Global Signiﬁcance of Mangrove Blue Carbon in Climate Change Mitigation (Version 1)

: Mangrove forests store and sequester large area-speciﬁc quantities of blue carbon (C org ). Except for tundra and peatlands, mangroves store more C org per unit area than any other ecosystem. Mean mangrove C org stock is 738.9 Mg C org ha − 1 and mean global stock is 6.17 Pg C org , which equates to only 0.4–7% of terrestrial ecosystem C org stocks but 17% of total tropical marine C org stocks. Seagrasses sequester more C org per unit area than mangroves (179.6 g C org m − 2 · a − 1 ) but twice the C org sequestered by mangroves globally (15 Tg C org a − 1 ). Mangroves sequester only 4% (range 1.3–8%) of C org sequestered by terrestrial ecosystems, indicating that mangroves are a minor contributor to global C storage and sequestration. CO 2 emissions from mangrove losses equate to 0.036 Pg CO 2 -equivalents a − 1 based on rates of C sequestration but 0.088 Pg CO 2 -equivalents a − 1 based on complete destruction for conversion to aquaculture and agriculture. Mangrove CO 2 emissions account for only 0.2% of total global CO 2 emissions but 18% of CO 2 emissions from the tropical coastal ocean. Despite signiﬁcant data limitations, the role of mangrove ecosystems in climate change mitigation is globally insigniﬁcant but may be more signiﬁcant and e ﬀ ective at the national and regional scale.


Introduction
The concept of blue carbon was introduced in 2009 in an assessment report to a special collaboration of the United Nations Environmental Programme (UNEP), Food and Agriculture Organization of the United Nations (FAO) and the Intergovernmental Oceanographic Commission of the United Nations Educational, Scientific and Cultural Organization (IOC/UNESCO) [1] with the idea that the role of coastal ecosystems such as salt marshes, mangroves and seagrass meadows in absorbing carbon (C) to reduce emissions is of global significance and they should be protected and, if necessary, restored in order to maintain and expand their ability as critical C sinks. 'Blue carbon', defined as the coastal carbon sequestered and stored by ocean ecosystems [1], has been increasingly used as a concept to justify numerous studies describing C stocks and rates of C sequestration, especially in salt marsh, mangrove and seagrass ecosystems.
A detailed assessment was commissioned by the International Union for Conservation of Nature (IUCN) [2] to document the C management potential of salt marshes, mangrove forests, seagrass meadows, kelp forests and coral reefs. The report found that these coastal habitats are quantitatively and qualitatively important for numerous reasons, including a high potential for C management [2]. The report concluded that (1) sediments and soils in these ecosystems, while small in geographical extent, sequester proportionally more C than terrestrial ecosystems due to lower potential for emissions of greenhouse gases (CH 4 , CO 2 ); (2) there is therefore a critical need for comprehensive C inventories from these habitats to properly assess their role in absorbing C emissions; (3) anthropogenic greenhouse gas emissions are being underestimated because such emissions from these coastal habitats are not being accounted for in national and international inventories, meaning their C savings from sequestration do not count towards meeting climate change commitments; and (4) these habitats continue to be destroyed and need to be protected and restored.
Subsequently published policy reports [3][4][5] indicated that when these habitats are converted their C is released back into the atmosphere, thus reversing the effect of fostering carbon sequestration in REDD+ (Reducing Emissions from Deforestation and Forest Degradation; + refers to conservation and sustainable management and enhancement of carbon stocks) and other rehabilitation projects. Policymakers need to understand that there are three components involved in C sequestration: (1) the annual sequestration rate, that is, the annual flux of organic carbon (C org ) transferred to anaerobic soils and sediments where it cannot undergo oxidation to CO 2 and be released into the atmosphere; (2) the amount of C stored in above-and below-ground biomass; and (3) the total ecosystem C stock stored below-ground as a result of prior sequestration, that is, historical sequestration over a habitat's lifetime.
Since the publication of these seminal publications, there has been an explosion of subsequent papers on blue carbon, with over 1000 papers published since 2009 [6]. This impressive growth reflects the need of NGOs and various agencies around the globe for more data, as well as a lot of enthusiasm for the idea that blue carbon storage and sequestration is of national and international significance in reducing carbon emissions.
Two publications have estimated that mangrove forests, especially if converted to aquaculture ponds, cattle pastures and infrastructure upon deforestation, would account for more than one half of the carbon lost (0.09-0.45 Pg CO 2 a −1 ) [7] from coastal ecosystems to the atmosphere and account for at least as much buried C as salt marshes and seagrasses [8]. However, two more recent publications [6,9] have cast doubt on the global significance of mangroves as C sinks, while at least one other publication [10] concluded that mangrove C is nationally important to Indonesia, due in part to the nation's large mangrove biomass and forest area.
This paper is an attempt to clarify the global and regional significance of mangrove forest C storage and sequestration in reducing and mitigating anthropogenic CO 2 gas emissions. The most recent data will be used to better pinpoint the range of rates of C sequestration, C stocks and potential and actual losses from deforestation.

Carbon Stocks
Mangrove C stocks have been measured in 52 countries in Africa, Southeast Asia, South and East Asia, Central and North America, the Caribbean, South America, the Middle East, Australia, New Zealand and some Pacific Islands (Table 1). Total ecosystem C org stocks average 738.9 ± 27.9 Mg C org ha −1 (± 1SE) with 224 measurements and a median value of 702.5 Mg C org ha −1 ; above-ground biomass C (living and dead) averages 109.3 ± 5.0 Mg C org ha −1 (± 1SE) with 272 measurements, below-ground biomass C (live and dead roots) averages 80.9 ± 9.5 Mg C org ha −1 (± 1SE) with 76.5% of total C stocks vested in mangrove soils (mean = 565.4 ± 25.7 Mg C org ha −1 ) to a depth of at least 1 m (Table 1). These values are considerably lower than the estimates of Alongi [11] and Kauffman et al. [12]. In most cases, minimum and maximum estimates varied by an order of magnitude. Above-ground and below-ground biomass C accounted for 14.8% and 8.7% of total ecosystem C stocks. There is considerable variability in these estimates, reflecting the wide range of ages and geomorphological types of forests, from young plantations to mature undisturbed forests. Also, it is highly likely that the soil C stocks are underestimated in most studies as other studies have measured considerable soil C stocks below 1 m depth (Supplementary Materials Table S1). Further, these data do not include possible inorganic C stocks, particularly in arid mangroves and those near coral reef and mixed terrigenous-carbonate environments [12]. Table 1. Estimates of organic carbon stocks (Mg C org ha −1 ) in mangrove above-ground (AGBC org ) and below-ground root biomass (BGBC org ) and soils (SC org ) to a depth of 1 m. SC org stock estimates taken from cores < 1 m depth are not presented. Some SC org stocks were taken from cores > 1 m depth (see Supplementary Table S1). ND = no data. References are provided in Supplementary Table S1. Using the median of 702.5 Mg C org ha −1 and the most recent estimate of global mangrove area of 83,495 km −2 [13], we derive a global C stock estimate for mangroves of 5.85 Pg C. This estimate is higher than the estimates of 5.0 Pg C by Jardine and Siilamäki [14] and 4.19 Pg C by Hamilton and Friess [15], lower than the estimates by Sanders et al. [16] of 11.2 Pg C and Alongi [6] but within the range (3.7-6.2 Pg C) estimated by Ouyang and Lee [17]. While some of these differences are due to the use of different ecosystem C stock estimates, the main difference is due to the large disparity in the use of estimates of global mangrove area. The higher estimates used the global area estimate of Giri et al. [18] of 137,760 km 2 while the lower estimates used the global area estimate of 83,495 km of Hamilton and Casey [13]. The latter estimate is based on the newest and most accurate databases of the Global Forest Change database, the Terrestrial Ecosystems of the World database and the Mangrove Forests of the World database to extract mangrove forest cover at high spatial and temporal resolutions.

Country
Regionally, total ecosystem C stocks are, on average, greatest on the Pacific Islands (mean = 987.4 Mg C org ha −1 ) of Kosrae, Yap and Palau, followed by mangroves in Southeast Asia (mean = 860.9 Mg C org ha −1 ), Central and North America and the Caribbean (mean = 777.7Mg C org ha −1 ) and Africa (mean = 664.2 Mg C org ha −1 ). Total ecosystem C stocks were considerably lower in Australia and New Zealand (mean = 563.4 Mg C org ha −1 ), South America (mean = 424.0 Mg C org ha −1 ), South and East Asia (mean = 395.5 Mg C org ha −1 ) and the Middle East (mean = 248.4 Mg C org ha −1 ). The size of mangrove C stocks is obviously related to climate, with higher estimates in forests of the humid tropics and lower estimates in the dry tropics and in subtropical and warm temperate regions. This interpretation is supported by the analysis of Sanders et al. [16] who found that 86% of observed variability in mangrove C stocks is associated with annual rainfall, which is the best predictor of mangrove ecosystem C stocks.
At the individual forest level, the smallest C stocks occur in small stands that occur in the arid tropics or are young plantation forests. As forests age, forest biomass and thus C stocks increase. A clear example is the mangrove forests of known age in French Guiana [19]. As the forests age, C stocks in above-and below-ground biomass, soil and the forest ecosystem increase with increasing age (Figure 1). Each of the four C stocks shows significant linear regression (r 2 = 0.959, p < 0.001 for AGBC org ; r 2 = 0.618, p = 0.039 for BGBC org ; r 2 = 0.982, p < 0.001 for soil C org ; and r 2 = 0.979, p < 0.001 for total ecosystem C org ). These data indicate that mangrove forests continue to accumulate organic carbon with increasing age, at least up to 66 years, suggesting that mangrove C is best preserved if mature mangrove forests are conserved and left undisturbed. Plantation data from Vietnamese and Indonesian [20][21][22] mangroves similarly indicate increased C storage with increased stand age.
Sci 2020, 3, x FOR PEER REVIEW 5 of 15 Figure 1. The relationship of mangrove above-(AGBCorg) and below-ground (BGBCorg) biomass C, soil Corg and total ecosystem Corg stocks in different aged forests in French Guiana [19].

Carbon Sequestration Rates
Rates of carbon sequestration, derived from soil accretion rates, in mangroves average 179.6 g Corg m −2 ·a −1 and a median of 103 g Corg m −2 ·a −1 , with rates varying widely from 1 to 1722.2 g Corg m −2 ·a −1 (Figure 2). Half of all observations were in the range of 1-100 g Corg m −2 ·a −1 (Figure 2). The mean value is greater than the estimates of Breithaupt et al. [23], McLeod et al. [24] and Alongi [11]. Assuming a global area of 83,495 km −2 [13] and multiplying by the median value, carbon sequestration in the world's mangrove forests equates to 8.6 Tg Corg a −1 . This value is lower than the 23-25 Tg Corg a −1 calculated by Twilley et al. [25], Jennerjahn and Ittekot [26] and Duarte et al. [27] and the recent estimate of 14.2 Tg Corg a −1 by Alongi [6]. The standard deviation is greater than the mean, reflecting the high level of variability in soil accretion rates and rates of carbon sequestration among mangroves of different ages, types and locations. There was no clear relationship with latitude as it is likely that these rates are a function of several interrelated factors such as forest age, tidal inundation frequency, tidal elevation, geomorphology, species composition, soil grain size, catchment and river input and extent of anthropogenic inputs; most of the highest rates were measured in mature forests in close proximity to river deltas and in forests in highly impacted catchments.

Carbon Sequestration Rates
Rates of carbon sequestration, derived from soil accretion rates, in mangroves average 179.6 g C org m −2 ·a −1 and a median of 103 g C org m −2 ·a −1 , with rates varying widely from 1 to 1722.2 g C org m −2 ·a −1 (Figure 2).
Half of all observations were in the range of 1-100 g C org m −2 ·a −1 (Figure 2). The mean value is greater than the estimates of Breithaupt et al. [23], McLeod et al. [24] and Alongi [11]. Assuming a global area of 83,495 km −2 [13] and multiplying by the median value, carbon sequestration in the world's mangrove forests equates to 8.6 Tg C org a −1 . This value is lower than the 23-25 Tg C org a −1 calculated by Twilley et al. [25], Jennerjahn and Ittekot [26] and Duarte et al. [27] and the recent estimate of 14.2 Tg C org a −1 by Alongi [6]. The standard deviation is greater than the mean, reflecting the high level of variability in soil accretion rates and rates of carbon sequestration among mangroves of different ages, types and locations. There was no clear relationship with latitude as it is likely that these rates are a function of several interrelated factors such as forest age, tidal inundation frequency, tidal elevation, geomorphology, species composition, soil grain size, catchment and river input and extent of anthropogenic inputs; most of the highest rates were measured in mature forests in close proximity to river deltas and in forests in highly impacted catchments.

Carbon Losses
Blue carbon storage in mangroves may be underestimated by considering soil Corg pools only to a depth of 1m but may be offset by losses of CH4 and oxidation of ancient Corg stored in deep soils [49,50]. Some of the soil Corg is decomposed and returned to the atmosphere as CH4. As CH4 has a higher global warming potential than CO2, it can offset the CO2 removed via Corg burial. Rosentreter et al. [49] calculated that high CH4 emissions from mangroves can partially offset blue carbon burial rates on average by 20% using the 20-year global warming potential. Corg buried in mangrove deposits not only releases CH4 but also century-old sequestered carbon in the form of exported dissolved inorganic carbon (DIC). In a subtropical mangrove system, ∆ 14 C was measured in the DIC exported from the pore water and soil ∆ 14 C profiles. Pore water exchange released isotopically depleted, old DIC to adjacent creek waters [50]. The DIC came from an average depth of 40 cm, equivalent to about a century of soil accumulation. Thus, 100-yr old DIC is still susceptible to remineralization and tidal export via pore water exchange or submarine groundwater discharge.
The loss of mangroves, irrespective of cause, results in significant loss of Corg inventory, especially if the soil horizon is removed or disturbed. This removal can be converted to CO2-eq (equivalent) emissions back to the atmosphere. Immediate removal of biomass and soil of destroyed mangrove forests to convert the area to aquaculture ponds, cattle pastures and other land uses results in extremely high losses (Table 2), with CO2 eq emissions averaging 1802.2 Mg ha −1 ·a −1 and ranging from 407.9 to 2781.5 Mg ha −1 ·a −1 [56,57] as estimated in Brazil, Mexico, the Philippines, Honduras, Dominican Republic, Indonesia and Costa Rica. Most of these emissions come from loss of the soil pool to a depth of 1 m. If soils deeper than 1 m are dredged, the estimated CO2 eq will be greater.

Carbon Losses
Blue carbon storage in mangroves may be underestimated by considering soil C org pools only to a depth of 1m but may be offset by losses of CH 4 and oxidation of ancient C org stored in deep soils [49,50]. Some of the soil C org is decomposed and returned to the atmosphere as CH 4 . As CH 4 has a higher global warming potential than CO 2 , it can offset the CO 2 removed via C org burial. Rosentreter et al. [49] calculated that high CH 4 emissions from mangroves can partially offset blue carbon burial rates on average by 20% using the 20-year global warming potential. C org buried in mangrove deposits not only releases CH 4 but also century-old sequestered carbon in the form of exported dissolved inorganic carbon (DIC). In a subtropical mangrove system, ∆ 14 C was measured in the DIC exported from the pore water and soil ∆ 14 C profiles. Pore water exchange released isotopically depleted, old DIC to adjacent creek waters [50]. The DIC came from an average depth of 40 cm, equivalent to about a century of soil accumulation. Thus, 100-yr old DIC is still susceptible to remineralization and tidal export via pore water exchange or submarine groundwater discharge.
The loss of mangroves, irrespective of cause, results in significant loss of C org inventory, especially if the soil horizon is removed or disturbed. This removal can be converted to CO 2 -eq (equivalent) emissions back to the atmosphere. Immediate removal of biomass and soil of destroyed mangrove forests to convert the area to aquaculture ponds, cattle pastures and other land uses results in extremely high losses (Table 2), with CO 2 eq emissions averaging 1802.2 Mg ha −1 ·a −1 and ranging from 407.9 to 2781.5 Mg ha −1 ·a −1 [51,52] as estimated in Brazil, Mexico, the Philippines, Honduras, Dominican Republic, Indonesia and Costa Rica. Most of these emissions come from loss of the soil pool to a depth of 1 m. If soils deeper than 1 m are dredged, the estimated CO 2 eq will be greater. Table 2. Losses of blue carbon via CO 2 eq (Mg ha −1 ·a −1 ) emissions from degraded mangroves worldwide. ND = no data. a = Mg CO 2 eq ha −1 lost immediately upon conversion/hurricane disturbance; b = above-ground biomass C losses only. Hurricanes and typhoons can destroy significant areas of mangroves, as estimated in the Philippines, Honduras, Vietnam and in Florida ( Table 2). Averaging the remaining estimates (n = 20), we derive an average emission of 65.2 ± 10.6 Mg CO 2 eq ha −1 ·a −1 (± 1 SE) with a median of 46 Mg CO 2 eq ha −1 ·a −1 ( Table 2). Assuming total deforestation of mangroves (biomass + soils to 1 m depth) and using the mean CO 2 eq emission of 1802.2 Mg CO 2 eq ha −1 ·a −1 and multiplying by an annual average deforestation rate of 0.16% [13,15] and a global mangrove area of 83,495 km −2 [13], we can estimate an annual loss of 24.08 Tg CO 2 eq a −1 or 0.0024 Pg CO 2eq a −1 . This estimate is considerably less than those of Pendleton et al. [7] and Alongi [6] mostly due to lower recent estimates of annual deforestation and less global mangrove area. Mangrove losses are small on a global scale, equating to just 2.2% of CO 2 losses due to losses (1.1 Gt C a −1 ) of the world's tropical terrestrial forests [69] and offsetting just 1.8% of the carbon sink (1.32 Pg a −1 ) in the global ocean's continental margins [70]. However, mangrove losses offset 148.6% of total CO 2 -air-sea exchange (−16.21 Tg C a −1 ) by the world's tropical coastal zone [71].

Disturbance
Are mangrove blue C stocks and C sequestration rates globally significant? The global mean C stock for mangroves is estimated to be 6.17 Pg C org , which is the largest C stock of any ecosystem in the global tropical ocean, constituting~17% of total tropical marine C stocks (Table 3). Although mean mangrove C stocks per unit area are the largest among the world's ecosystems (except tundra and peatlands), global mangrove C stocks equate to only 1.6% (range: 0.4-7%) of individual terrestrial ecosystem global C stocks (Table 3). Regarding C sequestration among coastal environments, seagrass meadows sequester slightly more than twice (35.3 Tg C org a −1 ) the amount of mangroves (15Tg C org a −1 ). Mangroves sequester~50% of tropical peatlands globally but only 4% compared to other terrestrial ecosystems (range: 1.3-8%). CO 2 emissions due to deforestation and other destructive land use practices result in large returns of CO 2 to the atmosphere, for a total of roughly 51 Pg CO 2 -eq a −1 ( Table 3). While the same calculations for mangroves result in an estimate of 0.036 Pg CO 2 -eq a −1 , in some regions mangrove biomass and soils are entirely removed (Section 4) resulting in mean C losses of 1802.2 Mg C org ha −1 ·a −1 . Assuming that all mangroves are so destroyed at a rate of 0.16% per year, total CO 2 emissions equate to 0.088 Pg CO 2 -eq a −1 rather than the lower estimate based solely on losses of global C sequestration (see footnote b in Table 3).
While there is no doubt that mangroves store and sequester large amounts of carbon relative to their small global area, a perusal of Table 3 indicates that they play only a minor global role in storing C org and in mitigating CO 2 emissions. Mangrove CO 2 emissions account for roughly 0.2% of total global CO 2 emissions and account for about 18% of CO 2 emissions from the tropical coastal ocean. It must be noted that these C stock and C rate estimates are crude and can only point to relative differences, as there are significant data limitations. As pointed out by Taillardat et al. [9] and Alongi et al. [10], climate change mitigation is likely to be more significant and effective at the national scale especially in countries losing mangroves rapidly, such as in Indonesia and Brazil. Table 3. Estimated area-specific and global C stocks, C sequestration rates and CO 2 emissions due to losses from mangrove forests, salt marshes, seagrass meadows, coral reefs, the tropical coastal ocean and terrestrial ecosystems.

Ecosystem
Area ( [117] a = from Tables 1 and 2; b = estimated assuming total forest biomass and soil losses to a depth of 1m (see Section 4). CO 2 emissions based on global sequestration rate are in parentheses. c = weighted average of conversion rates for mangroves, seagrasses and coral reefs.
Funding: This research received no external funding.

Conflicts of Interest:
The author declares no conflict of interest.