Mapping Chinese annual gross primary productivity with eddy covariance measurements and machine learning.

Annual gross primary productivity (AGPP) is the basis for grain production and terrestrial carbon sequestration. Mapping regional AGPP from site measurements provides methodological support for analysing AGPP spatiotemporal variations thereby ensures regional food security and mitigates climate change. Based on 641 site-year eddy covariance measuring AGPP from China, we built an AGPP mapping scheme based on its formation and selected the optimal mapping way, which was conducted through analysing the predicting performances of divergent mapping tools, variable combinations, and mapping approaches in predicting observed AGPP variations. The reasonability of the selected optimal scheme was confirmed by assessing the consistency between its generating AGPP and previous products in spatiotemporal variations and total amount. Random forest regression tree explained 85 % of observed AGPP variations, outperforming other machine learning algorithms and classical statistical methods. Variable combinations containing climate, soil, and biological factors showed superior performance to other variable combinations. Mapping AGPP through predicting AGPP per leaf area (PAGPP) explained 86 % of AGPP variations, which was superior to other approaches. The optimal scheme was thus using a random forest regression tree, combining climate, soil, and biological variables, and predicting PAGPP. The optimal scheme generating AGPP of Chinese terrestrial ecosystems decreased from southeast to northwest, which was highly consistent with previous products. The interannual trend and interannual variation of our generating AGPP showed a decreasing trend from east to west and from southeast to northwest, respectively, which was consistent with data-oriented products. The mean total amount of generated AGPP was 7.03 ± 0.45 PgC yr-1 falling into the range of previous works. Considering the consistency between the generated AGPP and previous products, our optimal mapping way was suitable for mapping AGPP from site measurements. Our results provided a methodological support for mapping regional AGPP and other fluxes.

Adaptive genetic diversity of dominant species contributes to species co-existence and community assembly.

The synthesis of evolutionary biology and community ecology aims to understand how genetic variation within one species can shape community properties and how the ecological properties of a community can drive the evolution of a species. A rarely explored aspect is whether the interaction of genetic variation and community properties depends on the species' ecological role. Here we investigated the interactions among environmental factors, species diversity, and the within-species genetic diversity of species with different ecological roles. Using high-throughput DNA sequencing, we genotyped a canopy-dominant tree species, Parashorea chinensis, and an understory-abundant species, Pittosporopsis kerrii, from fifteen plots in Xishuangbanna tropical seasonal rainforest and estimated their adaptive, neutral and total genetic diversity; we also surveyed species diversity and assayed key soil nutrients. Structural equation modelling revealed that soil nitrogen availability created an opposing effect in species diversity and adaptive genetic diversity of the canopy-dominant Pa. chinensis. The increased adaptive genetic diversity of Pa. chinensis led to greater species diversity by promoting co-existence. Increased species diversity reduced the adaptive genetic diversity of the dominant understory species, Pi. kerrii, which was promoted by the adaptive genetic diversity of the canopy-dominant Pa. chinensis. However, such relationships were absent when neutral genetic diversity or total genetic diversity were used in the model. Our results demonstrated the important ecological interaction between adaptive genetic diversity and species diversity, but the pattern of the interaction depends on the identity of the species. Our results highlight the significant ecological role of dominant species in competitive interactions and regulation of community structure.

Traits, strategies, and niches of liana species in a tropical seasonal rainforest.

Plant functional traits and strategies hold the promise to explain species distribution, but few studies have linked multiple traits to multiple niche dimensions (i.e., light, water, and nutrients). Here, we analyzed for 29 liana species in a Chinese tropical seasonal rainforest how: (1) trait associations and trade-offs lead to different plant strategies; and (2) how these traits shape species' niche dimensions. Eighteen functional traits related to light, water, and nutrient use were measured and species niche dimensions were quantified using species distribution in a 20-ha plot combined with data on canopy gaps, topographic water availability, and soil nutrients. We found a tissue toughness spectrum ranging from soft to hard tissues along which species also varied from acquisitive to conservative water use, and a resource acquisition spectrum ranging from low to high light capture and nutrient use. Intriguingly, each spectrum partly reflected the conservative-acquisitive paradigm, but at the same time, the tissue toughness and the resource acquisition spectrum were uncoupled. Resource niche dimensions were better predicted by individual traits than by multivariate plant strategies. This suggests that trait components that underlie multivariate strategy axes, rather than the plant strategies themselves determine species distributions. Different traits were important for different niche dimensions. In conclusion, plant functional traits and strategies can indeed explain species distributions, but not in a simple and straight forward way. Although the identification of global plant strategies has significantly advanced the field, this research shows that global, multivariate generalizations are difficult to translate to local conditions, as different components of these strategies are important under different local conditions.

Environmental and management controls of soil carbon storage in grasslands of southwestern China.

In order to predict the effects of climate change on the global carbon cycle, it is crucial to understand the environmental factors that affect soil carbon storage in grasslands. In the present study, we attempted to explain the relationships between the distribution of soil carbon storage with climate, soil types, soil properties and topographical factors across different types of grasslands with different grazing regimes. We measured soil organic carbon in 92 locations at different soil depth increments, from 0 to 100 cm in southwestern China. Among soil types, brown earth soils (Luvisols) had the highest carbon storage with 19.5 ±â€¯2.5 kg m-2, while chernozem soils had the lowest with 6.8 ±â€¯1.2 kg m-2. Mean annual temperature and precipitation, exerted a significant, but, contrasting effects on soil carbon storage. Soil carbon storage increased as mean annual temperature decreased and as mean annual precipitation increased. Across different grassland types, the mean carbon storage for the top 100 cm varied from 7.6 ±â€¯1.3 kg m-2 for temperate desert to 17.3 ±â€¯2.9 kg m-2 for alpine meadow. Grazing/cutting regimes significantly affected soil carbon storage with lowest value (7.9 ±â€¯1.5 kg m-2) recorded for cutting grass, while seasonal (11.4 ±â€¯1.3 kg m-2) and year-long (12.2 ±â€¯1.9 kg m-2) grazing increased carbon storage. The highest carbon storage was found in the completely ungrazed areas (16.7 ±â€¯2.9 kg m-2). Climatic factors, along with soil types and topographical factors, controlled soil carbon density along a soil depth in grasslands. Environmental factors alone explained about 60% of the total variation in soil carbon storage. The actual depth-wise distribution of soil carbon contents was significantly influenced by the grazing intensity and topographical factors. Overall, policy-makers should focus on reducing the grazing intensity and land conversion for the sustainable management of grasslands and C sequestration.

Comparison of infrared canopy temperature in a rubber plantation and tropical rain forest.

Canopy temperature is a result of the canopy energy balance and is driven by climate conditions, plant architecture, and plant-controlled transpiration. Here, we evaluated canopy temperature in a rubber plantation (RP) and tropical rainforest (TR) in Xishuangbanna, southwestern China. An infrared temperature sensor was installed at each site to measure canopy temperature. In the dry season, the maximum differences (Tc - Ta) between canopy temperature (Tc) and air temperature (Ta) in the RP and TR were 2.6 and 0.1 K, respectively. In the rainy season, the maximum (Tc - Ta) values in the RP and TR were 1.0 and -1.1 K, respectively. There were consistent differences between the two forests, with the RP having higher (Tc - Ta) than the TR throughout the entire year. Infrared measurements of Tc can be used to calculate canopy stomatal conductance in both forests. The difference in (Tc - Ta) at three gc levels with increasing direct radiation in the RP was larger than in the TR, indicating that change in (Tc - Ta) in the RP was relatively sensitive to the degree of stomatal closure.

Water use efficiency in a primary subtropical evergreen forest in Southwest China.

We calculated water use efficiency (WUE) using measures of gross primary production (GPP) and evapotranspiration (ET) from five years of continuous eddy covariance measurements (2009-2013) obtained over a primary subtropical evergreen broadleaved forest in southwestern China. Annual mean WUE exhibited a decreasing trend from 2009 to 2013, varying from ~2.28 to 2.68 g C kg H2O-1. The multiyear average WUE was 2.48 ± 0.17 (mean ± standard deviation) g C kg H2O-1. WUE increased greatly in the driest year (2009), due to a larger decline in ET than in GPP. At the diurnal scale, WUE in the wet season reached 5.1 g C kg H2O-1 in the early morning and 4.6 g C kg H2O-1 in the evening. WUE in the dry season reached 3.1 g C kg H2O-1 in the early morning and 2.7 g C kg H2O-1 in the evening. During the leaf emergence stage, the variation of WUE could be suitably explained by water-related variables (relative humidity (RH), soil water content at 100 cm (SWC_100)), solar radiation and the green index (Sgreen). These results revealed large variation in WUE at different time scales, highlighting the importance of individual site characteristics.

The effects of nitrogen fertilization on N2O emissions from a rubber plantation.

To gain the effects of N fertilizer applications on N2O emissions and local climate change in fertilized rubber (Hevea brasiliensis) plantations in the tropics, we measured N2O fluxes from fertilized (75 kg N ha(-1) yr(-1)) and unfertilized rubber plantations at Xishuangbanna in southwest China over a 2-year period. The N2O emissions from the fertilized and unfertilized plots were 4.0 and 2.5 kg N ha(-1) yr(-1), respectively, and the N2O emission factor was 1.96%. Soil moisture, soil temperature, and the area weighted mean ammoniacal nitrogen (NH4(+)-N) content controlled the variations in N2O flux from the fertilized and unfertilized rubber plantations. NH4(+)-N did not influence temporal changes in N2O emissions from the trench, slope, or terrace plots, but controlled spatial variations in N2O emissions among the treatments. On a unit area basis, the 100-year carbon dioxide equivalence of the fertilized rubber plantation N2O offsets 5.8% and 31.5% of carbon sink of the rubber plantation and local tropical rainforest, respectively. When entire land area in Xishuangbanna is considered, N2O emissions from fertilized rubber plantations offset 17.1% of the tropical rainforest's carbon sink. The results show that if tropical rainforests are converted to fertilized rubber plantations, regional N2O emissions may enhance local climate warming.

Size-class effect contributes to tree species assembly through influencing dispersal in tropical forests.

We have investigated the processes of community assembly using size classes of trees. Specifically our work examined (1) whether point process models incorporating an effect of size-class produce more realistic summary outcomes than do models without this effect; (2) which of three selected models incorporating, respectively environmental effects, dispersal and the joint-effect of both of these, is most useful in explaining species-area relationships (SARs) and point dispersion patterns. For this evaluation we used tree species data from the 50-ha forest dynamics plot in Barro Colorado Island, Panama and the comparable 20 ha plot at Bubeng, Southwest China. Our results demonstrated that incorporating an size-class effect dramatically improved the SAR estimation at both the plots when the dispersal only model was used. The joint effect model produced similar improvement but only for the 50-ha plot in Panama. The point patterns results were not improved by incorporation of size-class effects using any of the three models. Our results indicate that dispersal is likely to be a key process determining both SARs and point patterns. The environment-only model and joint-effects model were effective at the species level and the community level, respectively. We conclude that it is critical to use multiple summary characteristics when modelling spatial patterns at the species and community levels if a comprehensive understanding of the ecological processes that shape species' distributions is sought; without this results may have inherent biases. By influencing dispersal, the effect of size-class contributes to species assembly and enhances our understanding of species coexistence.

The role of stream water carbon dynamics and export in the carbon balance of a tropical seasonal rainforest, southwest China.

A two-year study (2009 ~ 2010) was carried out to investigate the dynamics of different carbon (C) forms, and the role of stream export in the C balance of a 23.4-ha headwater catchment in a tropical seasonal rainforest at Xishuangbanna (XSBN), southwest China. The seasonal volumetric weighted mean (VWM) concentrations of total inorganic C (TIC) and dissolved inorganic C (DIC) were higher, and particulate inorganic C (PIC) and organic C (POC) were lower, in the dry season than the rainy season, while the VWM concentrations of total organic C (TOC) and dissolved organic C (DOC) were similar between seasons. With increased monthly stream discharge and stream water temperature (SWT), only TIC and DIC concentrations decreased significantly. The most important C form in stream export was DIC, accounting for 51.8% of the total C (TC) export; DOC, POC, and PIC accounted for 21.8%, 14.9%, and 11.5% of the TC export, respectively. Dynamics of C flux were closely related to stream discharge, with the greatest export during the rainy season. C export in the headwater stream was 47.1 kg C ha(-1) yr(-1), about 2.85% of the annual net ecosystem exchange. This finding indicates that stream export represented a minor contribution to the C balance in this tropical seasonal rainforest.

Strong neutral spatial effects shape tree species distributions across life stages at multiple scales.

Traditionally, ecologists use lattice (regional summary) count data to simulate tree species distributions to explore species coexistence. However, no previous study has explicitly compared the difference between using lattice count and basal area data and analyzed species distributions at both individual species and community levels while simultaneously considering the combined scenarios of life stage and scale. In this study, we hypothesized that basal area data are more closely related to environmental variables than are count data because of strong environmental filtering effects. We also address the contribution of niche and the neutral (i.e., solely dependent on distance) factors to species distributions. Specifically, we separately modeled count data and basal area data while considering life stage and scale effects at the two levels with simultaneous autoregressive models and variation partitioning. A principal coordinates of neighbor matrix (PCNM) was used to model neutral spatial effects at the community level. The explained variations of species distribution data did not differ significantly between the two types of data at either the individual species level or the community level, indicating that the two types of data can be used nearly identically to model species distributions. Neutral spatial effects represented by spatial autoregressive parameters and the PCNM eigenfunctions drove species distributions on multiple scales, different life stages and individual species and community levels in this plot. We concluded that strong neutral spatial effects are the principal mechanisms underlying the species distributions and thus shape biodiversity spatial patterns.

[Seasonal variation of soil respiration and its components in tropical rain forest and rubber plantation in Xishuangbanna, Yunnan].

By using trenching method and infrared gas analyzer, this paper studied the seasonal variation of soil respiration (SR), including root respiration (RR) and heterotrophic respiration (HR), in tropical seasonal rain forest (RF) and rubber (Hevea brasiliensis) plantation (RP) in Xishuangbanna of Yunnan, China. The results showed that the SR and HR rates were significantly higher in RF than in RP (P < 0.01), while the RR rate had less difference between the two forests. Soil temperature and moisture were the key factors affecting the SR, RR and HR. The SR and HR rates in the two forests were rainy season > dry-hot season > foggy season, but the RR rate was rainy season > foggy season > dry-hot season in RF, and foggy season > rainy season > dry-hot season in RP. The contribution of RR to SR in RF (29%) was much lower than that in RP (42%, P < 0.01), while the contribution of HR to SR was 71% in RF and 58% in RP. When the soil temperature at 5 cm depth varied from 12 degrees C to 32 degrees C, the Q10 values for SR, HR, and RR rates were higher in RF than in RP. HR had the highest Q10 value, while RR had the lowest one.

Vegetation cover of forest, shrub and pasture strongly influences soil bacterial community structure as revealed by 16S rRNA gene T-RFLP analysis.

Bacterial community structure is influenced by vegetation, climate and soil chemical properties. To evaluate these influences, terminal restriction fragment length polymorphism (T-RFLP) and cloning of the 16S rRNA gene were used to analyze the soil bacterial communities in different ecosystems in southwestern China. We compared (1) broad-leaved forest, shrub and pastures in a high-plateau region, (2) three broad-leaved forests representing a climate gradient from high-plateau temperate to subtropical and tropical regions and (3) the humus and mineral soil layers of forests, shrub lands and pastures with open and restricted grazing activities, having varied soil carbon and nutrient contents. Principal component analysis of the T-RFLP patterns revealed that soil bacterial communities of the three vegetation types were distinct. The broad-leaved forests in different climates clustered together, and relatively minor differences were observed between the soil layers or the grazing regimes. Acidobacteria dominated the broad-leaved forests (comprising 62% of the total clone sequences), but exhibited lower relative abundances in the soils of shrub (31%) and pasture (23%). Betaproteobacteria was another dominant taxa of shrub land (31%), whereas Alpha- (19%) and Gammaproteobacteria (13%) and Bacteriodetes (16%) were major components of pasture. Vegetation exerted more pronounced influences than climate and soil chemical properties.