Pneumatophores and crab burrows increase CO 2 and CH 4 emission from sediments in two Brazilian fringe mangrove forests

: We assessed the release of greenhouse gases (CO 2 and CH 4 ) from air-exposed sediments and dissolved inorganic carbon (DIC) from inundated sediments in 2 Brazilian mangrove forests. Our focus was on the impact of biogenic structures, i.e. pneumatophores and crab bur-rows, on greenhouse gas emissions. Emission of CO 2 from air-exposed bare sediment (111−156 and 57−148 mmol m −2 d −1 in darkness and in light, respectively) was comparable to DIC release from inundated sediment (122−158 and 52−62 mmol m −2 d −1 , respectively). Pneumatophores and crab burrows increased dark CO 2 emission during air exposure by 113−123 and 49−91%, respectively. CH 4 emission from air-exposed bare sediment (0.22−0.25 mmol m −2 d −1 ) was increased 92−137 and 288−607%, respectively, by pneumatophores and burrows. Carbon loss in the form of CO 2 and DIC from sediments with biogenic structures can at the two study locations be extrapolated to 64.1 and 71.0 mol C m −2 yr −1 . These values fit well with literature values of litterfall in the studied area, providing carbon accretion of 28.5 and 21.6 mol C m −2 yr −1 . However, the budget will be unbalanced if the role of biogenic structures is not considered. In the presence of biogenic structures, CH 4 emissions of 2.8 and 3.3 mol C m −2 yr −1 (when converted to CO 2 units) will partly (10−15%) counteract the climate mitigation effect of the accumulated carbon. Carbon budgets in mangrove sediments may therefore be flawed if the contribution of biogenic structures to green-house gas emissions is ignored.


INTRODUCTION
Mangrove forests are considered important blue carbon ecosystems with a high capacity to sequester and store carbon (C) in anoxic sediments and thus mitigate global climate change (Alongi 2012, Jennerjahn 2020. Together with saltmarshes and seagrass beds, mangrove forests are responsible for a disproportionally large fraction of the marine blue carbon storage, despite the relatively small areal coverage of these ecosystems (Kristensen et al. 2017, Macreadie et al. 2021, Wang et al. 2021. Mangrove forests are largely restricted to tropical and subtropical areas, where they cover a total of only 130 000−150 000 km 2 (Giri et al. 2011, Spalding et al. 2011, Friess et al. 2019. The mangrove area in Brazil is estimated to be about 10% of the global coverage and extends along more than 90% of the coastline (Spalding et al. 2011, Schaeffer-Novelli et al. 2016. C inputs to mangrove sediments are dominated by primary production of mangrove trees and import of organic and inorganic particles from upstream rivers and the coastal ocean (Matos et al. 2020, Sasmito et al. 2020. Conversely, C losses occur through respira-*Corresponding author: ebk@biology.sdu.dk Pneumatophores and crab burrows increase CO 2 and CH 4 emission from sediments in two Brazilian fringe mangrove forests tion by trees, benthic fauna and microbial decomposers, as well as carbonate dissolution combined with emission of gases to the atmosphere and tidal export of solutes and particles (Alongi 2020a). A global average balance between these sources and sinks of C reveals that mangrove forests have the potential to accumulate and sequester around 170 g C m −2 yr −1 (Pérez et al. 2018, Alongi 2020b. However, there are still gaps in mangrove C budgets, since measured sinks such as burial, CO 2 emission and tidal export in many cases cannot fully balance the estimated C sources in the form of primary production and tidal import ). It has been suggested that the C budget deficiency of mangrove forests is caused by underestimated sinks in the form of excess emission of CO 2 via biogenic structures, such as pneumatophores, cable roots and crab burrows , Alongi 2020a. During low tide, these structures provide air-filled conduits from the sediment to the atmosphere driven by a diffusive transport that is 10 000 times faster than in sediment porewater (Kristensen et al. 2022). Although the contribution of pneumatophores and crab burrows to CO 2 emission into the atmosphere may increase the C loss, there is still room for substantial blue carbon formation in mangrove sediments (Alongi 2020b). In fact, these structures may instead close the gaps and complete mangrove C budgets.
There are indications, however, that the contribution of other greenhouse gases (CH 4 and N 2 O) which are more potent than CO 2 may threaten the warming mitigation capacity of the accumulated blue carbon in mangrove forests (Rosentreter et al. 2021). For example, CH 4 has a global warming potential that is up to 34 times higher than CO 2 on a 100 yr scale when climate C feedbacks are included (Gillett & Matthews 2010, Myhre et al. 2013. In a recent review, Kristensen et al. (2022) found that the global warming mitigation of blue carbon in mangrove forests is offset to almost zero when including the effect of pneumatophores, cable roots and crab burrows as channels for rapid emission of CH 4 and N 2 O.
The aim of the present study was to determine emissions of 2 greenhouse gases (CO 2 and CH 4 ) and the release of dissolved inorganic carbon (DIC) from sediments in 2 Brazilian fringe mangrove forests adjacent to and distant from the urban areas of Cananéia City. Focus was particularly on the influence of biogenic structures, i.e. pneumatophores and crab burrows, for the emission of CO 2 and CH 4 to the atmosphere. Gas emission during air exposure was determined using in situ techniques and during inundation using laboratory-based DIC exchange measurements. Based on the inundation frequency and duration during day and night over 1 mo, we quantified the total release of C-based greenhouse gases and compared it to known local estimates of C input. We used these results to assess the impact of biogenic structures on C budgets and the climate change mitigation potential of the explored Brazilian fringe mangrove forests.

Study sites
The Cananéia Lagoon is a mangrove-surrounded estuary located near the southern border of São Paulo State, Brazil (Fig. 1). The estuary has 110 km 2 of water area (mean depth of 6.5 m) and is bordered by 72 km 2 of mangrove forest. It is connected to the South Atlantic Ocean by Cananéia Inlet in the south and Icapara Inlet in the north. Water exchange in the estuary is primarily the result of inflow of offshore water by tidal currents and outflow of fresh water from several rivers (Schaeffer-Novelli et al. 1990). The climate is mild subtropical, with average annual temperature ranging from 20°C in the dry winter (July to August) to 28°C in the wet summer (January to March); average annual precipitation is 2200 mm (Cunha-Lignon et al. 2009). The tidal regime is semidiurnal and microtidal, with average tidal amplitudes of 0.7 ± 0.3 m (mean ± SD).
The study was conducted during February 2014 in fringe mangrove forests adjacent to 2 creeks near the Cananéia Inlet. Olaria Creek is located close to Cananéia City on Cananéia Island, while Nobrega Creek is located on the unpopulated Comprida Island ca. 2 km away from Cananéia Island (Fig. 1). The mangrove forest surrounding Olaria Creek is periodically exposed to anthropogenic discharges from suburban parts of Cananéia City, while no such exposure occurs at Nobrega Creek (Aidar et al. 1997, Barcellos et al. 2005. The mangrove forest along both creeks is dominated by 3 mangrove species (Laguncularia racemosa, Rhizophora mangle and Avicennia schauerianna; Schaeffer-Novelli et al. 1990). The most abundant burrowing brachyuran crabs are Ucides cordatus, fiddler crabs (Uca spp.) and sesarmid crabs (e.g. Aratus pisonii).
Parallel samplings were conducted at comparable mangrove locations in Olaria (25.0204° S, 47.9294° W) and Nobrega (25.0137° S, 47.9133° W) Creeks. One intertidal location dominated by the mangrove A. schauerianna was chosen 20 m inland of each creek with similar tidal elevation of the mangrove forest floor. The 2 locations were ~50 cm above mean sea level and were flooded 20−25 times monthly during high spring tides (Fig. 2). Salinity and water temperatures at both locations were 22.8−23.4 and 28−30°C at the time of sampling.

Density and size of pneumatophores and crab burrows
The density and size of visible biogenic structures (A. schauerianna pneumatophores and crab burrows) were determined from digital photographs taken of the mangrove forest floor simultaneously with the in situ incubations and sampling. Photographs at Olaria (n = 14) and Nobrega (n = 9), which included a known scale and covered on average 0.5 m 2 (range: 0.3− 1 m 2 ) of the forest floor, were taken perpendicular to the sediment surface. All visible pneumato phores and burrow openings were counted by image analysis. The diameter of all detected crab burrows was measured digitally from the photographed sediment surfaces. The burrow inhabitants were not identified. Pneumatophore height was obtained from horizontal photos flush with the sediment surface at Olaria (n = 3) and Nobrega (n = 3). The height of all visible pneumatophores was measured by image analysis.

Sediment characteristics
Sediment cores were collected from all sites by inserting core liners (30 cm long) into undisturbed sediment. The cores were dug out by hand and closed at both ends with rubber stoppers before transport to the laboratory at the University of São Paulo field station in Cananéia. Six sediment cores (20 cm deep, 5 cm inner diameter [i.d.]) were collected from each location for determination of solid phase granulometry and water content, as well as particulate organic carbon (POC), total nitrogen (TN) and surface sediment chlorophyll a (chl a) content.
Three sediment cores from each location were sectioned into 0−2, 4−6, 8−10 and 12−14 cm intervals. Sediment water content was determined from the weight loss after drying (60°C) subsamples to a constant weight. POC and TN were determined on dried samples with an elemental analyser (Flash EA 1112 Series, Thermo Finnigan) after acidification with HCl fumes to remove carbonates. Grain size distribution of other subsamples was analysed on a Malvern Mastersizer 3000 to obtain the median grain size and silt+clay content.
The upper 0.5 cm portions of 3 other sediment cores from each location were retained and immediately homogenized for chl a determination using the standard spectrophotometric method (Parsons et al. 1984). Subsamples of sediment were extracted overnight in 5 ml of 90% ethanol in darkness at 5°C. The extract was centrifuged at 1200 × g, and absorbance of the supernatant was measured at 665 and 750 nm before and after acidification with 2 drops of 1 M HCl.

In situ CO 2 and CH 4 exchange across the sediment−air interface
The exchange of CO 2 and CH 4 between airexposed sediment and the atmosphere at both study locations was determined during low tide on the mangrove forest floor with and without pneumatophores or crab burrows.
The exchange of CO 2 was measured with a portable LI-COR 6400 Photosynthesis System equipped with a custom-made dynamic CO 2 flux chamber (18 cm long, 8 cm i.d.). Measurements were first conducted in light and darkness on bare sediment without pneumatophores and crab burrows. Three 10 cm long transparent acrylic core tubes with 8 cm i.d. were pushed about 5−7 cm into the sediment at each location. The tubes were left open for at least 30 min to equilibrate and thereafter sealed gas-tight using a rubber septum with the transparent LI-COR 6400 measuring chamber for light incubations. The LI-6400 was set for closed-chamber measurement, while the circulation pump in the sensor head assured a recirculating gas phase in the chamber. Exchange of CO 2 gas with the chamber in each position was measured for 3 × 5 min with CO 2 measurements every 10 s. Chamber CO 2 typically changed linearly during measurements. The chamber was lifted from the core tubes for 1 min between each 5 min measurement series to allow re-equilibration with the atmosphere.
Light intensity at the mangrove forest floor was determined continuously by a built-in quantum sensor in the LI-6400. Each core tube was then darkened with a bucket placed upside-down for 15−20 min to reset the enclosed sediment surface to dark conditions, and the measuring chamber was enclosed by an opaque plastic cap.
Exchange of CO 2 gas in darkness was then measured for 3 × 5 min on each core tube as described above. Subsequently, the dark measurements were repeated for 3 × 5 min with each of the 3 core tubes enclosing sediment with 1 average sized A. schauerianna pneumatophore. The procedure was repeated with each of the 3 core tubes placed on top of 1 average sized crab burrow. Rates of CO 2 exchange across the enclosed 50 cm 2 surface area were calculated based on the linear concentration change over time and volume of gas captured within the chambers.
Exchange of CH 4 was measured at both Olaria and Nobrega with static flux chambers consisting of 15 cm long acrylic core tubes (8 cm i.d.). Core tubes were pushed 5 cm into the sediment and sealed gastight at the top with rubber stoppers. Nine core tubes were deployed at each location with 3 in bare sediment, 3 enclosing 1 pneumatophore and 3 on top of 1 average-sized crab burrow. Sampling of the headspace gas was accomplished by 5 ml syringes connected to 2 hypodermic needles traversing the rubber stopper. Gas samples were retrieved from the headspace of each chamber. While samples of about 5 ml gas were taken after 0, 45−60 and 100−120 min from the headspace with 1 syringe, constant pressure was maintained by simultaneously injecting 5 ml atmospheric air each time with the other syringe. A 2 ml sample was transferred to a 3 ml evacuated glass Exetainer ® . The Exetainer vials were stored cool until analysis on a gas chromatograph within 1 mo. Rates of CH 4 exchange across the enclosed 50 cm 2 surface area were calculated from the linear concentration change over time during incubation and corrected for chamber volume and dilution during sampling.

DIC exchange during submersion
Three sediment cores (8 cm i.d.) from the mangrove forest floor at both study locations were used for determination of DIC exchange in light and darkness during inundation. Transparent core liners (30 cm long) were pushed into sediment without biogenic structures and carefully dug out by hand. The sediment cores were closed at both ends with rubber stoppers and transported to the laboratory located at the University of São Paulo field station in Cananéia.
After return to the laboratory, the headspace of sediment cores was gently filled with aerated creek water collected at Olaria and Nobrega Creeks dur ing core sampling. The cores were equipped with a rotating magnet that was driven by a central stirring magnet and placed in a flow-through water bath kept at a constant temperature (29°C). The water bath was darkened with a black plastic sheet. Cores were left open in darkness to acclimatize for ~2 h before headspace water was sampled for DIC in 20 ml gas-tight glass vials. The cores were closed gas-tight with rubber stoppers and were left to incubate for 2−3 h before the rubber stoppers were removed and water from each core was sampled as described above. After the dark incubations, the headspace of each core was re-filled with aerated creek water and cores were left open to acclimatize for 1 h in ambient light before repeating the DIC exchange incubation in light with a transparent lid instead of the rubber seal. DIC was analysed immediately by potentiometric Gran titration (Talling 1973), and the ex change was calculated from the concentration change during incubations after correction for water volume and sediment surface area. Gross primary production (GPP) was calculated by subtracting dark rates from light rates.

Sediment environment and biogenic structures
Sediment granulometry and organic content showed no specific trends with depth in the upper 14 cm of the sediment and are pre-sented in Table 1 as averages for the entire sediment column. The 2 fringe mangrove locations, Olaria and Nobrega, also did not differ significantly among most sediment characteristics. The sediment at both locations consisted roughly of 50% silt+clay with a median grain size of 57−65 μm (Table 1). POC and TN content was higher by 38% (t-test, p < 0.01) and 20% (t-test, p < 0.05), respectively, at Nobrega compared to Olaria. Chl a content in the upper 0.5 cm of the sediment was almost 2 times higher at Olaria than at Nobrega (t-test, p < 0.001) ( Table 1).
Pneumatophores and crab burrows were abundant on the mangrove forest floor at both locations with densities of 180−254 and 89−125 m −2 , respectively ( . Chlorophyll a is given as the mean (SD) of the 0−0.5 cm layer of 3 cores (n = 3). POC: particulate organic carbon; TN: total nitrogen; ww: wet weight; dw: dry weight ence between locations, high standard deviations point to a patchy distribution of these biogenic structures. At both locations, average pneumatophore height was slightly above 8 cm with a diameter of 0.5−1 cm, while the average diameter of burrow openings was 1−1.4 cm ( Table 2).

Exchange of C gases during air exposure
In darkness, emission of CO 2 from air-exposed fringe mangrove sediments without biogenic structures (bare) ranged from 111 mmol m −2 d −1 at Olaria to a non-significantly higher level of 156 mmol m −2 d −1 at Nobrega (Fig. 3). The corresponding emissions during light exposure at the 2 locations were lower, 57 and 148 mmol m −2 d −1 , respectively, providing a wet season benthic GPP of 54 and 8 mmol m −2 d −1 , respectively. The light intensity at the mangrove forest floor during light incubations was (mean ± SD) 580 ± 133 mE m −2 s −1 at Olaria and 481 ± 445 μE m −2 s −1 at Nobrega. Since the im pact of biogenic structures on CO 2 emission was measured with the 8 cm i.d. incubation chambers enclosing either 1 pneuma tophore or 1 crab burrow, the reported emissions represent a sediment surface with 200 pneumatophores or burrows m −2 . The presence of pneumato phores increased the measured dark CO 2 emission significantly (t-test, p < 0.05) by 123% at Olaria and 113% at Nobrega (Fig. 4) compared with bare sediment. The corresponding increase of 91 and 49%, respectively, in the presence of burrows was significant for Olaria (t-test, p < 0.05) and marginally significant for Nobrega (t-test, p = 0.066). Accordingly, each pneumatophore was estimated to emit 682 μmol CO 2 d −1 at Olaria and 887 μmol CO 2 d −1 at Nobrega (Table 3) Table 3. Contribution of individual pneumatophores and burrows to CO 2 and CH 4 emission during air exposure in darkness. The estimated total emission of mangrove forests in Olaria and Nobrega is given for the observed density of pneumatophores and burrows ( Nobrega. By considering these individual rates and the observed average density of pneumatophores and burrow openings (Table 2), the total emission from air-exposed and darkened sediment at the 2 mangrove locations during the warm wet season was similar, around 350 mmol m −2 d −1 (Table 3). Emission of CH 4 from air-exposed bare fringe mangrove sediments was 0.22−0.25 mmol m −2 d −1 , with no significant difference between locations. The measured CH 4 emission in the presence of 200 pneumatophores m −2 increased significantly (t-test, p < 0.05) by 137% at Olaria and 92% at Nobrega (Fig. 4). The corresponding significant (t-test, p < 0.01) increase in the measured CH 4 emission with 200 burrows m −2 present was much larger, reaching 288% at Olaria and 607% at Nobrega. This leads to individual emissions of 1.16−1.49 μmol CH 4 d −1 from pneuma tophores and 3.15−7.65 μmol CH 4 d −1 from burrows (Table 3). When the observed average densities of these biogenic structures at Olaria and Nobrega are considered, the total CH 4 emission from air-exposed sediment was similar at the 2 locations in the warm wet season, with rates of 0.99−1.14 mmol m −2 d −1 .

Exchange of DIC during inundation
DIC release from darkened and inundated fringe mangrove sediments to the overlying water was 158 mmol m −2 d −1 in cores from Olaria, which was not significantly different from cores from Nobrega (122 mmol m −2 d −1 ). Inundated cores from Olaria and Nobrega incubated in light showed a DIC release of 62 and 52 mmol m −2 d −1 , respectively. This leads to a potential warm wet season GPP of 96 and 70 mmol m −2 d −1 , respectively. It should be noted that these cores were exposed to light intensities of 1427 ± 285 μE m −2 s −1 during incubation, which was about 3 times higher than the irradiance recorded in situ under the mangrove canopy.

DISCUSSION
The 2 studied fringe mangrove locations in Brazil were similar with respect to sediment characteristics and C gas fluxes during the warm wet season despite the proximity of Olaria creek to potential anthropogenic impacts from Cananéia City (Barcellos et al. 2005). Others have reported that greenhouse gas emissions from mangrove sediments are increased substantially by anthropogenic nutrient and organic matter inputs (Chen et al. 2011, Zheng et al. 2018, Gnanamoorthy et al. 2022. The anthropogenic impact at Olaria must therefore be considered negligible at the time of our study, confirming the resilience of mangrove environments to moderate eu trophication (Kristensen & Suraswadi 2002, Kristensen et al. 2011, Penha-Lopes et al. 2012, Valiela et al. 2020. The present study demonstrates instead that a major controlling factor for CO 2 and CH 4 emissions from these mangrove sediments is the presence of biogenic structures, like pneumatophores and crab burrows. Consequently, C gas emissions across the sediment− air interface measured at sites where these structures are absent may seriously underestimate the true rates (Kitaya et al. 2002, Lin et al. 2021. It is therefore apparent that the contribution from pneumatophores and crab burrows must be included to fill an often-missing gap in mangrove C budgets (Kristensen et al. 2022).
The background emission of C gases from longterm air-exposed and darkened sediment devoid of pneumatophores and crab burrows at Olaria and Nobrega was 2−3 times higher than the global average compiled by Kristensen et al. (2022), but comparable to that reported for other mangrove sediments (Alongi et al. 2012). The high rates recorded in the present study must therefore be partly related to the lability of sediment organic matter at Olaria and Nobrega. Furthermore, it is known that CO 2 and CH 4 emissions from mangrove sediments driven by microbial degradation of deposited litter and allochthonous materials is strongly enhanced in locations like Olaria and Nobrega, which have long air-exposure time (Alongi et al. 2012, Chanda et al. 2014, Lin et al. 2021. The increased emission of C gases from airexposed sediment in the presence of pneumatophores and crab burrows at Olaria and Nobrega (Fig. 3) substantiates their potential for stimulation of organic matter degradation and importance as airfilled conduits for gas exchange. The contribution from individual biogenic structures to CO 2 emission (381−887 μmol ind. −1 d −1 ) is comparable between locations and type of biogenic structures, and similar to the rates of 207−678 μmol ind. −1 d −1 reported by Kitaya et al. (2002), Kristensen et al. (2008) and Penha-Lopes et al. (2010). Emission of CH 4 from individual pneumatophores of ~1 μmol ind. −1 d −1 at Olaria and Nobrega is comparable to rates reported by Kreuzwieser et al. (2003) and Kristensen et al. (2008). However, the 10−20 times higher individual CH 4 emissions from crab burrows at Olaria and Nobrega than reported previously , Penha-Lopes et al. 2010) are surprising and may reflect extraordinarily extensive burrow networks and deep drainage during the long-term exposure at low tide at the 2 study locations.
Pneumatophores have numerous open lenticels during air exposure that allow for rapid diffusive exchange of gases between the air-filled aerenchyma tissue of roots and the atmosphere (Purnobasuki & Suzuki 2005). The CO 2 emitted from pneumatophores is partly derived from microbial C oxidation in sediments surrounding deep roots via transport across the root epidermis and partly from direct root respiration (Geissler et al. 2002, Kitaya et al. 2002, Lovelock et al. 2006. Drained crab burrows, on the other hand, augment gas exchange by providing a mosaic of deep subsurface sediment−air interfaces for rapid diffusive transport gases to and from the atmosphere. In this case, emitted CO 2 will originate from microbial action within the sediment and from respiration by the crab inhabitant (Datta 2005, Penha-Lopes et al. 2010. Furthermore, common to both pneumatophores and crab burrows is their capacity to stimulate aerobic microbial degradation of organic matter in excess of the background anaerobic rates by their translocation of oxygen deep into the sediment (Nielsen et al. 2003, Kristensen 2008, Thomas & Blum 2010, Cheng et al. 2020. Conversely, emission of CH 4 from pneuma tophores and burrows originates solely from methanogenesis in the surrounding anoxic sediment because plant roots and crabs residing inside burrows do not generate CH 4 (Kreuzwieser et al. 2003, Kristensen et al. 2008, Lin et al. 2021 ) used for microbial heterotrophic respiration are depleted (Canfield et al. 2005). Substantial efflux of CH 4 from mangrove sed-iments therefore requires pneumato phores with aerenchyma tissue or drained crab burrows as deep air-filled conduits, because most of the generated CH 4 otherwise is oxidized before reaching the atmosphere when diffusing slowly upward within the sediment.
The role of biogenic structures for gas exchange during inundation is strongly hampered because lenticels of pneumatophores are closed and burrows typically are plugged by their inhabitants or filled with stagnant water (De la Iglesia et al. 1994, Skelton & Allaway 1996, Allaway et al. 2001. Transport of gases during inundation is therefore limited to diffusion within sediment porewater and at the sediment−water interface. Accordingly, the DIC release from darkened and inundated Olaria and Nobrega sediments is remarkably similar to the CO 2 emission from darkened and air-exposed bare sediment (Fig. 3). Unfortunately, there are no measurements of CH 4 release from inundated sediment in the present study, but previous studies have revealed that the vertical diffusion of CH 4 across the sediment−water interface it is very low in mangrove forests (Alongi et al. 2000, Kreuzwieser et al. 2003, He et al. 2019, Kristensen et al. 2022. Following inundation, release of CO 2 and CH 4 to mangrove waters and subsequent emission to the atmosphere may instead occur by tidal flushing and lateral drainage into creeks via crab burrows and other crevices during ebb tides (Call et al. 2019. Since this contribution was not considered in the present study, the emission budgets for Olaria and Nobrega (Table 4) Rovai et al. (2021) measured litterfall to an equivalent of 30.9 ± 1.4 mol C m −2 yr −1 . This is an underestimate of the true mangrove tree net primary production (NPP) since the contribution from trunk (wood) and root growth is unaccounted for. The latter compartments are typically responsible for about 2 ⁄ 3 of the total tree NPP (Alongi 2020b). By correcting for this deficit, the Cananéia NPP estimate (92.6 mol C m −2 yr −1 ) approaches that of the global average reported by Alongi (2020b). The similarity in CO 2 release from Olaria and Nobrega sediments obtained in the present study indicates that the selected study sites are representative of the Cananéia fringe mangrove forest. With correction for inundation time, light/dark cycles and biogenic structures (Table 4), the CO 2 release measured during February in the warm wet season can be extrapolated to a net ecosystem loss of 8.70 and 9.38 mol C m −2 mo −1 , respectively. However, the fraction of CO 2 emission from pneumatophores that is de rived from root respiration (i.e. ~30%, Ouyang et al. 2018) should not be considered a C sink, as it has already been accounted for in NPP calculations. By correcting for this, the estimated February CO 2 loss is reduced to 7.22 and 8.00 mol C m −2 mo −1 , respectively.
Since monthly average temperature in the Cananéia area ranges from ~28°C in February to ~20°C in July (Cunha-Lignon et al. 2009), an annual emission estimate can only be obtained by correcting sediment metabolism for this temperature difference. By applying a typical Q 10 for sediment microbial processes of 2.5 (Thamdrup et al. 1998), the cold season (July) CO 2 loss at Olaria and Nobrega is 3.5 and 3.8 mol C m −2 mo −1 , leading to a full-year extrapolation of 64.1 and 71.0 mol C m −2 yr −1 . Accordingly, 28.5 and 21.6 mol C m −2 yr −1 of the above-mentioned NPP is not recycled, which is amazingly similar to the C accretion of 23.6 mol C m −2 yr −1 estimated by Rovai et al. (2021). This fringe forest estimate is somewhat higher than the net global mangrove C sequestration estimates of 8−17 mol C m −2 yr −1 reported previously (Alongi 2012, 2020b, Hutchison et al. 2014, Ouyang et al. 2017, Pérez et al. 2018, Jennerjahn 2020. It should be noted, however, that the present C budget does not consider import of allochthonous C and export of particulate and dissolved C to and from the Olaria and Nobrega mangrove forests. The values should therefore be considered rather crude and preliminary estimates. Nevertheless, if biogenic structures are not considered in the C budget, CO 2 emission at Olaria and Nobrega will be reduced by two-thirds, leading to a major C gap that cannot be explained only by the missing import and export. The integrated total CH 4 emissions for Olaria and Nobrega mangrove forests, when rates from February 2014 (Table 4) are extrapolated to a full year after correcting for temperature dependence, is equivalent to a loss of 2.8−3.3 mol C m −2 yr −1 after conversion to C-units using the global warming potential of this gas. The CH 4 contribution driven by pneumatophores and crab burrows will therefore counteract the climate mitigation effect of the buried blue carbon by 10−15%. The important role of biogenic structures for emission of greenhouse gases other than CO 2 was also confirmed by Al-Haj & Fulweiler (2020) and Kristensen et al. (2022). They argued that emission of CH 4 and N 2 O under the influence of these structures may completely expunge the climate effect of blue carbon in some mangrove forests. However, there are considerable uncertainties in the global mangrove climate budgets due to the large variability in NPP estimates, C sequestration, greenhouse gas emissions and tidal C export among forests. Some mangrove environments may have a large mitigation capacity, while others are neutral or even negative. More surveys at larger scales involving biogenic structures appropriately with high spatial and temporal resolution in various mangrove forest types at local, regional and global levels are therefore required to fully elucidate the climate mitigation capacity of these important ecosystems. Nevertheless, the present study emphasizes that C budgets will have serious gaps if the contribution of pneumatophores and crab burrows to greenhouse gas emissions is ignored.