Carbon Stocks Assessment in a Disturbed and Undisturbed Mangrove Forest in Ghana

: Mangroves and other blue carbon ecosystems have long been recognised for their carbon sink function, yet the organic carbon stocks of mangroves in many countries in Africa remain to be assessed. This study evaluates the impact of traditional forest conservation on long-term carbon sequestration in a non-degraded (Amanzule) and a degraded (Kakum) mangrove forest system in Ghana (West Africa). The amount of carbon stored in mangrove trees was estimated from diameter-based allometric equations. Tree (above- and below-ground) carbon was ~34-fold higher in the Amanzule forest (mean = 0.89 ± 0.10 t/ha) than in the Kakum forest (mean = 0.026 ± 0.019 t/ha). Soil carbon density was estimated as organic carbon and bulk density at speciﬁc depths in both forests. Soil organic carbon density was ~5-fold higher in the Amanzule forest (mean = 2935.79 ± 266 t/ha) than the Kakum forest (mean = 554.01 ± 83 t/ha). The variation in the vertical distribution of soil carbon was not signiﬁcant in either forest ( F = 0.57; p > 0.05). These ﬁndings underscore the role of traditional conservation on mangrove carbon stocks and the need to consider the governance of coastal ecosystems when estimating blue carbon.


Introduction
It has been well established that the level of carbon dioxide (CO 2 ) in the atmosphere is rising at an alarming rate ≈ 2.36 ppm per year over the past six decades [1]. This increasing amount of atmospheric CO 2 is the leading cause of global climate change, responsible for increased drought, intense flooding, and sea level rise with dire consequences for critical life-supporting ecosystems and agricultural production among other impacts on human livelihoods [2][3][4].
There is currently a broad consensus among the scientific community that conservation of ecosystems that serve as natural carbon sinks can reduce the level of CO 2 in the atmosphere [5][6][7][8][9]. These ecosystems include some, but not all, inland forests, mangroves, salt marshes, and seagrass beds which capture CO 2 through photosynthesis, thus trapping carbon in their living biomass above and below ground, as well as in dead tissues buried in the sediments. Together, the above-ground, below-ground and sediment-bound dead tissues of these blue ecosystems hold around 2000 Gt of carbon, which is over 2.5 times the amount of carbon that is currently held in the atmosphere [10].
In particular, mangrove ecosystems along the world's coastlines hold three billion metric tons of carbon. Except for the tundra and peatlands, mangroves store more blue carbon per unit area than any other ecosystem in the world [11][12][13][14]. This is possible because mangrove forest sediments do not become saturated with carbon [15]. Rather, the rate of carbon sequestration increases over time with the accretion of sediment within mangrove forests [11,16]. Furthermore, the carbon trapped within sediments underlying mangrove forests is held for a longer period than in other plant-dominated systems (see, for example, Kakum is located in the dry equatorial zone of Ghana with coastal savannah as the major vegetation type. The area experiences high rainfall with the wettest periods in May/June and September/October each year. The mean annual rainfall is about 1000 mm with the average monthly temperature ranging between 24°C to 30 °C [44]. The topography of the Kakum forest is flat with forest ochrosols as the dominant soil type [45,46].
The Amanzule forest is part of the Amanzule wetlands regarded as the dwelling place of the gods by local communities. As a consequence of this traditional belief, the Amanzule forest has been under protection for many years [44], managed by local traditional norms and customs even though the forest has no formal conservation status [37]. In contrast, the Kakum mangrove forest is intensely exploited for fuel wood and other domestic uses [42].
Three sampling plots (herein referred to as plots), each measuring 125 m × 40 m were established in each mangrove forest. Each plot was divided into 18 subplots of 10 m × 10 m per plot (herein referred to as subplots) ( Figure 2). The plots were sited in areas previously determined to have dense mangrove cover [47]. To ensure comparability, the plots in both forests were located close to the low water mark, with plot A furthest from the low water mark and plot C closest to the low water mark. Kakum is located in the dry equatorial zone of Ghana with coastal savannah as the major vegetation type. The area experiences high rainfall with the wettest periods in May/June and September/October each year. The mean annual rainfall is about 1000 mm with the average monthly temperature ranging between 24 • C to 30 • C [44]. The topography of the Kakum forest is flat with forest ochrosols as the dominant soil type [45,46].
The Amanzule forest is part of the Amanzule wetlands regarded as the dwelling place of the gods by local communities. As a consequence of this traditional belief, the Amanzule forest has been under protection for many years [44], managed by local traditional norms and customs even though the forest has no formal conservation status [37]. In contrast, the Kakum mangrove forest is intensely exploited for fuel wood and other domestic uses [42].
Three sampling plots (herein referred to as plots), each measuring 125 m × 40 m were established in each mangrove forest. Each plot was divided into 18 subplots of 10 m × 10 m per plot (herein referred to as subplots) ( Figure 2). The plots were sited in areas previously determined to have dense mangrove cover [47]. To ensure comparability, the plots in both forests were located close to the low water mark, with plot A furthest from the low water mark and plot C closest to the low water mark.

Data Collection and Analyses
Data were collected during low tide to ensure easy access to plots. Carbon stock within the selected mangroves was assessed based on three measures: total above-ground carbon (AC), total below-ground carbon (BC), and total organic carbon (SC) in different layers of soil following the method of Kauffman and Donato [48]. The above-and belowground carbon refer to the amount of carbon stored in living plant tissues above and below the soil, respectively [49]. The above-and below-ground carbons were estimated using Equation (1): where W is tree biomass, and f is the biomass-to-carbon conversion factor of 0.46 and 0.39 respectively, for above-and below-ground biomass [50]. The above-and below-ground biomass of each tree was determined using Equations (2) and (3), respectively [51]: and where Wtop is above-ground biomass, WR is below-ground biomass, D is tree diameter at breast height, and p is the specific wood density based on conservative estimates by Howard et al. [50] for different mangrove species. The tree diameter at breast height was determined at 1.37 m above ground using a Vernier calliper for smaller trunks. For larger tree trunks, the circumference was measured with a tape and the value divided by π to obtain the diameter. In the case of Rhizophora spp. the diameter was measured at 30 cm above the highest stilt root.
The organic carbon content of the soil (SC) was measured from soil samples collected with an auger and a rectangular soil sampler (volume: 120 cm 3 ) at depths of 0-15 cm, 16-30 cm, 31-50 cm, and 51-100 cm, following the method of Kauffman and Donato [8]. Samples were stored in opaque polyethene bags and transported to the laboratory for analysis.

Data Collection and Analyses
Data were collected during low tide to ensure easy access to plots. Carbon stock within the selected mangroves was assessed based on three measures: total above-ground carbon (A C ), total below-ground carbon (B C ), and total organic carbon (S C ) in different layers of soil following the method of Kauffman and Donato [48]. The above-and belowground carbon refer to the amount of carbon stored in living plant tissues above and below the soil, respectively [49]. The above-and below-ground carbons were estimated using Equation (1): where W is tree biomass, and f is the biomass-to-carbon conversion factor of 0.46 and 0.39 respectively, for above-and below-ground biomass [50]. The above-and below-ground biomass of each tree was determined using Equations (2) and (3), respectively [51]: and where W top is above-ground biomass, W R is below-ground biomass, D is tree diameter at breast height, and p is the specific wood density based on conservative estimates by Howard et al. [50] for different mangrove species. The tree diameter at breast height was determined at 1.37 m above ground using a Vernier calliper for smaller trunks. For larger tree trunks, the circumference was measured with a tape and the value divided by π to obtain the diameter. In the case of Rhizophora spp. the diameter was measured at 30 cm above the highest stilt root.
The organic carbon content of the soil (S C ) was measured from soil samples collected with an auger and a rectangular soil sampler (volume: 120 cm 3 ) at depths of 0-15 cm, 16-30 cm, 31-50 cm, and 51-100 cm, following the method of Kauffman and Donato [8]. Samples were stored in opaque polyethene bags and transported to the laboratory for analysis. In the laboratory, each layer of the soil was dried in the oven at 105°C to constant weight, and ground in a porcelain mortar. The homogenized samples were sieved through a 0.5 mm mesh to remove root parts and analysed for organic carbon content using the modified dichromate oxidation method of Bajgai et al. [52]. Following oxidation, the organic carbon concentration was calculated using Equation (4) [53]: where O C is the weight of carbon per unit weight of soil layer, A is the volume (mL) of dichromate consumed during boiling, N FAS is the normality (0.2 N) of the ferrous ammonium sulphate solution used in the oxidation, whilst Dw is the weight (g) of the dried homogenised soil. The total weight of organic carbon occurring within each soil layer (Kg C m −2 ) was calculated using Equation (5): where T is the thickness (m) of the soil layer and B is the bulk density (kg m −3 ) of the soil layer sampled. The total carbon stock (Kg C ha −1 ) from each of the plots was determined by summing the total above-ground carbon (A C ), below-ground carbon (B C ) and carbon in the different soil layers (S C ). The mangrove tree populations in the two study areas were characterised based on species density, relative density, total basal area, and relative dominance. The stand density of each subplot (number of trees per hectare) was calculated as: The mean total density (±s.d.) of trees in each plot was obtained based on the number of trees in the subplots.
The basal area (m 2 ) of each tree was computed as DBH 2 × 7.854 −5 following [42] and the total tree basal area (TBA m 2 ha −1 ) as:

TBA =
Sum of the basal area for all tree species Area of sampling plot (m 2 ) (7)

Statistics Used
The sampling was designed to investigate three major factors that are likely to influence the results. These factors were the type of forest management (conservation vs. exploitation), location of sampling area within the forests, and subplots. The vegetation structure in each mangrove forest was assessed using the density and basal area of the tree species. The basal area of tree species from the sampling locations was compared using a two-way analysis of variance (two-way ANOVA) with forest type and sampling plot as the sources of variance.
Possible differences in the biological characteristics of the individual species in different forest systems were assessed using a two-way ANOVA with forest type and sampling plot as the sources of variance. Total above-ground carbon (A C ), total below-ground carbon (B C ), and soil organic carbon (S C ) measured in the different forest systems were also compared using a two-way ANOVA. Organic carbon concentration in the different layers of soil from each forest was compared. Tukey's test (α = 0.05) was performed to determine which pairs of means were significantly different.

Vegetation Composition and Structure
Three mangrove tree species, namely Rhizophora mangle, Avicennia germinans and Laguncularia racemosa were found in the Kakum and Amanzule forests. Avicennia germi- nans was the dominant species in the Kakum forest with 86.8% cover whereas R. mangle dominated the Amanzule forest with a cover of 85.9%. Notably, R. mangle was the tallest tree in both forests while L. racemosa was the shortest tree (Table 1). Tree girth sizes were significantly different among species in the Kakum forest with A. germinans and L. racemosa having the biggest and smallest girth sizes respectively (F = 7.59; p < 0.05). Girth sizes differed significantly among trees in the Amanzule forest with R. mangle and L. racemosa having the biggest and smallest girth sizes respectively ( Table 1). The three mangrove species were taller and had greater trunk diameters in the Amanzule forest than in the Kakum forest. Basal area of A. germinans differed between forests and among plots (p = 0.000 for both). Basal area of R. mangle and L. racemosa differed significantly among plots but not between sites (p = 0.217; p = 0.253 respectively).
Kakum had a higher mean tree density than Amanzule ( Table 2). The mean tree biomass in the Amanzule forest (mean AG + BG biomass = 1.21 ± 4.6 t ha −1 ) was greater than in the Kakum forest (mean AG + BG biomass = 0.03 ± 0.02 t ha −1 ); t (1366) = 9.46, p = 0.00). In the Kakum forest, tree biomass was similar for plots A and B whereas plot C had the highest tree density and biomass. In Amanzule forest, plot C had the highest tree density and lowest tree biomass. The highest tree biomass was recorded in plot A in the Amanzule forest (Table 2). A. germinans and R. mangle had the highest densities in the Kakum and Amanzule forests respectively. This is reflected in the high total basal area observed for the respective species in both forests (Table 3). Species with the least densities were R. mangle (Kakum forest) and L. racemosa (Amanzule forest). Whereas R. mangle was about 10 times greater in Amanzule than in Kakum, L. racemosa was~14 times higher in Kakum than in Amanzule (Tables 3 and 4 Table 4. Two-way ANOVA comparing biological characteristics of mangrove species found in forest protected by traditional customs (Amanzule) and forest exposed to consumptive exploitation (Kakum).

Carbon Stock Estimates
Soil organic carbon densities were ~5-fold higher in the Amanzule forest than in the Kakum forest. There was no significant vertical variation in soil organic content in either Generally mean bulk density increased with depth in both forests. However, the increase was not statistically significant in either forest (F = 0.57; p > 0.05) (Figure 3).

Carbon Stock Estimates
Soil organic carbon densities were~5-fold higher in the Amanzule forest than in the Kakum forest. There was no significant vertical variation in soil organic content in either forest (Figure 4). Soil carbon density ranged from 462 ± 308 t/ha (at 0-15 cm depth) to 648 ± 378 t/ha (50-100 cm depth) in the Kakum forest and from 2797 ± 973 t/ha (0-15 cm depth) to 3119 ± 1009 t/ha (50-100 cm depth) in the Amanzule forest.
forest (Figure 4). Soil carbon density ranged from 462 ± 308 t/ha (at 0 −15 cm depth) to 648 ± 378 t/ha (50-100 cm depth) in the Kakum forest and from 2797 ± 973 t/ha (0-15 cm depth) to 3119 ± 1009 t/ha (50-100 cm depth) in the Amanzule forest. The amount of organic carbon stored at specific depths differed between Kakum and Amanzule forests and among plots within each forest (Table 5). Table 5. Two-way ANOVA comparing organic carbon in different layers of soil found in forest protected by traditional customs (Amanzule) and forest exposed to consumptive exploitation (Kakum).  Figure 4). Rhizophora mangle contained about 100 The amount of organic carbon stored at specific depths differed between Kakum and Amanzule forests and among plots within each forest (Table 5). Table 5. Two-way ANOVA comparing organic carbon in different layers of soil found in forest protected by traditional customs (Amanzule) and forest exposed to consumptive exploitation (Kakum). Rhizophora mangle had the highest carbon density per tree in both forests followed by A. germinans and L. racemosa (Table 6; Figure 4). Rhizophora mangle contained about 100 times; A. germinans contained about 36 times and L. racemosa contained about 5 times more carbon in the Amanzule forest than in the Kakum forest (Table 6). Estimated above-ground carbon density was higher than estimated below-ground carbon density for all tree species in both forests. Average above-and below-ground tree carbon densities were about 28 times and 17 times higher, respectively, in Amanzule forest than in Kakum forest ( Figure 5). Above-ground tree carbon was~1.9-fold and 2.7-fold higher than below-ground tree carbon for all species in Kakum and Amanzule respectively. Significant differences (F = 31.41; p = 0.00) occurred in carbon density among mangrove species in both forests. Tree carbon is lower in the Kakum forest (0.03 ± 0.02 t/ha) relative to the Amanzule forest (0.89 ± 1.65 t/ha). Spatially, tree carbon differed significantly across sampling plots in the Kakum (F = 11.49; p < 0.05) and Amanzule (F = 144.83; p < 0.05) forests (Tables 7 and  8). A two-way ANOVA revealed that tree carbon density was significantly affected by the Tree carbon is lower in the Kakum forest (0.03 ± 0.02 t/ha) relative to the Amanzule forest (0.89 ± 1.65 t/ha). Spatially, tree carbon differed significantly across sampling plots in the Kakum (F = 11.49; p < 0.05) and Amanzule (F = 144.83; p < 0.05) forests (Tables 7 and 8). A two-way ANOVA revealed that tree carbon density was significantly affected by the forest type and sampling plot for both above-ground, F (2, 3564) = 239.1, p = 0.00, and below ground, F (2, 3564) = 337.7, p = 0.00 carbon (Table 8). Soil carbon was lower in the Kakum forest than in the Amanzule forest. In both forests, soil carbon was higher than tree carbon. Soil carbon was~5-fold higher in Amanzule forest (2935.79 ± 266 t/ha) than in Kakum forest (554.01 ± 83) whereas tree carbon was~34-fold higher in Amanzule (0.89 ± 0.10 t/ha) than in Kakum (0.026 ± 0.019 t/ha).

Discussion and Conclusions
This study examined carbon storage in two mangrove ecosystems in Ghana-the protected Amanzule forest in the Western Region, and the unprotected Kakum forest in the Central Region of the country where there is unmitigated wood extraction. Above-and below-ground tree carbon was estimated from the diameter and density of the mangrove trees whereas soil organic carbon was calculated from soil samples.
The results show that the Amanzule forest had about thirty times more above-ground biomass than the Kakum forest. This corroborates the observation of Tamooh et al. [54] that logging and other anthropogenic disturbances reduce the health of mangrove forests and affect their overall carbon stock density. A greater part of the degraded Kakum mangrove forest is harvested for fuel wood and other subsistence uses due to lack of enforcement of traditional forest protection rules. Consequently, very few large trees remain in the forest. Mangrove trees in the Kakum forest were stunted as indicated by their low average height and diameter and as also observed by other workers (see [42,55]). In contrast, the Amanzule species were taller with wider girths, thus underscoring the beneficial effect of traditional protection on the health of natural ecosystems.
The high tree density observed in the Kakum forest is reflective of the open nature of the forest canopy. The trees are generally short and as such seedlings receive adequate warmth for growth. The apparent poor stature of the mangrove trees in the Kakum forest could be attributed to the prevailing soil type (ochrosols) which has low resistance to degradation, low nutrient levels and contains toxic concentration of aluminium see the reference [56]. These conditions, in concert with high bulk density (more mineral particles) observed in Kakum forest, do not provide the best conditions for mangrove growth performance. Studies have shown that sediments characterised by high soil density are less porous and rich in mineral particles [57]. Such sediments restrict tree growth via their impact on soil aeration and water penetration. The relatively low density of mangrove trees observed in the Amanzule forest is attributable to the high tree canopy with big stem sizes and therefore there was little chance for the survival of seedlings. The big stem sizes of mangrove trees encountered in the Amanzule forest influenced the high carbon density values recorded and support the view of Donato et al. [58] and Assefa et al. [59] that the quantity of carbon stored is primarily determined by the size of the stand, canopy height and stature. For instance, the high soil carbon values recorded are characteristics of the high organic matter content and the forest oxysols-ochrosols intergrade soils dominating the Amanzule forest. This study has also shown that several years of carbon sequestration, devoid of intensive tillage, in the Kakum and Amanzule mangrove forests may have stored a great amount of carbon below ground (as peat) and accrete. We recorded high values of soil carbon,~1000-fold higher than the values recorded for tree carbon in both forests. This is consistent with the general observation that soil C accounts for > 90% of total ecosystem C in mangrove forests [41,58,60]. The soil carbon recorded in this study is higher than global estimates [11] but similar to values reported for West-Central Africa see the reference [41].
The present study highlights the importance of non-degraded mangrove forests in capturing carbon. Forest degradation contributes to increasing levels of greenhouse gases in the atmosphere through the release of stored carbon in tree biomass. By implication, degraded coastal ecosystems such as mangroves are converted from net carbon sinks to net carbon sources. Therefore, carbon offsets through conservation of mangrove ecosystems could be far more cost-effective in addressing global carbon flux than current approaches focused on other terrestrial trees. Protection of mangrove ecosystems may also have enormous add-on benefits to fisheries and potentially limit coastal erosion.
This study contributes to carbon estimates for West-Central Africa as reported by Kauffman and Bhomia (2017). It confirms the carbon storage potential of mangrove ecosystems and highlights the impact of local cultural practices on mangrove carbon stocks. There is, therefore, an urgent need to consider the use of local tradition and governance approaches to improve conservation of mangrove ecosystems.
Given the high dependence of local inhabitants on services provided by mangroves, conservation measures and research towards the sustainable use of mangroves in Ghana must be prioritized in addition to sensitization and strengthening of local governance systems. The identification and scale-up of supplementary livelihood options for rural coastal communities, in the context of the blue economy, should be considered to reduce pressure on mangrove ecosystems. This study provides a snapshot of carbon stored in degraded and non-degraded mangrove forests without recourse to carbon flux caused by anthropogenic and ecological changes. Additionally, non-randomization of sampling plots in this study may account for a lower average carbon stocks value for the whole forest due to lower tree density in some areas outside of sampling plots. It will, therefore, be important to establish permanent plots to conduct longitudinal studies to generate data for the reporting of Ghana's nationally determined contribution of greenhouse gas emissions and removals as stipulated by the United Nations Framework Convention on Climate Change.