Seedling traits and community dynamics data in Xishuangbanna Forest Dynamic Plot

Study siteThe study was conducted in a seasonal tropical rainforest in Xishuangbanna, southwestern China (101°34′E, 21°36′N) (Fig. 1). This area is situated in the northern edge of tropics with a mean annual temperature of 21.8 ℃. A strong monsoonal climate, with a rainy season from May to October and a dry season from November to April influences the vegetation composition and structure (Cao et al. 2006). We used monthly precipitation data during the study period (2007-2017) obtained from the Mengla National Reference Meteorological Station (101°34′E, 21°28′N, 633.4 m above sea level), Mengla County, Yunnan, Province. These monthly data were summed by season and were used in our data analysis. Over the study period, the annual precipitation averaged 1562.1 ± 71.7 mm with ~80% occurring during the rainy season (Fig. S1).A 20-ha permanent plot was established in 2007 in Xishuangbanna to monitor the spatial and temporal dynamics of the forest. The plot was established following a standardized protocol (Condit 1998). All free-standing trees with ≥1 cm diameter at breast height (DBH), were tagged and identified to the species-level. In November 2007, a total of 150 seed-trap stations were established in the plot using a stratified random design in order to monitor seed rain and seedling dynamics in time and space. During seedling plot site selection, streams and large rocks were avoided (Fig. 1). Each seed-trap station consisted of a seed-trap and three 1 m × 1 m seedling plots positioned 2 m away from the sides of the seed trap. In each seedling plot, woody plants < 1 cm DBH were defined as seedlings and were tagged and identified to the species-level. The height of each seedling was measured from the ground to the highest apical bud. The first seedling survey was carried out in November 2007. We conducted subsequent surveys at the end of dry season (May) and at the end of rainy season (November) of the following 10 years (i.e. 20 census intervals from 2007 to 2017). We defined the data obtained in May as the dry season data, and the data obtained in November as the rainy season data.Functional traits collection of tree seedlingsWe quantified functional traits for all species in this study. Specifically, we randomly selected three seedlings per species from the area surrounding the forest plot and harvested them in November 2019. Harvested seedlings were < 50 cm in height, but no longer retained cotyledons. Leaf thickness (LT) was measured by micrometer. Leaf chlorophyll content (Chl; μmol/L) was measured with a Chlorophyll Meter (SPAD-502Plus, Konica Minolta Inc. Osaka, Japan). For these measurements we took three leaves from each individual, and three readings from the widest portion of each leaf blade to the narrowest portion. Harvested seedlings were divided into roots, stem, and leaves and their fresh masses were measured. Leaves were scanned using a digital scanner (Canon 5600F Canon Inc., Tokyo, Japan) and individual leaf area (LA; cm2) was determined using the R package LeafArea (Katabuchi 2017). Next, all material from each seedling was oven-dried for 48 h at 65 ºC and then weighed for dry mass. The leaf and stem dry matter content (LDMC and SDMC 100 × dry mass per unit fresh mass; %) was calculated using these measures. The stem density (SD; g/cm3) were calculated as the ratio of wood dry mass to fresh volume that was determined using the water displacement method. The specific leaf area (SLA; leaf area/unit dry leaf mass; m2/kg) was also calculated using these mass and area measurements. Stable carbon isotope composition (δC13; ‰) was measured with an isotope ratio mass spectrometer (DELTA V Advantage, Thermo Fisher Scientific，Inc., Bremen). Osmotic potential (πosm; μmol/L) was measured with a VAPRO 5560 vapour pressure osmometer (Wescor, Logan, UT), and these data were used to calculate leaf turgor loss point (πtlp; μmol/L) following the methods proposed by Bartlett et al. (2012a). Carbon (C) and nitrogen (N) concentration (mg/g) was measured with a Dumas-type combustion C-N elemental analyzer (Vario MAX CN, ElementarAnalysensysteme GmbH, Hanau, Germany). From those data we also calculated the carbon to nitrogen ratio (CN). We found collinearity between SDMC and SD. So, we only included SDMC in the following data analysis (Figure S3). All trait data were averaged at the species-level for further analyses and the trait collection followed standardized methodologies (Poorter & Markesteijn 2008).