Because evapotranspiration (ET) is the second largest component of the water cycle and a critical process in terrestrial ecosystems, understanding the inter-annual variability of ET is important in the context of global climate change. Eight years of continuous eddy covariance measurements (2003–2010) in a subtropical coniferous plantation were used to investigate the impacts of climatic factors and ecosystem responses on the inter-annual variability of ET. The mean and standard deviation of annual ET for 2003–2010 were 786.9 and 103.4 mm (with a coefficient of variation of 13.1%), respectively. The inter-annual variability of ET was largely created in three periods: March, May–June, and October, which are the transition periods between seasons. A set of look-up table approaches were used to separate the sources of inter-annual variability of ET. The annual ETs were calculated by assuming that (a) both the climate and ecosystem responses among years are variable (Vcli-eco), (b) the climate is variable but the ecosystem responses are constant (Vcli), and (c) the climate is constant but ecosystem responses are variable (Veco). The ETs that were calculated under the above assumptions suggested that the inter-annual variability of ET was dominated by ecosystem responses and that there was a negative interaction between the effects of climate and ecosystem responses. These results suggested that for long-term predictions of water and energy balance in global climate change projections, the ecosystem responses must be taken into account to better constrain the uncertainties associated with estimation.
Citation: Xu M, Wen X, Wang H, Zhang W, Dai X, Song J, et al. (2014) Effects of Climatic Factors and Ecosystem Responses on the Inter-Annual Variability of Evapotranspiration in a Coniferous Plantation in Subtropical China. PLoS ONE 9(1): e85593. https://doi.org/10.1371/journal.pone.0085593
Editor: Dafeng Hui, Tennessee State University
Received: July 16, 2013; Accepted: November 29, 2013; Published: January 22, 2014
Copyright: © 2014 Xu et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was funded by the National Key Basic Research Special Foundation Project (2010CB833501), Major Program of National Natural Science Foundation of China (31290221), National Natural Science Foundation of China (31070559, 31210103920) and the National Key Basic Research Program of China (2012CB416903). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Evapotranspiration (ET) is the second largest component of the water cycle, consuming a large proportion (60–90%) of annual precipitation –. ET plays an important role in terms of the energy and nutrient exchange in ecosystems – and affects many other important ecological processes . Forests cover approximately 30% of the global land surface, but their evapotranspiration accounts for more than 45% of the global ET across terrestrial ecosystems . In the context of climate change, forests are considered to provide important climate forcings and feedbacks . While climate change may adversely affect ecosystem functions, forests could be managed to mitigate climate change . As ET is an integral part of forest biogeophysical processes , meteorologists and ecologists require an understanding of how the forest ET responds to environmental drivers so that they can better predict how the biosphere will affect (or will be affected by) climate change –.
The factors that control ET have been well documented for many ecosystems on intra-annual scales –. Evapotranspiration is controlled by abiotic and biotic factors, many of which are projected to change on multiple spatial and temporal scales . Generally, ET is positively related to net radiation (Rn), air temperature (Ta), and the vapor pressure deficit (VPD) , –. However, soil water content (SW) sometimes becomes the main control factor under drought conditions . As for biotic factors, leaf area index (LAI)  and additional model parameters that have been used to represent the canopy characteristics have influences on ET. For example, canopy conductance, which can be computed by inverting canopy resistance from the Penman-Monteith equation –, can help estimate ET. These parameters were found to be closely related to ET, especially under drought stress , .
In contrast to intra-annual variability, the inter-annual variability of ET and its driving mechanisms are poorly understood. Although many forest evaporation datasets have been collected worldwide using the eddy covariance technique and compiled over years to decades , , only a small fraction of studies have published their long-term ET records. Furthermore, most of these studies have concentrated on tropical forests , , , and boreal forests , . There have also been a few studies reporting the inter-annual variability of ET based on 3–5 year measurements in temperate forests , , . Some of these studies have determined that the inter-annual variability of ET was small (coefficient of variation (CV) <10%) in contrast to the considerable year-to-year variation in annual rainfall , , . Others who held the opposite views have insisted that the inter-annual variability of ET was significant and could not be neglected . In addition, several models have been developed to study or predict long-term variation in ET . However, these models may be inadequate for evaluating the inter-annual variability of ET because of their dependency on climatic factors –.
The inter-annual variability of ET can be affected by biotic and abiotic factors, similar to those that operate on shorter time scales , . In addition to the factors mentioned above, the leaf phenology should be given careful consideration. As noted by Xue et al. , the reason that Ohta et al.  could not fully explain the inter-annual variability of ET in a Siberian larch forest was because they neglected the inter-annual variability in the dates of leaf expansion. Zha et al.  determined that the inter-annual variability of ET was controlled by early spring soil temperatures, which are a key factor affecting the growing season length (GSL).
In this study, the sources of the inter-annual variability of ET were separated into two parts: (1) those induced by climatic factors (extreme climatic events included) and (2) those produced by ecosystem responses including the changes in phenology (e.g., GSL), ecosystem structure (e.g., LAI), and physiological changes in plants (e.g., stomatal and canopy conductance) , . Some studies have proposed that ecosystem responses overwhelm the direct influences of environmental factors on the inter-annual variability of ET , , , but few researchers have separated the sources of inter-annual variability of ET into ‘biological’ and ‘climatological’ components because they are quite difficult to tease apart. However, in recent studies, the inter-annual variations of carbon fluxes were successfully separated into those driven by environmental factors or by biological factors with a novel statistical method , . These studies suggested that it would be possible to separate the sources of inter-annual variability of ET because water processes are highly correlated with carbon assimilation in an ecosystem.
Southern China is characterized by a humid monsoon climate and has large subtropical coniferous plantations, which account for 41% of the total subtropical forest area . Thus, it is important to conduct research on the carbon and water cycles in this region. Previous studies have indicated that the carbon fluxes are characterized by strong inter-annual variations due to seasonal droughts in the summer or low temperatures in the early spring , . Considering that water and carbon cycles are directly coupled within the stomata through transpiration, photosynthesis, and respiration , evapotranspiration may respond to climatic drivers as do the carbon fluxes. Therefore, the goal of this study is to investigate whether the inter-annual variability of ET in this region is strong and to determine the main source of the inter-annual variability of ET.
Eight years (2003–2010) of continuous eddy-covariance measurements that were recorded at a subtropical coniferous forest site were analyzed. Based on previous research on the carbon cycle , a set of look-up tables (LUTs) were employed to separate the sources of the inter-annual variability of ET. The objectives of this study were to (1) characterize the seasonal and inter-annual variation of ET and (2) evaluate the contributions of the two types of driving forces, climatic factors and ecosystem responses, to the inter-annual variability of ET.
Materials and Methods
This study was conducted in a subtropical coniferous plantation (26°44′29″N, 115°03′29″E, 102 m in elevation above sea level) at Qianyanzhou Ecological Research Station (QYZ), which is owned by the Institute of Geographic Sciences and Natural Resources Research of the Chinese Academy of Sciences. No specific permits were required for the described field studies in the research site. The field studies did not involve endangered or protected species. The site is located in a typical red soil hilly region in south China, with a subtropical monsoon climate. The prevailing wind direction of this climate regime is north-northwest in the winter and south-southeast in the summer. The coniferous trees were planted around 1985. The dominated species are Masson pine (Pinus massoniana Lamb.), Slash pine (Pinus elliottii Englem.) and Chinese fir (Cunninghamia lanceolata Hook.), with a tree density of about 1460 stems ha−1 and total biomass of 106 t ha−1. The red soil is weathered from red sand rock, and soil texture is divided into 2.0–0.05 mm (17%), 0.05–0.002 mm (68%) and <0.002 mm (15%). Further details of QYZ site can be referred to Wen et al.  and Wang et al. .
According to meteorological observations from 1989–2010 at QYZ, the mean annual temperature is 18.0°C and mean annual precipitation is 1509.0 mm, while the growing season (April–October) mean values are 23.9°C and 1050.8 mm, respectively (Fig. 1). Eight years (2003–2010) of flux data were observed with the eddy covariance method in a subtropical forest plantation at QYZ and used for this study. During the eight observation years, the dataset represented much of the typical inter-annual variability at this site, including both typical and extreme climatic events. In 2002–2003, a typical El Niño event occurred, which resulted in an extreme summer drought in 2003 , and in 2009–2010, there was a strong La Niña event .
Observation and instrumentation
The eddy covariance flux observation system was established in a subtropical coniferous plantation at QYZ site in 2002 (Fig. 2). The above-canopy flux was measured at a height of 39.6 m by instruments loaded on a ventilated tower. The wind velocity was detected by a 3-D sonic anemometer (Model CSAT3, Campbell Scientific Inc., USA), and the variations in CO2 and water vapor concentration were measured with a LI-7500 open-path CO2/H2O analyzer (Licor Inc., USA). All signals were sampled at a frequency of 10 Hz and the CO2 and H2O fluxes were calculated and recorded at 30 min intervals by a CR5000 datalogger (Campbell Scientific Inc., USA).
The interval between buffer circles is 200.
Air temperature and relative humidity sensors (Model HMP45C, Campbell) were mounted in ventilated shield at the heights of 23.6 m and 39.6 m above the ground. Soil temperature and soil water content were measured at the depths of 5 cm with thermocouples (105T and 107-L, Campbell), and TDR probes (Model CS615-L, Campbell), respectively. Radiation measurements were made using a four-component net radiometer (Model CNR-1, Kipp & Zonen, Netherlands), a pyranometer (Model CM11, Kipp & Zonen) and a quantum sensor of photosynthetically active radiation (Model LI190SB, Licor Inc.). Rainfall was monitored with a rain gauge (Model 52203, RM Young Inc., USA). Meteorological variables were sampled at 1 Hz, and 30 min average data were recorded with three CR10X dataloggers, and a CR23X datalogger with a 25-channel solid-state multiplexer (Campbell Scientific Inc., USA).
Flux correction and gap filling
This study adopted the methods of calculating and correcting the carbon dioxide and water vapor fluxes that were presented by Wen et al. . The CO2 and H2O fluxes were calculated every 30 minutes from the 10 Hz raw data. Processing of the flux data was performed using routine methods, including three-dimensional rotation , a correction for the effects of the fluctuations of air density (the Webb, Pearman and Leuning density correction (WPL correction)) , storage calculations and spurious data removal , , . Spurious data caused by rainfall, water condensation or system failure were removed from the dataset. To avoid the possible underestimation of the fluxes under stable conditions at night, nighttime data (solar elevation angle<0) were excluded when the value of friction velocity (u*) was below the relevant thresholds, which were identified with the method described by Reichstein et al. . The threshold values of u* ranged from 0.16 to 0.22 m s−1, with an average value of 0.20 m s−1 for the years from 2003 to 2010. Approximately 23.8% of the half-hour data were excluded for the daytime and 85.4% were excluded for the nighttime. The data gaps were filled using the look-up table method . Any data gaps in meteorological variables were filled using the mean diurnal variation method . Further details of data processing are presented in the previous studies of ChinaFLUX , , , . In this study, the H2O flux was considered to be ET of the whole ecosystem, as in previous studies , .
Partitioning the sources of inter-annual variation in ET
To separate the effects of climate drivers and ecosystem responses on the inter-annual variability of carbon fluxes, Marcolla et al.  established a set of look-up tables (LUTs) according to selected key factors that influence the ecosystem carbon assimilation, such as temperature (Ta) and photosynthetically active radiation (PAR). Using these LUTs, these researchers statistically separated the two sources of inter-annual variability in carbon fluxes. In this study, a similar methodology was applied, but the LUTs for ET were built with the environmental factors Rn and VPD, which have a strong influence on the ET and widely used in combination with Penman-Monteith Equation .
To create the LUTs, the half-hour flux values and corresponding meteorological data were integrated to daily scale for each year. Then, the data were classified according to Rn (class width 1 Wm−2) and VPD (class width 0.1 kPa). With the LUTs, we could then use the meteorological data (Rn and VPD) to estimate a corresponding ET value. Specifically, the ETs were calculated in three different ways:
- The yearly LUTs were established using the meteorological data (Rn and VPD) and the corresponding ET. Then, the yearly observed meteorological data from the entire time series (DOY (day-of-year) 1–365 (366 for leap year)) was used to reproduce the ET series according to the yearly LUTs. The ET series and the inter-annual variability of ET calculated with this method can be considered to include effects of both climatic factors and the ecosystem responses (Vcli-eco).
- To isolate the climate factors, an average table that included meteorological data and ET for the entire time series from DOY 1 to 366 (including leap years) was established based on the eight-year data. Then, the LUTavg based on the average table was established. In the LUTavg, the ETs represent the average ecosystem responses for the eight years that were measured. Finally, the LUTavg is looked up using the yearly observed meteorological data from the entire time series to reassign ET series. The inter-annual variability of ET that is calculated through this method can be considered to be the inter-annual variability caused by variable climatic factors (Vcli).
- To isolate the ecosystem responses, the eight-year averaged daily meteorological data with entire time series (DOY 1–366) can be used to look up the yearly LUTs established in approach (1) to reproduce ET series for each year. The inter-annual variability of ET calculated through this method can be considered to be the inter-annual variability caused by variable ecosystem responses (Veco).
The standard deviations of the yearly values of ET that were obtained through the three approaches (Vcli-eco, Vcli and Veco) were used as estimates of what portion of the inter-annual variability of ET was generated from which sources. Then the interaction of climate drivers and ecosystem responses was analyzed based on these results.
Correlation between climatic drivers and ecosystem evapotranspiration
To identify the relationship between climatic drivers and ET, we selected five climatic factors, net radiation (Rn), air temperature (Ta), precipitation (P), vapor pressure deficit (VPD) and soil water content (SW), that were important according to the previous studies , , , . We analyzed their correlations with the ecosystem ET on the annual scale, placing emphasis on the time lag effects between the climatic factors and ET.
For this purpose, annual average Ta, SW, VPD and cumulative Rn, precipitation and ET were calculated. Multiple “yearly” statistics (approximately 80 values from the eight years) were obtained using the 12-month intervals and shifting them one month at a time , . To investigate the lag of the ET's response to the climatic factors, we shifted the climatic series backward one month at a time (up to twelve months) to calculate the correlations between climate drivers and ET. Student's t–tests were applied to verify the statistical significance of correlation coefficients.
Seasonal and inter-annual variation of ecosystem evapotranspiration and the corresponding environmental factors
The ET, net radiation (Rn), air temperature (Ta), vapor pressure deficit (VPD), precipitation (P) and soil water contents (SW) are characterized in Figure 3. The ecosystem ET and corresponding Rn, Ta, and VPD values showed a single peak, and reached their maximums in July (Fig. 3). The precipitation also had a strong secondary peak. The precipitation decreased in July as the air temperature reached a maximum. The asynchronous seasonality of temperature and precipitation, an important characteristic of the eastern Asian monsoon climate, often results in summer droughts in southern China. In addition, precipitation in the first half year was usually higher than in the second half of the year. The seasonal pattern in soil water content (SW) (at 5 cm) was not as obvious as the other factors. It remained between 0.12 m3 m−3 and 0.22 m3 m−3 and decreased after July, with values in the second half year generally lower than in the first half year, in accordance with precipitation pattern.
The error bars represent the standard deviation.
Figure 4 and Table 1 show the environmental condition and ET for each year at QYZ. Over the eight years, the average annual ET was 786.9±103.4 mm (CV = 13.1%) with a substantial variation, from 567.9 mm in 2005 to 876.5 mm in 2009. The years of 2007 and 2009 had the highest ET, whereas 2005 and 2006 had the lowest values (Table 1). Correspondingly, the Ta was higher in 2007 and 2009 in the early spring, and the Rn in 2009 was higher, with positive anomalies for ten months (Fig. 4). In 2004 and 2008, the climatic factors were close to the multi-year averages, except for a lower Rn and Ta in the early spring of 2008. In 2010, although the Ta was lower, the precipitation was abundant. In 2003, the negative anomalies in precipitation were large, and the asynchrony between precipitation and temperature resulted in a severe summer drought in June and July. In 2005 and 2006, the Rn was low in the first half year, and its annual total was approximately 6.4% lower in both years compared to multi-year average.
The El Niño and La Niña events had significant impacts on the climatic factors. In the El Niño year (2003), the Ta was higher, but the precipitation was lower. In the La Niña year (2010), the precipitation was higher but the Ta was lower (Fig. 4b and c, Table 1). In addition, under the effects of El Niño events, the SW was lower in the second half year of 2003 and the first half of 2004, but the VPD was higher during this period. In contrast, under the influence of La Niña, the pattern in SW and VPD anomalies was opposite to that of the El Niño year (Fig. 4d and e). Correspondingly, in the El Niño year, the anomalies in the ET were negative for ten months, but in the La Niña year, the opposite trend was found, and the anomalies in the ET were positive for eleven months (Fig. 4f).
Partitioning the sources of inter-annual variation of ET
To separate the sources of the inter-annual variability in ET, a statistical method (LUTs) was used to calculate the ET under different conditions. These approaches allowed the climatic drivers and ecosystem responses to change separately or together.
The annual ETs calculated through the three LUT approaches are presented in Figure 5. The average annual ETs were 798.5, 786.9, 754.1 mm for Vcli, Vcli-eco, Veco, respectively, and the corresponding inter-annual variabilities were 28.13, 103.40 and 93.50 mm (Fig. 6). Vcli showed the highest average annual ET and the lowest inter-annual variability. Compared to Vcli-eco, which was almost equal to the measured ET, Vcli overestimated ET and dampened the inter-annual variability. This result implied that the ecosystem responses may offset climatic effects in the long term. In addition, the inter-annual variability of Vcli was only 27.2% of that produced by Vcli-eco. At the same time, the correlation coefficient between the ETs of Vcli and Vcli-eco was only 0.48 (p = 0.2316), which indicated that the ecosystem had a noticeable resistance to direct changes in the external drivers. In contrast, Veco underestimated ET and produced higher inter-annual variability. This indicated that the climatic factors drove the daily function of the ecosystems and maintained the normal ecosystem status, while the ecosystem responses furnished the main source of the inter-annual variability in ET.
The correlation coefficients of the cumulative values calculated with the three approaches are also reported.
Additionally, the inter-annual variability of ET was 28.13 mm as generated by variable climatic factors and was 93.50 mm as generated by ecosystem responses. If there were no interaction, the inter-annual variability of Vcli-eco should be 121.63 mm, but the combined value was 17.6% higher than the real result 103.40 mm (Fig. 6). Therefore, we can infer that the changes in ecosystem responses were counteracting the effects of climatic variations, an inference that agrees with the classical ecological theory.
Direct and indirect effects of climatic factors on ET
Previous researchers have confirmed that variations in ET can be attributed to climatic factors and the biotic responses induced by environmental forcing , , , . Although they are entangled and difficult to separate from each other, there are many studies that refer to the role biotic factors play in driving the variation in ET, especially in the long term , . In this study, we used a statistical method to separate the effects of climatic drivers and ecosystem responses and quantified their contributions to the inter-annual variability of ET. The results underscore the prominent role that ecosystem responses play in the inter-annual variability of ET (Fig. 5 and Fig. 6). The ecosystem responses include many aspects, such as variation in LAI between years and variability in the photosynthetic ability of plants. These factors are sometimes difficult to measure and quantify over longer time scales. Fortunately, the indirect effects of climatic factors can, to some extent, reflect the ecosystem responses and help us understand the effects of ecosystem responses on the inter-annual variability of ET.
ET is composed of two components: evaporation and transpiration , . When analyzing how biotic and abiotic factors affect ET, we should consider the effects on each part. Evaporation is mainly a physical process driven by radiation and VPD . Transpiration is a biotic process, which may also be driven by environmental factors but that is mainly controlled by biological variables . In this study, ET was found to be related to Rn, Ta, VPD, SW and precipitation, which is consistent with previous studies , –. The lag effects became apparent when the correlograms between annual ET and dominant climatic drivers were drawn. This phenomenon indicated that the ecosystem responses sometimes did not happen simultaneously with the environmental change.
Numerous connections between climatic factors and ET were observed in the short and long term . On a shorter time scale, the direct effects of climatic factors are obvious. The correlation coefficients between climatic factors on monthly time scale were larger than they were on the annual scale (Table 2), which is similar to findings in studies of carbon fluxes , . In addition, the ET showed a similar seasonal rhythm to the Rn, Ta and VPD (Fig. 3). On the annual scale, the ET is still related to the climatic drivers. The annual ET was sometimes linked to the temperature and precipitation anomalies caused by the extreme climate events such as the El Niño event in 2003, which caused a low ET (27.6 mm lower than the average) (Fig. 4 and Table 1). However, generally the relationships between ET and the climatic factors were not as close as they were on shorter time scales. Moreover, when the statistical methods were employed to further analyze their relationship, important lag effects emerged.
Two mechanisms may explain the lag effects and lower correlation coefficients on longer time scales. First, the rhythm signals of finer time scales are disguised when the analysis is conducted on longer time scales. Second, the ecosystem requires time to respond to climatic factors and its modulating effect is prominent over longer period. This is what we refer to as ecosystem responses. However, the lag effects did not exist for all climatic factors (e.g., Rn and VPD, Fig. 7). The evaporation from the soil and the canopy surface respond quickly to changes of Rn, as demonstrated by Monteith . Meanwhile, Rn and VPD act on transpiration as a pulling force, and their effects were instantaneous .
Correlograms were calculated by shifting the climatic driver backward one month at a time. The open signs represent that the coefficients are not significant while the closed signs mean significant (p<0.05).
As for Ta, precipitation and SW, there are prominent lag effects as shown in Figure 7. This result agreed with some previous studies , and it can be observed as the indirect effect of climatic factors. The lag may be helpful in understanding the ecosystem responses. Here, we try to interpret the lag effects of Ta, SW and precipitation. The relationship between ET and Ta, showed a parabolic lag effect. The highest correlation coefficient showed up at 8 month lags, which is likely because the growth of forests often requires more time to respond to changes in temperature. This may be partly because of its impact on the growing season length (GSL) . In this study area, the temperatures between April and November were generally above the threshold 5°C. Therefore, the GSL mainly depended on the temperature between December and March (Table 3). Zhang et al.  noted that lower temperatures in the early spring greatly reduce the annual net ecosystem exchange (NEE) of carbon by shortening the GSL in this study area. Because transpiration, which is the dominant component of ET, is coupled with photosynthesis, we tested the effect of GSL on ET. The resulting correlation coefficients between ET and GSL (Ta≥5°C, in days) were as high as 0.56 (p = 0.1471, n = 8). This may partially explain the lower ET in 2005 and higher ET in 2007, because the GSL was shorter in 2005 and longer in 2007 (Table 3). This result is comparable to the results obtained by Zha et al.  which showed that the inter-annual variability of ET was controlled by early spring soil temperatures in three boreal forests and a grassland ecosystem.
The highest correlation coefficients between ET and SW at 5 cm were found at time lags of 10–12 months. This indicated that the soil water conditions in a given year may have significant effects on ET in the following year. This phenomenon may be attributed to the rooting depth and available SW at deeper layers or potentially to plant responses that result in altered water use strategies under the soil water stress , , . In our study site, the dominant tree species are Slash pine, Masson pine and Chinese fir. These pioneer species have deep root systems that may be capable of absorbing water from deeper layers, much like the plants in Florida scrub oak and pine flatwoods ecosystems that Bracho et al.  described as able to obtain water from deeper sources.
The correlation coefficients between precipitation and ET were negative on the annual scale. When the precipitation was high, forest ET was likely reduced because of the lower solar radiation due to frequent cloud cover. In dry years, as Lundquist and Loheide  described, the ecosystem typically has greater potential ET due to higher Rn, higher Ta, higher VPD and longer growing seasons. All of these factors provide strong driving forces for ET. In our study area, the precipitation is abundant on the annual scale. Even in dry years, the trees can access water in deeper layers, which may lead to the relatively high ET. Another possibility as suggested by Pejam et al. , is that larger precipitation events would contribute more to the surface run-off than to soil water storage. Because our study site is hilly, with red earth characterized by high clay soils with relatively poor infiltration rates, there may be high runoff when the rain is heavy .
Contributions of seasonal variations to the inter-annual variation of ET
During the observed years, the ET of this ecosystem was positive throughout the year (Fig. 8a). Although the ET for all eight years had a similar cumulative trend, large inter-annual variability emerged when comparing the cumulative curves of different years (Fig. 8a). The inter-annual variability of ET in this subtropical coniferous plantation was also relatively large when compared to the other ecosystems (Table 4).
At the beginning of each year, from January to mid-March, the ET increased slowly at less than 2 mm d−1. From mid-March to mid-July, the cumulative ET rose sharply, up to 4 mm d−1 in June, and then gradually declined through the end of the year. With respect to the increasing rate of ET, the absolute difference in ET among years was not large in the beginning of the year. Since the growing season, accompanied with the intensive transpiration, the discrepancies in cumulative curves became larger (Fig. 8a). However, the increases in rates could not fully account for the contributions to the inter-annual variability in the ET. Therefore, to evaluate the contributions of different phenological phases to the variability in annual ET, the difference between the standard deviation of cumulative ET from a given month and its preceding one (DstDev) was calculated to quantify the contribution of each month to the observed inter-annual variability of the ET. Over the whole year, the DstDev had three peaks, in March, May–June and October (Fig. 8b), which coincide with the seasonal transitions.
In subtropical China, March is the beginning of the spring, and new twigs generally sprout as the plant leaves expand. Accompanying this phenological transition, the cumulative ET increased rapidly (Fig. 8a) because of the great increase in transpiration . However, in accordance with variations of early spring temperatures , the phenological calendars varied among years. For example, in 2005 the low temperature in spring postponed the twig and leaf emergence and caused phase lags in the enhanced vegetation index (EVI) time series . Therefore, the phenological transition and the advances or lags in growing seasons gave the ET in March a bigger role in the inter-annual variability of ET. Zha et al.  found a similar effect in the Siberian forests, years with warmer spring temperatures had a greater overall ET because of earlier leaf emergence. Hollinger et al.  also found a similar phenomenon in a carbon flux study in forests. They noted that the climatic conditions in the spring largely influenced the inter-annual variability of NEE. The ecosystem they studied absorbed more atmospheric carbon dioxide in years with warm spring and less carbon dioxide in years with cold spring.
The variations of ET in May and June had the largest contribution to the inter-annual variability of ET in this study. This may have occurred for the following reasons. First, there was substantial variability in the ecosystem structure and function in this period. The ecosystem behaved similarly in this period every year, but the differences were great between years. For example, the LAI or EVI would level off during this period each year, but when compared between years, large differences emerged . Transpiration, which accounts for 70–80% of ET , , became more active in this period . Therefore, under the coactions of these two factors, the ET varied greatly in this period between years. Second, this period marks the transition between wet and dry seasons, which occurs with the seasonal reversal of the monsoon circulation features . The variability in the onset of the monsoon ,  led to a high inter-annual variability in the climatic factors. The climatic factors and variable ecosystem responses combined to make this period the most influential on the inter-annual variability of ET.
In October, the DstDev increased abruptly, which might have been caused by the inter-annual variability in senescence time of old needles and the retreat time of the monsoon . After October, the lower Dstdev may have been driven by the lowering activity of the plants. Therefore, the underlying mechanisms that make these seasonal transitions important to the inter-annual variability of ET may be the coactions of climatic factors and ecosystem responses. This is consistent with the view that seasonal transitions have great influence on fluxes and energy exchange and that they play an important role in inter-annual variability , .
The role of ecosystem responses in long term studies
Previous studies have generally focused more on the effects of climatic factors. Although some researchers have considered the biotic factors such as canopy conductance to be as important as climatic factors , , , the biotic effects on ET have not generally received adequate consideration . In our study, we interpreted all biotic factors as an integrated part of ecosystem responses, and we separated the sources of the inter-annual variability of ET into those that were induced by climate and those produced by the ecosystem responses. The results suggested that ecosystem responses were largely responsible for the inter-annual variability of ET (Fig. 5 and Fig. 6). As discussed above, both of the indirect effects of the climatic factors and the three peaks in variability contributed to the inter-annual variability of ET as shown in Figure 8, underscore the importance of ecosystem responses. By taking ecosystem responses into consideration, more accurate estimations of ET may be made, particularly in years where climatic drivers seem unable to capture the variation in ET (Fig. 5). Our results indicate that the ecosystem ET may respond directly to the climatic forces, but it may also provide some buffering effects because the ecosystems requires time to respond to external stimulations and adjust its functions and structures .
Currently, several types of models are used to estimate and predict ET worldwide . However, most of those models tend to underestimate its inter-annual variability . Komatsu et al.  suggested that the models can be roughly classified into two groups: simple models that use only climatic factors as input variables  and complex models that employ more process-based formulations including the Penman–Monteith equation or the bulk equation . Although the process-based models give limited consideration to biotic factors, they are still limited in their ability to estimate the inter-annual variability of ET. Our results indicate that the ecosystem responses are more important to the inter-annual variability of ET than the environmental drivers, which implies that to more precisely evaluate ecosystem ET, the features of ecosystem responses should be considered alongside the climatic factors, especially for longer time scales.
There is also potential for an effect of CO2 fertilization on the inter-annual variation of ET under climate change context. It has been reported that the elevated CO2 concentrations may enhance photosynthesis in forests . They may also affect the transpiration, the dominant component of ET, because photosynthesis and transpiration are coupled in the stomata . In addition, elevated CO2 concentration may affect the growing season length – and induce changes in the ecosystem functions –. These changes in ecosystem responses may in turn increase the inter-annual variability of ET. Recently, the atmospheric CO2 concentration has been increasing by approximately 2 ppm per year , and it rose approximately 14 ppm during our study period (2003–2010) . Compared to CO2 enrichment experiments and models, this increase is minor, and therefore the effect of CO2 fertilization on ET might not have been captured by our study. However, in the context of climate change, the rising atmospheric CO2 concentrations may affect ET in the long term by affecting plants and by magnifying ecosystem responses (e.g., LAI, GSL) . Such scenarios should be discussed in the future.
However, ecosystem responses are complicated and difficult to quantify. To demonstrate the underlying mechanisms, accurate knowledge of how the components of ET (e.g., soil evaporation, canopy evaporation and transpiration) respond to climatic drivers is vital and should be targeted for future investigations of ET. These efforts will require more comprehensive models and additional data. Though some processes respond quickly to variation in climatic factors, other biogeochemical processes may need a longer time to respond to the environmental changes . Thus, long-term physiological and biogeochemical research is needed to elucidate the bioprocesses and the underlying mechanisms.
- The ET of a coniferous plantation in subtropical China showed a strong seasonal pattern, consistent with the single peak of climatic factors such as Rn, Ta, VPD,which reached their maximums in July. During the observation period from 2003 to 2010, the mean value of annual ET was 786.9 mm. A relatively large inter-annual variability with a standard deviation of 103.40 mm (CV = 13.1%) was detected.
- The main source of the inter-annual variability of ET was ecosystem responses, according to the LUT results. The inter-annual variability of ET caused by climate variability and that caused by changes in ecosystem responses had a negative interaction.
- On the annual scale, the main meteorological drivers for ET were Rn and Ta. The Rn affected ET directly without a lag effect. In contrast, the correlation coefficients between ET and Ta, SW and precipitation showed significant lag effects which reflect the ecosystem responses.
- The variation during seasonal transitions in March, May–June, and October contributed heavily to the inter-annual variability of ET, which could be attributed to the coactions of climatic factors and ecosystem responses.
The authors greatly thank Dr. Xianjin Zhu and Dr. Tao Zhang for their valuable comments and suggestions. We appreciate Dr. M. Luke McCormack for reviewing the English. We are greatly grateful to Dr. Dafeng Hui and the reviewers for their helpful comments that improved this paper.
Conceived and designed the experiments: MJX XFW HMW. Performed the experiments: MJX HMW WJZ XQD YDW XLF YFL. Analyzed the data: MJX HMW XFW. Contributed reagents/materials/analysis tools: YFL XMS GRY. Wrote the paper: MJX HMW JS.
- 1. Liu SG, Riekerk H, Gholz HL (1998) Simulation of evapotranspiration from Florida pine flatwoods. Ecological Modelling 114: 19–34.
- 2. Gholz HL, Clark KL (2002) Energy exchange across a chronosequence of slash pine forests in Florida. Agricultural and Forest Meteorology 112: 87–102.
- 3. Sumner DM, Belaineh G (2005) Evaporation, precipitation, and associated salinity changes at a humid, subtropical estuary. Estuaries 28: 844–855.
- 4. Powell TL, Starr G, Clark KL, Martin TA, Gholz HL (2005) Ecosystem and understory water and energy exchange for a mature, naturally regenerated pine flatwoods forest in north Florida. Canadian Journal of Forest Research-Revue Canadienne De Recherche Forestiere 35: 1568–1580.
- 5. Bracho R, Powell TL, Dore S, Li J, Hinkle CR, et al. (2008) Environmental and biological controls on water and energy exchange in Florida scrub oak and pine flatwoods ecosystems. Journal of Geophysical Research 113: G02004.
- 6. Tian FX, Zhao CY, Feng ZD (2011) Simulating evapotranspiration of Qinghai spruce (Picea crassifolia) forest in the Qilian Mountains, northwestern China. Journal of Arid Environments 75: 648–655.
- 7. Miao H, Chen S, Chen J, Zhang W, Zhang P, et al. (2009) Cultivation and grazing altered evapotranspiration and dynamics in Inner Mongolia steppes. Agricultural and Forest Meteorology 149: 1810–1819.
- 8. FAO (2006) Global Forest Resources Assessment 2005: Main Report. Progress Towards Sustainable Forest Management. p. 320. FAO Forestry Paper 147, Rome, Italy.
- 9. Bonan GB (2008) Forests and climate change: forcings, feedbacks, and the climate benefits of forests. Science 320: 1444–1449.
- 10. Parry ML, Canziani OF, Palutikof JP, van der Linden PJ, Hanson CE (2007) Climate change 2007: impacts, adaptation and vulnerability: Intergovernmental Panel on Climate Change.
- 11. Kumagai To, Katul GG, Porporato A, Saitoh TM, Ohashi M, et al. (2004) Carbon and water cycling in a Bornean tropical rainforest under current and future climate scenarios. Advances in Water Resources 27: 1135–1150.
- 12. Wang K, Dickinson RE (2012) A review of global terrestrial evapotranspiration: Observation, modeling, climatology, and climatic variability. Reviews of Geophysics 50: RG2005.
- 13. Fisher JB, Whittaker RJ, Malhi Y (2011) ET come home: potential evapotranspiration in geographical ecology. Global Ecology and Biogeography 20: 1–18.
- 14. Shukla J, Mintz Y (1982) Influence of land-surface evapotranspiration on the earth's climate. Science 215: 1498.
- 15. Ge Z-M, Zhou X, Kellomäki S, Peltola H, Wang K-Y (2011) Climate, canopy conductance and leaf area development controls on evapotranspiration in a boreal coniferous forest over a 10-year period: A united model assessment. Ecological Modelling 222: 1626–1638.
- 16. Huntington TG (2006) Evidence for intensification of the global water cycle: Review and synthesis. Journal of Hydrology 319: 83–95.
- 17. Scott RL, Edwards EA, Shuttleworth WJ, Huxman TE, Watts C, et al. (2004) Interannual and seasonal variation in fluxes of water and carbon dioxide from a riparian woodland ecosystem. Agricultural and Forest Meteorology 122: 65–84.
- 18. Tognetti R, Giovannelli A, Lavini A, Morelli G, Fragnito F, et al. (2009) Assessing environmental controls over conductances through the soil-plant-atmosphere continuum in an experimental olive tree plantation of southern Italy. Agricultural and Forest Meteorology 149: 1229–1243.
- 19. Kochendorfer J, Castillo EG, Haas E, Oechel WC, Paw UKT (2011) Net ecosystem exchange, evapotranspiration and canopy conductance in a riparian forest. Agricultural and Forest Meteorology 151: 544–553.
- 20. Li Z, Zhang Y, Wang S, Yuan G, Yang Y, et al. (2010) Evapotranspiration of a tropical rain forest in Xishuangbanna, southwest China. Hydrological Processes 24: 2405–2416.
- 21. Stoy PC, Katul GG, Siqueira MBS, Juang J-Y, Novick KA, et al. (2006) Separating the effects of climate and vegetation on evapotranspiration along a successional chronosequence in the southeastern US. Global Change Biology 12: 2115–2135.
- 22. Lundquist JD, Loheide SP (2011) How evaporative water losses vary between wet and dry water years as a function of elevation in the Sierra Nevada, California, and critical factors for modeling. Water Resources Research 47.
- 23. Kelliher FM, Hollinger DY, Schulze ED, Vygodskaya NN, Byers JN, et al. (1997) Evaporation from an eastern Siberian larch forest. Agricultural and Forest Meteorology 85: 135–147.
- 24. Li J, Liu Y, Yang X, Li J (2006) Studies on water-vapor flux characteristic and the relationship with environmental factors over a planted coniferous forest in Qianyanzhou Station. Acta Ecologica Sinica 26: 2449–2456.
- 25. Alfieri JG, Blanken PD, Yates DN, Steffen K (2007) Variability in the environmental factors driving evapotranspiration from a grazed rangeland during severe drought conditions. Journal of Hydrometeorology 8: 207–220.
- 26. Maximov T, Ohta T, Dolman AJ (2008) Water and energy exchange in East Siberian forest: A synthesis. Agricultural and Forest Meteorology 148: 2013–2018.
- 27. Monteith J. Evaporation and environment; 1965. p. 4.
- 28. Bernier PY, Bartlett P, Black TA, Barr A, Kljun N, et al. (2006) Drought constraints on transpiration and canopy conductance in mature aspen and jack pine stands. Agricultural and Forest Meteorology 140: 64–78.
- 29. Bucci SJ, Scholz FG, Goldstein G, Hoffmann WA, Meinzer FC, et al. (2008) Controls on stand transpiration and soil water utilization along a tree density gradient in a Neotropical savanna. Agricultural and Forest Meteorology 148: 839–849.
- 30. Pejam MR, Arain MA, McCaughey JH (2006) Energy and water vapour exchanges over a mixedwood boreal forest in Ontario, Canada. Hydrological Processes 20: 3709–3724.
- 31. Wilson KB, Baldocchi DD (2000) Seasonal and interannual variability of energy fluxes over a broadleaved temperate deciduous forest in North America. Agricultural and Forest Meteorology 100: 1–18.
- 32. Baldocchi DD, Hincks BB, Meyers TP (1988) Measuring biosphere-atmosphere exchanges of biologically related gases with micrometeorological methods. Ecology 1331–1340.
- 33. Baldocchi D (2008) TURNER REVIEW No. 15. ‘Breathing’ of the terrestrial biosphere: lessons learned from a global network of carbon dioxide flux measurement systems. Australian Journal of Botany 56: 1.
- 34. Tanaka N, Kume T, Yoshifuji N, Tanaka K, Takizawa H, et al. (2008) A review of evapotranspiration estimates from tropical forests in Thailand and adjacent regions. Agricultural and Forest Meteorology 148: 807–819.
- 35. Kume T, Tanaka N, Kuraji K, Komatsu H, Yoshifuji N, et al. (2011) Ten-year evapotranspiration estimates in a Bornean tropical rainforest. Agricultural and Forest Meteorology 151: 1183–1192.
- 36. Zha TS, Barr AG, van der Kamp G, Black TA, McCaughey JH, et al. (2010) Interannual variation of evapotranspiration from forest and grassland ecosystems in western canada in relation to drought. Agricultural and Forest Meteorology 150: 1476–1484.
- 37. Ohta T, Maximov TC, Dolman AJ, Nakai T, van der Molen MK, et al. (2008) Interannual variation of water balance and summer evapotranspiration in an eastern Siberian larch forest over a 7-year period (1998–2006). Agricultural and Forest Meteorology 148: 1941–1953.
- 38. Xue B-L, Komatsu H, Kumagai To, Kotani A, Otsuki K, et al. (2012) Interannual variation of evapotranspiration in an eastern Siberian larch forest. Hydrological Processes 26: 2360–2368.
- 39. Kosugi Y, Takanashi S, Tanaka H, Ohkubo S, Tani M, et al. (2007) Evapotranspiration over a Japanese cypress forest. I. Eddy covariance fluxes and surface conductance characteristics for 30 years. Journal of Hydrology 337: 269–283.
- 40. Komatsu H, Cho J, Matsumoto K, Otsuki K (2012) Simple modeling of the global variation in annual forest evapotranspiration. Journal of Hydrology 420–421: 380–390.
- 41. Oishi AC, Oren R, Novick KA, Palmroth S, Katul GG (2010) Interannual Invariability of Forest Evapotranspiration and Its Consequence to Water Flow Downstream. Ecosystems 13: 421–436.
- 42. Park H, Yamazaki T, Yamamoto K, Ohta T (2008) Tempo-spatial characteristics of energy budget and evapotranspiration in the eastern Siberia. Agricultural and Forest Meteorology 148: 1990–2005.
- 43. Zhang W, Wang H, Yang F, Yi Y, Wen X, et al. (2011) Underestimated effects of low temperature during early growing season on carbon sequestration of a subtropical coniferous plantation. Biogeosciences 8: 1667–1678.
- 44. Marcolla B, Cescatti A, Manca G, Zorer R, Cavagna M, et al. (2011) Climatic controls and ecosystem responses drive the inter-annual variability of the net ecosystem exchange of an alpine meadow. Agricultural and Forest Meteorology 151: 1233.
- 45. Luyssaert S, Janssens IA, Sulkava M, Papale D, Dolman AJ, et al. (2007) Photosynthesis drives anomalies in net carbon-exchange of pine forests at different latitudes. Global Change Biology 13: 2110–2127.
- 46. Hui DF, Luo YQ, Katul G (2003) Partitioning interannual variability in net ecosystem exchange between climatic variability and functional change. Tree Physiology 23: 433–442.
- 47. Wang Y, Li Q, Wang H, Wen X, Yang F, et al. (2011) Precipitation frequency controls interannual variation of soil respiration by affecting soil moisture in a subtropical forest plantation. Canadian Journal of Forest Research 41: 1897–1906.
- 48. Wen X, Wang H, Wang J, Yu G, Sun X (2010) Ecosystem carbon exchanges of a subtropical evergreen coniferous plantation subjected to seasonal drought, 2003–2007. Biogeosciences 7: 357–369.
- 49. Wen X-F, Yu G-R, Sun X-M, Li Q-K, Liu Y-F, et al. (2006) Soil moisture effect on the temperature dependence of ecosystem respiration in a subtropical Pinus plantation of southeastern China. Agricultural and Forest Meteorology 137: 166–175.
- 50. Wang Y, Wang Z-L, Wang H, Guo C, Bao W (2012) Rainfall pulse primarily drives litterfall respiration and its contribution to soil respiration in a young exotic pine plantation in subtropical China. Canadian Journal of Forest Research 42: 657–666.
- 51. McPhaden MJ (2004) Evolution of the 2002/03 El Niño. Bulletin of the American Meteorological Society 85: 677.
- 52. Kim WM, Yeh SW, Kim JH, Kug JS, Kwon MH (2011) The unique 2009–2010 El Niño event: A fast phase transition of warm pool El Niño to La Niña. Geophysical Research Letters 38: L15809.
- 53. Zhu Z, Sun X, Zhou Y, Xu J, Yuan G, et al. (2005) Correcting method of eddy covariance fluxes over non-flat surfaces and its application in ChinaFLUX. Science in China Series D: Earth Sciences 48.
- 54. Webb EK, Pearman GL, Leuning R (1980) Correction of flux measurement for density effects due to heat and water vapor transfer. Quarterly Journal of Royal Meteorological Society 106: 85–100.
- 55. Zhang LM, Yu GR, Sun XM, Wen XF, Ren CY, et al. (2006) Seasonal variation of carbon exchange of typical forest ecosystems along the eastern forest transect in China. Science in China Series D-Earth Sciences 49: 47–62.
- 56. Reichstein M, Falge E, Baldocchi D, Papale D, Aubinet M, et al. (2005) On the separation of net ecosystem exchange into assimilation and ecosystem respiration: review and improved algorithm. Global Change Biology 11: 1424–1439.
- 57. Falge E, Baldocchi D, Olson R, Anthoni P, Aubinet M, et al. (2001) Gap filling strategies for defensible annual sums of net ecosystem exchange. Agricultural and Forest Meteorology 107: 43–69.
- 58. Yu GR, Song X, Wang QF, Liu YF, Guan DX, et al. (2008) Water-use efficiency of forest ecosystems in eastern China and its relations to climatic variables. New Phytologist 177: 927–937.
- 59. Wilson KB, Hanson PJ, Mulholland PJ, Baldocchi DD, Wullschleger SD (2001) A comparison of methods for determining forest evapotranspiration and its components: sap-flow, soil water budget, eddy covariance and catchment water balance. Agricultural and Forest Meteorology 106: 153–168.
- 60. Admiral SW, Lafleur PM, Roulet NT (2006) Controls on latent heat flux and energy partitioning at a peat bog in eastern Canada. Agricultural and Forest Meteorology 140: 308–321.
- 61. Saugier B, Katerji N (1991) Some Plant Factors Controlling Evapotranspiration. Agricultural and Forest Meteorology 54: 263–277.
- 62. Yoshida M, Ohta T, Kotani A, Maximov T (2010) Environmental factors controlling forest evapotranspiration and surface conductance on a multi-temporal scale in growing seasons of a Siberian larch forest. Journal of Hydrology 395: 180–189.
- 63. Iida Si, Tanaka T, Sugita M (2006) Change of evapotranspiration components due to the succession from Japanese red pine to evergreen oak. Journal of Hydrology 326: 166–180.
- 64. Shuttleworth WJ, Wallace J (1985) Evaporation from sparse crops-an energy combination theory. Quarterly Journal of the Royal Meteorological Society 111: 839–855.
- 65. Penman HL (1948) Natural evaporation from open water, bare soil and grass. Proceedings of the Royal Society of London Series A Mathematical and Physical Sciences 193: 120–145.
- 66. Kume T, Komatsu H, Kuraji K, Suzuki M (2008) Less than 20-min time lags between transpiration and stem sap flow in emergent trees in a Bornean tropical rainforest. Agricultural and Forest Meteorology 148: 1181–1189.
- 67. McCaughey JH, Pejam MR, Arain MA, Cameron DA (2006) Carbon dioxide and energy fluxes from a boreal mixedwood forest ecosystem in Ontario, Canada. Agricultural and Forest Meteorology 140: 79–96.
- 68. Jassal RS, Black TA, Spittlehouse DL, Brummer C, Nesic Z (2009) Evapotranspiration and water use efficiency in different-aged Pacific Northwest Douglas-fir stands. Agricultural and Forest Meteorology 149: 1168–1178.
- 69. Si JH, Feng Q, Zhang XY, Chang ZQ, Su YH, et al. (2007) Sap flow of Populus euphratica in a desert riparian forest in an extreme arid region during the growing season. Journal of Integrative Plant Biology 49: 425–436.
- 70. Chang X, Zhao W, Zhang Z, Su Y (2006) Sap flow and tree conductance of shelter-belt in arid region of China. Agricultural and Forest Meteorology 138: 132–141.
- 71. Wu J, van der Linden L, Lasslop G, Carvalhais N, Pilegaard K, et al. (2012) Effects of climate variability and functional changes on the interannual variation of the carbon balance in a temperate deciduous forest. Biogeosciences 9: 13–28.
- 72. Richardson AD, Hollinger DY, Aber JD, Ollinger SV, Braswell BH (2007) Environmental variation is directly responsible for short-but not long-term variation in forest-atmosphere carbon exchange. Global Change Biology 13: 788–803.
- 73. Butler D (1976) Estimation of the transpiration rate in an apple orchard from net radiation and vapour pressure deficit measurements. Agricultural Meteorology 16: 277–289.
- 74. Ryu Y, Baldocchi DD, Ma S, Hehn T (2008) Interannual variability of evapotranspiration and energy exchange over an annual grassland in California. Journal of Geophysical Research-Atmospheres 113.
- 75. Granier A, Reichstein M, Breda N, Janssens IA, Falge E, et al. (2007) Evidence for soil water control on carbon and water dynamics in European forests during the extremely dry year: 2003. Agricultural and Forest Meteorology 143: 123–145.
- 76. Saleska SR (2003) Carbon in Amazon Forests: Unexpected Seasonal Fluxes and Disturbance-Induced Losses. Science 302: 1554–1557.
- 77. Tu J, Liu Q, Li H, LIN Y (2008) Dynamics of Sap Flow of Pinus elliottii and Its Relevant Factors in Rapid Growth Season at Qianyanzhou Experimental Station of Jiangxi Province. Scientia Silvae Sinicae 44: 46–51.
- 78. Hollinger DY, Aber J, Dail B, Davidson EA, Goltz SM, et al. (2004) Spatial and temporal variability in forest-atmosphere CO2 exchange. Global Change Biology 10: 1689–1706.
- 79. Granier A, Biron P, Lemoine D (2000) Water balance, transpiration and canopy conductance in two beech stands. Agricultural and Forest Meteorology 100: 291–308.
- 80. Wang B (2002) Rainy Season of the Asian-Pacific Summer Monsoon*. Journal of Climate 15: 386–398.
- 81. Wang B, Wu R, Lau K (2001) Interannual Variability of the Asian Summer Monsoon: Contrasts between the Indian and the Western North Pacific-East Asian Monsoons*. Journal of Climate 14: 4073–4090.
- 82. Ding Y, Chan JCL (2005) The East Asian summer monsoon: an overview. Meteorology and Atmospheric Physics 89: 117–142.
- 83. Richardson AD, Andy Black T, Ciais P, Delbart N, Friedl MA, et al. (2010) Influence of spring and autumn phenological transitions on forest ecosystem productivity. Philosophical Transactions of the Royal Society B: Biological Sciences 365: 3227–3246.
- 84. Ham JM, Knapp AK (1998) Fluxes of CO2, water vapor, and energy from a prairie ecosystem during the seasonal transition from carbon sink to carbon source. Agricultural and Forest Meteorology 89: 1–14.
- 85. Faver KL, Otoole JC, Krieg DR (1989) Short-Term Estimation of Sorghum Evapotranspiration from Canopy Temperature. Agricultural and Forest Meteorology 48: 175–183.
- 86. Betts AK, Goulden M, Wofsy S (1999) Controls on evaporation in a boreal spruce forest. Journal of Climate 12: 1601–1618.
- 87. Odum EP (1969) Strategy of ecosystem development. Science 164: 262–270.
- 88. Wei H, He H, Liu M, Zhang L, Yu G, et al. (2012) Modeling evapotranspiration and its components in Qianyanzhou plantation based on remote sensing data. Journel of Natural Resources 25: 778–789.
- 89. Zhang L, Dawes W, Walker G (2001) Response of mean annual evapotranspiration to vegetation changes at catchment scale. Water Resources Research 37: 701–708.
- 90. Bader MKF, Siegwolf R, Korner C (2010) Sustained enhancement of photosynthesis in mature deciduous forest trees after 8 years of free air CO2 enrichment. Planta 232: 1115–1125.
- 91. Piao SL, Friedlingstein P, Ciais P, Zhou LM, Chen AP (2006) Effect of climate and CO2 changes on the greening of the Northern Hemisphere over the past two decades. Geophysical Research Letters 33.
- 92. Taylor G, Tallis MJ, Giardina CP, Percy KE, Miglietta F, et al. (2008) Future atmospheric CO2 leads to delayed autumnal senescence. Global Change Biology 14: 264–275.
- 93. Urban O, Hrstka M, Zitova M, Holisova P, Sprtova M, et al. (2012) Effect of season, needle age and elevated CO2 concentration on photosynthesis and Rubisco acclimation in Picea abies. Plant Physiology and Biochemistry 58: 135–141.
- 94. Cernusak LA, Winter K, Dalling JW, Holtum JAM, Jaramillo C, et al. (2013) Tropical forest responses to increasing atmospheric CO2: current knowledge and opportunities for future research. Functional Plant Biology 40: 531–551.
- 95. Holtum JAM, Winter K (2010) Elevated CO2 and forest vegetation: more a water issue than a carbon issue? Functional Plant Biology 37: 694–702.
- 96. Dai LJ, Cui WH, Jiang YB (2012) Temporal and spatial distribution of tropospheric carbon dioxide from 2003 to 2010 in China. Ecology and Environmental Sciences 21(7): 1266–1270.