Spatial Model of above Ground Carbon Distribution of Mangrove in

Wildlife Reserve of Karang Gading dan Langkat Timur Laut using

Landsat 8 Satellite Imagery

Nurdin Sulistiyono

12*

, Selpandri G. Sitompul

, Desi Natalie Sinaga

, Pindi Patana

Faculty of Forestry, Universitas Sumatera Utara, Jl. Tridharma Ujung No.1 Kampus USU, Medan, Indonesia

Center of Excellence for Natural Resources-Based Technology, Mangrove and Bio-Resources Group

Universitas Sumatera Utara, Medan, Indonesia

Keywords: Mangrove Forests, Carbon Stock, NDVI, Landsat 8

Abstract: Wildlife reserve of KarangGading dan Langkat Timur Laut (KGLTL) in North Sumatra Province is

conservation forest area where mangrove forest is the dominant type of land cover. Mangrove forest is

important ecosystem because mangrove have rich-carbon stock, most carbon-rich forest among ecosystems

of tropical forest. This research was based on the lack of information on the carbon stock distribution in

Wildlife reserve of KGLTL. The objective of this study was to formulate the mangrove above ground

carbon stock estimation model using landsat 8 satellite imagery, as well as to develop a carbon stock

distribution map based on the selected model. The study found that the normalize different vegetation

index(NDVI) has a considerably high correlation with the above ground carbon is 0.8280. On the basis of

the values of aggregation deviation, mean deviation, bias, root mean square error, paired sample t test, and

R², the best model for estimating the mangrove above ground carbon is -172.00 + 552.89 NDVI with the R²

value of 68.48%. Potency of above ground carbon in Wildlife reserve of KGLTL is 10.71 to 122.10 ton per

ha.

1 INTRODUCTION

Mangrove ecosystem have many functions of

ecological and economical, so mangrove ecosystem

is important coastal ecosystem (Barbier et al., 2008).

Mangrove have rich-carbon stock about 1023 Mg

(Megagram) C ha

-1

, most carbon-rich forest among

ecosystems of tropical forest (Donato, et al., 2011).

The carbon sequestered in vegetated coastal

ecosystems, especially mangrove forests, salt

marshes and seagrass beds called blue carbon

(Mcleod et al., 2011).

The conversion of mangroves to other land uses

occurs mostly in the East Coast of North Sumatra,

such as at Wildlife reserve of KGLTL in Deli

Serdang and Langkat regency. Besides being rich in

marine products, Wildlife reserve KGLTL is also an

important habitat for various types of waterbirds and

has been identified as one of the migratory bird

habitats. The existence of mangrove forests in

Wildlife sanctuary of KGLTL continues to

experience pressure in the form of deforestation. The

results of the study mention that the rate of

mangrove deforestation in Deli Serdang Regency

was <1% while in Langkat Regency it was 2 - 3%

per year (Basyuni and Sulistiyono, 2018).

Deforestation of mangroves causes C emissions of

0.02 - 0.12 Pg (Petagram) per year, which is

estimated to be equivalent to 10% of emissions from

deforestation globally (Donato et al., 2011).

The utilization of remote sensing technologies

such as the use of Landsat 8satellite imagery can be

used to estimate the distribution of above ground

carbon in mangrove ecosystem. Among of mangrove

biomass estimation methods are based on vegetation

indices (Hamdan et al., 2013; Wicaksonoet al., 2011;

Winarso et al., 2015). This research was based on

the lack of information on the carbon stock

distribution in Wildlife reserve of KGLTL. The

objective of this study was to formulate the

mangrove above ground carbon estimation model

using landsat 8 satellite imagery, as well as to

develop a carbon stock distribution map based on

the selected model.

Sulistiyono, N., Sitompul, S., Sinaga, D. and Patana, P.

Spatial Model of above Ground Carbon Distribution of Mangrove in Wildlife Reserve of Karang Gading dan Langkat Timur Laut using Landsat 8 Satellite Imagery.

DOI: 10.5220/0010103301430146

In Proceedings of the International Conference of Science, Technology, Engineering, Environmental and Ramiﬁcation Researches (ICOSTEERR 2018) - Research in Industry 4.0, pages

143-146

ISBN: 978-989-758-449-7

143

2 MATERIALS AND METHOD

2.1 Study Area

This research was conducted in Wildlife reserve of

KGLTLwith the total area of 13,542.37 ha. The

geographical location of Wildlife reserve of KGLTL

lies between 3

51’30” – 3

59’45” North Latitude

and 98

30’ – 98

42’ East Longitude. Wildlife

reserve of KGLTL was located in Deli Serdang and

Langkatregency of North Sumatra Province (Figure

1).

Figure 1: Location of site research

2.2 Analysis of Land Cover and NDVI

The material of this study was landsat 8 satellite

imagerypath/row of 129/057 acquisition on 22

February 2018 which was downloaded from USGS.

Data of Deli Serdang and Langkat regency

administration were obtained from Central Agency

on Statistics in 2010. The method of onscreen

digitalizing is used to analyze of land cover

classification. Land cover validation was done by

confusion matrix to get the overall accuracy and

kappa precision.

Optical data approach is commonly used to

derive vegetation indices for mangrove biomass

estimation (Hamdan et al., 2013; Wicaksono et al.,

2011; Winarso et al., 2015).The formula for getting

NDVI values is:

NDVI = (band 5 - band 4)/(band 5 + band 4) (1)

Where :

Band 4: digital number (DN) of red band

Band 5: digital number (DN) of near infrared

2.3 Estimation of Carbon Stock

Distribution

Collecting data to determine amount of above

ground biomass obtained by taking direct

measurements in the field using a sample plot of 20

x 20 m. Placement of sample plots was done by

purposive sampling while still considering the

distribution and representation of the sample.

Measurements of above ground biomass are carried

out on vegetation with diameters more than 10 cm.

The value of the above ground carbon is 50% of the

above ground biomass. Measurements of above

ground biomass were measured using the equation

developed by Komiyama (Komiyama et al.,

2005)with the equation:

Wtop= 0.251ρD

2.46

(2)

Where :

Wtop = above ground biomass (ton/ha)

D = diameter of breast height (cm)

ρ = wood density (g/cm

)

Regression analysis using ordinary least square

(OLS) was used to determine the relationship

between above ground carbon stock (y) and NDVI

value (x). Testing of OLS regression requirements

such as data normality tests and heterokedasitas tests

is carried out on above ground carbon. Regression

models used in this study were:

1. Linear : y = a + bx (3)

2. logarithmic : y = a + b ln x (4)

3. Power : y = a x

(5)

4. Eksponensial : y = aexp

(6)

Validation tests were carried out to determine the

deviation of the estimated above ground carbon from

regression model with above ground carbon from

field. The model validation test was built using

paired samples t -test, aggregate deviation (AD),

relative deviation (SR), root mean square error

(RMSE) and bias (e). The number of sample plots

used to build model of carbon distribution is 45 plots

while the number of plots for validation tests is 18

plots.

ICOSTEERR 2018 - International Conference of Science, Technology, Engineering, Environmental and Ramiﬁcation Researches

144

3 RESULT AND DISCUSSION

3.1 Land Cover and NDVI

The composition of the land cover distribution in

Wildlife reserve of KGLTL based on landsat 8

satellite imagery (Figure 2a) wasthe area of land

cover water body 631.73 ha (4.66%), bare soil

317.56 ha (2.34%), mangrove forest 9,929.44 ha

(73.32%), settlement 13.25 ha (0.10%), oil palm

plantation 2,422.29 ha (17.89%) and fishpond

228.10 ha (1.68%). The validation test of land cover

analysis in Wildlife reserve of KGLTL showed

overall accuracy 92.40 % and kappa accuracy 89.7

%, respectively. This result showed the classification

of land cover in Wildlife reserve of KGLTL which

produced is good.

The results of the land cover classification in

Wildlife reserve of KGLTL are used as a basis for

determining NDVI distribution in mangrove forests.

Estimation value of NDVI distribution in mangrove

forest shows the values range between 0.31 to 0.54

(Figure 2b).

Figure 2: Land cover (a), NDVI distribution (b) in

Wildlife reserve of KGLTL in 2018

3.2 Carbon Stock Distribution

The statistical test and validation test of above

ground carbon distribution model can be seen in

Table 2. The use of NDVI in the four regression

models in table 2 can be used to estimate above

ground carbon distribution (sig of anova<0.05).

Based of R

value, linear regression model is the

best model for estimating the above ground carbon

distribution with the highest R

value of 68.48%.

The results of validation test using paired

samples t test on four regression models showed no

significant difference between the measurement data

in the field with estimated data from the model (sig

of paired sample t test > 0.05). The result of RMSE

test showed that the linear regression model has

smallest RMSE value while the logarithmic model

has the highest value. This shows the RMSE value

based on the linear model is the best model. The

results of testing the bias value showed the power

model is the best model because it has the smallest

bias value of 47.341.

The results of testing the aggregate deviation

values of all models have a qualifying value because

they are in the range of -1 to 1. While the results of

testing the relative deviation values of linear models

and logarithmic models is qualified, while the power

Table 2: Result of statistic model and validation

No Model

Sig of

anova R

Sig of

t paired RMSE Bias AD RD

y = 552.89x - 172

0.000 68.480 0.993 10.505 55.927 0.042 8.953

y = 2945.9x

4.6272

0.000 68.210 0.826 12.692 47.935 0.004 17.196

y = 226.04ln(x) + 257.58

0.000 67.210 0.687 37.424 58.176 0.044 7.319

y = 0.4788e

11.154x

0.000 67.490 0.755 12.364 47.341 0.005 17.531

Spatial Model of above Ground Carbon Distribution of Mangrove in Wildlife Reserve of Karang Gading dan Langkat Timur Laut using

Landsat 8 Satellite Imagery

145

and exponential models do not qualified because

they have values above 10%.

Figure 3: Above ground carbon distribution in Wildlife

reserve of KGLTL in 2018

Based on the results of statistical test of models

and validation tests, the linear model is the best

model that can be used to estimate the distribution of

above ground carbon in Wildlife reserve of KGLTL.

The results of above ground carbon distribution in

Wildlife reserve of KGLTL can be seen in Figure 3.

Based on the linear regression model, distribution of

above ground carbon in Wildlife reserve of KGLTL

can be classified into 3 classes: class < 50 (ton/ha)

covered 1,863.63 ha, class 50 -75 (ton/ha) covered

3,396.43 ha and > 75 (ton/ha) covered 3,581.82 ha.

This study resulted high correlation between

NDVI and field biomass and have good correlations

between vegetation indices and field biomass

according to another research(Wicaksono et al.,

2011 and Hamdan et al., 2013).The similarity of the

characteristicsof research location may be one of the

causes. Meanwhile, the empirical algorithm is

usually site specific that might not be applicable at

different part, even in the same country (Winarso et

al., 2015).

4 CONCLUSIONS

The utilization of NDVI on Landsat 8 satellite

imagery can be used to estimate the above ground

carbon distribution in Wildlife reserve of KGLTL.

The best regression model for estimating above

ground carbon distribution in Wildlife reserve of

KGLTL is linear regression model with R² value of

68.48%. Wildlife reserve of KGLTL has above

ground carbon stock level between 10.71 to 122.10

ton/ha.

ACKNOWLEDGEMENTS

This study was partly supported by an TALENTA

Grant 2018 (No. 414/UN5.2.3.1/PPM/KP-

TALENTA USU/2018) from the Universitas

Sumatera Utara, Ministry of Research, Technology

and Higher Education, Republic of Indonesia.

REFERENCES

Barbier, E.B., Koch, E.W., Silliman, B.R., Hacker, S.D.,

Wolanski, E., Primavera, J., Reed, D.J., 2008. Coastal

Ecosystem Based Management with Nonlinear

Ecological Functions and Values. Science 319 (5861):

321-323.DOI: 10.1126/science.1150349.

Basyuni, M., Sulistiyono, N., 2018. Deforestation and

reforestation analysis from land-use changes in North

Sumatran Mangroves, 1990-2015.IOP Conf. Series:

Materials Science and Engineering 309 (2018)

012018. doi:10.1088/1757-899X/309/1/012018.

Donato, D.C., Kauffman, J.B., Murdiyarso, D., Kurnianto,

S., Stidham, M.,Kanninen., M., 2011.Mangroves

among the most karbon-rich forests in the tropics,

Nature Geoscience, 4(5), 293-297.

Hamdan O., Khairunnisa M.R., Ammar A.A., Hasmadi

I.M., Aziz H.K., 2013. Mangrove Carbon Stock

Assessment by Optical Satellite Imagery.Journal of

Tropical Forest Science 25(4): 554–565.

Komiyama, A., Poungparn,S., Kato, S., 2005. Common

allometric equations for estimating the tree weight of

mangroves. Journal of Tropical Ecology, 21(04), 471-

477.

Mcleod, E., Chmura, G.L., Bouillon, S., Salm, R., Björk,

M., Duarte, C.M., Silliman, B.R., 2011. A Blueprint

for Blue Carbon: 4 Toward an Improved

Understanding of the Role of Vegetated Coastal

Habitats in Sequestering CO2. Frontiers in Ecology

and the Environment 9(10): 552-560.

Wicaksono, P., Danoedoro, P., Hartono, H., Nehren, U.,

Ribbe, L., 2011. Preliminary Work of Mangrove

Ecosystem Carbon Stock Mapping in Small Island

Using Remote Sensing: Above and Below Ground

Carbon Stock Mapping on Medium Resolution

Satellite Image. Remote Sensing for Agriculture,

Ecosystems, and Hydrology XIII, 8174, 81741B–

81741B–10. https://doi.org/ 10.1117/12.897926.

Winarso, G., Vetrita, Y., Purwanto, A.D., Anggraini, N.,

Darmawan, S., Yuwono, D.M., 2015. Mangrove

Above Ground Biomass Estimation Using

Combination of Landsat 8 And

AlosPalsarData.International Journal of Remote

Sensing and Earth Sciences Vol. 12 No. 2 December

2015: 85 – 96.

ICOSTEERR 2018 - International Conference of Science, Technology, Engineering, Environmental and Ramiﬁcation Researches

146