1 Introduction

The world is facing climate change, causing irregular weather patterns in the past decades. Global climate change is caused by an increase in atmospheric greenhouse gas (GHG) concentrations, including carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O) (Montzka et al., 2011; Kweku et al., 2018). The CO2 concentration in the atmosphere has increased to 409.9 ppm since the pre-industrial times (IPCC, 2021), while CH4 and N2O have increased by 5–10 and 1 ppb per year, respectively (Reay et al., 2018). Even though CH4 and N2O are emitted at lower concentrations, they have an average of a hundred-year global warming effect (100 GWP) of 29.8 and 273 times that of CO2, respectively (IPCC, 2021).

Mangroves are important sites for carbon accumulation, although their carbon-rich soils can also be sources of GHGs (Alongi, 2014; Chen et al., 2016). Mangrove vegetation fixes CO2 and stores carbon as aboveground biomass and roots (Bouillon et al., 2008; Murdiyarso et al., 2015), while some of the fixed carbon is exported as litter to adjacent coastal ecosystems (Adame & Lovelock, 2011). The rest is either decomposed by microorganisms or consumed by animals (Bardgett et al., 2008; Shiau & Chiu, 2020). The decomposition of organic carbon causes CO2 production in the soil oxic layers and CH4 production in the anoxic layers (Wang et al., 2009; Treat et al., 2015). Additionally, N2O is emitted through the nitrogenous pathway involving soil nitrification in the presence of O2 or denitrification in anoxic conditions, where soil carbon is high and oxygen is low (Zhu et al., 2013; Queiroz et al., 2019).

The GHG emissions in mangrove soils have been associated with environmental factors, including organic carbon, salinity, nutrients, and redox potential (Chen et al., 2010, 2014; Welti et al., 2017). GHG emissions increase with nutrient inputs and soil organic carbon (Chen et al., 2010, 2014) and decrease with salinity (Chen et al., 2010). For instance, CH4 is more likely to be produced under low-salinity conditions due to reduced competition with sulfate- and nitrate-reducing bacteria, which are more energy efficient than methanogenic bacteria (Purvaja & Ramesh, 2001; Biswas et al., 2007).

Mangroves exhibit zoning patterns across the intertidal area with characteristic redox and salinity gradients (Naidoo, 2016; Srikanth et al., 2016), which are likely to influence GHG emissions (Ulumuddin, 2019). Mangrove species with solid root structures, such as Sonneratia and Rhizophora, tend to dominate in the seaward or fringing forest (Tomlinson, 2016), where flooding is frequent, salinity is more less constant, and nutrient inputs are high (Ulumuddin, 2019). Variability across the intertidal zone also drives soil microorganism occurrence and metabolism, both of which are closely associated with GHG production (Chen et al., 2014). Other factors are also likely to affect mangrove characteristics and biogeochemical processes, such as climate and geomorphology (Lugo & Snedakers, 1974; Shih & Cheng, 2022), but at the local scale, mangrove zonation is likely to be one of the most important factors.

Measurements on GHG emissions in mangrove soils in many countries have been limited (Sasmito et al., 2019), including in Indonesia (Chen et al., 2014; Cameron et al., 2019; Sidik et al., 2019). Indonesia has one of the largest mangrove areas in the world, accounting for 19.5% of the total (Bunting et al., 2018). Indonesia has also one of the highest rates of mangrove loss (Richards & Friess, 2016), which are causing substantial GHG emissions (Maiti & Chowdhury, 2013). Thus, Indonesia has a high potential to reduce its emissions by reducing deforestation and supporting mangrove restoration (Buelow et al., 2022). However, to accurately estimate the carbon gains from mangrove protection and restoration, the baseline GHG emissions need to be excluded from their carbon mitigation potential (Rosentreter et al., 2021).

Indonesia is a tropical archipelagic country which a large variety of mangrove ecosystems, which are globally significant for its blue carbon and multiple co-benefits (Murdiyarso et al., 2015). Indonesia’s estuaries have significant freshwater inputs forming characteristic salinity gradients, which drive different types of mangrove species composition and structure, thus likely highly variable GHG emissions. One mangrove ecosystem experiencing high anthropogenic pressure is Benoa Bay in Bali. The Bay suffers from nutrient pollution (Raharja et al., 2018; Rahayu et al., 2018) and massive sedimentation caused by land reclamation, which has already caused the deterioration of the mangroves. Thus, this site is urgently needed for management and restoration activities, potentially through Blue Carbon projects.

The objectives of this study were as follows: first, to estimate the GHG fluxes (CO2, CH4, and N2O) in the mangroves of Benoa Bay along the intertidal zone (seaward, middle, and landward) and, second, to assess the relationship between GHG fluxes, forest structure, and environmental parameters (salinity, oxidation reduction potential, soil organic carbon). The wet or rainy season was chosen as the sampling time due to the maximum potential for soil GHG fluxes in tropical mangrove forests to occur (Kristensen et al., 2008; Tang et al., 2018; Otero et al., 2020; Kitpakornsanti et al., 2022). The main goal was to provide baseline data from these human-impacted mangroves in Indonesia. The result from this study will support the Indonesian government program FoLU (Forestry and Other Land Use) Net Sink 2030 by providing GHG fluxes data from the forestry sector. This information can be useful to accurately assess the carbon sequestration capacity and potential for reduced emissions of mangroves they are appropriately managed. We hypothesized that GHG fluxes during wet season will differ among mangrove community zones due to differences in vegetation structure and environmental conditions along the intertidal zone.

2 Methods

2.1 Site description

Benoa Bay, Bali (8°43′ to 8°42′S, 115°11′ to115°14′W; Fig. 1), has a large area of mangrove forest with 1132 ha. The mangroves have distinct zonation patterns, starting from the sea zone (seaward), which is dominated by Sonneratia alba, continuing to the middle zone, which is dominated by Rhizophora spp. and the landward zone, which has mixed mangrove species (Andiani et al., 2021; Sugiana et al., 2022). Seaward and middle zones are frequently flooded during high tides, while the landward zone is only inundated during large spring tides (Fig. 1). The soil texture ranges from fine sand to gravel but is dominated by coarse sand (Imamsyah et al., 2020; Prinasti et al., 2020). The porewater salinity of these mangroves generally increases, while the redox decreases, from the land to the seaward edge (Sugiana et al., 2021).

Fig. 1
figure 1

Study sites in Benoa Bay, Bali, Indonesia. 1 Landward plots, 2 middle plots, and 3 seaward plots

Full size image

We sampled nine plots within three forests close to three villages: Sanur, Serangan, and Pemogan, during the wet season at low tide of the new moon period (3–9 December 2021) at daylight (1 pm–3 pm; UTC + 8). During sampling, temperatures ranged from 23.2 to 34.6 °C (average: 30.1 °C), humidity levels ranged from 58.0 to 98.0% (average: 76.4%), and monthly rainfall was 549.0 mm (Denpasar Central Statistics Agency, 2022).

2.2 Greenhouse gas flux measurement

We used the closed chamber method adopted from Chen et al. (2016). At each plot, three 20 × 20 × 25 cm squared dark chambers were installed at 2 cm deep in the soil. Gas samples were collected using 10-mL syringes plot for 30 min at 10-min intervals (0, 10, 20, 30 min for total 108 samples). We only took the gas once on each plot due to access difficulties and the limitation of low tide time, which only lasted for 2 h, especially on the seaward edge. The GHG concentrations were analyzed with gas chromatography (450-GC Varian) equipped with a flame ionization detector (FID), a thermal conductivity detector (TCD), and a 63Ni electron capture detector (μECD) for CH4, CO2, and N2O consecutively. The 450-GC also has a PAL autosampler injector, which functions for auto-injection and uses Ar, H2, He, N2, and compressed air as carrier gas. The measurement process began at under 25 °C at room temperature. A standard curve was used as a reference in the analysis. The concentration of GHG was calculated by comparing the peak area of the sample with the standard curve. The relationships between deployment time and GHG concentrations were significant (R2 of 0.83–0.98, 0.76–0.94, and 0.79–0.94 for CO2, CH4, and N2O, respectively, Fig. 2), thus confirming that the mangrove sediment continuously resealed gases that were accumulated in the chambers during sampling. The GHG fluxes were calculated as in Chen et al. (2016):

$$Fm=\frac{V \,x\, \Delta M\, x\, {10}^{6}}{A\, x\, P}$$

where

Fig. 2
figure 2

The trend between GHG concentrations and deployment times for each mangrove zone

Full size image

Fm: GHG fluxes (µmol m−2 h−1)

ΔM: The slope of the linear regression line between GHG concentrations (ppm) and sampling frequency (10 min transformed to an hour)

V: Chamber volume (L)

A: Chamber area (m2)

P: Constant gas volume (m3 mol−1)

Based on the average gas fluxes of the three mangrove zones, the CH4 and N2O fluxes were converted to CO2-equivalent fluxes to indicate the gas warming effect (IPCC, 2021). The CO2-equivalent flux was calculated using the following formula:

$$Fe=Fm\, x\, M\, x\, GMP$$

where

Fe: CO2-equivalent flux (g CO2 m−2 h−1),

Fm: Interfacial gas flux (mol m−2 h−1)

M: Molecular weight of the GHG

GMP: Warming effect or the conversion of CH4 and N2O emissions to CO2 equivalents as 29.8 and 273, respectively, over a 100-year timeframe (IPCC, 2021).

2.3 Soil and porewater physicochemical characteristics

Temperature, pH, salinity, and oxidation reduction potential (ORP) in the water and sediment were measured during sampling with a pH meter EZODOPH5011, SCT meter YSI EC200, and multimeter COM-600 water quality tester. Soil samples were collected from each plot at a depth of 5 cm using a 10-cm-diameter core pipe, from which water content and soil organic carbon (SOC) were measured. Samples were dried at 70 °C until constant weight and reweighed to measure water content. For SOC, the soil was filtered through a 2-mm mesh and burned at 550° (Chen et al., 2014). Soil type was identified with the megascopic method and the Wentworth diagram (1922).

2.4 Forest structure

Forest structure was measured following the guidelines from COREMAP-CTI LIPI (Dharmawan et al., 2020). At each site, we measured tree and sapling density, canopy coverage, and diameter at breast height (DBH) within a 10 × 10 m plot. Each plant was classified either as a tree (DBH ≥ 5 cm) or a sapling (DBH < 5 cm), and its species was identified based on Giesen et al. (2007) and Tomlinson (2016). Density, dominance, and species were used to estimate the importance value index (IVI). Field data collection was run in the MonMang 2.0 application, similar to Sugiana et al. (2022).

Hemispherical photography was used to estimate mangrove canopy coverage in each zone (Jennings et al., 1999; Ishida, 2004). Squared output hemisphere photographs were captured from scattered positions on each plot using a smartphone with 16-MP resolution. Each photograph was analyzed using ImageJ software to count the number of pixels. Canopy coverage percentage was calculated by comparing the number of pixels with vegetation to the total number of pixels on each photograph. The mangrove health index (MHI) was calculated by combining data on canopy cover, density, and DBH (Dharmawan & Ulumuddin, 2021), and below-ground carbon (BGC) is also measured by converting DBH of each mangrove stands using allometric equation based on Kauffman and Donato (2012).

2.5 Statistical analyses

All univariate data were analyzed with the Shapiro-Wilk normality test. Environmental parameters were normally distributed (p > 0.05) and were compared against zones with an analysis of variance (ANOVA) and Tukey HSD test for each mangrove zone. The GHG flux data were not normally distributed (p < 0.05), hence were analyzed with the nonparametric Kruskal-Wallis and Spearman rank. The GHG emissions were compared among zones and against environmental parameters (forest structure, soil, and porewater properties). All the tests used a statistical level of p < 0.05 and were run with the R software version 4.0.2.

3 Results

3.1 Forest structure

Species composition was different among mangrove zones, with the landward zone being the most diverse, with six mangrove spp. dominated by Sonneratia alba (IVI = 107.1%). The middle zone had two mangrove species dominated by Rhizophora apiculata (IVI = 175.1%, Table 1), and the seaward zone was solely composed of S. alba (IVI = 300.0%). A significantly higher density of mangrove trees and saplings was found in the middle zone with 3900 ± 1654 tree ha−1 (ANOVA: F3.6 = 4.67, p < 0.05) and 1400 ± 1015 tree ha−1 (ANOVA: F3.6 = 5.52, p < 0.05), respectively, compared to the land and seaward zone. The seaward zone had higher DBH (13.6 ± 1.7 cm) and lower canopy cover (40.8 ± 14.3%) compared to the land and middle zones 7.9 ± 3.1 cm and 7.8 ± 0.9 cm (ANOVA: F3.6 = 22.38, p < 0.05) and 72.9 ± 5.7% and 78.7 ± 1.1%, respectively (ANOVA: F3.6 = 47.30, p < 0.05), respectively. Even though the seaward zone had the largest DBH, it had the lowest MHI values with 41.0 ± 3.0% compared to land and middle zones with 51.2 ± 5.7% and 58.0 ± 3.6% (ANOVA: F3.6 = 36.57, p < 0.05), respectively. The low MHI value in the seaward zone was caused by its low canopy cover and sapling density. Despite the significant difference among the MHI across zones (p < 0.05, Table 1), they were all in the same category, suggesting a “moderate” condition. Below-ground carbon (BGC) from mangrove roots was also significantly different between the middle zone and the other two zones (F3.6 = 40.19, p < 0.05), with the higest value found in middle zone at 229.5 ± 84.4 Mg ha−1.

Table 1 Mangrove community parameters across the landward, middle, and seaward zones
Full size table

3.2 Soil and porewater physicochemical characteristics

The mangrove soil was mainly mud, except for the seaward zone, which was composed of sand and mud (Table 2). Water content during sampling was significantly highest in the middle zone with 50.9 ± 7.5% (ANOVA: F3.6 = 29.83, p < 0.05), and SOC concentrations were similar among zones (F3.6 = 0.80, p > 0.05), with values ranging from 19.8 to 25.8 mg g−1. The temperature of the porewater was similar among zones, with a mean value of 30.4 °C (ANOVA: F3.6 = 0.38, p > 0.05). Water pH, salinity, and ORP were significantly different in the seaward compared to the rest of the zones with mean values of 7.0 ± 0.1 (ANOVA: F3.6 = 23.29, p < 0.05), 28.3 ± 1.3 ppt (ANOVA: F3.6 = 8.40, p < 0.05), and 21.1 ± 82.8 mV (ANOVA: F3.6 = 6.85, p < 0.05) as can be seen in Table 2.

Table 2 Mangrove soil and porewater physicochemical properties across the landward, middle, and seaward zones
Full size table

3.3 Greenhouse gas fluxes

The CO2 emissions were similar among mangrove zones (K-Wallis: H3,6 = 0.03, p > 0.05), although with a trend of decreasing emissions from the land (2,644.7 ± 3,338.8; range of 314.8 to 8,952.8 μmol m−2 h−1) to the middle (1,646.6 ± 1,335.6; 266.1 to 4,175.4 μmol m−2 h−1) and the seaward zone (1,563.5 ± 1,302.4; 97.8 to 3,578.3 μmol m−2 h−1; Fig. 3A). Similarly, CH4 and N2O flux patterns tended to decrease from the land to the sea, although the differences were not significant (CH4 K-Wallis: H3.6 = 5.89, p > 0.05 and N2O K-Wallis H3.6 = 5.04, p > 0.05). The CH4 flux in the landward zone was 34.7 ± 34.1 (5.0 to 88.0) μmol m−2 h−1, the middle zone was 21.36 ± 15.72 (5.0 to 44.4) μmol m−2 h−1, while the seaward zone was 10.0 ± 11.7 (0.9 to 38.5) μmol m−2 h−1 (Fig. 3B). For N2O, the landward flux was 1.4 ± 0.8 (0.43 to 2.96) μmol m−2 h−1, followed by the middle zone with 1.0 ± 0.8 (0.35 to 3.60) μmol m−2 h−1, and, finally, the seaward zone with 0.6 ± 0.3 (0.2 to 1.5) μmol m−2 h−1 (Fig. 3C). The CO2-equivalent fluxes from CH4 and N2O showed a similar decreasing trend from the land to the sea with a total mean GWP of 19.2 mg CO2 m−2 h−1 (Table 3).

Fig. 3
figure 3

A CO2, B CH4, and C N2O fluxes (μmol m−2 h−1) from mangrove soils in the landward (L), middle (M), and seaward (S) zone in Benoa Bay, Indonesia

Full size image
Table 3 CO2-equivalent greenhouse gas fluxes from mangrove soils within the intertidal zones in Benoa Bay, Indonesia
Full size table

3.4 Correlation with environmental parameters

Soil parameters were positively correlated with CO2 and CH4, but not with N2O fluxes. Higher soil organic carbon was associated with larger CO2 (p = 1.9 × 10−2) and CH4 fluxes (p = 5 × 10−4), while lower ORP and low porewater salinity were associated with larger CH4 fluxes (p = 2.3 × 10−2 and 2.9 × 10−3, respectively). Forest structure parameters were not significantly correlated with GHG fluxes (Table 4).

Table 4 Spearman-rank correlation coefficient values (r) among soil, porewater and mangrove structure, and greenhouse gases
Full size table

4 Discussion

The GHG fluxes in mangroves from Benoa Bay, Indonesia, had a decreasing trend from the land to the seaward zone. The CO2 and CH4 fluxes were significantly higher where SOC was highest, and CH4 flux were significantly highest where salinity and ORP were lowest. These results complement current literature showing that soil physicochemical characteristics are important drivers of GHG emissions in mangroves (Chen et al., 2010). Our results also prove that in addition to sequestering carbon, pollution-impacted mangroves release GHG from sediments, which need to be accounted for when estimating their full carbon mitigation potential.

Other studies have shown significant changes in GHG fluxes across intertidal zones, some finding higher emissions in the seaward (Wang et al., 2009) and others in the landward zone (Hirota et al., 2007; Chen et al., 2010; Lin et al., 2020). Each mangrove zone has unique vegetation characteristics, resulting mainly from tidal inundation variations, which drive interstitial salinity and nutrient availability. Similarly, in Benoa Bay, mangrove structure is quite different among zones, including the density of trees and saplings, the size of tree trunks, and the canopy cover. Additionally, pH was also significantly different among zones. Despite showing different values among zones, none of these parameters was significantly associated with GHG fluxes, suggesting that the effect of forest structure and pH was a less strong predictor than SOC and salinity. However, SOC was also influenced by vegetation density, where the denser the vegetation, the higher the carbon stores in the soil. However, this statement should be further studied, considering its weak relationship, and only applied if the measured vegetation density comes from the same species (Dermawan et al., 2023). Each vegetation species has a different contribution to the burial rate of SOC (Weiss et al., 2016), so this needs to be considered since SOC was closely correlated to GHG fluxes.

In Benoa Bay, the heterogeneous structure of the mangrove forest has caused differences in stand structure, such as the seaward zone, which is dominated by Sonneratia alba, with lower stand density due to its allelopathic compounds (Li et al., 2015; Zhang et al., 2018), compared to the middle zone, which is dominated by Rhizophora apiculata. Apart from SOC burial contribution, variations in mangrove species also indicate differences in salinity. Sonneratia alba could tolerate high water salinity, which makes this species dominate the seaward zone with more frequent seawater input (Pillai & Harilal, 2016). In addition, due to the more frequent washing of seawater through tides, SOC concentrations in the seaward zone tend to be lower than in the other zones. Therefore, differences in the zoning pattern of mangrove forests also indirectly indicate variations in GHG fluxes.

The higher fluxes at higher SOC make sense, as a higher carbon source for soil microorganisms results in higher rates of anoxic and anaerobic respiration (Bouillon et al., 2008; Morell et al., 2011). Similar results were found in South China (Chen et al., 2010) and Jiulong River estuary, China (Chen et al., 2016), where a positive correlation between SOC with CO2 and CH4 fluxes was found. Organic carbon sources in sediments come from within or outside the mangrove ecosystem, such as anthropogenic carbon carried by rivers or tides (Ulumuddin, 2019). Decomposition and respiration of organic carbon in mangrove soils result in CO2 or CH4 fluxes; however, the proportion of these gases depends on environmental conditions within the mangrove soil.

The CH4 fluxes were significantly correlated with low porewater salinity and ORP. Most mangrove soils have soils low in oxygen (Lu et al., 1999; Arai et al., 2016). However, in highly impacted mangroves, where carbon pollution is significant, anoxic conditions favor the production of CH4 through methanogenesis (Maltby et al., 2016; Sánchez-Carrillo et al., 2021). CH4 emissions are low in sites with high inputs of marine water, which favors sulfate-reducing bacteria that outcompete methanogens (Ulumuddin, 2019). Although N2O emissions followed a similar trend as CO2 and CH4, these differences were insignificant. Nevertheless, the decreasing N2O land-seaward trend is expected, as denitrification a likely source of N2O) is higher where soil organic carbon is high and pH is neutral, which sometimes can occur in the seaward zone (Fernandes et al., 2010; Marton et al., 2012).

Compared to other studies, our results are within the range of other tropical and subtropical regions (Table 5). We acknowledge that our sampling was limited in time; however, the wet season in Benoa Bay is where precipitation and temperature are highest; thus, our data could be representative of the higher end of annual emissions. In addition, these results were also supported by some literature which finds that the wet season becomes the peak of GHG flux in tropical mangrove forests such as CO2, CH4, and N2O in Thailand (Kitpakornsanti et al., 2022); CO2 in Tanzania and Brazil (Kristensen et al., 2008; Castellón et al., 2021); CH4 in Malaysia, Brazil, and Myanmar (Tang et al., 2018; Dalmagro et al., 2019; Cameron et al., 2021); and N2O in Brazil (Otero et al., 2020). In subtropical regions, similar results were also found for CH4 (Philipp et al., 2017; Zhu et al., 2021). Therefore, if this holds true, CH4 and N2O GWP of mangroves in Benoa Bay is about 1.7 MgCO2 ha−1 year−1, which is lower than CO2 GWP on natural riverine mangroves in Perancak, Indonesia, with 12.2 MgCO2 ha−1 year−1 (Sidik et al., 2019) and of rehabilitated mangroves in Sulawesi, Indonesia, with overall GWP of 17 MgCO2 ha−1 year−1 (Cameron et al., 2019). However, the emission of mangrove soils is low compared to their potential for carbon sequestration, which is estimated at 52.9 MgCO2 ha−1 year−1 (LIPI, 2018). Thus, regardless of their soil emissions, even in pollution-impacted regions, mangroves still maintain a capacity to mitigate carbon emissions. This result is particularly important given the high national deforestation rate of mangrove forests in Indonesia, which reached 182,091 ha, contributing to 136.9 MgCO2 ha−1 year−1 (Arifanti et al., 2021). Thus, restoration of mangroves such as those in Benoa Bay could contribute to lowering emissions in Indonesia.

Table 5 Comparison of GHG fluxes in mangroves soil of Benoa Bay with other tropical and subtropical regions
Full size table

5 Conclusion

GHG emissions in mangroves in Bali slightly varied across mangrove zones, with a decreasing land-to-seaward trend. The main parameters that controlled the emissions were salinity and SOC. Despite the GHG emissions at their maximum value during wet season, mangroves in Benoa Bay are still likely to remove and store significant quantities of carbon. Our data contribute to accurately estimating the carbon mitigation of mangroves in Indonesia. The low GHG emissions in these mangroves support the idea that even disturbed mangroves have potential to act as a carbon sinks. Further monitoring throughout the year could improve our results and identify additional drivers of GHG emissions in Indonesian mangroves and better ways to manage and protect these valuable ecosystems.