Effects of Pneumatophore Density on Methane Emissions in Mangroves

: Mangroves play an important role in carbon sequestration. However, mangroves can be sources of greenhouse gas (GHG) emissions. In this study, methane (CH 4 ) emissions and related soil properties were determined in multiple mangroves in Taiwan, including Kandelia obovata and Avicennia marina mangroves. K. obovata possess prop roots, whereas pneumatophores are found in A. marina . Our results showed that mangrove soils were signiﬁcant sources of CH 4 emissions, which should be accounted for in mangrove carbon budgets. In particular, CH 4 emissions in the A. marina mangroves were approximately 50- to 100-fold those of the K. obovata mangroves and the adjoining mudﬂats. Multiple regression analyses indicated that the soil salinity and pH in K. obovata mangroves and the soil redox potential and organic content in the mudﬂats were the key factors affecting CH 4 emissions. However, the pneumatophore density alone explained approximately 48% of the variation in CH 4 emissions in the A. marina mangroves. More pneumatophores resulted in higher CH 4 emissions in the A. marina mangroves. Thus, compared with the assessed soil properties, the contribution of pneumatophores to the transportation of CH 4 from soil was more signiﬁcant. In addition to soil properties, our results demonstrated that the root structure may also affect GHG emissions from mangroves.


Introduction
Anthropogenic activities, such as deforestation and fossil fuel combustion, have generated a great amount of carbon dioxide (CO 2 ) emissions, with accompanying greenhouse effects and global warming [1,2]. To alleviate climate change, it is essential to develop mitigation strategies to increase the carbon storage capacity and capture ability in natural ecosystems [3,4]. Vegetated coastal ecosystems (VCEs), including mangroves, seagrass meadows, and tidal marshes, are known as the major "blue carbon" sinks, and hold the potential for higher organic carbon (OC) sequestration than terrestrial forests [4,5].
Mangroves provide multiple ecosystem functions and services, such as coastal protection, water quality improvement, nutrient cycling, carbon sequestration, and fisheries [6][7][8]. Expressed as economic value, the mangrove payment of ecosystem services (PES, 91,000 $US dollars ha −1 ) is much greater than that of seagrass meadows (~13,000 $US dollars ha −1 ) and salt marshes (~1500 $US dollars ha −1 ) [9]. Serrano et al. [10] reported that the soil carbon stock in 1-m-thick mangroves in Australia was approximately 251 Mg C ha −1 , which was 1.5~2.5 times greater than the levels in seagrass meadows (112 Mg C ha −1 ) and salt marshes (168 Mg C ha −1 ). In addition, Donato et al. [11] demonstrated carbon storage in mangrove soils to be approximately three to five times higher than that in upland forests in tropical zones. Thus, mangroves play an essential role in sequestering carbon and alleviating climate change.
Recently, mangroves were found to be not only carbon sinks, but also greenhouse gas (GHG) sources [4,12,13]. Frequent tidal inundation has been found to restrict the

Site Description
Nine research sites ( Figure 1) were established along the western coast of Taiwan, from north to south: A-Wazihwei (WZ); B-Zhuwei (ZW); C-Xinfeng (XF); D-Zhunan (ZN); E1-Fangyuan Kandelia obovata (FY-K); E2-Fangyuan Avicennia marina (FY-A); F-Budai (BD); G-Beimen (BM); H-Chiku (CK). The climate is dry and mild in winter and humid and hot in summer (Table 1). K. obovata mangroves dominated at Site A, B, C, and D. A. marina mangroves dominated at Site F, G, and H. At Site E, both species were codominant in the mangroves (E1 and E2). Compared with the A. marina mangroves, a higher density and tree height were observed in the K. obovata mangroves ( Table 1). The tides at all of the mangroves were semidiurnal tides, and the tidal range was highest at Site E and decreased gradually to the north and south. The mean immersion times at the sites during flood periods ranged from 1.2-18.2 h/day. The soil texture for all of the mangroves was silt. The median grain size was the greatest at Site C (0.058 ± 0.017 mm) and the smallest at Site E1 (0.015 ± 0.001 mm). Table 1. Climatic conditions, mangrove characteristics and soil texture at the mangrove sites: A-Wazihwei (WZ); B-Zhuwei (ZW); C-Xinfeng (XF); D-Zhunan (ZN); E1-Fangyuan Kandelia obovata (FY-K); E2-Fangyuan Avicennia marina (FY-A); F-Budai (BD); G-Beimen (BM); H-Chiku (CK). Site E and decreased gradually to the north and south. The mean immersion times at the sites during flood periods ranged from 1.2-18.2 h/day. The soil texture for all of the mangroves was silt. The median grain size was the greatest at Site C (0.058 ± 0.017 mm) and the smallest at Site E1 (0.015 ± 0.001 mm).

Methane Emissions Measurements
The CH 4 emissions and soil properties were determined for a complete seasonal cycle during 2019-2020 (Table 1) within the two species of mangroves and adjoining mudflats (located in the open space outside of mangroves). Mudflats were available for the CH 4 measurements at Site C, D, E2, F, and G. The CH 4 emission measurements were adapted from Lin et al. [13]. Briefly, we used an ultraportable greenhouse gas analyzer (LGR915-0001, Los Gatos Research, San Jose, CA, USA) connected to an in situ closed-path chamber through a polyvinyl chloride (PVC) tube to determine the CH 4 emissions from the soils of the mangroves and mudflats. A semicircular transparent acrylic chamber attached to a stainless-steel ring (with a diameter of 30 cm and height of 15 cm) was placed to avoid crab holes and was inserted into a 10-cm depth of soil to create the air volume (10.6 L) over a 0.071 m 2 surface area. The CH 4 fluxes within the chamber were monitored by the gas analyzer and were documented by a data logger using a 20-s recording interval in 10-min sessions. The CH 4 emissions were determined by Equation (1).
where F is the CH 4 emission in terms of µmol CH 4 m −2 h −1 , S is the slope of linear regression line between CH 4 concentration (ppm) and logging time point (20 s), V is chamber volume (L), R is ideal gas constant = 0.082 (L atm K −1 mol −1 ), T is the absolute temperature (K), A is the bottom area of chamber (i.e., 0.071 m 2 ), and 180 is time conversion constant = (1 h × (60 min/hour) × ((60 s/min)/20 s)). The CH 4 detection limit was 0.01 to 100 ppm. The CH 4 emissions were measured from the soils of the mangroves and mudflats with three to five replications (five in most cases) during emersion periods (2 h before and after the low tide) at each site in each season. The distance between each randomly selected soil sample was at least 5 m in order to eliminate any potential interactions between the soil samples resulting from the site disturbance.
Based on our previous research [13], significant CH 4 emissions were observed in spring and summer. To determine the effects of pneumatophore density on CH 4 emissions, we quantified CH 4 emissions under different densities of pneumatophores in A. marina mangroves at Sites E2 and G in spring or summer. The density of pneumatophores was classified into the following four levels: (1) 0-20, (2) 20-50, (3) 50-80, and (4) at least 80 pneumatophores within the chamber (0.071 m 2 ; Figure 2). The measurements of the CH 4 emissions at each level were carried out in three replicates. To represent the density of pneumatophores, the number of pneumatophores was divided by the surface area (0.071 m 2 ) of the chamber. after the low tide) at each site in each season. The distance between each randomly selected soil sample was at least 5 m in order to eliminate any potential interactions between the soil samples resulting from the site disturbance. Based on our previous research [13], significant CH4 emissions were observed in spring and summer. To determine the effects of pneumatophore density on CH4 emissions, we quantified CH4 emissions under different densities of pneumatophores in A. marina mangroves at Sites E2 and G in spring or summer. The density of pneumatophores was classified into the following four levels: (1) 0-20, (2) 20-50, (3) 50-80, and (4) at least 80 pneumatophores within the chamber (0.071 m 2 ; Figure 2). The measurements of the CH4 emissions at each level were carried out in three replicates. To represent the density of pneumatophores, the number of pneumatophores was divided by the surface area (0.071 m 2 ) of the chamber.

Soil Sample Analyses
The soil properties in the top 10-cm layer were evaluated after CH4 measurements with five and two replications in the mangroves and mudflats, respectively, at each site in each season. In the field, a portable pH meter (WD-35634-40, OAKTON Instruments, Vernon Hills, IL, USA) was applied to measure the soil pH, and a redox potential meter (ORP30, CLEAN L'eau, Taoyuan City, Taiwan) was used to determine the soil ORP and temperature. As the mangrove soils were wet and free of particles, the instrument manual [28] suggested that the soil pH could be measured in situ. The soil pH and ORP were measured by inserting the meters into the soil at a depth of 5 cm. The soil samples were then collected using stainless-steel cores (with a diameter of 7 cm and length of 80 cm), and syringes (with a diameter of 2.9 cm and length of 5 cm) were used to retrieve the top 10 cm of the soil core samples for two subsamples. One subsample was used for the soil bulk density, water, and organic matter content analysis [13,29], while the other was used for the salinity measurement. These subsamples were transferred into 50-mL centrifuge tubes, stored in a portable cooler, and then transported to the laboratory for further analysis. The subsamples were stored in a −20 ℃ freezer and were analyzed within two weeks at the laboratory. At the laboratory, the pore water was derived from the soil subsample with a syringe, and the salinity was then measured using a portable refractometer (Refractometer FG-201, Hangzhou Chincan Trading Co., Ltd., Hangzhou, Zhejiang Province, China).

Statistical Analyses
As the Shapiro-Wilk test results demonstrated that the CH4 emissions and soil property datasets were nonnormally distributed (p-value < 0.05), the nonparametric Kruskal-Wallis test was utilized to determine the differences in CH4 emissions and soil properties among the studied sites and habitat types. When the results showed a significant difference (p-value <0.05), Dunn's test was implemented as a post hoc analysis to determine

Soil Sample Analyses
The soil properties in the top 10-cm layer were evaluated after CH 4 measurements with five and two replications in the mangroves and mudflats, respectively, at each site in each season. In the field, a portable pH meter (WD-35634-40, OAKTON Instruments, Vernon Hills, IL, USA) was applied to measure the soil pH, and a redox potential meter (ORP30, CLEAN L'eau, Taoyuan City, Taiwan) was used to determine the soil ORP and temperature. As the mangrove soils were wet and free of particles, the instrument manual [28] suggested that the soil pH could be measured in situ. The soil pH and ORP were measured by inserting the meters into the soil at a depth of 5 cm. The soil samples were then collected using stainless-steel cores (with a diameter of 7 cm and length of 80 cm), and syringes (with a diameter of 2.9 cm and length of 5 cm) were used to retrieve the top 10 cm of the soil core samples for two subsamples. One subsample was used for the soil bulk density, water, and organic matter content analysis [13,29], while the other was used for the salinity measurement. These subsamples were transferred into 50-mL centrifuge tubes, stored in a portable cooler, and then transported to the laboratory for further analysis. The subsamples were stored in a −20 • C freezer and were analyzed within two weeks at the laboratory. At the laboratory, the pore water was derived from the soil subsample with a syringe, and the salinity was then measured using a portable refractometer (Refractometer FG-201, Hangzhou Chincan Trading Co., Ltd., Hangzhou, Zhejiang Province, China).

Statistical Analyses
As the Shapiro-Wilk test results demonstrated that the CH 4 emissions and soil property datasets were nonnormally distributed (p-value < 0.05), the nonparametric Kruskal-Wallis test was utilized to determine the differences in CH 4 emissions and soil properties among the studied sites and habitat types. When the results showed a significant difference (p-value <0.05), Dunn's test was implemented as a post hoc analysis to determine which sites, habitat types, or levels differed. Stepwise multiple regression analysis was used to evaluate the effects of the soil properties on the CH 4 emissions across the studied sites in the mangrove ecosystems for both species. R software (Version 3.6.1, https://www.r-project.org/) [30] and SigmaPlot 14.0 (Systat Software, San Jose, CA, USA) were applied for the statistical analyses, and p < 0.05 was considered statistically significant in this study.

CH 4 Emissions and Soil Properties
In the mangroves and adjoining mudflats, CH 4 emissions were significantly different among the studied sites ( Figure 3, and Tables 2 and 3). The mean CH 4 fluxes from the mangrove soils ranged from 0.8 to 94.1 µmol-CH 4 m −2 h −1 . Higher CH 4 emissions were observed in the soils of the Avicennia marina mangroves (Figure 3 and Table 2). Compared with the Kandelia obovata mangroves (−1.7 to 16.6 µmol-CH 4 m −2 h −1 ) and mudflats (−0.7 to 7.6 µmol-CH 4 m −2 h −1 ), CH 4 emissions were significantly greater in the A. marina mangroves (2.1 to 765.9 µmol-CH 4 m −2 h −1 ; Figure 4). Soil properties other than the soil temperature varied distinctly among the mangroves (Table 2). However, only the soil bulk density, organic matter, and water content differed distinctly among the mudflats. which sites, habitat types, or levels differed. Stepwise multiple regression analysis was used to evaluate the effects of the soil properties on the CH4 emissions across the studied sites in the mangrove ecosystems for both species. R software (Version 3.6.1, https://www.r-project.org/) [30] and SigmaPlot 14.0 (Systat Software, San Jose, CA, USA) were applied for the statistical analyses, and p < 0.05 was considered statistically significant in this study.

CH4 Emissions and Soil Properties
In the mangroves and adjoining mudflats, CH4 emissions were significantly different among the studied sites ( Figure 3, and Tables 2 and 3). The mean CH4 fluxes from the mangrove soils ranged from 0.8 to 94.1 µ mol-CH4 m −2 h −1 . Higher CH4 emissions were observed in the soils of the Avicennia marina mangroves (Figure 3 and Table 2). Compared with the Kandelia obovata mangroves (−1.7 to 16.6 µ mol-CH4 m −2 h −1 ) and mudflats (−0.7 to 7.6 µ mol-CH4 m −2 h −1 ), CH4 emissions were significantly greater in the A. marina mangroves (2.1 to 765.9 µ mol-CH4 m −2 h −1 ; Figure 4). Soil properties other than the soil temperature varied distinctly among the mangroves (Table 2). However, only the soil bulk density, organic matter, and water content differed distinctly among the mudflats.

Effects of Soil Properties and Pneumatophores on Methane Emissions in Mangrove Ecosystems
The results of the multiple regressions demonstrated that the soil pH and salinity exerted significant negative effects on CH 4 emissions (CH 4 flux = −3.19 × pH − 0.50 × salinity + 24.69; p-value < 0.01; R 2 = 0.43) in the K. obovata mangroves. There was an exponential relation between salinity and CH 4 flux, but a linear relationship between pH and CH 4 flux ( Figure 5). However, nonsignificant correlations were found between the soil properties and CH 4 emissions in the A. marina mangroves. In the mudflats, the CH 4 emissions were negatively related to the soil ORP, but positively related to the organic matter content (CH 4 flux = −0.01 × ORP + 0.65 × organic matter content − 0.45, p-value < 0.01, R 2 = 0.57) (Figure 6). The results demonstrate that the CH 4 flux ranged from 4.8 to 161.3 µmol-CH 4 m −2 h −1 with 57 to 1814 pneumatophores m −2 , and greater numbers of pneumatophores produced higher CH 4 emissions in the A. marina mangroves (Figure 7).
The current study found that CH4 emissions in the A. marina mangroves were approximately 50-to 100-fold those of the K. obovata mangroves and adjoining mudflats (Figure 4). The CH4 emissions in vegetated soils were generally higher than those in unvegetated soils (mudflats), as more labile organic compounds originating from plants were provided for methanogenic communities [32][33][34]. However, this observation was not necessarily recapitulated in the K. obovata mangroves in this study. This suggests that differences in mangrove species and root structure may lead to various patterns of GHG emissions.
As GHGs were generated mainly as a result of microbial activities that were greatly affected by soil characteristics, GHG emissions were expected to be regulated by the soil properties [31]. Indeed, the multiple regression analyses indicated that the soil salinity and pH in the K. obovata mangroves, and the soil redox potential and organic matter content in the mudflats were the main soil properties influencing CH4 emissions. Previous findings have reported salinity to be inversely related to CH4 emissions in mangroves, with relationships illustrated by either linear [35] or logarithmic regressions [18]. Sulfate is among the most abundant ions in seawater. In high salinity environments, the higher availability of sulfate ions enhanced the activities of sulfate-reducing bacteria, which might inhibit methanogenic bacteria obtaining organic substrates; thus, less CH4 emissions were produced [19,[36][37][38][39]. Methanogens have been found to be sensitive to soil pH, and demonstrated better growth in the pH range of 6 to 8 [40,41]. It appears that the soil pH in the K. obovata mangroves was favorable for methanogenesis. The soil organic matter content exhibited a significantly positive correlation with CH4 emissions, as high levels of organic matter provide additional carbon sources for methanogens, generating a greater quantity of CH4 [14,23,42]. In addition, methanogenic bacteria are highly active in lower redox potential conditions [26,30,43]. As a result, higher CH4 emissions were observed from anaerobic soils with a high organic matter in the mudflats.

Discussion
Large variations in CH 4 emissions were observed at the nine mangroves (−1.7 to 765.9 µmol-CH 4 m −2 h −1 ; Figure 4) as a result of spatial and habitat differences. However, compared with other studies, CH 4 emissions in the studied mangrove ecosystems were still within the range of previous estimates (Kandelia obovata, 0.8 to 72.5 µmol-CH 4 m −2 h −1 , Chen et al. [31]; Avicennia marina, 0.2 to 1087.5 µmol-CH 4 m −2 h −1 , Allen et al. [21]; mudflat, 5.6 to 87.5 µmol-CH 4 m −2 h −1 , Xiang et al. [32]). As a result, the studied mangrove soils were significant sources of CH 4 emissions, which should be accounted for in the carbon budgets of mangrove ecosystems.
The current study found that CH 4 emissions in the A. marina mangroves were approximately 50-to 100-fold those of the K. obovata mangroves and adjoining mudflats (Figure 4). The CH 4 emissions in vegetated soils were generally higher than those in unvegetated soils (mudflats), as more labile organic compounds originating from plants were provided for methanogenic communities [32][33][34]. However, this observation was not necessarily recapitulated in the K. obovata mangroves in this study. This suggests that differences in mangrove species and root structure may lead to various patterns of GHG emissions.
As GHGs were generated mainly as a result of microbial activities that were greatly affected by soil characteristics, GHG emissions were expected to be regulated by the soil properties [31]. Indeed, the multiple regression analyses indicated that the soil salinity and pH in the K. obovata mangroves, and the soil redox potential and organic matter content in the mudflats were the main soil properties influencing CH 4 emissions. Previous findings have reported salinity to be inversely related to CH 4 emissions in mangroves, with relationships illustrated by either linear [35] or logarithmic regressions [18]. Sulfate is among the most abundant ions in seawater. In high salinity environments, the higher availability of sulfate ions enhanced the activities of sulfate-reducing bacteria, which might inhibit methanogenic bacteria obtaining organic substrates; thus, less CH 4 emissions were produced [19,[36][37][38][39]. Methanogens have been found to be sensitive to soil pH, and demonstrated better growth in the pH range of 6 to 8 [40,41]. It appears that the soil pH in the K. obovata mangroves was favorable for methanogenesis. The soil organic matter content exhibited a significantly positive correlation with CH 4 emissions, as high levels of organic matter provide additional carbon sources for methanogens, generating a greater quantity of CH 4 [14,23,42]. In addition, methanogenic bacteria are highly active in lower redox potential conditions [26,30,43]. As a result, higher CH 4 emissions were observed from anaerobic soils with a high organic matter in the mudflats.
Compared with the K. obovata mangroves and mudflats, however, the multiple regression results demonstrated that none of the soil properties explained the large proportion of variation in CH 4 emissions from the A. marina mangroves. Similar results were also observed by Lin et al. [13]. As the pneumatophores provided gas exchange conduits between soils and the atmosphere, CH 4 was likely emitted through the pneumatophores from the deep soil [19,25,26]. Our results showed that the density of pneumatophores in the A. marina mangroves explained approximately 48% of the variation in CH 4 emissions (Figure 7). Livesley et al. [26] reported that 24% of the variation in CH 4 flux resulted from pneumatophores. Thus, the significance of pneumatophores in transporting CH 4 from the soil to the atmosphere in mangroves cannot be neglected. In summary, the soil properties and pneumatophores were the primary factors affecting CH 4 emissions in K. obovata and A. marina mangroves, respectively.
To further evaluate the effects of pneumatophore density on CH 4 emissions in different climatic regions in A. marina mangroves, our results were compared with other data derived from previous research [25,26,36,44]. The linear regression lines of CH 4 emissions on the pneumatophore density presented in previous studies were transformed into the same units (µmol-CH 4 m −2 h −1 ) and were redrawn in this study (Figure 8). Even though the sampling times were different between the studies (Figure 8), the results indicate that the slopes of the regression lines were always steeper in tropical regions than in subtropical or temperate regions, despite the high variation in slopes within tropical regions. As higher temperatures facilitate methanogen activity in soils [18,21], these comparisons suggest that with the same density of pneumatophores, more CH 4 is emitted from A. marina mangroves in tropical regions. In addition, CH 4 emissions were positively correlated with pneumatophore density in A. marina mangroves across all climate regions.
Compared with the K. obovata mangroves and mudflats, however, the multiple regression results demonstrated that none of the soil properties explained the large proportion of variation in CH4 emissions from the A. marina mangroves. Similar results were also observed by Lin et al. [13]. As the pneumatophores provided gas exchange conduits between soils and the atmosphere, CH4 was likely emitted through the pneumatophores from the deep soil [19,25,26]. Our results showed that the density of pneumatophores in the A. marina mangroves explained approximately 48% of the variation in CH4 emissions (Figure 7). Livesley et al. [26] reported that 24% of the variation in CH4 flux resulted from pneumatophores. Thus, the significance of pneumatophores in transporting CH4 from the soil to the atmosphere in mangroves cannot be neglected. In summary, the soil properties and pneumatophores were the primary factors affecting CH4 emissions in K. obovata and A. marina mangroves, respectively.
To further evaluate the effects of pneumatophore density on CH4 emissions in different climatic regions in A. marina mangroves, our results were compared with other data derived from previous research [25,26,36,44]. The linear regression lines of CH4 emissions on the pneumatophore density presented in previous studies were transformed into the same units (µ mol-CH4 m −2 h −1 ) and were redrawn in this study (Figure 8). Even though the sampling times were different between the studies (Figure 8), the results indicate that the slopes of the regression lines were always steeper in tropical regions than in subtropical or temperate regions, despite the high variation in slopes within tropical regions. As higher temperatures facilitate methanogen activity in soils [18,21], these comparisons suggest that with the same density of pneumatophores, more CH4 is emitted from A. marina mangroves in tropical regions. In addition, CH4 emissions were positively correlated with pneumatophore density in A. marina mangroves across all climate regions.  [25]; subtropical: Kreuzwieser et al. [44]; temperate: Livesley et al. [26]. Sampling time: tropical 1: pre-monsoon, monsoon, post-monsoon, and summer; tropical 3: dry and wet; subtropical: spring and winter; temperate: spring, summer, fall, and winter.

Conclusions
In this study, CH 4 emissions and soil characteristics were determined in multiple mangroves of Taiwan, including in the Kandelia obovata (subtropical zone) and Avicennia marina (tropical zone) mangroves. Our results showed that CH 4 emissions in the A. marina mangroves were significantly greater than those in the K. obovata mangroves and the adjoining mudflats. The multiple regression analyses indicated that the soil salinity and pH in the K. obovata mangroves, and the soil redox potential and organic matter content in the mudflats, were the main soil properties affecting CH 4 emissions across the studied mangroves. However, this study also showed that pneumatophore density alone in A. marina man-groves explained approximately 48% of the variation in CH 4 emissions. Thus, compared with the assessed soil properties, the contribution of pneumatophores to the transportation of CH 4 from the soils was more significant. In addition to the soil properties, our results demonstrated that the root structure may also affect GHG emissions in mangroves.