Estimation of above ground carbon stocks of mangrove forests from remote sensing and field data in hai ha district quang ninh province during 2016 2019

VIETNAM NATIONAL UNIVERSITY OF FORESTRY FOREST RESOURCES & ENVIRONMENTAL MANAGEMENT FACULTY ========================= STUDENT THESIS ESTIMATION OF ABOVE-GROUND CARBON STOCKS OF MANGROV

VIETNAM NATIONAL UNIVERSITY OF FORESTRY FOREST RESOURCES & ENVIRONMENTAL MANAGEMENT FACULTY

=========================

STUDENT THESIS

ESTIMATION OF ABOVE-GROUND CARBON STOCKS OF MANGROVE FORESTS FROM REMOTE SENSING AND FIELD DATA IN HAI HA DISTRICT, QUANG NINH

PROVINCE DURING 2016 – 2019

Major:Natural Resources Management

Code:D850101

Faculty: Forest Resources and Environmental Management

Supervisor: Assoc Prof Dr Hai-Hoa Nguyen

Student: Vu Hong Son Student ID: 1553090678 Class: K60 Natural Resources Management Course: 2015 – 2019

Advanced Education Program Developed in collaboration with Colorado State University, USA

Ha Noi, 2019

Secondly, I sincerely thank to Assoc Prof Dr Hai-Hoa Nguyen, my supervisor who gave me aplenty of useful guidance and advice that help me to finish my study

Thirdly, I would like to express my gratitude and appreciation to the Managing Board

of Forest Protection Department of Hai Ha district because of the permission and enthusiasm support for me to implement the field survey

Last but not least, I am greatly thankful to my group members for accompanying me in the hard but meaningful time

In short, I really thank to all people helping me to finish this study

Hanoi, September 25 th , 2019

Author

Vu Hong Son

CONTENTS

ACKNOWLEDGMENTS

Chapter I 1

INTRODUCTION 1

Chapter 2 LITERATURE REVIEW 3

2.1 Overview of coastal mangrove 3

2.1.1 Status and distribution of mangrove forest in the world 3

2.1.2 Significance of mangroves carbon stock 5

2.1.3 Status and distribution of mangrove forest in Viet Nam 6

2.1.4 Status and distribution of mangrove in Hai Ha, Quang Ninh 7

2.2 Application of remote sensing data and GIS to mangrove mapping and carbon stocks estimation 9

2.2.1 Mangrove biomass and carbon pools estimation approach in the world 9

2.2.2 Advantages of applying remote sensing and GIS in mangrove forest studies 11

2.2.3 Application of remote sensing data and GIS to mangrove studies in the world 12

2.2.4 Application of remote sensing data and GIS to mangrove studies in Viet Nam 14

Chapter 3 GOAL, OBJECTIVES AND METHODOLOGY 15

3.1 Goal 15

3.2 Objectives 15

3.3 METHODOLOGY 16

3.3.1 Remote sensing data 16

3.3.2 Investigating the status of mangrove forests and management scheme in Hai Ha district, Quang Ninh province 16

3.3.3 Calculating above-ground biomass and carbon stocks of mangrove forests during 2013- 2019 in Hai Ha district, Quang Ninh province 19

3.3.4 Quantify the changes in above-ground biomass and carbon stocks of mangrove forests during 2016 - 2019 in Hai Ha district, Quang Ninh province 23

Chapter 4 NATURAL AND SOCIO-ECONOMIC CONDITIONS 25

4.1 Study site 25

4.2 Natural conditions 25

4.3 Socio-economic conditions 26

Chapter 5 RESULTS AND DISCUSSIONS 28

5.1 Status of mangrove forests in Hai Ha district, Quang Ninh province 28

5.1.1 Spatial distribution of Hai Ha mangrove forests 28

5.1.2 Species identification in study area 34

5.1.3 Current management scheme of Hai Ha mangrove forests 34

5.2 Above-ground biomass and carbon stocks of mangrove forests during 2016- 2019 in Hai Ha district, Quang Ninh province 35

5.2.1 Above-ground biomass and corresponding NDVI value of sub-plots in 2019 35

5.2.2 Development of regression models 35

5.2.3 Above-ground biomass and carbon stocks of study area 37

5.3 Changes in above-ground biomass and carbon stocks of mangrove forests 42

5.3.1 Changes in above-ground biomass during 2016 – 2019 42

5.3.2 Change in above-ground carbon 42

5.4 Solutions to enhance carbon stocks of mangrove forests in studies sites 45

5.4.1 Mechanism and policy solutions 45

5.4.2 Technical solutions 45

5.4.3 Implementation of carbon sequestration payment mangrove forests 46

Chapter 6 CONCLUSION, LIMITATION AND FURTHER STUDY 48

6.1 Conclusion 48

6.2 Limitations and further study 48

REFFERNCES 49

LIST OF TABLES

Table 2.1: Mangroves extent in the world (ha) 3

Table 2.2: The 15 mangrove-rich countries 4

Table 2.3: Mangrove forest distribution in Hai Ha district, Quang Ninh province in 2017 8

Table 2.4: Mangrove afforestation projects in Hai Ha district from 1999 to 2016 8

Table 2.5: Allometric equations for biomass calculation of mangrove trees 9

Table 3.1: Satellite images information 16

Table 3.2: Field investigation plots 20

Table 3.3: Field data collection tables 21

Table 3.4: Wood density of mangrove species studied 22

Table 3.5: Regression models acquired for the test 23

Table 5.1: Accuracy assessment of land cover map in 2019 28

Table 5.2: Above-ground biomass and corresponding NDVI of sub-plots in 2019 35

Table 5.3: R-square values and parameters estimated of regression model tested 36

Table 5.4: Above-ground biomass and above-ground carbon of study area in 8/2019 and 12/2016 37

Table 5.5: Money from C-PFES for above-ground carbon of mangrove forest in Hai Ha district on 12/2019 46

LIST OF FIGURES Fig 2.1 : Comparison of mangrove C storage (mean ±95% confidence interval) with that of

major global forest domains 5

Fig 3.1: Polygon of study area 17

Fig 3.2: Spatial distribution of field investigation plots 20

Fig 4.1: Map of study area 25

Fig 5.1: Land cover map of study area in 2019 29

Fig 5.2: NDVI values for Hai Ha mangrove forests in 2019 30

Fig 5.3: Land cover map of study area in 2016 32

Fig 5.4: NDVI values for Hai Ha mangrove forests in 2016 33

Fig 5.5: Quadratic regression model of biomass and corresponding NDVI value 36

Fig 5.6: Spatial distribution of above-ground biomass of Hai Ha mangrove forests in 2019 38 Fig 5.7: Spatial distribution of carbon stocks of Hai Ha mangrove forests in 2019 39

Fig 5.8: Spatial distribution of above-ground biomass of Hai Ha mangrove forests in 2016 40 Fig 5.9: Spatial distribution of carbon stocks of Hai Ha mangrove forests in 2016 41

Fig 5.10: Above-ground biomass changed from 2016 to 2019 of Hai Ha mangrove forests 43 Fig 5.11: Carbon stocks changed from 2016 to 2019 of Hai Ha mangrove forests 44

LIST OF DIAGRAMS Diagram 3.1: Establishment processes of land covers and NDVI map 17

Diagram 3.2: Total above-ground biomass calculation process 19

Diagram 3.3: Processes of quantifying the changes in above-ground biomass 24

and carbon stocks 24

Diagram 5.1: Mangrove management scheme in Hai Ha district, Quang Ninh province 34

Chapter I INTRODUCTION

As the result of climate change, sea level rise has been a global issue According to The Fourth Assessment Report (AR4) of the Intergovernmental Panel on Climate Change (IPCC), global sea level will rise by up to ~60 cm by 2100 in response to ocean warming and glaciers melting However, the recently identified accelerated decline of polar ice sheet mass (Allison

et al., 2009; Rignot et al., 2008; Velicogna, 2009) raises the possibility of future sea-level rise

by more than 1 m in 2100 (Pfeffer et al., 2008; Lowe et al., 2009)

The coastal areas are considered as the most vulnerable areas by sea level rise (Nicholls

et al., 2010) The immediate effect is that flooding occurs frequently and higher intensity, as well as saltwater intrusion of surface waters Meanwhile, longer-term effects also occur as the coast adjusts to the new conditions, including soil erosion and saltwater intrusion into groundwater Coastal wetlands such as saltmarshes and mangroves will also decline unless they have a sufficient sediment supply to keep pace with sea level rise These physical impacts in turn have both direct and indirect socioeconomic impacts, which appear to be overwhelmingly negative (Nicholls et al., 2007) With over 3,260 kilometers of coastline and about 50 percent of the population living in lowland areas (General Statistics Office of Vietnam, 2018), Vietnam is considered as one of the most vulnerable and being negatively impacted by sea level rise In 2007, the World Bank estimated that a one-meter rise in sea level could affect 10 percent of Vietnam’s population, with the GDP loss about 10 percent Mangroves are considered one of the most effective solutions to deal with the effects of sea level rise due to their important roles (Kathiresan, 2012) Specifically, mangroves have important roles in minimizing the fury of cyclones and tsunami, controlling the flood, preventing the coastal erosion, trapping the sediments and recycling nutrient Moreover, mangrove forests also play very important role in climate change mitigation due to the ability

to absorbing huge amount of CO2 from atmosphere

Over the past five decades, Vietnam has lost 51% (about 408.000 ha) of mangrove areas compare to 1943 (Hoa et al., 2018) Large areas of mangrove forest have been cut down and converted into other land use purposes, such as aquaculture, industrial zone, infrastructures, port, etc Moreover, several consequences of climate change, especially sea level rise and increasing salinity level, are also considered as the main factors which lead to decreasing mangrove areas and mangrove degradable in Vietnam Fortunately, there are many great efforts to increase mangrove areas from Vietnamese government and local authorized in recent years Besides, Vietnam has been supported by many organizations around the world in

the restoration and development of mangroves, such as JICA, Japanese Red Cross, Denmark Red Cross, etc

It is important to assess the growth of mangrove forests in order to assess the efficiency

of those efforts Currently, remote sensing is very common and useful tool to detecting mangrove forests extent and calculating carbon stock There are a lot of researches working in the topic of mangrove forests and carbon stocks of mangrove forests which using the remote sensing technique These researches have been conducted in many northern provinces of Vietnam, but just a few case researches in Quang Ninh province Therefore, the research of

“Estimation of above-ground carbon stocks of mangrove forests from remote sensing and field data in Hai Ha district, Quang Ninh province during 2016 – 2019” was

conducted with the expectation that it could provide important data for the locality, mangrove development projects and information for future studies

Chapter 2 LITERATURE REVIEW 2.1 Overview of coastal mangrove

2.1.1 Status and distribution of mangrove forest in the world

The term mangrove has been discussed by experts and scientists for many years

(Tomlison, 1986) It is commonly used to identify trees and shrubs that have developed morphological adaptation to the tidal environment (e.g aerial roots, salt excretion glands and

vivipary of seeds), as well as the ecosystem itself

Mangrove forests are distributed in the inter-tidal region between the sea and the land in the tropical and subtropical regions of the world between approximately 30° N and 30° S latitude (Giri et al., 2010) Their global distribution is believed to be delimited by major ocean currents and the 20° C isotherm of seawater in winter (Alongi, 2009)

The world's first research of the total area of mangroves was carried out by FAO and

UNEP in 1980 In the report, they estimated that the total area of worldwide mangrove forests

were 15,642,673 ha In the following years, many other studies have been conducted to estimate the area of mangroves worldwide The results of these researches were presented in the Table 2.1

Table 2.1: Mangroves extent in the world (ha)

The recent research about the distribution of mangrove forests was conducted in 2010

by Giri, et al In the report, total area of mangrove forests worldwide in 2000 was estimated about 137,760 km2 in 118 countries and territories The largest extent of mangroves was found in Asia (42%) followed by Africa (20%), North and Central America (15%), Oceania (12%) and South America (11%) Approximately 75% of mangroves were concentrated in 15 countries (Table 2.2)

Table 2.2: The 15 mangrove-rich countries

No Country Area (ha) % of global

Source: Giri et al (2011)

The mangrove forests of the world is less than half of what it once was (Spalding et al., 1997; Spiers, 1999) and much of what remains is in a degraded condition (UNEP, 2004) Coastal habitats across the world are under heavy population and development pressures, and are subjected to frequent storms The continued decline of the forests is caused

by conversion to agriculture, aquaculture, tourism, urban development and overexploitation (Alongi, 2002; Giri et al., 2008) About 35% of mangroves were lost from 1980 to 2000 (MA, 2005), and the forests have been declining at a faster rate than inland tropical forests and coral reefs (Duke et al., 2007) Relative sea-level rise could be the greatest threat to mangroves (Gilman et al., 2008) Predictions suggest that 30–40% of coastal wetlands (IPCC, 2007) and 100% of mangrove forests (Duke et al., 2007) could be lost in the next 100 years if the present rate of loss continues

On the positive side, the rate of loss has recently been decreasing from 187 000 ha lost annually in the 1980s (- 1.04 percent per year) to 102 000 ha annually (- 0.66 percent per year) during the 2000 – 2005 period (FAO, 2007) This reflects an increased in people's awareness of the value of mangroves and the benefits of protecting mangroves Most countries have now minimized or even banned the conversion of mangrove areas for aquaculture purposes and require environmental impact assessments prior to large-scale conversion of these areas for other uses This has led to new legislation, better protection and management and, in some countries, to an expansion of mangrove areas through active planting or natural regeneration

2.1.2 Significance of mangroves carbon stock

Mangroves are among the most carbon-rich forests in the tropics, containing on average 1,023 Mg carbon per hectare, and exceptionally high compared to mean carbon storage of the world’s major forest domains (Fig 2.1) (Donato et al., 2011)

Source: Donato et al., (2011)

Fig 2.1 : Comparison of mangrove C storage (mean ±95% confidence interval) with that of

major global forest domains

Forest ecosystems capture carbon in two ways: (1) via carbon fixation and growth of tree biomass (wood) and (2) via accumulation of carbon in soil accumulating with time (Alongi, 2016) Thus, carbon stocks of mangrove ecosystems is contained in two different pools, include: above-ground pools (e.g living trees, dead trees, shrubs and dwarf-trees, liter

and fallen) and ground pools (root systems and soil) Most carbons are stored ground, can reach over 50% and sometimes over 90% of the total ecosystem carbon stock of mangroves (Kaufmann and Donato, 2011), with 75–95 % of tree carbon stored below-ground

below-in dead roots (Alongi et al., 2003, 2004) Most above-ground biomass is eventually lost due to clear-cutting and human use, decomposition, and export to the adjacent coastal zone

The contribution of mangrove to global forest carbon sequestration is small It accounts for about 3 % of carbon sequestered by the world’s tropical forests although accounting for

<1 % of total tropical forest area (Alongi, 2016) However, if mangroves are disturbed, their high area-specific carbon stocks suggest the potential for significant greenhouse gas emissions For example, deforesting mangroves on peat soils results in CO2 emissions (2900 tC/km2/year) comparable to rates estimated from collapse of terrestrial peats (150–3200 tC/km2/year) (Lovelock et al., 2011) A plausible range of CO2 emissions of 112 to 392 tC/ha from cleared mangroves and soils gives a global emissions range of 2–10 % of global deforestation emissions and up to 50 % of emissions from the world’s tropical peatlands (Donato et al., 2011)

In addition, preventing a ton of carbon emissions from mangrove deforestation is less costly than reducing a ton of carbon emissions from currently regulated Greenhouse Gases (GHG) sources in developed countries (Siikamaki et al., 2012)

2.1.3 Status and distribution of mangrove forest in Viet Nam

According to MARD of Viet Nam in 2018, Viet Nam had about 200.000 ha of mangrove forest, which divided into 4 main regions and 12 sub-regions (Hong et al., 1991) included:

(I) Northeast coast area, from Mui Ngoc (Quang Ninh province) to Do Son (Hai Phong city): This area has complicated condition about geomorphology, hydrology and climate In which, some condition are suitable for mangrove development such as the protecting of island system from big waves and storms But there are also many factors which limit the growth and diversity of mangrove forest such as low rainfall, low temperature in winter, low sediment deposition Mangrove forest concentrated in the estuary area such as Tien Yen – Bai Che, where have good conditions for the growth of mangrove forest

(II) Coastal areas of the Northern Delta, from Do Son (Hai Phong city) to Lach Truong (Thanh Hoa province): Natural mangrove forest developed in the estuary area with the appearance of sandy islands in front of estuary area, which prevent the influence of tidal In

the south of this region, the development of natural mangrove forest is prevented due to the great influence of tide and storm The mangrove community is mainly brackish water plants, including the species of Sonneratia caseolaris distributed in estuaries such as Tien Lang (Hai Phong)

III) Central Coast region, from Lach Truong (Thanh Hoa province) to Vung Tau (Ba Ria - Vung Tau): Natural mangroves do not develop along the coast because of empty terrain which vulnerable to large waves, steep coastline, short rivers, less sediment Therefore, natural mangrove forests are mainly concentrated in estuary area and uneven distribution due

to the influence of topography and sandy wind

(IV) Southern coastal area, from Vung Tau (Ba Ria - Vung Tau) to Mui Nai, Ha Tien (Kien Giang province): Natural condition of this region are very favorable for the development of mangroves such as high temperatures, high rainfall, fertile silt and less storms Mangroves in this area are very diverse, with the presence of most mangrove species

in Vietnam In the early 20th century, this area had over 250 thousand hectares of mangrove forest However, a large area of mangroves was destroyed by the war and the pressure of population growth

The change in Vietnam's mangrove areas have not been different from the general trend

of the world From 1943 to 2000, the total area of mangrove forest was significantly reduced from 400,000 ha (Hong et al, 1997) to 156,608 ha (MARD, 2000) The main reasons of mangroves loss come from human activities However, the Vietnamese government and local authorities have been made great efforts in recovering and developing mangrove forests Moreover, a lot of NGOs have also supported mangroves rehabilitation projects of Viet Nam Consequently, mangroves have been well protected and developed, total 14 000 ha of mangrove plantation in various provinces in the last decade (FAO, 2007) The success of rehabilitation efforts in Viet Nam is mainly due to the close cooperation between funding agencies, local authorities, the government and people

2.1.4 Status and distribution of mangrove in Hai Ha, Quang Ninh

According to Cuc et al (2008), total area of mangrove forests in Quang Ninh province

is about 20,818.40 ha, which includes 17,456.2 ha of natural forest (11,774.27 ha of mixed

forest, 894.34 ha of Anvicennia sp., 3,112.44 ha of Rhizophora stylosa, 1,297.75 ha of

Aegiceras corniculatum, and 377.4 ha of Sonneratia caseolaris) and 326,35 ha of planted

forest In the period of 1975 to 2008, Quang Ninh has lost large area of mangrove forest

Specifically, there were 34.000 ha (half of total area) of mangrove forest lost during 1975 to

1990 From 1990 to 2000, there were 300 ha of mangrove forest lost per year

Coastal area of Hai Ha district belongs to Tien Yen – Ha Coi gulf, Quang Ninh province Where, according to Hong et al (1991), has suitable condition for mangrove development such as the protecting of island system from big waves and storms Total mangrove forest area of Hai Ha in 2017 was about 1630.93 ha (Forest Protection Department

of Hai Ha, 2017), and distributed in 6 coastal communes as presented in Table 2.3

Table 2.3: Mangrove forest distribution in Hai Ha district, Quang Ninh province in 2017

Source: Forest Protection Department of Hai Ha district (2017)

Since 1999, many mangrove forests planting projects has implemented by Red Cross of Hai Ha district, which based on the funding from ACMANG organization-Japan Some major mangrove planting campaigns were listed in Table 2.4

Table 2.4: Mangrove afforestation projects in Hai Ha district from 1999 to 2016

1 1999 150 Rhizophora stylosa Quang Minh and Quang Phong

2 2000 50 Rhizophora stylosa Quang Minh and Quang Phong

5 2016 30 Rhizophora stylosa Quang Minh and Quang Phong

Source: Forest Protection Department of Hai Ha district (2017)

However, area and quality of mangroves have not increased over the years due to the low efficiency of afforestation projects and, especially, the conversion of mangrove areas to other uses

2.2 Application of remote sensing data and GIS to mangrove mapping and carbon stocks estimation

2.2.1 Mangrove biomass and carbon pools estimation approach in the world

Above ground biomass (AGB) estimation

To determine the carbon pool of above-ground components, it is necessary to first determine the biomass of each component of the forests However, almost researches have shown that AGB was mainly contained in living tree and the rest components (dead tree, palms, shrubs, seedling, down wood) have just contributed a small proportion in AGB (normally range from 1% - 5%)(Kauffman and Cole, 2010; Kauffman et al., 2011; and Murdiyarso et al, 2010)

A number of allometric equations for mangrove biomass calculation were published Some typical equations were summarized by Kauffman and Donato (2011) and were presented in Table 2.3

Table 2.5: Allometric equations for biomass calculation of mangrove trees

N = 84

Chave et al (2005) Komiyama et al (2008)

Where: B = biomass (kg), h = height (m), D = diameter at breast height (cm), ρ = wood density (g/cm3), N = sample size of trees to develop the equation and R2 is the coefficient of determination

Due to the differences in structure and wood density among species, specific equations for each species are likely to achieved greater accuracy than general equations Nevertheless, the accuracy of any equation can be affected by the varying of wood density, morphology and height diameter relationships among sites (Kauffman and Donato, 2011) Thus, different equations can yield very large differences in biomass predictions

For trees less than 20 cm dbh differences in biomass estimates are not very great However, it is critical to note that differences in estimates of biomass of the largest individuals can lead to significant errors (Chave et al, 2005; Komiyama et al, 2005; Komiyama et al, 2008; Kauffman and Cole, 2010; Kauffman and Donato, 2011) For example, the biomass prediction of the largest Bruguiera tree (69 cm dbh) in Yap (Federated States of Micronesia) mangrove forests (Kauffman et al 2011) was 2588 kg using the Kauffman and Cole (2010) equation, and 7014 kg using the general equation of Komiyama et

al (2008) For the largest trees in this stand the differences were even more dramatic The biomass estimate for an 80 cm dbh tree was 3034 kg using the Kauffman and Cole (2010) equation, but over 3-fold higher (9434 Kg) using the Komiyama et al (2008) general mangrove equation (Kauffman and Donato, 2011)

Above ground carbon (AGC) estimation

Above-ground carbon pools are then determined by multiplying the biomass of individual components by their specifc carbon concentration (percentage) To determine the carbon concentration of above-ground biomass, samples of each component can be analysed via dry combustion or using published carbon concentrations data (Kauffman and Donato, 2011) For example, Kauffman et al (2011) reported the carbon concentration of the wood of

Bruguiera gymnorrhiza as 46.3%, Rhizophora apiculata as 45.9%, and Sonneratia alba as

47.1%

Since the carbon concentration of wood is usually a little less than 50%, it is also common practice to convert biomass to carbon by multiplying by 0.46–0.5, if local or species-specifc values are not available (Kauffman and Donato, 2011)

Below ground biomass (BGB) estimation

Belowground biomass is an important component in mangroves because it comprises a relatively high proportion of the ecosystem compared to upland forests (Komiyama et al., 2008) However, there is small number of allometric equations for belowground biomass of

mangrove forest (Kauffman and Donato, 2011) because of the extreme difficulty in below ground objects collection and field work measurement

Dharamawan and Siregar (2008) provided a belowground biomass equation for

Avicennia marina in Indonesia as:

B TB = 0.199 * ρ 0.899 * D 2.22

Where: BBT = Tree belowground biomass (kg),

ρ = wood density (g/cm3),

D = tree diameter at breast height (cm)

Below ground carbon (BGC) estimation

Below ground carbons are mainly contained in root systems and soils (Kauffman and

Donato, 2011)

The carbon mass of roots is calculated as the product of root biomass and root carbon

concentration Carbon concentrations of roots are typically lower than that of aboveground tree components For example Jaramillo et al (2003) reported carbon concentration of roots in tropical forests as 36– 42%

Soil carbon pools are determined by summing the mass of each sampled soil depth, for

example, the soil horizon was divided into depth intervals of 0–15 cm, 15–30 cm, 30–50 cm, 50– 100 cm and 100–300 cm (Kauffman et al, 2011; Donato et al, 2011; and Murdiyarso et al, 2010) and taking measurements of bulk density and carbon concentration at each layer The soil carbon mass per sampled depth interval is calculated as follows:

Soil carbon (Mg/ha) = bulk density (g/cm 2 ) * soil depth interval (cm) * %C

Where %C is the carbon concentration expressed as a whole number

The total soil carbon pool is then determined by summing the carbon mass of each of

the sampled soil depths

2.2.2 Advantages of applying remote sensing and GIS in mangrove forest studies

Remote sensing is the phenomenon of observing objects from a remote place without

being in direct contact with the object (Lillesand, 2015) The ability of remote sensing to

provide continuous temporal information has been utilized to monitor ecosystem changes (Rogan et al., 2003) and to build up essential knowledge to predict ecosystem response under future environmental changes The main advantage of remote sensing data is that they provide

a synoptic view over a large area, whose measurement would not be possible by using ground based techniques (Tharku et al., 2019)

Remote sensing techniques are remarkably suitable for above ground biomass (AGB)

and, eventually, above ground carbon (AGC) estimation as compared to traditional field measurement methods, the method can even estimate AGB at different scales (Patil et al., 2013) The use of remotely sensed images, such as a Sentinel-2 image, can be cost and time-efficient Moreover, it is also possible to estimate AGB with reasonable accuracy, compared

to AGB estimates through conventional field measurements (Hayashi and Bettinger, 2006)

2.2.3 Application of remote sensing data and GIS to mangrove studies in the world

Numerous researchers have worked on mangrove forest carbon stock estimation using

RS and GIS technique approach

Fatoyinbo et al (2008) was combined image classification, tree height estimation and

ground data to obtain the biomass of mangroves in Mozambique In which, Landsat ETM+ and Shuttle Radar Topography Mission (SRTM) data were used The SRTM data were calibrated using the Landsat derived land-cover map and height calibration equations Stand-specific canopy height-biomass allometric equations developed from field measurements and published height-biomass equations were used to calculate aboveground biomass of the mangrove forests on a landscape scale The results showed that mangrove forests covered a total of 2909 km2 in Mozambique, a 27% smaller area than previously estimated and the total mangrove dry aboveground biomass in Mozambique was 23.6 million tons and the total carbon was 11.8 million tons However, this combination

of optical and radar satellite data can be applied only for above ground biomass estimation (Patil et al., 2015)

Hamdan et al (2013) assess carbon stock of Matang mangrove forest, Malaysia

Landsat-TM and SPOT-5 satellite images for 1991 and 2011 respectively were utilised

to identify mangroves Vegetation indices (NDVI, SAVI, OSAVI) generated from the images was used as a variable to indicate carbon stock and it was correlated to forest inventory information through regression As the result, arbon stocks of Matang Mangroves ranged from 1.03 to 263.65 (tC/ha) and 1.01 to 259.681C (tC/ha) for the years 1991 and 2011 respectively Total carbon stock in Matang mangroves was estimated at about 3.04 (MtC) in year 1991 and 2.15 (MtC) in 2011 This study also suggested that the use of NDVI is relevant and viable in vegetative studies However using NDVI alone can significantly underestimate the biomass of some woody mangroves because NDVI represents canopy properties rather than trunk properties that are crucial for accurate biomass retrieval (Foody et al 2001) Thus, the combination of vegetation indices such as SAVI and OSAVI along with NDVI index can help us obtain higher accuracy (Hamdan et al 2013)

Normalized Difference Vegetation Index (NDVI) quantifies vegetation by measuring

the difference between near-infrared (which vegetation strongly reflects) and red light (which

vegetation absorbs) NDVI always ranges from -1 to +1 But there isn’t a distinct boundary

for each type of land cover Overall, NDVI is a standardized way to measure healthy

vegetation Specifically, high NDVI values present healthy vegetation In contrast, a less

healthy or no vegetation is display by low NDVI values

Another research which also combines NDVI with other vegetation indices for

mangrove carbon stock assessment was conducted by Patil et al (2015) This research was

implemented in the Thane creek of the Mumbai Metropolitan Region, India and using

Resourcesat-2 LISS4 Standard Multispectral Images from National Remote Sensing Centre

(NRSC) Researchers were used allometry equation of Field et al (1995) to calculate biomass:

Biomass = NDVI * PAR * LUE

Where: NDVI is Normalize Different Vegetation Index

PAR is Photosynthesis Active Radiation

LUE is Light use efficiency

Next, an equation was generated to express the relationship between Leaf+Branch

carbon vs Stem carbon Because NDVI represents only the canopy biomass and, eventually,

the carbon stock of canopy rather than trunk biomass (Foody et al., 2001)

The equation is: Stem-carbon = 08784 * (Leaf + Branch)-carbon 0.981

Then, above ground carbon was calculated based on above ground biomass and below

ground carbon was estimated by equation of Patil et al (2013):

BGB = 0.3335 * DBH 1.7294

Finally, the mean AGC and BGC stocks were 21.66 and 18.06 t/ha, respectively,

thereby yielding a total carbon stock of 39.72 t/ha

This study was used only remote sensing and GIS technique approach with no ground

sampling was involved This method brings to researchers a lot of advantage (Patil et al.,

2015) For example, RS and GIS techniques are cost effective, less time consuming and less

manpower requiring methods for estimation of biomass, moreover, the tedious fieldwork and

sampling is avoided It also provides spatial and temporal data for vast areas The

conventional biomass estimation methods involve intensive ground sampling, which are more

error-prone, while using satellite images, effect of such errors is removed

However, the RS and GIS approach cannot measure below ground biomass, a drawback

due to which ground-based exercises become necessary Thus, a carbon stock estimation

study must combine both the approaches (Patil et al., 2015)

2.2.4 Application of remote sensing data and GIS to mangrove studies in Viet Nam

According to the report of GIS Conference of Viet Nam in 2019, remote sensing and

GIS have been widely applied in many fields (98 research papers were submitted in 2019 only) In which, monitoring mangrove forest using remote sensing and GIS is also very developed in recent years However, almost research just focus on detecting mangrove forest extent changed (Thi et al., 2014; Hai-Hoa, 2016; Son and Anh, 2016; and Bao & Hoa, 2018) with no carbon stock calculation involve Thus, the researches in mangrove forest carbon stock calculation just use ground sampling method in fieldwork (Hanh et al., 2014; Hung et al., 2015; Dung et al., 2016; Luong and Nga, 2017; Hanh et al., 2018) Which also bring to researcher a lot of disadvantage as debated

Therefore, approaching and applying remote sensing and GIS in mangrove forest study, especially in mangrove forest carbon stock calculation in Vietnam are very essential

Chapter 3 GOAL, OBJECTIVES AND METHODOLOGY 3.1 Goal

This study aims to provide an additional scientific foundation of estimating

above-ground biomass and carbon stocks of mangrove forests based remote sensing and field survey data

3.2 Objectives

Objective 1: To investigate the status of mangrove forests and management scheme in

Hai Ha district, Quang Ninh province

This objective is going to answer the questions of how much mangrove forest extents there are; how many species there are; where the mangrove forest has spatially distributed; and how the current management scheme of mangrove forest works in study area

Objective 2: To estimate above-ground biomass and carbon stocks of mangrove forests

during 2013- 2019 in Hai Ha district, Quang Ninh province

This objective is going to answer the questions of how much carbon stocks per hectare

there are in mangrove forest; where carbon stocks of mangrove forest are the highest or the lowest, and what the main influential drivers of carbon stock fluctuation are

Objective 3: To quantify changes in above-ground biomass and carbon stocks of

mangrove forests during 2010- 2019 in Hai Ha district, Quang Ninh province

This objective is going to answer the questions of how much biomass and carbon stocks

of mangrove forest have been changed during selected periods; and what drivers of change there are

Objective 4: To propose solutions to enhance carbon stocks of mangrove forests in Hai

Ha district, Quang Ninh province

This objective is going to answer the question of what the solutions should be given to

enhance carbon stocks of mangrove forests in studies sites

3.3 METHODOLOGY

3.3.1 Remote sensing data

Sentinel-2A images, which have been developed and are being operated by European Space Agency (ESA), with the spatial resolution 10 m were obtained from Earth Explore web page

Specific remotely sensed images used in this study were presented in Table 3.1

Table 3.1: Satellite images information

Source: United States Geological Survey (2019)

3.3.2 Investigating the status of mangrove forests and management scheme in Hai Ha district, Quang Ninh province

 To estimate how much mangrove forest extents there are and determine where the

mangrove forest has spatially distributed

Sentinel-2 image with different bands combination (10 m resolution) we use to classify land cover by using NDVI classification method in ArcGIS 10.4 software The field survey data and Google Earth software will be used to evaluate the accuracy of classified map The steps for mapping were arranged in order and illustrated by Diagram 3.2

Step 1: Drawing polygon of mangrove forest area of Hai Ha district

Google Earth software (GES) has provided us high resolution images with the range of 0.1 to 2.5 m depending on the source of the data and the image quality provided by the satellite Images of study area were taken from Maxar Technologies (has been known as DigitalGlobal) with images resolution was 0.3 m, thus, objects of image could be clearly defined

Mangrove polygon was drawn directly on GES Image in 2016 was also considered for drawing process to ensure that no mangrove areas were missed during the entire study period

At the same time, ground-truthing points were used to identify suspicious points on the map, thereby increasing the accuracy of the polygon These processes were demonstrated by Fig 3.1

Diagram 3.1: Establishment processes of land covers and NDVI map

Fig 3.1: Polygon of study area

(1) Create polygon of mangrove from Google Earth Pro software

(2) Clip band 4 and band 8 of satellite images in 2016 and 2019

by polygon created

(3) Calculate NDVI value based

on band 4 and band 8 clipped

(6.1) Google Earth

Pro software

(7) Reclassify map => Create land

cover map => Calculate area of

each land cover type

(8) Separate mangrove object => Create NDVI map of mangrove

Step 2: Clip band 4 (RED band) and band 8 (NIR band) of satellite data by polygon

created

NDVI value of natural forest and other vegetation are usually equal or higher than that

of mangrove forest Then, areas of mangroves and other vegetation with the same NDVI value will be classified into the same layer Moreover, large NDVI values of other vegetation will result in a large range of NDVI values of the map, consequently, the NDVI values of smaller classes are also large That leads to errors in the map classification process In contrast, if the range of NDVI values of maps is small, then the NDVI values of classes are also smaller and the accuracy of the classification map will be higher Therefore, it is necessary to clip the image from the satellite according to the polygon created

Clipping process could be done automatically by ArcGIS software

Step 3: Calculate NDVI value for study area by using band 4 (RED band) clipped and

band 8 (NIR band) clipped according to the equation:

NDVI =

The calculation can be done automatically by ArcGIS software

Step 4: Composite band 1 to band 6 of satellite image data by ArcGIS software to

create map of study area

Step 5: Classify land cover corresponding to each class of NDVI value into 4 types,

which including Water, Bare land, Mangrove, and Other vegetation

Step 6: Map accuracy assessment

(6.1) Classified map was exported from ArcGIS and overlaid into GES Then, the

incorrect classification areas of the classification map have been reclassified

(6.2) Collect ground-truthing points

The undefined points on GES were determined by field survey and the coordinates of the points were determined by Garmin Oregon 750t (GPS device) After that, the coordinates

of the points are uploaded to the GES and the classification map at the same time to assess the accuracy of the classification map

Step 7: Land cover map were created by reclassify NDVI map In which, all classes

represented the same object were combined into one class Then, total area of each land cover type was automatically calculated by ArcGIS software

Step 8: The NDVI value range of mangrove forest was obtained from step 5 Then, the

lowest and highest values of NDVI were used to isolate the mangrove object As the result,

NDVI map of mangrove forest only was achieved and this map was used to calculate mangrove biomass and mangrove carbon later

 To determine how many species there are

The study tends to conduct field survey in order to determine the name and number of mangrove species Then, 10 sample plots will be established At the same time, we measure forest structure, include: total height, diameter-at-breast-height (DBH), status of tree to obtain other objectives

 To investigate current management scheme of local community in Hai Ha district,

Quang Ninh province

The current mangrove forest management scheme of local community was obtained through directly interview with forest rangers in Forest Protection Department of Hai Ha district

3.3.3 Calculating above-ground biomass and carbon stocks of mangrove forests during 2013- 2019 in Hai Ha district, Quang Ninh province

 To calculate above-ground biomass of mangrove forest

The steps for calculation were arranged in order and shown by Diagram 3.2

Diagram 3.2: Total above-ground biomass calculation process

(4) Develop regression model:

- Dependent variable (y): AGB

- Independent variable (x): NDVI

(5) Choose the regression model

have highest R2 value

NDVI map of mangrove forest

Step 1: Field measurement

(1) Determining plot location: 10 plots were created for the study

Mangrove areas of Quang Phong commune accounts for 56.28% of the entire district's mangroves area Therefore, 5 plots (50% of the total plot) were decided to be established in Quang Phong And 5 other plots were equally divided among the remaining 5 communes A general survey of the whole study area was conducted to make a preliminary assessment about the species composition, tree height and stem diameter of mangroves Then collected data was used to determine plot locations to ensure that the mangrove characteristics were represented As the results, the order and spatial location of 10 plots were presented by Table 3.2 and Fig 3.2

Table 3.2: Field investigation plots

time Latitude Longitude

1 Duong Hoa 187.8 21°24'5.08"N 107°39'39.02"E May 29, 2019

2 Tien Toi 65.46 21°24'5.22"N 107°40'13.37”E May 30, 2019

8 Quang Thang 165.22 21°30'19.33"N 107°45'31.21"E January 3, 2019

9 Quang Thanh 44.38 21°30'17.89"N 107°46'37.09"E January 3, 2019

10 Quang Minh 234.61 21°26'57.01"N 107°46'14.02"E January 4, 2019

Fig 3.2: Spatial distribution of field investigation plots

(2) Plot establishment

Fig 3.3: Dimension of surveyed plot

Each plot had dimension of 30m*30m (equivalent 900 m2) Within each plot, three plots were created with dimension of 10m x 10m and the forest structures were measured in sub-plot Fig 3.3 expressed the shape of each sub-plot

(3) Data collection

- Mangrove species identification (Maeda et al., 2016): Object-based image analysis

was applied for identification

- Mangrove structure measurement: species name, trees high, stem diameter: Protocols

for the measurement, monitoring and reporting of structure, biomass and carbon stocks in mangrove forests, which developed by Kauffman and Donato (2012), were applied for the measurement

In which, stem diameter of mangrove tree was measured by diameter tape as:

+ Main stem tree diameters are typically measured at 1.37 m above the ground, which

is also called the diameter at breast height (DBH)

+ For trees with tall buttresses exceeding 1.37 m above ground level, stem diameter was

measured at the point directly above the buttress

+ For stilt rooted species (e.g Rhizophora spp.), stem diameter was measured above the

highest stilt root (Clough and Scott 1989, Komiyama et al., 2005)

- Data recording: Collected data were recorded into Table 3.3

Table 3.3: Field data collection tables

Name of plot: Plot 1 Name of investigator:

Remark Scientific name Local name

Sum

Step 2: Above-ground biomass calculation for each sub-plot

AGB of each sub-plot was estimated by equation:

AGB = 0.251*ρ*(D)2.46 (Komiyama et al., 2005)

Where: AGB is above-ground biomass (kg)

ρ is wood density (g/cm3)

D is stem diameter (cm)

Wood density of mangrove species was obtained from database of World Agroforestry

Centre (2011) and presented by Table 3.4

Table 3.4: Wood density of mangrove species studied

Rhizophora stylosa 0.94 Seng et al (1951)

Avicennia marina 0.73 Louppe et al (2008)

Aegiceras coniculatum 0.64 Seng et al (1951)

Source: Agroforestry Centre (2011)

Step 3: Determine NDVI value of each sub-plot

Each sub-plot have dimension of 10m x10m, which equal to the size of a pixel of the

NDVI map Therefore, each sub-plot had a corresponding NDVI value

NDVI values of sub-plot were obtained automatically from NDVI map of mangrove

forest

Step 4: Regression model development

NDVI and AGB value of each sub-plot from field survey were acquired to develop

regression model In which, AGB played as a dependent variable and NDVI value was an independent variable The 9 types of regression model were acquired for the determining which one had the highest correlation coefficient(Myeong et al., 2006) and SPSS software was used for the test Name of regression models and corresponding allometric equations were shown in Table 3.5

Step 5: The regression model, which had the highest value of R2,was chosen for the biomass calculation

Step 6: Calculate AGB for each pixel of map

The allometric equation of corresponding chosen regression model was used to

calculate AGB for each pixel of NDVI map of mangrove forest in both 2019 and 2016

Step 7: Calculate total AGB of study area

The biomass of the study area was calculated by the total biomass of the all pixels

Table 3.5: Regression models acquired for the test

No Regression model Allometric equation

2 Logarithmic AGB = A + B*log(NDVI)

Source: Myeong et al (2006)

 To calculate how much above – ground carbon per hectare there are in mangrove

forest

It was common practical to covert above-ground biomass to above-ground carbon by

multiplying AGB by 0.46 to 0.5 (Kauffman & Donato, 2012)

In this study, above-ground carbon was calculated from above-ground biomass multiplied with the carbon conversion factor of 0.47 (Barceclete et al., 2016; IPCC, 2006)

 To determine where carbon stocks of mangrove forest are the highest or the lowest

Carbon stock map were created to determine the spatial distribution of carbon pool 3.3.4 Quantify the changes in above-ground biomass and carbon stocks of mangrove forests during 2016 - 2019 in Hai Ha district, Quang Ninh province

 To determine how much biomass and carbon stocks of mangrove forest have been

changed during selected periods

The steps for mapping the change in biomass and carbon stock were arranged in order and shown by Diagram 3.3

Diagram 3.3: Processes of quantifying the changes in above-ground biomass

and carbon stocks

Step 1: Quantify the change of biomass and carbon stock between 2019 and 2016

Biomass and carbon stock map in 2019 and 2016 were used to quantify the change of them through selected period The change was automatically calculated by ArcGIS software

by the equation: “Change” = “biomass/carbon of 2019” – “biomass/carbon of 2016” Each pixel of the map contained a corresponding value of biomass/carbon So, the essence of the above subtraction is the difference of the biomass/carbon value of the same pixel between 2019 and 2016

Step 2: Result analysis

Three ranges of result were considered to analyze the change of biomass and carbon stock:

 “change” > 0: amount of biomass/carbon increased from 2016 to 2019

 “change” = 0: amount of biomass/carbon unchanged from 2016 to 2019

 “change” < 0: amount of biomass/carbon decreased from 2016 to 2019

Step 3: Mapping the change of biomass/carbon from 2016 to 2019

Biomass and carbon stock map

in 2019 and 2016

(1) Quantify the change of biomass and carbon stock between 2019 and 2016:

“Change” = (“biomass/carbon of 2019” - ”biomass/carbon of 2016)

(2) Result analysis:

+ “Change” > 0: Biomass/carbon increase + “Change” = 0: Biomass/carbon unchanged + “Change” < 0: Biomass/carbon decrease

(3) Mapping the change of biomass/carbon

from 2016 to 2019