Introduction

Terrestrial ecosystems are major carbon reservoirs and important sources or sinks of greenhouse gases (GHG) (IPCC et al. 2013). Wetlands have higher carbon storage than other ecosystems, accounting for about 20–30% of the earth’s soil carbon (Mitsch et al. 2013). Stored carbon is decomposed into CO2 and CH4 through soil respiration and methanogenesis (Li et al. 2018). The wetland area in Northeast China is about 1.06 × 104 km2, accounting for about 16% of total wetland area in China (Di et al. 2004). Activities such as blind reclamation, overgrazing, and construction of artificial reservoirs and ponds have exacerbated the salinization process of wetlands in Northeast China, and more than two-thirds of the wetlands in the Songnen Plain have experienced secondary salinization. As a typical inland alkaline wetland in Northeast China, Zhalong wetland is very sensitive to climate change. In the past 50 years, the annual and seasonal average temperature of Zhalong Wetland has shown an upward trend, and the rainfall has decreased. This has led to an increase in evapotranspiration in the Zhalong Wetland, and the water shortage in the wetland has become increasingly serious. The annual evapotranspiration capacity of Zhalong Wetland is 1506.2 mm (The Ministry of Forestry of the People’s Republic of China 1997). Soil alkalization caused by massive evaporation of inland alkaline wetlands is suitable for the growth of methanogenic bacteria and is an important source of methane flux (Liu et al. 2019).

Seasonal freeze–thaw is an important meteorological event in Northeast China, and it impacts the soil ecology and GHG flux characteristics in this area, controls the transfer and transform of the soil nutrient greatly, and affects the chemistry process of mass and energy cycles in the global ecosystems (Liu et al. 2019). Currently, global warming and human activities are changing the structure and function of wetland ecosystems. They can affect GHG fluxes during freeze–thaw periods by water levels, soil properties, and the freeze–thaw process. Water levels are an important determinant of GHG fluxes in a spring-fed forested wetland (Koh et al. 2009). The high water levels in wetlands lead to a hypoxic environment in the soil, which inhibits the autotrophic respiration of plants leading to lower CO2 fluxes (Koh et al. 2009). The high soil moisture obviously enhanced soil microbial heterotrophic activities and soil microbial respiration in low tidal flats (Hu et al. 2016). In studies of alpine grassland ecosystems on the Tibetan Plateau, the intensity of CO2 flux reduction decreases as the water levels rises (Zhao et al. 2017). In alpine peatlands, CO2 fluxes increased significantly with decreasing water levels (Zhang et al. 2020). While CO2 fluxes were not affected by changes in water levels during a 10-day anaerobic incubation (Toczydlowski et al. 2020), and CH4 fluxes were mainly limited by water levels, as CH4 fluxes require an anaerobic environment created by high water levels (Natali 2015). In alpine peatlands, decreasing water levels reduce CH4 fluxes (Zhang et al. 2020). However, decreasing water levels in chronically flooded wetlands increased CH4 fluxes (Ding et al. 2002). In general, the effect of water levels on CO2 and CH4 fluxes varies according to region differences. Freezing soil can form a better anaerobic environment with unpredictable effects on GHG production (Xiangwen et al. 2019). Spring freeze–thaw cycles inhibit CO2 fluxes in forests and agroecosystems (Kurganova & Gerenyu 2015). In a German farmland ecosystem, CO2 fluxes increased during the spring freeze–thaw period instead (Sehy et al. 2004). CH4 fluxes were also suppressed under spring freeze–thaw cycles in farmland systems of Northern China (Liang et al. 2007), while no significant effect of spring freeze–thaw cycles on CH4 fluxes in the Zhalong wetland (Liu et al. 2019). The variation of soil properties during the freeze–thaw period could dramatically affect GHG fluxes. The freeze–thaw cycling process during the spring freeze–thaw period leads to a pulsed release of GHG fluxes. Nevertheless, the fluctuation of GHG fluxes and the effect of water levels on GHG sources and sinks have been rarely reported during the autumn freeze–thaw period.

The objectives of this study are to reveal the characteristics of CO2 and CH4 fluxes in Zhalong wetlands during the autumn freeze–thaw period and their relationships with water levels. Daily variation of CO2 and CH4 fluxes in Zhalong saline wetlands reveals the key environmental drivers causing the differences in GHG fluxes. During the autumn freeze–thaw period, the rainfall in Zhalong wetland decreases, and the water level was lower than that in the growing season. This study contributes to a comprehensive and in-depth understanding of the characteristics of CO2 and CH4 fluxes in saline wetlands at different water levels during the autumn freeze–thaw period and clarifies the response of CO2 and CH4 fluxes to water levels. This can improve the understanding of GHG sources and sinks in inland saline wetlands and contribute to the construction of regional and even global climate models. The value of the contribution of storage and drainage processes to saline wetland GHG fluxes will be accurately evaluated in the global warming process.

Material and methods

Site description

Zhalong Wetland Nature Reserve (46°52′–47°32′N, 123°47′–124°37′E), a saline wetland, is located in the Songnen Plain, Heilongjiang Province, China. The wetland has a total area of approximately 2100 km2, 80% of which are reeds (Phragmites australis), swampy wetlands. It has a mid-temperate climate, a mean annual air temperature of 3.9 °C, a freezing period of 7 months, and a mean annual precipitation of 420 mm (Gao et al. 2018). The wetland area is low-lying and flat, with numerous marshes distributed. High water levels, poor drainage, and high evaporation lead to soil salinization.

Three water levels points were chosen as study points. The water level above ground at high flooded (HF) point was 7.9–24.8 cm, the main vegetation type is Phragmites australis (Table 1). And the water level above ground at middle flooded (MF) point was 1.0–9.5 cm; the main vegetation type is Phragmites australis (Table 1). Dry (D) point had no surface water, and the main vegetation types are Axonopus compressus, Medicago Sativa Linn, and Imperata cylindrica (Table 1). The salinities in HF, MF, and D were 87.7 ± 0.08, 101.1 ± 0.12, and 61.8 ± 0.07 mg L−1, respectively (Table 1). The gas was collected from 14 October to 23 November, 2021.

Table 1 The water level and dominant vegetation of the three points
Full size table

Greenhouse gas flux measurements

The static closed-chamber technique was applied to measure the GHG flux rate (Liu et al. 2019). The chamber consisted of two parts: an open-bottom chamber (50 cm × 50 cm × 50 cm) and a permanent collar (50 cm × 50 cm × 20 cm high). The cubic chambers were made of polypropylene, insulated with expanded polystyrene to minimize temperature changes, and equipped with a battery-driven fan for air circulation through the chambers. During the experimental period, the gutter of the base collar was filled with water to form a water seal. Gas samples were collected every 2 days from 14 October to 6 November, and every 4 days from 11 to 23 November.

Gas sampling was conducted from 9:00 a.m. to 11:00 a.m. (Liu et al. 2019). A full-day sampling started on 23 October, from 7:30 to 19:30 (Xu et al. 2017). A syringe equipped with a three-way screw plug was used to collect 25 mL of gas into a 12-mL vacuum gas bottle at 0, 15, 30, and 45 min (Liu et al. 2019). The concentrations of CH4 and CO2 were analyzed by a gas chromatograph (Agilent 7890A) equipped with a methanizer (Ni-catalyst at 350 °C) and a flame ionization detector. The detector temperature was 300 °C, the hydrogen flow was 60 mL min−1, and the air flow was 300 mL min−1. The separation of CH4 and CO2 was carried out on a 60/80 mesh 13XMS column with a length of 2 m and an inner diameter of 2 mm. The oven temperature was 55 °C, and the carrier gas was high-purity nitrogen at a flow rate of 20 mL min−1. The soil temperatures at 0 cm, 5 cm, 10 cm, and 15 cm were measured by a portable digital thermometer (JM624, Imin Instruments Ltd., Tianjin, China).

Soil sampling and analysis

Soil samples were collected every 4–5 days from 14 October to 11 November. Soil samples were collected at top (0–10 cm) and bottom (10–20 cm) soil layers, then were screened with a 2-mm sieve. The fresh soil was extracted with 1 mol L−1 KCl. NO3−-N content and NH4+-N content were determined using a continuous flow analyzer (SealAnalyticalAA3, Norderstedt, Germany). Soil moisture was measured using the desiccation method. The air-dried soil was used to measure total soil organic carbon, pH, soil salinity, and moisture content (Liu et al. 2019). All experiments were repeated three times.

Statistical analysis

The GHG flux calculation method and data analysis method refer to the previous study (Gao et al. 2019). The data were statistically and analytically analyzed using one-way ANOVA (one-way analysis of variance). It is mainly used to analyze the correlation between the parallel sampling points of each plot and environmental factors and establish a linear model to analyze the significance of the correlation, so as to obtain the interpretation degree of environmental factors to the greenhouse gas emission flux. Fisher’s least significant difference method (LSD) was used for multiple comparisons (α = 0.05), and the data in the graphs were means ± standard deviations.

Results

Diurnal changes of CO2 and CH4 fluxes

The diurnal air temperature varied significantly in the three points. Compared to CH4, the CO2 fluxes are more sensitive to temperature, in agreement with the surface air temperature variation (Fig. 1). The air temperature was relatively high at noon, and so did CO2 fluxes which peaked at 11:30 and were 306.5 (MF), 220.7 (HF), and 173.8 mg m−2 h−1 (D). The lowest air temperature was 3.3 °C, 3.6 °C, and 1.6 °C at MF, D, and HF, corresponding to CO2 fluxes (132.3, 98.6, 191.9 mg m−2 h−1) and CH4 fluxes (2.9, 0.0, 10.3 mg m−2 h−1).

Fig. 1
figure 1

Diurnal variation of CO2 and CH4 fluxes and air temperature from 7:30 to 19:30 during the autumn freeze–thaw period in three types of points. a CO2 fluxes, b CH4 fluxes (n = 3)

Full size image

In HF, the diurnal variation of CO2 fluxes ranged from 167.3 to 220.7 mg m−2 h−1, which maintained a small variation range due to the thermal insulation effect of water. In MF, the CO2 fluxes ranged from 84.6 to 306.5 mg m−2 h−1, which ranged from 89.0 to 173.8 mg m−2 h−1 in D (Fig. 1a). The CH4 fluxes changed little with temperature in D (3.3 mg m−2 h−1) (Fig. 1b). The peak of CH4 fluxes of MF and D appeared at the same time (9.2 mg m−2 h−1). Due to water insulation, the peak CH4 fluxes appeared later in HF (16.8 mg m−2 h−1).

In MF, CO2 fluxes were more sensitive to temperature and consistent with surface air temperature changes (R2 = 0.962, P < 0.01) (Fig. 2). In D, CO2 fluxes were significantly positively correlated with 0-cm (R2 = 0.695, P < 0.01) and 5-cm (P < 0.05) soil temperature (Fig. 2).

Fig. 2
figure 2

Model of relationship between the diurnal variation of CO2 fluxes versus temperature in MF and D. T0: 0-cm soil temperature, T5: 5-cm soil temperature

Full size image

Characteristics of CO2 and CH4 fluxes changes

The CO2 fluxes of the three points all showed a fluctuating downward trend, which represented the source of CO2 release (Fig. 3a). The range of CO2 fluxes was 11.3–179.9 mg m−2 h−1, 7.6–143.6 mg m−2 h−1, and 8.1–164.2 mg m−2 h−1 in HF, MF, and D, respectively (Fig. 3a). The average and cumulative CO2 fluxes of HF (83.9 mg m−2 h−1 and 1594.0 kg hm−2) > D (78.0 mg m−2 h−1 and 1482.3 kg hm−2) > MF (66.2 mg m−2 h−1 and 1258.4 kg hm−2) (Fig. 3b).

Fig. 3
figure 3

CO2 and CH4 fluxes during the autumn freeze–thaw period. a CO2 fluxes, b Cumulative and average CO2 fluxes, c CH4 fluxes, d Cumulative and average CH4 fluxes

Full size image

The ranges of CH4 fluxes were 3.1–11.9, 2.0–5.4, and 0.0–1.8 mg m−2 h−1 in HF, MF, and D. The average fluxes are 7.6, 2.5, and 0.3 mg m−2 h−1 in HF, MF, and D. The trends of CH4 fluxes were basically the same, showing a gradual decrease in MF and D (Fig. 3c). While HF showed an increasing trend, first reached a peak and then decreased slightly (Fig. 3c). The CH4 flux changes coincide with the water levels (Table 3). The average and cumulative CH4 fluxes of three points varied widely, showing that HF (7.6 mg m−2 h−1 and 143.6 kg hm−2) > MF (2.5 mg m−2 h−1 and 46.8 kg hm−2) > D (0.3 mg m−2 h−1 and 4.9 kg hm−2) (Fig. 3d). It indicates the effect of vegetation on CH4 fluxes at various water levels.

Correlation of soil CO2 and CH4 fluxes with environmental factors

In HF, CO2 fluxes were only significantly and positively correlated with 0–10 cm soil NH4+-N content (P < 0.05) (Tables 2 and 3). In MF, CO2 fluxes were significantly and positively correlated with air temperature (P < 0.05) and soil temperature (P < 0.01). In D, CO2 fluxes were significantly positively correlated with air temperature (P < 0.05), soil temperature, and 0–10-cm soil moisture content (P < 0.05). Soil NH4+-N content (0–10 cm) could explain 85.5% of the CO2 fluxes in HF (Table 4). Air and soil temperature could explain 24.5–61.1% of the CO2 fluxes in MF and D (Table 4). Soil moisture content (0–10 cm) could explain 98.9% of the CO2 fluxes in D. In both MF and D, the CO2 fluxes decreased with the decreasing temperature. In D, the CO2 fluxes were also influenced by soil moisture content, and the CO2 fluxes could also increase with the increased moisture content (P < 0.05).

Table 2 Temperature and other soil physicochemical characteristics of three points
Full size table
Table 3 Pearson’s correlation between GHG fluxes and environmental factors
Full size table
Table 4 Correlation of CO2 fluxes environmental factors
Full size table

In MF, the CH4 fluxes were significantly and positively correlated with the surface soil temperature (P < 0.05) and with 0–10-cm soil NO3−-N content (P < 0.05). In D, the CH4 fluxes were significantly and positively correlated with 10–20-cm soil moisture content (P < 0.01). Zero-centimeter soil temperature could explain 48.1% of the CH4 fluxes in the MF. The 10–20-cm soil moisture content could explain 95.6% of the CH4 fluxes from D. The factors influencing CH4 fluxes were different in all three points, indicating that CH4 fluxes were related to point types.

Discussion

Influencing factors of CO2 and CH4 diurnal fluxes

The highest CO2 fluxes of the three points were all at 11:30 a.m., earlier than the highest temperature (13:30 a.m.) (Fig. 1). In the wetlands of northern Jiangsu Province, maximum CO2 fluxes were also observed slightly earlier than the maximum temperature in October (Xu et al. 2017). The temperature difference from 7:30 a.m. to 19:30 in MF (11.7 °C) was significantly higher than that in HF (9.9 °C) and D (9.0 °C). Accordingly, the fluctuation of CO2 fluxes was most severe in MF (221.9 mg m−2 h−1) than that in HF (53.5 mg m−2 h−1) and D (84.9 mg m−2 h−1). Due to the multiple influences of plant respiration, plant photosynthesis, soil respiration, and adsorption, the diurnal variation form presented a multimodal pattern (Yuesi et al. 2000).

In this study, the diurnal variation of CH4 fluxes was generally high in the early afternoon. The largest CH4 fluxes were also observed in the early afternoon in wetlands in northern Jiangsu Province and the West Siberian peatlands (Veretennikova & Dyukarev 2017, Xu et al. 2017). Anaerobic decomposition of organic matter is the main factor in fluxes of CH4 (Tsai et al. 2020). Elevated daytime temperatures may promote anaerobic decomposition of organic matter, while lower nighttime temperatures may inhibit the rate of decomposition (Tsai et al. 2020). CH4 is mainly convective transport with higher transport efficiency during the day, while diffusion transport with lower transport efficiency is the dominant model in the dark (Käki et al. 2001; Van Der Nat et al. 1998). In addition, the CH4 accumulated at night is released into the atmosphere by convective transport during the day, which also contributes to the high daytime CH4 fluxes (Turetsky et al. 2014). The three points showed that the CH4 fluxes increased with the deepening of the water levels, and higher CH4 fluxes were also observed from fully submerged soils (Turetsky et al. 2014). Moreover, part of the CH4 is oxidized to CO2 at night when the water levels are low; the rising CH4 bubbles also could be oxidized to CO2 when the water levels are high, so that reducing the CH4 fluxes (Schrier-Uijl et al. 2011).

Influencing factors of the CO2 fluxes during the autumn freeze–thaw period

In HF, CO2 fluxes were significantly positively correlated with 0–10-cm soil NH4+-N content (P < 0.05). The CO2 fluxes in the remaining two sample points are independent of 0–10 cm soil NH4+-N content. Positive correlations between CO2 fluxes and total CO2-equivalent fluxes and soil NH4+-N content were also observed in mangrove wetlands (Chen et al. 2016). Ammonia nitrogen is conducive to the absorption of nutrients and water by plant roots and promotes soil respiration (Ma et al. 2019). In MF and D, CO2 fluxes were significantly positively correlated with soil and air temperature (Tables 2 and 3). There was also a significant positive correlation between CO2 fluxes and soil and air temperatures in the wetlands where reeds were harvested in Zhalong wetland (Liu et al. 2019). The temperature was relatively high in the early stage of the autumn freeze–thaw period. A significant cooling was experienced in early November, resulting in a low point in CO2 fluxes (18.4 mg m−2 h−1, MF; 17.0 mg m−2 h−1, D). Then, there was slight warming, and the CO2 fluxes increased to 94.5 mg m−2 h−1 (MF) and 88.53 mg m−2 h−1 (D). Soil temperature enhances soil respiration by accelerating microbial activity and promoting plant root growth, thereby increasing carbon dioxide fluxes (Tang et al. 2020). Moreover, repeated freeze–thaw caused by temperature changes in autumn could increase soil active organic carbon for microbial use (Oztas &Fayetorbay 2003). Soil temperature at 10–15 cm and 0–10 cm had a greater effect on CO2 fluxes in MF and D, respectively. This suggests that the 10–15-cm and 0–10-cm soil depths are the focal areas for CO2 production at MF and D, respectively. The CO2 fluxes at HF are less controlled by temperature. Only in D, CO2 fluxes were influenced by the 0–10-cm soil moisture content (Tables 2 and 3), which showed a positive correlation. Moisture content impacts soil respiration intensity in agricultural fields and meadows (Cleveland et al. 2010). There was no surface water in D. The oxygen content decreased significantly when the soil moisture content increased; more CO2 is produced under anaerobic conditions than under aerobic conditions (Walz et al. 2018).

The mean and cumulative CO2 fluxes were not significantly different in the three points, indicating that water levels and vegetation had no direct effect on CO2 fluxes, but it can indirectly affect the fluxes of CO2 by changing soil environmental factors (Fig. 3c). The average CO2 fluxes is less than the average flux of CO2 in the northeast permafrost (105.5 mg m−2 h−1) (Gao et al. 2022). The reason might be that the temperature drop reduced the respiration activities of microorganisms and plants during the autumn freeze–thaw period, which in turn reduced the CO2 fluxes. Moreover, the hysteresis between CO2 fluxes and temperature is higher when the soil moisture content is higher (Gaumont-Guay et al. 2009). Higher water levels delay the effect of temperature on CO2 fluxes when comparing the CO2 fluxes in lower water levels.

Influencing factors of the CH4 fluxes during the autumn freeze–thaw period

In MF, there was a positive correlation between NO3−-N content of 0–10-cm soil with CH4 fluxes. The CH4 fluxes in the remaining two points are independent of NO3−-N content of 0–10-cm soil. In the anaerobic environment of wetland, there are associated CH4 oxidizing bacteria that can use nitrate as an electron acceptor to oxidize CH4, resulting in lower CH4 fluxes (Ettwig et al. 2016; Shen et al. 2018). The NO3−-N can promote root growth and root secretion function of marsh plants, and the effective substrate for CH4-producing bacteria in the roots of marsh plants will increase accordingly (Wang et al. 2012). This promotes the metabolic activity of CH4-producing bacteria and increases CH4 fluxes (Wang et al. 2012).

In natural environments, soil moisture content is also an important factor for CH4 production. Soil moisture can increase the activity of CH4-producing bacteria and provide anaerobic conditions (Ma et al. 2012). The significant positive correlation between CH4 fluxes with 10–20-cm soil moisture content only in D. The CH4 fluxes in the remaining two points are independent of soil moisture content. CH4 was presented as a flux source in HF, MF, and D with mean values of 7.5, 2.5, and 0.3 mg m−2 h−1, respectively. It is much larger than the maximum value of CH4 flux in northeast paddy fields (0.1 mg m−2 h−1) (Zhang et al. 2017). There is a significant positive correlation between water level and CH4 fluxes. The prolonged overwater condition created an anaerobic environment that is conducive to the growth and reproduction of anaerobic CH4-producing bacteria, leading to an increase in CH4 fluxes (Turetsky et al. 2014). The water levels were the key factor affecting the type of methanogens in the soil. In the wetlands of the Qinghai-Tibet Plateau, Methylobacter (90.0%) of type I methanotrophs were overwhelmingly dominant in the high water level, while Methylocystis (53.3%) and Methylomonas (42.2%) belonging to types II and I methanotrophs were the predominant groups in the low water level (Cui et al. 2018). CH4-producing bacteria are strictly anaerobic, and CH4-producing bacteria populations are larger in water covered lands (Šťovíček et al. 2017). The higher the water levels and the longer the inundation time, the higher the CH4 fluxes (Henneberg et al. 2016; Sha et al. 2015). Greater CH4 fluxes in water-covered wetlands.

Conclusion

Water levels affect the physicochemical properties of wetland soil during the autumn freeze–thaw period. However, water levels could not directly significantly affect the cumulative CO2 fluxes, but it can indirectly affect the fluxes of CO2 by changing soil environmental factors. CO2 fluxes decreased with decreasing air and soil temperatures in MF and D. While CO2 fluxes were positively correlated with 0–10-cm soil NH4+-N content in HF. The water level significantly affects the CH4 fluxes, the higher the water level, and the higher CH4 fluxes. Wetlands at lower water levels (below 10-cm water levels) did not show much difference in CH4 fluxes compared to drylands. CH4 fluxes increased with decreasing water levels in HF. In MF, CH4 fluxes were positively correlated with surface temperature and 0–10-cm soil NO3−-N content. In D, CH4 fluxes were positively correlated with 10–20-cm soil moisture content. All in all, water level has a significant effect on wetland methane flux, but not on carbon dioxide flux.