00051000977 Analysis of blue carbon sequestration potential in mangrove ecosystems A case study in Tien Lang, Hai Phong
The necessity of research
Mangrove ecosystems are vital for both environmental and socio-economic functions in coastal regions, acting as natural barriers against storms, preventing coastal erosion, supporting biodiversity, and providing nursery grounds for aquatic species (Alongi, 2002) They are among the most efficient carbon sequestration systems, storing carbon in both above-ground biomass and soil, with Blue Carbon playing a critical role in mitigating climate change (Donato et al., 2011).
However, mangrove degradation is a growing global issue, with over 35% of mangrove cover lost in recent decades due to human activities like urbanisation, aquaculture, and land development (Alongi, 2012; FAO, 2007) Vietnam, particularly, has seen a significant decline in mangrove areas due to shrimp farming, land reclamation, and industrial expansion, which threaten both the ecosystem services provided by mangroves and their carbon sequestration potential (Nam et al., 2016; Tue et al., 2016) Moreover, human activities such as aquaculture, particularly shrimp farming, have been a major driver of mangrove loss in Vietnam Land-use change not only reduces mangrove cover but also releases previously stored carbon into the atmosphere, thus intensifying greenhouse gas emissions (Nguyen et al., 2021; Tong et al., 2004)
Blue Carbon refers to the carbon sequestered by coastal and marine ecosystems, including mangroves, salt marshes, and seagrass meadows These habitats are highly effective at capturing and storing atmospheric CO2 due to the anaerobic conditions in waterlogged sediments that slow the decomposition of organic matter This enables long-term storage of carbon, making Blue Carbon a vital natural solution for climate change mitigation (Mcleod, 2011; Nellemann, 2009).
Vietnam's mangroves play a critical role in carbon sequestration, with focused studies in the Mekong Delta and Red River Delta over the past two decades These studies highlight the carbon storage potential of mangrove ecosystems and their vulnerability to increasing anthropogenic pressures such as aquaculture development and land reclamation (Nguyen et al., 2021) However, most research concentrates on natural mangrove ecosystems, leaving a research gap regarding planted mangrove forests, which are a key feature of the Tien Lang study site.
Tong et al., 2004 emphasizes that although planted mangroves initially contribute less to carbon sequestration compared to mature natural stands, they can accumulate carbon efficiently as they mature This observation is supported by case studies in Ca Mau and Can Gio, where restoration and afforestation efforts have successfully enhanced local Blue Carbon stocks
Although substantial progress has been made in understanding Blue Carbon in natural mangroves, research on planted mangrove forests remains limited, particularly in northern Vietnam The Tien Lang mangrove ecosystem is largely composed of plantation-based restoration, offering a unique opportunity to evaluate the carbon potential of managed reforestation projects and their contribution to carbon sequestration and ecosystem services.
In northern and central Vietnam, Tue et al., 2014 reported ecosystem carbon values ranging between 150–450 tC/ha in Kandelia obovata plantations, highlighting how monoculture replanting efforts can lead to substantial but variable carbon stocks Quang Tran et al., 2018 also observed comparable values in mixed mangrove systems across northern provinces, suggesting that species diversity may enhance carbon accumulation, particularly in early plantation stages
Comparative studies from outside Vietnam provide useful contrast For example, Meng et al., 2021 reported a broader range (150–538.7 tC/ha) in southern China’s
Bruguiera sexangula mangrove systems demonstrate that geomorphic variability and site disturbance are major drivers of carbon stock variability A global synthesis of natural mangrove forests reports exceptional carbon stocks, reaching up to 1023 tC/ha in some cases, especially in undisturbed and long-established stands in equatorial zones (Donato et al., 2011).
These comparisons reveal two key implications for mangrove carbon management: first, planted mangroves contribute to carbon storage but may take decades to reach the storage potential of mature natural mangrove ecosystems; second, local environmental conditions—including sediment type, tidal flushing, and nutrient dynamics—significantly influence soil organic carbon (SOC) depth profiles and overall carbon retention in mangrove systems.
At Tien Lang, the co-occurrence of functional mangrove species—Sonneratia caseolaris, Aegiceras corniculatum, and Kandelia obovata—supports steady biomass growth and progressive carbon sequestration, aligning with He et al (2020), who demonstrated that species mixing in plantation systems enhances carbon performance and ecological resilience, especially under climate stress.
While extensive research exists on natural mangrove ecosystems in the Mekong and Red River Deltas, empirical estimates of the carbon potential of planted mangrove forests in northern Vietnam remain scarce Planted mangroves exhibit a distinct carbon recovery trajectory, shaped by species selection, reforestation techniques, and ongoing land-use conflicts Framing the analysis within regional (Nguyen et al., 2017) and international (Donato et al., 2011; Meng et al., 2021) frameworks, this study aims to illuminate how reforested mangroves contribute to long-term carbon sequestration.
Figure 1.1 Land cover map for the year 2010 in Hai Phong (Pham et al., 2017)
This study assesses the carbon sequestration capacity of mangrove forests in Tien Lang District, Hai Phong, a region experiencing significant environmental pressure from expanding aquaculture and changing land use Unlike many studies that focus on natural mangrove ecosystems, this research highlights the potential of planted mangroves, which are more prevalent in the area.
These human-initiated forests offer an important opportunity to evaluate how reforestation and restoration efforts contribute to blue carbon storage, especially during the early stages of ecosystem recovery
This assessment measures above-ground carbon (AGC) and soil organic carbon (SOC) using an integrated approach that combines field data with remote sensing techniques By analyzing satellite imagery from 1996, 2020, and 2024, the study traces the spatial development of mangrove coverage over time and links these changes to local environmental factors and human activities The integration of ecological data with geospatial analysis enables a more comprehensive understanding of how such factors influence carbon accumulation in mangrove ecosystems The results are intended to support improved mangrove management strategies, guide restoration planning, and contribute to Vietnam’s national climate change mitigation goals.
Review of methods used in Calculation of Blue-Carbon Sequestration
Accurately estimating above-ground biomass (AGB) in mangrove ecosystems is essential for assessing blue carbon storage Over the past two decades, numerous allometric equations have been developed to estimate AGB based on tree traits such as diameter at breast height (DBH), tree height, and crown diameter, but their accuracy varies with species morphology, geographic region, and whether forests are planted or natural To address this, a literature review of established mangrove AGB models was conducted and those best suited to the characteristics of planted mangroves in Tien Lang, Hai Phong, were selected The chosen equations offer scientific credibility, align with IPCC guidelines, and reflect local field conditions, including species structure and measurement feasibility.
This study focused on three dominant mangrove species in the area, Sonneratia caseolaris, Aegiceras corniculatum, and Kandelia obovata For each, relevant allometric models were reviewed and applied as follows
Several models have been proposed for Sonneratia caseolaris to reflect its wide distribution across Southeast Asia (Komiyama et al., 2005) The early Komiyama equation using DBH and height has been widely cited, but more recent work in Vietnam employs remote sensing–integrated approaches For instance, Tran et al (2017) applied ALOS-2 PALSAR data with neural networks to estimate biomass in Hai Phong with high accuracy, while Quang et al (2022) compared multiple machine learning methods for biomass estimation, highlighting the potential of integrating satellite imagery with field measurements In the present study, we selected the Komiyama et al (2005) equation for Sonneratia caseolaris due to its simplicity, consistency with IPCC methods, and direct applicability to field measurements of DBH and height; AGB is estimated using the formula: AGB = 0.0673 × (0.54)^0.976 × (DBH)^2.004 [Eq.1] (Komiyama et al., 2005), where 0.54 is the wood density.
For Aegiceras corniculatum (Su), presents structural challenges due to its often shrub-like form and irregular trunk geometry Conventional DBH-based models tend to under- or over-estimate their biomass To address this, (Fu & Wu, 2011) developed a crown diameter and height-based equation for Aegiceras corniculatum and similar mangrove species, AGB= 3.1253x(((Crown Diameter) 2 ) x Height) 0.9063 [Eq.2] This study chose this equation due to its greater sensitivity to the growth form of Aegiceras corniculatum in Vietnam’s northern deltas It allows for more accurate field-based estimates, especially in dense, low-stature stands
Kandelia obovata stands as one of the most widely planted mangrove species in northern Vietnam due to its strong adaptability and ecological importance Studies have used diverse allometric approaches to quantify its biomass accumulation A well-known example is the DBH-based model proposed by Komiyama et al (2005), which has gained broad application in tropical mangrove biomass assessments because of its generality Nevertheless, this model may not fully capture the structural traits of planted mangroves in northern Vietnam, particularly in young stands characterized by low canopy closure and irregular trunk forms.
In contrast, (Tue et al., 2016) developed site-specific allometric equations tailored for K obovata plantations in the Red River Delta, using diameter measured at
30 cm above the base of the trunk (D₃₀) This choice of measurement height accounts for the basal flaring and low stem bifurcation that are typical in northern Vietnamese mangrove stands, which often make DBH measurements less reliable or inconsistent in young plantations Their model, derived from destructive sampling of 101 trees, produced the following equation for Above Ground Biomass (AGB)=0.04975×𝐷 1.94748 [Eq.3], where D is the stem diameter (cm) measured at 30 cm above the widening base These models were specifically calibrated for K obovata in Vietnamese coastal conditions and were preferred over generalized models due to their higher relevance to local plantation characteristics
After biomass estimation, conversion to Above-Ground Carbon (AGC) was carried out by applying a standard carbon fraction of 0.47 as recommended by the (IPCC, 2006)
The Below-Ground Biomass (BGB) was also species-specific: For Sonneratia caseolaris, BGB was estimated using BGB = 0.199 × (DBH) 2.22 [Eq.4] (Komiyama et al., 2005) For Kandelia obovata, BGB was calculated using BGB = 0.01420 × D 2.12146 [Eq.5] (Nguyen et al., 2016) For Aegiceras corniculatum (Su), BGB was estimated by applying a root-to-shoot ratio of 0.24 multiplied by the species' AGB, BGB = AGB x 0.24 [Eq.6] (Fu & Wu, 2011)
Subsequently, BGB was converted to Below-Ground Carbon (BGC) using a carbon fraction of 0.47 This method yields species-specific carbon stock estimates that are aligned with international mangrove carbon accounting standards (Howard et al., 2014; IPCC, 2006).
Loss on Ignition (LOI) is the widely applied method in Vietnam to assess Soil Organic Carbon (SOC) The approach first determines Organic Matter (OM) from combusted samples and then converts OM to SOC using the Van Bemmelen factor of 1.724, expressed as SOC (%) = OM (%) / 1.724 Vietnamese national guidelines, including TCVN 6647:2007 and TCVN 8941:2011, provide standardized procedures for LOI and Walkley-Black techniques in local soil analysis.
Remote sensing has become a cornerstone of mangrove monitoring, enabling timely assessment of forest health and land-use change Tools such as Landsat 8/9 OLI-TIRS and Sentinel-2A underpin mangrove classification and NDVI analysis, providing insights into vegetation density and trends in deforestation or afforestation over time The Normalised Difference Vegetation Index (NDVI) is a widely used metric for tracking mangrove vigor and changes in cover In the Vietnamese context, studies report high accuracy for supervised classification that combines NDVI with Landsat imagery to detect mangrove dynamics, with notable work by Nguyen, Hardy, et al (2021) and Pham et al (2017).
In Vietnam, mangrove loss is primarily driven by human activities, especially aquaculture and shrimp farming Land-use change not only reduces mangrove cover but also releases stored carbon into the atmosphere, intensifying greenhouse gas emissions (Tong et al., 2004; Nguyen et al., 2021).
Time-series analysis of satellite data from 1996 through 2020 and including 2024 enables the detection of patterns and the quantification of carbon losses or gains resulting from these changes This methodology aligns with the IPCC Good Practice Guidance for LULUCF (IPCC, 2003), supporting robust, policy-relevant assessments of land-use carbon dynamics.
Blue Carbon sequestration was estimated using standardized, cross-validated methodologies, with Above-Ground Carbon (AGC) and Below-Ground Carbon (BGC) calculated through species-specific allometric equations based on diameter at breast height (DBH), tree height, and crown diameter For Sonneratia caseolaris and Kandelia obovata, DBH- and height-based models were applied, while Aegiceras corniculatum required crown diameter and height-based models to better reflect its growth form Biomass estimates were then converted to carbon stocks using a carbon fraction of 0.47 following IPCC (2006) guidelines Soil Organic Carbon (SOC) stock was determined using Loss on Ignition (LOI) and Walkley-Black methods in accordance with national standards (TCVN 8941:2011), with SOC stock computed from bulk density, TOC (%) and depth intervals based on formulas recommended by Howard (2014) and Kauffman & Donato (2012).
Integrating remote sensing tools such as Landsat and Sentinel imagery with supervised classification and NDVI analysis allowed for temporal assessments of mangrove area change Time-series analyses following (IPCC, 2003) LULUCF guidelines were used to detect carbon stock changes from 1996 to 2024
This study addresses a gap in mangrove carbon research by estimating above-ground carbon (AGC), below-ground carbon (BGC), and soil organic carbon (SOC) in planted mangroves using updated field measurements and standardized calculation methods It also integrates geographic information system (GIS) analysis with historical satellite data to enable robust temporal change detection of mangrove areas and carbon stocks Based on these analyses, the study offers evidence-based recommendations for restoration and management policies to enhance carbon sequestration, ecosystem resilience, and informed decision-making in coastal conservation.
Overview of the research site
Tien Lang District, located in Hai Phong City in northern Vietnam, forms a critical coastal buffer along the Red River Delta The region faces intensive land-use competition between aquaculture, agriculture, and mangrove restoration projects Over the past two decades, significant efforts have been made to rehabilitate mangrove cover through plantation programs.
Sonneratia caseolaris, Kandelia obovata, and Aegiceras corniculatum
The mangrove forests of Tien Lang in Hai Phong are situated in a humid subtropical climate (Köppen Cwa) with distinct wet and dry seasons, where the area averages about 23.7°C annually and winter temperatures range from 12°C to 20°C while summer highs reach up to 35°C Annual rainfall is 1,800–1,900 mm, with nearly 90% of precipitation falling in the May–October wet season and a peak in August of about 240 mm; during the dry season (November–April), monthly rainfall drops below 60 mm, often increasing soil salinity and reducing freshwater input The region also receives roughly 2,987 hours of sunshine per year, supporting vigorous mangrove growth under favorable conditions These climatic variables—particularly seasonal rainfall and temperature—critically regulate mangrove biomass productivity and soil organic carbon accumulation, thereby shaping the long-term blue carbon sequestration potential of the Tien Lang mangrove ecosystem.
The region experiences semi-diurnal tides, brackish estuarine conditions, and seasonal freshwater input, which influence mangrove growth dynamics Due to its transitional ecological setting and human-altered landscape, Tien Lang offers a unique opportunity to assess carbon storage potential in reforested mangrove systems Field data collection was conducted across 15 representative plots to cover varying species
Comparative Study
To situate this study within the broader scientific context, the following table summarizes key findings from previous research on blue carbon storage in both planted and natural mangrove systems, particularly focusing on Above-Ground Carbon (AGC), Below-Ground Carbon (BGC), and Soil Organic Carbon (SOC) The table highlights regional and global variability and establishes a comparative framework for assessing the carbon dynamics in Tien Lang
200–1,023 50–200 39–450 (Alongi, 2014; Donato et al., 2011; Tue et al., 2014)
(He et al., 2020; Nguyen et al., 2021)
High SOC depth and density
Lower SOC, AGC recovery slow
(He et al., 2020; Quang Tran et al., 2018)
The table highlights the diverse carbon storage potential of mangroves, underscoring their significance in carbon sequestration It also provides a basis for evaluating the performance of planted forests in Tien Lang, benchmarked against national and global benchmarks to guide policy and conservation efforts.
Research questions, hypotheses, objectives, tasks
Table 1.2 Research questions and hypotheses
Q1: What is the current amount of
Blue Carbon stock in the planted mangrove forests of Tien Lang?
H1: The planted mangrove forests in Tien
Lang contain substantial Blue Carbon stocks, contributing significantly to climate change mitigation
Q2: How have mangrove land cover changes from 1996 to 2024 influenced
Blue Carbon stock, and to what extent has recent restoration enhanced carbon recovery?
H2: Historical land-use change (1996–2008) caused around 70 ha loss in mangrove area and a approximately 13% decline in carbon density Restoration efforts post-2010 increased area by ~100 ha and boosted carbon density from 70 to 85 tC/ha, with species composition—especially Aegiceras corniculatum and Kandelia obovata— driving recovery rates
Based on the research questions and hypotheses, the research objectives and the sub-division tasks are presented in the following table
Table 1.3 Research objectives and tasks
O1: Evaluate the Blue Carbon potential of the planted mangrove forests in Tien Lang using ground data and remote sensing
Task 1: Conduct field sampling in 15 plots to measure Above-Ground Carbon (AGC) and Soil Organic Carbon (SOC) using standardized methods
O2: Analyze carbon sequestration rates by estimating AGC and SOC across multiple species and soil layers
Task 2: Use allometric equations and LOI/Walkley-Black methods to estimate biomass carbon and soil carbon; calculate Total Organic Carbon
O3: Assess mangrove land cover change impacts on Blue Carbon from
1996 to 2024 using satellite imagery and GIS
Task 3: Utilize Landsat (1996, 2020) and Sentinel/OLI data (2024) to classify mangrove cover, clean data (cloud masking), compute NDVI, and map land-use changes
O4: Produce spatial maps of carbon stock and accumulation across the study area
Task 4: Interpolate AGC and SOC values across the site using GIS tools; generate accumulated carbon and classification maps.
Scope of the research
Geographical Scope
This study was conducted in Tien Lang District, Hai Phong City, Vietnam, in the Red River Delta coastal region, where large-scale planted mangrove ecosystems have been established to curb coastal erosion and mitigate climate-change impacts; despite these benefits, mangroves face growing pressure from aquaculture expansion, urbanization, and infrastructure development The research focuses on 15 field plots within the planted mangrove zones, selected to represent a range of mangrove conditions, ages, and species compositions Dominant species examined include Sonneratia caseolaris, Kandelia obovata, and Aegiceras corniculatum The plots span different locations within the reforested and restored zones to enable assessment of spatial heterogeneity in carbon stock across Tien Lang.
Scientific Scope
The study focuses on evaluating the Blue Carbon potential of the planted mangrove forests by integrating: a Above-Ground Carbon (AGC) Estimation
Tree measurements, including Diameter at Breast Height (DBH), Crown Diameter, and total Height, were used with species-specific allometric equations to estimate Above-Ground Biomass (AGB) Above-Ground Carbon (AGC) was calculated as 47% of AGB in accordance with IPCC conversion guidelines In addition, Below-Ground Carbon for Roots (BGC) was estimated to provide a complete assessment of the forest carbon stocks.
Below-Ground Biomass (BGB) was estimated using species-specific allometric equations based on tree diameter and, where available, supplementary parameters such as tree height or crown diameter For each species, established models were applied to predict root biomass without relying solely on generalized root-to-shoot ratios BGC was calculated by converting BGB to carbon stock, assuming a carbon content of 47%, in line with IPCC (2006) guidelines for forest carbon reporting The assessment also includes Soil Organic Carbon (SOC) estimation.
Soil samples were collected at 20 mm intervals in each plot and analyzed for organic matter by the Loss on Ignition (LOI) method and for Total Organic Carbon (TOC) by the Walkley-Black method (TCVN 8941:2011) Bulk density (BD) was estimated from empirical equations reported in Vietnamese literature and field data, using BD = 1.539 × e^(−0.289 × TOC (%)) [Eq 8] (Tue et al., 2016).
TOC was obtained by summing AGC (in tC/ha), BGC (in tC/ha), and SOC (in tC/ha) for each plot e Remote Sensing and GIS Analysis
Satellite imagery analysis complemented field-based carbon estimations Landsat 8–9 (OLI/TIRS) surface reflectance bands—Red (Band 4) and NIR (Band 5)—were used to calculate the Normalized Difference Vegetation Index (NDVI) for vegetation assessment, providing a robust proxy to support interpretations of vegetation cover and potential carbon stocks.
2024 Cloud and shadow masking were applied as pre-processing steps to ensure the quality and reliability of the data
Long-term mangrove extent changes were assessed using Global Mangrove Watch (GMW) datasets for the years 1996 and 2020, allowing the detection of mangrove gain or loss over a ~30-year period
Additionally, NDVI values were utilized to estimate Above-Ground Biomass (AGB) following a regression model developed by (Quang Tran et al., 2018) for
Sonneratia caseolaris plantations in Vinh Quang Commune, Hai Phong The biomass estimation formula applied was: Biomass (t/ha) = 378.81×NDVI 1.0191 [Eq.9]
This approach enabled the integration of field-measured carbon stock data with remote sensing outputs, enhancing the spatial extrapolation of biomass and carbon distribution across the study area Finally, NDVI maps, classified mangrove cover maps, and interpolated carbon stock distribution maps were produced to visualize vegetation health, land cover changes, and blue carbon accumulation patterns over time.
Temporal Scope
Spanning 1996–2024, this study captures nearly three decades of changes in the mangrove landscape It integrates remote-sensing data from the Global Mangrove Watch (GMW) for 1996 and 2020 and Landsat 8–9 OLI/TIRS Level-2 Surface Reflectance (2024) to perform a time-series analysis of land-cover change and mangrove area dynamics In December 2024, field data on Above-Ground Carbon (AGC) and Soil Organic Carbon (SOC) were collected, with laboratory analyses conducted soon after to quantify current blue carbon storage.
This temporal range provides insight into the impact of human-induced land-use change and the effectiveness of mangrove replanting and restoration efforts over time.
Methodological Scope
This study uses an interdisciplinary and integrative methodology that combines, field methods, Laboratory Analysis, Carbon Calculation Models, and GIS and Remote Sensing Analysis
Field data collection involved direct measurement of key mangrove tree parameters, including diameter at breast height (DBH), tree height, and crown diameter across all 15 sampling plots Additionally, soil core samples were collected at each plot to evaluate below-ground carbon storage All field sampling procedures followed the Vietnamese National Standards (TCVN) to ensure consistency and methodological rigor
Laboratory analyses quantified soil organic matter and total organic carbon (TOC) in the samples Organic matter was determined by the Loss on Ignition (LOI) method, while TOC was measured using the Walkley-Black titration method Bulk density was calculated using both empirical models and direct measurements of the sample dry weight and volume, ensuring a robust assessment of soil properties.
Carbon stock assessment relies on species-specific allometric equations to estimate above-ground carbon (AGC) and below-ground carbon (BGC) Soil organic carbon (SOC) is computed with standard formulas that integrate TOC percentage, bulk density, and sampled soil depth.
In addition to field and laboratory methods, remote sensing and GIS techniques were used to assess historical land cover and carbon distribution Satellite data preprocessing, including cloud removal and NDVI mapping, was performed to enhance classification accuracy Time-series land-use classification was conducted using Global Mangrove Watch (GMW) datasets and Landsat imagery Spatial interpolation of carbon stock estimates was carried out using QGIS software to generate continuous surface maps of blue carbon distribution across the study area
By addressing these dimensions, this research aims to provide a detailed scientific foundation for understanding the Blue Carbon potential of planted mangrove ecosystems in Tien Lang It will also contribute to strategies for sustainable coastal management, focusing on how replanted mangroves can play a role in climate change mitigation through carbon sequestration.
Framework of research
Carbon potential of planted mangrove forests in
Field data from 15 plots in Tien Lang: Tree DBH, height, crown diameter;
Soil samples (0–20 cm, 20–40 cm, 40–80 cm, 80–100 cm)
Measure tree parameters for AGB and BGB estimation Task 1.2: Collect and analyze soil samples for SOC
- Allometric equations (Komiyama et al., 2005; Tamai et al., 1986)
- Biomass formulas for Su based on crown diameter
- LOI and Walkley-Black methods (TCVN 8941:2011)
- AGC (kgC) and BGC (kgC) per plot
- SOC (tC/ha) per depth layer
- Total carbon stock (tC/ha)
2 Estimate carbon sequestratio n rates and historical changes
Past literature data Field measuremen ts from current study
Task 2.1: Review historical studies Task 2.2:
Compare historical and present carbon stock data
- Descriptive statistics and trend analysis
- Carbon sequestration rates over time
3 Assess land-use change impacts on
Task 3.1: Image pre-processing and cloud masking
Blue Carbon using remote sensing
Task 3.2: NDVI calculation and supervised classification Task 3.3:
Biomass estimation using NDVI estimation:
- GIS-based land cover change analysis
- Biomass and carbon estimation from NDVI
4 Visualize and quantify spatial distribution of Blue
Plot coordinates Field and remote sensing- based carbon data
Task 4.1: Spatial interpolation of AGC, BGC, SOC Task 4.2:
Generate Blue Carbon distribution maps
- Spatial analysis of carbon stock hotspots
- Spatial patterns of carbon accumulation and loss
5 Provide recommend ations for mangrove restoration and policy
Integrated spatial and quantitative data results
Interpret spatial and quantitative findings
Task 5.2: Draft restoration and policy recommendations
- Best practices for planted mangrove management
- Carbon conservation strategies for planners and policymakers
Quantified Blue Carbon stocks in Tien Lang Trends in carbon sequestration over a decade Spatial changes in mangrove coverage
Recommendations for sustainable mangrove management and conservation efforts
Land-Use Impact Analysis using spatial change detection
SOC to quantify carbon stocks
Secondary data to compare historical sequestration rates
Sentinel-2A imagery to map mangrove extent and changes
Combine all datasets for holistic evaluation of carbon potential
Evaluate the Blue Carbon potential using field and remote sensing data Analyze carbon sequestration rates
(AGC,BGC & SOC stocks) Assess impact of land-use changes over time on Blue Carbon
Study Area: Planted mangrove forests in Tien Lang, Hai Phong, Vietnam Timeframe: 1996, 2020, and 2024
Field measurements (AGC & BGC)Satellite data (Sentinel-2A imagery)Literature/secondary data
Research Methods and Methodology
This study uses a mixed-methods approach that integrates field-based carbon stock measurements, laboratory analysis, secondary data review, and geospatial analysis via remote sensing to quantify the Blue Carbon sequestration potential of planted mangrove forests in Tien Lang, Hai Phong, Vietnam, and to assess changes in mangrove cover over time attributable to human activities.
This study centers on Vietnam’s critical mangrove ecosystem in Tien Lang, Hai Phong, identified as the primary site for field data collection due to its ecological significance and heterogeneous land-use changes driven by human activity, providing an ideal case for evaluating Blue Carbon potential The research involves direct field data collection with measurements of Above-Ground Carbon (AGC) and Soil Organic Carbon (SOC) to quantify carbon stocks, while secondary data from existing literature and remote sensing techniques are used to analyze spatial patterns and compare results across the Tien Lang mangrove area.
Field Data Collection
Spatial Data Collection of Mangrove Forests
The objective of the field data collection in the Tien Lang mangrove forest was to estimate the Above-Ground Carbon (AGC) and Below-Ground Carbon (BGC) stocks
A total of 15 sampling plots, each measuring 10 m x 10 m, were established using a random sampling method to ensure spatial representativeness across the plantation area
The selection of sampling plots in this study was guided by both ecological composition and practical field accessibility within the mangrove forests of Tien Lang and Bang La communes A total of 15 plots, each 100 m² in size, were established to represent the dominant vegetation structures across the area Nine plots were located in mono-species plantations dominated by Sonneratia caseolaris, a species widely planted in northern Vietnam due to its high salinity tolerance, rapid growth, and adaptability to intertidal environments The remaining six plots were established in mixed-species mangrove areas, which included Kandelia obovata as the predominant species The species composition in the study area was previously documented in a 2023 ecological survey, which reported a distribution of approximately 47.2% Kandelia obovata, 46.9%
Sonneratia caseolaris, and 5.9% other mangrove species (Bang La Forest Management
As noted by Board (2023), the sampling design uses a 9:6 ratio of monoculture to mixed-species plots, ensuring the framework mirrors actual ecological proportions and enables robust comparisons of carbon stocks across plot types Site selection also accounted for tidal accessibility, the representativeness of restoration stages, and proximity to historical land-use transition zones, balancing ecological validity with operational feasibility.
The data collection procedure was as follows; Within each plot, tree measurements were conducted For Sonneratia caseolaris and Kandelia obovata, both
Diameter at Breast Height (DBH, cm) and tree height (m) were recorded For Aegiceras corniculatum, crown diameter (m) and height (m) were measured instead, due to its shrub-like growth form Species-specific allometric equations were applied to estimate the Above-Ground Biomass (AGB) and Below-Ground Biomass (BGB) Biomass estimates were subsequently converted into carbon stocks by applying a standard carbon fraction of 0.47, as recommended by the (IPCC, 2006)guidelines for tropical forest biomass
Figure 2.1 Sampling Plots at Mangrove Forest areas at Tein Lang District, Hai Phong,
Table 2.1 List of Sampling Plots
Plot No Latitude Longitude Species
St 10 20.628549 106.664954 Kandelia obovata, Sonneratia caseolaris & Aegiceras corniculatum
St 11 20.622985 106.648033 Kandelia obovata, Sonneratia caseolaris & Aegiceras corniculatum
St 12 20.622786 106.64822 Kandelia obovata, Sonneratia caseolaris & Aegiceras corniculatum
St 13 20.622811 106.648557 Kandelia obovata, Sonneratia caseolaris & Aegiceras corniculatum
St 14 20.622608 106.648743 Kandelia obovata, Sonneratia caseolaris & Aegiceras corniculatum
Soil Sample Collection
To estimate Soil Organic Carbon (SOC) in the study area, soil samples were collected from the centre of the plots to elucidate the vertical distribution of SOC in mangrove ecosystems Samples were taken at five depth intervals—0–20 cm, 20–40 cm, 40–60 cm, 60–80 cm, and 80–100 cm—covering the upper to deeper soil layers These intervals were chosen based on methodological standards (IPCC, 2006; Kauffman & Donato, 2012) and the ecological processes known to influence carbon accumulation and decomposition at different depths.
0–20 cm (Surface Layer) is the most biologically active soil zone, where fresh litter deposition, microbial activity, and root biomass are typically highest This layer closely reflects recent organic matter inputs and is especially sensitive to surface disturbances, making it a critical indicator of soil health and ecosystem function.
40 cm: Represents the upper subsoil, where the influence of root turnover and humified organic matter becomes more prominent This layer starts to show reduced biological activity but still retains significant carbon due to the downward movement of organic inputs 40–60 cm: A transitional zone between the upper and deeper soil strata Organic matter here is more decomposed and reflects long-term storage under lower oxygen conditions 60–80 cm: This deeper subsoil zone accumulates older, more stabilized carbon The processes of leaching and compaction often dominate here, with less root penetration and microbial activity.80–100 cm: The deepest sampled layer reflects long- term carbon storage Although carbon concentrations are typically lower, this depth is essential for estimating total ecosystem carbon stocks and understanding deep-soil carbon stabilization dynamics
Sampling at these specific intervals allows for accurate estimation of SOC across the vertical profile and facilitates comparisons with international studies that use similar depth frameworks It also supports improved carbon accounting under programs like REDD+ and national greenhouse gas inventories (UNFCCC, 2014)
The procedure was; (1) Soil samples were taken in addition to above-ground measurements in order to assess soil organic carbon (SOC) (2) Within each plot, soil cores were taken from five predetermined depths (0–20 cm, 20–40 cm, 40–60 cm, 60–
80 cm and 80–100 cm) (3) Each sample was labeled and transported to the laboratory for SOC analysis (4) Utilized the Loss on Ignition (LOI) method, following TCVN 6647:2007, the samples were dried and combusted at 550°C to estimate the organic matter content, to get the LOI (%) (5) The Total Organic Carbon (TOC) content of each soil sample was determined using the Walkley-Black method according to TCVN 8941:2011 standards This method is widely regarded for its simplicity, reliability, and suitability for SOC estimation in mangrove soil studies
Table 2.2 Primary Data Collection Methods for Spatial Data and Soil Sample
Data Methodology Data Collection method
Sampling Method Established sample plots within the mangrove forest
- Used GPS to accurately map the location of each plot
Defined a grid system to systematically place plots throughout the study area Each plot was 10m x 10m capture high-resolution spatial data
- Measured tree height, diameter at breast height (DBH), and canopy cover within each plot
15 Sample Plots (10m x 10m) – Random Sampling Method
- Each plot will contain three subplots of 100 square meters (10m x 10m) for detailed measurements
- Soil samples were collected from the centre of each plot at different depths (0-20 cm, 20-40 cm, 40-60 cm, 60-80 cm, and 80-100 cm) and assessed the Soil Organic Carbon (SOC)
- Used soil augers to extract soil samples
- Recorded GPS coordinates for each soil sampling point
- Labelled and stored soil samples properly for laboratory analysis.
Remote sensing data
Data Sources
Global Mangrove Watch (GMW_N21E106) datasets were used to extract mangrove forest extent in 1996 and 2020 at 30 m spatial resolution Landsat 8–9 OLI/TIRS Collection 2 Level-2 imagery were used for NDVI computation in 2024 to evaluate vegetation health and biomass estimation.
Procedure
GMW vector layers for 1996 to 2020 were obtained and clipped to the Tien Lang study boundary Landsat 8–9 surface reflectance data (2024) were downloaded from the USGS Earth Explorer portal Preprocessing steps included; Cloud and shadow masking using the Quality Assessment Band (QA_PIXEL) Atmospheric correction to ensure accurate surface reflectance values Image alignment and clipping to the study area boundary
NDVI was computed from Landsat 8–9 surface reflectance data to assess mangrove canopy health in 2024 and to identify areas of dense or degraded vegetation These NDVI values were then used to estimate above-ground biomass (AGB) through a regression model derived from Quang Tran et al (2018) for mangrove plantations in Hai Phong (Eq 9) The resulting biomass distribution maps span the study area and enable extrapolation of field measurements to a landscape-scale assessment of AGB across the mangrove ecosystem.
Using mangrove cover layers from the Global Mangrove Watch (GMW), we mapped historical land-cover changes from 1996 to 2020 and conducted change detection to identify mangrove gain, loss, and stable areas For 2024, a land-use classification was developed based on NDVI thresholds to categorize the landscape into mangroves, non-mangroves, and water bodies The resulting classification maps were validated through visual interpretation and by overlaying field data to confirm accuracy.
Classified mangrove extents and NDVI-based biomass estimates were integrated with field-sampled aboveground carbon (AGC) and soil organic carbon (SOC) data to support a comprehensive carbon-stock assessment of mangrove ecosystems A planned correlation analysis between NDVI-derived biomass and field-measured AGC aimed to evaluate the accuracy of remote-sensing estimations for Blue Carbon Spatial patterns of carbon stock distribution and mangrove health were visualized to understand Blue Carbon dynamics and their temporal changes.
Final Outputs
NDVI maps, biomass distribution maps, classified mangrove cover maps (1996,
2020, 2024), and interpolated carbon stock maps were produced to visualise vegetation health, land-use change, and carbon accumulation patterns in Tien Lang.
Data Analysis
Laboratory Analysis
To assess the Soil Organic Carbon (SOC) content in the Tien Lang mangrove ecosystem, two standardized laboratory methods were applied following Vietnamese national standards (TCVNs), including the Loss on Ignition (LOI) Method — TCVN 6647:2007 The LOI method determines LOI (%) as the total weight loss of soil after combustion, reflecting the soil’s organic matter content A procedure summary was followed to execute the LOI analysis, aligning with guidance from the Vietnam Ministry of Science and Technology (2011).
Loss on ignition (LOI) is used to determine the organic matter content in soil samples In the procedure, an empty crucible is weighed to establish its tare mass, and roughly 2 g of soil is added The crucible with soil is dried in an oven at 105°C for 24 hours to remove moisture, and the combined weight is recorded The dried sample is then combusted in a muffle furnace at 550°C for 4 hours to oxidize and remove organic matter, after which the crucible containing the ash is cooled and weighed The organic matter content is expressed as LOI% and calculated from the weight difference before and after ignition using the standard LOI formula.
● Mass of soil before burning = Mass of crucible + dried soil
● Mass of soil after burning = Mass of crucible + soil after combustion
● Mass of soil = Difference between the weight of the dried soil and the empty crucible.
Soil Organic Carbon (TOC%) Determination — Walkley-Black Method (TCVN 8941:2011)
Total Organic Carbon (TOC%) in soil samples collected from the Tien Lang mangrove area was determined using the Walkley-Black method, applied in accordance with the Vietnamese national standard TCVN 8941:2011 (Vietnam Ministry of Science and Technology, 2011).
Total organic carbon (TOC) content in soil samples was determined using the Walkley-Black titration method, which oxidizes organic carbon with potassium dichromate (K2Cr2O7) in the presence of concentrated sulfuric acid (H2SO4) To ensure homogeneity, soil samples were air-dried, ground, and sieved through a 0.5 mm mesh For the oxidation process, 0.5 g of prepared soil was placed in an Erlenmeyer flask, followed by the addition of 10 mL of 1N potassium dichromate.
20 ml of concentrated sulfuric acid The mixture was gently agitated and allowed to react for 30 minutes to ensure complete oxidation of the organic carbon
To stabilize the solution, 200 mL of distilled water and 10 mL of orthophosphoric acid were added A diphenylamine indicator (1 mL) enabled visual identification of the titration endpoint The mixture was titrated with 0.5 N iron(II) sulfate (FeSO4) until the color changed from purple to green, indicating the endpoint The TOC percentage was calculated based on the volume of FeSO4 used, accounting for a standard blank This procedure was applied across multiple soil depths in each sampling plot to determine the vertical distribution of soil organic carbon (TOC%).
Soil Organic Carbon Calculation
To derive SOC in tons per hectare, will follow these steps after calculating the TOC% from the LOI method a Calculate Bulk Density
After drying the soil sample at 105°C, measure the dry mass (m) of the soil Bulk density (ρ) is the mass of dry soil per unit volume and is expressed in g/cm³, which is required to estimate soil organic carbon (SOC) per unit area For a soil volume V (cm³) corresponding to the sampling depth, bulk density is calculated as ρ = m / V, where m is the dry weight in grams and V is the soil volume This relationship enables converting SOC measurements from a mass basis to an area basis for accurate soil carbon assessments.
Using the TOC (%) and soil bulk density, we calculate soil organic carbon (SOC) per hectare for each soil depth We then sum the SOC across all measured soil depths to obtain the total SOC per hectare.
• TOC% is the Total Organic Carbon concentration from lab analysis,
• Bulk density is in g/cm³,
• Soil depth is in cm,
• The factor 10 converts g/cm² to ton/ha
Carbon Stock Calculation
a Above-Ground Carbon (AGB) Calculation
The AGB data obtained from field measurements were converted to carbon stock using a following allometric equations for Sonneratia caseolaris (Bần chua), Kandelia obovata (Trang), and Aegiceras corniculatum (Sú) which are the dominant species in
Tien Lang’s planted mangrove area is represented by the proportion of biomass that constitutes carbon, enabling an estimate of the mangrove forest’s above-ground carbon stocks In the above-ground carbon calculation, species-specific allometric equations are applied: Sonneratia caseolaris uses Equation 1 (Komiyama et al., 2005); Kandelia obovata uses Equation 2 (Nguyen et al., 2016); and Aegiceras corniculatum (Su) uses Equation 3 (Fu & Wu, 2011) These equations translate measured biomass into carbon stock estimates for each species, contributing to accurate carbon accounting and informing climate mitigation strategies in the mangrove ecosystem.
AGC (above-ground carbon) is derived from AGB (above-ground biomass) by applying a carbon conversion factor of 0.47, a value established from empirical measurements of carbon content across a broad range of plant species This carbon fraction is widely adopted in forest carbon accounting, and the IPCC guidelines for national greenhouse gas inventories specify 0.47 as the default carbon fraction for tropical forests, mangroves, and other ecosystems.
It is suggested by IPCC as a widely accepted global average c Below-Ground Biomass (BGB) Estimation
Below-Ground Biomass (BGB) was estimated using species-specific models based on DBH or biomass relationships: Equation 4 for Sonneratia caseolaris, Equation 5 for Kandelia obovata, and Equation 6 for Aegiceras corniculatum, as recommended by Fu & Wu (2011) Subsequently, Below-Ground Carbon (BGC) was calculated from these BGB estimates.
Below-Ground Carbon (BGC) was then calculated similarly by applying the 0.47 carbon fraction: BGC=BGB×0.47 e Carbon Sequestration Calculation
Calculate the total carbon sequestration by summing the above-ground and soil organic carbon stocks
Total Carbon (ton/ha) = AGC+BGC+SOC [Eq.13]
Spatial Estimation of Carbon Stocks Based on Remote Sensing Data
To estimate the spatial distribution of Above-Ground Biomass (AGB) and, subsequently, Above-Ground Carbon (AGC) across the mangrove forest of Tien Lang, a regression model based on satellite-derived NDVI values was applied rather than traditional spatial interpolation methods such as Inverse Distance Weighting (IDW) The model exploits the well-established relationship between NDVI and biomass to produce continuous AGB and AGC maps, offering potentially higher accuracy and better coverage in mangrove ecosystems By focusing on NDVI-driven regression, the study demonstrates an efficient approach for assessing biomass and carbon stocks across coastal wetlands like Tien Lang.
Using Landsat 8–9 OLI/TIRS imagery from 2024, NDVI was calculated after cloud and shadow masking to enable accurate vegetation assessment The regression equation developed by Quang Tran et al (2018) for Sonneratia caseolaris in Vinh Quang Commune (Eq 9) was applied to estimate biomass Biomass values were then converted to aboveground carbon (AGC) stocks by multiplying by a standard carbon fraction of 0.47, in line with IPCC guidelines Finally, spatial maps of biomass and carbon stocks were produced by applying the NDVI–biomass relationship pixel by pixel across the study area, yielding detailed distributions of vegetation biomass and carbon storage.
Advantages of This Method is that this approach allowed for a more continuous and spatially explicit estimation of carbon stocks across the entire mangrove landscape
It avoided the bias and spatial limitation of interpolating from only 15 field plots, ensuring more reliable large-scale carbon estimation.
Remote Sensing Analysis
This study assesses land cover change dynamics and their influence on Blue Carbon sequestration potential in the Tien Lang mangrove forest area for the periods 1996, 2020, and 2024 By integrating Global Mangrove Watch (GMW) datasets with Landsat 8–9 imagery, the analysis maps mangrove extent and evaluates vegetation health to support robust biomass estimates and carbon stock calculations The results illuminate how changes in mangrove extent and health over time affect carbon sequestration potential and inform targeted conservation and climate mitigation strategies in the region.
Step 1: Historical Mangrove Extent Analysis (1996–2020)
To assess mangrove coverage changes over the long term, Global Mangrove Watch (GMW) raster data from 1996 and 2020 were used GMW raster files for the years 1996,
From the Global Mangrove Watch (GMW) website, we downloaded 30-meter resolution rasters for 2007, 2010, 2015, and 2020 and imported them into QGIS We clipped each raster with a manually digitized polygon shapefile representing the Tien Lang mangrove boundary to isolate the study area The clipped rasters were converted to vector format (polygon shapefiles) for each year using the Raster to Vector tool, and the DN field in each vector layer was dissolved to merge all mangrove areas into a single polygon per year.
We calculated the mangrove area in hectares using the Field Calculator and added a time attribute to each shapefile (e.g., 1996, 2007, 2010) All annual shapefiles were merged with the Merge Vector Layers tool to create a single multitemporal shapefile The Data Plotly plugin was then used to generate a line graph that shows the trend of mangrove cover change over time.
Step 2: NDVI Calculation for 2024 Using Landsat 8–9
To assess current vegetation health, NDVI for 2024 was computed from Landsat 8–9 Surface Reflectance imagery (Collection 2, Level-2) downloaded from USGS EarthExplorer The analysis utilized Band 4 (Red) and Band 5 (NIR), with cloud and shadow masking performed prior to calculation using the QA_PIXEL band and the Fmask algorithm in QGIS or Google Earth Engine to ensure high-quality imagery.
NDVI was calculated using the formula:
Where, NIR = Near-infrared light reflected by vegetation, and RED = Red light absorbed by vegetation
NDVI values range from -1 to 1, with higher values indicating healthier, denser vegetation This index will allow for an assessment of the health of the mangroves over time
The NDVI raster was then generated to assess vegetation vigor across the mangrove area
Step 3: NDVI Threshold Classification for Land Cover Types
After computing the NDVI, we applied a threshold-based land cover classification to identify the dominant landscape types This approach follows thresholds adapted from Tran & Le (2018): mangrove forests correspond to NDVI > 0.2, non-mangrove vegetation to 0.02 < NDVI ≤ 0.2, and water bodies to NDVI ≤ 0.02.
By applying a classified NDVI map, we quantified the area in hectares covered by mangroves, non-mangrove vegetation, and water bodies, delivering a precise land-cover assessment This classification provides a spatial overview of current land use and mangrove health status, informing conservation planning, habitat monitoring, and environmental management.
Figure 2.2 Land Cover Classification Map of Tien Lang District (2024)
(Mangrove forests, non-mangrove vegetation, and water bodies were classified to facilitate targeted NDVI analysis and carbon stock estimation.)
Step 4: Biomass Estimation from NDVI Values
NDVI values were further used to estimate Above-Ground Biomass (AGB) for the mangrove forests using a regression model developed between field-measured biomass (AGB from 10m × 10m plots) and NDVI values
Two approaches were evaluated for estimating mangrove above-ground biomass (AGB) in Hai Phong Option 1 uses the Published Model based on the regression equation proposed by Quang Tran et al (2018) (Eq 8) for mangrove plantations in Vinh Quang Commune, Hai Phong Option 2 relies on an Own Field Data Model developed from 15 field plots, linking plot NDVI to measured AGB values, with the best-fitting form (linear or polynomial) chosen according to R² and residual analysis The model with the superior fit and sound residual behavior was selected for AGB estimation from NDVI data.
Based on model performance, the most appropriate model was chosen to predict AGB across all mangrove pixels for 2024.
Remote Sensing Analysis Validation of NDVI-Derived Biomass with Field
To ensure the accuracy and reliability of biomass and carbon stock estimations derived from satellite-based NDVI data, a validation process was conducted using field- measured data a Regression Model Development
Field-collected aboveground biomass (AGB) values were obtained from 15 sampling plots, each measuring 10 m × 10 m, and these AGB values were matched with corresponding NDVI values extracted from Landsat 8–9 imagery at 30 m × 30 m resolution Linear and polynomial regression analyses were performed to develop predictive models relating NDVI to the field-measured biomass.
The relationship was evaluated using multiple metrics: a high Coefficient of Determination (R²) to assess goodness of fit, a low Root Mean Square Error (RMSE) to quantify prediction error, and visual inspection of scatter plots and residual plots to detect patterns and validate model assumptions Outlier considerations were also addressed as part of the evaluation to ensure robust model performance.
Because NDVI pixel size may differ from plot size, regression analyses prioritized plots dominated by a single species (Sonneratia caseolaris) with consistent NDVI values Plots showing extreme AGB values or containing mixed species were carefully assessed for their influence on model accuracy However, no artificial outlier removal was performed unless justified statistically c Model Application.
The selected regression model (either published or field-derived) was applied pixel-wise across the classified mangrove areas to generate continuous biomass maps
Estimated biomass values were converted to AGC, BGC, and subsequently combined with SOC to derive full ecosystem carbon stocks across the Tien Lang mangrove study area
To explore the relationship between NDVI and biomass in mixed-species plots comprising Sonneratia caseolaris, Kandelia obovata, and Aegiceras corniculatum, a separate regression analysis was conducted specifically for Plots 10–15 Both linear and second-degree polynomial regression models were developed to capture and compare linear versus nonlinear NDVI–biomass relationships in these plots.
Due to the small sample size (n=7), the regression models were treated as exploratory and not used for full carbon stock extrapolation The primary biomass estimation model was based on mono-species plots of Sonneratia caseolaris Limitations of the study include the restricted sample and the single-species approach, which may limit the extrapolation and broader applicability of the results.
Although 10 m × 10 m ground plots sample only a portion of a 30 m × 30 m NDVI pixel, this mixed-pixel effect can introduce uncertainty Nevertheless, field validation serves as essential calibration to improve NDVI-based carbon stock assessments, particularly for planted mangrove ecosystems that exhibit relatively uniform structure.
Limited numbers of mixed-species plots constrained the development of a statistically robust NDVI-biomass model for heterogeneous stands Although preliminary regression analyses for mixed-species plots were conducted, they displayed lower R² values and higher RMSE relative to mono-species models.
Therefore, the mono-species-derived model was cautiously applied for the estimation of carbon stocks across the study area, acknowledging potential biases in areas dominated by mixed mangrove species.
Final Biomass and Carbon Stock Estimation
Following model development and validation, the most reliable regression model (mono-species Sonneratia caseolaris model) was selected for extrapolating biomass and carbon stocks across the mangrove forest area
The final step involved selecting the regression model; a linear regression model developed for mono-species plots was chosen because it demonstrated better performance, with higher R² and lower RMSE, than the mixed-species model The regression equation y = 934.59 x − 327.98 was used to estimate Above-Ground Biomass (AGB) from NDVI values Biomass prediction across the entire mangrove area was then performed by inputting NDVI values for each 30 m × 30 m Landsat pixel covering the study area into the regression model.
A continuous biomass map was generated by applying the regression formula pixel-by-
Above-ground carbon (AGC) was estimated by applying a standard carbon fraction of 47% to the predicted above-ground biomass (AGB) values, in accordance with the IPCC 2006 Guidelines Below-ground biomass (BGB) was estimated using species-specific root-to-shoot ratios (Komiyama et al., 2005) Soil organic carbon (SOC) was interpolated from ground samples using soil bulk density and total organic carbon (TOC) analysis.
Total Ecosystem Carbon (TEC) stock was calculated by summing the aboveground carbon (AGC), belowground carbon (BGC), and soil organic carbon (SOC) components for each pixel (Eq.13) All carbon pools (AGC, BGC, SOC) were mapped across the study area using GIS, providing a comprehensive, pixel-level view of carbon storage Hotspot analysis identified areas with high and low carbon density, informing targeted conservation and carbon management strategies.
Blue Carbon Stock in Tieng Lang Mangrove Forest
Above-Ground Carbon (AGC)
Field measurements were conducted across 15 sampling plots, each measuring
10 m × 10 m (100 m²), covering a total sampled area of 1,500 m² in the planted mangrove zones of Tien Lang District, Hai Phong Above-Ground Biomass (AGB) was estimated using species-specific allometric equations for Sonneratia caseolaris, Aegiceras corniculatum, and Kandelia obovata A standard carbon conversion factor of 0.47 was applied to biomass values to derive Above-Ground Carbon (AGC)
Plot-level above-ground carbon (AGC), expressed as tonnes of carbon per hectare (tC/ha), ranged from 15.71 tC/ha in Plot 7 to 128.53 tC/ha in Plot 12, revealing substantial spatial heterogeneity driven by variations in species composition, tree density, and stand maturity.
An analysis of above-ground carbon (AGC) stock across 15 sampling plots reveals notable variation between mono-species and mixed-species stands, with Plot 7, a Sonneratia-dominated monoculture, recording the lowest AGC at 157.05 kgC per plot (15.71 tC/ha), while Plot 12, a mixed-species stand dominated by Kandelia obovata and Aegiceras corniculatum, shows the highest AGC at 1,285.31 kgC (128.53 tC/ha); the overall average AGC across all plots is 620.13 kgC per plot (62.01 tC/ha), highlighting substantial variability in carbon storage capacity between vegetation compositions and indicating that mixed-species plots generally exhibit higher carbon sequestration potential than mono-species stands.
Additionally, Below-Ground Carbon (BGC) estimates and Soil Organic Carbon (SOC) values were recorded for every plot, enabling a thorough evaluation of Total Organic Carbon (TOC) stock in the subsequent assessment.
Species Composition, Biomass and Below-Ground Carbon (BGC)
Field survey results reveal a clear distinction in species composition and biomass distribution across the 15 sampling plots in Tien Lang Mangrove Forest Plots 1–8 were dominated almost exclusively by Sonneratia caseolaris, forming mono-species stands, while plots 9–15 showed mixed-species stands comprising Kandelia obovata, Aegiceras corniculatum, and Sonneratia caseolaris.
Below-Ground Biomass Carbon (BGC) distribution differed significantly between mono-species and mixed-species mangrove plots In Sonneratia-dominated plots (Plots 1–8), BGC ranged from 3.77 tC/ha (Plot 7) to 17.04 tC/ha (Plot 2), with an average of approximately 10.11 tC/ha By contrast, mixed-species plots (Plots 9–15), which include combinations of Kandelia obovata, Sonneratia caseolaris, and Aegiceras corniculatum, exhibited higher BGC values, ranging from 13.83 tC/ha at the lower end.
9) to 42.85 tC/ha (Plot 12) The average BGC across these mixed-species plots was 27.05 tC/ha This pattern highlights the enhanced below-ground carbon storage capacity in structurally and floristically diverse mangrove ecosystems
Table 3.1 Comparison of AGB, BGB, AGC, and BGC between Mono-Species and
Mixed-Species (Plots 9–15) Average AGB (t/ha) 85.52 t/ha 149.29 t/ha
Average AGC (tC/ha) 42.76 tC/ha 74.65 tC/ha
Average BGB (t/ha) 21.58 t/ha 57.15 t/ha
Average BGC (tC/ha) 10.15 tC/ha 26.86 tC/ha
The field data revealed that mixed-species plots stored over five times more Below-Ground Biomass (BGB) and Below-Ground Carbon (BGC) compared to
In mangrove systems, monoculture plots of Sonneratia show moderate above-ground biomass and limited below-ground growth, whereas mixed-species stands containing large individuals of Aegiceras corniculatum and Kandelia obovata accumulate substantially more biomass both above and below ground The enhanced root biomass in diverse plots underscores the role of species richness and structural diversity in boosting ecosystem carbon storage Consequently, diverse mangrove stands not only sustain higher above-ground productivity but also improve below-ground carbon sequestration potential, a critical mechanism for climate-change mitigation in coastal forests.
Figure 3.1 Comparison of AGB and BGB (Mono-Species vs Mixed Species)
Validation of NDVI-Derived Above-Ground Biomass (AGB) with Field
To assess the accuracy of NDVI-derived biomass estimates for the planted mangrove forests in Tien Lang, field-collected Above-Ground Biomass (AGB) values were compared with biomass estimated from NDVI values using the regression model developed by Quang Tran et al., 2018 The validation focused on mono-species plots (Plots 1 to 8), where Sonneratia caseolaris dominated, minimizing species structural variability
Table 3.2 summarizes the comparison between ground truth AGB values and NDVI-estimated biomass values for each plot, alongside the variance percentage
Table 3.2 Comparison between Ground-Measured AGB and NDVI-Derived Biomass
The variance between ground-measured AGB and NDVI-derived biomass estimates ranged from +45.69% (Plot 2) to +476.50% (Plot 7) Although some plots, such as Plot 2 and Plot 5, demonstrated relatively lower discrepancies, the majority exhibited significant overestimation This trend suggests that NDVI-based models tend to substantially inflate biomass values, particularly in plots where NDVI saturation, species-specific reflectance properties, and sub-pixel variability may compromise estimation accuracy
These discrepancies were expected, considering: Spatial mismatch between the 10m × 10m ground plots and the 30m × 30m satellite pixels NDVI sensitivity limits in high biomass mangrove environments Age and structure variability within Sonneratia caseolaris plantations, even in visually homogeneous stands
Despite the overestimations at individual plots, the regression analysis demonstrated a reasonable general trend between NDVI and field-measured AGB, supporting the use of NDVI-derived biomass estimates for broader-scale carbon stock assessments
Figure 3.2 Comparison of Own Field Model vs Published Model (Mono-Species)
(Comparison between the Own Field-Derived Regression Model and the Published
Model (Tran & Le, 2018) for Estimating Above-Ground Biomass (AGB) based on NDVI in Mono-Species (Sonneratia caseolaris) Plots)
Figure 3.2 shows the relationship between NDVI values and field-measured above-ground biomass (AGB) for mono-species Sonneratia caseolaris plots, comparing two predictive models The regression model built from this study’s field data (red solid line) exhibits a positive linear trend with an R² of 0.24 and an RMSE of 35.77 t/ha, indicating a moderate fit between NDVI and AGB in mono-species mangrove stands In contrast, the published model from Quang Tran et al (2018) (green dashed line) consistently overestimates biomass across the observed NDVI range, underscoring the limitations of applying external models without local calibration Overall, the field-derived model better captures actual ground conditions, reinforcing the need for site-specific models to improve blue carbon stock assessments in the mangrove ecosystem of Tien Lang.
Observations and Notable Trends in AGC
Field measurements and NDVI-derived biomass estimations revealed distinct spatial patterns of carbon distribution across the Tien Lang mangrove plots Plot 12 recorded the highest Above-Ground Carbon (AGC), reaching approximately 128.53 tC/ha, followed closely by Plot 11 at 100.20 tC/ha These plots also had the highest total biomass, exceeding 364 t/ha and 294 t/ha respectively The presence of mature individuals of Kandelia obovata and Aegiceras corniculatum in these mixed-species plots contributed significantly to their enhanced carbon stocks, both above and below ground
Conversely, the lowest AGC values were observed in Sonneratia-dominated plots, particularly Plots 7 and 8, which recorded only 15.71 and 21.58 tC/ha, respectively These low values reflect the early successional stage and limited biomass accumulation of Sonneratia caseolaris, which, despite its fast growth, develops less woody mass compared to other mangrove species
The disparity between mono- and mixed-species stands was evident across all biomass compartments Mixed-species plots (Plots 9–15) consistently exhibited higher AGB, BGB, and AGC values than the Sonneratia-only plots (Plots 1–8) Average
Below-Ground Biomass (BGB) in the mixed plots surpassed 99 t/ha, compared to just 12.8 t/ha in monospecific stands This reflects more developed root systems and structural complexity in diverse communities, which enhance both carbon storage and ecosystem stability
NDVI-derived aboveground biomass (AGB) estimates aligned moderately with field observations in low-biomass sites; however, they substantially underestimated AGB in high-biomass plots—such as plots 11 and 12—by more than 70% This discrepancy highlights the limitations of NDVI, particularly its saturation in dense canopies and its inability to fully capture vertical structure and species-specific traits.
Implications of Carbon Sequestration
The variation in AGC across the study plots highlights the critical role of species composition, vegetation maturity, and stand structure in determining blue carbon potential Monospecific plots dominated by Sonneratia caseolaris (Plots 1–8) consistently recorded lower AGC values, ranging from approximately 15.71 to 71.01 tC/ha Despite the species’ rapid early growth, its limited wood density and relatively smaller stature result in lower long-term carbon accumulation In contrast, mixed- species plots (Plots 9–15), particularly those containing Kandelia obovata and Aegiceras corniculatum, demonstrated substantially higher AGC values, with Plot 12 reaching a maximum of 128.53 tC/ha
High-performing mixed mangrove stands exhibited more mature vegetation and structurally complex canopies, which enhanced below-ground carbon (BGC) storage In particular, Plots 11 and 12 recorded BGC values of 38.13 and 42.85 tC/ha, respectively—more than double those of the best-performing Sonneratia plots These results provide strong empirical support for mangrove restoration approaches that emphasize species diversity and natural succession, showing that such strategies not only boost carbon sequestration but also enhance ecosystem resilience and multifunctionality over time.
Soil Organic Carbon (SOC) Stock Estimation
Overview and Rational
Soil Organic Carbon (SOC) represents a significant fraction of total carbon in mangrove ecosystems, often exceeding the carbon stored in biomass In this study, SOC was quantified across 15 plots using soil samples collected at five depth intervals (0–
100 cm) The analysis was performed using the LOI method and Walkley-Black method (TCVN standards), and results were expressed in tonnes of carbon per hectare (tC/ha)
Soil Organic Carbon (SOC) was estimated using the Loss on Ignition (LOI) method, a commonly applied approach in Vietnam for coastal soils In this method, organic matter content (OM%) is derived from combusted soil samples and then converted to SOC using the Van Bemmelen factor of 1.724, in accordance with the national standard TCVN 8941:2011 (Vietnam Ministry of Science and Technology).
2011) This approach has also been validated by Nguyen, An, et al., 2021 in studies of mangrove soils in Vietnam
For methodological robustness, the LOI results were cross-referenced with IPCC
(2006) guidelines and SOC estimation procedures detailed in Howard, 2014 The
Walkley–Black wet oxidation method was reviewed, as presented in the Coastal Blue Carbon manual (Howard, 2014), and was considered a benchmark for evaluating LOI applicability under field conditions The IPCC (2003) and IPCC (2006) documents provided guidance for integrating soil organic carbon (SOC) into land-use change and carbon accounting frameworks, ensuring consistency with international standards.
TOC and LOI Relationship
To explore the relationship between Loss on Ignition (LOI) and Total Organic Carbon (TOC), a linear regression analysis was performed LOI (%) was used as the independent variable, while TOC (%), determined using the Walkley-Black method, served as the dependent variable The aim was to assess how well LOI could predict TOC, which is essential for estimating Soil Organic Carbon (SOC) in cases where full chemical analysis may be limited The resulting regression model is expressed as:
TOC (%) = 0.599×LOI (%) − 2.9049 a Model Performance and Interpretation
Coefficient of Determination (R²): 0.7285 This indicates that approximately 72.85% of the variance in TOC can be explained by LOI values, suggesting a moderately strong predictive capability
The regression analysis examining the relationship between Loss on Ignition (LOI) and Total Organic Carbon (TOC) demonstrated statistically significant results The p-value for the intercept was 0.0043, and the p-value for the LOI coefficient was 3.38 × 10⁻⁶, both of which are below the conventional significance threshold of α = 0.05 These results confirm that both the intercept and the slope are statistically significant contributors to the model The 95% confidence interval for the intercept ranged from - 4.77 to -1.04, while the confidence interval for the LOI coefficient ranged from 0.41 to 0.79 Notably, the confidence interval for the LOI coefficient does not include zero, further reinforcing its importance as a predictor of TOC
In terms of model interpretation, the slope coefficient of 0.599 indicates that for every 1% increase in LOI, the TOC increases by approximately 0.599%, on average The intercept value of -2.9049 represents the theoretical TOC when LOI is zero; however, this value has limited practical relevance given that natural soils typically contain some organic matter Overall, the model confirms a strong and positive relationship between LOI and TOC in the sampled mangrove soils
Figure 3.3 A linear regression model where Total Organic Carbon (TOC) is predicted using Loss on Ignition (LOI)
To assess the reliability of the developed model (TOC (%) = 0.599 × LOI (%) − 2.9049), a comparison was made with previous studies conducted in Vietnam, particularly the work of Tue et al., 2016, who proposed the relationship; TOC (%) 0.543 × LOI (%) − 0.451; for mangrove soils Although both models exhibit a consistent and strong positive linear relationship between LOI and TOC, slight variations in slope and intercept are evident These differences likely arise from several factors First, regional variation in soil composition, including clay content, mineralogy, and organic matter quality, can influence the proportion of combusted organic matter that actually reflects true organic carbon Second, vegetation type and root biomass vary between sites and species, influencing the deposition and stability of organic carbon in the soil Third, differences in salinity, hydrology, and microbial activity across mangrove zones can alter the decomposition rates and organic matter accumulation, thus affecting the LOI–TOC relationship
Finally, analytical protocols and instrumentation precision, even when based on standard methods, may lead to slight systematic biases in TOC or LOI measurements Despite these variations, the close alignment in overall trends across studies confirms the robustness of LOI as a practical proxy for estimating TOC in mangrove soils under Vietnamese conditions.
Variation of Total Organic Carbon (TOC) by Soil Depth
To investigate the vertical distribution of Total Organic Carbon (TOC) in mangrove soils, TOC concentrations were analysed across five standard depth intervals: 0–20 cm, 20–40 cm, 40–60 cm, 60–80 cm, and 80–100 cm The TOC (%) values were derived using the Walkley-Black method, and descriptive statistics are summarized in Table 3.3
According to Table 3.3, the mean TOC concentration is highest in the surface soil layer (0–20 cm) at 3.78% and declines steadily with increasing depth, reaching a low of 1.76% in the 80–100 cm layer.
Table 3.3 Descriptive statistics of TOC (%) across soil depths (Mean, SD, Min, Max)
Statistical Analysis: ANOVA and Tukey HSD
A one-way ANOVA was performed to test whether SOC concentrations significantly differed among the five depth intervals The analysis yielded: F(4,70)=5.78, p=0.000443
The analysis reveals significant differences in total organic carbon (TOC) across different depths at the 95% confidence level To pinpoint the specific depth intervals responsible for these differences, a Tukey Honest Significant Difference (HSD) post-hoc test was performed.
Table 3.4 Significant pairwise differences in SOC across depths (Tukey HSD results)
80–20 cm -1.63 0.019 Significant ↓ TOC at 80 cm vs 20 cm
100–20 cm -2.02 0.0018 Highly significant ↓ at 100 cm vs 20 cm 100–40 cm -1.82 0.0064 Significant ↓ at 100 cm vs 40 cm 80–40 cm -1.43 0.053 Marginal significance
Other depth comparisons yielded p-values greater than 0.05 and were not statistically significant
Figure 3.4 Boxplot showing the variation of Total Organic Carbon (TOC %) at different soil depths
The figure illustrates a clear decreasing trend in TOC with increasing depth, with the highest median and variability observed at 20 cm and 40 cm.
Discussion and Interpretation
Soil organic carbon (SOC) decreases with depth in mangrove soils, a pattern that matches earlier findings showing the surface is richer in organic matter due to litter deposition, root activity, and microbial communities (Alongi, 2014; Donato et al., 2011) In these systems, the top 40 cm of soil stores significantly more carbon than deeper subsoil layers such as 80–100 cm, indicating that surface soil layers play a pivotal role in carbon sequestration within mangrove ecosystems.
This trend has practical implications for carbon monitoring and conservation, as a large share of soil organic carbon (SOC) is stored in the upper soil layers, and disturbances such as land-use change, surface erosion, or aquaculture expansion can lead to substantial carbon release.