Tidal influence on carbon dioxide and methane fluxes from tree stems and soils in mangrove forests

Mangroves are critical blue carbon ecosystems. Measurements of methane (CH₄) emissions from mangrove tree stems have the potential to reduce uncertainty in the capacity of carbon sequestration. This study is the first to simultaneously measure CH₄ fluxes from both stems and soils throughout tidal cycles. We quantified carbon dioxide (CO₂) and CH₄ fluxes from mangrove tree stems of Avicennia marina and Kandelia obovata, which have distinct root structures, during tidal cycles. Tree stems of both species served as net CO₂ and CH₄ sources. Compared to fluxes in the soils, the mangrove tree stems exhibited remarkably lower CH₄ fluxes but no difference in CO₂ fluxes. The stems of A. marina exhibited an increasing trend in CO₂ flux from low to high tides. However, CH₄ fluxes showed high temporal variability, with the stems of A. marina functioning as a CH₄ sink before tidal inundation and becoming a source after ebbing. In contrast, the stems of K. obovata showed no consistent pattern in the CO₂ or CH₄ fluxes. Based on our findings, the stem CH₄ fluxes in A. marina may vary by up to 1200 % when considering tidal influence, compared to when tidal influence is ignored. Therefore, sampling only during low tides might underestimate stem CO₂ and CH₄ fluxes on a diurnal scale. This study highlights the necessity of considering tidal influence and species when quantifying greenhouse gas (GHG) fluxes from mangrove tree stems. Further study is needed to explore the underlying mechanisms driving the observed flux variations and improve the understanding of GHG dynamics in mangrove ecosystems.

1 Introduction

Global methane (CH₄) emissions have reached a record-high level (Saunois et al., 2020). Currently, there are two primary methods utilized for assessing global CH₄ emissions: the bottom-up method and the top-down method. The bottom-up method relies on compiling data from greenhouse gas (GHG) inventories and biogeochemical models to infer the sources of emissions. On the other hand, the top-down method involves measuring atmospheric CH₄ concentrations and utilizing transport models to infer the sources of emissions in order to estimate and assess CH₄ emissions on a global scale. CH₄ emissions estimated by the bottom-up method are significantly higher than those estimated by the top-down method, indicating a high degree of uncertainty and suggesting that some sources may be overlooked or not well understood (Jackson et al., 2020). CH₄ generated in wetlands can be released into the atmosphere not only through diffusion, ebullition, and transport mediated by herbaceous plants but also through the stems of woody plants (Gauci et al., 2010; Terazawa et al., 2007). Pangala et al. (2017) demonstrated that the difference between top-down and bottom-up estimates of CH₄ emissions could be accounted for by the upscaled CH₄ flux from tree stems, emphasizing the necessity of considering this pathway in carbon budgets (Carmichael et al., 2014). Furthermore, forest wetlands account for approximately 60 % of the global wetland area, highlighting the potential contribution of woody stems to global GHG emissions (Barba et al., 2019a; Covey and Megonigal, 2019). While carbon dioxide (CO₂) exchange at the stem–atmosphere interface has been examined (Teskey et al., 2008), little is known regarding the sources and mechanisms of CH₄ emissions originating from tree stems relative to those originating from other pathways. CH₄ emitted by tree stems may originate from microorganisms/cryptogams within the stem bark (Jeffrey et al., 2021; Lenhart et al., 2015) or from the soil, where it is produced and enters the roots before being transported in either liquid or gaseous form through xylem or aerenchyma tissue (Kutschera et al., 2016; Vroom et al., 2022).

GHG emissions from tree stems exhibit temporal and spatial variations, with different influencing mechanisms found in various studies. Firstly, tree stem GHG fluxes tend to be higher during the growing season and lower during the dormant season, but there may also be no significant differences among seasons (Barba et al., 2019b; Köhn et al., 2021; Pangala et al., 2015; Pitz et al., 2018; Wang et al., 2016; Zhang et al., 2022). Secondly, significant variations in GHG fluxes from tree stems have been observed at different heights above ground level, with a decreasing trend along the tree stem height (Moldaschl et al., 2021; Pangala et al., 2013, 2014, 2015; Sjögersten et al., 2020), although some studies have not reported this phenomenon (Machacova et al., 2021; Wang et al., 2016). Thirdly, tree stem GHG emissions may be regulated by various environmental factors, such as temperature, moisture, and redox potential (Barba et al., 2019b; Gao et al., 2021; Jeffrey et al., 2019; Pitz et al., 2018; Schindler et al., 2020, 2021; Sjögersten et al., 2020; Terazawa et al., 2015), which can be affected by fluctuations in water table height due to seasonal changes and hydrological processes (Jeffrey et al., 2023; Peacock et al., 2024; Terazawa et al., 2021). Finally, the physiological factors of trees, such as lenticel density, wood density, water content, and stem bark structure, may also influence GHG fluxes originating from tree stems (Jeffrey et al., 2024; Pangala et al., 2013, 2014, 2015; Wang et al., 2016; Zhang et al., 2022).

However, most related studies have focused on freshwater wetlands and upland forests, while relatively limited research has focused on mangrove forests. Jeffrey et al. (2019) reported that dead mangrove trees may account for approximately 26 % of CH₄ emissions in mangrove ecosystems. However, He et al. (2019) reported inconsistent results, revealing a relatively small contribution from tree stems. The contribution of mangrove tree stems to the ecosystem GHG flux is generally less than the contribution of soil (Gao et al., 2021; He et al., 2019; Jeffrey et al., 2019), but it still has the potential to exceed 50 % (Zhang et al., 2022). Additionally, GHG fluxes from mangrove tree stems vary among tree species (Zhang et al., 2022) and may even differ within a single tree species (Gao et al., 2021), highlighting the uncertainty in GHG emissions from mangrove tree stems and emphasizing the need for further investigation.

Mangroves are primarily distributed in tropical and subtropical coastal regions and are regarded as critical ecosystems with a high capacity for sequestering blue carbon (Li et al., 2018; Duarte de Paula Costa and Macreadie, 2022). The anaerobic conditions resulting from tidal inundation, along with the abundant organic matter, turn mangrove soil into a source of CH₄ emissions (Lin et al., 2020). This, in turn, impacts their role in mitigating global warming. Moreover, several studies have demonstrated the influence of tides on the emission of GHGs in coastal wetlands. In both seagrass meadows and tidal marshes, the CH₄ flux tends to peak when tidal water reaches the sampling site (Bahlmann et al., 2015; Capooci and Vargas, 2022). The sudden release of CH₄ can occur through physical force under the influence of tidal movement (Li et al., 2021), resulting in the advective exchange of groundwater or soil pore water with the overlying surface water (Billerbeck et al., 2006; Rosentreter et al., 2018). CH₄ emissions during tidal inundation may be higher if tidal water contains high concentrations of dissolved CH₄, which can increase the emissions of CH₄ through diffusion due to the concentration gradient (Sturm et al., 2017; Tong et al., 2013). Yamamoto et al. (2009) reported a positive correlation between the water table and GHG fluxes in the flooded littoral zone with vegetation, suggesting that water pressure, rather than gas diffusion, primarily affects the emissions of CO₂ and CH₄ across the water–atmosphere interface by ejecting gases from pore spaces. This finding is contrary to previous results in which lower CH₄ fluxes were observed during high tides, which may have been caused by the higher water pressure limiting CH₄ diffusion in soil pore spaces filled with water and plant-mediated transport (Tong et al., 2010, 2013). Additionally, CH₄ may be oxidized during diffusion in water (Tong et al., 2013). Furthermore, if the dissolved oxygen concentration, sulfate concentration, and salinity are high in tidewater, this may inhibit CH₄ production and/or promote CH₄ oxidation (Huang et al., 2019), resulting in lower CH₄ emissions during high tides. The variation in the CH₄ flux across the water–atmosphere interface during tidal inundation could be driven by current- or wind-induced turbulence (Sturm et al., 2017). CH₄ emissions even exhibited different trends during spring and neap tides (Huang et al., 2019; Tong et al., 2013). However, to our knowledge, there is only one study on GHG fluxes from mangrove tree stems during tidal cycles (Epron et al., 2023).

This study aimed to quantify the CO₂ and CH₄ emissions from the tree stems of K. obovata and A. marina, which are dominant mangrove species with distinct root structures, distributed along the northern and southern coasts of Taiwan, respectively. We investigated the temporal variations in stem GHG fluxes during tidal cycles and assessed the influence of tides on the upscaled flux. We also simultaneously measured the GHG emissions from mangrove soil, even during tidal inundation, to compare the temporal dynamics of GHG fluxes between the tree stems and soil. We hypothesized that GHG fluxes from mangrove tree stems and soil exhibit synchronized temporal and species variation during the tidal cycle and that the tidal cycle may exert a significant impact on GHG emissions on a larger scale.

2 Materials and methods

2.1 Site description

This study focused on mangroves at four sites along the western coast of Taiwan (Fig. 1). The dominant mangrove species in Wazihwei (K-WZW; 25°10^′ N, 121°25^′ E) and Xinfeng (K-XF; 24°55^′ N, 120°58^′ E) is Kandelia obovata, while Avicennia marina is the dominant species in Fangyuan (A-FY; 23°56^′ N, 120°19^′ E) and Beimen (A-BM; 23°17^′ N, 120°6^′ E). K-WZW and K-XF are situated in northern Taiwan, a subtropical region, with average annual precipitation values of 2023 and 1537 mm, respectively. A-FY and A-BM are located in southern Taiwan, a tropical region, with average annual precipitation values of 1162 and 1603 mm, respectively. A-BM has the largest forest area (75.3 ha), while K-XF has the smallest (8.12 ha). Mean tree height across all sites ranged from 1.8 to 5.1 m, and tree density and diameter at breast height (DBH) averaged 0.6–2.4 tree m⁻² and 5.6–10.5 cm, respectively (Table 1). The tides were semidiurnal at all sites. The soil texture at all sites was silty, with an average grain size of 0.046 mm. During summer (the study period), the average air temperature was 28.4 °C for K. obovata and 29.4 °C for A. marina (Lin et al., 2023). The sampling campaign was conducted from 1 June 2022 to 29 July 2022, with each site sampled for 3 d during the spring tide (Table 1). This period was chosen mainly because there is a higher GHG flux in summer compared to other seasons, as indicated by preliminary studies conducted at the same sites (Lin et al., 2020).

Figure 1 Sample sites along the western coast of Taiwan. The blue dots represent the locations of sampled trees. K-WZW: Wazihwei. K-XF: Xinfeng. A-FY: Fangyuan. A-BM: Beimen. The dominant mangrove species in K-WZW and K-XF is Kandelia obovata, while Avicennia marina is the dominant species in A-FY and A-BM. The map sources are Natural Earth (left) and © Google Earth (right).

2.2 Flux measurements

At each sampling site, a mangrove tree was selected for tree stem CO₂ and CH₄ flux measurements obtained at approximately 110 cm above the ground. This specific height was chosen with consideration for the potential maximum tidal height, which may reach up to 80 cm above the ground (Table 1). Due to differences in stem morphology, two distinct stem chambers – a semirigid chamber and a cylindrical chamber – were used in this study to measure the GHG emissions of K. obovata and A. marina, respectively (Fig. S1 in the Supplement).

The semirigid chamber was adapted from Siegenthaler et al. (2016) and was constructed from transparent recycled polyethylene terephthalate (rPET) bottles. A plastic sheet measuring 14 cm in length and 11 cm in width was cut from a bottle, and 2 cm wide and 1.5 cm thick chloroprene rubber (CR) foam tape was attached around the edges and center of the plastic sheet. Two holes were drilled and fitted with adapters to connect the tubing, resulting in a chamber with a 16 cm² surface area and a 0.2 L volume. The chamber was installed on the tree stem with a strap prior to measurement and subsequently removed. The second cylindrical chamber was constructed from a 0.2 L white polypropylene (PP) bottle. A 16 cm² square was cut from the lid, and two small holes were drilled at the bottom of the bottle; these holes were fitted with adapters to connect the tubing. The lid was fixed to the stem and sealed with silicone prior to measurement. After each measurement, the chamber was removed, but the lid remained on the trunk (Fig. S1).

Two locations within 2 m of the sampled tree were selected to measure CH₄ and CO₂ fluxes at the soil–atmosphere and water–atmosphere interfaces during the tidal cycle, using the static chamber (Lee et al., 2011) and the floating chamber method (Lin et al., 2024), respectively. The soil chamber comprised a semicircular, transparent polymethyl methacrylate (PMMA) cover (30 cm in diameter) and a stainless steel ring (16 cm in height and 30 cm in diameter), with an adapter on the cover for connecting the tubing. The ring was pressed into the soil before the cover was placed over it, and a long-tailed clip was used to secure and cover the steel ring tightly to achieve an airtight seal (Fig. S1). During high tide, if the water level exceeded the height of the soil chamber (16 cm), the floating chamber was used (Fig. S1).

In this study, a portable gas analyzer (LI-7810 from LI-COR Biosciences, Lincoln, NE, USA) was used to simultaneously measure CO₂ and CH₄ fluxes. The chamber was connected to the analyzer through tubing, and the gas inside the chamber was drawn into the analyzer with a pump, with each measurement lasting approximately 5 min for the stem and 7 min for the soil. During the tidal cycle, stem and soil GHG fluxes were measured consistently. After each measurement was completed, the airtight sealed chamber was opened for approximately 3 min to allow the GHG concentration within the chamber to stabilize. Simultaneously, at the beginning of the flux measurement, the water level adjacent to the sampled trees was measured using a tape measure fixed to a PVC pipe (Fig. S1). To minimize soil disturbance, the researcher remained stationary at one location during the sampling campaign and avoided walking around. Sampling was mainly conducted during daylight hours. Soil GHG flux data were mainly derived from Lin et al. (2024). The GHG flux (F) was calculated using the following equation:

$$\begin{array}{}\text{(1)}& F=(S\times V\times c)/(RT\times A),\end{array}$$

where S is the slope obtained from the linear regression of GHG concentration changes over time (expressed in ppb CH₄ s⁻¹ or ppm CO₂ s⁻¹), V is the chamber volume (liters), c is the conversion factor from seconds to hours, R is the ideal gas constant (0.082 L atm K⁻¹ mol⁻¹), T is the air temperature inside the chamber (kelvin), and A is the surface area of the chamber (m²). If the R² value of the linear regression was <0.7, the GHG flux was removed from further statistical analysis. The surface area and volume of the semirigid chamber were calculated as described by Siegenthaler et al. (2016).

Different upscaling methods were applied to the tree stem GHG fluxes. First, the average fluxes during low and high tides were multiplied by the non-inundation time and the inundation time (in hours), respectively. These values were then summed to calculate the daily fluxes, accounting for tidal influence, which is denoted as F_BothTide. Since sampling in mangrove forests was mostly conducted during low tide, the average fluxes during low tide were multiplied by 24 h to scale up to daily fluxes, denoted as F_LowTide, for comparison with the fluxes accounting for tidal influence. The equations are shown below:

$$\begin{array}{}\text{(2)}& {\displaystyle }{F}_{\mathrm{BothTide}}=\left(F_{\mathrm{high}}\times t_{\mathrm{inundated}}\right)+\left(F_{\mathrm{low}}\times \left(\mathrm{24}-t_{\mathrm{inundated}}\right)\right),\text{(3)}& {\displaystyle }{F}_{\mathrm{LowTide}}=F_{\mathrm{low}}\times \mathrm{24},\end{array}$$

where F_low and F_high are the average fluxes during low and high tides, respectively, and t_inundated is the average inundation time per day, acquired by multiplying the hours per day when the water level was higher than 0 cm by 2 (since the tides are semidiurnal tides).

2.3 Statistical analysis

All the statistical analyses were performed using R 4.2.2 software. All the data were assessed for a normal distribution using the Shapiro–Wilk test. The Kruskal–Wallis test on ranks was used to evaluate the differences in CO₂ and CH₄ fluxes between sites. To determine which sites differed, Dunn's multiple comparison test was applied as a post hoc analysis when differences were significant (p<0.05). The relationships between the CO₂ and CH₄ fluxes during rising and falling tides were analyzed using a simple linear regression model. The results were considered statistically significant when the p value was <0.05. Data are primarily presented as the mean ± standard deviation (SD).

3 Results

During the study period, the mangrove tree stems served as net CO₂ sources, but there was distinct variation between sites (Fig. 2). In the K. obovata mangroves, the average CO₂ fluxes from the tree stems during the tidal cycle were 1.21 ± 0.10 mmol m⁻² h⁻¹ at the K-WZW site and 1.06 ± 0.20 mmol m⁻² h⁻¹ at the K-XF site (Fig. 2a). The stem CO₂ fluxes were significantly higher at the A-FY and A-BM sites, averaging 10.62 ± 2.35 and 16.00 ± 9.41 mmol m⁻² h⁻¹, respectively (Fig. 2a). Across all sites, only the tree stem at the A-FY site functioned as a net CH₄ sink (−0.17 ± 0.52 µmol m⁻² h⁻¹). However, the stem CH₄ fluxes at the K-WZW and K-XF sites showed no significant difference compared to the A-FY site, averaging 0.05 ± 0.06 and 0.04 ± 0.04 µmol m⁻² h⁻¹, respectively (Fig. 2b). The stem CH₄ fluxes were significantly higher at the A-BM site (0.48 ± 1.17 µmol m⁻² h⁻¹; Fig. 2b). Compared to those of the tree stems, the soils of the K. obovata and A. marina mangrove forests exhibited remarkably high CH₄ fluxes, averaging 7.59 ± 8.74 and 42.23 ± 62.95 µmol m⁻² h⁻¹, respectively. The average CO₂ flux from the soil was 1.73 ± 2.31 mmol m⁻² h⁻¹ in the K. obovata mangroves and 3.42 ± 3.36 mmol m⁻² h⁻¹ in the A. marina mangroves, but it did not differ significantly from that of the tree stems.

Figure 2 Differences in stem (a) CO₂ and (b) CH₄ fluxes among sites. Each data point represents a flux measurement taken during the tidal cycle, with 88 replicates for K-WZW, 82 replicates for K-XF, 75 replicates for A-FY, and 152 replicates for A-BM. The different letters above the boxplots indicate significant differences among sites, as determined by the Kruskal–Wallis test and Dunn's test (p<0.05).

The mean inundation time and highest tidal height at each sampling site are provided in Table 1. During the tidal cycle, the CO₂ fluxes from the mangrove tree stems exhibited different trends across all sampling sites (Fig. 3). The emissions remained relatively constant during the tidal cycle, ranging from 1.01 to 1.43 mmol m⁻² h⁻¹ and from 0.85 to 2.59 mmol m⁻² h⁻¹ at the K-WZW and K-XF sites, respectively (Fig. 3a). However, a sharp emission peak (2.59 mmol m⁻² h⁻¹) was observed at the K-XF site on day 2, when the tide was falling; this peak was 3-fold higher than the lowest flux (0.85 mmol m⁻² h⁻¹) measured on the same day (Fig. 3a). The CO₂ flux at the A-FY and A-BM sites generally showed an increasing trend throughout the tidal cycle, ranging from 4.54 to 14.00 mmol m⁻² h⁻¹ and from −1.68 to 39.15 mmol m⁻² h⁻¹, respectively (Fig. 3a). However, this trend was observed at the A-FY site only on day 1, when there was a distinct temporal trend in the increase in CO₂ flux relative to that at the A-BM site. Specifically, the former started to increase before the flood current entered and stabilized after high tide, reaching a peak flux (10.36 mmol m⁻² h⁻¹) at the end of the measurement. Conversely, the latter showed no significant change during the rising tide, followed by a steep rise toward high tide and a slight decrease during the falling tide; however, the CO₂ flux still remained higher than that during the pre-flood tide, ranging from −1.68 to 33.24 mmol m⁻² h⁻¹ during the rising tide and from 8.74 to 39.15 mmol m⁻² h⁻¹ during the falling tide (Fig. 3a).

Table 1 Comparison of upscaling methods with and without considering tidal influences on CO₂ and CH₄ fluxes in mangroves.

^a For F_BothTide, the average fluxes during low and high tides were added after being multiplied by the corresponding time length. ^b For F_LowTide, the average fluxes during low tides were multiplied by 24 h. The sampling date and time are provided in ISO 8601 format. ^c The data were derived from Lin et al. (2021).

The CO₂ flux pattern observed during the tidal cycle differed between the tree stems and soils. Generally, the soil CO₂ flux peaked before and after high tide at all sites, during either the rising tide or the falling tide, when the flood current was just entering or leaving the sampling site (Fig. 3b).

Figure 3 Variations in (a) stem CO₂ fluxes and (b) soil CO₂ fluxes during the tidal cycle. Time was standardized based on the time of the highest water level during the high-tide period (set as 0) and then adjusted by decrementing the time by 0.1 for every 10 min interval prior to the peak and incrementing the time by 0.1 for every 10 min interval after the peak. The average values of the flux and water level were calculated for each standardized time interval. The shaded area denotes the water level at the sampled tree. On days 1, 2, and 3, the plant data were arranged chronologically according to the sampling date.

Similar to the CO₂ fluxes, the CH₄ fluxes in K. obovata and A. marina exhibited distinct temporal trends during the tidal cycle (Fig. 4). In the K. obovata mangroves, there was significant variation in the stem CH₄ flux during the tidal cycle, ranging from −0.14 to 0.38 µmol m⁻² h⁻¹ and from −0.05 to 0.18 µmol m⁻² h⁻¹ at the K-WZW and K-XF sites, respectively, while consistent patterns were lacking between the sampling campaigns (Fig. 4a). The stem CH₄ flux of A. marina increased throughout the tidal cycle, ranging from −1.92 to 0.55 µmol m⁻² h⁻¹ and from −4.13 to 2.67 µmol m⁻² h⁻¹ at the A-FY and A-BM sites, respectively. Specifically, the tree stems of A. marina functioned as CH₄ sinks before tidal inundation (−0.53 ± 0.73 µmol m⁻² h⁻¹ at the A-FY site and −0.64 ± 1.51 µmol m⁻² h⁻¹ at the A-BM site), but the CH₄ flux gradually increased thereafter, eventually becoming a CH₄ source during low tide (0.18 ± 0.24 µmol m⁻² h⁻¹ at the A-FY site and 1.54 ± 0.56 µmol m⁻² h⁻¹ at the A-BM site). However, this pattern was not observed across all sampling campaigns (Fig. 4a).

For both mangrove species, the soil CH₄ flux during high tide (21.65 ± 45.29 µmol m⁻² h⁻¹) was lower than that during low tide (47.70 ± 63.27 µmol m⁻² h⁻¹) (Fig. 4b). Furthermore, there was a peak in the soil CH₄ flux during both tidal increases and decreases on all 3 sampling days, similar to that observed for the soil CO₂ flux (Figs. 3b; 4b).

Figure 4 Variations in (a) stem CH₄ fluxes and (b) soil CH₄ fluxes during the tidal cycle. Time was standardized based on the time of the highest water level during the high-tide period (set as 0) and then adjusted by decrementing the time by 0.1 for every 10 min interval prior to the peak and incrementing the time by 0.1 for every 10 min interval after the peak. The average values of the flux and water level were calculated for each standardized time interval. The shaded area denotes the water level at the sampled tree. On days 1, 2, and 3, the plant data were chronologically arranged according to the sampling date.

During the tidal cycle, the CO₂ flux from the mangrove tree stems was positively correlated with the CH₄ flux during both the rising and falling tides. However, a significant relationship was detected only for A. marina (Fig. 5a; p<0.001). The CO₂ and CH₄ fluxes from both the stems and soils were simultaneously measured, and a negative correlation between the stem and soil fluxes was observed across the two mangrove species. However, a significant relationship was detected only for A. marina during the falling tide (Fig. 5b, c; p<0.001).

Figure 5 Relationships between (a) stem CO₂ fluxes and CH₄ fluxes, (b) stem CO₂ fluxes and soil CO₂ fluxes, and (c) stem CH₄ fluxes and soil CH₄ fluxes. The shaded areas denote the 95 % confidence intervals of the regression lines. $R_{\mathrm{adj}}^{\mathrm{2}}$: adjusted R².

Since the tides at the sample sites were mainly semidiurnal tides, the average inundation time per day was calculated by multiplying the average time of high tide (when the water level was higher than 0 cm) during each sampling event by 2. The A-BM site exhibited the longest inundation time (15.33 h), while the inundation time during the sampling campaign was 6.69 h at the K-WZW and K-XF sites and 5.19 h at the A-FY site. The average highest tidal height (determined by the distance between the soil and water surface during high tide) was 58.1 cm at the K-WZW site, 70.5 cm at the K-XF site, 47.3 cm at the A-FY site, and 77.5 cm at the A-BM site. Different upscaling methods were applied to determine the tidal influence on the diurnal variation in the fluxes, where F_BothTide denotes the sum of the average fluxes during low and high tides, each multiplied by the corresponding time length, and F_LowTide denotes the average flux during low tides multiplied by 24 h. The GHG fluxes exhibited notable differences when tidal influences were considered (Table 1). Based on our findings, sampling only during low tide may underestimate the stem CO₂ and CH₄ fluxes on a diurnal scale, except at the K-WZW site, where the stem CO₂ and CH₄ fluxes were 0.03 % and 3.68 % lower when considering tidal influences (Table 1). At the K-XF, A-FY, and A-BM sites, the differences in the stem CO₂ fluxes between the upscaling methods were smaller than the differences in the stem CH₄ fluxes, ranging from 0.81 % to 9.40 % (Table 1). The stem CH₄ fluxes at the K-XF site were approximately 6 % higher when considering tidal influences, as opposed to when tidal influences were ignored (Table 1). If the tidal influences were not accounted for, the mangrove tree stems at the A-FY and A-BM sites both acted as net CH₄ sinks, while the CH₄ sink capacity was 8 % and 1200 % lower, respectively, after accounting for tidal influences, resulting in the mangrove tree stem at the A-BM site turning into a net CH₄ source (Table 1).

4 Discussion

This study revealed distinct spatial and temporal variations in the CO₂ and CH₄ fluxes originating from tree stems and soils. Specifically, the sites dominated by A. marina exhibited CO₂ fluxes up to 15 times higher than those at the sites dominated by K. obovata. The tree stems of A. marina at the A-FY site acted as a net CH₄ sink, while the A-BM site emitted CH₄ at a rate approximately 3 times higher. In contrast, the tree stems of K. obovata at the K-WZW and K-XF sites were weak CH₄ sources compared to the tree stems at the A-BM site. The temporal dynamics during the tidal cycle also differed between the two mangrove species. Regarding K. obovata, the stem CO₂ and CH₄ fluxes at the K-WZW and K-XF sites lacked a consistent pattern between the sampling campaigns. In contrast, A. marina exhibited an increasing trend in CO₂ flux throughout the tidal cycle, whereas the CH₄ flux exhibited high temporal variability, functioning as a sink before tidal inundation and becoming a source during low tide at both the A-FY and A-BM sites. Therefore, our results suggest that different mangrove species – in this case, K. obovata and A. marina – may have varying capacities for CO₂ and CH₄ exchange with the atmosphere through tree stems during tidal cycles. Further investigation with a larger sample size is needed to examine the hypothesis of mangrove species variation in GHG fluxes.

In terms of biological factors, A. marina contains pneumatophores, while K. obovata does not. Pneumatophores may facilitate the transport of oxygen to the rhizosphere and increase the oxidation-reduction potential, thereby inhibiting the methanogenesis process (Dušek et al., 2021). However, they can also serve as pathways for deep-soil-layer CH₄ emissions, facilitating CH₄ transport (He et al., 2019; Lin et al., 2021). In this study, pneumatophores were not intentionally avoided during the measurements. Therefore, the presence of pneumatophores may contribute to the increased soil CH₄ flux in the A. marina mangrove forest.

GHG emissions from stems, whether originating from the soil or the stems themselves, require radial diffusion through the bark or lenticel to reach the atmosphere (Barba et al., 2019a). Radial diffusion is primarily influenced by biological factors, such as wood density, wood moisture content, and lenticel density (Covey and Megonigal, 2019). A higher lenticel density, in particular, creates more pathways for GHG emissions, resulting in increased emissions (Zhang et al., 2022). Based on visual observations in situ, we found that the tree stems at the A-FY and A-BM sites exhibited a significantly higher lenticel density than those at the K-WZW and K-XF sites (Table 1). Therefore, it is speculated that the higher lenticel density of A. marina facilitates the emission of GHGs from stems, resulting in a higher stem GHG flux at the A-FY and A-BM sites.

Previous studies on GHG emissions originating from mangrove tree stems were mostly conducted during low tide and under daylight conditions. Gao et al. (2021) showed that the stems of Kandelia obovata can both absorb and release CH₄, with average fluxes of −5.69 and 1.84 µmol m⁻² h⁻¹, respectively. Zhang et al. (2022) reported higher CH₄ emissions from K. obovata stems (7.04 µmol m⁻² h⁻¹), which dominated the ecosystem's CH₄ flux in mangroves without pneumatophores. These results contradict the findings of this study, in which the CH₄ emissions of K. obovata stems contributed less than the soil emissions. Liao et al. (2024) measured lower stem CH₄ fluxes from K. obovata during winter (0.54 µmol m⁻² h⁻¹); these fluxes were 10 times higher than the average fluxes observed in this study. In the case of A. marina, the average stem CH₄ fluxes were 1.56 µmol m⁻² h⁻¹ (Jeffrey et al., 2019) and 2.79 µmol m⁻² h⁻¹ (Zhang et al., 2022) at mangrove sites located in Australia and China, respectively. The tree stems of A. marina also exhibited CH₄ consumption capacity, with fluxes ranging from −33.96 to 48.83 µmol m⁻² h⁻¹, as reported in Gao et al. (2021). Regarding other mangrove species, Kandelia candel exhibited a stem CH₄ flux of −1.81 µmol m⁻² h⁻¹, while Sonneratia apetala, Laguncularia racemosa, and Bruguiera spp., all of which have the same specialized root structure as A. marina, exhibited stem CH₄ fluxes of 2.62, 0.87, and −0.49 µmol m⁻² h⁻¹, respectively (He et al., 2019). Epron et al. (2023) measured the CH₄ flux of the stems of Bruguiera gymnorrhiza throughout a 24 h cycle, with fluxes ranging from −0.36 to 263.16 µmol m⁻² h⁻¹. In this study, the CH₄ fluxes in the stems of A. marina and K. obovata ranged from −0.14 to 0.38 µmol m⁻² h⁻¹ (0.05 ± 0.06 µmol m⁻² h⁻¹ at the K-WZW site and 0.04 ± 0.04 µmol m⁻² h⁻¹ at the K-XF site) and from −4.13 to 2.67 µmol m⁻² h⁻¹ (−0.17 ± 0.52 µmol m⁻² h⁻¹ at the A-FY site and 0.48 ± 1.17 µmol m⁻² h⁻¹ at the A-BM site), respectively; these values are at the low end of the reported range of stem CH₄ fluxes from previous studies (Table 2). Although CH₄ fluxes from mangrove tree stems generally decreased with increasing height (Epron et al., 2023; Gao et al., 2021; Jeffrey et al., 2019; Liao et al., 2024), average stem CH₄ fluxes in A. marina and K. obovata at similar heights to those in this study (>1 m) were still higher. This may be due to site-specific variations in environmental conditions, tree physiology, and microbial activity, all of which can influence the production and consumption of methane by mangrove trees (Barba et al., 2019a; Covey and Megonigal, 2019). Further research is needed to delve into the underlying mechanisms, which were not fully elucidated in this study due to limited data availability.

Table 2 Comparison of stem methane (CH₄) fluxes in mangrove ecosystems reported in this study with previous literature. The values are presented as the range (the minimum value to maximum value) and/or the mean ± standard deviation (given in parentheses).

GC stands for gas chromatography, CRDS for cavity ring-down spectroscopy, CEAS for cavity-enhanced absorption spectroscopy, and ND for no data.

The tree stem CO₂ and CH₄ fluxes exhibited similar temporal patterns during the tidal cycle. A significant positive relationship was also found between these fluxes, indicating that CO₂ and CH₄ emitted by mangrove tree stems may originate from the same source or be influenced by the same mechanism during the tidal cycle (Liao et al., 2024). According to previous studies, CO₂ emissions primarily occur through root respiration and stem respiration, as well as via internal plant metabolism and transport from soils (Teskey et al., 2008). In contrast, CH₄ may be emitted or absorbed by methanogens and methanotrophs present in tree bark or heartwood (Feng et al., 2022; Jeffrey et al., 2021). CH₄ emitted by tree stems may also originate from the soil, where CH₄ produced in the soil enters the root system, moves into the tree's aerenchyma tissues or xylem, and is subsequently directly released into the atmosphere through lenticels or tree stems (Barba et al., 2019a; Covey and Megonigal, 2019). Therefore, the fixation of CO₂, oxidation of CH₄, and emission of both GHGs by the tree stem may originate from the tree stem itself or from the soil. In this study, the transformation of tree stems from CH₄ sinks to CH₄ sources was observed in the A. marina mangrove forest. This observation indicates that CH₄ emitted by tree stems may be affected by different sources during different periods of the tidal cycle.

The transport mechanism of GHGs in the stem is similar to that in herbaceous plants, occurring mainly via diffusion or evaporation, either jointly or individually. The diffusion direction mainly depends on the CH₄ concentration gradient. For example, if the gas concentration in the rhizosphere is high, GHGs can enter the plant root system in either gaseous form or liquid form, thus entering the aerenchyma or xylem tissue (Vroom et al., 2022). Aerenchyma is a specialized tissue found in many mangrove tree species (Evans, 2004). It comprises air-filled spaces that create gas transport pathways within the plant. Aerenchyma facilitates gas movement, including that of CO₂ and CH₄, within stems. Within the aerenchyma, CO₂ and CH₄ can diffuse or passively flow along concentration gradients. This transport pathway allows gases to move vertically within the plant – from the roots, through the stem, and ultimately into the atmosphere. Aerenchyma tissue is particularly important for CH₄ transport because CH₄ is produced in oxygen-limited soils or in the rhizosphere by methanogens. The aerenchyma provides a direct pathway for CH₄ to move upward through the stems and be emitted into the atmosphere (Yáñez-Espinosa and Angeles, 2022). CO₂ and CH₄ can also dissolve during dilution and be transported within the xylem via sap flux (Takahashi et al., 2022). This study revealed the transition of mangrove tree stems from CH₄ sinks to CH₄ sources within the tidal cycle, a phenomenon that has not been observed in other studies, even with high-frequency measurements of upland tree stems at 1 h intervals (Barba et al., 2019b). We speculate that tree stems of A. marina may absorb CH₄ through the presence of methanotrophs during low tide (Jeffrey et al., 2021). During inundation, the diffusion of CH₄ produced in the deep soil layer may be restricted by water pressure (Tong et al., 2013) since the pore spaces are filled with water. Tong et al. (2010) also reported a significantly lower CH₄ flux during inundation than during low tide. Therefore, we hypothesize that CH₄ produced in the soil during inundation periods may be primarily emitted into the atmosphere through tree stems (Vroom et al., 2022; Yáñez-Espinosa and Angeles, 2022) rather than across the water–atmosphere interface via diffusion or ebullition (Li et al., 2021), resulting in the observed gradual increase in CH₄ flux throughout the tidal cycle. This hypothesis is also supported by the negative relationship between the soil and stem CH₄ fluxes in A. marina observed during both rising and falling tides in this study. However, the CH₄ flux in the tree stems of Bruguiera gymnorrhiza peaked after the tide receded (Epron et al., 2023), which does not support this hypothesis. It is critical to note that the specific mechanisms driving the observed peaks may vary depending on factors such as mangrove species, environmental conditions, tidal dynamics, and site-specific characteristics. However, further research is necessary to fully comprehend the underlying mechanisms.

To our knowledge, this study is the first to simultaneously measure CH₄ fluxes in both stems and soils throughout the tidal cycle, even during tidal inundation. When quantifying the GHG emissions of mangrove tree stems, discrete and continuous methods are two common measurement approaches. Discrete measurements involve sampling at specific time points with a lower temporal resolution and are practical and cost-effective. Continuous measurements provide real-time monitoring with a high temporal resolution, accurately capturing short-term fluctuations and peak emissions, but require specialized equipment and technical expertise. When considering tidal influences through continuous measurements, CH₄ emitted by mangrove tree stems was significantly higher, with differences of up to 1200 % in the stem CH₄ fluxes. Conversely, the upscaled CH₄ flux accounting for tides in tidal salt marshes was lower (Huang et al., 2019). When quantifying the GHG emissions of mangrove tree stems, discrete measurements are commonly used due to sampling difficulty at night and during high tide. Although discrete measurements can still provide reliable estimates of the average emission rate over a specific period, they are useful only for broader-scale quantification and for carbon and CH₄ budgeting models. This study highlights the need for continuous measurements of GHG fluxes in coastal ecosystems, which can provide a more detailed understanding of emission patterns, aid in overall emission quantification, help individuals identify key drivers and mechanisms, reduce uncertainty in GHG emissions, and facilitate the assessment of the impacts of specific events or environmental variables (Capooci and Vargas, 2022). However, in contrast to practical, feasible, and cost-effective discrete measurements, continuous measurements require specialized equipment, technical expertise, and intensive labor. It should also be noted that considerable differences were mainly observed at the A-BM site, which had the longest inundation time and the highest water table.

This study provides insights into the potential tidal influence on GHG fluxes from mangrove tree stems. However, several uncertainties require further investigation. First, the study was conducted during summer and daylight hours, which may have resulted in higher fluxes due to the effects of higher temperatures and the sap-flux-dependent transport mechanism within the tree stems (Barba et al., 2019b; Köhn et al., 2021; Pangala et al., 2015; Pitz et al., 2018; Takahashi et al., 2022; Wang et al., 2016; Zhang et al., 2022). Second, the sampling campaign was conducted during the spring tide, and CH₄ fluxes in tidal wetlands may differ between spring and neap tides (Huang et al., 2019; Tong et al., 2013). Third, sampling only at a height of 110 cm may have resulted in height-related GHG flux variations within mangrove tree stems being overlooked, as observed in other studies (Epron et al., 2023; Jeffrey et al., 2019; Moldaschl et al., 2021; Pangala et al., 2013, 2014, 2015; Sjögersten et al., 2020). Finally, with the limited data availability, it is still uncertain whether there is a significant difference in GHG emissions from tree stems between the two mangrove species.

5 Conclusion

This study revealed distinct temporal variations in CO₂ and CH₄ fluxes in the tree stems of A. marina and K. obovata throughout tidal cycles. While GHG fluxes in K. obovata stems displayed inconsistent patterns, CH₄ fluxes in A. marina stems suggested a transition in the stems from being a sink to a source, indicating the influence of different sources and mechanisms during different tidal phases. When considering tidal influences, the stem CH₄ flux in A. marina may vary by up to 1200 %, causing the stem to transition from a net CH₄ sink to a source. This study highlights the need to consider tidal influences and species when quantifying GHG fluxes in mangrove tree stems and the potential limitations of discrete measurements relative to continuous measurements. However, further study is needed to fully understand the underlying mechanisms driving the observed flux variations and to improve our understanding of, and reduce uncertainty in, GHG dynamics in mangrove ecosystems.