A Comparison of Soil Carbon Stocks of Intact and Restored Mangrove Forests in Northern Vietnam

Background and Objectives: In northern Vietnam, nearly 37,100 hectares of mangroves were lost from 1964–1997 due to unsustainable harvest and deforestation for the creation of shrimp aquaculture ponds. To offset these losses, efforts in the late 1990s have resulted in thousands of hectares of mangroves being restored, but few studies to date have examined how effective these efforts are at creating restored mangrove forests that function similarly to the intact mangroves they are intended to replace. Materials and Methods: We quantified and compared soil carbon (C) stocks among restored (mono and mixed species) and intact mangrove forests in the provinces of Quang Ninh, Thai Binh, Nam Dinh and Thanh Hoa in northern Vietnam. A total of 96 soil cores up to a depth of 200 cm were collected every 25 m (25, 50, 75, 100, 125, and 150 m) along 16 linear transects that were 150 m long and perpendicular to the mangrove upland interface (six cores along each transect) at Quang Ninh (four transects), Thai Binh (five), Nam Dinh (four) and Thanh Hoa (three). Five-cm-long soil samples were then collected from the 0–15 cm, 15–30 cm, 30–50 cm, 50–100 cm, and >100 cm depth intervals of each soil core. Results: The study confirmed that the soil C stock of 20–25-year-old restored mangrove forest (217.74 ± 16.82 Mg/ha) was not significantly different from that of intact mangrove forest (300.68 ± 51.61 Mg/ha) (p > 0.05). Soil C stocks of Quang Ninh (323.89 ± 28.43 Mg/ha) were not significantly different from Nam Dinh (249.81 ± 19.09 Mg/ha), but both of those were significantly larger than Thai Binh (201.42 ± 27.65 Mg/ha) and Thanh Hoa (178.98 ± 30.82 Mg/ha) (p < 0.05). Soil C stock differences among provinces could be due to their different geomorphological characteristics and mangrove age. Soil C stocks did not differ among mangroves that were restored with mixed mangrove species (289.75 ± 33.28 Mg/ha), Sonneratia caseolaris (L.) Engl. (255.67 ± 13.11 Mg/ha) or Aegiceras corniculatum (L.) Blanco (278.15 ± 43.86 Mg/ha), but soil C stocks of those mangroves were significantly greater than that of Kandelia obovata Sheue, Liu & Yong (174.04 ± 20.38 Mg/ha) (p < 0.05). Conclusion: There were significant differences in the soil C stocks of mangrove forests among species and provinces in northern Vietnam. The soil C stock of 20–25-year-old restored mangrove forest was not significantly different from that of intact mangrove forest.


Introduction
Mangrove forests are well known for their ability to fix and store large amounts of atmospheric CO 2 [1][2][3][4], one of the main greenhouse gases that contributes to global warming [5]. This carbon (C) is assimilated and stored in aboveground tree biomass (i.e., stems, branches, leaves and roots) and in soils as both belowground roots and trapped sediments [6]. Up to 90% of mangrove C stocks are stored in anaerobic and waterlogged soil conditions that inhibit microbial breakdown of C and can result in long-term storage if left undisturbed [7][8][9][10]. This also results in mangroves storing 10-15% of coastal sediment C stocks despite only representing 0.5% of the world's coastlines [11,12]. The ability to remove and store large amounts of C has significantly increased the role mangroves play in the reduction of greenhouse gas emissions and thus climate change mitigation and adaptation [13].
Mangroves in Vietnam are mainly distributed in the estuarine areas of 29 coastal provinces. Total mangrove forest area of Vietnam was estimated at 408,500 ha in 1943 [14]. However, the area of from sediment inputs from the Thai Binh, Hong and Ma rivers. These accreted mudflats in Thai Binh, Nam Dinh and Thanh Hoa are larger than in Quang Ninh, but are subjected to strong winds and waves. The watersheds in Thai Binh, Nam Dinh and Thanh Hoa are larger than Quang Ninh and, as a result, these systems tend to have lower salinities that range from 0.5% -25% [25,26].
The total area of mangroves in northern Vietnam decreased from 47,000 ha in the 1980s to 11,000 ha in the 1990s. This was mainly due to deforestation for the creation of shrimp aquaculture ponds. Today, the area of mangroves in northern Vietnam has increased to more than 31,200 ha. The recent increase in mangrove area is largely due to restoration efforts and protection policies; nearly 75% of the mangrove area has been intentionally restored [27,28]. Mangrove forests are dominated by trees of S. caseolaris, Bruguiera gymnorrhiza (L.) Lam., K. obovata, Aegiceras corniculatum (L.) Blanco, R. stylosa and Avicennia marina (Forssk.) Vierh. The restored mangroves were established mainly in Thai Binh, Nam Dinh, and Thanh Hoa provinces during 1993-1998, through planting of K. obovata, S. caseolaris, or both of these species (mixed).
A total of 96 soil cores were collected using a 6.4-cm open-face auger every 25 m (25, 50, 75, 100, 125, and 150 m) along 16 linear transects, that were 150 m long and perpendicular to the mangrove upland interface (6 cores along each transect) at Quang Ninh (4 transects), Thai Binh (5), Nam Dinh (4) and Thanh Hoa (3) in May and June of 2018. The sampling sites and transects are described in Figure 1 and Table 1. Five-cm-long soil samples were then collected at the 0-15-cm, 15-30-cm, 30-50-cm, 50-100-cm, and >100-cm intervals [29]. All wet soil samples were dried to a constant mass at 60 • C to determine the bulk density (BD) for each interval. The dried samples were ground by hand using a mortar and pestle, then sieved through a 2-mm mesh to remove any large roots or rocks before being analyzed for C content (%C). Bulk density was obtained by the ratio of dry mass per volume of wet sample. Due to limited resources, C content in all soil samples was determined using the Walkley-Black method [30]. Carbon content of 50 randomly selected soil samples was also analyzed using the combustion method on a Thermo Scientific™ FlashSmart™ Elemental Analyzer (EA) to calculate correction coefficients for C contents from the Walkley-Black method (WB). The regression relationship of WB with EA is y = 0.667x + 0.361 (R 2 = 0.841, p < 0.001).
Forests 2020, 11, x FOR PEER REVIEW 3 of 10 Nam Dinh and Thanh Hoa are larger than in Quang Ninh, but are subjected to strong winds and waves. The watersheds in Thai Binh, Nam Dinh and Thanh Hoa are larger than Quang Ninh and, as a result, these systems tend to have lower salinities that range from 0.5‰-25‰ [25,26]. The total area of mangroves in northern Vietnam decreased from 47,000 ha in the 1980s to 11,000 ha in the 1990s. This was mainly due to deforestation for the creation of shrimp aquaculture ponds. Today, the area of mangroves in northern Vietnam has increased to more than 31,200 ha. The recent increase in mangrove area is largely due to restoration efforts and protection policies; nearly 75% of the mangrove area has been intentionally restored [27,28]. Mangrove forests are dominated by trees of S. caseolaris, Bruguiera gymnorrhiza (L.) Lam., K. obovata, Aegiceras corniculatum (L.) Blanco, R. stylosa and Avicennia marina (Forssk.) Vierh. The restored mangroves were established mainly in Thai Binh, Nam Dinh, and Thanh Hoa provinces during 1993-1998, through planting of K. obovata, S. caseolaris, or both of these species (mixed).
A total of 96 soil cores were collected using a 6.4-cm open-face auger every 25 m (25, 50, 75, 100, 125, and 150 m) along 16 linear transects, that were 150 m long and perpendicular to the mangrove upland interface (6 cores along each transect) at Quang Ninh (4 transects), Thai Binh (5), Nam Dinh (4) and Thanh Hoa (3) in May and June of 2018. The sampling sites and transects are described in Figure 1 and Table 1. Five-cm-long soil samples were then collected at the 0-15-cm, 15-30-cm, 30-50cm, 50-100-cm, and >100-cm intervals [29]. All wet soil samples were dried to a constant mass at 60 °C to determine the bulk density (BD) for each interval. The dried samples were ground by hand using a mortar and pestle, then sieved through a 2-mm mesh to remove any large roots or rocks before being analyzed for C content (%C). Bulk density was obtained by the ratio of dry mass per volume of wet sample. Due to limited resources, C content in all soil samples was determined using the Walkley-Black method [30]. Carbon content of 50 randomly selected soil samples was also analyzed using the combustion method on a Thermo Scientific™ FlashSmart™ Elemental Analyzer (EA) to calculate correction coefficients for C contents from the Walkley-Black method (WB). The regression relationship of WB with EA is y = 0.667x + 0.361 (R 2 = 0.841, p < 0.001).   Soil C stocks were then determined for each soil core interval by multiplying the interval length by the BD of the subsample and the whole %C value of the subsample. Intervals were then summed up for the entire 2-m-long core from each plot and then averaged across mangrove type, mangrove species, and province. One-way analysis of variance (ANOVA) was used to compare C stocks between mangrove type (intact and restored mangroves) as well as across mangrove species (mixed species, S. caseolaris, A. corniculatum and K. obovata) and the different provinces (Quang Ninh, Thai Binh, Nam Dinh and Thanh Hoa). Fisher's least significant difference (LSD) test was used to perform all pairwise comparisons among group means. All values reported are mean values ± standard error (SE).

Soil Carbon in Intact vs. Restored Mangroves
The average soil BD of intact mangrove forests in northern Vietnam (0.91 ± 0.04 g/cm 3 ) was significantly smaller (p < 0.05) than that of restored mangrove forests (1.03 ± 0.03 g/cm 3 ). In contrast, the average %C and soil carbon density of intact mangrove forests (2.43 ± 0.47% and 19.89 ± 2.92 mg/cm 3 , respectively) were significantly greater (p < 0.05) than that of restored mangrove forests (1.46 ± 0.13% and 13.34 ± 1.31 mg/cm 3 , respectively). However, in the upper 100 cm, the BD and %C of the intact mangrove forests were significantly different from the restored mangrove forest (p < 0.05). Below this depth (>100 cm), the difference was not significant (p > 0.05) ( Figure 2). The results suggest that both intact mangroves and restored mangroves originally had equivalent sedimentary substrate. This is also revealed by our field observation that the bottom layers of both mangrove types were mostly composed of sand. Similar to %C, soil carbon density in the upper 100 cm depth in the intact mangrove forests was significantly greater than in restored mangrove forests (p < 0.05). Below this depth (>100 cm), the difference was not significant (p > 0.05). Soil carbon density in both mangrove types increased from 15.42 to 27.20 mg/cm 3 between 0 and 15 cm depth, then declined rapidly (27.20-10.69 mg/cm 3 ) in the middle layers (15-100 cm) and slowly at the bottom of the core (10.69-8.63 mg/cm 3 ). The differences in BD, %C and soil carbon density in the upper layers may result from the different stand ages and development of mangrove roots in the upper 50 cm. Ha et al. [22] reported a linear relationship between carbon stocks in the soil and in the roots and suggested that the roots are an important source of organic carbon accumulation in mangrove soils. Moreover, Komiyama et al. [31], Tamooh et al. [32] and Castañeda-Moya et al. [33] demonstrated that mangrove root densities as well as their carbon content not only varied with depth but also with mangrove stand age.
Forests 2020, 11, x FOR PEER REVIEW 5 of 10 al. [31], Tamooh et al. [32] and Castañeda-Moya et al. [33] demonstrated that mangrove root densities as well as their carbon content not only varied with depth but also with mangrove stand age.  Table 2 shows that the average soil C stocks between mangrove types at the depth intervals of 0-15 cm, 15-30 cm, and >100 cm were similar (p > 0.05), but significantly different at the depth intervals of 30-50 cm and 50-100 cm (p < 0.05). The total soil C stock of intact mangrove forest (300.68 ± 51.61 Mg/ha) was not significantly different from that of restored mangroves (217.74 ± 16.82 Mg/ha) (p > 0.05). Gao et al. [34] reported that forest age could be one of the key factors affecting soil C stock. Donato et al. [1] and Lovelock et al. [35] also reported that mature mangrove plants could contribute much more soil C stock and biomass than younger plants. However, Lunstrum and Chen [36] found no significant variation in soil C stock between S. apetala and K. obovata after six years of planting. Our study indicated that 20-25-year-old restored mangrove forest could have similar levels of soil C stock as intact mangrove forest.

Soil Carbon among Different Species
In northern Vietnam, average soil BD was the highest in S. caseolaris (1.07 ± 0.07 g/cm 3 ), followed by K. obovata (1.03 ± 0.04 g/cm 3 ), A. corniculatum (0.95 ± 0.04 g/cm 3 ) and mixed-species mangrove forest (0.90 ± 0.05 g/cm 3 ). In contrast, average %C and soil carbon density were the highest in mixed-species mangrove forest (2.33 ± 0.39 g/cm 3 and 18.76 ± 3.26 mg/cm 3 , respectively), followed by A. corniculatum (1.84 ± 0.53 g/cm 3 and 16.65 ± 3.34 mg/cm 3 , respectively), S. caseolaris (1.51 ± 0.15 g/cm 3 and 14.85 ± 1.88 mg/cm 3 , respectively) and K. obovata (1.28 ± 0.18 g/cm 3 and 11.42 ± 1.81 mg/cm 3 , respectively). The soil BD of all species increased with depth, while %C and carbon density decreased with depth ( Figure 3). Among different species, there was no significant difference of soil BD, %C and carbon density in the upper 100 cm (p > 0.05). However, below this depth, the differences were statistically  Table 2 shows that the average soil C stocks between mangrove types at the depth intervals of 0-15 cm, 15-30 cm, and >100 cm were similar (p > 0.05), but significantly different at the depth intervals of 30-50 cm and 50-100 cm (p < 0.05). The total soil C stock of intact mangrove forest (300.68 ± 51.61 Mg/ha) was not significantly different from that of restored mangroves (217.74 ± 16.82 Mg/ha) (p > 0.05). Gao et al. [34] reported that forest age could be one of the key factors affecting soil C stock. Donato et al. [1] and Lovelock et al. [35] also reported that mature mangrove plants could contribute much more soil C stock and biomass than younger plants. However, Lunstrum and Chen [36] found no significant variation in soil C stock between S. apetala and K. obovata after six years of planting. Our study indicated that 20-25-year-old restored mangrove forest could have similar levels of soil C stock as intact mangrove forest.
Forests 2020, 11, x FOR PEER REVIEW 6 of 10 significant (p < 0.05), in which soil BD, %C and carbon density of K. obovata mangroves were significantly lower than for other species or mixed mangroves (p < 0.05). Hien et al. [17] suggested that, in K. obovata mangroves, root biomass may be an important source of soil organic carbon enrichment, but the contribution of mangrove products in deeper layers than 100 cm was negligible; meanwhile, Marchand et al. [37,38] also reported that, in some mangroves, the root system was considered to be the main contributor of the accumulation of organic matters in mangrove soils. However, Alongi [12], Bouillon et al. [39], Barr et al. [40] and Li et al. [41] suggested that soil organic carbon in mangrove forests mainly originates from the decomposition of mangrove litter or from adjacent coastal waves, tides and rivers. The species with different root system would also impact soil organic carbon. Table 3 shows that average soil C stocks among species at depth intervals of 0-15 cm, 15-30 cm, and 50-100 cm were similar (p > 0.05), but significantly differed at deeper soil depth intervals of 30-50 cm and >100 cm (p < 0.05). In northern Vietnam, mixed and A. corniculatum mangrove stands have a longer history of establishment compared to other species and also have pneumatophores (i.e., vertical roots arising from shallow, horizontal roots) and thus a more complicated tertiary structure. This could explain the higher C stocks in the upper intervals, which would be influenced by the higher trapping efficiencies of these root systems and thus greater deposition and sedimentation rate but lower erosion rate compared to the younger S. caseolaris and K. obovata mangroves [25]. Table 3 also shows a significant difference in total soil C stocks among species (p < 0.05). Fisher's LSD analysis indicated that the differences in total soil C stock among mixed, S. caseolaris and A. corniculatum mangroves (289.75 ± 33.28 Mg/ha, 255.67 ± 13.11 Mg/ha and 278.15 ± 43.86 Mg/ha, respectively) were not significant (p > 0.05). However, the total soil C stock of K. obovata (174.04 ± 20.38 Mg/ha) was significantly lower than that of other species (p < 0.05). Gao et al. [34] commented that mangrove forests with different species would have different capabilities of soil carbon sequestration and different contributions to carbon stock. Ha et al. [22], Komiyama et al. [31], Tamooh et al. [32] and Castañeda-Moya et al. [33] also demonstrated that mangrove root density as well as soil C stock not only varied with depth and stand ages but also with mangrove species.  Hien et al. [17] suggested that, in K. obovata mangroves, root biomass may be an important source of soil organic carbon enrichment, but the contribution of mangrove products in deeper layers than 100 cm was negligible; meanwhile, Marchand et al. [37,38] also reported that, in some mangroves, the root system was considered to be the main contributor of the accumulation of organic matters in mangrove soils. However, Alongi [12], Bouillon et al. [39], Barr et al. [40] and Li et al. [41] suggested that soil organic carbon in mangrove forests mainly originates from the decomposition of mangrove litter or from adjacent coastal waves, tides and rivers. The species with different root system would also impact soil organic carbon. Table 3 shows that average soil C stocks among species at depth intervals of 0-15 cm, 15-30 cm, and 50-100 cm were similar (p > 0.05), but significantly differed at deeper soil depth intervals of 30-50 cm and >100 cm (p < 0.05). In northern Vietnam, mixed and A. corniculatum mangrove stands have a longer history of establishment compared to other species and also have pneumatophores (i.e., vertical roots arising from shallow, horizontal roots) and thus a more complicated tertiary structure. This could explain the higher C stocks in the upper intervals, which would be influenced by the higher trapping efficiencies of these root systems and thus greater deposition and sedimentation rate but lower erosion rate compared to the younger S. caseolaris and K. obovata mangroves [25]. Table 3 also shows a significant difference in total soil C stocks among species (p < 0.05). Fisher's LSD analysis indicated that the differences in total soil C stock among mixed, S. caseolaris and A. corniculatum mangroves (289.75 ± 33.28 Mg/ha, 255.67 ± 13.11 Mg/ha and 278.15 ± 43.86 Mg/ha, respectively) were not significant (p > 0.05). However, the total soil C stock of K. obovata (174.04 ± 20.38 Mg/ha) was significantly lower than that of other species (p < 0.05). Gao et al. [34] commented that mangrove forests with different species would have different capabilities of soil carbon sequestration and different contributions to carbon stock. Ha et al. [22], Komiyama et al. [31], Tamooh et al. [32] and Castañeda-Moya et al. [33] also demonstrated that mangrove root density as well as soil C stock not only varied with depth and stand ages but also with mangrove species.

Soil Carbon among Provinces
The BD in mangrove soil was the highest in Thanh Hoa, followed by Thai Binh, Nam Dinh, and then Quang Ninh, with mean values of 1.12 ± 0.05, 1.01 ± 0.05, 0.97± 0.02, and 0.89 ± 0.04 g/cm 3 , respectively. In all provinces, BD increased with depth. However, among the provinces, there was a significant difference in BD in the upper 50 cm (p < 0.05); below this depth, there was no significant difference (p > 0.05). Fisher's LSD analysis indicated that the differences in BD in the upper 50 cm across sites were driven by differences between Thanh Hoa and Quang Ninh (Figure 4).

Soil Carbon among Provinces
The BD in mangrove soil was the highest in Thanh Hoa, followed by Thai Binh, Nam Dinh, and then Quang Ninh, with mean values of 1.12 ± 0.05, 1.01 ± 0.05, 0.97± 0.02, and 0.89 ± 0.04 g/cm 3 , respectively. In all provinces, BD increased with depth. However, among the provinces, there was a significant difference in BD in the upper 50 cm (p < 0.05); below this depth, there was no significant difference (p > 0.05). Fisher's LSD analysis indicated that the differences in BD in the upper 50 cm across sites were driven by differences between Thanh Hoa and Quang Ninh (Figure 4). In contrast to the BD, soil %C in all sites decreased with depth ( Figure 4). The mean value of %C in Quang Ninh (2.89 ± 0.30%) was almost twice the %C values in remaining provinces: Thanh Hoa (1.20 ± 0.09%), Thai Binh (1.29 ± 0.15%) and Nam Dinh (1.65 ± 0.23%). Mangrove soil carbon density in the four provinces increased from 12.17 ± 2.38 to 31.78 ± 3.04 mg/cm 3 between 0 and 30 cm depth, then declined rapidly to 9.03 ± 2.13 mg/cm 3 in the middle layers (30-100 cm), and declined slowly to 8.33 ± 1.59 mg/cm 3 at the bottom of the core. In the upper 50 cm, significant differences in carbon density among the different provinces were also observed (p < 0.05). Below this depth, the differences were not significant ( Figure 4). The differences in BD, %C and soil carbon density were likely due to the different environmental conditions of the sites previously described. Quang Ninh mangroves are protected by nearshore islands and thus not affected by strong waves. As a result, less organic matter washes away. The BD, %C and soil carbon density from this study were close to those measured in Xuan Thuy National Park of Vietnam [17] and in other Asian mangroves, such as in China [36,42] and Thailand [43], but lower than those measured in the Indo-Pacific [1].
There was a significant difference in the total soil C stocks among provinces (p < 0.05). However, Fisher's LSD analysis indicated that the total soil C stocks of Quang Ninh (323.89 ± 28.43 Mg/ha) were not significantly different from Nam Dinh (249.81 ± 19.09 Mg/ha), but were significantly greater than Thai Binh (201.42 ± 27.65 Mg/ha) and Thanh Hoa (178.98 ± 30.82 Mg/ha). Quang Ninh also had significantly greater soil C stocks than the other three provinces in almost soil layers, while soil C stocks did not differ among the remaining three provinces at each depth interval (Table 4). This is probably because of their different geomorphological characteristics, as described in the previous section. Quang Ninh and Nam Dinh mangroves have also developed over a longer period of time In contrast to the BD, soil %C in all sites decreased with depth ( Figure 4). The mean value of %C in Quang Ninh (2.89 ± 0.30%) was almost twice the %C values in remaining provinces: Thanh Hoa (1.20 ± 0.09%), Thai Binh (1.29 ± 0.15%) and Nam Dinh (1.65 ± 0.23%). Mangrove soil carbon density in the four provinces increased from 12.17 ± 2.38 to 31.78 ± 3.04 mg/cm 3 between 0 and 30 cm depth, then declined rapidly to 9.03 ± 2.13 mg/cm 3 in the middle layers (30-100 cm), and declined slowly to 8.33 ± 1.59 mg/cm 3 at the bottom of the core. In the upper 50 cm, significant differences in carbon density among the different provinces were also observed (p < 0.05). Below this depth, the differences were not significant (Figure 4). The differences in BD, %C and soil carbon density were likely due to the different environmental conditions of the sites previously described. Quang Ninh mangroves are protected by nearshore islands and thus not affected by strong waves. As a result, less organic matter washes away. The BD, %C and soil carbon density from this study were close to those measured in Xuan Thuy National Park of Vietnam [17] and in other Asian mangroves, such as in China [36,42] and Thailand [43], but lower than those measured in the Indo-Pacific [1].
There was a significant difference in the total soil C stocks among provinces (p < 0.05). However, Fisher's LSD analysis indicated that the total soil C stocks of Quang Ninh (323.89 ± 28.43 Mg/ha) were not significantly different from Nam Dinh (249.81 ± 19.09 Mg/ha), but were significantly greater than Thai Binh (201.42 ± 27.65 Mg/ha) and Thanh Hoa (178.98 ± 30.82 Mg/ha). Quang Ninh also had significantly greater soil C stocks than the other three provinces in almost soil layers, while soil C stocks did not differ among the remaining three provinces at each depth interval (Table 4). This is probably because of their different geomorphological characteristics, as described in the previous section. Quang Ninh and Nam Dinh mangroves have also developed over a longer period of time compared to the other two provinces [25]. Moreover, the differences in the bioturbation, climatic and physicochemical factors among provinces could also influence the soil carbon C stock [2,34,44,45]. Table 4. Soil carbon pools (mean ± SE) (Mg/ha) according to provinces in the study area. The letters in each row denote the differences among provinces.

Conclusions
The results of the present study show that the soil C stock of mangrove forests in northern Vietnam varies depending on geographical location and mangrove species. The highest soil C stock was estimated for Quang Ninh, where intact mangroves have grown steadily for a long time, while the lowest value was at Thanh Hoa, where restored mangroves have been developed since 1995-1998. Soil C stocks of S. caseolaris, A. corniculatum and mixed species were significantly larger than that of K. obovata. The study also confirmed that soil C stock of 20-25-year-old restored mangrove forest was not significantly different from that of intact mangrove forest. This demonstrated that mangrove restoration can help to sequester blue carbon, which is a crucial aspect to mitigating climate change in northern Vietnam.