The spatial pattern of climate factors has been shown to influence the patterns of carbon fluxes. However, mechanisms underlying this climate control are not well established. We synthesize global carbon flux data to propose new biogeographic mechanisms for the climate control. The mechanisms include three linked core controls: (1) “climate-constrained community photosynthetic capacity mechanism” that spatial patterns of temperature and precipitation define growing season length and mean carbon uptake capacity to control the annual production; (2) “climate-constrained dual substrates respiration mechanism” that spatial patterns of production and stored organic matters control the annual respiration; (3) “climate and disturbance-constrained ecosystem carbon allocation mechanism” that the spatial patterns of production and disturbances impactedproduction consumption efficiency control the annual net production. Our results clarify how the BioGeographic Reactor (BGR), which links patterns of climate with carbon fluxes operate, and that provide new insight into the terrestrial carbon budget and its controls.

On a global or regional scale, low-latitude (tropical and subtropical areas) ecosystem play important roles in exchanges between the biosphere and the atmospheric surface layer, but few studies have focused on these ecosystems to investigate how they are influenced by the shifts of environmental variables. In order to quantify how the environmental controls influence the biosphere-atmosphere interactions in the ecosystem scale in such unique environmental region, we analyzed the long-term eddy-covariance data along with the environmental variables collected from two typical land-cover types in Northern Taiwan. The first one, Guandu site (Asiaflux code: GDP), is a marsh grassland ecosystem located in a flat floodplain in the northwest of Taipei Basin. The second ecosystem, Chi-Lan Mountain site (Asiaflux code: CLM), is an evergreen coniferous forest at a mountainous terrain at the altitude of 1600m in the north section of the central mountain ridge in Northwestern Taiwan. The distance between these two study sites is only about 55 km, however these two ecosystem experience different but very representative climate and environment in the subtropical region of East Asia. 

In this study, we used the eddy-covariance data to analyze the temporal variation in exchanges of scalars and energy components, and further determined the effects of environmental factors on the dynamics of the budgets in these two different ecosystems. We aimed to (1) quantify the scalar and energy budgets by examining eddy-covariance data in terms of environmental variables; (2) analyze the effect of environmental factors on the abovementioned budgets; and (3) interpret possible shifts and trends in biosphere-atmosphere interactions in these two representative subtropical ecosystem under the influence of environmental changes in the future.

Our preliminary results showed that the different patterns do exist in these two subtropical ecosystem when compared with the data with middle to high latitude region in East Asia. This diversity led to different potential shifts and trends of biomass accumulation and distribution of these two typical low-latitude vegetation types in this region under different seniors of environmental changes in the future.

Terrestrial ecosystems in Asia cover large land area and represent various biomes including tundra, boreal, temperate, subtropical and tropical forests, wetlands, grasslands, crop fields, and the world largest rice paddies extending from the arctic circle to equator. These biomes are significantly affected by the Monsoon Asia climate as well as under various pressures such as land-use and climate changes. Knowledge of carbon dioxide (CO₂), methane (CH₄) and other greenhouse gases (e.g. N₂O) budgets of the terrestrial ecosystems in Asian region is essential to advance our understanding of global carbon cycle and prediction of the impacts of climate change. Since the mid-1990s, we have been installing multichannel automated chamber systems at tundra in the West Siberian lowland, boreal forests in central Alaska, cool-temperate and temperate forests in Japan, Korea and China, subtropical forests in Japan, Mainland China and Taiwan, tropical seasonal forests in China and Thailand, tropical rainforests in China and Malaysia, and even arid grassland in Inner-Mongolia and wetland on the Tibetan Plateau, for continuous measurements of forest floor CO₂ budget as well as NEP. Among the sites, eight of the systems are using for conducting soil warming experiments. Currently, the chamber network is expanding rapidly in the Asian region. Our ultimate objective is to estimate the carbon budget of Asian terrestrial ecosystems as well as its response and feedback to regional climate change.

In recent years, with the rapid technical development of new sensors (e.g. CH₄ and N₂O analyzers), the chamber network is potentially applied for simultaneously measurement of major target GHGs (e.g. CO₂, CH₄, N₂O) budget together. This talk will present CO₂/CH₄ fluxes and their controls of representative Asian terrestrial ecosystems by using multichannel automated chamber systems.

Forest ecosystem is the major stock of biomass in terrestrial ecosystems. Elucidating the carbon cycle mechanism in forest ecosystem is vital for understanding global carbon cycle and predicting future carbon budget along with global climate change. Understory carbon budget is very important component of forest carbon cycle, but the detailed information about the processes and control factors are limited. Especially, influence of forest management (tree thinning) on understory carbon budget is not known well.

Multi-channel automated chamber measurement system was set in larch forest on northern foot of Mt. Fuji in 2006. The control unit of chamber system mainly consisted of a data logger (CR1000, Campbell Scientific), an infrared gas analyzer (LI820, LI-COR) and an air compressor. We set soil chambers (90 cm × 90 cm × 50 cm) for soil CO₂ flux measurement. Surroundings of the half of those soil chambers were root cut with chainsaw until 25 cm depth for the measurement of heterotrophic respiration (Rh), and the remaining control chambers were used for soil respiration (Rs) measurement. We also set plant chambers (90 cm × 90 cm × 100 cm) which included understory vegetation to measure understory net CO₂ exchange (NUE), understory respiration (Ru) and understory gross primary production (GPP_u). In 2014 and 2015, stepwise tree thinning was applied to this larch forest, and 30% of total larch trees were cut down in the end (March 2015).

Comparison of estimated understory carbon fluxes between before (the average of 2006-2013) and after 30% thinning (2015) suggested that understory photosynthetic photon flux density (PPFD) increased 64.5%, resulting 68.8% increase of GPP_u. On the other hand, Rs, Rh and Ru increased 10.8%, 10.2% and 21.8% mainly due to slight increase of soil temperature (7%). As a whole, NUE increased 5.8%, and did not change obviously.

The United Nations Food and Agriculture Organization pursues the climate smart agriculture (CSA) with a hope of triple wins: (1) sustainably increasing agricultural productivity and incomes; (2) adapting and building resilience to climate change; (3) reducing and/or removing greenhouse gases emissions, where possible. Using the time series of the long-term KoFlux data at Haenam farmland site (HFK) in AsiaFlux, a suite of biotic, network, and thermodynamic indicators are evaluated. These ecological indicators are then integrated to scrutinize this agricultural ecosystem in terms of the CSA’s triad strategies (i.e., efficiency, resilience, and carbon footprint).

Grassland occupies about 50% of the TP and acts as a carbon sink nowadays. Climate warming may increase the productivity of the grassland on the Plateau. It may also accelerate carbon releasing at the same time, especially when grassland degradation occurs. Since August 2001, intensive field observations of energy balance and CO2 flux has been induced on the TP at Haibei Alpine Meadow Ecosystem Research Station  (37°37'N, 101°19'E, 3250m a.s.l.) by a Japan–China cooperation project. Preliminary analyses of the first three years data suggest that (1) the Qinghai-Tibetan Plateau plays a potentially significant role in global carbon sequestration, because alpine meadow covers about one-third of this vast plateau, and (2) the annual NEP in the alpine meadow was comprehensively controlled by the temperature environment, including its effect on biomass growth (Kato et al., 2006).

Here, 14 years of eddy covariance measurements were used to characterize the statistical features of NEE and its relationship with temperature variations during the 14 years. We fitted the Michaelis–Menten equations of the light–response curve to the alpine meadow ecosystem getting the ecosystem-scale light-response curve and then get the relationship between the parameters of the light-response curve and temperature.

The parameter values obtained described the shape and amplitude of the temperature responses of the maximum gross photosynthesis (Amax) and the initial slope of the light response curve (a).

Predicted shifts in the magnitude and frequency of precipitation events (larger but fewer) and enhanced nitrogen (N) deposition in the future may interactively affect productivity of grasslands. In this study, we quantified the effects of N addition on the responsive patterns of gross primary productivity (GPP) to different sizes of individual precipitation events (Psize) in a temperate grassland, China. Results showed that the duration of GPP-response (τR) and the maximum absolute GPP-response (GPPmax) increased linearly with Psize, driving a corresponding increase in time-integrated amount of the GPP-response (GPPtotal) because variations of GPPtotal were largely explained by τR and GPPmax. Meanwhile, GPPtotal was greatly stimulated by N enrichment, and the N-induced GPP-stimulation increased with increasing Psize. For a given individual event, N enrichment rarely affected the duration of the GPP-response induced by the event, but significantly stimulated the maximum absolute GPP-response. GPPtotal in both the N-addition and control treatments increased linearly with Psize with similar Psize-intercepts (around 5 mm, indicating a similar lower threshold of Psize that begin to stimulate GPP-response) but steeper slope due to N-addition. The higher foliar N content might play an important role in the N-induced GPP-stimulation. We also analyzed the effects of varied sizes of rainfall events on the soil inorganic N content, by which GPP might be affected. Our work has important implications for the modeling community to obtain an advanced understanding of productivity-response of grassland ecosystems to altered precipitation regimes and increasing N deposition.

Commercial fishponds are widespread in the coastal regions of subtropical China, but little is known regarding their exchange of energy and greenhouse gases (GHGs) with the atmosphere. This study aims to quantify the dynamics of GHG (including carbon dioxide, CO₂ and methane, CH₄) as well as energy (sensible heat, H and latent heat, LE) fluxes quasi-continuously at the commercial fishponds of subtropical Hong Kong at the ecosystem scale using the eddy covariance technique. Our results showed that the mean daily concentrations of CO₂ and CH₄ in the atmosphere varied between 378 and 435 µmol mol^-1, and between 1.9 and 3.1 µmol mol^-1, respectively. The mean daily H at the fishponds ranged from 57 to 158 W m^-2, while that of LE fluctuated between 2.7 and 14.4 MJ m^-2 d^-1. The mean daily values of net ecosystem exchange of CO₂ (NEE) and CH₄ at the ponds exhibited annual ranges of 0.27-11.5 g CO₂ m^-2d^-1 and 31-303 mg CH₄ m^-2 d^-1. On average, the fishpond ecosystem behaved as a net CO₂ sink with negative NEE values only for few hour during the day, which could be attributed to the photosynthetic activities carried out by the phytoplankton and the green algae present in pond water. For the majority of time, the fishponds acted as a net source of CO₂ with a positive NEE, indicating the dominance of ecosystem respiration over primary production. They were also a considerable CH₄ source, with the methanogens in organic-rich pond soils possibly becoming a governing factor of CH₄ production. Many of the cultured fish species in these ponds such as carps and tilapia are bottom dwellers and could enhance the release of GHGs from soils to the water column and eventually to the atmosphere through bioturbation. Our findings suggest that the commercial fishponds in subtropical China could potentially exert considerable effects on the greenhouse gas balance and hence climate change in the region.

Dynamic changes in evapotranspiration (ET), energy fluxes, shoot elongation rate and stem diameter change of cotton and sweet corn were studied in the open field in relation to rapid changes in radiation caused by moving clouds. ET, energy fluxes were measured with the Bowen ratio method, using a signal averaging time of 5 min. Elongation rate and stem diameter change were monitored with position transducers using a signal averaging time of 1 min. Temperature of the canopy surface (T_s), photosynthetically active radiation (PAR) and wind velocity were also monitored simultaneously. Canopy conductance for water vapor (g_cw) was calculated with the Penman-Monteith equation. ET, sensible heat flux, and T_s were all found to respond to change in radiation within minutes or sooner. The results indicate that radiation controlled directly canopy gas exchanges, and indirectly affected shoot growth and stem diameter by its effects on transpiration rate. A novel finding was that g_cw also responded within minutes or sooner to fluctuations in radiation. The results are discussed in terms of the concept of soil-plant-atmosphere continuum, and interpreted in terms of dynamic interactions between transpiration and plant water status. 

The main purpose of this study was to investigate the litterfall dynamics and nutrient returns, litter layer and nutrient accumulations among four native tree species, i.e..Pongamia pinnata, Liquidambar formosana, Cinnamomum camphora and Bischofia javanica which were planted at the Wanlong Farm in Pingtung County. In order to understand the impact of litterfall on the nutrient internal cycling, this study also estimated the rate of decomposition of each species and the time required for the nutrients decomposition. The litterfall was collected once a month for 24 months (from May 2012 to April 2014). The experimental results showed that the annual litterfall was 5.03 and 5.09 Mg ha^-1 yr^-1 for P. pinnata stand in the first and second year respectively, 3.36 and 3.64 Mg ha^-1 yr^-1 for L. formosana, 5.24 and 5.30 Mg ha^-1 yr^-1 for C. camphora, and 5.70 and 4.86 Mg ha^-1 yr^-1 for B. javanica. The highest concentrations of N and Mg were found in P. pinnata and B. javanica respectively. Annual nutrient returns in litter roughly showed the decreasing content of C>N>Ca>K>Mg>P. The daily instantaneous decay rates (k) were the greatest in L. formosana stand, followed by B. javanica, P. pinnata and C. camphora. Using N and P nutrient content in the leaves of four native tree species as the retranslocation thresholds, results showed that N was incomplete retranslocation except for B. javanica, but P retranslocation was complete in four tree species. An amount of nutrient cations input via litterfall to the exchangeable stock in the soil was the highest in P. pinnata (20%) and the lowest in C. camphora (7.7%). These results suggested that even on the same site, the litterfall and its nutrient return varied from different vegetation, and this cause changes in nutrient cycling and soil properties.

Soil carbon cycling processes are paid much attention by ecological scientists and policy makers because of the possibility of carbon being stored in soil via land use management. Soil respiration contributed large part of terrestrial carbon flux, but the relationship of soil respiration and climate change was still obscurity. Reforestation is one of solutions to mitigate CO₂ increase and to sequestrate CO₂ in tree and soil. Therefore, the objective of this study is to assess soil carbon and soil respiration impact during forest restoration process in subtropical broad-leaves plantation in central Taiwan. The research site located on central Taiwan was bamboo plantation before 2010. The soil belongs to Inceptisol with 40% of sandstone. The research site were divided to three treatments: Bamboo plantation area, reforestation area (broad-leaves plantation), and 30 years Taiwania cryptomerioides area. Each area were set 4 plots for soil sampling and soil respiration measurement from 13 to 2014. The data was showed that soil respiration rate was lowest in winter and highest in summer. Soil respiration rate was highly related to soil temperature, but was not significantly related to soil water content. However, soil respiration rate was decreased in rainy season, even soil temperature was still high. The annual highest soil respiration amount was reforestation area among three different treatments, indicating that when bamboo plantation was removed and planted broadleaves trees, solar radiation directly irradiated to soil surface resulting soil temperature high and therefore soil respiration increased.

Eddy covariance technique was used to indicate the seasonal dynamic of carbon and water vapor fluxes over two rubber (Hevea brasiliensis Müll. Arg.) plantations in Thailand, (1) The old rubber plantation, 21-year-old rubber plantation, at the Chachoengsao Rubber Research Station in Chachoengsao province (Eastern, Thailand) and (2) the young rubber plantation, 6-year-old rubber plantation, in Bueng Kan province (Northeastern, Thailand).  The both observation sites are monoclonal stand of rubber trees clone RRIM 600.  The pattern of seasonal variation of net ecosystem exchange (NEE) between both sites were similar.  In 2014, the annual NEE were -10.08 and -6.95 ton C ha^-1 year^-1 for the old and young rubber plantation, respectively.  The NEE reached it peaks (net source of CO₂) in January – February, because of serious drought and leaf fall period.  The NEE decreased in the following month, the system becoming a net sink of CO₂, due to increase of green area of the plantations.  The magnitude and distribution of net ecosystem production (NEP), gross primary production (GPP) and ecosystem respiration (Re) were controlled by seasonal changes in solar radiation, drought, as well as leaf emergence and senescence.  The pattern of seasonal variation of evapotranspiration (ET) between both sites were similar.  In 2014, the annual ET were 1,060 and 848 mm year^-1 for the old and young rubber plantation, respectively.  The annual ET of old plantation was significantly greater than the other, (young plantation).  In addition, we found daily evapotranspiration (ET) for both sites varied greatly during the growing seasons.  Overall, both plantations were a carbon sink and sensitive to changes in environmental conditions.

As heterotrophic respiration (RH) has great potential to increase atmospheric CO2 concentrations, it is important to understand warming effects on RH for a better prediction of carbon–climate feedbacks. However, it remains unclear how RH responds to warming in subtropical forests. Here, we carried out trenching alone and trenching with warming treatments to test the climate warming effect on RH in a subtropical forest in southwestern China. During the measurement period, warming increased annual soil temperature by 2.1 °C, and increased annual mean RH by 22.9%. Warming effect on soil temperature(WET) showed very similar pattern with warming effect on RH (WERH), decreasing yearly. Regression analyses suggest that WERH was controlled by WET and also regulated by the soil water content. These results showed that the decrease of WERH was not caused by acclimation to the warmer temperature, but was instead due to decrease of WET. We therefore suggest that global warming will accelerate soil carbon efflux to the atmosphere, regulated by the change in soil water content in subtropical forests.

To improve our understanding of responses of soil temperature to global warming in subtropical forests, we conducted a soil warming experiment in a subtropical evergreen broad-leaved forest in Ailao Mountains, Yunnan, SW China. Based on measurements from 2011 to 2013, we examined warming effects on seasonal and diurnal patterns of soil temperature and soil water content. The warming effects in dry season were greater than in rainy season. The warming effects showed seasonal variations, but not diurnal variations. Soil-surface temperatures increased between January to April, with a 3℃ maximum in February. Warming increased soil temperature by 2℃ at 5 cm. The warming effect decreased exponentially with soil depth. Based on a 0.5℃ temperature increase, soil warming could reach 3.82 m in the dry season, 12.04 m in the rainy season, with an annual mean of 6.58 m. warming did not change the seasonal and diurnal patterns of soil water content of the forest. The warming effects in winter and nighttime were greater than in summer and daytime. Warming decreased soil water content, more greatly in rainy season than that in dry season. Temperature increase and soil water decrease resulting from warming had seasonal variation, but no diurnal variation. The experimental warming increased soil temperature about 2 ℃, which could be considered meet the demand of 2 ℃ in warming experiment, and thereby provided good bases for the response of soil respiration to warming.

Rapid climate changes and intensifying human activities have resulted in water table lowering and enhanced nitrogen deposition in alpine wetlands on the Tibetan Plateau. The changes may affect greenhouse gas (GHG) emissions by influencing biological processes of production or consumption. However, the magnitude and pattern of CH4 emissions and the responses of GHG emissions to water table lowering and nitrogen deposition in alpine wetlands remain poorly understood, because of technical limitations and the harsh environments. We conducted the first in situ year-round observation using Eddy Covariance method and a mesocosm experiment simulating changes in water table depth (+3 and -20 cm soil depth relative to the surface) and nitrogen deposition (0 and 3 g N m-2 yr-1) to investigate the GHG fluxes in an alpine wetland. Results from in situ observation showed that the annual CH4 emissions were 26.4 and 33.8 g CH4 m-2 in 2012 and 2013, respectively, and a two-peak seasonal variation in CH4 fluxes was observed, with a small peak in the spring thawing period and a large one in the peak growing season. Furthermore, the non-growing season CH4 emissions accounted for 43.2-46.1% of the annual emissions, highlighting an indispensable contribution that was often overlooked by previous studies. Mesocosm experiment showed that water table lowering largely reduced CH4 emissions but did not affect net CO2uptake and N2O fluxes, while N deposition increased net CO2 uptake and N2O emissions but had little influence on CH4 emissions. As a result, both water table lowering and N deposition reduced the overall global warming potential. Our findings suggested that the CH4 emissions cannot be ignored during non-growing season in alpine wetlands on the Tibetan plateau, and future environmental changes may shift a typical alpine wetland from a GHG source to a GHG sink.

Plants litter is an important source of soil organic matter (SOM). Transformation and input of plant organic matter to soils during litter decomposition is a key process for understanding how SOM was formed and stored. In this study, a two-year litterbag experiment was set up in the Baiwang Mountain, northwest of Beijing, China. We chose seven types of plants, including leaves from three angiosperms (Quercus mongolica, Acer saccharum and Broussonetia papyrifera), one gymnosperm (Sabina chinensis) and three grasses (Setaria viridis, Amaranthus retroflexus and Chenopodium glaucum). The litterbags were collected monthly and the remained litters were weighed to calculate degradation rate. We also measured lipid biomarkers, such as alkanes (LCA), long-chain fatty alcohols and long-chain fatty acids, and solid ¹³C NMR in the plants litters to understand the progress of litter decomposition. After the two years decomposition, the leaves of all seven plants have lost 50-90 % of initial mass. Meanwhile, the changes of long-chain alkane concentration showed similar pattern among different species. Their LCA showed a general decreasing trend during two years decomposition. However, two or three apparent peaks of LCA depending on vegetation species were observed, corresponding to rain season during summer Asian monsoon. A possible explanation to this phenomenon is that selected loss of water soluble components (such as carbohydrate) by heavy rain or labile components (by pigment of carbohydrate) by intense microbial activity during summers with higher temperature and higher precipitation. Further studies such as lignin, NMR and stable isotope analyses are needed to confirm our hypothesis.  

In 2014-2015, improved open-path and enclosed-path gas analyzers were developed, based on established LI-7500A and LI-7200 analyzers, with the focus on improving stability in the presence of contamination, refining temperature control over the infrared light source, and providing more accurate gas concentration measurements. In addition to optical and electronic redesign, both analyzers incorporate automated on-site flux calculations using EddyPro® software run by a weatherized remotely-accessible microcomputer, called  SmartFlux. The ultimate goal of such development was to reduce errors in CO₂ and H₂O fluxes measurements. Field tests of both analyzers were conducted over six periods, each 5-14 months long, at 6 sites with diverse environments, setups, and types of contamination, using 26 gas analyzers. CO₂ and H₂O drift characteristics were consistently better in the new RS models than in the original analyzers.  Improvements in H₂O contamination-related drifts were particularly significant, with modified models often drifting many times less than the original. While the enclosed LI-7200RS performed substantially better than the original LI-7200 in terms of the drifts in H₂O. Improvements in CO₂ contamination-related drifts were modest. Results from field tests suggest that both RS systems can help improve flux data coverage and likely reduce site maintenance: §  Frequency of cleaning and site visits for service and maintenance should decrease, especially for the open-path design §  Amount of high quality of flux data is expected to increase for both  open-path and enclosed-path designs §  Amount of total data coverage over long periods of deployment should also increase substantially. The presentation will describe details and results from field tests of these new models in comparison with older models and control reference instruments. 

Up to now, it is not clear how environmental factors limitated physiological process of ecosystem carbon budget. The large uncertainties still existed in simulating the impact of climate change on photosynthetic capacity with process-based models. Particularly, it is difficult to describe the impact of drought on leaf photosynthetic capacity. In this study, a throughfall exclusion experiment was established in the subtropical evergreen coniferous plantation. Leaf photosynthetic activity and water status in control and dry treatment are monitored. We estimate leaf photosynthetic capacity (maximum carboxylation rate-V_cmax; the maximum electron transport rate-J_max; dark respiration-R_d; and stomatal conductance-g_s) by using Farquhar model based on the measured CO₂ response curve. We further describe the seasonal pattern of photosynthetic capacity and determine how drought modifies leaf photosynthetic capacity and its sensitivity to water limitation. Meanwhile, we discuss the effect of water deficit under different drought conditions on the key processes of typical forest ecosystem carbon cycle. The research will be useful for the development of ecosystem models, which improve the accuracy of simulated forest carbon cycle responding to drought.

The soil effective depth should be different due to regional features of various soils and vegetation, which impacts spatial and temporal distribution of soil moisture storage capacity and land surface water-carbon flux. In this study, the soil effective depth was calibrated using LPJ dynamic vegetation model on the target of remaining watershed water balance in three climate regions (Dongjiang watershed in humid areas, Huaihe watershed in humid，semi-humid areas and Jinghe watershed in semi-humid, semi-arid areas). It was used to analyze soil moisture storage capacity and land surface water-carbon flux (runoff R, actual evapotranspiration ET and net primary productivity NPP) due to variation of the soil effective depth, The results indicated that the estimated soil effective depth is 70cm in Dongjiang watershed, 90cm in Huaihe watershed and 140cm in Jinghe watershed. The soil effective depth and soil moisture storage capacity increases with drought index of the climate. It appears that the correction of the soil effective depth in terms of water balance effectively reduces the simulation error, and affects the simulated results of the land surface water-carbon flux. However, the large or small effect is related to climatic conditions. The annual mean runoff and actual evapotranspiration change significantly in the humid areas ,while NPP changes significantly in the humid and semi-humid areas. The results provide a reference for improving the reliability of application of the LPJ model in different climate regions.

The young rubber plantation which the canopy not closed together allow light to penetrate to the soil then weed can to grow under rubber canopy. The weed which is the vegetation under the canopy may play an important role in CO₂ and H₂O flux over the plantation. The rubber farmers usually control the weed by chemical or mechanical every year after rubber was planted. The weed usually growth and cover under canopy in rainy season and dry out in dry season. We have developed the tool that is used to measure CO₂and H₂O exchange of vegetation under rubber canopy. It was conducted in a 5-year-old (in 2014) rubber tree plantation in Pak Khat district, Bueng Kan province (Northeastern, Thailand). The rubber tree plantation was a monoclonal stand of rubber trees clone RRIM 600. The rubber tree density was 462 trees.ha^-1.  CO₂ exchange was measured by 0.5 x 0.5 x 0.5 m chamber with an open system. We found variation of CO₂ and H₂O exchange within row of rubber and seasonal variation due to variation in amount of vegetation. Diurnal variation also found due to change in light and temperature along the day.

Accurate quantification of the contribution of environmental variability and functional changes to the interannual variability of net ecosystem production (NEP) and evapotranspiration (ET) in coniferous forests is needed to understand global carbon and water cycling. This study quantified these contributions to the interannual variability of NEP and ET for a subtropical coniferous plantation in southeastern China, and the effect of drought stress on these contributions was also investigated. The NEP and ET were observed from eddy covariance measurements during 2003–2012. A homogeneity-of-slopes model was adopted to quantify the contribution to the interannual variability of these fluxes. Environmental variability accounted for 71 % and 85.7 % of the interannual variability of NEP and ET, respectively; however, correspondingly, functional changes accounted for only 11.3 % and 5.9 %, respectively. Furthermore, functional changes explained more of the interannual variability of NEP in dry years (16.3 %) than in wet years (3.8 %), but there was no obvious change in the contribution of functional changes to the interannual variability of ET in dry (4.7 %) or wet (5.5 %) years. Thus, environmental variability rather than ecosystem functional changes dominated the interannual variability of both ET and NEP. However, different environmental variables controlled the interannual variability of NEP and ET. The results also indicated that, compared with NEP, ET was more resistant to drought stress through the self-regulating mechanisms of this plantation.