Introduction

Rising concentrations of carbon dioxide (CO2) and methane (CH4), which are known greenhouse gases (GHGs) in the atmosphere, are major contributors to global warming (IPCC 2014). Coastal wetlands, widely distributed along coastlines of the world, have been recognized as important sources of GHGs (Poffenbarger et al. 2011; Chen et al. 2012). A distinguishing feature from other wetlands is that coastal wetlands are periodically affected by tides. Tidal cycling has the potential to impact coastal wetlands soil hydrology and salinity characteristics, which are important factors controlling soil CO2 and CH4 production, diffusion and emission (Smith et al. 2003; Poffenbarger et al. 2011). For instance, several laboratory incubation studies indicated that artificially simulated seawater addition increased soil CO2 and CH4 production compared with freshwater controls (Weston et al. 2011), and induced a brief stimulation of soil CO2 production but continuous suppression of CH4 production, as compared with dilute seawater controls (Chambers et al. 2011). However, in situ measurements of GHG effluxes in most studies were made while tidal flats were not inundated by tidewater, such as in Chen et al. (2012), Livesley and Andrusiak (2012) and Wang et al. (2009), probably because of operational difficulties during tidal inundations. Thus, estimated GHG emissions from coastal wetland soils have large uncertainties because of insufficient data reflecting tidal effects. This restricts our ability to understand the role of coastal wetlands in mitigating global climate change.

Tides usually present two types of hydrological cycles on different time scales, semi-diurnal and semi-lunar tidal cycles (Kvale 2006). A semi-diurnal tide, including an entire flooding and ebbing process, recycles in every 12.4 h. Therefore, there are commonly two high tides and two low tides each day. Heights of high and low tides change from day to day, with a period of 14.8 days. This periodic variation is called a semi-lunar tidal cycle, including neap and spring tide periods. In the neap tide period (NTP), tidal heights are relatively low, such that most of the vegetated zone is unaffected by tidewater for several days. During the spring tide period (STP), that zone can be inundated by tidewater for several days because of high tide heights. Therefore, tidal effects on soil CO2 and CH4 emissions of the vegetated zone may vary greatly between tidal cycles, because of different inundation frequencies and durations.

Contradictory results of tidal effects on soil CO2 and CH4 emissions have been reported in several recent studies. For instance, no significant difference in CH4 emission was observed before and after tidal inundation (Tong et al. 2010, 2012). However, the temporal variations in CO2 and CH4 effluxes were positively correlated with change in the water table in the littoral zone of Lake Obuchi, Japan (Yamamoto et al. 2009, 2011). Also, CO2 emission was reported to be greatly inhibited by submersion in the Kurose River estuary of Japan (Sasaki et al. 2009). These studies were conducted across the semi-diurnal cycle, mainly in STP when soils were inundated or water-saturated most of the time. We speculate that temporal changes in soil conditions across a semi-diurnal cycle should be less profound than those across a semi-lunar tidal cycle. Along with the transition from NTP to STP or vice versa across a semi-lunar cycle, significant and complex variations in soil CO2 and CH4 emissions should be observed. However, there has been no study on whether or how the semi-lunar tidal cycle affects soil CO2 and CH4 emissions from coastal wetlands.

The present study was conducted from the mid-NTP to the end of STP in Dongtan wetland of the Yangtze River estuary, China. Our major objectives were to: (1) examine whether a semi-lunar tidal cycle could significantly affect CO2 and CH4 emissions; (2) explore possible underlying mechanisms of temporal variations in soil CO2 and CH4 emissions of coastal wetlands in relation to the semi-lunar tidal cycle.

Materials and methods

Study site

Dongtan wetland (230 km2, 31°25′–31°38′N, 121°50′–122°05′E) is on the eastern side of Chongming Island in the Yangtze River estuary (Fig. 1). The island has a typical subtropical monsoon climate, with mean annual temperature of 15.3 °C. Mean annual precipitation is 1022 mm, with most falling between April and September (Xu and Zhao 2003). Given a huge volume of sediments from the Yangtze River, the wetland continues to expand at an annual rate of 150–200 m in length or 4.06 km2 in area toward the East China Sea (Zhao et al. 2009). Vegetation of the wetland is dominated by Spartina alterniflora, Phragmites australis and Scirpus mariqueter (Li et al. 2009).

Fig. 1
figure 1

Locations of sampling sites in Dongtan wetland, Chongming Island in the Yangtze River estuary, China

Full size image

Tides in the study area have typically semi-diurnal and semi-lunar tidal cycles, with mean and maximum tidal ranges of 2.0–3.1 and 4.6–6.0 m, respectively, above the Wusong Datum Plane (Yang et al. 2001). Figure 2 shows field records of maximum tidal heights of each day across a semi-lunar tidal cycle in July at Hengsha (the tidal station nearest the Dongtan wetland). Tidal heights were relatively low from 13 to 20 July (Fig. 2), such that the wetland vegetation zone was not flooded by tidewater or standing water, according to our field observation. A part or nearly all of the vegetation zone was frequently subjected to tidal inundation from 21 to 27 July, when tidal heights were high.

Fig. 2
figure 2

Temporal variations in maximum tidal height within a semi-lunar tidal cycle (13–27 July) at Hengsha (tidal station nearest Dongtan wetland)

Full size image

Experimental design

A transect from a mudflat to a dike (built in 1998), was established north of Dongtan wetland (Fig. 1). Three sites from a low (sites S1 and S2) to high tide zone (site S3) were selected on the transect to examine tidal effects on soil CO2 and CH4 emissions from various tidal positions. The three sites were all in S. alterniflora stands to avoid interaction of plant species and tide. Three plots at least 5 m apart were chosen at each site for gas and soil sampling.

The field campaign was conducted in July 2009, when soil CO2 and CH4 emissions were the highest of the year, from the 17 July (mid-NTP with the lowest tide heights) to 27 July (end of STP) (Fig. 2). S1 and S2 were affected by tides beginning 21 July, and S3 was reached beginning 23 July. There was no raining event during NTP, so changes in soil water were only affected by the tidal cycle.

Gas sampling and analysis

A static closed chamber method was used to measure CO2 and CH4 emissions from the soil. Chambers were made of plexiglass plate of dimension 40 × 40 × 45 cm. A thermometer was installed inside the chamber, to determine temperature. Air pressure inside and outside the chamber was balanced by a stainless steel tube with internal diameter 2 mm. A battery-driven fan was used to mix air inside the chamber. A three-way valve with a needle 22-cm long was installed at the middle of the top wall of the chamber for sampling gas. The chamber was wrapped with aluminum foil to minimize air temperature changes inside it during measurement.

Ten days prior to field sampling, aboveground litter was carefully removed and plants were clipped without disturbing the soil. PVC collars, with a groove (5-cm depth) on the top edge for filling with water to seal the chamber, were installed in all plots prior to the field campaign. A hole with diameter 3 cm was driven on each side of the collar to allow tidewater to move in and out of the collar. These holes were sealed with rubber plugs during gas sampling and then opened after measurement. During STP, gas samples were taken after the tides had receded, when there was only a shallow layer of water left on the soil surface (water depth less than 2 cm). About 50 ml of gas was manually extracted from the chamber, using 50 ml syringes at 0, 10, 20, and 30 min after chamber closure. Three rounds of gas sampling were conducted between 10:00 and 15:00 each day. All gas samples were analyzed in the laboratory within 5 days after collection. Simultaneous with gas sampling, air and soil temperature at 5 cm depth were determined using Pt 1000 temperature sensors, and soil redox potential (Eh) was measured at 5-cm depth following the method of Zou et al. (2009) to examine tidal effects on the soil redox environment.

Concentrations of CO2 and CH4 in gas samples were determined using a gas chromatography (Agilent 6890 N, Agilent Technologies, USA) equipped with a FID detector (200 °C). Nitrogen (99.999 %) was used as a carrier gas. CO2 in gas samples was separated by a Porapak-Q packed column (2-m length, running at 25 ml/min and 55 °C) and reformed to CH4 by a nickel catalyst reformer (375 °C). CH4 was separated using a carbon molecular sieve packed column (2-m length, running at 30 ml/min and 55 °C).

Soil sampling and chemical analysis

In coastal wetland, if soil properties are affected by tidewater across a semi-lunar tidal cycle. Changes will likely occur only in top soils, because the tidal period is a relatively short. Therefore, soil samples were collected only at 0–20 cm depths, using stainless steel tubes of inner diameter 5 cm and length 50 cm. Samples of fresh soil for analyzing dissolved organic carbon (DOC) were stored at 4 °C prior to analysis.

Soil samples were oven-dried at 105 °C to constant weight to determine gravimetric moisture. Using a multi-parameter water quality analyzer (YSI 556 MPS, YSI Inc., USA), soil pH and salinity were measured with a 1:2.5 and 1:5 (weight:volume, soil:distilled water) slurry in the laboratory, respectively. Soil sulfate was analyzed using a Barium chloride turbidimetric method with an automated discrete analyzer (SmartChem2000, Westco Scientific Instruments Inc., France). Soil moisture, pH, salinity and sulfate were determined following the methods of Lu (1999). Soil DOC was measured according to Martin–Olmedo and Rees (1999) using a TOC analyzer (Multi N/C 3100, Analytik Jena AG, Germany).

Data analyses

A t test was used to analyze differences in soil properties, DOC, CO2 and CH4 emissions between NTP and STP at each site. Regression analysis was performed to test the effects of soil moisture on soil Eh, and was also used to determine the effects of soil properties on soil CO2 and CH4 emissions within a semi-lunar tidal cycle. Statistical significance was determined with a level of P < 0.05. All data are reported as mean ± SE. All statistics were performed using the software of SPSS 13.0 software for Windows.

Results

Soil properties and DOC

Mean soil moistures were significantly higher in STP than in NTP (Table 1). Increased mean soil moistures between STP and NTP were greater in the low tide zone than in the high tide one. No significant change in soil pH values was observed between NTP and STP (Table 1). Soil Eh declined significantly along with the transition from NTP to STP, and change in soil Eh was greater in the high tide zone than in the low tide one (Table 1). Regression analysis showed that soil Eh was negatively correlated with moisture (R 2 = 0.45, P < 0.001; Fig. 3).

Table 1 Soil properties in neap and spring tidal periods
Full size table
Fig. 3
figure 3

Negative relationship between soil Eh and moisture

Full size image

Mean soil salinity and sulfate content were significantly greater in STP than in NTP, owing to the influence of tides (Table 1). Compared with NTP, increase in soil salinity and sulfate content in STP decreased gradually from the low to the high tide zone.

Mean soil DOC content did not change significantly with the transition from NTP to STP, except at site S1 where the mean DOC was significantly lower in NTP than in STP.

Soil CO2 and CH4 emissions

CO2 effluxes from the soil in STP were significantly lower than those in NTP (P < 0.001; Fig. 4a). Tidal position did not significantly affect soil CO2 effluxes in STP, mainly because soils were over-saturated or submerged under water in that period. There was a similar temporal pattern of soil CH4 emission, which were also significantly lower in STP (P < 0.001; Fig. 4b).

Fig. 4
figure 4

Temporal dynamics in soil CO2 (a) and CH4 (b) emissions, along with semi-lunar tidal cycling. Sites S1 and S2 impacted by tides from 21 to 27 July, and site S3 impacted by tides from 23 to 27 July. Asterisks indicate significant differences (P < 0.05) between means of neap and spring tide periods

Full size image

Additionally, the decreases in soil CO2 and CH4 emissions began earlier in the low tide zone (21 July at sites S1 and S2) than in the high tide zone (23 July at site S3) (Fig. 4). Therefore, total soil CO2 and CH4 emissions were greater in the high tide zone than in the low tide zone, because the duration of strong emissions was longer in latter zone.

Regression indicated that soil CO2 effluxes were negatively correlated with soil moisture (R 2 = 0.43, P < 0.001; Fig. 5a), but positively correlated with Eh across the sites (R 2 = 0.64, P < 0.001, Fig. 5b). Similarly, there was negative correlation between soil CH4 effluxes and soil moisture (R 2 = 0.11, P = 0.01; Fig. 6a) but positive correlation between soil CH4 effluxes and Eh (R 2 = 0.37, P < 0.001; Fig. 6b). Moreover, soil CH4 effluxes were negatively correlated with salinity (R 2 = 0.24, P < 0.001; Fig. 6c) and sulfate (R 2 = 0.33, P < 0.001; Fig. 6d).

Fig. 5
figure 5

Regression analyses of soil CO2 emissions with soil moisture (a) and Eh (b), within a semi-lunar tidal cycle

Full size image
Fig. 6
figure 6

Regression analyses of soil CO2 emissions with soil moisture (a), Eh (b), salinity (c) and sulfate (d), within a semi-lunar tidal cycle

Full size image

Discussion

Tidal effects on soil properties

During the cycling of semi-lunar tides, periodical neap and spring tides significantly affected soil physical and chemical properties of coastal wetlands. A major change caused by the transition between NTP and STP was in soil water conditions. Compared with NTP, soil was over-saturated or submerged nearly all the time during STP because of frequent tidal inundation, especially in the low tide zone. This suppressed the diffusion of O2 from atmosphere into soil, and existing soil O2 was rapidly consumed during STP (de Mars and Wassen 1999; Seybold et al. 2002). Therefore, soil Eh decreased dramatically with the transition from neap to spring tides. Furthermore, variations of soil Eh with that transition increased gradually from the low to the high tide zone, probably because the frequency of inundation by tidewater is greater and the duration is longer in the former zone (Tang et al. 2010). Our result is supported by Mendoza et al. (2012), who reported that soil Eh under high inundation frequency was significantly lower than that under low inundation frequency in a mangrove swamp of Northern Brazil.

Soil pH is an important in regulating the soil inorganic carbon (SIC) pool (Mi et al. 2008). SIC may account for more than 60 % of soil total carbon in wetlands of the Yangtze River estuary (Cheng et al. 2006). Although S. alterniflora could reduce soil pH through secreting low molecular weight organic acids by roots (Hines et al. 1994) and preferential uptake of \( {\text{NH}}_{4}^{ + } \)-N (Mozdzer et al. 2011), mean pH value of the water near the Yangtze River estuary was ~8.2 (Patra et al. 2007). Thus, periodic tidal inundation can maintain the alkaline environment (pH > 8.0) of soils, ensuring the stability of the SIC pool in the estuary.

Soil salinity and sulfate were significantly greater in STP than in NTP, suggesting that tides transport substantial nutrients to coastal wetlands. Because of greater inundation frequency and longer duration, this exchange of salts was more efficient in the low tide zone than in the high tide zone. Consequently, nutrient input via this mechanism may be vital in supporting a high plant productivity in coastal wetlands (Morris and Haskin 1990; Trilla et al. 2010). We speculate that positive effects of nutrient input on plant productivity were stronger in the low tide zone than in the high tide zone.

Tidal effects on CO2 and CH4 emissions

In Dongtan wetland, soil CO2 emission in STP was only 29–34 % of that in NTP. Two mechanisms might be responsible for the decrease of CO2 emission in STP, an anaerobic environment inhibiting soil CO2 production (Glatzel et al. 2004), and high water content impeding CO2 diffusion from the soil into the atmosphere (Smith et al. 2003). Glatzel et al. (2004) showed that anaerobic respiration produced CO2 only 20–25 % of that from aerobic respiration in wetland soils. Tides may also affect soil CO2 emission through other pathways. For example, tides can import substantial nutrients including sulfate to coastal wetlands, possibly increasing CO2 production (Howes et al. 1984; Kostka et al. 2002). Sulfate reduction has been considered one of the major pathways of carbon mineralization in anaerobic saltmarshes. The relative contribution of sulfate reduction to total anaerobic carbon mineralization varies greatly, from a limited portion to ~50 % of mineralized C in coastal wetlands (Roychoudhury et al. 2003; King 1988). This contribution and its variation caused by tidal cycling are unlikely to alter the pattern of significantly lower CO2 emission in STP across a semi-lunar tidal cycle.

In coastal wetlands, processes of CH4 production are more complicated than those of CO2 production, because the former can be simultaneously affected by both promoting and inhibiting factors (Smith et al. 2003; Poffenbarger et al. 2011). CH4 is produced from anaerobic decomposition of substrates (e.g., soil organic matter, litter and root exudates) by methanogens, so CH4 emission should be positively correlated with soil moisture (Jungkunst et al. 2008) and negatively correlated with soil Eh (Koh et al. 2009). In our study, soil CH4 emission in STP constituted for 28–35 % of those in NTP, a change opposite that of soil moisture and a trend similar to soil Eh across a semi-lunar tidal cycle. These findings suggest that other factors may have influenced CH4 production in the soil.

As alternative electron acceptors, the presence of nitrate, Mn4+, Fe3+ and sulfate all have the potential to inhibit CH4 production in anaerobic decomposition of soil organic matter (Chapin III et al. 2002). Soil Eh maintained relatively low values in STP, so Mn4+ and Fe3+ were reduced to Mn2+ and Fe2+, respectively. Inhibitions of Mn4+ and Fe3+ on CH4 production were insignificant or nonexistent in STP. Peng et al. (2011) reported that tides imported inorganic nitrogen including nitrate to wetlands of the Yangtze River estuary. However, most inorganic nitrogen was absorbed and utilized by plants because the wetland had high plant productivity, especially in S. alterniflora stands (Liao et al. 2007). The content of soil sulfate in our study was more than ten times higher than the nitrate content reported in Peng et al. (2011), so nitrate may not be a major inhibitor of methanogenesis. Several studies have demonstrated that high soil salinity and sulfate were the most important inhibitors of CH4 production in coastal wetlands (Freeman et al. 1994; Purvaja and Ramesh 2001; Poffenbarger et al. 2011). In Dongtan wetland, tidewater in STP significantly increased soil salinity and sulfate, thereby strengthening the inhibition of methanogenesis. However, there remained a certain amount of CH4 production or emission in STP. This is likely attributed to at least two causes. First, fine-scale heterogeneity of electron acceptors and electron donors typically existed in the soils (Hojberg et al. 1994), suggesting simultaneous sulfate reduction and methanogenesis in coastal wetlands (Weston et al. 2011). Second, there were also noncompetitive substrates that can be only used by methanogens such as methanol, methionine and trimethylamine (Lyimo et al. 2002). Yuan et al. (2014) reported that CH4 production together S. alterniflora invasion in a coastal saltmarsh could be increased by an enhanced level of noncompetitive substrate trimethylamine.

Our study highlighted that both CO2 and CH4 emissions were significantly lower in STP than in NTP during a semi-lunar tidal cycle. This suggests that the overall, soil carbon emissions would probably be overestimated if soil CO2 and CH4 effluxes are only measured during NTP in wetlands of the Yangtze River estuary, such as in Wang et al. (2009).

Implications

Although our study was for in a limited area of the Yangtze River estuary, observed tidal cycle effects on soil carbon emissions are very likely common phenomena in coastal wetlands. There was significant variation in soil CO2 and CH4 emissions with semi-lunar tidal cycling, with STP effluxes only about a third of those in NTP. This temporal variation caused by the semi-lunar tidal cycle must be considered in modeling GHGs emissions from coastal wetlands under future scenarios of climate change.

Recent studies have shown that soils of coastal wetlands have relatively high carbon sequestration rates and acted as important carbon sinks (Chmura et al. 2003; McLeod et al. 2011). These findings were mainly attributed to high plant productivity, and thus high carbon input to the soil (McLeod et al. 2011). Our study suggests that semi-lunar tidal inundation significantly reduces soil CO2 and CH4 emissions, that is, it decreases vertical carbon output from the soil into the atmosphere. These are likely important mechanisms underlying high soil carbon accumulation in coastal wetlands. Overall, tides can enhance the strength of carbon sinks in such wetlands by promoting carbon input and inhibiting carbon decomposition. Furthermore, enhanced soil carbon accumulation by tides is likely greater in the low tide zone than in the high tide one, owing to greater inundation frequency and longer duration in the former zone.

In China and other countries, many engineering projects (such as levee, floodgates, and sea wall construction) have been implemented to regulate tidal flow in river estuaries or estuarine wetlands (Wang et al. 2006; Williams et al. 2013). These projects modify tidal patterns and/or hydrology of seawater and freshwater, thereby altering carbon cycling and GHG emissions in those wetlands. For example, CH4 emission from soil submerged by seawater is relatively low because of sulfate inhibition. However, that emission will greatly increase if wetland soil is submerged by freshwater, leading to a significant increase of global warming potential. Nitrogen supply from river flow may also be substantially reduced by these engineering projects, thereby reducing NPP of estuarine ecosystems. Large-scale engineering projects in river estuaries have potentially great impacts on coastal wetland ecosystems, in terms of carbon cycling and GHG emissions.

Conclusions

During the cycling of semi-lunar tides, periodic neap and spring tides significantly affected soil properties and carbon emissions in the Yangtze River estuary. Soil moisture, salinity, and sulfate are significantly increased, whereas soil Eh is significantly declined with the transition between NTP and STP. Soil CO2 effluxes in STP were only 29–34 % of those in NTP, owing to an anaerobic environment inhibiting soil CO2 production and tidal inundations impeding CO2 diffusion from the soil into atmosphere. During STP, the soil anaerobic environment had the potential to promote methanogenesis, but CH4 production was significantly inhibited by increased soil salinity and sulfate (Hawkins and Freeman 1994; Dowrick et al. 2006). Significantly reduced carbon emissions during the STP across the semi-lunar tidal cycle may be one of the potential mechanisms underlying the high soil carbon accumulation in coastal wetlands.