica forest (BDF), Juglans mandshurica forest (JMF) and mixed forest (MF). Comprehensive research on species composition and community structure of these four forest types will contribute to a more thorough knowledge of the successional dynamics and ecological functions in warm temperate deciduous broad-leaved forests. Following the field protocol of the 50-ha plot on Barro Colorado Island in Panama, one 1-ha plot was established in each of these forest types in Dongling Mountain in 2009. In these plots, all free-standing trees >1 cm in diameter at breast height (DBH, 1.3 m above ground) were mapped, tagged, and identified to species. As for species composition, there were 22 species, belonging to 20 genera and 20 families in the QWF plot; 22 species, belonging to 18 genera and 16 families in the BDF plot; 30 species, belonging to 23 genera and 21 families in the JMF plot; and 19 species, belonging to 15 genera and 14 families in the MF plot. Data on species abundance, mean DBH, basal area and importance value indicated that the dominant species were obvious within each community. The size-class structure (DBH) of all species in the four plots generally conformed to a reverse “J” distribution, indicating good community regeneration. The size-class structure of a given species varied among plots, suggesting that size-class structure may be related to community type. For example, the size-class structure of Quercus wutaishanica showed a skewed normal distribution in the QWF plot, but a reverse “J” distribution in the BDF plot. The size-class structure of Betula dahurica showed a normal distribution in both the QWF plot and the BDF plot, but differed greatly in the range of DBH’s.

Tree mortality, usually resulting from interactions among multiple factors, is a crucial process in forest dynamics. Using two census datasets (2004 and 2009) from a 25 ha plot in the Changbai Mountains, we analyzed the composition, size class structure and spatial distribution of individual trees (DBH ≥ 1 cm) that died during the 5-year period. The number of species went from 52 in 2004 to 51 in 2009, with 3 species disappearing and 2 others appearing. The number of individuals changed from 36,908 to 34,926, with 4,030 dying and 2,048 being recruited. The number of dead individuals accounted for 10.9% of total individuals in 2004. Species with high mortality also tended to have high recruitment. Compared with tree species, shrub species had both higher mortality and recruitment rates. In addition, 44 species showed an increased mean DBH in 2009, while the mean DBH of 5 other species decreased. Mortality decreased as DBH increased. Size class distributions of dead dominant species in different vertical layers were similar in the two censuses. Spatial distributions of dead individuals were species-specific. Dead individuals from smaller size classes were spatially clumped at small scales and became randomly spaced at larger scales. However, dead individuals from larger size classes tended to show random distribution at various scales.

Both unified neutral theory and niche theory have played an important role in understanding the mechanisms of species coexistence in tropical rain forests. Using Torus-translation tests, we examined the relationships between microtopography (elevation, convexity, and slope) and the distribution of 89 wood species with ≥5 individuals and diameter at breast height (DBH) ≥ 1.0 cm in a 5-ha permanent plot in a subtropical broad-leaved evergreen forest in Baishanzu, Zhejiang Province. We classified plants into three growth stages (sapling, juvenile and mature stages) based on their diameter-classes, and compared relationships between microtopography and the distribution of plants in different growth stages for each species. Species with less than 40 individuals always showed few associations with their habitats, while most common species in the 89 studied showed significant associations with their habitats. The majority of species exhibited shifts in habitat preference among growth stages; the exceptions were Rhododendron latoucheae, Cyclobalanopsis stewardiana and Cleyera pachyphylla, whose associations with habitats were similar at all stages. Our study suggests that topographical differentiation is important for maintaining species diversity in this subtropical broad-leaved evergreen forest, and that habitat associations of species change ontogenetically.

Evergreen broad-leaved forest (EBLF) is a typical vegetation type in low elevation regions in eastern China. The permanent plot-based approach to community analysis is fundamentally important for revealing mechanisms of biodiversity maintenance. In this study, community structure and species composition were investigated and analyzed using a 20-ha permanent plot in Tiantong National Forest Park, Zhejiang Province. Our results were as follows: (1) we counted a total of 94,603 individuals, belonging to 152 species, 94 genera and 51 families. The three most dominant families were Theaceae, Lauraceae and Fagaceae. The genera present were dominated by tropical (52.1% in total) and temperate (42.6% in total) floras; (2) species in the evergreen life form were dominant with a community importance value of 80.3%. The three most dominant species were Eurya loquaiana, Litsea elongata and Choerospondias axiliaris. Fifty five species were considered rare; (3) the size distribution of all trees followed a reverse “J” shape; and (4) evergreen species showed a greater resprouting ability than deciduous species. In conclusion, with rich species composition and a mature community structure, the EBLF in Tiantong region is typical of this vegetation type.

Point pattern analysis of species in a community is important in gaining a better understanding of the underlying ecological processes controlling the observed structure. In this paper, univariate and bivariate spatial point pattern analysis based on the pair-correlation function were used to evaluate the spatial patterns of Quercus variabilis. We compared the spatial patterns and associations of Q. variabilis at four different growth stages, and the relationships between standing trees and snags of Q. variabilis in two fully-mapped 1-ha forest plots in the Baotianman National Nature Reserve. We found that: (1) Aggregation was the main pattern type of Q. variabilis in the plots, with one plot showing stronger aggregation and the other exhibiting more randomness. (2) At different growth stages of Q. variabilis in the two plots, aggregation was found in younger stages and randomness in the older. (3) Positive interactions were found between the different growth stages of Q. variabilis in Plot I, and negative interactions in Plot II. (4) There are positive associations between standing trees and snags of Q. variabilis in Plot I, and spatial segregation in Plot II. Our study suggests that the spatial distribution of Q. variabilis may be influenced by environmental heterogeneity. This is important because the spatial distribution of this species in turn affects self regeneration of Q. variabilis populations in Baotianman National Nature Reserve.

To explore the composition, distribution, and height structure of tree seedlings in an evergreen broad-leaved forest in the Dinghu Mountain, 149 seed traps were set up, surrounded by seedling plots, in a 20-ha dynamic plot. Seedling plots were censused in March, 2008 and a total of 2,632 seedlings were recorded. Multiple linear regression was used to analyze the relationship between the abundance of seedlings for nine species and six variables, including the number of conspecifics within 10 m away from a seed trap, the sum of conspecific basal area within 10 m away from a seed trap, and convexity, slope, aspect, and elevation. Factors influencing seedling abundance differed among species. Three species were related only to topographic variables, one species was related only to the conspecifc variables, four species were influenced by both topography and conspecific abundance, and one species was influenced by neither. For all species except Memecylon ligustrifolium, significant partial regression coefficients were positive, indicating that seedling abundance increases with the number or the basal area of conspecifics within 10 m away from a seed trap. This research provides indirect evidence for dispersal limitation and niche theory.

Exploring tree population distribution patterns and interspecific spatial associations are helpful in elucidating the mechanisms underlying species coexistence in forest communities. We analyzed population distribution patterns and interspecific adult-adult spatial associations of common tree species at scales of 0-50 m in five 1-ha warm temperate secondary forest plots near Beijing, China. We found that: (1) all species showed aggregated spatial patterns at some scales; aggregation occurred mainly at neighborhood scales of < 15 m, tended to peak within a 1-m radius around focal conspecific trees, and the percentage of species exhibiting a random or regular pattern increased with scale, mainly occurring at scales of > 15 m; (2) the proportion of species pairs showing non-significant associations was high (~50%), and even in those species pairs that showed significant associations, segregation and partial overlap were dominant association types. Few species pairs (~4%) showed mixing. We feel that population spatial distribution of trees, particularly the observed prevalence of conspecific aggregation, in these plots was regulated by seed dispersal limitation and environmental heterogeneity. Moreover, aggregated distributions also promoted interspecific segregation and partial overlap. It is possible that distribution patterns were associated with habitats. Few species pairs showed interspecific mixing, in this case, interspecific competition exclusion difficultly occured, but in the interior of conspecific aggregation, density dependence should be a dominant mechanism regulating population distributions. Our findings contribute to a clearer understanding of the mechanisms influencing the structure of these forests.

Community phylogenetic structure contains information about evolutionary relationships among coexisting species in a community. Studies on community phylogenetic structure provide new insights into ecological processes and the mechanisms underlying community restoration. We compared the phylogenetic structure of forest communities along a human disturbance gradient using data from 12 1-ha plots in secondary forests within the Gutianshan National Nature Reserve, Zhejiang Province, China. We found that, at the 20 m×20 m scale, the plantation forest community (I) was phylogenetically overdispersed. Forest communities II (naturally-restored forest 50 years after clear cutting and 20 years after heavy selective cutting), III (naturally-restored forest 50 years after clear cutting), and IV (old-growth forest) were phylogenetically clustered. Forest types II and IV were the most clustered. We also conducted the same analyses for four forest types grouped into different DBH classes, and found that forests II, III, IV with DBH less than 5 cm and 5-10 cm were phylogenetically clustered, but forest I was overdispersed. In contrast, all the forests with DBH more than 10 cm were phylogenetically overdispersed except the forest IV. Our results suggest that seed dispersal is the main process resulting in phylogenetic overdispersion at early stage of succession, and habitat filtering is main process leading to phylogenetic clustering when habitat heterogeneity increases during succession. Possibly as a result of habitat filtering, forest IV was always phylogenetically clustered regardless of which size class was examined.

To elucidate the driving factors behind plant functional traits, especially in mountainous areas, we explored how variation in topography and soil characteristics affects ecophysiological and morphological traits of woody plants within the Gutianshan 24-ha plot in Zhejiang Province. This site is in a typical subtropical broad-leaved forest. During the summers of 2008 and 2009, we measured three ecophysiological traits (chlorophyll content, chlorophyll fluorescence parameters F_v/F_m and PI_ABS, and stem sapwood xylem specific hydraulic conductivity) and four morphological traits (stomata density, specific leaf area, leaf thickness, and ratio of leaf length to leaf width) among 115 woody species in the field. Redundancy analysis was conducted to identify the most influential environmental factors from our topographical factors (mean elevation, convexity, slope, and aspect) and five soil parameters (moisture, total nitrogen content, total phosphorus content, total carbon content, and pH). Leaf chlorophyll content was negatively correlated with elevation and convexity, but positively correlated with soil moisture and total nitrogen content. The two chlorophyll fluorescence parameters (F_v/F_m and PI_ABS) were strongly negatively correlated with soil total nitrogen and phosphorus content; these variables were likely interrelated with low soil pH values. Stem sapwood xylem specific hydraulic conductivity was positively correlated with soil moisture and specific leaf area was positively correlated with elevation. Topographical variables explained 10.4% of total variation in functional traits; mean elevation was the most powerful explanatory variable, followed by convexity, slope and finally, aspect. Soil parameters explained 13.9% of the total functional trait variation; and soil moisture was the most powerful factor, followed by total nitrogen content, pH value, total phosphorous content, total carbon content and the ratio of nitrogen and phosphorous. At this scale and at this site elevation and convexity were the two most influential topographical factors, and moisture and nitrogen were the most influential soil factors, on plant functional traits. However, because of a lack of phosphorous in the soil, some important physiological processes including photosynthesis may have been limited, thereby leading to some confusing trait/environment relationships.