Potential effects of sea level rise on the soil-atmosphere greenhouse gas emissions in Kandelia obovata mangrove forests

1. Introduction

The accelerating rate of sea level rise (SLR) is causing coastal ecosystem changes on a global scale (Meeder et al., 2021). The global mean sea level is projected to rise between 0.28 m and 0.98 m by the end of the 21st century based on the sea level of 2000 according to the Representative Concentration Pathway Scenarios, hence posing a major challenge for coastal regions worldwide (IPCC, 2013; Jayanthi et al., 2018; Perera et al., 2018; Mafi-Gholami et al., 2020). Mangrove forests as the only forests between land and sea are vulnerable to SLR (Krauss et al., 2014; Lovelock et al., 2015; Chen and Wang, 2017; Langston et al., 2017; Minick et al., 2019; Kirwan and Gedan, 2019). If soil accumulation rate is less than the rate of sea level rise, SLR will change water movement, alter tidal variation, and increase seawater intrusion into mangrove forests, resulting in both deeper flooding level, longer flooding duration and more anaerobic conditions (Ye et al., 2010; Wang et al., 2019). Therefore, SLR potentially has a substantial effect on the mangrove ecosystem.

The global wetland area accounts for 10% of the earth’s surface area, but stores 15% of earth’s total terrestrial carbon (Davidson et al., 2018; Sheng et al., 2021; Zhang et al., 2021). Although mangrove forests occupy a limited area compared with terrestrial forests, mangrove forests are increasingly seen as highly productive and most carbon-rich ecosystems among all coastal wetlands (Donato et al., 2011; McLeod et al., 2011). The average global carbon stock in the mangrove soils is approximately 720 t/hm² (Alongi, 2014). Carbon stored in the mangrove ecosystems, along with salt marshes and seagrass beds, is termed “blue carbon” and has become a topic of global interest (Capooci et al., 2019). Meanwhile, mangrove ecosystems are known to constitute a net source of greenhouse gases including carbon dioxide (CO₂), methane (CH₄) and nitrous oxide (N₂O) (Chen et al., 2010, 2012, 2015, 2016). Emissions of these three greenhouse gases exchanged between mangrove soils and the atmosphere, reduce soil carbon stock and the contribution of mangrove plants to the atmospheric cooling effect (Chen et al., 2015, 2016). For CO₂, this emission rates from mangrove soil tended to be higher when soils were exposed to the atmosphere than when they were immersed by tides, due to the faster molecular diffusion for gases than fluids, as well as the increased surface area for aerobic respiration and chemical oxidation (Chen et al., 2012). For CH₄, the increase in the ${{\rm {SO}}_4^{2-}}$ carried by tidal flooding due to SLR could suppress CH₄ production (Stumm and Morgan, 1981). In addition, N₂O production is favored due to the alternation between dry and wet conditions and is more likely to occur during low levels of O₂ (Liikanen and Martikainen, 2003; Jørgensen and Elberling, 2012; Capooci et al., 2019). Changes of mangrove soil conditions due to SLR will affect mangrove greenhouse gas emissions, further influencing mangrove blue carbon capability. Although previous studies (Rogers et al., 2019; Chen et al., 2020b) have found that SLR will increase the stocks of organic carbon (OC) in mangrove soils, the direct field observations for responses of greenhouse gas emissions from mangrove ecosystems to SLR remain unexplored.

Therefore, in the present study, potential effects of SLR on the soil-atmosphere greenhouse gas emissions (CO₂, CH₄ and N₂O) were quantified among three sites under different intertidal elevations in 9 years-old Kandelia obovata mangrove forests, representing tidal flooding situations of different SLR values. We posed the following research questions: (1) How higher sea levels could affect mangrove soil-atmosphere greenhouse gas emissions? (2) What are possible mechanisms of changes in mangrove soil-atmosphere greenhouse gas emissions with SLR?

2. Materials and methods

2.1 Study area and plot arrangements

The study area is located along the coastline of Zengying (24°33′9.24″N, 118°02′3.71″E) in Xiamen, Fujian, China (Fig. 1). This region is characterized by a subtropical marine monsoonal climate, with annual mean rainfall of 1097 mm and the rainy season from April to June. Mean annual air temperature is 20.8℃ and average annual insolation duration is 2276 h (Chen and Ye, 2013). March to May, June to August, September to November and December to February are categorized as spring, summer, autumn and winter months, respectively. Tides are semi-diurnal, with an average range of 4 m. Salinity of seawater adjacent to mangrove area ranges from 26 to 30 (Chen and Ye, 2013, 2014). Altitude of local mean sea level in Xiamen Bay is 365 cm (Ruan et al., 2010).

Figure  1.  Maps of Zengying, the study area. SLR: sea level rise.

DownLoad: Full-Size Img PowerPoint

Many aquaculture ponds with boulder strips were used to breed crabs in this intertidal zone before 2002. These ponds were deserted and cofferdams around them were demolished after 2002. Soil with similar soil physico-chemical properties beside these ponds were used to transform three adjacent ponds for mangrove planting sites (Chen and Ye, 2013, 2014; Chen et al., 2020b). The final steady differences in elevation between two nearby sites were 40 cm, with sea level altitudes of 335 cm, 375 cm and 415 cm for the lower (SLR 80 cm), middle (SLR 40 cm) and upper (SLR 0 cm) elevation, respectively. According to the Ocean Blue Book on Climate Change in China (Liu et al., 2019), the average increase in sea level in China was 3.3 mm/a (33 cm per 100 a, 66 cm per 200 a) from 1980 to 2018. Elevation differences among the three sites were used to represent different tidal flooding situations that might be expected with rising sea level. Sea levels at Sites SLR 0 cm, SLR 40 cm and SLR 80 cm represented tidal flooding time of the current (Site SLR 0 cm), future ~100 a’s (Site SLR 40 cm) and future ~200 a’s SLR (Site SLR 80 cm) of mangrove forests along the Jiulong River Estuary, approximately 22 km away from the study area. Sites SLR 80 cm, SLR 40 cm and SLR 0 cm had mean inundation time of 10 h/d, 8 h/d and 6 h/d (i.e., 5 h, 4 h and 3 h per tide), respectively, measured during one whole tide cycle of 15 d in August 2004 (Chen and Ye, 2014). According to our previous study (Chen et al., 2020b), before mangrove planting, the soils among the three sites have similar soil physico-chemical properties including pH (p=0.452, one-way ANOVA), water content (p=0.566, one-way ANOVA), soil salinity (p=0.628, one-way ANOVA) and soil OC content (p=0.977, one-way ANOVA). In May 2004, the healthy mature K. obovata propagules were planted at the three sites and K. obovata among the three sites became the dominant species with 9 years-old at the sampling time of the present study. In the field observation, pollutants (e.g., wastewater from aquaculture) which affected mangrove soil-atmosphere greenhouse gas emissions were not found to enter into the three sites from land and sea.

2.2 Gas sampling and flux measurement

Methods of collecting soil-atmosphere greenhouse gas fluxes have been reported in the previous studies (Chen et al., 2010, 2015, 2016). In brief, soil-atmosphere greenhouse gas fluxes were sampled monthly throughout the whole year from June 2013 to May 2014 at the three sites. Three replicates (each was approximately 10 m apart) for each site were applied in gas sampling using static closed chamber technique, and all samplings were conducted two hours before the lowest tide during the daytime. The static transparent chambers with a diameter of 17 cm (total area coverage of 0.023 m²) and a headspace volume over the soil of 1.47 L were applied in sampling. Each chamber was lightly inserted 2.5 cm into the soil at locations with no aboveground vegetation or litter fall. At 0 min, 15 min and 30 min after closure, a 5 mL gas sample was collected by passing a 10 mL glass syringe through the sampling outlet and was then immediately injected into dark airtight containers which were made of rubber stopper and aluminum foil.

During sample analysis, 1 mL gas sample was analyzed within 12 h using a gas chromatography system (Agilent 7890A, Santa, Clara, CA, USA) to determine concentrations of N₂O, CH₄ and CO₂ and the standard was analyzed in every 6 samples (Chen et al., 2010, 2015, 2016). N₂O concentration was determined with a ⁶³Ni electron capture detector (μECD). CH₄ and CO₂ concentrations were analyzed with a flame ionization detector (FID). Gas concentrations were determined by comparing the sample peak areas against the standard curves (Agilent Greenhouse Gas Checkout Sample). Each soil-atmosphere greenhouse gas flux was calculated using the difference between the two points in the time series of concentration for each plot.

The soil-atmosphere flux (F, mol/(m²·h)) of each greenhouse gas was calculated by following formula:

where F was the interfacial gas flux (mol/(m²·h)), V was the internal air volume (m³) in the container after the pot was placed, $\Delta $M (h^–1) was the change in gas concentration in the container, A was the surface area of the soil (m²) and P was the volume of per molar gas (m³/ mol).

2.3 Soil sampling and analysis

Soil samples (5 cm depth) were immediately collected for measurement of soil properties including organic carbon (OC), total nitrogen (TN) and total phosphorus (TP) after gas sampling was completed. Rapid dichromate oxidation procedure was used to determine soil OC content and a continuous flow analyzer (Futura II, Alliance Instruments, France) was used to determine soil TN and TP after Kjeldahl digestion (Allen et al., 1974). All data were expressed as the 105℃ dry weight. Moreover, air temperature which was used to calculate soil-atmosphere greenhouse gas fluxes was measured by atmosphere thermometer (digital thermometer, TP101) during sampling period.

2.4 Estimates of annual greenhouse gas emission and CO₂-equivalent fluxes

Each annual greenhouse gas flux was further estimated based on the summation for monthly measurement during the entire year from June 2013 to May 2014. Annual gas fluxes were then converted to CO₂-equivalent fluxes (F_e, g/(m²·a)) by the following equation according to Myhre et al. (2013):

where $R_{{\rm{N}}_2{\rm{O}}} $, $R_{{\rm{CH}}_4} $ and $R_{{\rm{CO}}_2} $ denote annual emission rates of N₂O (g/(m²·a)), CH₄ (g/(m²·a)) and CO₂ (g/(m²·a)), respectively.

2.5 Statistical analyses

Normality of each variable was examined using Kolmogorov–Smirnov test, and homoscedasticity was examined with Kaiser-Meyer-Olkin index and Bartlett’s test of sphericity. Main and interactive effects of SLR and sampling months on each greenhouse gas flux and soil properties were analyzed using two-way analysis of variance (ANOVA), with pairwise comparisons of interactive effects conducted using post hoc Tukey’s HSD test. Pearson correlation was used to describe relationships between greenhouse gas fluxes and soil properties. All statistical analyses were performed using SPSS 22.0 for Windows (SPSS, Chicago, IL, USA).

3. Results

3.1 Greenhouse gas fluxes

Both SLR (F=9.522, P<0.001) and months (F=4.479, P<0.001) had significant impacts on the N₂O fluxes (Table 1). This parameter showed a decreasing order of Site SLR 0 cm ((0.57±0.28) μmol/(m²·h)), Site SLR 80 cm ((0.41±0.26) μmol/(m²·h)), Site SLR 40 cm ((0.28±0.14) μmol/(m²·h)). Significantly higher N₂O fluxes were determined either in spring or summer months than that in autumn or winter months (Fig. 2). The lowest values were obtained in October at Sites SLR 40 cm and SLR 0 cm while in November at Site SLR 80 cm.

Table  1.  The interfacial gas flux (F) values of two-way ANOVA tests showing the effects of sea level rise (SLR) and month on the soil-atmosphere greenhouse gas fluxes and soil properties

Sources of variation	df	Parameter
		OC	TN	TP	C/N	N2O flux	CH4 flux	CO2 flux
SLR	2	257.387***	168.866***	74.294***	31.207***	9.522***	0.273	62.963***
Month	11	14.318***	22.277***	17.372***	2.588**	4.479***	9.563***	56.088***
Interaction	22	7.284***	12.441***	3.778***	3.312***	1.357	2.195*	6.589***
Note: *, ** and *** indicate significant difference at p<0.05, p<0.01 and p<0.001, respectively. OC: organic carbon; TN: total nitrogen; TP: total phosphorus; C/N: C/N content ratio; SLR: sea level rise.

 | Show Table

DownLoad: CSV

Figure  2.  Temporal variation of N₂O, CH₄ and CO₂ fluxes from soils at the three Kandelia obovata forests (mean±SE, SLR: sea level rise).

DownLoad: Full-Size Img PowerPoint

In contrast to N₂O fluxes, significant differences of CH₄ fluxes were only found among different months (F=9.563, P<0.001) (Table 1). The lowest values were found in December at Sites SLR 80 cm and SLR 0 cm while in October at Site SLR 40 cm, and this parameter peaked in August at Sites SLR 0 cm and SLR 40 cm, and September at Site SLR 80 cm (Fig. 2). Contrary to the trend in the N₂O fluxes, annual CH₄ fluxes were highest at Site SLR 40 cm ((3.56±2.57) μmol/(m²·h)), followed by Site SLR 80 cm ((3.41±2.88) μmol/(m²·h)) and Site SLR 0 cm ((3.13±2.81) μmol/(m²·h)).

Annual CO₂ fluxes were (0.87±3.48) mmol/(m²·h), (3.56±3.58) mmol/(m²·h) and (3.47±1.71) mmol/(m²·h) at Sites SLR 80 cm, SLR 40 cm and SLR 0 cm, respectively. SLR (F=62.963, P<0.001), month (F=56.088, P<0.001) and their interaction (F=6.589, P<0.001) had significant effects on the CO₂ fluxes (Table 1, Fig. 2). This parameter showed an obvious temporal variation at all three sites, with lower values in autumn or winter months than that in spring or summer months. Although mangrove soils at Sites SLR 40 cm and SLR 80 cm were sources of CO₂ over the entire year, they occasionally showed consumption of atmospheric CO₂ especially during the cold months.

3.2 Soil properties

SLR (F=257.387, P<0.001) and month (F=14.318, P<0.001) had a significant interaction in terms of soil OC content at the three sites, which gradually decreased with SLR (Table 1, Fig. 3). Soil OC contents were (5.48±1.07) mg/g, (9.55±2.31) mg/g and (11.21±2.06) mg/g at Sites SLR 80 cm, SLR 40 cm and SLR 0 cm, respectively. At the three sites, a significant temporal trend of soil OC content was detected, with the highest value occurring either in spring or summer months.

Figure  3.  Temporal variation in surface (0−5 cm) soil organic carbon (OC), total nitrogen (TN), total phosphorus (TP) and C/N content ratio at the three Kandelia obovata forests (mean±SE, SLR: sea level rise).

DownLoad: Full-Size Img PowerPoint

Soil TN contents were (0.58±0.10) mg/g, (0.83±0.25) mg/g and (0.97±0.22) mg/g at Sites SLR 80 cm, SLR 40 cm and SLR 0 cm, respectively. SLR (F=168.866, P<0.001) and month (F=22.277, P<0.001) had a significant interaction (F=12.441, P<0.001) on the soil TN contents, which decreased with SLR (Table 1, Fig. 3). Significantly higher soil TN contents were determined in summer at Site SLR 40 cm, in autumn at Site SLR 0 cm while in spring at Site SLR 80 cm. Differences were observed throughout the whole year, with the values exhibiting the following order: SLR 80 cm < SLR 40 cm < SLR 0 cm. C/N content ratios were also significantly affected by SLR (F=31.207, P<0.001) and month (F=2.588, P<0.01) (Table 1), and the values were comparable between Sites SLR 40 cm and SLR 0 cm, with lower values observed at Site SLR 80 cm (Fig. 3).

Soil TP contents were at a less variable level throughout the entire year with mean values of (0.45±0.08) mg/g, (0.56±0.10) mg/g and (0.59±0.07) mg/g at Sites SLR 80 cm, SLR 40 cm and SLR 0 cm, respectively. SLR (F=74.294, P<0.001) and month (F=17.372, P<0.001) had a significant interaction (F=3.778, P<0.001) on this parameter which decreased with SLR (Table 1, Fig. 3). Significant temporal trends of soil TP content were detected with the highest value occurring either in spring or summer months at the three sites.

3.3 Relationships between greenhouse gas and soil parameters

As shown in Table 2, there was a significantly positive correlation between N₂O flux and soil TP content. CH₄ flux was significantly positively correlated with soil TN content, but not with soil TP content. In comparison to N₂O and CH₄ fluxes, CO₂ fluxes were strongly positively correlated with soil OC, TN and TP contents.

Table  2.  Pearson correlation coefficients (r) between soil-atmosphere greenhouse gas fluxes and soil properties

Greenhouse gas	OC	TN	TP	C/N
N2O	0.133	0.168	0.232*	0.009
CH4	0.171	0.299**	0.143	–0.163
CO2	0.534**	0.537**	0.536**	0.121
Note: * and ** indicate significant r value at p<0.05 and p<0.01, respectively. OC: organic carbon; TN: total nitrogen; TP: total phosphorus; C/N: C/N content ratio.

 | Show Table

DownLoad: CSV

3.4 Annual emission of greenhouse gases and CO₂-equivalent fluxes

Annual N₂O emissions at Sites SLR 80 cm and SLR 40 cm were 159.83 mg/(m²·a) and 108.98 mg/(m²·a), which were lower than that at Site SLR 0 cm (218.65 mg/(m²·a)). On the contrary, compared with annual CH₄ emissions at Site SLR 0 cm, the values at Sites SLR 80 cm and SLR 40 cm increased by 7.4% and 13.5%. Furthermore, annual CO₂ emission was comparable between Sites SLR 40 cm and SLR 0 cm, which was approximately 4 times higher than that at Site SLR 80 cm (Table 3).

Table  3.  Estimation of annual emission of N₂O, CH₄ and CO₂ based on monthly greenhouse gas flux measurements from June 2013 to May 2014

Emission	SLR 80 cm	SLR 40 cm	SLR 0 cm
N2O/(mg·m–2·a–1)	159.83	108.98	218.65
CH4/(mg·m–2·a–1)	475.87	502.77	442.92
CO2/(g·m–2·a–1)	332.81	1373.72	1340.00
Note: SLR: sea level rise.

 | Show Table

DownLoad: CSV

The CO₂-equivalent fluxes were estimated by the summation of three gas fluxes with values of 396.61 g/(m²·a), 1423.29 g/(m²·a) and 1420.21 g/(m²·a) at Sites SLR 80 cm, SLR 40 cm and SLR 0 cm, respectively. The minimum contributor to greenhouse effect was CH₄, only accounting for 1.06%–4.08% of CO₂-equivalent fluxes at the three sites. When it came to the greatest contributor to greenhouse effect, it was CO₂ for Sites SLR 0 cm, SLR 40 cm and SLR 80 cm, with a proportion reaching 83.91%–96.52% (Fig. 4).

Figure  4.  CO₂-equivalent fluxes at the three Kandelia obovata forests. SLR: sea level rise.

DownLoad: Full-Size Img PowerPoint

4. Discussion

Mangrove soils are known to constitute a net source of CO₂, CH₄ and N₂O (Chen et al., 2010, 2012, 2015, 2016). As shown in Table 4, in the previous studies, the fluxes of CO₂, CH₄ and N₂O ranged from 0.027 mmol/(m²·h) to 20.56 mmol/(m²·h) (Poungparn et al., 2009; Chen et al., 2010, 2015), from –130 μmol/(m²·h) to 5168.6 μmol/(m²·h) (Chauhan et al., 2008; Chen et al., 2010, 2015), from 0.13 μmol/(m²·h) to 126 μmol/(m²·h) (Corredor et al., 1999; Chen et al., 2010, 2015, 2020a), respectively. In the present study, the fluxes of CO₂, CH₄ and N₂O at Site SLR 0 cm were 3.47 mmol/(m²·h), 3.13 μmol/(m²·h) and 0.57 μmol/(m²·h), respectively, which were in broad agreement with records from previous studies. However, the fluxes of CO₂, CH₄ and N₂O at Sites SLR 40 cm and SLR 80 cm were changed due to SLR. Compared with Site SLR 0 cm, N₂O decreased by approximately 27.3% at Site SLR 80 cm and 50.2% at Site SLR 40 cm. This flux is closely controlled by water level (tidal influence), soil oxygen availability, soil salinity and soil nutrient environments (Menyailo et al., 1997; Inubushi et al., 1999; Chen et al., 2010; Jørgensen and Elberling, 2012). On the one hand, N₂O may be converted to N₂ or consumed before it gets emitted to the atmosphere depending upon the soil saturation status (Heincke and Kaupenjohann, 1999; Clough et al., 2005; Chapuis-Lardy et al., 2007; Chen et al., 2012). Higher soil water content found at Sites SLR 40 cm and SLR 80 cm resulted in soil in the saturation status (Chen et al., 2022b). N₂O may be converted to N₂ or consumed, resulting in lower N₂O fluxes at Sites SLR 80 cm and SLR 40 cm. On the other hand, salinity has been shown to inhibit complete denitrification (N₂O was converted to N₂ or consumed) due to the decrease in the proportion of N₂O reducing microorganisms with increasing salinization, potentially allowing for a build-up of N₂O (Menyailo et al., 1997; Inubushi et al., 1999). Low salinity levels at Sites SLR 80 cm and SLR 40 cm caused by SLR (Chen et al., 2022b) did not inhibit complete denitrification, resulting in less build-up of N₂O in the soils. Meanwhile, lower soil TN content, including inorganic nitrogen, which indirectly represented soil nutrient environments at Sites SLR 40 cm and SLR 80 cm, may limit microbes to generate N₂O which could explain the phenomenon in the present study.

Table  4.  Soil-atmosphere fluxes of CO₂, CH₄ and N₂O from some mangrove forests

Location	CO2 flux/(mmol·m–2·h–1)	CH4 flux/(μmol·m–2·h–1)	N2O flux/(μmol·m–2·h–1)	Reference
Bay of La Parguera, Southwest Puerto Rico;Bird Island, Southwest Puerto Rico;Magueyes Island, Southwest Puerto Rico	ND	8.57–188.89	0.13–8.58	Corredor et al. (1999)
Bhitarkanika, eastern coast of India	ND	5.09–179.44	0.23–20.79	Chauhan et al. (2008)
The secondary mangrove forests of eastern Thailand	0.027–0.052	ND	ND	Poungparn et al. (2009)
Maipo Natural Reserve, Hong Kong, South China;Futian mangrove forest, Shenzhen, South China	0.69–20.56	11.9–5168.6	0.14–23.83	Chen et al. (2010)
Jiulong River Estuary Mangrove Reserve	0.85–3.20	2.00–21.70	0.13–1.96	Chen et al. (2015)
Jiulong River Estuary Mangrove	ND	–130–1800	4–126	Chen et al. (2020)
Note: ND: no determination of the in situ gas flux.

 | Show Table

DownLoad: CSV

Contrary to N₂O fluxes, CH₄ fluxes at Sites SLR 80 cm and SLR 40 cm were higher than that at Site SLR 0 cm. CH₄ fluxes are controlled by water levels and CH₄ fluxes will be discharged in large quantities at high water levels (Christensen et al., 2003; Wei et al., 2020). In the present study, higher soil water levels were found at Sites SLR 80 cm and SLR 40 cm compared with Site SLR 0 cm (Chen et al., 2022b). Water levels at Sites SLR 80 cm and SLR 40 cm may surpass the threshold which controlled CH₄ fluxes, resulting in higher CH₄ fluxes at Sites SLR 80 cm and SLR 40 cm. On the other hand, CH₄ fluxes in the mangrove environments were enhanced at the low salinity due to the intense oxidation or an alleviation of competition by the more energetically efficient sulfate- and nitrate-reducing bacteria than methanogenic archaea (Chen et al., 2010, 2018; Poffenbarger et al., 2011). Low soil salinity found at Sites SLR 80 cm and SLR 40 cm potentially decreased ${\rm{SO}}_4^{2-} $ availability which may be beneficial to methanogenic archaea, and stimulated CH₄ production, as reported in the other study (Stumm and Morgan, 1981). Finally, pH also affects methanogenic activity and the methanogenic archaea microbes are pH sensitive, and most of them grew over a relatively narrow pH range of about 6−8, with an optimum pH of 7.7 for activity of methane production methanogens in coastal wetlands (Chang and Yang, 2003). Soil pH increased with SLR (Chen et al., 2022b) and this parameter maybe another factor which stimulated CH₄ production at Sites SLR 40 cm and SLR 80 cm in the present study.

In the former studies, a negative relationship between water content and CO₂ fluxes was found (e.g., Alongi, 2009; Chen et al., 2010, 2012; Chambers et al., 2013). In the present study, mean inundation times at Sites SLR 80 cm, SLR 40 cm and SLR 0 cm were 10 h/d, 8 h/d and 6 h/d (i.e., 5 h, 4 h and 3 h per tide), respectively. Although higher soil water content was determined at Site SLR 40 cm, annual CO₂ fluxes at Sites SLR 40 cm and SLR 0 cm were similar. CO₂ fluxes decreased by approximately 75% only when the sea level rose to 80 cm. There may be a threshold of water levels that controls CO₂ fluxes and this threshold was opened at Site SLR 80 cm, resulting in low CO₂ fluxes at Site SLR 80 cm. Meanwhile, sulfate effect was a main biogeochemical driver that influenced OC cycling in the wetland soils (Chambers et al., 2013). The lowest soil salinity at Site SLR 80 cm decreased ${\rm{SO}}_4^{2-} $ availability and decrease of ${\rm{SO}}_4^{2-} $ availability inhibited sulfate reduction, which may result in a decrease of CO₂ fluxes. Furthermore, microbial biomass plays an important role in mangrove OC stocks, such as key role of root symbiotic microbiota in the decomposition of organic matter (Alongi, 2009). CO₂ emissions from soils raised from a variety of processes including autotrophic respiration, decomposition of organic matter via aerobic and anaerobic microorganisms (heterotrophic microbial respiration) (Poungparn et al., 2009; Capooci et al., 2019). Microbial biomass decreased due to anaerobic conditions and low nutrient environments, resulting in the low decomposition rate (Breithaupt et al., 2012). In the present study, soil TN contents which indirectly represented soil nutrient environments decreased with SLR and soil environments were under anaerobic conditions due to long periods of inundation, which would resulted in low CO₂ fluxes at Site SLR 80 cm. In addition, availability of OC is an important biogeochemical factor for soil CO₂ emissions. On the one hand, aboveground and belowground biomass, and leaf photosynthesis of K. obovata were suppressed at higher levels of flooding (Chen and Ye, 2013). Annual vegetation OC stock increments and vegetation OC stocks decreased with SLR (Chen et al., 2022a). These suggested that low vegetation biomass or OC stock may be contributing to the low soil CO₂ emissions at Site SLR 80 cm. On the other hand, soil CO₂ emissions were correlated with soil OC content and the lowest soil OC content (0–5 cm soil depth) was found at Site SLR 80 cm, resulting in low soil CO₂ emissions at Site SLR 80 cm.

In order to estimate the respective contributions of these three fluxes to CO₂-equivalent fluxes, all three greenhouse gas fluxes were converted to CO₂-equivalent fluxes on a 100 a time scale based on annual greenhouse gas fluxes. Mangrove soils at the three sites had a positive CO₂-equivalent fluxes, and CO₂-equivalent fluxes were comparable between Sites SLR 40 cm and SLR 0 cm due to CO₂ fluxes. However, the estimation of CO₂-equivalent fluxes from mangrove soils decreased under the influence of rising sea-level by 80 cm, which was attributed more to the decrease of N₂O and CO₂ fluxes at Site SLR 80 cm. On the other hand, CO₂ was the primary contributor of mangrove soils to CO₂-equivalent fluxes at Sites SLR 0 cm, SLR 40 cm and SLR 80 cm, as reported in the previous studies (Chen et al., 2010, 2012, 2015, 2016). However, SLR changed the contribution rates of CH₄ and N₂O fluxes to CO₂-equivalent fluxes. At Site SLR 80 cm, the contribution rate of CH₄ and N₂O to CO₂-equivalent fluxes were up to 4.08% and 12.01%. The warming potential of trace CH₄ and N₂O was non-negligible and should be considered in the evaluation of the warming effect on the soil gas emissions especially for mangrove forests and other coastal wetlands subjected to SLR.

Different intertidal elevations are used to represent different tidal flooding situations that might be expected with rising sea level. This study found the differences in the soil-atmosphere greenhouse gas emissions with SLR. However, some questions were not considered in this study. Firstly, three intertidal elevations were used to represent different tidal flooding situations of SLR and SLR impacts were not directly observed in this study. Some mangroves have adjusted to sea level rise and avoided submergence by accreting soil (Krauss et al., 2014). Adaptation of mangroves to SLR was not considered in the present study. Long-term observations of effect of SLR on the soil-atmosphere greenhouse gas emissions and even the whole mangrove ecosystem should be considered in the future studies. Secondly, sea levels were different at the different areas, and sea level may even reach 200 cm by the end of the 21st century based on Representative Concentration Pathway Scenarios (Mafi-Gholami et al., 2020). Data of SLR was only based on the average increase in sea level in China. Considered with complexity of SLR, without a doubt, more samples should be collected at the different areas with different SLR conditions to fully clarify effects of SLR on the soil-atmosphere greenhouse gas emissions and even the whole mangrove ecosystem in the future studies. Finally, although potential effects and possible mechanisms of SLR on the soil-atmosphere greenhouse gas emissions were found through different tidal flooding situations of SLR, specific mechanisms of SLR on the three greenhouse gas emissions were not considered in the present study and specific mechanisms should be considered in the next studies.

5. Conclusions

In the present study, three intertidal elevations were used to represent different tidal flooding situations of SLR to clarify potential effects of SLR on the soil-atmosphere greenhouse gas emissions in mangrove soils. On the one hand, higher soil water content, and lower salinity and soil nutrition environments caused by SLR may limit microbes to generate N₂O which resulted in significantly lower N₂O fluxes at Site SLR 80 cm. Lower salinity, higher soil water content and soil pH caused by SLR stimulated CH₄ production at Sites SLR 80 cm and SLR 40 cm. Significantly lower CO₂ fluxes were found at Site SLR 80 cm due to higher soil water content, lower salinity, lower OC availability and soil nutrient environments. On the other hand, the lowest CO₂-equivalent fluxes were determined at Site SLR 80 cm due to lower N₂O and CO₂ fluxes. Contribution rates of N₂O and CH₄ to CO₂-equivalent fluxes were up to the highest value at Site SLR 80 cm and warming potential of trace CH₄ and N₂O was non-negligible with SLR. Therefore, potential effects of SLR on the soil-atmosphere greenhouse gas emissions and even blue carbon capability of the whole mangrove ecosystem should be attended continuously in the future studies.