4. DISCUSSIONS
4.1. Seasonal variations of GPP, RES and NEE
The result showed that GDD was the primary control factor affecting the change of monthly NEE and GPP in the alpine wetland (Figs. 3 and 4). In addition, to a certain extent, Ts had a relatively strong control over monthly GPP and monthly NEE in the alpine wetland (Figs. 3). In the alpine ecosystem of the QTP, photosynthetic radiation is relatively strong, and vegetation contains a large amount of aboveground biomass. As a result, temperature in the growing season has a strong impact on photosynthesis and carbon sequestration of vegetation in alpine wetland ecosystem (Marcolla et al., 2011; Shen et al., 2015). Furthermore, the accumulation effect of temperature had important effects on dormancy, leaf phenology and late growth of plants (Kato et al., 2006; Ueyama et al., 2013). Therefore, the metabolism of alpine plants, especially in response to temperature, had rich phenotypic plasticity (Körner, 1999). In addition, temperature could affect microbial activity, enzyme activity and decomposition of soil organic matter, so that the thermal condition was indirectly positively correlated with the supply of available nutrients in the soil of the ecosystem, thus making temperature a key control factor for the photosynthetic and carbon fixation capacity of alpine vegetation in the wetland of QTP (Street et al., 2007).
Ts played the most important role in the change of monthly RES in the alpine wetland of QTP (Fig. 3b). Many studies have shown that low soil temperature limits the soil enzyme activity and soil microbial biomass, and had an important impact on soil respiration and carbon balance in the alpine ecosystem (Van den Bos., 2003). This might be because the alpine wetland contained a large amount of soil organic matter, which was extremely sensitive to the change of soil temperature. Higher soil temperature could stimulate the activity of microorganisms and the decomposition of soil organic matter, so that soil temperature was the dominant control factor of RES (Moore et al., 1997; Updegraff et al., 2001). Furthermore, during the growing season, monthly NEE had significant correlation with monthly GPP (r2= 0.88, p < 0.001) and monthly RES (r2= 0.44, p < 0.001). Therefore, GPP played a more controlling role on the variation of NEE compared with RES during the growing season, which was consistent with other studies on temperature-restricted ecosystems (Groendahl et al., 2007; Ueyama et al., 2013).
4.2. Inter-annual variations of GPP, RES and NEE
There was a significant positive correlation between Ts of the non-growing season and annual GPP of the following year (p <0.05). This might be because the higher Ts in the non-growing season stimulated the activities of microorganisms, enhanced the soil enzyme activity, accelerated the decomposition of litters and soil organic matter, and thus produced rich nutrients in the soil, promoted the growth of vegetation in the following year, and improved the photosynthesis and carbon fixation capacity of vegetation in alpine wetland (Yu et al., 2003; Shen et al., 2011). In addition, the increase of Ts in the non-growing period promoted the melting of snow and ice, thus reducing the surface albedo, which was conducive to the early re-greening of vegetation and ultimately improved the photosynthetic productivity of vegetation in alpine wetland (Li et al., 2014; Reverter et al., 2010; Yu et al., 2003; Zhang et al., 2013). Meanwhile, the findings suggested that annual GPP had significant negative correlation with the growing season PPT (p <0.05). One possible was that PPT raised the water level of the wetland, thereby limiting the diffusion of oxygen in the atmosphere to the soil, thus inhibiting the activity of microorganisms and reducing the decomposition rate of soil organic matter, thus reducing the nutrients in the soil, and consequently reducing the photosynthetic and carbon fixation capacity of vegetation in alpine wetland (Chimner and Cooper., 2003).
Annual RES had a significant positive correlation with the non-growing season Ts (r2 = 0.44, p = 0.037), which had no significant correlation with growing season Ts (r2 = 0.006, p = 0.83) and annual Ts (r2 = 0.01, p = 0.74). Furthermore, the non-growing RES had significant correlation with non-growing season Ts (r2 = 0.45, p = 0.034). This might due to the long cold weather, the wetland contained a large amount of soil carbon that was extremely sensitive to temperature changes (Hao et al., 2011; Kang et al., 2014). The higher temperature during the non-growing season was conducive to the decomposition of soil organic matter and the improvement of soil respiration (Zhao et al., 2005; Kang et al., 2014). These results also suggested that the non-growing RES was crucial for annual RES. In addition, the non-growing RES had significant correlation with non-growing season PPT (r2 = 0.54, p = 0.015). This might due to rainfall and snowfall cause snow and ice cover on the surface of wetland, which had a certain insulation effect on soil temperature and was conducive to promoting soil respiration (Hao et al., 2011; Song et al., 2018; Zhang et al., 2008).
Annual PPT was significant positive correlated with annual NEE (Table 1). This might be due to the increase of PPT in the growing season would reduce GPP, while the increase of PPT in the non-growing season would increase RES. As a result, the increase of inter-annual PPT leaded to accelerate the carbon loss in wetland ecosystem. The increase of the non-growing season temperature promotes soil respiration, lead to have significant positive correlation with annual NEE. In this study, the non-growing season RES was significantly correlated with annual RES and NEE (p <0.05), which might be because the alpine wetland ecosystem on the QTP had a long non-growing period due to high altitude and low temperature, and the alpine wetland was covered by snow and ice for nearly half one year, so CO2 emission in the non-growing season had a crucial impact on the variations of annual CO2 fluxes (Hao et al., 2011; Song et al., 2018; Zhao et al., 2005). Our results also suggested that under the context of global warming, especially the increase of temperature in the non-growing season of the QTP, thus alpine wetland will might have a positive feedback on carbon loss caused by climate warming. In addition, the water table depth in the alpine wetland could also affect the soil thermal conductivity, thus impact on the change of soil temperature. In the context of climate change with increasing temperature and precipitation on the QTP, which is likely to through the interactive effects of Ts and water table depth on CO2 production (Gao et al., 2012; Zhu et al., 2013). Further research is needed to make this clearly.