GIS-Based Land Suitability Mapping for Rubber Cultivation ... · land evaluation techniques (Thomas...

International Journal of Applied Engineering Research ISSN 0973-4562 Volume 12, Number 20 (2017) pp. 9420-9433

© Research India Publications. http://www.ripublication.com

9420

GIS-Based Land Suitability Mapping for Rubber Cultivation in Seremban,

Malaysia

Goma Bedawi Ahmed1,2*, Abdul Rashid M. Shariff 4, Mohammad Oludare Idrees1

Siva Kumar Balasundram 3, Ahmad Fikri bin Abdullah4

1Geospatial Information Science Research Centre (GISRC), University Putra Malaysia. 2Arab Center for Desert Research and Development of Desert Communities Morzok- Libya.

3Department of Agriculture Technology, University Putra Malaysia. 4Department of Biological and Agricultural Engineering, University Putra Malaysia.

*Corresponding Author*Orcid: 0000-0002-5387-0523 & *Scopus Author ID: 57190944215

Abstract

Land suitability evaluation (LSE) is a valuable tool for land

use planning in major countries of the world, including

Malaysia. Previous LSE studies focused mostly on the use of

biophysical and ecological datasets for the design of equally

important socio–economic variables. This study presents sub

national level estimation of suitable agricultural lands for

Rubber crops in Seremban, Malaysia, combining physical,

biophysical and ecological variables. The objective of the

study was to provide an up–to-date GIS-based agricultural

land suitability evaluation (ALSE) for determining suitable

agricultural land for Rubber crops in the study area.

Biophysical and ecological factors that influence agricultural

land use were assembled and the weights of their respective

contributions were assessed using analytic hierarchical

process. For Rubber cultivation, Lenggeng, Pantai and Setul

are the most suitable; while Ampangan, Seremban, Rasah, and

Rantau are moderately suitable. Since Seremban, Labu, and

Pantai are not suitable for growing rubber. The total suitable

land for rubber cultivation in Seremban district is 35575

hectares distributed among the classes as: 16048 hectares is

highly suitable, 15399 hectares moderately suitable and 4128

hectares is marginally suitable. These values represent 45%,

43% and 12% respectively Quantitative assessment of the

model’s detection accuracy reveals to what extent it is

sensitive to suitable and non-suitable land with sensitivity and

specificity values of 84.14% and 76% respectively and overall

accuracy (area under the ROC curve) of 0.80 (80%) with p-

value of <0.0001 at 95% confidence interval.

Keyword: multi criteria, Analytical hierarchy proses, GIS,

Rubber

INTRODUCTION

The growing population in developing countries have

intensified the pressure on both natural and agricultural

resources (Department of Statistics Malaysia, 2015). To meet

the economic demands of the growing world population, an

increased economic return is required. Both population

increases and the process of urbanization have increased the

pressure on agricultural resources (Hedges et al., 2015). This

increased pressure on the available land resources may result

in land degradation. Dependable and accurate land evaluation

is, therefore, indispensable to the decision-making processes

involved in developing land use policies that will support

sustainable rural development (Department of Statistics

Malaysia, 2016). If self-sufficiency in agricultural production

is to be achieved, land reliable evaluation techniques will be

required to develop models for predicting the land suitability

for different types of agriculture crops (Malaysia Digital

Association, 2016). Land evaluation is the systematic

assessment of land potential to find out the most suitable area

for the intended application at hand (Economic Planning Unit,

2015).

Multi-criteria evaluation processes are widely used in some

regional planning processes since they aim at estimating the

potential of land for alternative land uses (Ho et al., 2010).

Among which agricultural land use may be the most important

area where it is applied (Eastman, 1999). This method could

play a key role in future land-use planning (Govindan et al.,

2015). Not forgetting that agricultural land suitability

classification based on indigenous knowledge is also vital to

land use planning (Macharis & Bernardini, 2015). The aim of

systematic assessment of land and water resources is to

identify and put into practice future alternative land uses that

will best meet the needs of the people while at the same time

safeguarding resources for the future (Sola et al., 2016).

Theoretically, the potential of land suitability for agricultural

use is determined by evaluating the process of climate

parameter, soil, water resources and topographical, as well as

the environmental components under the criteria given and the

understanding of the local biophysical restraints (Vaidya &

Hudnurkar, 2013). The use of GIS Multi-Criteria Decision

Making (MCDM) methods allows the user to derive

knowledge from different sources to support land use

planning and management (Alanne et al., 2007). One multi-

attribute technique that has been incorporated into the GIS-

mailto:[email protected]



9421

based land use suitability procedure is the Analytical

Hierarchy Process (AHP)(T.l. Saaty, 2003). MCDM methods,

such as the AHP method, have been successfully applied to

land evaluation techniques (Thomas Saaty, 1977). AHP

allows for transparent decision-making process, however, the

method is rarely used in developing countries such as

Malaysia (Sambasivan & Fei, 2008). We have used a GIS-

based MCDM land suitability analysis method to classify the

study area (Seremban District) with respect to the potential for

Rubber cultivation. We assumed that this goal could

effectively help agricultural insurance through the

identification and separation of land-based capabilities with

regards to environmental, Biophysical and Socio-economical

potential.

MATERIALS AND METHODS

Study area and dataset

The study area is Seremban district, an administrative unit of

the state of Negeri Sembilan in Peninsula Malaysia (Figure 1).

Geographically, the district lies between longitudes 101o 45'

0" E to 102o 6' 0" E and latitudes 3o 0' 0" N to 2o 30' 0" N,

occupying a total land area of about 951.682 kilometer. The

elevation range between 0 – 1180.2 meters above mean sea

level. The study area is typical tropical rainforest with average

precipitation of 173 mm per annum. It is characterized with

two major seasons – wet and dry season. The wet season is

usually between the months of May and September while the

dry months includes October to March (Malaysian

Meteorological Department, 2009), The district comprises of

eight sub-administrative units called “Mukims” with a total

population of 536147 people according to Malaysia statistics

record (Labour force Survey, 2015).The mukims are

Lenggeng, Setul, Pantai, Labu, Seremban, Rantau, Rasah and

Ampangan. The main occupation of the people is agriculture

dealing in rubber farming, oil palm, coconut plant

(Department of Statistics Malaysia, 2015).

Figure 1: Location of the study area, Seremban district.

Data set

Different data sets obtained from different sources in different

projection systems were used in this study. To facilitate data

integration and spatial analysis, all the data were projected

into the same geographic coordinate system; Rectified Skew

Orthomorphic Malaya Grid (meters) - Kertau RSO Malaya

(m) used in west Malaysia. A list of the data and their sources

is presented in Table 2.

Table 2: List of datasets used in the study

No Data Type Description Source

1 Soil chemical

and Physical

values

Profile data for

each type of soil

Department of

Agriculture

(DOA), Kuala

Lumpur

2 Soil map Soil semi detail

at 1:25000

DOA 2010

3 Terrain Topographic

map

Jabatan Ukur

dan Pemetaan

Wilayah

Persekutuan

Kuala Lumpur

(JUPEM, 2010)

4 Land use Scale 1:50000 DOA

5 Rainfall

precipitation

Monthly rainfall

from 9 stations

for 10 years

2005-2015

(DOA)

6 Drainage

network

Scale 1:25000 JUPEM, 2010

7 Length of dry

season

Scale 1:50000 DOA

For each of these data and their derivatives, selection of

criteria and weight analysis are based on the feedback of the

Malaysian Rubber board experts obtained via the administered

questionnaire and interviews. A total number of 25

questionnaires were sent out to respondents in Malaysian

rubber board out of which 19 (76%) were return. Personal

interviews were conducted with the experts to seek their

opinion on the selected criteria and sub-criteria.

Identification and selection of land qualities

Selection of the appropriate criteria for rubber growing area in

Seremban is identified based on expert opinion . A number of

land qualities for evaluating land suitability based on the

recommendation of FAO framework 1976 and 1983 were

identified (Holthuis, 1976). Those selected qualities (Table 3)

and their criteria with respect to rubber cultivation in Malaysia



9422

are chosen based on the opinion of experts in the rubber

industry. Suitability assessment for a land use type includes

four processes:

i. Selection of related land qualities to the land use

within the study area

ii. Selection of land characteristics to be used to

measure each of the selected land qualities

iii. Determination of the threshold values which will

form the boundaries of suitability classes

iv. Determination of how ratings based on individual

qualities are to be combined into overall

suitability.

Each land quality was examined against the following three

criteria: i) The effect of land quality upon use, ii)occurrence of

the critical value of land quality within the study area and

iii) practicability of obtaining information on individual.

Each land quality is given a numerical score 1 to 3 based on

the effect of land quality upon use (Table 4). The number 1

implies that the land quality has significant effect on land use,

2 indicate a moderate effect and 3 shows that it has slight

effect. The second criterion is the occurrence of critical

values within the study area. Three categories were defined:

frequent, infrequent and rare occurrence, similarly with scores

1, 2 and 3, respectively. The third criterion is the

practicability of obtaining information classified as (a)

available (Score 1), (b) not available, but obtainable by

research (Score 2) and (c) not obtainable (Score 3). Decision

on which land quality is important, and which is not, was

made using the scores. If the score of significance is 1 or 2,

then the land quality is used for analysis; but, where the score

is 3 meaning that it is less important, the land quality is

excluded from further analysis. Table 4 depicts the required

land qualities selected for the study area and their associated

land characteristics. They have been aggregated into five main

groupings (climate, soil, topographic and Hydrology). Then,

the criteria that most directly affect crop growth of a

particular land utilization type were identified.

Table 3: Land qualities and characteristics in the study area

Grouping Quality Characteristics Unit

Climate Moisture availability - Annual precipitation

- Length of dry season

mm

Soil

Nutrient availability

Nutrient retention

Rooting conditions

Soil workability

Soil productivity

- pH H2O - Depth to sulfuric horizon

- Soil Organic Matter (SOM)

- Cation Exchange Capacity (CEC)

- Base Saturation

Gravel and stones

Effective soil depth

Texture and structure

Soil class

cm

%

%

%

%

cm

%, class

Kg/h/year

Topographic

Potential for mechanization

Slope

Elevation

Degree

Height

Hydrology

Oxygen availability

Soil drainage class

class



9423

Table 4: Selection of land qualities through spreadsheet

Land Qualities Selection Criteria

(A) Crop Requirements Importance

for

the use

Existing

critical

values in the

study area

Availability of

data in the

study area

Significance

1-Radiation Regime 2 2 3 3

2-Temperature Regime 1 1 2 2

3-Moisture Availability 1 1 1 1

4-Oxygen (soil drainage) 1 1 1 1

5-Nutrient availability 1 1 2 2

6-Nutrient Retention 1 1 1 1

7-Rooting conditions 1 1 1 1

8-Conditions affecting

germination

1 1 1 1

9-Air humidity as affecting

growth

2 3 3 3

10-Conditions for ripening 2 2 3 3

11-Climatic hazards 2 2 3 3

12-Pest and diseases 1 1 3 3

(B) Management Requirements

13-Soil workability 1 2 2 2

14-Potential for mechanization 1 2 2 2

15-Conditions for land preparation

and clearance

2 2 3 3

16-Conditions affecting storage

and

Clearance

1 2 3 3

17-Conditions affecting timing of

Production

2 2 3 3

18-Access within the production

unit

1 3 3 3

19-Size of potential management

Units

2 2 3 3

20-Location: existing/ potential

Accessibility

2 2 3 3

(C ) Conservation Requirements

21-Erosion hazard 1 1 1 1

22-Soil degradation hazard 1 2 3 3

Data preparation

Each data requires some level of processing as an independent

data layer into the modeling. Prior to data integrattion, the

respective input parameters were obtained indidvidually

through a number of processing tasks with each resulting into

specific suitability classifications (based on expert opinion

and previous studies e.g. as presented in Table 5. The

processes into the classification are explained in the following

sub-section.



9424

Table 5: Crop Requirement and sub-criteria

Land use requirements/Land

Characteristics

Land suitability class

S1 S2 S3 N1 N2

Average annual rainfall (mm) 2500 - 3000 2000 - 2500 1500 - 2000 < 1500 -

Dry months (month) 1-2 2-3 3-4 >4

Soil drainage class Good Moderate Mod poor, poor Very poor, rapid

Soil texture (surface) SSiL, SiL SCL, CL LS - s

Soil Series

Soil depth < 60 60 - 140 140 - 200 > 200

Slope (%) <8 8-16 16-30 >30

Elevation (m) 200 200-400 400-600 >600

Magnesium Mg >54 36-54 18-35 <18 -

K available >312.8 156.4-312.8 78.2-156.3 <78.2 -

Cation Exchange Capacity CEC > 10 10-8 8-6 < 6 -

pH (H2O) 5.6-7.0 7.1-8.0 4.5-5.5 <4.0, >8.0 -

P available >25 15-25 6-14 <6

Electrical conductivity (EC) (mS/cm) 0-4 4-8 8-15 >15

Soil productivity(kg/ha/year) >1350 1250-1350 1050-1150 1000-1050 <1000

Rainfall – The rainfall index map (Figure 3 a) provides five

indices based on the average intensity of the rainfall recorded

over 10 years (2005-2015) (Camerlengo & Somchit, 2000).

The result indicates that rainfall intensity decreases towards

the north-east to south-west in the range of average

precipitation of 230 mm to less than 138 mm. An interesting

observation of the result shows that the classes a north-west to

south-east trending which probably due to monsoonal wind

direction and orographic controlled rainfall.

Dry-wet month - Based on expert opinion, dry season should

not be more than four months and should not be less than a

month (Lintner et al., 2012). The classification outcome

(Figure 3 b) shows that almost 80 percent of the study area

has one month or less dry season. The northern part has

approximately a month of dryness while in most of the south

record shows it is less than a month. The remaining part

which falls within the central zone has between two to four

months of dryness.

Soil series - There are nine group of soil series identified in

Seremban by DOA, each with varying characteristics and

chemical composition (Gasim et al., 2011). Four soil series

group chosen as best classes for rubber are: Gajah Mati-

Munchong-Malacca; Serdang-Bungor-Munchong; Durian -

Malacca – Tavy; and Rengam - Jerangau / Munchong -

Seremban (arranged in order of soil series suitability) (Sujaul

et al., 2012) (Figure 3 c). It can be observed that the best two

classes dominates the southern area. Also, a tiny north-south

trending strip east of the northern part of Seremban contains

the best two soil series suitability classes. The subsequent two

suitable classes are dominant in the northern side while the

least class is found in isolated locations within other classes.

Studies have shown that depth of sulfuric horizon is important

to rubber growth. Wong (1986) reported that a considerable

extent of marine alluvial soils in Malaysia is considered

highly acidic due to the presence of excessive quantities of

oxidizable sulfur compounds (Paramananthan & Zauyah,

1986). The presence of the acid sulfate layer within the first

25 cm depth of the soil surface layer constitutes a serious

limitation for rubber (Lin et al., 2011). At depths between 50

cm and 100 cm, the limitation can be considered as minor for

rubber plant. Generally, nutrient retention is identified as one

of the main soil properties for rubber plants to grow well

(Melling et al., 2005). For example, CEC is the most

important soil chemical property that exhibits the degree of

mineral nutrient retention and bioavailability. It is also used as

a reference in qualifying the standard soil fertility.

Soil texture - Texture and structure are considered together

because in combination they influence the workability,

aeration, drainage, root penetrability and water-releasing

capacity of soil (Hambirao et al., 2014) .Texture classes are

defined by the relative contents of the amount of sand, silt,

and clay contents. It can be seen that the study area has

exclusively two distinct soil texture classes, the fine texture

and the coarse texture (Figure 3 d). The fine texture group is

found predominantly in the study area except along the north-

eastern side which contains the second category, coarse



9425

texture soil. The coarse group also extends to the central

region and north-south trending strip in the north-eastern part.

A general observation is that the lower elevation are made up

of fine soil texture while the coarse group is majorly along

high altitude within areas that are still densely forested.

Drainage – The index map comprises of three main

indicators: well drained, moderately well drained and

imperfect (Paramananthan & Zauyah, 1986) (Figure 3 e).

Based on the classification process, regions close to the main

river show considerable level of drainage unsuitability.

Looking at the drainage pattern of Seremban, drainage

suitability factors are somewhat evenly distributed between

the first two best drainage suitability classes. The well drained

class is the best for rubber cultivation followed by the

moderately well drained class(Ahmad et al., 2009). The last

one, somewhat imperfect, is considered not good at all.

Elevation and slope – elevation (in geographic term) defines

the height of locations relative to a fixed reference (Manap et

al., 2010). It has been established that planting rubber trees at

elevation below 200 m above m.s.l produces better yield

(Streutker & Shrestha, 2011). On this basis, the elevation

index map was classified (Figure 3 f). The map shows that

over 90% of the study area fulfills this requirement. Similar to

elevation, the relative change in elevation with respect to the

change in distance defines the direction and steepness of a line

called slope or gradient. This factor affects both rubber

growth and other activities such as the use of machineries and

application of fertilizer. The slope index map classified into

four slope classes (Figure 3 g) shows that about 80% are

below 8 degrees which is favorable for rubber cultivation.

Productivity – This economic evaluation index is directly

linked to soil composition, nutrient, clay content and the

chemical properties (Howell et al., 2005). In Malaysia,

productivity is usually estimated on the basis of how much

yield (kg/ha/year) is obtainable from a parcel of land. No

wonder in the productivity index map (Figure 3 h), the most

profitable areas take a pattern similar to what is obtained in

(Figure 3c)

Figure 2: Methodology flowsart

Criteria weightage and data normalization

At the first level of evaluation, the relative importance of the

criteria weightage was examined using AHP. AHP has been

extensively utilized for multi-criterion decision making of

land suitability in many fields (T. L. Saaty, 2004).It

determines the relative significance of input parameters based

on pairwise comparisons . In pairwise comparison matrix,

criteria level of importance is presented in a range of values

from 1 to 9 where the former indicates “equal importance”

and the latter signifies “Extremely more

importance”(Alexander, 2012). AHP incorporates an effective

technique for checking the consistency (Equation 1) of the



9426

evaluations made by the decision maker when building each

of the pairwise comparison matrices(R. W. Saaty, 2003).

Consistency ratio (CR) indicates the probability that the

matrix ratings were randomly generated(T. L. Saaty, 2013).

The random indices (RI) for the matrices are listed in Table 6

.The rule of thumb is that a CR less than or equal to 0.1

indicates an acceptable reciprocal matrix, while a ratio over

0.1 indicates that the matrix should be revised.

𝐶𝑅 =𝐶𝐼

𝑅𝐼 𝐶𝑅 =

𝐶𝐼

𝑅𝐼 𝑐𝑜𝑠𝑖𝑠𝑡𝑎𝑛𝑐𝑦 𝑟𝑎𝑡𝑖𝑜 1

Where CI= (λmax-n) / (n-1)

RI=Random Consistency Index

n=Number of Criteria.

λmax is the priority vector multiplied by each column total.

Normalization

Combination of spatial data requires standardization to avoid

conflicts that may arise from disparity in data format. Here,

statistical normalization technique which adopts linear

transformation scale (LTS) was utilized(Jiang et al., 2011).

This method proportionally transforms raw data into range of

values usually 1-9 (Table 3). The normalization process is

expressed as:𝑋

𝑋𝑖𝑗′ =

𝑋𝑖𝑗

𝑋𝑗𝑚𝑎𝑥

𝐿𝑇𝑆 𝑚𝑎𝑡𝑚𝑎𝑡𝑖𝑐𝑎𝑙 2

where, 𝑋𝑖𝑗′ = standardized score for ith and Jth objects and

attribute respectively, 𝑋𝑖𝑗 is the raw scor and 𝑋𝑗𝑚𝑎𝑥 is the

maximum score for the Jth attribute. Thus, the normalize value

is range from 0 to 1.

Model building: Data Conversion, overlay and validation

For data compatibility and ease of analysis, the data were

converted from vector to raster to facilitate selection of

criteria in GIS environment(ESRI, 2013). This operation was

followed by reclassification to match the ranking output of the

pairwise comparison result. Using the converted data, the

weightage criteria was used to develop a GIS-based model

(Figure 4) and implemented for the overlay analysis that

resulted to the land suitability map. With 105 ground control

points (GCP) recorded with Garmin handheld GPS during the

field work, the resulting map was validated using the receiver

operating characteristic’s (ROC) area under the ROC curve

measure

RESULTS AND DISCUSSION

Production Criteria inex map

The AHP output was used to generate data layers in GIS

environment representing suitability index map layers . of the

initail data preparation and processing produced the suitability

index map of the individual data layer (Figure 3).



9427

Figure 3 : Land quality index map (a) Rainfall (b) Dry-wet months (c) Soil series (d) Soil texture Soil texture (e) Drainage (f)

Elevation (g) Slope (h) soil productivity.

Wieghtedg evaluation and normlization

The process of pairwise comparison generates normalized

value for each critria for computational efficiency of the

consistency ratio. The final output is accepted or rejected by

assessing the consistency ratio value. For this study, the

consistency ratio of each factor are within the acceptable

threshold of not more than 10% (Table 7). The result of the

top level evaluation indicates that soil productivity (C1) is the

most important land quality for rubber suitability evaluation

while rainfall (C8) is the least important (Figure 6).

Table 6: Pairwise comparison for the main criteria

C1 C2 C2 C4 C5 C6 C7 C8 Priority Vector

C1 1 3 3 3 3 7 9 7 0.34

C2 1/3 1 2 2 2 5 7 5 0.19

C3 1/3 ½ 1 3 3 2 5 5 0.16

C4 1/3 ½ 1/3 1 3 5 3 3 0.13

C5 1/3 ½ 1/3 1/3 1 5 5 3 0.10

C6 1/7 1/5 ½ 1/5 1/5 1 3 3 0.05

C7 1/9 1/7 1/5 1/3 1/5 1/3 1 3 0.03

C8 1/7 1/5 1/3 1/3 1/3 1/3 1/2 1 0.04

3 6 8 10 13 26 27 36 1

Not: C1 = soil productivity, C2 = Soil texture, C3 = Soil series, C4 = Drainage, C5 = Slope, C6 = Elevation, C7 = Rainfall, C8 = Dry-wet months

Table 7: Consistency Index and ratio consistency for Criteria

Lambda CI CI/RI Consistency check

C1 7.791866 0.131978 0.099983 10%

C2 5.362868 0.090717 0.080997 8%

C3 4.252778 0.084259 0.093621 4%

C4 4.176679 0.084259 0.093621 7%

C5 4.255397 0.084259 0.093621 9%

C6 4.252778 0.084259 0.093621 4%

C7 3.070379 0.035185 0.060664 8%

C8 5.430224 0.107556 0.096032 10%

The results generated from the data layer are represented in

raster values which is not usable in the GIS environment for

layer combination. Not only that, each data layer has different

range of values, usually in decimal which is not appropriate

for weighted overlay. To bring them into common data

structure, the layer were standardized by a normalization

process. Weighted Overlay requires only numerical value, for

example 1 – 9 as used in this study. In addition to the data

structure, the layers are in varying dynamic range which

needed to be standardized. Data standardization and

normalization was done using statistical Linear

Transformation Scale (LTS). TLS was chosen because it has

the advantages of proportional transformation from the raw

data using Pairwise Comparison Matrix Table 8.



9428

Table 8: Criteria Spatial Value for Weighted Overlay Analysis and normilzation

Criteria Sub – criteria AHP value ‘x’ value

Normalized value ‘x’/ x max Multiplied value

n(x / x max),n = 9

Rainfall >230

138-168

198 -168

168 -138

<138

0.382

0.298

0.153

0.118

0.048

1

0.780

0.400

0.308

0.125

9

7

4

3

1

Dry-wet months Four months

Three months

Two months

One months

0.477

0.226

0.189

0.108

1

0.473

0.396

0.226

9

4

4

2

Soil Series Gajah Mati

Bungor

Jerangau

Rengam

Melaka

0.405

0.205

0.197

0.097

0.097

1

0.506

0.486

0.239

0.239

9

5

4

2

2

Soil Texture Sandy silty loam (SSiL)

silty loam (SiL)

sandy clay loam (SCL)

loam sandy (LS)

0.477

0.226

0.189

0.108

1

0.473

0.396

0.226

9

4

4

2

Drainage Well drained

Moderately well drained

Somewhat imperfectly

0.589

0.252

0.159

1

0.451

0.269

9

4

2

Slope <8

8 -16

16-30

>30

0.558

0.268

0.133

0.042

1

0.480

0.238

0.075

9

4

2

1

Elevation <200

200-400

400-600

>600

0.558

0.263

0.122

0.057

1

0.471

0.218

0.102

9

4

2

1

productivity >1350

1350-1250

1250-1150

1150-1050

1050-1000

0.455

0.283

0.104

0.112

0.046

1

0.621

0.228

0.246

0.101

9

6

2

2

1

Based on the previous data conversion and reclassification,

the overall process was then compiled in the ArcGIS model

builder for the weighted overlay analysis. ArcGIS model

builder offers a numbers of advantages particularly

progressive processing and easier database management

(Figure 4). In the weighted overlay analysis, the “scale

value” was given based on the raster value from the

reclassification process. These values are required to be

similar to avoid conflicts with the existing pixel values in the

raster format. Besides, overlaying raster data with more than

two layers is much more easier compared to vector data

because it does not include any topological operation

(Bagheri, Sulaiman, & Vaghefi, 2012). The process compares

the different criteria in each layer and their respective degree

of relevance to arrive at an objective suitability decision for

the entire area. Model builder is simple, visually discerning

with graphic architecture, provides easy database management

and scripting for repetitive tasks (Jankowski, 2015).



9429

Figure 4: Rubber land suitability evaluation model structure



9430

Figure 5: Rubber land suitability map of Seremban

The output of the model was reclassified into three suitability

classes: highly suitable, moderately suitable and marginally

suitable (Figure 5). From the figure, the most suitable land

area is in green colour, followed by the moderate class in

brown and the least suitable areas in red. It can be observed

that the southern part of the area is almost entirely suitable at

varying degrees of suitability. The highly suitable class

majorly lies in the south with isolated areas at the central part.

Similarly, the moderately suitable class is more dominant to

the east and central area. While the marginally suitable lands

falls in the extreme end of the southern area and marginally at

the west. The high elevation areas in the east and north are

totally evaluated unsuitable for rubber cultivation. This is not

unusual because, most of the indices of the individual layer

indicate that those particular parts of the district are not

suitable , Quantitative evaluation gives the total area of the

three classes to be 35575 hectares distributed thus (Table 9):

highly suitable (16048 hectares, (45%)), moderately suitable

(15399 hectares (43%)) and marginally suitable (4128

hectares (12%)).

Table 9: Distribution of Suitable Areas

Three sub-districts, Setul, Lengeng and Seremban at the north

and central area are almost entirely found to be unsuitable for

rubber cultivation as opposed to the other sub-districts.

Particularly Seremban district is void because it is almost

entirely developed with anthropogenic interference in the

physical landscape, drainage and surface moisture. Labu

district is has all the classes of suitability but lacking

suitability potential in the northern part, the same scenario is

witnessed in Pantai district (both of which fall at the central

zone. This study shows that Rasa, Rantau and Ampangan

Type Hectare Perc. (%)

highly suitable

moderately suitable

marginally suitable

16048

15399

4128

45

43

12

Total 35575 100



9431

districts have the most suitable land for rubber cultivation.

Result of the model accuracy assessment is presented in

Figure 6 and Table 10 The graph of sensitivity-specificity

plots provide visual and statistical understanding of how well

the model formed. Interactively, the sensitivity measures how

good the model is at defining suitable land (the true positive),

while the specificity indicates ability of the classification to

correctly identify unsuitable land (true negative) The analysis

shows that the model is sensitive to detection suitable land for

rubber cultivation and detecting non-suitable land with

sensitivity and specificity values of 84.14% and 76%

respectively . Overall, area under the ROC curve of 0.80

(80%) and p-value of <0.0001 was achieved at 95%

confidence interval. Since the P is far less than 0.05, it

indicates that the area under the ROC curve is significantly

different from 0.5 (the null hypothesis); meaning that the

model successfully differentiates between suitable and

unsuitable lands for rubber farming in Seremban with high

level of accuracy.

Figure 6: Graph of the ROC curve analysis

Table 10: Statistical AUC assessment parameters

Assessment parameter Measure

Sensitivity 84.14

Specificity 76.00

Area under the ROC curve (AUC) 0.80

Standard Error 0.03

95% Confidence interval 0.761 to 0.837

Significance level P (Area=0.5) <0.0001

CONCLUSION

The GIS-based Multi-Criteria evaluation technique results to

RLSM, a geospatial model to identify and categorize land

suitable for rubber cultivation based on land qualities and crop

growth requirements. The technique produced objective

assessment of the spatial distribution and quantity of the

suitability classes taking into consideration soil and rubber

growth requirements. It is observed that suitable lands are

more concentrated in the southern part of Seremban district.

Most of the existing rubber plantations are not located on

optimal productive land which may account for low

productivity. This RLSM can be applied in other parts of the

country with consistent result for comprehensive nation-wide

evaluation to allow decision makers and farmers select

appropriate site for rubber farming and equally estimate

potential yield.

REFERENCES

[1] Ahmad Fariz, M., Wan Yaacob, W. Z., Mohd Raihan,

T., & Abdul Rahim, S. (2009). Groundwater and Soil

Vulnerability in the Langat Basin Malaysia. European Journal of Scientific Research, 27(4), 628–635.

https://doi.org/ISSN 1450-216X

[2] Alanne, K., Salo, A., Saari, A., & Gustafsson, S.-I.

(2007). Multi-criteria evaluation of residential energy

supply systems. Energy and Buildings, 39, 1218–1226.

https://doi.org/10.1016/j.enbuild.2007.01.009

[3] Alexander, M. (2012). Decision-Making using the

Analytic Hierarchy Process (AHP) and SAS/ IML. The United States Social Security Administration Baltimore,

1–12.

[4] Bagheri, M., Sulaiman, W. N. A., & Vaghefi, N.

(2012). Land Use Suitability Analysis Using Multi

Criteria Decision Analysis Method for Coastal

Management and Planning: A Case Study of Malaysia.

Journal of Environmental Science and Technology.

https://doi.org/10.3923/jest.2012.364.372

[5] Camerlengo, A. L., & Somchit, N. (2000). Monthly and

annual rainfall variability in Peninsular Malaysia.

Pertanika Journal Science and Technology, 8(1), 73–

83.

[6] Department of Statistics Malaysia. (2015a). Current

population estimates, Malaysia 2014-2016.

https://doi.org/10.1017/CBO9781107415324.004

[7] Department of Statistics Malaysia. (2015b).

Department of Statistics Malaysia Official Portal.

Department of Statistics, Malaysia, (November), 1–5.

https://doi.org/2016-08-11

[8] Department of Statistics Malaysia. (2016). Current Population Estimates, Malaysia, 2014-2016



9432

POPULATION. Department of Statistics Malaysia.

https://doi.org/10.1017/CBO9781107415324.004

[9] Eastman, J. (1999). Multi-criteria evaluation and GIS.

Geographical Information Systems, 493–502.

[10] Economic Planning Unit. (2015). Eleventh Malaysia Plan : Anchoring Growth on People. Rancangan Malaysia Kesebelas (Eleventh Malaysia Plan) : 2016-2020. https://doi.org/10.1017/CBO9781107415324.004

[11] ESRI. (2013). ArcGIS Desktop: Release 10.2. Redlands CA.

[12] Gasim, M. B., Ismail, B. S., Mir, S. I., Rahim, S. A., &

Toriman, M. E. (2011). The physico-chemical

properties of four soil series in Tasik Chini, Pahang,

Malaysia. Asian Journal of Earth Sciences, 4(2), 75–

84. https://doi.org/10.3923/ajes.2011.75.84

[13] Govindan, K., Rajendran, S., Sarkis, J., & Murugesan,

P. (2015). Multi criteria decision making approaches

for green supplier evaluation and selection: A literature

review. Journal of Cleaner Production, 98, 66–83.

https://doi.org/10.1016/j.jclepro.2013.06.046

[14] Hambirao, G. S., & Dr. P.G. Rakaraddi (Basaveshwar

Engineering College, Bagalkot, I. (2014). Soil

Stabilization Using Waste Shredded Rubber Tyre

Chips. IOSR Journal of Mechanical and Civil Engineering (IOSR-JMCE), 11(1), 20–27.

[15] Hedges, L., Lam, W. Y., Campos-Arceiz, A., Rayan, D.

M., Laurance, W. F., Latham, C. J., … Clements, G. R.

(2015). Melanistic leopards reveal their spots: Infrared

camera traps provide a population density estimate of

leopards in malaysia. Journal of Wildlife Management, 79(5), 846–853. https://doi.org/10.1002/jwmg.901

[16] Ho, W., Xu, X., & Dey, P. K. (2010). Multi-criteria

decision making approaches for supplier evaluation and

selection: A literature review. European Journal of Operational Research, 202(1), 16–24.

https://doi.org/10.1016/j.ejor.2009.05.009

[17] Holthuis, L. B. (1976). FAO Species Catalogue - Vol.1

- Shrimps and Prawns of the World. FAO Fisheries Synopsis No. 125, Volume 1, 1(125), 38–50.

https://doi.org/10.1111/j.1365-2664.2009.01751.x

[18] Howell, C. J., Schwabe, K. A., Haji Abu Samah, A.,

Graham, R. C., & Taib, N. I. (2005). Assessment of

Aboriginal Smallholder Soils for Rubber Growth in

Peninsular Malaysia. Soil Science, 170(12), 1034–1049.

https://doi.org/10.1097/01.ss.0000187346.30201.40

[19] Jankowski, P. (2015). International Journal of

Geographical Information Systems Integrating

geographical information systems and multiple criteria

decision-making methods Integrating geographical

information systems and multiple criteria decision-

making methods. https://doi.org/10.1080/026937

99508902036

[20] Jiang, H., Yu, S. X., & Martin, D. R. (2011). Linear

scale and rotation invariant matching. IEEE Transactions on Pattern Analysis and Machine Intelligence, 33(7), 1339–1355.

https://doi.org/10.1109/TPAMI.2010.212

[21] LabourforceSurvey. (2015). Department of Statistics

Malaysia Press Release. Department of Statistics Malaysia, (June), 5–9.

https://doi.org/10.1017/CBO9781107415324.004

[22] Lin, X.-H., Chen, Q.-B., Hua, Y.-G., Yang, L.-F., &

Wang, Z.-H. (2011). Soil moisture content and fine root

biomass of rubber tree (Hevea brasiliensis) plantations

at different ages. Ying Yong Sheng Tai Xue Bao = The Journal of Applied Ecology / Zhongguo Sheng Tai Xue Xue Hui, Zhongguo Ke Xue Yuan Shenyang Ying Yong Sheng Tai Yan Jiu Suo Zhu Ban, 22(2), 331–336.

[23] Lintner, B. R., Biasutti, M., Diffenbaugh, N. S., Lee, J.

E., Niznik, M. J., & Findell, K. L. (2012).

Amplification of wet and dry month occurrence over

tropical land regions in response to global warming.

Journal of Geophysical Research Atmospheres,

117(11). https://doi.org/10.1029/2012JD017499

[24] Macharis, C., & Bernardini, A. (2015). Reviewing the

use of multi-criteria decision analysis for the evaluation

of transport projects: Time for a multi-actor approach.

Transport Policy, 37, 177–186. https://doi.org

/10.1016/j.tranpol.2014.11.002

[25] Malaysia Digital Association. (2016). Digital

Landscape Exploring the Digital Landscape in

Malaysia-. 2016 Malaysia Digital Landscape,

(August), 1–47.

[26] Malaysian Meteorological Department. (2009). Climate Change Scenarios For Malaysia (2001 - 2099). Malaysian Meteorological Department (Vol. January,

S). https://doi.org/978-983-99679-1-3

[27] Manap, M. A., Ramli, M. F., Azmin Sulaiman, W. N.,

& Surip, N. (2010). Application of remote sensing in

the identification of the geological terrain features in

Cameron highlands, Malaysia. Sains Malaysiana,

39(1), 1–11.

[28] Melling, L., Hatano, R., & Goh, K. J. (2005). Soil CO2

flux from tree ecosystems in tropical peatland of

Sarawk, Malysia. Tellus, 57B, 1–11.

[29] Paramananthan, S., & Zauyah, S. (1986). Soil

landscapes in Peninsular Malaysia. Geol. Soc. Malaysia, Bulletin, 9(April), 565–583.

[30] Saaty, R. W. (2003). The Analytic Hierarchy Process

(AHP). Super Decisions. https://doi.org/10.3414/



9433

ME10-01-0028

[31] Saaty, T. (1977). [Saaty, 1977]. Journal of Mathematical Psychology, 15, 234–281.

[32] Saaty, T. l. (2003). Saaty T.L. (2003), The Analytic

Hierarchy Process (AHP) for Decision Making and the

Analytic Network Process (ANP) for Decision Making

with Dependence and Feedback. Creative Decisions Foundation, 114.

[33] Saaty, T. L. (2004). Decision making — the Analytic

Hierarchy and Network Processes (AHP/ANP). Journal of Systems Science and Systems Engineering, 13(1), 1–

35. https://doi.org/10.1007/s11518-006-0151-5

[34] Saaty, T. L. (2013). The Modern Science of

Multicriteria Decision Making and Its Practical

Applications: The AHP/ANP Approach. Operations Research, 61(5), 1101–1118. https://doi.org/10.1287

/opre.2013.1197

[35] Sambasivan, M., & Fei, N. Y. (2008). Evaluation of

critical success factors of implementation of ISO 14001

using analytic hierarchy process (AHP): a case study

from Malaysia. Journal of Cleaner Production, 16(13),

1424–1433.

https://doi.org/10.1016/j.jclepro.2007.08.003

[36] Sola, I., Gonz??lez-Aud??cana, M., & ??lvarez-Mozos,

J. (2016). Multi-criteria evaluation of topographic

correction methods. Remote Sensing of Environment, 184, 247–262. https://doi.org/10.1016/j.rse.2016.07.002

[37] Streutker, D. R., Glenn, N. F., & Shrestha, R. (2011). A

Slope-based Method for Matching Elevation Surfaces.

Photogrammetric Engineering & Remote Sensing,

77(7), 743–750. https://doi.org/10.14358/PERS.77.7

.743

[38] Sujaul, I. M., Muhammad Barzani, G., Ismail, B. S.,

Sahibin, A. R., & Mohd Ekhwan, T. (2012). Estimation

of the Rate of Soil Erosion in the Tasik Chini

Catchment, Malaysia using the RUSLE Model

Integrated with the GIS. Australian Journal of Basic and Applied Sciences, 6(12), 286–296.

[39] Vaidya, O., & Hudnurkar, M. (2013). Multi-criteria

supply chain performance evaluation. International Journal of Productivity and Performance Management, 62(3), 293–316. https://doi.org/10.1108/17410

401311309195

GIS-Based Land Suitability Mapping for Rubber Cultivation ... · land evaluation techniques (Thomas...

Documents

Transcript of GIS-Based Land Suitability Mapping for Rubber Cultivation ... · land evaluation techniques (Thomas...