Physics and Chemistry of the Earth 120 (2020) 102934

Contents lists available at ScienceDirect

Physics and Chemistry of the Earth

journal homepage: http://www.elsevier.com/locate/pce

Projected ecosystem response to the anticipated effects of sea level rise

D.S.M. Ishak a, *, A.K. Abd Wahab a, b, S.Y. Teh c, M.H. Jamal a, b
a
Center for Coastal and Ocean Engineering, Research Institute for Sustainable Environment (RISE), Universiti Teknologi Malaysia, 81310, Skudai, Johor, Malaysia
b
School of Civil Engineering, Faculty of Engineering, Universiti Teknologi Malaysia, 81310, Skudai, Johor, Malaysia
c
School of Mathematical Sciences, Universiti Sains Malaysia, Penang, 11800, Malaysia

A R T I C L E I N F O A B S T R A C T

Keywords: The impact of sea level rise (SLR) on a mangrove island ecosystem was investigated by projecting future increase
Inundation change in tidal inundation at the island. The aims of this study is to simulate the potential mangrove response under
Mangroves several conditions of SLR scenarios for the year 2050 and 2100. Processed WordView-2 and IFSAR images were
Zonation pattern
used in the analysis. Segmentation and classification process were conducted using SPRING 5.2 software while
inundation was simulated using ArcGIS 9.3. A series of mangrove migration map for 2050 and 2100 SLR sce­
narios showed an adjustment to tidal inundation classification due to the increased seawater level. Mangroves
from the higher zone (Z4) migrated into the lower zone (Z2 and Z3) as a response to SLR. In the worst-case
scenario, more than 25% of the mangrove forest is predicted to be permanently submerged due to a 1.3 m
sea level rise. The total land loss in the year 2100 was estimated at 73.52ha (Case study 1) and 148.92ha (Case
study 2). The worst-case scenarios will possibly lead to the extinction of Xylocarpus moluccensis when Z4 starts to
be completely inundated during the 2050 SLR projection. In conclusion, the findings showed that the tolerance to
seawater inundation of different mangrove species determines its responses and migration pattern.

1. Introduction level and distances, period of submergence and frequencies will change
accordingly. These modifications will affect the intertidal ecosystem,
Natural threats originating from the ocean waters have become a especially the low-lying wetland ecosystem such as the mangroves.
major concern to coastal communities. Sea-Level-Rise (SLR) phenome­ Mangroves cannot keep pace with rise in sea level when the intensity of
non was identified as one of the expected problem faced by intertidal changes is drastic. Mangrove migration is an adaptation response by
communities resulting from global climate change (IPCC, 2001).The mangroves to SLR relative to the mangrove surfaces. The migration is
mean annual temperature for selected monitoring stations in Malaysia calculated based on the total movement of seaward mangrove margin to
showed significant increase in warming trend for the past 40 years. In landward position (Gilman et al., 2008a). This phenomenon occurs as
addition, Malaysia also experienced changes in rainfall and rainfall ex­ the mangrove species zones migrate inland to maintain their preferred
tremes, in line with global climate change, with the recurrent severe exposure to tidal period, frequency and depth of the sea (Ellison, 1993).
floods and water stress in recent years (Pour et al., 2020). The projection Mangrove zones migrate landward through seedling recruitment and
of global sea level rise by several models indicated an increase from 12 vegetative reproduction. Mangrove response will consequently change
to 22 cm in seawater surface during the 20th century (Solomon et al., the future mangrove composition, the width of the buffer zone and the
2007). The long-term effects from SLR will reduce the function of mangrove zonation pattern. Basic information and prediction of local
intertidal ecosystems due to habitat deterioration (Gilman et al., 2007). mangrove response to the long-term effect of SLR are therefore signifi­
Tidal inundation change is the most obvious physical impact arising cant. In the end, the prediction of mangrove migration will provide
from sea level rise (Cooper et al., 2008). The inundation risk to coastal useful information for site planning and conservation strategies to offset
communities had been extensively studied by many authors from all the anticipated losses. The basic understanding on mangrove habitat and
over the world, such as Marfai and King (2008a), in Indonesia (Kuhn and its response towards the threat is equally important in mangrove
Tuladhar, 2011), in Australia (Tian et al., 2010), in China, and later replanting efforts to ensure the success of the whole conservation stra­
(Di-Nitto et al., 2014) in Kenya. As the sea level rises, the inundation tegies. Therefore, this study aimed to determine the vegetation map and

* Corresponding author.
E-mail address: [email protected] (D.S.M. Ishak).

https://doi.org/10.1016/j.pce.2020.102934
Received 27 December 2019; Received in revised form 17 August 2020; Accepted 1 September 2020
Available online 7 September 2020
1474-7065/© 2020 Elsevier Ltd. All rights reserved.
D.S.M. Ishak et al. Physics and Chemistry of the Earth 120 (2020) 102934

potential mangrove migration for several SLR conditions of the year Gram Schmidt pan-sharpening in ENVI 4.8. Object-based classification
2050 and 2100. was performed in SPRING 5.2 within region growing segmentation and
Bhattacharyya classification. Mapping process developed in ARCGIS 9.3
2. Methodology software. Finally, the accuracy test was introduced to the classified
output by performing the error matrix assessment. An overview of the
2.1. Study area image processing procedures is given in Fig. 2.

The study area covers the whole Kukup Mangrove Island (KMI) 2.2.1. Remote sensing dataset and image processing
(Fig. 1). Kukup Mangrove Island had been recognized as one of the The study used WordView-2 (WV-2) image acquired on May 2011. It
largest uninhabited mangrove islands in the world. It was designated as was commercially available by Digital Globe. The image is readily
a state national park in 1997. The mangrove island was then had been orthorectified and georeferenced to World Geodetic System (WGS-84)
promoted into the RAMSAR sites on January 31, 2003 by RAMSAR projection system. The 25km2 scene covering the whole KMI is a bundle
Convention due to its international importance (Ramsar List, 2019). image consists of 8-band multispectral image with spectral resolution
Kukup Mangrove Island located 1-km from the mainland and contains 2m × 2m and panchromatic image with the resolution up to 0.5 m.
647ha pristine mangrove ecosystem which is naturally established. WorldView-2 image offers wide range of band combination selection
Rhizhopora apiculata, Rhizophora mucronata, Brugueira parviflora and which is an advantage for classification and mapping process. According
Xylocarpus moluccensis among the most common mangrove species to Salehi et al. (2012), WV-2 satellites carries two sensors to capture the
found at the island. It was also reported that Kukup Mangrove Island is a blue, green, red, yellow, coastal, red-edge, infrared-1 and infrared-2
home for critically endangered mangrove species, Brugueira hainaseii or band. Other satellite image used in the study is InterFerometric Syn­
locally known as Berus Mata Buaya (IUCN, 2019). Kukup island is sig­ thetic Aperture Radar (IFSAR) data obtained on 2008. IFSAR images
nificant as shelter and physical protection from the sea for Kukup town. generate digital elevation model for the use of contour, terrain genera­
It helps to block the wave energy reaching Kukup town and provides a tion and, slopes maps from the Digital Terrain Model (DTM) and Digital
favourable condition for maricultural activities located close to the is­ Surface Model (DSM) included in the package. The spatial resolution for
land. Kukup Mangrove Island experience hot and humid weather with both data is 5 m and referenced to the vertical datum Earth Gravitational
annual average temperature is 30-degree celsius while annual rainfall Models (EGM96). In this study, WV-2 images provide data for vegetation
was recorded at average 2455 mm. The region is subject to a semidiurnal mapping process while the IFSAR data were mainly used for inundation
micro-tidal regime with tidal range recorded at 1.83 m. It is also re­ mapping.
ported that KMI was fully submerged by seawater during Mean Higher
High Water (MHHW), provides the favourable condition for the
mangrove growth and expansion in whole island. 2.3. Data analysis

2.3.1. Field data collection

2.2. Materials and methods Field survey was conducted on June 2014 to record the ground
control points (GCPs) and mangrove species inventory. Point Centre
The following section provide an overview for mapping the Quarter Method (PCQM) was adopted during the survey is developed by
mangrove migration at the study area. The image processing involved Cottam and Curtis (1956). It is a plotless method for forest survey which

Fig. 1. Location of the study area.

2
D.S.M. Ishak et al. Physics and Chemistry of the Earth 120 (2020) 102934

Fig. 2. Flowchart for the analysis process.

increase the efficiency by reducing the time and labour cost. A 100m Segmentation is a grouping process of individual pixel with attri­
transect line was randomly constructed across the island. At each tran­ butes similarities and behaviour into the group pixels (Verma et al.,
sect, a centre point is introduced at every 30 m interval. In which four 2011) Segmentation algorithm in SPRING 5.2 uses the region growing
quarter is identified at every 90◦ angles at the centre point (Fig. 3). A method which iteratively merge the pixels of similar mean grey level to
mature mangrove tree nearest to the sampling point is sampled to satisfy the homogeneity criteria. Two segmentation parameter values
represent the species composition at every quarter. Tree height, Diam­ were set-up as the threshold level for the homogeneity. ‘Similarity’ value
eter at the Breast Height (DBH), coordinate location, distance to the determines if two neighbouring pixels can be grouped together, whereas
centre point and species identification was recorded for the sample tree. ‘Area’ value rule out the objects smaller than the determinant value
As access to the deeper part of the forest by walks is not advisable by the (Camara et al., 1996). Based on works by Joshi et al. (2012) (Joshi et al.,
authority, the data collection was focus on the area which is accessible 2012), three sets of parameters were used to determine the best-fit
by the boat through small creeks. Over the field data collection, 116 segmented region. Table 1 show the sets of segmentation parameter
GCPs had been established (Fig. 3). In which, 30% of the GCPs will be and segmented region for the study area. Afterwards, eight categories
used to validate the classification map. Mangrove forest composition is were developed in SPRING 5.2 for classification process using the
calculated based on Cintron and Novelli (1984) (Cintrón and best-fit segmented region. To define threshold for the classification, 116
Schaeffer-Novelli, 1984), formulation to report the species basal area, ground control points (GCPs) associated with mangrove canopies was
relative density, relative dominance, absolute density and absolute used to confine the element to specific classes. Multispectral band
dominance. combination 7 (near infrared-1), 6 (red-edge) and 5 (red) is selected for
classification. The different wavelength combination is capable to
2.3.2. Pan-sharpening, segmentation and classification process discriminate specific vegetation reflectance (Digital Globe, 2009). It is
Pan-sharpening is a pixel fusion process to enhance the low- sensitive to chlorophyll transition that capture the subtle variation in the
resolution multispectral image with the high-resolution panchromatic tree canopies from different mangrove species.
image for WV-2 image. The transformation using Gram-Schmidt pan Bhattacharya coefficients applied to measure the separability of
sharpening tool involved the component substitution fusion to generate spectral classes in segmented region. It measures the overlap between
a new set of transformed bands through three transformation and in­ two statistical population and determine the relative closeness from the
verse stage between multispectral band and panchromatic image. two populations (Khalid et al., 2006). High threshold acceptance level
Transformation was performed in ENVI 4.8 had retain the multispectral was set-up to run the classifier. Other multispectral combinations were
information while acquire the high-resolution information from tested to justify the applicability of the combination. Fig. 4 shows the
panchromatic image. The transformation process optimizes the spatial classified output from Bhattacharya classifier using different multi­
resolution of multispectral band from 2m to 0.5 m for image enhance­ spectral band combinations. The accuracy assessment was performed to
ment. Pan-sharpened multispectral image undergoes segmentation the classified output in error matrix assessment. From the analysis,
process in SPRING 5.2 for the object-based classification process. vegetation band combination was the most fitted classified output by

3
D.S.M. Ishak et al. Physics and Chemistry of the Earth 120 (2020) 102934

ellipsoidal, geoidal and orthometric height. Corrected vertical datum

dataset was then transferred to the local chart datum height by adding
the conversion coefficient derived by National Geodetic Vertical Datum
of study area. Step two was the DEM development in ARCGIS 9.3 using
3D-Analyst tool to form the Triangulation Irregular Network (TIN)
dataset. TIN dataset is a continuous faceted surface triangular mosaic of
distributed point that formed the elevation height known as DEM. Five
elevation height boundaries were determined for DEM at 0.5 m interval
according to local chart datum height. The interval creates elevation
boundaries for mangrove zone as mentioned in Watson Hydrological
Classification.
To analyse the mangrove migration for selected SLR cases, an in­
crease of seawater level projected by National Hydraulic Research
Institute of Malaysia (NAHRIM) was added to current tidal height
boundaries for the study area. Three SLR case scenarios was imposed to
the present DEM namely Case Study 1 (relative SLR), Case Study 2
(average SLR) and Case Study 3 (Worst-case SLR). New features classes
were added to an existing TIN dataset to create a new height boundary to
simulate the amount of water level increased. TIN surface analysis is
repeated to all SLR scenarios defined in Case Study 1, 2 and 3 for the
year 2050 and 2100. New TIN datasets were developed to demonstrate
the changes in mangrove zonation pattern under projected SLR sce­
narios. The analyses were conducted in 3D-Analyst tool of ARCGIS 9.3.

3. Results and discussion

3.1. Field survey and map classification

From the results, Rhizophora apiculata and Rhizhopora mucronata

show strong domination at the study area. The two species displayed a
major contribution to the mangrove species composition by 73% (rela­
tive density) and dominates more than 89% from the total number of
sampled trees (Tables 3 and 4). The relative density was calculated to
show the qualitative measure of the area occupied by the mangrove
species in terms of mangrove stands per hectare. On the other hand, the
Fig. 3. Ground control points, transect line and centre point constructed at the relative dominance refers to the exertion of major controlling influence
study area. of one species upon all others in terms of their size, productivity and
number. Rhizophora sp was found to be the most productive (basal area
94.05% accuracy level. Whereas modified false and visible band com­ = 2.63 m2) and abundant (1867 ind/ha) species as KMI.
bination resulted at 69.64% and 54.62% respectively. Mangrove species To relate the field survey data with the classified map, the percentage
and feature classes is defined as RA (Rhizophora apiculata), RM (Rhizo­ distribution of mangrove species in the classified map were compared to
phora mucronata), BC (Brugueira cylindrical), BP (Brugueira parviflora), percentage relative density from field survey data. Both attributes
SA (Sonneratia alba), XM (Xylocarpus moluccensis), NM (Non-mangrove) represent the general composition of the study area therefore, it is
and WATER (water bodies) in the classified map. The most fitted clas­ appropriate to compare the results. Table 5 shows the result for relative
sified map was exported in raster format and submitted to ARCGIS 9.3 density from field survey and the mangrove distribution illustrated in
for mangrove map development. classification map. From the analysis, the pattern of species composition
was similar for both data. The percentage differences between both data
2.3.3. Mangrove zone classification and migration analysis shows in positive and negative value. A positive value indicates in­
Watson Hydrological Classification (1928) (Watson, 1928) is the creases and a negative value indicate decreases of species composition of
basis for mangrove zonation at the study area. The classification was classification map compare to the field data. Rhizophora apiculata (RA)
developed based on the mangrove species tolerance to the frequency and showed the highest misclassification. The differences between the field
submergence period of tidal inundation. The differences between equi­ and classified map were in the range between 0% and 10%. The most
noctial tides and mean low water neap were divided into 5 intervals fitted mangrove composition was shown in RM (0% differences). The
where each zone has a value of 0.5 m. As the sea water level rises, the differences recorded in RA (10%) and BC (8%) was slightly higher than
tidal range interval will rise at all category and classes. Table 2 shows in the other species. However, mean differences between all species
the mangrove species classification and tidal range for several scenarios. were calculated at 4.5%, which is comparable to the misclassification
Digital Elevation Model (DEM) from IFSAR data provide the elevation percentage recorded for Overall Accuracy (OA) assessment produced in
contour used for mangrove zone classification. The data is critical in­ SPRING 5.2. Accuracy level for forest map classification was considered
formation to model the current and projected tidal inundation under minimally accepted when the percentage of classification is 70% fitted
several sea level rise scenarios. Geoid system used by IFSAR data is the to the reference data. Meanwhile, from a practical point of view, the
world vertical datum, EGM96 whereas, the local geoid for the study area accuracy level of 90%–95% is recommended to properly represent the
was define as Malaysia vertical datum known as MyGEOID field data in any map classifications (Ward et al., 1999). Therefore, in
(WMGEOID04). Two-step procedure was conducted in the analysis to this study the percentage differences. between the classified and field
develop the DEM for the local study area. First order of data analysis is data for each species (0%–10%) was considered as highly acceptable and
the transformation of geoid system using linear relationship of the classified map produced from the analysis is best to represent the
study area.

4
D.S.M. Ishak et al. Physics and Chemistry of the Earth 120 (2020) 102934

Table 1
Segmentation parameters and segmentation output.
Set Parameter Original land-cover image Segmented image

1 Similarity (5), Area (10)

Over-segmentation
Object 1: 4 elements
Object 2: 3 elements
2 Similarity (10), Area (15) Object 1: 1 element
Object 2: 3 elements

Best-fit segmentation
Object 1: 1 element
Object 2: 3 elements
3 Similarity (20), Area (30)

Under-segmentation
Object 1: 1 element
Object 2: 1 element

3.2. Mangrove zonation pattern mangrove island which represents Zone 3. These types of species
flourish at silt clay soil within the range of high and mid tidal zone
Six types of mangrove species were identified from classified map normally near to the downstream. According to Watson (1928), Zone 3
and field survey conducted for KMI. They are Rhizophora mucronata species received less inundation frequency (around 20–45 times per
(RM), Rhizophora apiculata (RA), Sonneratia alba (SA) (Zone 1), Bruguiera month) and was adapted to shorter submergence period. Therefore, the
parviflora (BP) (Zone 2), Brugueira cylindrica (BC) (Zone 3) and Xylo­ species was equipped with knee-root system which is smaller than Zone
carpus moluccensis (XM) (Zone 4). The domination of RA and RM at the 1 roots system. Xylocarpus moluccensis (Zone 4) lives slightly inland (at
outer part of the mangrove island relates to the species tolerance to higher elevations), and is closer to the riverbanks for freshwater supply
longer inundation period and higher salt water content at lower eleva­ (Tomascik et al., 1997). It is always found within the middle to the upper
tion. The mangroves of Zone 1 are normally found to have the stilt or tidal limits of estuarine reaches. The species usually grows individually
prop roots system (Tomascik et al., 1997). This type of root system rather in assemblages. Therefore, in our studies, the XM stands were
provides the stability and breathing apparatus to the species (Ellison and found in the patches located near to the freshwater sources.
Zouh, 2012). The root configuration system increased the stability by
anchoring to the substrate while trapping more sediment within the
system. This is an important feature for mangroves at Zone 1, which has 3.3. Projection of mangrove migration
the longest period of submergence in water, compared to other species
from the inner zone. The Zone 2 mangrove species represented by an The main objective of mangrove zone migration studies is to seek the
occurrence of Bruguiera parviflora. According to Ng and Sivatoshi (1999) information of mangrove zones changes over time towards the year
(Ng and Sivatoshi, 1999), this type of species is commonly found at 2100. This study was carried out to understand the contribution of each
intermediate estuarine zones in the mid-intertidal region. Bruguiera zone to the mangrove forest composition at KMI in the year 2050 and
parviflora thrives on newly formed ground usually on the stiff clay 2100 under projected SLR scenarios. Southeast Asia mangrove
behind pioneer species like Rhizophora sp. which are more salt tolerant. arrangement discussed by (Watson, 1928), (Tomascik et al., 1997) and
Bruguiera cylindrica was found dominating the inner part of the (UNEP, 2004) was used as a comparison to the mangrove classification.
According to UNEP (2004) the arrangement of mangrove zonation

5
D.S.M. Ishak et al. Physics and Chemistry of the Earth 120 (2020) 102934

Fig. 4. Multispectral band combination and classified output a) visible band combination (5, 3,2) b) modified-false band combination (7,3,2) c) vegetation band
combination (7,6,5). d) Mangrove vegetation map derived from classified vegetation band combination.

Table 2
Mangrove species classification and tidal range for several scenarios.
Case Study 1 Case Study 2 Case Study 3

Zone Mangrove Species Height Boundaries (Above 2050 (+0.12 2100 (+0.24 2050 (+0.14 2100 (+0.28 2050 (+0.65 2100 (+1.3
(Occurrence) CD, m) m) m) m) m) m) m)

0 None 1.21–1.71 1.32–1.83 1.45–1.95 1.35–1.85 1.49–1.99 1.86–2.36 2.51–3.01

1 RA-RM-SA 1.72–2.21 1.84–2.33 1.96–2.45 1.86–2.35 2.00–2.49 2.37–2.86 3.02–3.51
2 RA-RM-SA-BP 2.22–2.71 2.34–2.83 2.46–2.95 2.36–2.85 2.50–2.99 2.87–3.36 3.51–4.01
3 RA-RM-SA-BP-BC 2.72–3.21 2.84–3.33 2.96–3.45 2.86–3.35 3.00–3.49 3.37–3.86 4.02–4.51
4 RA-RM-SA-BP-BC-XM 3.22–3.71 3.34–3.83 3.45–3.95 3.36–3.85 3.50–3.99 3.87–4.36 4.52–5.01

Table 3 Table 5
Relative density of mangrove species distribution at the study area. Comparison between the mangrove distribution (classification map) and relative
density (field data) for Kukup Mangrove Island.
Species No. of Total no. of Relative Total Absolute
Individual Individual Tree Density (%) Density (ind/ha) Species Relative Density Mangrove Distribution (Map Differences,
Tree (Field Data), % Classification), % %
RA 60 116 52 1114/ha RA 52 42 − 10
RM 24 21 753/ha RM 21 21 0
BP 6 5 399/ha SA 3 2 − 1
BC 10 9 337/ha BP 5 10 +5
XM 13 11 678/ha BC 9 17 +8
SA 3 3 28/ha XM 11 8 − 3

will affect the assemblages of mangrove species and domination. As the

Table 4
result, mangrove domination migrated from less salt tolerance species
Relative dominance of mangrove species at the study area.
lives at Z4 to more tolerance species at lower zone, Z2 and Z3 (Fig. 5).
Species Basal Area Absolute Density Absolute Relative From the study. six mangrove migration maps were produced
(m2) (ind/ha) Dominance Dominance (%)
accordingly. Changes to mangrove area was demonstrated for each case
RA 1.91 1114 2127.74 71.72 study as stated below:
RM 0.72 753 540.66 18.22
BP 0.12 399 47.88 1.61
BC 0.20 337 67.74 2.28 3.3.1. Case study 1: relative SLR projection (increase seawater level by
XM 0.26 678 178.99 6.03 0.12 m–0.24 m)
SA 0.14 28 3.78 0.13 Case Study 1 (CS1) described SLR scenarios for the local or relative
SLR projection adopted from NAHRIM (2010) (NAHRIM, 2010). Two
SLR scenarios was illustrated in Case Study 1. They are 2050CS1 (0.12 m
pattern from the most seaward to the inner forest ecosystem was identify
increased) and 2100CS1 (0.24 m increased). The total mangrove area
by the tidal inundation and salinity preferences of the species. As sea
migrated from the current condition is summarized in Table 6. Fig. 5
level rises, the proportion of mangrove composition changed to main­
shows the mangrove migration process that change the forest
tain its preferences. From the studies, increased tidal height boundaries

6
D.S.M. Ishak et al. Physics and Chemistry of the Earth 120 (2020) 102934

Fig. 5. Projected mangrove ecosystem response for Case Study 1, 2 and 3.

composition at KMI. was comprised of selected species. The water body is classified as Zone
0 (Z0). Based on the analysis of mangrove zonation for current condi­
3.3.2. Case study 2: average SLR projection (increase seawater level by tion, Z4 dominated KMI by 35% from the total mangrove area. It is
0.14 m–0.28 m) followed by Z3 (29%), Z2 (22%), Z1 (18%), Z0 (2%). Mangrove Z4, is
Case Study 2 (CS2) was conducted for average SLR projection. The characterized by its infrequent inundation and a high diversity of
scenarios were labelled as 2050CS2 (water level increased by 0.14 m) mangrove species. It is located at the highest tidal range boundaries in
and 2100CS (water level increased by 0.28 m). Total mangrove area the study area. High biodiversity of mangrove species within Z4, due to
migrated in this case study was stated in Table 6.. Total mangrove area due to the availability of fresh water and nutrient inputs with lower
migrated from Z1, Z2, Z3 and Z4 into Z0 described the total land loss wave energy. Migration process in case study 1 experience gradually
cause by permanent submergence throughout the projection (Fig. 5). reduction for zone 4 domination from 35% to 16% from 2050 towards
2100. However, zone 3 area had increased from 29% to 40%, in similar
3.3.3. Case study 3: worst case SLR projection (increase seawater level by time period. Similar trend was observed, in the case study 2, which
0.65 m–1.3 m) domination of Zone 4 decreased from 35% to 10% whereas the per­
Case Study 3 (CS3) was conducted to illustrate the worst case SLR centage of zone 3 increased from 29% to 42%. In the same case study,
projection scenario. The projected SLR scenarios was defined as mangrove area converted to water bodies gradually increased from 2%
2050CS3 (water level increased by 0.65 m) and 2100CS3 (water level to 14%. Finally, an increase by 1.3 m for the worst-case scenario (Case
increased by 1.3 m). Results of analysis was summarized in Table 6 and study 3) had change the domination landscape of mangrove zone 4 to
Fig. 5. zone 2 with significant increase from 16% to 44% towards the year
The mangrove zone migration looked at four mangrove zones 2100.
namely Zone 1 (Z1), Zone 2 (Z2), Zone 3 (Z3) and Zone 4 (Z4). Each zone The landward migration is one of the possible scenarios stated by

7
D.S.M. Ishak et al. Physics and Chemistry of the Earth 120 (2020) 102934

Table 6 Currently, mangrove zone 4 (Z4) is expected to comprise the most

Forest composition for Case Study 1,2 and 3. diverse mangrove species at the study area. Zone 4 mangroves experi­
Quadrate Z1-Z4 (ha) Zone 0 (ha) ence less saline water from tidal inundation and received more fresh­
water supply compared to the lower zones (Zone 1, 2, 3). Although
2050CS1
Q1 161.61 4.43 mangrove species such as Rhizophora mucronata, Rhizophora apiculata,
Q2 112.18 2.26 Bruguiera cylindrica and Brugueira parviflora can be found at the lower
Q3 158.49 0.94 zones, the productivity and growth of those species increase towards
Q4 152.6 48.35 Zone 4. For example, a study conducted by (Kitaya et al., 2003) to
Total 584.88 55.98
Rhizophora apiculata found that the trees near to the sea (Zone 1) had
2100CS1
lower mean tree height compared to those in Zone 4. He concluded that
Q1 158.19 7.85 a shorter inundation period and a lower inundation level is the more
Q2 101.69 12.75 preferable. It is explained by the linear relationships between net
Q3 157.29 2.14 photosynthetic rates, leaf conductance and transpiration rates (Kitaya
Q4 137.6 63.35
et al., 2002). The findings showed that the measured of leaf conductance
Total 554.77 86.09
was affected by elevations. The lower elevation would suppress photo­
2050CS2 synthesis and thus the growth of Rhizophora apiculata. Kukup island is
Q1 160.92 5.12 potentially impacted by seawater level increased by 0.12 m–1.3 m from
Q2 111.96 2.48 the year 2050–2100. The total mangrove area is expected to be reduced
Q3 158.04 1.39
in 2100 as shown above. Under the basic assumption of no elevation
Q4 151.57 49.38
Total 582.49 58.37 changes or mitigation measures, the mangrove species will response
naturally to the changes of hydroperiod level due to SLR. The mangrove
2100CS2 response in term of mangrove zone migration was predicted to migrate
Q1 157.52 8.52 from higher mangrove zone (zone 4) to the lower zones (zone 3 and zone
Q2 100.49 13.95 2) under the projected SLR scenarios.
Q3 156.65 2.78
Q4 136.34 64.61
Total 551.00 89.86 4. Conclusion

2050CS3 The total land mass converted into water bodies were used as a
Q1 139.12 26.92
parameter to derive the mangrove landward migration due to SLR.
Q2 98.03 16.41
Q3 141.1 18.33 Kukup mangrove island potentially experiences land loss to water bodies
Q4 124.18 76.77 from 2.55% to 25.79% in 2100 for the worst case SLR scenario. The
Total 502.43 138.43 trend indicated that the mangrove shorelines at Kukup mangrove island
migrated landward as a response to the projected SLR. The mangrove
2100CS3 zone migration was analysed based on the percentage of mangrove zone
Q1 124.8 41.24
Q2 94.21 20.23
composition for the projected scenarios. Mangroves from the higher
Q3 138.45 20.98 zone (Zone 4) migrated into the lower zone (Zone 2 and Zone 3), thus
Q4 118.14 82.81 changing the domination of Xylocarpus moluccensis to Bruguiera sp or
Total 475.6 165.86 Rhizophora sp. In conclusion, the projected SLR scenarios introduce to
Z1-Z4(ha), denotes the total area of land masses. Kukup island shows the pattern of mangrove migration due to SLR. The
Zone 0 (ha), denotes the total area of land masses converted to water bodies. potential mangrove response was successfully investigated and expected
to provide useful knowledge to promote the conservation activities to
Gilman et al. (2008) (Gilman et al., 2008b) for mangrove response. Kukup mangrove island.
Ellison and Zouh (2012) (Ellison and Zouh, 2012 discussed the
mangrove landward migration as the mangrove relationship with SLR Declaration of competing interest
and sediment. The landwards area become available for mangrove
migration when the coastal inundation starts to irrigate and change the The authors declare that they have no known competing financial
tidal extent of the areas. When sedimentation process is maintained at interests or personal relationships that could have appeared to influence
the current condition, the mangrove soil may not be available as fast as the work reported in this paper.
the seawater increase. Seaward edge of mangrove vegetation will
potentially face the mass fatality and the surviving mangroves species Acknowledgement
are likely to invade (via seedling and reproduction) to the landwards
area. From the results it was clearly shows that the KMI will experience This paper has been made possible through the research work
landward mangrove migration when the domination changes from Z4 to granted by, Universiti Teknologi Malaysia through the Centre for
Z2 and Z3. The predicted mangrove zone migration at KMI is summarize Coastal and Ocean (COEI) using vote number A. K091105.5500.07108.
in Fig. 5. Author also wishing to acknowledge the financial support from Ministry
Although the impact of climate change induced by SLR is a long-term of Higher Education (MOHE) to Research Management Centre (RMC),
crisis, the altered hydrological pattern would lead to habitat change for UTM through grant with the vote number R. J130000.7851.5F153
plants and natural sources. The physiographical change would gradually (FRGS/1/2019/WAB13/UTM/02/4). The fieldwork at Kukup island
reshape the mangrove ecosystem and services. The migration of was partially funded by L’oreal UNESCO for Women in Science
mangrove zone correlates to the population decline in the mangrove Fellowship (2017) awarded to Associate Prof. Dr Teh Su Yean of Uni­
forest. Like flora, wetland fauna will experience habitat losses, and non- versiti Sains Malaysia (Vote Number: S. J130000.3322.4Y128).
resident fauna (such as migratory birds) potentially affected by the loss
of mangrove area. Vegetation provide keystone ecosystem functions References
(such as habitat structure and primary production) therefore, the
reduction will exacerbate stresses on wetland fauna populations. Camara, G., Souza, R.C.M., Freitas, U.M., 1996. SPRING: integrating remote sensing and
GIS by object oriented data modelling. Comput. Graph. 20, 395–403.

8
D.S.M. Ishak et al. Physics and Chemistry of the Earth 120 (2020) 102934

Cintrón, G., Schaeffer-Novelli, Y., 1984. Methods for studying mangrove structure. In: Kuhn, M., Tuladhar, R., 2011. Visualising the spatial extent of predicted coastal zone
The Mangrove Ecosystem: Research Methods. UNESCO, United Kingdom, pp. 1–251. inundation due to sea level rise in south-western Australia. Ocean Coast Manag. 54,
Cooper, M.J.P., Beevers, M.D., Oppenheimer, M., 2008. The potential impacts of sea level 796–806.
rise on the coastal region of New Jersey. USA Climate Change 90, 475–492. Marfai, M., King, L., 2008. Tidal inundation mapping under enhanced land subsidence in
Cottam, G., Curtis, J.T., 1956. The use of distance measure in phytosociological Semarang, Central Java Indonesia. Nat. Hazards 44, 93–109.
sampling. Ecology 37, 451–460. NAHRIM, 2010. The Study of the Impact of Climate Change on Sea Level Rise in Malaysia
Di-Nitto, D., Neukermans, G., Koedam, N., Defever, H., Pattyn, F., Kairo, J.G., Dahdouh- Final Report. National Hydraulic Research Institute Malaysia, NAHRIM), Malaysia,
Guebas, F., 2014. Mangroves Facing Climate Change: Landward Migration Potential p. 50.
in Response to Projected Scenarios of Sea Level Rise Biogeosciences, 11, pp. 857–871. Ng, P.K.L., Sivatoshi, N., 1999. A Guide to the Mangrove of Singapore 1: the Ecosystem
Digital Globe, 2009. Digital Globe Core Imagery Products Guide (United State : Longmont and Plant Diversity. Singapore Science Centre, Singapore, pp. 101–103.
Colorado), p. 30. Pour, S.H., Abd Wahab, A.K., Shahid, S., 2020. Physical-empirical models for prediction
Ellison, J.C., 1993. Mangrove retreat with rising sea-level. Bermuda Estuar. Coast.Shelf S. of seasonal rainfall extremes of Peninsular Malaysia. Atmos. Res. 233 https://doi.
37, 75–87. org/10.1016/j.atmosres.2019.104720.
Ellison, J.C., Zouh, I., 2012. Vulnerability to climate change of mangroves: assessment Ramsar List. Ramsar.org [Retrieved on 13 Nov 2019].
from Cameroon. Central Africa’, Biology 1, 617–638. Salehi, B., Zhang, Y., Zhong, M., Dey, V., 2012. Object-based classification of urban areas
Gilman, E.L., Ellison, J., Coleman, R., 2007. Assessment of mangrove response to using VHR imagery and height points ancillary data. Rem. Sens. 4, 2256–2276.
projected relative sea-level rise and recent historical reconstruction of shoreline Solomon, S., Quin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M.,
position’. Environ. Monit. Assess. 124, 105–130. Miller, H.I., 2007. Climate Change 2007: the Physical Science Basis. Cambridge
Gilman, E.L., Ellison, J.C., Coleman, R., 2008a. Threats to mangroves from climate University Press, Cambridge, United Kingdom.
change and adaptation options Aquat, 89 (2), 237–250. Tian, B., Liquan, Z., Xiangrong, W., Yunxuan, Z., Wen, Z., 2010. Forecasting the effects of
Gilman, E.L., Ellison, J., Coleman, R., 2008b. Threats to mangroves from climate change sea-level rise at chong ming dongtan nature reserve in the yangtze delta, shanghai,
and adaptation options. Aquat. Bot. 89 (2), 237–250. China. Ecol. Eng. 36, 1383–1388.
IPCC, 2001. Climate Change 2001: The Scientific Basis. Cambridge University Press, Tomascik, T., Mah, A.H., Nontyi, A., Moosa, M.K., 1997. The Ecology of the Indonesian
Cambridge, United Kingdom, p. 881. Seas (Part 2), Periplus Edition. HK) Ltd, pp. 1–950p.
IUCN, 2019. International union for conservation nature red list. http://www.iuc UNEP, 2004. Mangroves in the South China Sea. United Nations Environments
nredlist.org/search [27 Nov 2019]. Programme (UNEP). Available from: [January 2004].
Joshi, A., Joshi, J.R., Shrestha, N., Shrestha, S., Gautham, S., 2012. object based land Verma, O.P., Madasu, H., Seba, S., Muralidhar, K., 2011. A simple single seeded region
cover extraction using open source software. Nepalese J. Geoinf. 12, 26–30. growing algorithm for color image segmentation using adaptive thresholding’. In:
Khalid, S.M., Umar, M.I., Saquib, M.S., Asim, A.M., 2006. Bhattacharyya coefficient in International Conference on Communication System and Network Technologies,
correlation of gray-scale objects. J. Multimed. 1 (1), 56–61. India, pp. 500–503.
Kitaya, Y., Jintana, V., Pinyayotha, S., Jaijing, D., Yabuki, K., Izutani, S., 2002. Early Ward, T.J., Vanderklift, M.A., Nicholls, A.O., Kenchington, R.A., 1999. Selecting marine
Development of Seven Species Planted at Different Elevation in a Thai Estuary Tree, reserves using habitats and species assemblages as surrogates for biological diversity.
16, pp. 150–154. Ecol. Appl. 9 (2), 691–698.
Kitaya, Y., Suzuki, K., Miyagi, T., Asano, T., Vien, N.G., Le, V.S., 2003. Growth Watson, J.G., 1928. Mangrove Forests of the Malay Peninsula (Singapore : Fraser and
Characteristics of Dominantly Planted Species. Rhizophora Apiculata, and Neave).
Microclimate in the Mangrove Forest in Can Gio District Vietnam. University Annual
Report Osaka University Knowledge Archive [January 2003].