Aims New, powerful statistical techniques and GIS tools have resulted in a plethora of methods for modelling species distribution. However, little is known about the relative performance of different models in simulating and projecting species’ distributions under future climate. Our objective is to compare novel ensemble learning models with other conventional models by modelling the potential distribution of Masson pine (Pinus massoniana) and identifying and quantifying differences in model outputs.
Methods We simulated Masson pine potential distribution (baseline 1961-1990) and projected future potential distributions for three time periods (2010-2039, 2040-2069 and 2070-2099) using three global circulation models (GCM) (MIROC32_medres, JP; CCCMA_CGCM3, CA, and BCCR-BCM2.0, NW), one pessimistic SRES emissions scenario (A2), three ensemble learning models (random forest, RF; generalized boosted model, GBM, and NeuralEnsembles) and three conventional models (generalized linear model, GLM; generalized additive models, GAM, and classification and regression model, CART). The environmental envelope method was used to select absence of species. The area under the curve (AUC) values of receiver operator characteristic (ROC) curve, Kappa and true skill statistic (TSS) were used to objectively assess the predictive accuracy of each model. National standards for seed zone of Masson pine (GB 8822.6-1988) was employed to intuitively assess model performance. We developed ClimateChina software to downscale current and future GCM climate data and calculate seasonal and annual climate variables for specific locations based on latitude, longitude and elevation.
Important findings Ensemble learning models (GBM, NeuralEnsembles and RF) achieved a higher predictive success in simulating the distribution of Masson pine compared to other conventional models (CART, GAM and GLM). RF had the highest predictive accuracy, and CART had the lowest. Masson pine shows a globally consistent pattern in response to climate change for the three GCMs and six models, i.e., Masson pine will likely gradually shift northward and expand its distribution under altered future climate, with the magnitude of range changes dependent on model classes and GCMs. RF predicts a greater magnitude of range changes than other models. Projections of Masson pine distribution by NW are more conservative than JP and CA climate scenarios. In the case of Masson pine, range changes are mainly attributed to the colonization of newly available suitable habitat in high-latitudes and unchanging habitat suitability in the south-central part of its baseline range. Differences among 18 projections (6 models × 3 GCMs) increase with increasing time, and the greatest spatial uncertainty in projections is mainly in the north and west borders of the potential distribution range.

Aims Roots play an important role in the acquisition of soil resources and ecosystem processes. In this study, roots of nine subtropical tree species were studied to: 1) examine root diameter, specific root length and tissue N concentrations of different root branch orders in various tree species; 2) identify seasonal patterns of root N concentration; and 3) test the hypothesis that roots of lower orders have more marked seasonal patterns of N as they belong to a foundational module that is functionally different from higher order roots.
Methods Nine tree species in Dinghushan, Guangdong were studied. Roots were sampled from at least three trees of each species in the same plot at six different times during one year, and individual roots were separated according to branch orders. We measured root diameter, specific root length and tissue N concentrations of each order of all nine species. Variations of root N concentrations among orders in each species were analyzed by one-way ANOVA.
Important findings There were significant differences in root N concentration among different branch orders, with first and second orders having consistently higher N concentrations than higher orders, as expected. However, contrary to our hypothesis, the first two orders, which should belong to the rapid-cycling fraction of tree root systems, did not demonstrate the most marked seasonal changes in N concentration. Thus, the marked seasonal patterns in N concentrations in lower order roots found in temperate trees may not occur in subtropical trees. We suggest this contrast between temperate and subtropical trees may be related to N storage patterns in roots in temperate trees due to distinct phenology and the lack of distinct seasonality in subtropical trees. More study is needed to identify specific mechanisms that regulate N concentrations in roots and other plant tissues in subtropical vs. temperate trees.

Aims Climate change characterized by global warming has posed a great threat to terrestrial ecosystems and sustainable development of human society. Our objective was to examine the response of major vegetation types to climate change.
Methods We used the biweekly dataset of the National Oceanographic and Atmospheric Administration/Advanced Very High Resolution Radiometer (NOAA/AVHRR) normalized difference vegetation index (NDVI) and climatic data from 752 standard stations in China for 1982-2006 to study the responses of 12 major vegetation types to changes in air temperature and precipitation.
Important findings The NDVI trend was controlled by air temperature and precipitation at the transect scale, while the NDVI trend showed large spatial heterogeneity, possibly associated with changes in regional climate, land use and vegetation type. At the biome scale, annual NDVIof temperate deciduous shrubland (TDS) was significantly correlated with air temperature and that of temperate grass steppe (TGS) and subtropical and tropical coniferous forest (STCF) was significantly correlated with both air temperature and precipitation. No significant relationships were detected between NDVI and climatic factors in other types of vegetation. NDVI was most significantly correlated to the air temperature of the preceding four months. In addition, the time lag of NDVI responses to air temperature gradually decreased from January to April. Negative correlations were found between NDVI and air temperature from May to August in temperate coniferous forest (TCF), temperate deciduous broad-leaved forest (TDBF), TDS, STCF and subtropical and tropical grassland (STG). Results indicated a positive relationship between NDVIand precipitation of the same month in shrub and grassland types and a negative relationship between NDVIand precipitation of the same month in forest.

Aims Mass loss in wintertime is one of the key processes in litter decomposition in cold biomes. The contribution of soil fauna to litter decomposition has been unclear, and the contribution might be different in different winter periods (OF: the onset of the freeze-thaw season, DF: the deep frozen period and TS: the thawing stage). The objective was therefore to quantify the contributions of soil fauna to fir (Abies faxoniana) leaf litter decomposition during the freeze-thaw season.
Methods A field experiment using litterbags was conducted in an alpine fir forest from October 2010 to April 2011. Samples of air-dried fir leaf litter were placed in nylon litterbags (20 cm × 20 cm, 10 g per bag), and the edges were sealed. We investigated mass loss rates in litterbags with different mesh sizes (0.020, 0.125, 1.000 and 3.000 mm) and simultaneously analyzed contributions of micro-, meso- and macro-fauna to leaf litter decomposition.
Important findings Over the freeze-thaw season, the mass loss rates in litterbags with different mesh sizes were 12.13% (0.020 mm), 13.07% (0.125 mm), 14.95% (1.000 mm) and 18.74% (3.000 mm). Contribution percentages of all three body-size soil faunas were about 35.28%. Mass loss rates of fir leaf litter appeared consistent with the taxa and individual relative density of soil fauna. Contribution percentages of mass loss rates showed the order as micro- < meso- < macro-fauna during the three stages of the freeze-thaw season. The highest contribution percentages of micro-fauna (6.56%), meso-fauna (11.77%) and macro-fauna (21.94%) were detected at DF, OF and TS stage, respectively. However, the contribution percentages of meso- and macro-fauna during the DF stage were lower than the other two stages. The results demonstrated that soil fauna played an important role in litter decomposition during the freeze-thaw season in alpine forests of western China.

Aims Although much attention has been paid for the relationship between the biodiversity and ecosystem functioning, little is known about the ecosystem consequences of changes. In addition, the question whether species diversity (SD) can co-vary with functional diversity (FD) is unanswered. We used the Rao index to quantify FD in order to explore this question through comparison of different components of biodiversity and their effects on primary productivity (PP).
Methods A field manipulation experiment was conducted in alpine meadow at the Haibei Research Station of the Chinese Academy of Sciences from 2007 to 2010. The experiment used a split-plot design with clipping treatment in the whole plot using three clipping levels (stubbled 1 cm, 3 cm and unclipped). Subplots were treated with fertilizer (urea 7.5 g·m^-2·a^-1+ ammonium phosphate 1.8 g·m^-2·a^-1and unfertilized) and watering (20.1 kg·m^-2·a^-1and unwatered). General linear model univariate ANOVA, correlation analysis and regression analysis were used to analyze the effects of the treatments on plant functional traits, biodiversity and PP, as well as the relationship among them.
Important findings The clipping and fertilizing treatments had important effects on plant traits and PP. There were no significant differences in SD among the clipping treatments, whereas the FD quantified by distinct traits showed different trends across clipping treatments. Fertilization increased PP but decreased both SD and FD(except FD_H). Watering affected weakly only the Richness index (R). Correlation analysis indicated that the FD_{6 traits}reflected functional divergence better than the FDcalculated by the single trait. Regression analysis showed that there was a significantly positive and no correlation between SD and FD, respectively, and the relationship among SD, FD and PP followed three patterns: significantly positive, negative or no correlation, i.e., the relation forms were not only trait-dependent but also were related to disturbance and soil resources. Plant functional traits have a more direct influence on PP.

Aims Our objective is to study changes of phytoliths in Leymus chinensis quantitatively under saline-alkali stress to assist reconstruction of saline paleoenvironments.
Methods We selected similarly sized leaves ofL. chinensis under six different saline-alkali levels and extracted phytoliths with the wet ashing method.
Important findings We identified and named 10 categories and 3 387 phytoliths: flat-topped rondel, peaked rondel, barbed rondel, smooth elongate, sinus elongate, horn-ends elongate, pointed, plate, favose and other types. With increased salt concentration, rondel, elongate, pointed, plate and favose changed. The percentage of rondel increased in all groups. The same type of phytolith had differences in morphology and percentage among groups. All types of phytoliths increased in size with saline-alkali stress.

Aims In grassland ecosystems, decomposition of litter is an important process in nutrient cycling and soil organic matter formation. Our objectives are to explores (1) seasonal dynamics of decomposition of root tissues and shoot litter of Stipa baicalensis, a dominant species in a typical meadow steppe in Hulun Buir, Inner Mongolia, China, and (2) the effects on decomposition of placing litterbags on the soil surface vs. in the 15-cm soil layer.
Methods We placed litterbags with root tissues and shoot litter of S. baicalensis on the soil surface and in the 15-cm soil layer in May 2010. We collected the litterbags after 1, 2, 3, 5, and 12 months and determined mass loss and carbon and nitrogen content in the remaining litter. Microbial carbon and nitrogen contents were determined in the laboratory.
Important findings The mass loss rates of root and shoot litter were smaller on the soil surface than in the 15-cm soil layer, but the mass loss rates of root litter were not significantly different at the two positions. Irrespective of decomposition position, shoot litter decomposed faster than root tissues. Seasonal variation of carbon release showed a similar pattern with mass loss during decomposition. But seasonal variation patterns for nitrogen release were significantly different between shoot litter and root tissues. Nitrogen release of shoot litter exhibited a pattern of release―increase―release and that of root a pattern of release―increase. Nitrogen content difference in shoot or root litter was relatively small at both decomposition positions. Initial chemical compositions and soil moisture played dominant roles in seasonal dynamics of decomposition of root tissues and shoot litter. Findings suggest that future changes of temperature and soil moisture will markedly influence shoot decomposition, but not root decomposition in this meadow steppe.

Aims Spatial pattern of vegetation is determined by a variety of environmental factors on different scales. The influence of landform on vegetation pattern is receiving growing attention. This study examined the influence of slope location on vegetation pattern in Inner Mongolia grassland to provide a scientific basis for vegetation classification, vegetation mapping, rational utilization and ecosystem restoration and reconstruction.Methods In each of three typical sites in Stipa breviflora grassland, we established three parallel line transects of unequal length and placed 0.5 m × 0.5 m herb plots at 2-m intervals along each line. Vegetation and soil were surveyed for each plot. Cluster analysis, binomial test, significant difference test, canonical correspondence analysis and Pearson correlation analysis were used to analyze the influence of slope location on vegetation pattern and soil effects.
Important findings Vegetation types were clearly differentiated between upper and lower slopes, with a climatic climax community on the upper slope and a terrain community on the lower slope. Slope locations affected the species distribution pattern along the slope, mainly in species composition and some dominant species. The productivity of the lower slope is higher than that of the higher slope, and two different functional areas had formed. The character of the surface soil (especially 0-5 cm) was largest difference between upper and lower slope. Various geomorphic processes resulted in great habitat heterogeneity in a very small region. It provides an important mechanism for formation and maintenance of biological diversity at the local scale. Mechanical composition difference of surface soil (especially 0-5 cm) affected vegetation pattern, and soil moisture affected the productivity difference along the slope.

Aims Our objective was to investigate the soil organic carbon stock, fine root turnover and its contribution to soil carbon cycling of Reaumuria soongorica community in an arid ecosystem in Xinjiang Uygur Autonomous Region, China.
Methods We selected a R. soongorica community, a typical local community, in natural areas around the Ecological Station of Xinjiang Institute of Ecology and Geography, Chinese Academy of Sciences. We investigated the distribution and dynamics of soil organic carbon, monthly fine root dynamics, production and turnover rate throughout the 2010 growing season by means of sequential coring and ingrowth core. We calculated the soil organic carbon pool, fine roots carbon stocks and contribution of fine roots.
Important findings Soil water content increased with depth, but soil organic carbon decreased. The mean fine root mass in the community was 54.51 g·m^-2, and annual fine root production was estimated to be 82.76-136.21 g·m^-2·a^-1. Fine root turnover rate in the community was 2.08 times·a^-1, and annual inputs from fine root mortality to belowground soil organic carbon was 17 g·m^-2·a^-1. These results demonstrate that due to the high turnover rate of shrub fine root in the arid region, shrub fine root carbon is a crucial portion of soil organic carbon inputs.

Aims Precipitation and temperature are considered the limiting factors to vegetation growth in arid and cold areas, respectively. Our objective was to study the trends in vegetation cover and productivity and their relationships with prevailing hydrothermal factors in order to better understand climatic constraints on regional vegetation dynamics.
Methods We zoned dry and wet climates using daily meteorological data from 1959 to 2008 at 16 locations in the headwaters of the Yellow River Basin to calculate the ratio of precipitation to evapotranspiration estimated with the FAO Penman-Monteith model. Based on this zoning, the temporal dynamics of vegetation and changes in climate were analyzed using the normalized difference vegetation index (NDVI) obtained from the NASA AVHRR sensor (1982-2006) and the GLOPEM net primary productivity (NPP) (1981-2000) and corresponding annual meteorological data. Then we developed the relationships between NDVI and NPP and climatic factors.
Important findings The region studied is semi-humid in the southeast and semi-arid elsewhere, with a boundary nearly matching the 450 mm rainfall contour. The weather became wetter and especially warmer in most parts of the region during 1981-2006. The regional average NDVI and NPP markedly increased in semi-humid areas and slightly increased in semi-arid areas. Overall, NDVI had a significantly positive correlation with mean annual air temperature and a mainly negative correlation with mean annual precipitation in most semi-humid areas. The most important factor influencing NDVI was heat. In contrast, the key climatic factor affecting vegetation changes in the semi-arid areas was water, the relationship between NDVI and precipitation was strong and vegetation was more sensitive to changes in precipitation. The impact of climate change on NPP was similar to that on NDVI. Findings suggest that the responses of vegetation to climate change depend in part on the hydrothermal conditions of the region, and the zoning of dry and wet climate illustrates the spatial difference of vegetation feedback to climate change.

Aims Bulbophyllum ambrosia is an epiphytic orchid growing on rocks and trees in limestone seasonal rainforests and is commonly distributed in southern China. We found no fruit set of B. ambrosia during long-term monitoring of species diversity, reproductive phenology and pollinators of local orchid species in Green Stone Forest Park. Our objective was to determine barriers to fruit set in B. ambrosia.
Methods Flowering phenology and natural fruit-set were monitored once a week for three years. Floral morphology was studied during flowering time. We also investigated the rate of pollinia removal and deposition over two flowering seasons. To determine the self-compatibility system, different hand-pollination treatments were conducted over three years at the study site. We also observed floral visitors for 150 h over two years.
Important findings Unlike other myophilous species in the genus, B. ambrosia is fragrant and presents nectar as reward. Observations of floral visitors showed that honeybee (Apis cerana cerana) was the only efficient pollinator and regularly visited flowers with high frequency. In natural populations, the rates of pollinia removal and deposition were high, but there was no fruit set. The hinged labellum plays an important role in the process of honeybee pollination of B. ambrosia. No fruit was found in the selfing pollination treatment, and fruit set of crossing pollination treatments was >90% over two years, suggesting that B. ambrosia is self-incompatible. To our knowledge, this is the first report of a self-incompatible species in Bulbophyllum. Individuals of B. ambrosia have a mass of flowers opening simultaneously during anthesis because of its strong ability in cloning. Honeybees visited 4.29 ± 0.40 flowers ( n = 66) in average per visitation, which may lead to geitonogamy. Self-incompatibility causes no fruit set as the result of geitonogamy. These results provide useful implications for future conservation of B. ambrosia.