Table of Contents

MODELLING & EARLY WARNING SYSTEM (MEWS)

Emergency response analysis for humanitarian logistics ............................................................................. 3
Control of turbidity currents to reduce reservoir sedimentation using obstacles .......................................... 6
Tropical flood estimation model derived from terminal doppler weather radar information ........................ 10
Disaster safety net: evaluating MKN20 directive ........................................................................................ 13
Flood hazard assessment using multi criteria evaluation (MCE) method in Penampang area, Sabah...... 17
Aplikasi media sosial sebagai sistem amaran awal bencana banjir kepada penduduk dalam
lembangan Sungai Kelantan ....................................................................................................................... 22
Emergency and response planning (ERP): simulation on coordination of inter agency in flood
catastrophic event ....................................................................................................................................... 26
Underwater ground mapping for flood disaster using ultrasonic sensor ..................................................... 29
Rapid assessment method of flood damage using spatial-statistical models ............................................. 32
A feasibility study of a disaster management hub using predictive-analytics and modelling technique
based on crowdsourced data ...................................................................................................................... 45
Resq banjir: a mobile apps for emergency rescue, evacuation and relief center management ................. 47
Sistem dan aplikasi mobile pantas amaran bencana banjir ........................................................................ 50
Developing long-range, auto rechargeable radio sensor network .............................................................. 53
Integrated river basin monitoring and flood warning system ...................................................................... 57
Inspire: interactive dam safety decision support system for flood disaster reduction ................................. 63
Integrated mobile solution for alert, search & rescue (A-SAR) ................................................................... 72
Modeling the effect of land use/cover changes in flood source areas on downstream flood peak of
Kelantan river basin .................................................................................................................................... 76
Flood water level prediction modeling using NNARX structure for Sg Pahang basin ................................ 81
A coupled hydrologic-hydraulic model for simulation of flood inundation extent and depth in Kelantan
river basin .................................................................................................................................................... 86
Formulation of a trust evaluation algorithm using trust model for peer communications to improve
information reliability during the disaster environment ................................................................................ 96
A community-based study on the effectiveness of flood emergency warning system in Malaysia ............ 99
Flood occurrence reduction measures by runoff prediction based on land use scenario analysis using
SCS-CN method and GIS for Kelantan river basin ................................................................................... 103
Long term analysis of tidal range, high tides and selected estuarine water level in east coast of
Malaysia .................................................................................................................................................... 107
Flood hazard map utilizing public domain inundation hydrological (RRI) and hydraulic (HECRAS) models
and GIS ..................................................................................................................................................... 112
Modeling flood hydrograph in Sg Kelantan river basin using functional concept ..................................... 117
Hydrodynamic simulation of Kelantan river bank erosion and channel change during heavy floods ....... 122
Determination of floodplain chain and its impact on flood modelling using multi-temporal dem and
land use/land cover from insar/sar technique………………………………………………………………………….Lampiran 1
Significance of seasonal river mouth sandbar formation to flood levels in the Kelantan river

1
basin………………………………………………………………………………………………………………………………………..Lampiran 2
Urbanization and industrialization: a teaching learning based optimization algorithm for efficient
routing of the emergency flood evacuation process…………………………………………………………………Lampiran 3
The potential of parameter estimation through regionalization for flood simulations in ungauged
mesoscale catchments……………………………………………………………………………………………………………..Lampiran 4

2
 MODELLING & EARLY WARNING SYSTEM (MEWS)

EMERGENCY RESPONSE ANALYSIS FOR HUMANITARIAN LOGISTICS

Project Information
Project Leader : A/P. Dr. Joshua Ignatius
University : Universiti Sains Malaysia
Address : 11800 Minden, Penang, Malaysia
Contact number : 04-6534769
Email : josh@usm.my
Project Members : Dr. Wong Wai Peng
A/P. Adam Baharum
Dr. Siti Amirah Abd Rahman

1.0 Introduction
Weather-related disasters are becoming increasingly frequent, due largely to a sustained rise in the
number of flood and storms. Flooding alone accounted for 47% of all weather related disasters from the
year 1995 to the year 2015, affecting 2.3 billion people, with majority of 95% who live in Asia. In Malaysia,
there are four main types of disasters, which are landslide, hurricane/strong winds, flood and flash flood.
Among the four disasters, flood has the highest occurrence from the year 2010 to the year 2015. In the
state of Kelantan on the east coast of Peninsular Malaysia, flood is an annual occurrence of varying
severities. Malaysia’s National Security Council (NSC) confirmed that the flood in 2014 was the worst
recorded in the history of the state. It was described as a ‘tsunami-like disaster’. Due to the annual
occurrence of the flood and its disastrous damage in Kelantan, our scope of this paper is on Kelantan
flood in 2014. We chose Kuala Krai as our case study due to its high population density and it was hit by
the flood twice. Our focus is on managing disaster from the aspect of food delivery as flood directly
affects the food supply by destroying its storage and infrastructure.

2.0 Methodology
This research uses the systems dynamics methodology. We collected data through several state
government agencies in Kelantan because most of the data access applications are not available through
online and each state government agency is responsible for different kinds of emergency responses.
Malaysian National Security Council (MKN) is the leader for the series of emergency responses. Social
Welfare Department (JKM) is responsible for the placement of evacuation centres, the registration of
evacuees in each evacuation centre and distribution of food. Federal Development Department of
Kelantan (JPP) is responsible for the road infrastructure quality including the bridges. Economic Planning
Unit (UPEN) is responsible for disaster relief efforts such as economy recovery and repair of road
infrastructure. Health campus of Universiti Sains Malaysia Kubang Kerian (USMKK) is one of the
evacuation centres and the only functional hospital during the Kelantan flood in 2014. Therefore, we
approached two departments of USMKK which are Pertubuhan Gabungan Bantuan Bencana NGO
Malaysia (BBNGO) and Hal Ehwal Pelajar & Alumni (HEPA) for the collection of data. In addition, we
approached some non-governmental organization (NGO) such as Malaysian Red Crescent Society
(MRCS) as they provided assistance in distributing food to victims. In order to simulate the model, we
construct the stock and flow diagram as follows:

3
 Figure 1 Stock and Flow Diagram of the Food Delivery System

3.0 Results and Discussion

After simulation of model, we tabulate the values and compare the graph of the stock variables, which are
Warehouse, Mini Storehouse, Bases, Stores in Evacuation Centers and Food to Victims as Figure 2.

Figure 2 Stock Level Comparisons

For the food delivery rate and food distribution rate, we tabulate the values and plot the graphs for
comparison purposes as in Table 2.

Table 2 Level of Each Delivery/Distribution Rate

Food Delivery Food Delivery Food Delivery Rate Food Distribution
Time (Day)
Rate Rate 2 3 Rate
0 46.3554 46.3554 46.3554 154.518
1 53.7921 53.8671 53.8671 156.018
2 54.24 54.39 54.39 157.518
3 54.6879 54.9129 54.9129 159.018
4 55.1358 55.4358 55.4358 160.518
5 55.2117 55.5367 55.5367 161.018
6 55.2876 55.6376 55.6376 161.518

4
 7 55.3635 55.7385 55.7385 162.018
8 55.4614 55.8614 55.8614 162.518
9 55.5593 55.9843 55.9843 163.018
10 55.6572 56.1072 56.1072 163.518
11 55.7632 56.2382 56.2382 164.018
12 55.8692 56.3692 56.3692 164.518
13 55.8502 56.3502 56.3502 164.518
14 56.2311 56.7311 56.7311 164.518
15 56.612 57.112 57.112 164.518
16 56.9929 57.4929 57.4929 164.518
17 729.249 1041.27 1041.27 25.8814
18 10883.3 2525.94 2525.94 26.1414
19 19288.9 1211.5 1211.5 26.4014
20 30328.1 50.4301 50.4301 26.6614
21 31560.8 1693.97 1693.97 26.8564
22 57456.3 3491.13 3491.13 27.0514
23 90188.5 4103.71 4103.71 27.2464
24 136472 4716.22 4716.22 27.4414
25 203572 4706.24 4706.24 27.8964
26 302948 4704.83 4704.83 185.518
27 450662 4627.61 4627.61 189.018
28 670943 4548.99 4548.99 192.518
29 1.00067e+006 6385.07 6385.07 195.518
30 1.49127e+006 8219.95 8219.95 198.518
31 2.22155e+006 10053.9 10053.9 202.518

4.0 Conclusion

4.1 Capacity of evacuation centers do not match the population of victims.

4.2 Food distribution and aid comes in spikes. It is either too much or too little throughout the
affected period.
4.3 Standard operating procedures of emergency response assumes that the road
infrastructure is still in working condition.
4.4 Centralized kitchen concept was not realized earlier.
4.5 Lack of coordination due to inability to forecast between Water Level  Road Conditions
 Food Distribution Rate.

5
 CONTROL OF TURBIDITY CURRENTS TO REDUCE RESERVOIR SEDIMENTATION USING
OBSTACLES

Project Information
Project Leader : Dr Mohamad Hidayat bin Jamal
University : Universiti Teknologi Malaysia (UTM)
Address : Department of Hydraulics and Hydrology, Faculty of Civil Engineering,
81310 Johor Bahru, Johor
Contact number : 0196582880
Email : mhidayat@utm.my
Project Members : Prof Ahmad Khairi Bin Abd Wahab, Dr Zulhilmi Bin Ismail
Dr Nazri Ali, PM Dr. Mohd Fadhil Md Din
Dr Zulkiflee Ibrahim, Dr Ilya Khairanis Othman
Nurfarhain Binti Mohamed Rusli

1.0 Introduction
Density currents are generated when fluid of one density is released into fluid of a different density
(Marino et al., 2005). The density difference can be resulted from temperature gradients, dissolved
contents, suspended particles or a combination of them. The currents are known as turbidity currents in
case the main driving mechanism is obtained from suspended sediments (Simpson, 1999).
Sediment discharge of rivers flowing into reservoirs is typically very high during flood events. As
the turbid flood flows to fresh water of the reservoir, the turbid inflow displaces the ambient water until it
reaches a balance of forces and plunges under the water surface (Oehy and Schleiss, 2007), as seen in
Figure 1. As the sediments accumulate, reservoir loses its storage volume which leads to elimination of
flood and energy regulation capacity. Sediments can also damage dam power plants and block bottom
outlets, thereby decreasing efficiency and increasing maintenance costs (Cesare et al., 2001).

Figure 1:. Turbid inflow entering a reservoir

The leading edge of density currents is called head (also known as front) which is deeper than
the following flow (i.e. body) and has a raised nose at its foremost point. The schematic of a density
current propagating over a sloping bed and under a layer of stationary ambient fluid having a density (ρa)
less than that of the density current (ρd) illustrated in Figure 2. The highest point of the front is known as
its height (Hf) travelling with the velocity of Uf. For the body, the height and velocity are shown with h and
, respectively.

6
 Figure 2:. Sketch of a density current advancing over a slope

The loss of storage capacity in dam reservoirs due to sedimentation caused by turbidity currents
has been an issue of great concern and a topic of research (Fan and Morris, 1992;Kostic and Parker,
2003;Khavasi et al., 2012). Nevertheless, there has not been a substantial amount of work in the area of
controlling these currents with arrays of obstacles, in particular, regarding the continuous part (i.e. body)
of density currents. This research aims to develop a comparison between density currents dynamics over
smooth bed and surfaces covered with arrays of obstacles. Also, the influence of different configurations
of obstacles on the velocity structure of density current is investigated herein.

2.0 Methodology
To model density currents in the laboratory, a specific setting is required that could prepare dense fluids
and maintain the steady state of density currents during the course of experiments. An appropriate
experimental apparatus was prepared consisting of five main parts: water supply system, mixing tanks,
head tank, flume and drainage system as seen in Figure 3.

Figure 3:. Experimental set up

The flume was 10 m long, 0.3 m wide and 0.7 m deep. A sliding vertical gate divided the flume to
two sections of unequal length. Upstream of the gate was filled with dense fluid and the long downstream
section simulates a reservoir (see Figure 4) where a density current was propagating. Salt was dissolved
in tap water inside the tanks until the required density was obtained and the solution was homogeneously
mixed.
The dense flow discharge could be adjusted by the means of a valve and using an elec-
tromagnetic flow meter prior entering the flume. The experiments started with the sudden removal of the
gate. The gate was opened 7 cm in all experiments. Experiments were performed with two discharges
(Q= 0.5 and 1 L/s), having different bottom slopes (S=0.25%, 1% and 1.75%) and initial concentrations
(Cin=5, 15 and 25 gr/lit). Rough beds (see Figure 5) were made of square beams of height (D) 1.2 cm
perpendicular to the flow direction. Different spacing between the beams were chosen (i.e. λ=1.2, 4.8,
9.6, 19.2, 38.4, 76.8 and 153.6 cm) yielding seven rough beds having λ/D= 1, 4, 8, 16, 32, 64 and 128.
A Nortek Acoustic Doppler Velocimeter (ADV) with 10 MHZ acoustic frequency was used to
record the velocity profiles in the body of density currents at three locations (X= 3 m, 4 and 5 m from the

7
gate) along the centerline of the channel. Samples with SNR values less than 15 dB and correlation less
than 70% were filtered.

Figure 4: Density current flowing over smooth bed

Figure 5:. A rough bed (λ/D =8)

3.0 Results and Discussion

The main aim of this study is to investigate the influence of roughness elements on controlling density
currents. Therefore, density currents flowing over roughness elements with various spacing were tested
and compared. Stopping turbidity currents in a reservoir or influencing them in a way that the sediments
are not deposited in important zones (e.g. in front of water intake structures and bottom outlets) increases
the sustainability of reservoir operation significantly (Bühler et al., 2012).
Non-dimensional numbers are used to determine density current regimes. The densimetric
0.5
Froude Number (Frin) is defined as Frin=uin/(g’hin cosθ) where uin is the inlet velocity, hin is the inlet
opening height, g’ is reduced gravitational acceleration and  is the bottom slope. The reduced
gravitational acceleration (g’) is g’=g(ρd-ρa)/ρa where g is the gravitational acceleration. In our experiments
0.29<Frin<0.97. The inlet Reynolds number in these currents is defined as Rein=uin hin/υ where υ is the
kinematic viscosity of inlet mixture. This can be rewritten as Rein=Qin/b υ where the b is the flume width.
Hence, in our experiments, the inlet Reynolds Number for Qin=0.5 lit/s and Qin= 1 lit/s was equal to 1956
and 3912, respectively.The vertical velocity profiles were collected in laboratory experiments to
investigate velocity structure in the body of saline underflows. The typical velocity profiles of density
currents has an inner (wall) and outer (jet) region as seen in Figure 6 (Smooth bed, X= 3 m, S=1%, Qin= 1
lit/s, Cin=25 gr/lit) where u is the velocity of the current at the depth of z. These regions are separated by
maximum velocity (um). The distance above the bed where the maximum velocity occurs is referred to as
hm. In the inner region, the flow is mainly controlled by rigid boundary and bottom friction is the main
influential parameter (Khavasi et al., 2012). The outer region is related to the interface or diffusion
boundary at the top of velocity maximum.

Figure 6:. Characteristics of the collected velocity profiles

The velocity profiles for selected number of experiments with the same initial conditions are
illustrated in Figure 7 to discuss the influence of different rough beds. Rough beds caused two different
behaviours in velocity profiles of the currents. For λ/D =1, 4, 8, 16; the retardation of the density current

8
increased with the increase in the beams spacing. For these rough beds, increasing the spacing between
elements increased hm. This general rule was held up to a point λ/D =16. Density current propagating
over λ/D =16 had the least velocity and maximum hm.For λ/D =32, 64, 128; the controlling effect of the
rough beds was reduced and hence the maximum velocity of the currents increased as the spacing
between the beams became more. For these rough beds, increasing the spacing between elements also
decreased hm.
The maximum velocities for λ/D =32, λ/D =64 and λ/D =128 were similar to that of λ/D =8, λ/D =4,
λ/D =1, respectively. However, the depths where the maximum velocities happened were less than
them.In our experiments, there was a critical spacing (λ/D =16) above which increasing the distance
between the elements had little effect on controlling the velocity of the currents. When the elements
became too far apart (λ/D =128) the flow dynamics became similar to that over a smooth bed with the
highly dispersed elements representing very small individual obstacles in the path of the current.

Figure 7:. Velocity profiles of density currents flowing over smooth and rough beds
(S=1.75%, Cin=5gr/lit, X=4 m, Qin=1 lit/s)

4.0 Conclusion
In dam reservoirs, flood-induced turbidity currents are the main driving mechanism for sediment transport.
This study investigates controlling of turbidity currents with arrays of roughness elements. The
conclusions of this study are as follows:

4.1 For density currents flowing over rough beds with λ/D =1, 4, 8, 16; increasing the spacing
between the beams decreased the velocity of the currents.
4.2 The maximum retardation of density currents occurred for the currents propagating over
the rough bed with λ/D=16.
4.3 For density currents travelling over rough beds having λ/D =32, 64, 128, the controlling
effect of the rough beds was reduced and hence the velocity of the currents increased as
the spacing between the beams became more.
4.4 Density currents had a velocity similar to that of the smooth bed when flowing over the
rough bed with λ/D=128.

References
Bühler, J., C. Oehy and A. Schleiss (2012). "Jets Opposing Turbidity Currents and Open Channel Flows." Journal of
Hydraulic Engineering 139(1): 55-59
Cesare, G. D., A. Schleiss and F. Hermann (2001). "Impact of turbidity currents on reservoir
sedimentation." Journal of Hydraulic Engineering 127(1): 6-16
Khavasi, E., H. Afshin and B. Firoozabadi (2012). "Effect of selected parameters on the depositional behaviour of
turbidity currents." Journal of Hydraulic Research 50(1): 60-69.
Kostic, S. and G. Parker (2003). "Progradational sand-mud deltas in lakes and reservoirs. Part 2. Experiment and
numerical simulation." Journal of Hydraulic Research 41(2): 141-152.
Marino, B., L. Thomas and P. Linden (2005). "The front condition for gravity currents." Journal of Fluid Mechanics
536: 49-78

9
 TROPICAL FLOOD ESTIMATION MODEL DERIVED FROM TERMINAL DOPPLER WEATHER
RADAR INFORMATION

Project Information
Project Leader : Assoc Prof Ir Dr Ahmad Fadzil Ismail
University : Universiti Islam Antarabangsa Malaysia (UIAM)
Address : Department of Electrical and Computer Engineering
Contact number : 0123283175
Email : af_ismail@iium.edu.my
Project Members : Dr Farah Nadia Mohd Isa
Dr Wahidah Hashim

1.0 Introduction
Floods are among the most frequent and costliest natural disasters. Conditions that cause floods include
heavy or long-steady rain for several hours or days that saturates the ground. Since the long term
precipitation forecast is still not reliable enough, an accurate estimates degree of the extremity for
upcoming flood events that might cause dangerous meteorological situations. The information from a rain
gauge and radar data could be useful for decision maker as an additional information for flood warning
system. This type of flood estimating technique is one of the techniques derived from an algorithm
generated from rainfall rate, horizontal and vertical profile of radar reflectivity values. This algorithm was
developed to estimate flood phenomenon derived from rain gauges and weather radar. The rainfall rate,
cloud thickness value, the size of the cloud during the flood disaster was measured. In this research the
rain gauge data and radar data on the duration of time over the flood area covered by meteorological
radar and rain gauge was analyzed. The specific time can be the whole duration for the rain event before,
during and after the flood tragedy. The procedure was applied to 14 days precipitation phenomenon
observed in Kota Bharu, Kelantan (Malaysia) from 13 December 2014 until 26 December 2014 and was
validate using the precipitation phenomenon observed in Kuantan, Pahang (Malaysia) during the extreme
flood tragedy in December 2013. The derived algorithm acquired in this research is very useful to forecast
the flood tragedy in the future and as a development model to be integrated into the radar system.

2.0 Methodology
The following assignments have to be performed in order to achieve the aforementioned objectives. In
accordance with that purpose, it is anticipated that the radar data for the duration of at least 2 weeks will
be acquired from Malaysia Meteorological Department (MMD). This duration is chosen to consider the
precipitation activity before, during, and after the flood incident. This information about rainfall activity is
important in order to identify relevant parameter of cloud characteristics that cause flood to occur.
Relevant weather radar and available rain gauge data will be acquired (procured to be exact) accordingly.
Rain gauge data is measured in mm/hr. All acquired radar data must first be converted into a readable
format. The reason is that the radar data is in raw data format. This conversion must go through special
software identified as IRIS that is used specifically to process the raw radar data. The rainfall rate during
the extreme rainfall event is measured from the rain gauges. The size and the thickness of the cloud is
identified from the radar data.
Subsequent task will then involve the development of algorithms capable of making forecasts of
spatial rainfall for flood forecasting using the stated characteristics. The parameter of cloud
characteristics, including the thickness and the size of the cloud together with the rainfall rate must be
compared and analysed before the new proposed model capable of predicting flood in Malaysia be
verified. The radar derived proposed algorithms of flood model will be assessed using field data.
In order to achieve the third objective, which is to validate the development model that is
expected to be integrated into radar system. The proposed algorithm must be validated at other tropical
flood event. Radar data for another flood incidents during the year 2013 at Kuantan is being utilized. This
information justify and support the new proposed model. The flow process for Derivation Rainfall rate and
cloud size is shown in Figure 1.1.
The phases of the work that will be involved in this proposed project are illustrated in Figure 1.2
below.

10
 START

PRELIMINARY INVESTIGATION

PROBLEM ANALYSIS

REQUIREMENT ANALYSIS

DECISION ANALYSIS

FORMULATION

VALIDATION

END

Figure 1.1 Flow Process for Derivation Rainfall Rate and Cloud Size

Quantify the
Acquirement for Compilation of Analyse the size and
rainfall rate from Generate the
rainfall and radar rainfall rate thickness of the
the rain gauge
data values and radar cloud from the radar algorithm
during the flood
reflectivity reflectivity
event

Figure 1.2 Flowchart for Methodology

3.0 Results and Discussion

The final model of the analysis would be,

Y’ = B0 + B1X1 + B2X2 + B3X3

-5
Y’ = -1.43 + 0.06X1 + 5.28x10 X2 + 0.0044 X3

Figure 3.1 Tested model

11
From the tested algorithm result, it shows that the cloud appears during the flood incident is the cloud that
can lead to flood disaster and the percentage for the flood to happen be more compared to the flood
incident during the year 2014. As display in the Figure 3.2, the CAPPI view at 7.6 km height level and
200 km range distance for the cloud observed in Kuantan during the flood incident. The cloud is
2
considered as absolutely huge about 60560.5 km .

Figure 3.2 CAPPI View

4.0 Conclusion

4.1 The parameter of the cloud characteristics such as size and the thickness along with the
rainfall rate during the Malaysian flood disaster 2014 was able to identify and
characterize
4.2 Developing the coefficient that applicable to predict any rainfall activity that capable to
cause flood disaster.
4.3 Algorithm modelled was capable to be integrated into radar system

References
Basri, A.B., Ismail, A.F., Khairolanuar, M.H. & Badron, K. (2016). Analyses of Meteorology Information during
Malaysian Flood Disaster 2014. Advanced Science Letter - Scopus, (Accepted for publication).
Basri, A.B., Ismail, A.F., Khairolanuar, M.H. & Badron, K. (2016). Analyses of Cloud Characteristic during Malaysian
2014 Flood Event. Indian Journal of Science and Technology - Scopus, (Accepted for publication).
Basri, A.B., Ismail, A.F., Khairolanuar, M.H. & Badron, K. (2016). Analyses of Rainfall Rate during Malaysian 2014
Flood Event. International Journal of Software Engineering and Its Applications - Scopus, (Accepted for
publication).
Bertoldo, S., Lucianaz, C., & Allegretti, M. (2015). Extreme rainfall event analysis using rain gauges in a variety of
geographical situations.Atmospheric And Climate Science, 5(2), 82-90.
Beven, K. J., & Carling, P. (1989). Floods: hydrological, sedimentological and geomorphological implications. John
Wiley and Sons Ltd.

12
 DISASTER SAFETY NET: EVALUATING MKN20 DIRECTIVE

Project Information
Project Leader : Roslina binti Kamaruddin
University : Universiti Utara Malaysia
Address : 181 Bangunan Ekonomi, Pusat Pengajian Ekonomi, Kewangan dan Perbankan,
Kolej Perniagaan, Universiti Utara Malaysia
Contact number : 0195110977
Email : roslina_k@uum.edu.my
Project Members : Dr Soon Jan Jan
Prof Madya Abdul Rahim Anuar
Dr Siti Aznor Hj Ahmad
Dr Rabiul Islam

1.0 Introduction
The 2014 flood was the worst in the history of flood in Malaysia in 40 years with more than 300,000
people have been evacuated. This massive flood has caused severe damage to villager’s residence,
properties, livestock and crops, public infrastructure and business premises. In Malaysia National Security
Council (NSC) has responsibility for controlling the national disaster management system including flood.
The flood management system is based on National Policy and Mechanism in Disaster Management
known as MKN 20 that was established in May 1997 and it was revised in 30 March 2012 to describe the
role of various related stakeholders in more comprehensive and integrated, since the scope of the
disaster management increasingly complicated and complex. This mechanism seems very systematic in
managing the flood disaster, however historically it has commonly been considered as a government
function and is largely based on top-down government-centered machinery (Chan, 1995). Thus there are
questions arise, are these procedures are well understood by all levels of implementer, especially at the
district level and how effective the procedures of Arahan MKN 20 in managing 2014 massive flood if the
knowledge of implementer about the procedure is doubted? In addition of government agencies, this
disaster also attracted the participation of many volunteers to help flood victims. But do all these
volunteers are trained in dealing with disasters of this magnitude? This scenario shows that there are
many persons either from government agencies or volunteers in channeling the various types of
assistances to flood victims. However how far the assistances reach the right target groups in a fair
contribution? And if not where are the gaps that need to be improved to increase efficiency in the
management of flood disaster? Therefore the objective of this study is to evaluate the effectiveness of
Arahan MKN 20 in managing flood victims especially during and post-flood disaster 2014. Specifically
research objectives are (i) to evaluate the response and emergency actions from responsible agencies in
managing the flood victim; (ii) to assess the coordination of kinds distribution to the flood victims; (iii) to
review other countries’ mechanism in managing flood disaster especially Thailand and Japan and (iv) to
propose suggestions for sustainable management of flood disaster in the East Coast of Peninsular
Malaysia

2.0 Methodology
This study will use both secondary and primary data to elicit the study objectives. The secondary data
from government reports, books, articles, newsletters and internet sources was used to capture the
baseline information about the flood disaster in East Coast area. The primary will come from community
survey. An interview with officers from agencies involved in flood management was conducted first in
order to understand the scenario of flood in the study area. Then survey by using structured questionnaire
were carried out in two states of Peninsular Malaysia namely Kelantan and Pahang since these states are
mostly affected by flood disaster 2014. Survey on affected local communities in the district of Kuantan in
Pahang and district of Gua Musang in Kelantan has been done to find out their perception on the
effectiveness of the flood management and the coordination of kinds. These two districts was purposely
chosen to analyse the scenario of flood management in urban and remote area, whereas Kuantan will
represent the urban and Gua Musang will represent the remote area. The number of respondents that
have been interviewed was 213 in Kuantan, Pahang and 159 in Gua Musang, Kelantan. The focus group
discussion with various agencies involved in flood management such as MKN, local community
representative, police, fire department, welfare department, education department, health department and

13
other related agencies then have been conducted to discuss about the issues raised by the flood victims
in order to improve the Standard Operational Procedure (SOP) of MKN 2O Directive. Finally descriptive
analysis will be applied for primary data that has collected. The study will employ an analytical tool -
Statistical Package for Social Scientist (SPSS) to process the responses from primary sources. With the
aid of SPSS software the household responses will be coded and entered into a data analysis. Coding
will be done to classify answers into meaningful categories and bring out essential patterns and make
deductions from answers collated.

3.0 Results and Discussion

The evaluation of the effectiveness of Arahan MKN 20 in managing 2014 flood in the districts of Gua
Musang and Kuantan are divided into pre, during and post flood situation that related to community. Pre
and during-flood management represent the evaluation of response and emergency actions from
responsible agencies in managing the flood victim. While post-flood situation represent the evaluating of
coordination of kinds distribution to the flood victims.
As for pre-flood management procedure, respondents were asked with dichotomous question
(yes or no) with regard to the pre-flood procedures listed in the PTO such as dissemination of information
about weather condition, river/dam water level, location of flood warning board and siren, location of
evacuation center, safe vehicle area, area patrol, alternative routes and road closure, and water/electricity
supply disruption. Finding shows that the dissemination of pre-flood information is less efficient in remote
area such as in Lebir, Gua Musang with only 32% to 50% of respondents state ‘yes’ for all questions
listed imply that most of them didn’t informed about these information.
During the flood, the relocation rate of flood victims is low in remote area such as Sungai
Lembing in Kuantan District (52%) and Lebir in Gua Musang District (34.9%) as compared with urban
area such as Sungai Isap (100%) and Kampung Tiram (88%) in Kuantan district. This is due to limited
capacity of evacuation center and asset to relocate the victims. For example at Lebir the capacity of the
evacuation center is only for 100 people while the flood victims in that area is more than 400 families. On
the contrary, the relocation rate in Bertam area, one of the remotest areas in Kelantan, is high (88.1%)
due to a lot of villagers in Bertam had their own boats which formed the bulk of rescuer boats. At the
evacuation center, based on flood victims’ self-perceived on procedures listed in the SOP for flood
disaster management based on a scale of 1 = very dissatisfied to 5 = very satisfied, result showed the
overall score mean values for all procedures is more than 3, i.e. in the range of modest and satisfied. This
value is acceptable in the case of disaster situation at evacuation center. However flood victims in Bertam
and Lebir in Gua Musang district were dissatisfied on some flood management procedures such as space
allocation in evacuation center, distribution of food and non-food necessities, basic amenities, medical
treatment, counselling service, and role played by evacuation center committee shown by their score
means 2.36 to 2.95. This flood victim’s satisfaction level reflects the incompetency of management at the
evacuation center. However from 22nd until 25th December 2015, the Gua Musang area was
inaccessible as it was inundated. Communication, electricity and water supplies were cut off. Instead the
place for food storage also sank causing insufficient food supply to flood victims. Gua Musang area was
only accessible after three days of flooding. In this situation, the Penghulu had taken the initiative to
manage the flood victims.
For the case on non-evacuees, the flood management procedures is only limited to food, medical
and monetary aids. Even these assistance will reach to the victims when the flood began to recede.
However most of them claimed that they did not received those kind of assistances especially flood
victims from Bertam, Gua Musang and Sungai Lembing, Kuantan. This is because they were surrounded
by water and no access for outsiders to get into the area to deliver aids for the victims. In order to ensure
the flood management system is more holistic that can cover all scenarios of the flood, there is a need for
an SOP for non-evacuees in MKN 20 Directive.
In term of kind distribution coordination, the Welfare Department plays a role in coordinating all
assistance channeled to the evacuation centers. However, not everyone is aware about the aid delivery
procedure and most of them approach the victims directly in distributing aids. The lack of information on
aid coordination and distribution had led some victims to receive dumping assistances and some victims
not receiving anything. In this case, volunteers should be provided with information, knowledge and
training on the management of flood victims. The government also distributed Wang Ehsan of RM500 to
the flood victims. Most victims received the Wang Ehsan payment within the specified period except
victims in Sungai Lembing which only 34% of them answered ‘yes’ to the statement, even though many of

14
them submitted their claim forms within a week after evacuation centers closed. Generally the reason why
they did not receive the Wang Ehsan within the prescribed period was that aid distribution location is far
away, no transportation provided, staying in rural areas, information lag, and overcrowded aid distribution
venues, and old/fragile victims. However as an agency that responsible in registration of the flood victims,
the Department of Welfare (JKM) facing problems relating to the overlapping of household head’s name
due to inefficient registration system that allow the victims to cheat by entering more than one name per
household in order to obtain extra assistances. In order to avoid double payment to the affected
households, JKM had to spend extra time to produce the final name list devoid of name-overlapping. This
problem calls for a need to establish an integrated information system to allow the system to be more
effective disaster management without any resource waste.
This experience has provided guidelines toward improvement of sustainable flood management
especially on MKN20 Directives. Current MKN20 Directives only covers the case of 'normal' floods.
Suggestions for improvement of flood management procedure can be categorized according to the
stakeholders: (1) the community of victims, (2) government rescue agency, (3) non-governmental
organization (NGO). Mechanisms in the floods management can be categorized as (1) education on
disaster management, (2) the use of ICT in disaster management coordination between rescue agencies.
Communities need to cooperate and engage in programs related to enhancing their awareness and
preparedness capabilities organized by rescue agencies during the pre-flood phase. Every family should
be accomplished with Family Disaster Plan and Disaster Supplies Kit that lists the procedures to rescue
their family and property in the incident of flooding. However the success of this program depends on
sufficient funds allocated by the government to finance this program and the active involvement of local
communities. This program can also be co-sponsored by the NGO involved in the rescue work in the
incident of flooding.
Government rescue agencies like the Department of Irrigation and Drainage needs to generate a
mapping of flood risk areas. The ratio of rescue assets with the risk of flooding must be appropriate so
that there was no shortage of assets in the flooded areas. Programs such as Spatial Decision Support
System (SDSS), a comprehensive flood management plan that combines remote sensing technology,
geographic information system and global positioning system can be used by rescue agencies since the
coordination of flood management is critical aspect in the extreme case. To ensure the coordination of aid
distribution to flood victims NGOs or individual that want to deliver their aids need to register with the
rescuer agency at the scene. NGOs - such as the Red Crescent Society - also can get involved by
organizing Disaster Preparedness Education at the school level in order to nurture students become a
savior to their families through 1 Rescue for 1 family program. University students must not only involved
during post-flood but also can help increase awareness and preparedness of disaster through Co-
curricular program at university. The involvement of NGO can also focus on pre-disaster level, in addition
to the during-flood situation.

4.0 Conclusion

4.1 Findings from the literature on flood disaster management (FDM):

(i) FDM in developing countries is still of a reactive-response nature, i.e. more to
emergency response, rescue, and recovery
(ii) what is preferable is disaster management of a proactive and integrated nature,
pointing to Japan’s exemplary disaster management system

4.2 Information and data for this research were gathered from:
(i) interviews with agencies directly involved with flood disasters
(ii) questionnaire survey on 372 flood victims from Kelantan and Pahang
(iii) focus group discussion with all relevant agencies and flood victims to verify our
findings and to solicit policy recommendations

4.3 Findings on pre-flood management:

(i) inefficient information dissemination in remote areas
(ii) low level of awareness and capacity-building of the flood-prone community
4.4 Findings on during-flood management:
(i) low relocation rate in remote areas due to capacity limitation of evacuation centres

15
 (ii) victims from 2 of the 5 districts surveyed reported dissatisfaction with the inefficient
management of evacuation centres
(iii) search-and-rescue operations hampered by lack of equipment/assets, where the lack
is partly due to the sudden occurrence of such a large-scale flood

4.5 Findings on post-flood management:

(i) not all NGOs distributing in-kind aids were aware of the proper procedure, resulting in
poor coordination of in-kind distributions
(ii) problems of inefficient registration system for the RM500 Wang Ehsan assistance due
to overlaps of registered names of household head, and intentional fraudulent
behavior of victims

4.6 Policy improvement recommendations for MKN Directive 20 and the Standard Operating
Procedure for Flood Disaster Handling (Peraturan Tetap Operasi Pengendalian Bencana
Banjir):
(i) improvement on these two policies are necessary because these policies are only
suitable to deal with floods of the ‘normal’ category
(ii) need to readdress the issue and definition of the ‘terkepung’ category due to
physical/geographical inaccessibility of flood victims; this term has only been coined
during the 2014 year-end massive flood; the aforementioned standard operating
procedure is inadequate in dealing with this category of victims in terms of search-
and-rescue operations and food supply
(iii) improvement on the mechanisms of FDM can be made in terms of disaster
management education, and the use of ICT for better coordination between rescue
agencies
(iv) flood-prone communities need to be more involved in flood awareness programmes
for better preparedness on their part
(v) rescue agencies such as the Department of Irrigation and Drainage needs to produce
mappings of high-risk and flood-prone areas in order to come up with a more
accurate food supply and equipment/asset allocation for each area
(vi) NGOs distributing in-kind aids should be registered to avoid duplication and
redundancy in assistance/resources

16
 FLOOD HAZARD ASSESSMENT USING MULTI CRITERIA EVALUATION (MCE) METHOD IN
PENAMPANG AREA, SABAH

Project Information
Project Leader : Dr Rodeano Roslee
University : Universiti Malaysia Sabah
Address : Universiti Malaysia Sabah, Faculty of Science and Natural Resources,
Complex of Science and Technology, Jalan UMS, 88400 Kota Kinabalu,Sabah
Contact number : 010-9008007 / 088-320000 (Ext: 5651)
Email : rodeano@ums.edu.my
Project Members : Prof Dr Felix Tongkul
Dr Norbert Simon
Mustapha Abd. Talip

1.0 Introduction
The Penampang District of Sabah, East Malaysia is subjected to development pressure as the urban
centre of Kota Kinabalu expands onto the Sungai Moyog floodplain. The subsequent transition of land
use from rural development and cultivation of rice paddy to intensive urban development presents a
range of social and environmental issues. Of particular concern to the area are the issues associated with
flooding. In 2014 from October 7 to October 10, Penampang suffered its worse flood ever, since the last
big flood in 1991. According to the District Officer of Penampang as many as 40,000 people from 70
villages were affected by the flood. The flood coincided with continuous heavy rainfall due to typhoon
Phanfone and typhoon Vongfong. Another recent flood disaster in Penampang occurred on September
2007 and May 2013, affecting several villages.
The main objectives of this study are: a) to determine the Flood Hazard Level (FHL); b) to
determine the factors contributing to the flood occurrences; and c) to recommend mitigation measures in
order to minimize flood vulnerability & risk. It is hopes that the outcomes from this study can be an
important reference document for the local authority and other relevant agencies for the purpose of urban
planning and flood mitigation. An ad hoc, or reactive, approach to floodplain management has previously
been standard practice. Insufficient control over floodplain development practice has led to a worsening of
the flood problem. Until recently, floodplain management has only involved structural approaches to
modifying flood behaviour. However, without planning, the structural flood modification only compensates
for the poor development practice by restoring the flood behaviour to pre-development conditions.
Ultimately, there is no net benefit.

2.0 Methodology
There are four (4) main phases involved, namely: a) Phase I: Selection and evaluation of criteria; b)
Phase II: Multi-Criteria Evaluation (MCE); c) Phase III: Flood Susceptibility Analysis (FSAn); and d) Phase
IV: Flood Hazard Analysis (FHAn)

2.1 Selection and evaluation of criteria

The main purpose in Phase I are database development. Firstly, soil samples were collected from the
field will be analyzed their types in accordance with BS1377-1990. The next step is secondary data
compilation and literature review. Lastly observation of Flood Hazard Identification (FHI) parameters was
conducted through the fieldwork study.

2.2 Multi-Criteria Evaluation (MCE) technique

In Phase II, the choice of criterions that has a spatial reference is an important and profound step in Multi-
Criteria Evaluation (MCE) technique. Hence, the criteria consider in this study is based on their
significance in causing flood in the study area. Eigth factors are considered in relation to the causative
factors, which are rainfall, slope gradient, topography, drainage density, landuse, soil textures, slope
curvatures and flow accumulation.
Several questionnaires were distributed among experts in hydrology and hydraulics. The inputs
obtained from those experts were further used in carrying out the pair-wise comparison technique in order
to calculate the weights of each criterion. Pair-wise comparison is more appropriate if accuracy and
theoretical foundations are the main concern. The technique involves the comparison of the criteria and

17
as allows one to compare the importance of two criteria at a time. This very technique, which was
proposed and developed by Saaty (1980) within the framework of a decision making process known as
Analytical Hierarchy Process (AHP) is capable of converting subjective assessments of relative
importance into a linear set of weights. The criterion pair-wise comparison matrix takes the pair-wise
comparisons as an input and produces the relative weights as output. Further the AHP provides a
mathematical method of translating this matrix into a vector of relative weights for the criteria. Moreover,
because of the reason that individual judgments will never be agreed perfectly, the degree of consistency
achieved in the ratings is measured by a Consistency Ratio (CR) indicating the probability that the matrix
ratings were randomly generated. The rule-of-thumb is that a CR less than or equal to 0.10 signifies an
acceptable reciprocal matrix, and ratio over 0.10 implies that the matrix should be revised, in other words
it is not acceptable.

2.3 Flood Susceptibility Analysis (FSAn)

The initial step in Phase III is the delineation and conversion processes of data from the radar images
(IFSAR). Phase III also covers the integration between criteria weights and maps, producing a Flood
Susceptibility Analysis (FSAn) using spatial analyst, which determine the Flood Susceptibility Level (FSL)
in different period (2002, 2008 and 2014).
All of the thematic maps produced were analyzed through the spatial analyst technique (raster
calculator) based on Eq. (1) for LSL estimation and classification. The FSL calculation was carried out
through a combination of input parametric maps in Eq. (1) with the GIS operations using a grid base.

 [(32.53*Rainfall) + (22.74*Drainage Density) + (15.84*Flow Accumulation) + (11.08*Landuse) +

(7.19*Elevation) + (4.89*Slope Gradient) + (3.35*Soil Textures) + (2.38*Slope Curvatures)]
(1)

3.0 Results and Discussion

3.1 Flood Susceptibility Level (FSL)
Fig. 1 shows the results of Flood Susceptibility Level (FSL) data in year 2002, 2008 and 2014.The results
of the FSL level for the Penampang suggest that in 2002, 65% of the area have very low susceptibility
(VLS), 17% as low susceptibility (LS), 6% as moderate susceptibility (MS), 11% as high susceptibility
(HS) and 1% as very high susceptibility (VHS). While in 2008, 62.74% of the area have VLS, 15.62% as
LS, 15.49% as MS, 3.57% as HS and 2.58% as VHS. In year 2014, 40.49% of the area can be
categorised as having VLS, 35.08% as LS, 18.21% as MS, 5.50% as HS and 0.71% as VHS.
In general, the VLS to LS areas refer to stable conditions from flood vulnerability/risk. In contrast,
MS to HS areas are basically not recommended to be developed due to high flood vulnerability/risk.
However, if there is no choice or the developer or the local authorities really want to develop these areas,
some mitigation procedures to be introduced. VHS areas are strictly not recommended to be developed
and provisions for suitable structural and non-structural works planning controlare recommended.

18
 FIGURE 1 : Flood Susceptibility Analysis (FSAn) Maps of the study area (Year 2002, 2008 and 2014)

3.2 Flood Hazard Analysis (FHAn)

Flood Hazard Analysis (FHAn) is important for planning development activities in an area and can be
used as supplementary decision making tools. Rainfall, elevation, slope, drainage density, flow
accumulation, land use, soil texture, slope gradient and slope curvature were chosen as the most
influential factors for evaluating the flood hazard to the municipality.
Fig. 2 shows the FHAn map of the study area. The map was prepared by overlaying the flood
susceptibility maps in year 2002, 2008 and 2014 (Fig. 1). Based on the map, approximately 3.69% of total
study area classified as very high hazard (VHH), 2.38% as high hazard (HH), 15.93% as moderate
hazard (MH) and 15.16% as low hazard (LH), as vey low hazard (VLH) respectively.

FIGURE 2 : Flood Hazard Analysis (FHAn) map of study area

19
 2
The Sg. Moyog catchment covers an area of approximately 295km . The upper reaches of the
catchment extend into the Crocker Range, with elevation exceeding 1,800m. From the headwaters, the
Sg. Moyog meanders in a westerly direction through steep mountainous terrain, until it reaches the
expansive lower floodplain at Dongongon. Fig. 3 shows the floodplain map of the study area. From this
figure, most of the floodplain area is located at the western part of the study area. Across the Sg. Moyog
floodplain, the main towns are Dongongon and Putatan. The largest village is Kampung Petagas with
3,500 people. Any kind of development and activities should be minimizes as the area is more prone to
flood disaster.

Sources: (Department of Drainage and Irrigation, 2014)

FIGURE 3 : Floodplain map of study area

3.3 Validation for Flood Hazard Analysis (FHAn)

For validation of flood hazard models, two basic assumptions are needed. One is that flooded areas are
related to spatial information, and the other is that future flooded areas will affected by a specific factor
such as rainfall. In this study, the two assumptions are satisfied because the flooded areas were related
to the spatial information and the flooded areas were triggered by heavy rainfall in the study area. The
Flood Hazard Analysis (FHAn) result was validated using the Floodplain map (Fig. 3). The validation
process is carried out by comparing the floodplain data with the Flood Hazard Analysis (FHAn) map (Fig.
2) through the Prediction Rate Curve (PRC) (Fig. 4). In order to create PRC, the flood index values in
FHAn were calculated and all its cells were sorted in decreasing order. Then, the ordered cell values
were divided into 100 classes. All of these 100 classes were weighted and were overlaid with the
floodplain map data and then converted into the percentage format for each class. As a result of the
calculation and interpretation, the average ratio of the areas under the curve was 0.839, and thus can be
argued that validation prediction accuracy was 83.90% (Fig. 4). This means that the FHAn result that
were carried out in this study have a good reliability (0.8 < AUC < 0.9) (Zhu et al. 2010).

20
 FIGURE 4 : Illustration of cumulative frequency showing flood hazard index rank (y-axis) occurring in cumulative
percentages of flood occurrences (x-axis)

4.0 Conclusion

4.1 The results of this study indicate that the integration of MCE and GIS techniques provides
a powerful tool for decision making procedures in FSL mapping, as it allows a coherent
and efficient use of spatial data. The use of MCE for different factors is also
demonstrated to be useful in the definition of the risk areas for the flood mapping and
possible prediction. In overall, the case study results show that the GIS-MCE based
category model is effective in flood risk zonation and management.
4.2 The developed framework model (Fig. 4) will be a very valuable resource for consulting,
planning agencies and local governments in managing hazard/risk, land-use zoning,
damage estimates, good governance and remediation efforts to mitigate risks. Moreover,
the technique applied in this study can easily be extended to other areas, where other
factors may be considered, depending on the availability of data.
4.3 The main causes of flooding in the study area are: a) Increased runoff rates due to the
urbanisation; b) Loss of flood storage – development in flood plains and drainage
corridors; c) Inadequate drainage systems; d) Constriction at bridges; e) Undersized
culverts; f) Siltation in waterway channels from indiscriminate land clearing operations; g)
Localised continuous heavy rainfall; h) Tidal backwater effect; and i) Inadequate river
capacity.
4.4 Recognition that unplanned and uncontrolled development can increase the risk to life
and damage to property is fundamental to successful floodplain management.
Awareness of this issue is not just the responsibility of the local authorities, but all
stakeholders, covering both public and private sectors. Whilst the land developer has the
social responsibility for flood compatible development, the approving agencies share a
portion of that responsibility through effective floodplain management, excised in a
transparent, impartial manner.

References
Department of Drainage and Irrigation, 2014. Malaysia Water Resources Management Forum 2014.
http://www.sumo.my/index.php/happenings/667-malaysian-water-resources-management-mywrm-forum-
2014
Saaty, 1980. The Analytic Hierarchy Process. New York: Mc Graw Hill.
Zhu, W., Zeng, N. & Wang, N. 2010. Sensitivity, specificity, associated confidence interval and ROC analysis with
practical SAS implementation. NESUG

21
 APLIKASI MEDIA SOSIAL SEBAGAI SISTEM AMARAN AWAL BENCANA BANJIR KEPADA
PENDUDUK DALAM LEMBANGAN SUNGAI KELANTAN

Project Information
Project Leader : Kamarul bin Ismail
University : Universiti Pendidikan Sultan Idris
Address : Jabatan Geografi dan Alam Sekitar, Fakulti Sains Kemanusiaan.
Contact number : 011-262-477-63
Email : kamarul.ismail@fsk.upsi.edu.my
Project Members : Dr Mazdi bin Marzuki
Dr Mohd Hairy bin Ibrahim
Dr Nor Kalsum binti Mohd Isa
Encik Muhammad Hasbi bin Abdul Rahman
Encik Muhammad Nadzir bin Ibrahim

1.0 Introduction
Bencana alam merupakan suatu kesan bahaya daripada bencana semula jadi seperti banjir, puting
beliung, gempa bumi, tanah runtuh dan letusan gunung berapi. Hakikat hari ini bahawa dunia menjadi
saksi kepada banyak bencana banjir berlaku yang mengorbankan banyak nyawa akibat kurang tindak
balas awal dan persediaan untuk menyelamat (Ibrahim, 2007). Selain itu, dunia telah dilanda dengan
pelbagai bencana seperti Taufan Katrina di Amerika Syarikat, tsunami di Indonesia dan Jepun yang
mengakibatkan kerosakan harta benda. Oleh itu, bencana alam yang semakin kerap berlaku diramalkan
terus meningkat dari semasa ke semasa dan merosakkan sistem kitaran hidup manusia (Velev & Zlateva,
2006). Menurut Majlis Keselamatan Negara (2015) dan Ibrahim dan Fakru’l-Razi (2006) bencana
merupakan sesuatu ganggu kepada kelangsungan aktiviti komuniti, melibatkan kehilangan dan
kerosakan harta benda, kerugian ekonomi dan urusan negara serta kemusnahan alam sekitar di luar
daripada kemampuan komuniti setempat. Hal ini menunjukkan bahawa bencana adalah sesuatu
fenomena alam sekitar yang cukup serius. Pusat Penyelidikan di Epidemiologi Bencana (CRED)
mengkasifikasikan bencana apabila melibatkan mangsa seramai 10 orang atau lebih dilaporkan
meninggal, 100 orang dilaporkan terjejas dan membawa maksud kepada kerosakan ke atas masyarakat
dan alam sekitar (United Nations, 2015; Ibrahim, 2007).

2.0 Methodology
Kajian ini menggunakan data primer melalui kaedah kaji selidik untuk mendapatkan maklumat mengenal
pasti penggunaan aplikasi media sosial dalam kalangan masyarakat di dalam lembangan Sungai
Kelantan. Selain itu, ia juga tertujuan untuk mendapatkan maklum balas mengenai penggunaan media
sosial semasa bencana banjir. Seramai 250 orang responden telah telah dipilih untuk menjawab borang
soal selidik yang diedarkan. Kaedah pensampelan yang telah dipilih adalah pensampelan rawak mudah
dengan mengedarkan kepada masyarakat dalam kawasan mukim yang terlibat dengan bencana banjir.
Teknik pensampelan rawak mudah merupakan kaedah pensampelan yang paling mudah dan ringkas
(Fauzi et al., 2015). Tambahan lagi, pensampelan rawak mudah digunakan dalam penyelidikan untuk
memastikan populasi mempunyai ruang dan peluang yang sama dan bebas untuk dipilih sebagai sampel
kajian (Mohamad Suhaily Yusri et al., 2015; Chua, 2011). Kajian ini menggunakan analisis deskriptif iaitu
kekerapan dan min serta dilakukan perbincangan mengenai respon komuniti melalui aplikasi media sosial
semasa bencana banjir di kawasan Kuala Krai, Kelantan. Kekerapan digunakan untuk mendapatkan
maklumat penggunaan media sosial semasa banjir. Manakala, min pula digunakan untuk mendapatkan
tahap penggunaan aplikasi media sosial untuk berkomunikasi, berkongsi maklumat dan kolaborasi
(Jadual 1).

22
 Jadual 1: Penetapan skala min bagi komunikasi, berkongsi maklumat dan kolaborasi
3.0
Tahap Skor
Tinggi 3.33 – 5.00
Sederhana 1.67 – 3.32
Rendah 0.00 – 1.66
Sumber: Mohammad Suhaily Yusri et al., 2015

3.0 Results and Discussion.

Hasil analisis yang dijalankan menunjukkan bahawa penggunaan aplikasi media sosial dalam kalangan
bukan staf kerajaan hampir sama berbanding staf kerajaan (Jadual 2). Ini menunjukkan penggunaan
aplikasi media sosial tidak terlalu bergantung kepada jenis pekerjaan seseorang responden kajian ini.
Data yang ditunjukkan bagi keseluruhan aplikasi media sosial yang digunakan hampir sama sahaja
terutama antara pekerjaan kerajaan, swasta, suri rumah, buruh dan penoreh getah yang menjadi
responden dominan antara responden dalam kajian ini.

Jadual 2: Penggunaan aplikasi media sosial berdasarkan pekerjaan

Pekerjaan Media sosial
Faceboo WhatsAp SMS Twitter Instagra Flickr MySpace YouTube WeCha
k p m t
Kerajaan 20 27 28 4 8 1 1 4 9
Pesara 2 4 7 0 0 0 0 0 0
Swasta 27 30 36 2 8 0 0 2 17
Suri rumah 20 25 39 5 8 3 3 5 16
Buruh 9 12 26 0 9 0 0 2 5
Penternak 2 2 0 1 9 0 0 1 1
Penoreh getah 22 29 44 5 6 3 3 5 13
Peniaga 9 14 18 1 2 0 0 2 6
Pelajar 2 2 5 0 0 0 0 1 2
Pemandu 1 2 3 0 0 0 0 0 1
Tukang jahit 2 3 3 2 2 1 1 1 1
Sendiri 1 2 6 0 0 0 0 0 1

Maklum balas masyarakat terhadap bencana banjir dengan menyebarkan berita mengenai
bencana banjir semasa berlaku di Kuala Krai paling banyak menggunakan aplikasi WhatsApp iaitu
seramai 96 orang (38.4%) dan diikuti aplikasi SMS seramai 70 orang (28%), Facebook seramai 56 orang
(22.4%), WeChat seramai 14 orang (5.6%), dan masing-masing lima orang (2%) bagi Twitter dan
Instagram, dua orang (0.8%) bagi MySpace dan YouTube (Jadual 3). Tambahan lagi, jikalau melihat
kepada kekerapan bagi aspek keberkesanan dan efektif aplikasi media sosial dalam menyampaikan
berita mengenai bencana banjir turut didominasi oleh aplikasi WhatApps iaitu seramai 115 orang (46%)
dan diikuti oleh SMS seramai 68 orang (27.2%) dan seramai 48 orang (19.2%) bagi Facebook (Jadual 3).
Bagi aplikasi media sosial yang lain hanya mewakili 7.6 peratus iaitu Twitter (6 orang), Instagram dan
YouTube masing-masing 2 orang, MySpace seorang dan WeChat (8 orang).

Jadual 3: Penggunaan dan maklum balas komuniti Kuala Krai melalui media sosial
Media sosial Penggunaan Maklum balas komuniti
Kekerapan Peratusan (%) Kekerapan Peratusan (%)
SMS 215 86.0 70 28.0
WhatsApp 152 60.8 96 38.4
Facebook 117 46.8 56 22.4
WeChat 72 28.8 14 5.6
Instagram 34 13.6 5 2.0
YouTube 23 9.2 2 0.8
Twitter 20 8.0 5 2.0
Flickr 8 3.2 0 0.0
MySpace 8 3.2 2 0.8

23
Analisis penggunaan aplikasi media sosial dalam kalangan masyarakat melibatkan aspek komunikasi,
berkongsi maklumat dan membentuk kumpulan (kolaborasi) pula menunjukkan bahawa aplikasi Sistem
Pesanan Ringkas (SMS) mendominasi ketiga-tiga aspek berkaitan semasa bencana banjir di Kuala Krai,
Kelantan. Tambahan lagi, respon komuniti terhadap ketiga-tiga aspek komunikasi, berkongsi maklumat
dan membentuk kumpulan (kolaborasi) didominasi melalui penggunaan aplikasi Facebook, WhatsApp
dan SMS. Maklum balas awal daripada komuniti melalui penggunaan aplikasi media sosial semasa
bencana banjir di Kuala Krai, Kelantan sangat penting untuk memberikan kesedaran dan persediaan
awal kepada komuniti setempat supaya bersiap sedia dalam sebarang kemungkinan. Namun begitu,
kekerapan penggunaan aplikasi media sosial untuk aspek-aspek berikut berapa pada tahap jarang
(Jadual 3) dan analisis min berada pada tahap sederhana (Jadual 4). Selain itu, ia juga memberikan
amaran awal dan bertindak sebagai sistem amaran awal yang efektif melalui proses penglibatan komuniti
secara optimum (Collins & Kapucu, 2008). Hal ini demikian kerana, media sosial menjadi medium
perbincangan dan perkongsian maklumat yang sangat pantas dan efisien, mencipta kepada kandungan
perbincangan dan berkongsi maklumat. Velev dan Zlateva (2006) menjelaskan terdapat empat cara yang
masyarakat boleh gunakan teknologi media sosial semasa bencana alam iaitu komunikasi antara kawan
dan keluarga, mengemaskini situasi semasa bencana, menyebarkan kesedaran berkaitan bencana dan
akses bantuan perkhidmatan semasa bencana.
Hasil analisis bagi aspek-aspek penggunaan aplikasi media sosial dalam kajian ini menunjukkan
bahawa kesemua aspek penggunaannya berada tahap sederhana sahaja (Jadual 5). Walau
bagaimanapun, berbanding antara ketiga-tiga aspek penggunaan tersebut mendapati bahawa aspek
komunikasi berada pada tahap lebih tinggi (min=2.40, SD=0.89) berbanding aspek lain. Aspek berkongsi
maklumat juga berada pada tahap sederhana (min=2.36, SD=103). Manakala, aspek kolaborasi
(membentuk kumpulan) turut sederhana (min=2.30, SD=1.29).

Jadual 4: Kekerapan penggunaan media sosial bagi setiap aspek penggunaan

Kekerapan Komunikasi Berkongsi maklumat Kolaborasi
Bil % Bil % Bil %
Jarang-jarang 87 34.8 116 46.4 111 44.4
Kadang-kadang 67 26.8 83 33.2 70 28.0
Setiap hari 60 24.0 30 12.0 35 14.0
Setiap minggu 7 2.8 4 1.6 8 3.2
selalu 29 11.6 17 6.8 26 10.4

Jadual 5: Nilai min bagi aspek komunikasi, berkongsi maklumat dan kolaborasi
Aspek Min Sisihan Piawaian Tahap
Komunikasi 2.40 0.89 Sederhana
Berkongsi maklumat 2.36 1.03 Sederhana
Kolaborasi 2.30 1.29 Sederhana

Walaupun ketiga-tiga aplikasi media sosial SMS, WhatsApp dan Facebook yang paling dominan
digunakan oleh masyarakat dalam lembangan Sungai Kelantan sebagai respon awal terhadap bencana
banjir bagi aspek yang dinyatakan, terdapat juga aplikasi media sosial lain digunakan. Sebagai contoh
aspek komunikasi turut melibatkan aplikasi YouTube dan WeChat, aspek kolaborasi (membentuk
kumpulan) melibatkan Twitter, Instagram, MySpace dan WeChat, dan respon komuniti Kuala Krai
semasa bencana banjir melalui aplikasi Twitter, Instagram, YouTube dan WeChat dalam aspek berkongsi
maklumat. Hasil analisis yang dijalankan ini mengambarkan bahawa kepelbagaian aplikasi media sosial
digunakan dalam kalangan komuniti semasa bencana banjir berlaku.

4.0 Conclusion
Kajian yang dijalankan secara keseluruhannya mendapati bahawa:

4.1 Tiga aplikasi media sosial yang boleh digunakan oleh pihak berkuasa untuk digunakan
sebagai alat sebaran amaran awal bencana banjir ialah Whatapps, Facebook dan Sistem
Pesanan Ringkas (SMS).
4.2 Ini kerana penggunaan ketiga-tiga media sosial yang dominan ini adalah menyeluruh dari
aspek pendidikan dan pekerjaan responden. Kebarangkalian ketersampaian maklumat
adalah lebih tinggi berbanding dengan penggunaan media-media sosial yang lain.

24
 4.3 Namun begitu penyebaran maklumat melalui aplikasi media sosial yang melibatkan
penggunaan jalur lebar seperti Facebook, Whatapps dan WeChat hanya berkesan
sebelum bencana banjir berlaku. Dapatan yang diperolehi menunjukkan penyebaran
maklumat melalui SMS lebih berkesan semasa dan selepas bencana banjir berlaku.
4.4 Taburan penggunaan aplikasi media sosial berdasarkan ruangan pula menunjukkan
bahawa terdapat perbezaan yang ketara antara mukim-mukim dalam lembangan Sungai
Kelantan.
4.5 Hasil analisis menunjukkan penduduk yang berada dalam daerah Gua Musang lebih
cenderung menggunakan SMS berbanding dengan penduduk dalam daerah Kuala Krai
yang majoritinya menggunakan aplikasi Facebook dan Whatapps untuk berkolaborasi,
berkomunikasi dan bertukar maklumat.
4.6 Dari aspek penyebaran berita palsu pula terdapat perbezaan corak pengetahuan
kewujudan berita palsu dalam kalangan responden mengikut perbezaan tahap
pendidikan. Majoriti responden yang mempunyai tahap pendidikan pada peringkat SPM,
PMR dan UPSR tidak mengetahui kewujudan penyebaran berita palsu semasa bencana
berlaku.
4.7 Penggunaan media sosial oleh masyarakat dalam lembangan Sungai Kelantan sebelum
bencana berlaku majoritinya adalah untuk berkomunikasi dan berkongsi maklumat.
Semasa bencana berlaku bentuk penggunaan media sosial lebih cenderung kepada
mendapatkan pertolongan, mengetahui situasi semasa banjir dan penggunaan media
sosial selepas bencana juga adalah untuk mendapatkan pertolongan dan mengetahui
situasi terkini di lokasi bencana.
4.8 Secara keseluruhannya, kajian ini mencadangkan penggunaan tiga media media utama
iaitu Facebook, SMS dan Whatapps sebagai alat penyebaran maklumat atau sistem
amaran awal kepada masyarakat di lembangan Sungai Kelantan. Namun begitu, untuk
mengelakkan berlaku penyebaran maklumat palsu penggunaan SMS adalah lebih
berkesan berbanding media-media sosial yang lain.

References
nd
Chua Yan Piaw (2011). Kaedah Penyelidikan. (2 Ed.). Kuala Lumpur: McGraw-Hill Education.
Fauzi Hussin, Jamal Ali & Mohd Saiful Zamzuri Noor (2014). Kaedah penyelidikan dan analisis data. Sintok: Penerbit
Universiti Utara Malaysia.
Ibrahim Mohamaed Shaluf & Fakhru’l-Razi Ahmadun (2006). Disaster types in Malaysia: an overview. In Disaster
Prevention and Management: An International Journal, 15 (2): 286-298.
Ibrahim Mohamed Shaluf (2007). Disaster types. In Disaster Prevention and Menagement: An International Journal,
16 (5): 704-717.
Majlis Keselamatan Negara (2015). Portal Bencana. Putrajaya: Majlis Keselamatan Negara. Diperoleh daripada
http://portalbencana.ndcc.gov.my/portal pada November 2, 2015.
Schulz, E. F., Koelzer, V. A., & Mahmood, K. (1972). Floods and Droughts. United States: Water Resources
Publications.
Mohamad Suhaily Yusri Che Ngah, Hanifah Mahat & Koh Liew See (2015). Tahap kesedaran terhadap sistem
penuaian air hujan dalam kalangan komuniti Tanjong Malim, Perak. Dalam Mohmadisa Hashim,
Zullyadini A. Rahman, Nasir Nayan, Yazid Saleh, Hanifah Mahat, Zainudin Othman et al., (2015). Persidangan
Kebangsaan Geografi & Alam Sekitar Kali Ke-5, 6-7 Oktober 2015 di Universiti Pendidikan Sultan Idris.
United Nations (2015). United Nations Office for Disaster Risk Reduction. Diperoleh daripada
http://www.unisdr.org/who-we-are/what-is-drr pada Jun 26, 2015.
Velev, D., & Zlateva, P. (2006). Use of media social in natural disaster management. Bulgaria: Institute of System
Engineering and Robotics.
Volker, A., & Boekelman, R. D. (1993). Hydrology and Water Management of Deltaic Areas. Netherland: Center for
Civil Engineering Research and Codes.

25
 EMERGENCY AND RESPONSE PLANNING (ERP): SIMULATION ON COORDINATION OF INTER
AGENCY IN FLOOD CATASTROPHIC EVENT

Project Information
Project Leader : Prof Madya Dr Abdul Mutalib Bin Leman
University : Universiti Tun Hussein Onn Malaysia
Address : Faculty of Engineering Technology, 86400Parit Raja, Batu Pahat, Johor
Contact number : 012-2168640 / 07-4537794
Email : mutalib@uthm.edu.my
Project Members : 1. Prof Madya Dr Mohd Najib Bin Mohd Salleh (UTHM)
2. Prof Madya Dr Ishak Bin Baba(UTHM)
3. Dr Johnson Lim Soon Chong (UTHM)
4. Khairunnisa A.Rahman (Politeknik Merlimau)

1.0 Introduction
In Malaysia, flood is the most significant of natural hazard (WECAM, 2013) and has continued escalate
while the country is more developed and the current flood is occur at 2014. It affecting more than 15% of
the total population in Malaysia and damage cost is estimated to be a million of Ringgit Malaysia. Thamer
et al., (2011). According to EM-DAT (2011) , the frequency of major flooding in Malaysia for the past 50
years, frequency of occurrence, the number of killed, affected and damage loss. Highest frequency of
flood has occured in Kelantan or Terengganu area with the percentage area is 38%, followed by Johor of
19% Kedah of 14% and the other state is below than 10% from the total area. However, flooding that
occurred on 2014 is higher than 2010 such as reported by awani (2014) that involve 100,000 peoples
flood victims. This study focus on data for communication and how to coordinate the interagency were
really need to ensure the efficiency and the proper management to cater the flood victims.Thus the
delivery of information system for residents’ indispensable. Existing detection system not able to deliver
the information directly to resident. So it will take more longer times to deliver the information through
siren, media such as newspaper, television, or radia to resident. It makes resident lack of time to respond
to safe their life and their important goods. Nowadays Mobile wireless devices such as smartphones have
become a widespread and typical asset. Flood warning using mobile application that able to give sign via
mapping will help the early warning system. This project were identify and listed down all the contact
number of state, district and area (Whole country). System was develop and it can found and free access:
http//interagensibanjirmalaysia2.weebly.com. System for mobile application has been develop to ensure
the data information were given by using GPS.

26
2.0 Methodology

Detection System

Resident or traffic able

to respond

Exiting detection system

3.0 Results and Discussion

Table 1: Summary of Role, Confidence and Satisfaction level of respondent toward agencies

Role Confidence Satisfaction

MKN 3.33 3.40 3.17

POLIS 3.66 3.62 3.43

27
 BOMBA 3.63 3.94 3.83

HOSPITAL 3.61 3.54 3.63

JPAM 3.77 3.80 3.65

4.0 Conclusion

4.1 This project were identify and listed down all the contact number of state, district and
area (Whole country)
4.2 1 council and 4 related agency were evaluate by the despondence and show that the
role, confidence level and satisfaction were between 3.17-3.94.
4.3 System were develop and it can found and free access:
http//interagensibanjirmalaysia2.weebly.com
4.2 System for mobile application has been develop to ensure the data information were
given by using GPS.

References
Thamer Ahmed Mohammed, Saleh Al-Hassounand Abdul Halim Ghazali.Prediction of Flood Levels Along a Stretch
of the Langat River with Insufficient Hydrological Data. Pertanika J. Sci. & Technol. 19 (2): 237– 248 (2011)
Shaluf, I.M., Ahmadun, F. 2006. Disaster Types in Malaysia: An Overview. Disaster Prevention and Management,
15(2), 286-298.
Water and Energy Consumer Association of Malaysia (WECAM). Flood Mitigation and Adaptation Memorandum
2013.
Department of Irrigation and Drainage (DID). Flood Management. Vol. 1, 2009.
EM-DAT: The OFDA/CRED International Disaster Database www.em-dat.net – Universite Catholique de Louvain,
2011 – Brussels – Belgium
KemasKini. JumlahMangsaBanjir2014. Retrieved February 27 2015, http://www.astroawani.com/berita
Wong, W. S., Mazura N. Z. and Ceon H. S. Flood Mapping as a Planning and Management Tool. MyWRM Forum,
2012.
Billa, L., Shattri, M., Mahmud, A.R., Ghazali, A.H. 2006. Comprehensive Planning and The Role of SDSS in Flood
Disaster Management in Malaysia. Disaster Prevention and Management.,15(2): 233-240.
Badruddin A. Rahman. Issues of Disaster Management Preparedness: A Case Study of Directive 20 of National
Security Council Malaysia. Int. Journal of Business and Social Science.Vol.3 No.5; March 2012
Shaluf, I.M., Ahmadun, F. 2006. Technological Emergencies Expert System (TEES). Disaster Prevention and
Management,. 15(3), 414-424.
A.Fakhru’l-Razi. DisasterManagement in Malaysia, 2009,Universiti Putra Malaysia (UPM) Organization World
Meteorology (WMO). 2006. Social Aspects and Stakeholders Involvement in Integrated Flood Management.
APFM Technical Document No. 4, Flood Management Policy Series, Associated Programme on Flood
Management (WMO), Geneva,
Chan N. W. and Parker D. J. 1996. Response to dynamic flood hazard in peninsular Malaysia,” in Proc. The
Geographical Journal. 162(3): 313-325.
Khalid M. K. and Shafiai S. 2015. Flood Disaster Management In Malaysia: An Evaluation Of The Effectiveness Flood
Delivery System, International Journal of Social Science and Humanity. 5(4).
Sahu S. 2006. Guidebook on technologies for disaster preparedness and mitigation. Asian and Pacific Centre for
Transfer of Technology (APCTT).
Noor H. M., Ghazali H. and Mustapha F. 2012. Public INFOBANJIR: towards people centered flood information
dissemination. Water Resources and Hydrology Division: Department of Irrigation and Drainage Malaysia.
ID: 122.
Chan N. W. Impacts of disasters and disasters risk management in Malaysia: the case of floods, “in Economic and
Welfare Impacts of Disasters in East Asia and Policy Response. Sawada and S. Oum (eds). ERIA
Research Project Report 2011-8, Jakarta: ERIA: 497-545.
Malaysia Institute of Architects (PAM). 2015. Strategic Initiative in Flood Disaster Preparedness and Mitigation for
Malaysia.
Safiza S. K. B., Abdul S. S. and Zahriah O. 2009. Disaster Management in Malaysia: An Application Framework of
Integrated Routing Application for Emergency Response Management System.SOCPAR.Soft Computing
and Pattern Recognition, International Conference of, Soft Computing and Pattern Recognition, International
Conference of 2009:716-719.

28
 UNDERWATER GROUND MAPPING FOR FLOOD DISASTER USING ULTRASONIC SENSOR

Project Information
Project Leader : Dr. Mohd Hafiz Bin Fazalul Rahiman
University : Universiti Malaysia Perlis
Address : Pusat Pengajian Kejuruteraan Mekatronik, Universiti Malaysia Perlis, Kampus
Pauh Putra, 02600 Arau, Perlis
Contact number : 019-5754010
Email : hafiz@unimap.edu.my
Project Members : Prof. Dr. Ruzairi Bin Abdul Rahim
Dr. Zulkarnay Zakaria

1.0 Introduction
Ultrasound or sonar propagation in waters is greatly influenced by the interaction with both the water
surface and the bottom surface. This interaction can be used as a tool for detecting small targets on or
beneath the seabed or for examining the physical properties of the surface [1].
Visualization of acoustic wave fronts started to be the object of intensive research in the 1960s.
The understanding of acoustic wave fronts and their interaction with objects is important for optimizing
both the performance of acoustic sources and detectors and for the generation of structures, surfaces
and materials with particular acoustic absorption and scattering characteristics [2]. Moreover, the
visualization of acoustic wave fronts represents a reliable test for transducer design or periodic control,
i.e., to check if the properties of the generated sound beam still persist over a long-term period [3].
The normal occurring flood water in Malaysia is atypically muddy and full of debris. It is not as
clear as normal water. It comprised of sand, mud and many other floating and submerged materials which
will hinder the soundwave to not bounce back to the receiver. Other than the flood water condition, the
flood current, temperature of the water and the distance of flood-bed are other factors affecting the
ultrasonic speed. The stronger the occurring flood current, more noises recorded ultrasonically which will
affect the reading. In term of temperature, ultra wave can move smoother if the temperature is warmer.
The higher the temperature, better waveform will go through. Lastly, the flood-bed will affect the reading if
it comprised of soft sedimentation such as moss or mud, the wave would probably not bounce back.

2.0 Methodology
The concept of the system will be the concept of the autonomous underwater vehicle (AUV) which uses
acoustic transmission to map the underwater ground. The same concept used for this project using
ultrasound transmission.The prototype uses a 200 kHz frequency which is suitable for the condition on
flooded area and the condition of the boat used by the search and rescue (SAR). The ultrasonic are able
to penetrate to the smallest resolution object which in this case the minimum resolution is 7.5mm.
λ= /
Where λ is the wavelength, is the speed of sound in water and is the ultrasound frequency

3.0 Results and Discussion

Three experiments were conducted to find the viability of mapping the underwater depth information and
to convert them into mapping data before relaying them to ground station. The results from the three
experiments conducted using materials as shown below.

29
 Figure 1: Prototype sensor

3.1 No object
The first experiment is to ensure the prototype does not have any error. The prototype will move along the
water tank.

Figure 2: Sensors with empty water tank

In this result we can see that the sensors are functioning. The depth of the water in the tank was
confirmed by the sensors.

3.2 Brick as an object

The second experiment is to test the prototype with an obstacle in the water. We used brick because it’s a
solid material and will easily bounce back the sound signal.

Figure 3: Sensors with brick as object

30
In this mapping result, the sensors detected the brick at the frequency of 200 kHz due to the thickness of
the brick which is more than 3.75mm. The data was processed by the software program to create a
shallow part on the map where in figure 6.2 the blue section is deeper than the orange section.

3.3 Brick and PVC Pipe as an object

The final experiment is to test out the brick and PVC pipe together as obstacles.

Figure 4: Sensors with brick and PVC pipe as object

In this experiment, the sensors need to pass by the two obstacles, namely PVC pipe and the brick. The
result showed that the brick is clearly mapped to be shallow while the PVC pipe, the data indicated that
there are some parts that are not detectable. This is due to strength of the sensors frequency of 200 kHz
which will penetrate an object that is less than 7.5mm, not bounced back.

4.0 Conclusion
From the present study, the followings can be concluded:

4.1 The results collected from the three experiments using the water tank as simulated flood
situation showed the system can be applied to actual flood area.
4.2 More sensors can be installed to get wider map view of depth.
4.3 Mapping of obstacles in the deep and shallow areas can be improved using multi-angle
shapes to be more accurate.

References
M. Raju, 2001. Ultrasonic Distance Measurement With the MSP430. TAG: Pizo Drive Circuit. Mixed Signal
Processors 1(3): 1–18.
M. H. F. Rahiman, R. A. Rahim, and H. A. Rahim, Feb. 2011 “Gas Hold-Up Profiles Measurement Using Ultrasonic
Sensor,” IEEE Sens. J., 11(2): 460–461.
R. Longo, S. Vanlanduit, G. Arroud, and P. Guillaume, Jan. 2015. Underwater Acoustic Wavefront Visualization by
Scanning Laser Doppler Vibrometer for the Characterization of Focused Ultrasonic Transducers. Sensors
(Basel),15(8): 19925–36.

31
 RAPID ASSESSMENT METHOD OF FLOOD DAMAGE USING SPATIAL-STATISTICAL MODELS

Project Information
Project Leader : Abdul Hamid Mar Iman
University : Universiti Malaysia Kelantan
Address : Fakulti Sains Bumi, 17600 Jeli, Kelantan
Contact number : 09-9472914 / 019-7798287
Email : hamid.m@umk.edu.my
Project Members : Edlic Sathiamurthy
Alia Atika Asyikin
Muhammad Hanis Rashidan

1.0 Introduction
The December 2014’s flood has caused huge damage of close to RM 1 billion to the country, exclusive of
RM 78 million for cleaning operations in Kelantan. A report quoted that about RM 200 million was
estimated for the damage of infrastructure in Kelantan (The Star, 2/2/2015). According to Urban
Wellbeing, Housing and Local Government Minister, Datuk Rahman Dahlan, between 2,000 and 3,000
houses in Kelantan were destroyed in the worst flood ever in decades (Azura, 2015). More than 200,000
victims were affected by the massive flood which claimed 21 lives (Anon, 2015).
One of the main concerns of flood is to estimate the extent of damage to properties and other
assets. It is an intricate task to perform since damage assessment needs itemized identification and
estimate of affected objects. Some studies resort to only assessing flood impact without being able to
provide the monetary estimate of the damage (see for e.g. Ab-Jalil and Aminuddin, 2006; Pradan, 2009).
Therefore, it is vitally important to devise a rapid assessment method that can provide a reliable method
for estimating the monetary loss as soon as flood strikes in a particular location. Flood damage
assessment itself is not a new thing; there has been a substantial body of literature dealing with it.
However, the techniques are difficult to generalize since they vary and case-to-case.
By applying empirical damage or loss functions meant for compensation, relief, and/or insurance
purposes, flood damage rapid assessment method (FD-RAM) seeks to estimate the expected monetary
damage as soon as a disaster strikes (Poser and Dransch, 2010). In case of flood, these models
calculate the expected damage as a function of inundation depth, building characteristics, and possibly
further parameters such as water contamination (Poser and Dransch, 2010).

2.0 Theoretical Background

2.1 Flood Damage Model
For any property, expected physical damage (EPD) is generally modelled as:
EPD = f(SD, CD) (1)
where SD is structural damage and CD is content damage. SD comprises damage to land/soil and
building while CD can refer to any type and/or amount of ‘content’ asset. Therefore, ‘content’ can
comprise any moveable asset inside or outside a building such furniture, radio, television, appliance,
vehicle, clothes, money, etc. Damage to land/soil is difficult to ascertain. For example, the eroded soil of a
land parcel may need to be replaced. Consequently, it incurs re-fill cost. However, the amount of
nutrients that is being washed away from a farm as well as re-fill cost are difficult to measure. In the same
manner, the number of trees/crop damaged by flood is not easy to quantify.
To overcome the above difficulty, a survey based approach is proposed adopting the model as
shown in equation (1). A sample survey needs to be conducted to collect data on the quantum of damage
of each property or item at a particular site. For landed properties such as residential, office, and
commercial, structural as well as content damages are taken as some percentages of property value. In
general, equation (1) can be re-expressed as:
EPD = SD + CD
= (.p1*ALV +.p2*ABV) +.p3*(ALV + ABV) (2)
where ALV = assessed land value; ABV = assessed building value; .p1, .p2, and .p3 = certain defined
“proportion” or “percentage” property component’s damage in decimal form. ALV, ABV, and any other
‘content’ asset can be estimated by replacement cost approach. Alternatively, market value (MV) of
property can be used in place of ALV and ABV if sales data are available.

32
 For agricultural properties, damage can occur to land/soil (structure) and tree/crop (content).
Again, it is difficult to ascertain damage to these elements. For compensation purposes, land/soil damage
can be estimated as a percentage of market value of a particular type of agricultural property but
tree/crop damage is much more difficult to estimate. The general formula for damage estimation of
agricultural properties with immature trees/crop is modified from equation (2) as follows:
EPD = SD + CD
= land/soil + tree/crop
t
= .q1*MV + n[(c-d)(1+i) ] (3)
where MV = market value of a particular type of agricultural property (alternatively, actual replacement
cost can be used); .q1 = a defined proportion in decimal form; c = cost of replacement new of the
tree/crop; i = discounting rate; t = age of immature crop; n = number of damage trees/crop.
However, this formula cannot be used directly without modification based on the type of
agricultural property under view. For example, damage to annual and perennial crop such as banana,
maize, rubber, oil palm, cocoa, and orchard trees need to be estimated by “individual” tree counting – a
daunting, if not impossible, task in FD-RAM. As another example, the immaturity period is different for
different crops. For instance, the immature period for oil palm is four years, rubber five years, while for
some orchard trees, this period may be up to seven years.
A sample survey in the disaster area is needed in order to compute the reasonable figures of all
the above damage components. Specifically, a priori information is needed to compute .p1, .p2, and .p3.

2.2 Rapid Damage Assessment Procedure

The whole procedure of rapid assessment of flood damage is part of the general concept of decision
support system promoted by Malczewski (1997). Ideally, it should become part of national disaster
management programs of any country troubled by the disaster. The actual implementation of flood
damage rapid assessment method is rather complex. It has two main components, namely mapping
component and spatial modelling component.
The mapping component has the following mapping activities: boundary of study area; and
distribution of poor population; sampling points to compute asset value, particularly building and land
value. Geographic Information System (GIS) is a standard method for flood mapping through various
kinds of software such as ArcGIS, MapInfo, Idrisi, etc. One of the most widely used GIS software is
Environmental System Research Institute’s (ESRI) ArcGIS 10.x.
The spatial modelling process has the following modelling activities: flood inundation
coverage/flood modelling based on rainfall-runoff method; spatial flood damage-estimating model; and
general damage estimate. Fundamentally, we can specify flood damage-estimating model in a number of
ways (Messner et al., 2007; Merz et al., 2010; Green et al., 2011). Factors such as flood depth, velocity,
duration, water contamination, precaution, and warning time can be included. However, inclusion of flood
factors cannot be generalized and is very much determined by data availability.
One potential spatial flood damage-estimating model is Geographically Weighted Regression
(GWR) originally developed by Fotheringham et al. (2000; 2002; 2005). Suppose we had some location in
the study area, perhaps one of the data points, where (x,y) are the coordinates of its position. We can
rewrite the model, in vector form as:

V(x,y)W = W(x,y)a + W(x,y)Z + W(x,y)e (4)

where V is value of damage, a is regression’s intercept, Z represents hydrological, physical,

environmental, and socio-economic variables/factors, W is spatial weight matrix, e is error term, and is
some measure of spatial component of data points. This relationship is fitted by least squares to give an
estimate of the parameters at the location (x,y) and a predicted value. This is achieved through the
implementation of the geographical weighting scheme. The weighting scheme is organized such that data
nearer (x,y) is given a heavier weight in the model than data further away.
Using OLS, the parameters for a linear regression model is obtained by solving:
T -1 T
β = (Z Z) Z V (5)

The parameter estimates for GWR are solved using a weighting scheme:

33
 T -1 T
β(g) = (Z W(g)X) Z W(g)V (6)

The weights are chosen such that those observations near the point in space where the parameter
estimates are desired have more influence on the result than observations further away. Two functions
we have used for the weight calculation have been (a) bi-square and (b) Gaussian. In the case of the
Gaussian scheme, the weight for the ith observation is:
2
wi(g) = exp(-d/h) (7)

where d is the Euclidean distance between the location of observation i and location g, and h is a quantity
known as the bandwidth. (There are similarities between GWR and kernel regression). One characteristic
that is not immediately obvious, is that the locations at which parameters are estimated need not be the
ones at which the data have been collected.
The resulting parameter estimates are mapped in order to examine local variations in the
parameter estimates. One might also map the standard errors of the parameters estimates as well.
Hypothesis tests are possible - for example one might wish to test whether or not the variations in the
values of a parameter in the study area are due to chance. The bandwidth may be either supplied by the
user, or estimated using a technique such as cross validation technique. The (x,y)s are typically the
locations at which data are collected. This allows a separate estimate of the parameters to be made at
each data point. The resulting parameter estimates can them be mapped.
Flood Loss Estimation Model for the private sector (FLEMOps) on the meso scale (Thieken et al.,
2008) is applied with some adaptation to the location situations. This model calculates the damage ratio
for residential buildings as a function of inundation depth classified into five classes and building
characteristics, i.e. three buildings types and two building qualities. To be applicable on the meso scale,
mean building composition and the mean building quality per municipality were derived and the resulting
damage ratios are multiplied by total asset values disaggregated to land use units (Thieken et al., 2005).
Spatially assessed flood damage by kriging technique is used in performing data analysis. A
modified Ordinary Least Squares technique, kriging adopts weights to the surrounding measured values
to derive a prediction for an unmeasured location. The general formula for both interpolators is formed as
a weighted sum of the data:
Zˆ ( S0 )  i 1 i Z ( Si )
N
(8)

where Zˆ ( S0 ) = weighted sum of values; Z ( S i ) = the measured value at the ith location; i = an unknown
weight for the measured value at the ith location; s0 = the prediction location; N = the number of
measured values.
In the kriging technique, the weights (represented by i ) are based on both the distance between
the measured points and the prediction location and also the overall spatial arrangement of the measured
points. To use the spatial arrangement in the weights, the spatial autocorrelation must be quantified.
In the ordinary kriging, the weight, i depends on a fitted model to the measured points, the
distance to the prediction location, and the spatial relationships among the measured values around the
prediction location. The following section briefly discusses how the ordinary kriging formula is used to
create a map of the prediction surface and a map of the accuracy of the predictions.
There are a number of kriging techniques discussed in the literature. However, to avoid
cumbersome discussion, we would only adopt ordinary kriging in this study. Ordinary kriging estimates
the unknown value using weighted linear combinations of the available sample (Isaaks and Srivastava,
1989):
n n
vˆ   w j * v w
i 1
i 1 (9)
j 1
The error of ith estimate, ri, is the difference of estimated value and true value at that same location:
ri  vˆ  vi (10)
The average error of a set of k estimates is:

34
 1 k 1 k
m   i k
k i 1
r 
i 1
vˆi  vi (11)

The error variance is:

2
1 k 1 k  1 k 
   ( ri  mR ) 2   vˆi  vi   (vˆi  vi )
2
R (12)
k i 1 k i 1  k i 1 
However, we cannot use the equation because we do not know the true value V 1,...,Vk. In order
to solve this problem, we apply a stationary random function that consists of several random variables,
V ( xi ) . Xi is the location of observed data for i > 0 and i  n. (n is the total number of observed data). The
~
unknown value at the location X0 we are trying to estimate is V ( x0 ) . The estimated value represented by
random function is:
n
V ( x0 )   wi *V ( xi )
~
i 1
~
R( x0 )  V ( x0 )  V ( x0 ) (13)
The error variance is:
n n n n
~ ~ ~ ~
 R2   2   wi w j Cij  2 wi Ci 0  2  (  wi  1) (14)
i 1 j 1 i 1 i 1
~2
 is the covariance of the random variable V(X0) with itself and we assume that all of our random
variables have the same variance while  is the Lagrange parameter.
In order to get the minimum variance of error, we calculate the partial first derivatives of the
equation (11) for each w and setting the result to 0. The example of differentiation with respect to w is:
 (~R2 ) n
~ ~ n
 2 w j C1 j  2C10  2   0
~ ~
w1 j 1
w C
j 1
j 1j    C10 (15)

All of weight wi can be represented as:

n
~ ~
w C
j 1
i ij    Ci 0 (16)

For each i, 1  i  n
We can get each weight W i through equation (13). After getting the value, we can estimate the
value located in X0. We can use variogram instead of covariance to calculate each weight of equation
(12). The variogram and minimized estimation variance are:
~ ~
 ij   2  Cij (17)
n
~
 R2   wi i 0  
i 1
The kriging module includes two variogram models:

Spherical
  h  h 
3

C  C 1.5  0.5   if |h|  a

~(h)    a  a  (18)

C0  C1 if |h|  a
Exponential

35
 0 if |h| = a
~ 
 (h)      3 | h | (19)
C0  C1 1  exp a   if |h|  a
   

Nugget effect (c0)

Though the value of the variogram for h = 0 is strictly 0, several factors, such as sampling error
and short scale variability, may cause sample values separated by extremely small distances to be quite
dissimilar. This causes a discontinuity at the origin of the variogram. The vertical jump from the value of 0
at the origin to the value of the variogram at extremely small separation distances is called the nugget
effect (Isaaks and Srivastava, 1989).

Range (a)

As the distance of two pairs increases, the variogram of those two pairs also increases.
Eventually, the increase of the distance cannot cause the variogram to increase. The distance which
causes the variogram to reach plateau is called range (see Figure 1).

Sill (C0 + C1)

It is the maximum variogram value which is the height of plateau (see Figure 1).

Distance h

It is the distance between estimated location and observed location.

C0+C1

C0
a

Figure 1: An Example of Exponential Variogram Model

Equation (16) can be written in matrix notation as V * W = D where V is (n+1) x (n+1) matrix
which contains the variogram of each known data. The components of last column and row are 1 and the
last component of the matrix is 0; W is (n+1) matrix which contains the weight corresponding to each
location. the last of component of matrix is Lagrange parameter; and D is (n+1) matrix which contains the
variogram of known data and estimated data. The last component of the matrix is 1.
Since V and D is known, we can get the unknown matrix W by W = invert(V) * D. Applying
equation (13), we can get the estimated value on a specific location. We also can get the error variance
from the square root of equation (17).

3.0 Methodology
A sample-based flood damage survey was conducted in early 2015 in Kuaka Krai and Dabong. This
study area was chosen because it was the most severely-hit sub-region of the state of Kelantan.
Furthermore, state-wide FD-RAM was not possible due to data and financial limitations. Sample-based
field inspections were conducted to estimate flood damage to buildings, trees, and other items. Since it
was very difficult to account for each item damaged by flood, this study was confined only to estimating
damages of residential and agricultural properties. Some moveable assets categorised as “contents” (e.g.
furniture, house appliance, equipment), were, however, accounted for. As many as 336 geo-referenced
sites (longitude and latitude in metres) within the flood inundated river corridors were sampled and
mapped as “survey points” shape file (see Figure 2).

36
 Survey points
Hard core poor

Survey points
Hard core poor

1:400,000

Figure 2: Survey points shape file in the selected study area. These survey points include
locations of some hard core poor’s homes (smaller dots)

Data on flood-related factors were collected at each sampled location, namely land value (asking
price) (RM/acre); building value (replacement cost new) (RM/unit); proportion of structural damage (%);
proportion of content damage (%); current use (forest, agriculture, natural vegetation, urban, transport,
built-up); use activity (rubber, oil palm, orchard, water body, road, vacant, residential); structural type (soil,
building); content type (tree, building, miscellaneous items); and flood depth (feet). All of the information
was formatted as attribute table of the “survey points” shape file in ArcGIS software. The purpose of this
shape file was to enable spatial modelling of flood damage using Geographically W eighted Regression
(GWR) technique based on the following specification:

TotDmg = f(Curuse, Activ, Structy, Contyp, Flo_dep) (18)

where TotDmg = Total flood damage (RM); Curuse = Current use; Acti = Use activity; Structy = Structural
type; Contyp = Content type; and Floo_dep = Flood depth (feet).

Damage estimation according to of property type is given as in equations (1) and (2) above.
Spatially assessed flood damage by kriging technique was used in performing data analysis.. Flood
damage was calculated as follows:

TotDmg = ContDmg1 + ContDmg2 + StrDmg1 + StrDmg2

where ContDmg1 = CD_P x Buildv x ef; ContDmg2 = CD_P x Landv x ef; StrDmg1= SD_P x Buildv x ef;
StrDmg2 = SD_P x Landv x ef. [ef = 1 IF sampled point = Residential/building; ef = 0 IF sampled point =
Agriculture/forest]

where CD_P = % of content damage; SD_P = % of structural damage; Landv = land value (RM/unit); and
Buildv = building value (RM/unit).
In the damage assessment process, the following guide was used:

ContDmg1 = content damage for residential/building

37
 ContDmg2 = content damage for agricultural crop/forest
StrDmg1 = structural damage for Residential/building
StrDmg2 = structural damage for agricultural crop/forest

Building value was estimated based on replacement cost new (RCN) of the original building. This
was a challenging process since RCN cannot easily and accurately be estimated. Although the ideal
method was to base value estimates on official government valuation, this was not possible due to
resource constraints. The regression procedure for the above specification followed the steps as outlined
in equations (4) through (17). GWR was run to relate flood damage (content & structural) with their
influencing factors, namely current land use (Curuse), land use activity (Activ), property structural type
(Structy), and flood depth (Flo_dep).
Once outputs were generated, superimposition process was performed whereby land use map
was overlaid on modelled flood, and GWR-kriged flood damage estimate. A manual process of
identifying, listing, and estimating damages of various types of properties was carried out using this
superimposed map. (See example in Figure 3.)

Figure 3: Screen shot of an overly of land use, modelled flood, and GWR-kriged flood damage estimate

38
4.0 Results and Discussion
Figure 4 shows flood hazard superimposed on GWR-kriged flood damage map over the study area.

Figure 4: Flood hazard superimposed on GWR-kriged flood damage map

Flood damage and, thus, flood risk is higher in densely populated locations such as urban or
residential areas. Besides, sites closer to river banks (say, less than 1 km) were mostly exhibited greater
flood depth. Other factors also contribute to the magnitude of damage.
The regression results are shown in Table 1. The performance of the GWR was very modest with
2
a local R of only 0.58. This reflects the shortcoming in modelling spatial relationship of flood damage
since, apart from land use factors, many other hydrological and geomorphological factors were not
included in the model specification due to data limitation.
From Table 1, flood depth was found to be significantly influencing flood damage. Content type
was also significant to property damage while other land use factors did not show statistical significance.
With respect to content type, miscellaneous contents of moveable property such as furniture and
appliances could have incurred damage of RM 31,221 more than other types of contents such as trees
and vegetation whenever there was flood inundation in the study area.

Table 1: General statistical results of Geographically Weighted Regression (GWR)

VARNAME VARIABLE DEFINITION
Bandwidth 26,115.340592
Residual Squares 506,075,220,856.614
Effective Number 15.870035

39
 Sigma 3,9697.86226
AICc 8,105.709384
Dependent: TotDmg Total flood damage (RM)

2
Local R 0.58
2
R Adjusted 0.56
Residual 1,480.77
Standard Error 38,692.17
Std. Residual 0.03
Sample size 336
95%
Coefficient Std. error t-value Min max confidence
Intercept -26,585.93 5,681.39 -4.68 -31,286.39 -16,581.87 332.69
Current use (Curuse) 24,545.44 21,285.12 1.15 18,093.14 42,160.54 574.24
Activity (Acti) 13,029.86 17,976.61 0.72 -4,936.35 21,905.10 523.05
Structural type (Structy) 8,150.71 19,128.44 0.43 -5,129.26 32,765.70 1,103.62
Content type (Contyp) 31,221.04 18,104.81 1.72 15,234.98 36,765.76 396.62
Flood depth (Floo_dep) 5,547.52 873.03 6.35 3,659.86 6,808.20 97.33

By manually using the GIS map, various types of properties were identified and listed together
with their corresponding damage (see Table 2). Many places were severely inundated, more than 70% in
some cases.

Table 2: Flood inundation over some selected land uses in the study area – GIS analysis
Area Area
Total Content
Total Structural Affected Affected
Land use inundated Approx Damage
area (ha.) Damage (structural) (content)
area (ha.) (%)
(%) (ha.) (ha.)
(%)
Kediaman:
Kampung Felda 310.16 92.26 30 0 45 0.00 1.45
Kampung Setinggan 0.67 0.18 27
Kampung Tersusun 112.36 90.33 80 55 61 0.76 0.85
Kampung Tradisi 147.21 128.23 87 70 72 6.00 6.18
Perumahan Strata 0.03 0.03 100
Perumahan Bukan Strata 56.89 43.01 76 80 94 1.00 1.17
Perumahan Kakitangan 10.54 10.27 97
Perumahan Ladang/Estet 42.11 11.52 27
Perniagaan dan Perkhidmatan:
Perniagaan Terancang 31.71 20.91 66 50 60 0.02 0.02
Perniagaan Tidak Terancang 30.26 22.17 73
Pertanian:
Getah 74233.19 46415.61 63 30 7 5326.82 1242.92
Kelapa Sawit 8825.22 5947.63 67 0 11 0.00 354.53
Padi 373.67 172.03 46
Dusun 5671.7 2795.32 49 63 90 3.71 5.30
Tanah Terbiar (Pertanian tidak 746.82 643.37
0.00 0.39
diusahakan) 86 0 5
Industri:

40
Industri Terancang 94.44 66.68 71
Industri Tidak Terancang 71.77 51.42 72
Infrastruktur dan Utiliti:

Bekalan Air 8.38 7.1 85

Bekalan Elektrik 352.68 270.79 77
Pengairan dan Perparitan 12.4 6.87 55
Telekomunikasi 2.06 1.11 54
Institusi dan Kemudahan
Masyarakat:
a) Keagamaan
Masjid 21.45 10.46 49
Surau 2.6 1.31 50
Tokong 0.36 0.36 100
Kuil 0.27 0.16 59
b) Kegunaan Kerajaan/Badan
Berkanun:
Pejabat Kerajaan/Agensi Kerajaan 46.15 35 76 50 60 0.03 0.03
Badan Berkanun 37.5 37.4 100 50 60 0.03 0.03
c) Keselamatan
Balai Polis 5.02 5.02 100
Balai Bomba 0.4 0.4 100
Pondok Polis 4.71 3.89 83
Kem Tentera 7.3 7.3 100
d) Kesihatan
Klinik Kesihatan 9.07 9.07 100
Klinik Desa 6.3 2.23 35 0 0 0.00 0.00
Hospital 4.23 4.23 100
e) Pendidikan
Tadika 4.82 2.02 42
Sekolah Rendah 130.07 91.36 70 30 50
Sekolah Menengah 55.8 49.24 88
Sekolah Agama 13.76 9.31 68
Institut Latihan 0.33 -
f) Perkuburan
Islam 46.3 34.75 75 60 65
Cina 11.06 11.06 100
Hindu/Sikh 0.58 0.58 100
g) Lain-lain Kemudahan
Masyarakat
Balai Raya 1.75 1.25 71
Dewan Serbaguna Awam 0.79 0.26 33
Dewan Orang Ramai 0.95 0.67 71
Perpustakaan Awam 1.66 0.39 23
Lain-lain: 0.85 0.21 25
Pusat Rukun Tetangga Kuala Krai

41
 Pusat Aktiviti Rukun Tetangga
Pusat Sumber KEMAS / Pusat
Literasi Komputer
Pusat Kominiti Desa
Dewan Rukun Tetangga Taman
Gucil Jaya
Pengangkutan:
Jalan 1338.65 1082.19 81 60 56
Stesen Bas 0.29 0.29 100
Stesen Keretapi 4.36 4.36 100
Penternakan dan Akuakultur 44.67 2.37 5
Tanah Kosong 836.17 633 76 63 46
Hutan 128916.39 70860.78 55 0 0
Tanah Lapang dan Rekreasi 900.69 835.07 93 55 54
* Expressed as number of units rather than area of land (ha.)
No data were available on the map

To further illustrate the use of FD-RAM, Figure 5 took a group of hard core poor people as a
case. The map indicates that the hard core poor group experienced low to severe flood damage. Most of
them experienced a total flood damage of about RM 10,000/household. This a was quite small figure and
was not surprising as many of them did not own high-value property. Nonetheless, this damage was
about 26 times their monthly income and can be considered a huge suffering for a hard core poor family.
The model, however, suffered from prediction inaccuracy and, thus, overstressing on damage figure may
not be desirable due to possible over- or underestimation in the assessment process.
Not all of hard core poor in the study area were affected by flood and, thus, those hit must be
identified. This was done by picking the affected hard core poor’s homes from the map via clipping menu
available in ArcGIS. In this case, modelled “flood polygon” layer was clipped onto “survey points” layer.
The resulting clipped layer was then superimposed on another layer, namely kriged estimated total flood
damage (ETFD). Figure 5 shows the locational distribution of hard core poor which was superimposed
over kriged values of estimated total flood damage (ETFD) modelled using Geographically Weighted
Regression based on equation (17). By this way, the hydrological and physical aspects of flood were
factored into flood damage-estimating model.

42
 ETFD  RM 10,000
20,000

ETFD < RM 2,000

20,000

ETFD  RM 10,000
20,000 ETFD  RM 10,000
20,000

ETFD  RM 45,000
20,000 Scale 1:600,00

ETFD  RM 28,000
20,000
Figure 5: Kriged value of estimated total flood damage (ETFD) based on Geographically
Weighted Regression among hard core poor (black dots) in the study area.
Figures shown are middle-values of ETFD.

1:400,000

Figure 6: Identification of flood-hit hard core poor by ArcGIS procedure

43
4.0 Conclusion
Although accurate estimate was not the focus of this study, being able to derive some initial figure of flood
damage is an important aspect of emergency relief and recovery program by the authority. The ability of
knowing the ‘possible’ amount of damage at a specific site is an additional useful piece of information to
the government.
The usefulness of rapid damage assessment of flood disaster largely depends on the
completeness of data and accuracy of damage-estimating model. The correct GWR model specification
that will result in satisfactory results was rather difficult and the available body of literature was not that
useful to identify all the correct variables to include. Trial and error specification and test of the candidate
variables such as those of geomorphological, hydrological, physical demanded a lot of data collection that
was not possible due to resource constraint.
Accurate identification of ‘itemised objects’ affected by flood is always a problem of flood damage
estimation. In this study, only content and structural damage of certain types of property/asset were quite
conveniently accounted for their respective owners their respective owners their respective owners.
Moveable assets such as vehicle, machinery, agricultural tools, etc. were not easily taken into account for
various technical reasons. Assignment of damages of crops and animals to their respective owners was
also difficult especially for those whose properties/assets were located on different sites away from their
living premise.
Estimating flood damage was very challenging particularly in choosing the most appropriate
approach of valuation. Cost, market and investment approaches are legitimate bases of asset valuation
but none can be suitable for all situations and for all property types. Detailed examination of the property
is thus necessary before deciding on the appropriate approach to valuation. This was simply not possible
in rapid damage assessment procedure.

References
Abd Jalil Hassan and Aminuddin Abd Ghani (2006). Development of flood risk map using GIS for Sg. Selangor Basin.
Accessible at http://www.redac.eng.usm.my.html.
Azura Abas (2015). RM78mil to clean post-flood Kelantan. New Straits Times, 7 January.
Fotheringham, A.S., Brunsdon, C., and Charlton, M.E. (2000). Quantitative Geography, London: Sage.
Fotheringham, A.S., Brunsdon, C. and Charlton, M (2002). Geographically Weighted Regression the Analysis of
Spatially Varying Relationship. John Wiley & Sons, LTD.
Fotheringham, A.S., Brunsdon, C. and Charlton, M. (2005). Geographically Weighted Regression. ESRC National
Centre for Research Methods, June, University of Leeds.
Green, C. H., Viavattene, C., and Thompson, P. (2011). Guidance for Assessing Food Losses. CONHAZ Report,
Flood Hazard Research Centre – Middlesex University, Middlesex, WP6 Report.
Isaaks, E. H. and Srivastava, R. H. (1989). An Introduction to Applied Geostatistics.
Merz, B., Kreibich, H., Schwarze, R. and Thieken, A. (2010). Assessment of Economic Flood Damage. Natural
Hazards and Earth System Science, 10: 1697–1724.
Messner, F., PennningRowsell, E. C., Green, C., Meyer, V., Tunstall, S. M., and van der Veen, A. (2007). Evaluating
flood damages: guidance and recommendations on principles and methods, FLOODsite, Wallingford, UK,
T09-06-01.
Malczewski, J. (1997) Spatial Decision Support Systems, NCGIA Core Curriculum in GIScience,
http://www.ncgia.ucsb.edu/giscc/units/u127/u127.html, posted October 6, 1998.
Poser, K. and Dransch, D. (2010). Volunteered Geographic Information for Disaster Management with Application to
Rapid Flood Damage Estimation. Geomatica, 64(1): 89-98.
Pradhan, B. (2009). Flood Susceptible mapping and risk area delineation using logistic regression, GIS and remote
sensing. Journal of Spatial Hydrology, 9 (2): 1–18.
Thieken, A. H., Muller, M., Kreibich, H. and Merz, B. (2005). Flood damage and influencing factors: New insights
from the August 2002 flood in Germany. Water Resources Research, 41(12): W12430+, 2005.
Thieken, A. H., Olschewski, A., Kreibich, H., Kobsch, S., and Merz, B. (2008). Development and evaluation of
FLEMOps a new Flood Loss Estimation Model for the private sector. In D. Proverbs, C. A. Brebbia, and E.
Penning-Rowsell, editors, Flood Recovery, Innovation and Response I.

44
 A FEASIBILITY STUDY OF A DISASTER MANAGEMENT HUB USING PREDICTIVE-ANALYTICS
AND MODELLING TECHNIQUE BASED ON CROWDSOURCED DATA

Project Information
Project Leader : Md. Nabil Ahmad Zawawi
University : Universiti Tenaga Nasional
Address : College of Computer Science and Information Technology, Universiti Tenaga
Nasional, Jalan IKRAM-UNITEN, 43300 Kajang, Selangor
Contact number : 013-3790907
Email : MdNabil@uniten.edu.my
Project Members : Norziana Jamil

1.0 Introduction
Information are crucial to assess situations during disaster time. Furthermore, a clear assessment and
judgment call can be established as a result of a clear flow of information. The use of social media has
enabled its users with near real-time update of happenings around the globe. One of the benefits of a
socially connected community is that pools of user generated information are abundant during an event of
a natural disaster. Studies from recent disaster events such as the Queensland and Australian floods, the
Christchurch earthquake and the Japan earthquake have shown that crowd source information could be
used and can be treated as the first response point to gain important information of the disaster in terms
of crowd, aid and recovery. In our study, we have examined the possibility of applying localized crowd
source information pool to manage flood disaster in Malaysia. The study considered the feasibility of
implementing a crowdsourced based on two factors which are community preparedness and
infrastructure readiness with regard on the use of the Ushahidi framework as the crowd source engine.
We also presented a standardized operating procedure that could be applied in conjunction with the
implementation of the proposed framework to effectively manage the situation should the need arise.

2.0 Methodology
Briefly, our study followed this for steps extensively:
1. Identification and preparation phase
- Identify flood disaster management requirement based on prevention, preparedness,
response and recovery of previous event.
- Identify participants ranging from volunteers, NGO and local authorities to set up the test
case crowd sourcing framework.
- Identify proper method of information verification for each pool of data collected based on
previous events.
2. Framework deployment based on the identified requirement
3. Simulated testing based on scenarios.
- Alpha testing
- On site testing with collaborators
4. Feasibility Analysis
- Produce a manual of standard operating procedure for crowd source volunteer and also
the authorities on how the collected data need to be used and utilised during the actual
event.

3.0 Results and Discussion

3.1 The crowdsource hub could assist the related authorities to visualize and assess the
situation on the field in near real time.
3.2 Reports visualization are done based on the number of response for call and rescue,
supply and resource allocation and the capacity of the relief centers. This could greatly
enhance the decision making of the officials during resource allocation and rescue effort
during the disaster. It can also assist the restoration works based on the visualized
information displayed on the map.
3.3 Although it is based on the concept of crowdsourcing, the center of control should still be
maintained by the authority. Proper standard operating procedures still need to be
followed according to the MKN arahan 20 document.

45
 3.4 Its constraints are the portability of the system and accessibility to the application for
technologically challenged area. This must be considered for future enhancement and
research possibilities.

4.0 Conclusion
The hub can be utilized based on gathered information from the community and also the relevant
authorities to visualize the situation and channel relevant assistance effectively during disaster. Although
the main focus of the paper is for flood disaster, the hub can also be used for other disaster under the
MKN 20 jurisdiction that can benefit the community and also nation as a whole. The factors affecting its
feasibility are mainly technological and logistics. Those factors must be considered to ensure its optimal
execution during disaster.

References
Bessaleva, L. I and Weaver A.C (2013). Applications of Social Networks and Crowdsourcing for Disaster
Management Improvement. SocialCom 2013. IEEE Computer Society. (pp 213-219)
Degrossi, L. C., de Albuquerque, J. P., Fava, M. C., & Mendiondo, E. M. (2014). Flood Citizen Observatory: a
crowdsourcing-based approach for flood risk management in Brazil. In Proc. of the 26th Int. Conf. on Soft.
Engineering and Knowledge (pp. 1-3).
Goodchild, M.F.,(2007), Citizens as sensors: the world of volunteered geography, Geojournal 69: 211-21.
Kuhn, W. (2007), Volunteered Geographic Information and GIScience. NCGIA, UC Santa Barbara, 13-14 December.
MKN 20, Majlis Keselamatan Negara, Arahan No. 20 (Semakan Semula), Jabatan Perdana Menteri
McDougall, K. (2012), An assessment of the contribution of the volunteered geographic information during recent
natural disasters, Global Geospatial Conference 2012, 14–17 May, Quebec, Canada
McDougall, K. and Temple-Watts, P. (2012), The use of LIDAR and volunteered geographic information to map flood
extents and inundation, ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information
Sciences, I-4, XXII ISPRS Congress 25 August–1 September, Melburne, Australia, 251–256
Triglave-Cekada, M. and Radovan, D. (2013), Using volunteered geographical information to map the November
2012 floods in Slovenia, Natural Hazards and Earth System Sciences Discuss, 13,2753-2762.

46
RESQ BANJIR: A MOBILE APPS FOR EMERGENCY RESCUE, EVACUATION AND RELIEF CENTER
MANAGEMENT

Project Information
Project Leader : Assoc. Prof. Muhammad Rafie Hj Mohd Arshad
University : Universiti Sains Malaysia
Address : Pusat Pengajian Sains Komputer, Universiti Sains Malaysia, 11800 USM, Pulau
Pinang
Contact number : 04-6533616 (Pej), 013-4214669
Email : rafie@usm.my

1.0 Introduction
A quick disaster response for relief needs after such disaster is vital to alleviate a disaster’s impact in the
affected areas. In Malaysia, the management of disaster is executed through the committee system aptly
called The Disaster Management and Relief Committee (A. Bahari et.al, 2007). One of the main issues
faced during a disaster especially for the SOS notification is the difficulties that victims faced in sending
emergency requests from their current location due to infrastructure failure. Such limitation has led to the
difficulty in locating the victims and coordinating the required rescue operation with the rescue teams (N.
W. Chan, 1995, M. Abdul Malek, 2005, SMSBanjir, 2014). Such limitation led to the difficulty in locating
the victims, which affected the effectiveness of coordinating a rescue and evacuation operation.
In this research, we develop ResQ Banjir application which consist of Flood Rescue & Evacuation
Operation Management (FREOM) and Flood Relief Centre Management (FRCM) to assist in flood
disaster management. In these system, we exploits the existing and emerging technologies on smart
phones and tablet such as sensors, cameras, GPS, SMS, Location Based System (LBS) and Augmented
Reality (AR).

2.0 Methodology
The system was developed based on spiral model System Development Life Cycle (SDLC) approach,
where a number of prototypes were produced at the end of each cycle. Each prototype was reviewed and
improved until completed. In this research 5 different apps will be developed individually: SOS/SMS
banjir, Relief Center Management, Rescue & Evacuation Management (Server), Rescue & Evacuation
Operation and AR Tumpat. This research involves 5 stages to develop the proposed ResQ Banjir apps:
system analysis, existing apps enhancement and extension, design and development of new module,
pilot implementation and testing, and finally assessment and documentation stage.
The initial stage of the development will focuses on the understanding of the existing rescue
operations and relief center management with its SOPs. In this stage, users system needs will be
identified and analysis of the present system will be done. Next, customization, enhancement and
extension of the existing apps particularly “Sistem Penjejakan Jemaah Haji Hilang SMS/GPS” and
“Sistem Penjejak Lokasi Makkah AR” will be done based on the users requirements identified. Unit testing
will be conducted after the enhancement completed.
For the development of Relief Center Management; we will develop it based on Sahana system
architecture and database design. This will enable the apps to be integrated with Sahana especially for
data sharing. Users system needs will be identified and analysis of the present system will be done. A
total of 11 modules will be developed. Debugging and unit testing process will be involved in the new
module development stage. After the completion of system testing; a pilot implementation on the
proposed system will be conducted. During the pilot implementation, the proposed system will be tested
from various aspects such as performance testing, usability testing, functionality testing, security testing,
etc. The pilot testing was done in Tumpat, Kelantan with the assistant of Pejabat Tanah & Jajahan
Tumpat, Kelantan. Finally, in the last stage of the research, an evaluation report, system manual and
documentation will be prepared.

47
3.0 Results and Discussion
We have developed the ResQ Banjir System consists of two major applications, which are:
1) the Flood Relief Center Management (FRCM), and
2) the Flood Rescue & Evacuation Operation Management (FREOM).

The FREOM system is a web and mobile apps based system consist of the Rescue & Evacuation
Management System (server) for rescue and evacuation operation monitoring & management at the flood
operation centre; SOS/SMS Flood apps for sending SOS help by victims; ResQ Banjir Skuad
Penyelamat apps which is used by rescue teams to locate and rescue flood victims, and an Augmented
Reality (AR) guidance system for location direction navigation guide. The system module consists of the
following functions:
- support rescue and evacuation activities;
- flood victims are able to alert rescue and evacuation center their GPS position via SMS
(show on map)
- tracking of rescue unit position (show on OSM map)
- tracking of rescue vehicles position (show on OSM map)
- direction and distance guide using AR and smart phone compass technology
- navigation guide
- database of flood information and point of interest (POI) which will be display on the map.
- coordinates rescue and evacuation operation.

FRCM is a web based system which integrates a number of features such as flood relief center registry,
shelter activation and management, inventory management and disbursement, and relief aid and goods
supplies management. The FRCM consists of the following functions:
- Registration – relief center, staffs, volunteers, victims
- Relief Aid and Goods Supplies
- Asset, Resource and Inventory
- Relief and Human Resource Request
- Shelter Activation and Management
- Information and resource sharing between relief center
- Dashboard info.

We conducted a simulation testing that involves the overall capability of the applications in Jajahan
Tumpat, Kelantan. Both systems seem to be working in an efficient manner although there are some
issues that exists. The two system (FREOM and FRCM) works in tandem during a rescue and evacuation
operation. The system’s mapping component is capable of showing an overall map of the flood affected
area, provide additional map information such as POI (point of interests), favorite points and access to
specified places in the area. The system can also provide compass support and navigation guide to the
identified places/victims. For tracking the flood victims, a victim can send an SOS alert via their
Subscriber Identity Module (SIM) card number as its identity, along with other data such as their GPS
coordinates and message to the Rescue & Evacuation Management System server using a mobile
phone. The FREOM application provides two alternative connections for the user to use depending on
the available communication channels: mobile-internet or SMS. The system then indicates the nearest
rescue team to the victim and identifies the suitable team for the current operation. The central operation
centre would then send a notification on the location of the victim as well as the instruction to proceed
with the rescue operation.
The FREOM system would also send a notification to the victim by providing the status of the
rescue operation such as estimated arrival time and vehicle type. During the rescue operations, the
devices utilized by the rescue teams would send intermittent location notifications to the FREOM system
that enables the operation centre to track the team’s locations at any given time. Such tracking capability
would enable a more efficient rescue operation to be executed based on the victim’s location.
Once the rescue team arrives at the victim’s location, they could trigger a safety notification to the
FREOM system. The FRCM system would then notify the rescue team the location of the closest shelter
that could receive the victims. The system identifies the shelters based on relevant information such as
available space and necessary resources that would allow sufficient assistance to be provided to the

48
victims upon arrival to the centre. Such information would mitigate the issue of overflowing rescue centres
and low food / medical supplies that may affect the wellbeing of the rescued victims.
In order to better support the rescue mission, we developed an Augmented Reality (AR)
technology based apps; AR Tumpat. This application can help the rescue team and associates for better
tracking the flood victims. This application will show the direction and distance of the target object.
Besides that, it can also provide information about the target object. At the same time, the flood victims
can use this application to find the nearest rescue centres and other point of interest (landmarks) around
them.
For the development of Relief Centre Management application, we developed it based on the
Sahana system architecture and database design. This would enable the application to be integrated with
the Sahana architecture especially for wide data sharing between centres. The Flood Relief Centre
Management application is a standalone decentralized management system. The data entered into the
tablet will be sent/updated to the database server of Disaster Management System at the state or national
level whenever internet service is available. This intermittent updates will allow a centralized data
collection about existing relief centres to be collected. The data collection enables an effective
coordination of the relief operation at a higher management level. Within this application, an application
Info dashboard would be developed that allow users to get real time information about the relief centres
such as statistical information, current needs and more. The application would also allow users to input
information about the flood or response to the needs of the relief centre. These information will be
updated to the application through the Application server.

4.0 Conclusion
It is predicted that there will be a positive impact on the effectiveness and efficiency of rescue operation
during flood with the implementation of the ResQ Banjir apps. Our application is capable to operate on
mobile devices such as smartphones and tablets which are more suitable for most rescue and evacuation
operations. With the use of SMS, GPS and Augmented Reality (AR) technology on smart phones;
locating victims and tracking rescue teams will speed up the rescue and evacuation operations; besides
improving the coordination of the rescue and relief efforts.
The Relief Center Management apps could enhanced the operations of the center. With the use
of ICT, related decision making processes would be more effective as various information such as the
status of relief centers and surrounding areas could be updated in real-time to various rescue agencies
and to the public. With the support of the system, most of the issues faced by Relief Center Management
such as shortage of food, waters and medical supplies, crowded evacuation centers and inaccurate
information could be mitigated or even solved.

References
A. Bahari, Z. Azhar, Z. Jamal, K. Hussin, Government Mechanism in Managing Natural Disasters in Malaysia:
Achievement Factors; In Proceedings of the 4th International Conference on Disaster Prevention and
Rehabilation 2007 (September 2007) http://dspace.unimap.edu.my:80/xmlui/handle/123456789/2427.
J.Zander, P. J. Mosterman. ‘Mode-based Design of a Smart Emergency Response System.’ Internet:
http://www.designnews.com/author.asp?doc_id=274577, 28 August 2014, [20 August 2015].
M.Abdul Malek, "Disasters Management System in Malaysia", JURUTERA, September, 2005.
http://dspace.unimap.edu.my/dspace/bitstream/123456789/13828/1/010‐012_disaster%20mgmt.pdf
N. W. Chan, (1995), A Contextual Analysis of Flood Hazard Management in Peninsular Malaysia, Ph.D. Middlesex
University (UK)
Sahana Software. ‘Sahana Eden.’ Internet: http://sahanafoundation.org/products/eden/ , 2014, [10 April 2015]
SMS Short Message Services (SMsBanjir) Flood Warning System for TTDI Jaya, Shah Alam and Its Surrounding
Areas (Affected By Sungai Damansara), Department of Irrigation and Drainage(DID), 2014.

49
 SISTEM DAN APLIKASI MOBILE PANTAS AMARAN BENCANA BANJIR

Project Information
Project Leader : Prof Madya Dr Hj Issham Bin Ismail
University : Universiti Sains Malaysia
Email : issham@usm.my
Project Members : lzham Bin Mohamed

1.0 Pengenalan
Bencana banjir yang berlaku telah meninggalkan kesan yang mendalam kepada mangsa. Kerosakan
harta benda dan kehilangan nyawa orang yang tersayang terjadi kerana mangsa gagal mendapat
maklumat pantas dan tepat tentang bencana banjir yang akan melanda. Pihak berkuasa tidak
mempunyai mekanisma pantas untuk memaklumkan status dan amaran banjir. Siren yang dibunyikan
tidak difahami dan tidak meliputi semua kawasan banjir, manakala makluman melalui tv dan radio gagal
memberi apabila tiada bekalan elektrik. Ditambah pula maklumat yang disampaikan hanya berbentuk
satu hala sahaja sedangkan mangsa juga perlu menyalurkan maklumat pantas seperti meminta bantuan.
Pada ketika ini terdapat kajian yang telah dijalankan menggunakan kaedah yang hampir sama iaitu
menggunakan SMS utk memaklumkan amaran banjir namun sistem tersebut hanya bergantung kepada
SMS satu hala sahaja dan ini merupakan antara faktor kajian lepas gagal menjadi realiti. Sedangkan
sistem yang dibangunkan ini boleh berhubungan dengan sistem kawalan banjir agensi kerajaan yang
sediada dan iaberhubungan secara dua hala, iaitu boleh menghantar SMS kepada bakal mangsa banjir
malahan is membenarkan mangsa banjir menghantar SMS kecemasan kepada sistem bagi
membolehkan mereka menerima bantuan. Pada ketika ini aplikasi mobile yang ada hanya berbentuk 'pull'
iaitu semua maklumat mestilah diminta oleh pengguna dan maklumat yang diberi juga tidak semuanya
relevan kepada pengguna. Sistem yang dibangunkan ini menggunakan Sistem Pengkalan Data
Perkomputeran iaitu pengkalan data 'database' yang Iengkap untuk tujuan penghantaran yang lebih
cekap dan efisyen dan berhubungan secara langsung dengan aplikasi mobile yang telah dibangunkan ini.

2.0 Metodologi
Pembangunan sistem dilakukan dengan 2 fasa. Fasa pertama melibatkan pembangunan Pembangunan
Sistem Pengkalan Data Perkomputeran manakala fasa kedua adalah melibatkan pembangunan aplikasi
mobile untuk telefon pintar.
Pembangunan Sistem Pangkalan Data Perkomputeran dibangunkan dengan setelah semua data
yang terlibat dikumpul dan dibentuk semula. Pangkalan data yang digunakan untuk menyimpan semua
data ialah MySQL. MySQL dapat menampung data yang banyak dan pantas. Selain itu ia juga boleh
digunakan pada banyak platform.
Pembangunan Aplikasi mobile juga dibangunkan. Ini adalah untuk memudahkan bagi pengguna
telefon pintar. Aplikasi ini amat mesra pengguna. Pengguna juga boleh daftar baru dengan menggunakan
Aplikasi ini selain dari melihat rekod sms yang telah dihantar dan menghantar SOS.

50
 Rajah 1 : Proses Sistem dan aplikasi mobile pantas amaran bencana banjir

Bagi penggunaan SMS, satu modem PE, telah dicipta untuk tujuan menghantar SMS. Dengan
menggunakan Wavecomm Chip, modem ini dapat menghubungkan server dan pengguna melalui SMS.
Terdapat dua pengguna iaitu penduduk dan admin dalam sistem ini.

3.0 Hasil dan perbincangan

Web

Dengan adanya platform Web, pengguna dapat mendaftar,

mengguna dan melihat segala maklumat berkaitan banjir di
web ini. Pengguna juga boleh mengakses profil mereka
untuk tujuan konfigurasi.

Android/iOS

Pengguna juga boleh menggunakan aplikasi untuk

telefon pintar, mereka perlu masukkan perisian
banjir ini ke dalam telefon mereka terlebih dahulu.
Aplikasi ini juga mesra pengguna di mana ia
mempunyai antaramuka yang ringkas dan mudah
untuk digunakan.

Khidmat Pesanan Ringkas (SMS)

Untuk penggunaan SMS pula, anda perlu mendaftar
terlebih dahulu dengan menggunakan laman web
atau aplikasi. Anda juga boleh mendaftar secara
SMS dengan menaip REG [email] dan hantar ke
server. Untuk menghantar mesej kecemasan((SOS), anda boleh menggunakan format berikut SOS-

51
[msg]-[lokasi] dan hantar ke server. Semua maklumat disimpan pada satu pangkalan data bagi
memudahkan pengguna mengakses sistem walaupun berada dimana saja dan bila bila masa. Maklumat
yang disimpan telah dienkrip bagi tujuan keselamatan.

4.0 Kesimpulan
Dengan adanya sistem amaran banjir ini, perkara-perkara seperti berikut dapat dicapai: -

4.1 Pembangunan prototaip mampu mengurangkan risiko bencana banjir.

4.2 Amaran banjir lebih awal dapat menyelamatkan benda-benda penting daripada musnah
akibat banjir
4.3 Keselamatan harta benda dan nyawa dapat diselamatkan terlebih dahulu.

Rujukan
Ammar, N. S. (2010) – http://eprints2.utem.edu.my
Badri, M.A. & Halim, A.K., "Design of moving massage LCD display system (MMDS) via Short Message Service
(SMS) entry using Rabbit 2000 microcontroller," RF and Microwave Conference, 2008.
RFM 2008. IEEE International , vol., no., pp.81,85, 2-4 Dec, 2008 doi: 10.1109/RFM.2008.4897364
Chan Ngai Weng, (1995) "Flood disaster management in Malaysia: an evaluation of the effectiveness of government
resettlement schemes", Disaster Prevention and Management: An International Journal, Vol 4 Iss: 4, pp.22 –
29
Chan Ngai Weng, (1997) "Increasing flood risk in Malaysia. causes and solutions", Disaster Prevention and
Management: An International Journal, Vol. 6 Iss: 2, pp.72 – 86
Ghazali, (2006) "Comprehensive planning and the role of SDSS in flood disaster management in Malaysia", Disaster
Prevention and Management: An International Journal, Vol. 15 Iss: 2, pp.233 – 240
Izahanis, M. (2008) – http://eprints2.utem.edu.my
Norliza Katuk, Ku Ruhana Ku-Mahamud, Norita Norwawi, Safaai Deris, (2009) "Web-based support system for flood
response operation in Malaysia", Disaster Prevention and Management: An International Journal, Vol. 18
Iss: 3, pp.327 – 337
Parker, D. J. and Handmer, J. W. (1998), The Role of Unofficial Flood Warning Systems. Journal of Contingencies
and Crisis Management, 6: 45-60. doi: 10.1111/1468-5973.00067
Portal Bencana. (2015, January 1). Retrieved February 23, 2015, from http://portalbencana.mkn.gov.my/Portal
Publicinfobanjir.(2015, January 1). Retrieved February 23, 2015, from http://publicinfobanjir.water.gov.my/
Raihan Fitri Agustina, (2013) Hubungan Komitmen, Pengetahuan dan Sistem Operasi dengan Kesiapsiagaan
Menghadapi Bencana Banjir: Satu Kajian Kes di Pejabat Daerah Kubang Pasu. Masters thesis, Universiti
Utara Malaysia.
Zarron, A. & AzIan, M. (2010) – http://eprints2.utem.edu.my

52
 DEVELOPING LONG-RANGE, AUTO RECHARGEABLE RADIO SENSOR NETWORK

Project Information
Project Leader : Hassan Chizari
University : Universiti Teknologi Malaysia
Address : 02-04-01, N28a, Faculty of Computing, Universiti Teknologi Malaysia,
Contact number : 0142741033
Email : chizari@utm.my
Project Members : Shukor bin Abd Razak
Anazida binti Zainal
Mohd Aizaini bin Maarof
Maheyzah binti Sirat @ Md Siraj
Syed Zainudeen bin Mohd Shaid

1.0 Introduction
Predictive environmental sensor networks provide complex engineering and systems challenges. These
systems must withstand the event of interest, remain functional over long time periods when no events
occur, cover large geographical regions of interest to the event, and support the variety of sensor types
needed to detect the phenomenon. Current technologies, either suffer from limitations of sensor networks
such as low communication range and energy usage, or use large-size solutions which are both
expensive and hard for installation and maintenance.

2.0 Project Background

Early flood detection technologies are in two categories: telemetry and sensor networks. Most of the early
commercial solutions for flood detection is based on telemetry. In telemetry, each device on the field,
senses a phenomena and sends it to the central management. To have long range communication,
usually these devices are equipped with radio. These sensor devices usually do not have in-device
decision making or process on the sensed data. Moreover, they are not able to communicate with each
other to pass the data and have decision if anything happen to them (Basha, Ravela, & Rus, 2008). An
example of telemetry is (ICIMOD, 2012) where two very big devices have been used for sensing. In
contrast to telemetry, sensor network is a decentralized network, where sensors communicate with each
other to pass the data toward the main station. This modern technology is more robust to errors and
changes inside the network. However, there are several problems for current solutions on sensor
networks such as energy, short range communication and the size as discussed in following.
Wireless communication is short range and is not suitable when the area of monitoring is vast.
One way to tackle the problem of wireless communication is to used wired sensor network. This solution
is applicable in areas which cabling is possible and an instance of such solution is (Aziz, Hamizan, Mehat,
& Haron, 2009). The best way to tackle the wireless problem is to use radio rather than wireless. Although
the bandwidth of radio is limited compared to wireless, it provides much longer communication range.
(Basha & Rus, 2007) developed a radio sensor network, where each device has the size of a 1m2 with
fence around it for protection. It also needed to be connected to electricity power line. It is clear that such
work is not applicable in rainforests for flood detection. Another solution is provided by (Bielsa, 2012),
where sensors are able to survive for about a week. These devices are more practical when they are
used as mobile sensor devices for short term installation.

3.0 Methodology
In previous researches which have been done by this group, we developed an environmental sensing
device for indoor data collection. Nine devices have been created and tested in a green mosque for data
collection. In another research using the same concept, an indoor building inhabitance localization
method has been developed. Using the experience we have for developing sensor devices and the
available program for routing, the steps needs to be taken for current proposal is as follow:

1. To develop a low cost radio sensor device

2. To develop ad hoc network routing program in the device.
3. To develop energy maintenance program inside the sensor.
4. To evaluate the developed prototype.

53
4.0 Findings
All the objectives for this grant have been achieved. A new way of detecting flood based on the saturation
of soil moisture combined with AI algorithm to detect the possibility of flood. The features of the device
includes cheap radio sensor network, early flood detection on its origins and long lasting service and fault
tolerable against disasters.
The developed product features are: Early flood detection, Self - organized ad hoc network,
Hybrid with GSMS network for reliable message delivery, Long lasting network for long term area
monitoring, Online soil moisture presentation in web. The novelty in developing such product lies in
developing data collector nodes and the sink node, using the water saturation level in soil for predicting
the flood, hybrid network: radio sensor network and SMS packet delivery, highly optimized energy
efficient devices and packet optimization and compression for fast and energy efficient data delivery. The
developed early flood detection devices have been used in real environment using artificial rain
(firefighting pipelines and rain itself) and it showed that devices are able to detect the saturation of water
inside the soil and predict the flood.

FIGURE 1: The developed prototype for Data Harvesting FIGURE 2: The developed prototype for Data Collector
Device (DHD) Device (DCD)

FIGURE 3: DHD Schematic Board Scheme FIGURE 4: DHD Connection Scheme

54
 FIGURE 5: DCD Schematic Board Scheme

FIGURE 6: DCD Connection Scheme

The developed prototype contains three devices. The first one is Data Harvesting Device (DHD) which will be
installed ono the river basin to collect the soil moisture of the land (Fig. 1). For an area of 10km2 1000 DHD are
needed to be installed in the area. DHD is very small and cheap in price and can be easily be hidden from the eye of
trespassers. DHD is collecting the soil moisture with weather information and temperature of the soil. Then, using a
new algorithm compress the data and send it to nearest DHD in order to be passed to Data Collector Device (DCD).
DCD (Fig. 2), is the second developed device which will be placed close to the area of data collection. For an area of
10Km2, 50 DCD are needed. DCD collects the data sent by DHDs, compresses them and sends it to the server
computer to be analyzed for early flood detection warning system. The schematic presentation of developed
hardware for DCD and DHD are presented in Fig.3 to Fig.6.

55
5.0 Conclusion
In this project, an early flood detection system has been developed consists of two different devices for
data harvesting and collection as well as a sink for receiving data at the main server. The simulation and
evaluation of the devices showed that the developed prototype has the ability to collect and send the soil
moisture in order to be analyzed for flood warning system. The prototype won several innovation prizes
during the years 2015 and 2016 including INATEX 2015 Bronze Medal, MRCIE 2015 Silver Medal and
MTE 2016 Bronze Medal.

References
Aziz, I. A., Hamizan, I. A., Mehat, M., & Haron, N. S. (2009). Prototype Implementation of Flood Detection and Early
Warning System via SMS. Presented at the Proceedings of World Academy of Science Engineering
Technology.
Basha, E. A., Ravela, S., & Rus, D. (2008). Model-based monitoring for early warning flood detection (pp. 295–308).
Presented at the the 6th ACM conference, New York, New York, USA: ACM Press.
doi:10.1145/1460412.1460442
Basha, E., & Rus, D. (2007). Design of Early Warning Flood Detection Systems for Developing Countries (pp. 1–10).
Presented at the Information and Communication Technologies and Development.
Bielsa, A. (2012, February 10). Smart Water project in Valencia to monitor Water Cycle Management. Retrieved
March 3, 2015, from http://www.libelium.com/smart_water_cycle_monitoring_sensor_network/
ICIMOD. (2012, December 28). Prototype of Community-Based Flood Early Warning System Installed at Godavari
Knowledge Park. Retrieved March 3, 2015, from http://www.icimod.org/?q=9204
Krzhizhanovskaya, V. V., Shirshov, G. S., Melnikova, N. B., Belleman, R. G., Rusadi, F. I., Broekhuijsen, B. J., et al.
(2011). Flood early warning system: design, implementation and computational modules. Procedia
Computer Science, 4, 106–115. doi:10.1016/j.procs.2011.04.012
Libelium. (2011, September 5). Smart Water: wireless sensor networks to detect floods and respond. Retrieved
March 3, 2015, from http://www.libelium.com/smart_water_wsn_flood_detection/
Rus, D., & Basha, E. (n.d.). Wireless Sensor Network Provides Early Flood Detection for Underserved Countries.
Microsoft Research. Retrieved from http://research.microsoft.com/en-
us/collaboration/papers/early_warning_flood.pdf
Seal, V. (2012). A Simple Flood Forecasting Scheme Using Wireless Sensor Networks. International Journal of Ad
Hoc, Sensor & Ubiquitous Computing, 3(1), 45–60. doi:10.5121/ijasuc.2012.3105

56
 INTEGRATED RIVER BASIN MONITORING AND FLOOD WARNING SYSTEM

Project Information
Project Leader : Suriza Ahmad Zabidi
University : International Islamic University Malaysia
Address : Faculty of Engineering, International Islamic University Malaysia, Jalan
Gombak 53100 Kuala Lumpur
Contact number : 0361964484
Email : suriza@iium.edu.my
Project Members : Assoc. Prof. Dr Ahmad Fadzil Ismail
Assoc. Prof. Dr Sany Izan Ihsan
Dr Norazlina Saidin
Nor Bazilah Bopi (RO)

1.0 Introduction
Flood is a common natural disaster that occurs all over the world. Giant floods can bring massive
destruction and damages to the affected areas. During a state of emergency, time management is very
crucial in order to minimize the impacts of the disaster. Lack of data such as precise locations of affected
areas may cause interruptions and delay in conveying information and executing all the necessary
actions.
Flood has been studied under various considerations and methodologies such as wireless
sensors network (Chang et. al, 2006; Hughes et. al, 2003; DeRoure et. al, 2005), embedded system with
middleware (Hughes et. al, 2005), Internet-based real-time data acquisition (Chang et. al, 2002) and flood
modelling and forecasting (Creutin et. al, 2003; Sapphaisal, 2007; Zhang et. al, 2002). All these methods
are proposed to improve the accuracy of flood monitoring systems. In addition, implementation of an
improved real-time flood monitoring system can reduce the damage caused by floods (Islam et. al, 2014).
Monitoring water level, or river stage and rainfall rate at river basin give a huge advantage as early
predictions of flood occurrence can be done before the water reaches the general populations. Using the
data and information from the monitoring process, flood warnings can be generated to notify the public
about the likelihood of flood occurrence.
Integrated River Basin Monitoring and Flood Warning System is an integrated system with the
abilities to monitor coastal, river water level and rainfall rate as well as to give warnings when there are
high possibilities of flood occurrence to the related authorities and the public by exploiting all the
information. It is an interactive and informative system which can integrate different types of data from
Department of Irrigation and Drainage (DID) as well as tidal forecasting data from Department of Survey
and Mapping Malaysia (JUPEM) in order to deliver high performance and reliable information on the flood
prediction.

2.0 Methodology
This research used river stage and rainfall rate data from telemetry stations in Kelantan, which are
available on DID website. Hourly data from a total of 25 stations were retrieved from the website and
recorded in the database. Two solutions were provided throughout the research period; desktop-based
application and web-based application. Both solutions were developed using open-source software and
tools. Desktop-based application was developed using Visual Basic language in Visual Studio Community
2015 whereas the web-based application was assembled using PHP language and PostgreSQL was
used as the database.
A collection of online hydrological data which consists of Kelantan’s rivers water level from DID
can be retrieved from the department’s website. The regularly updated data are stored in a database, in
which it is then exported to be displayed in the desktop application and web application. Desktop
application is designed for the use of specific registered users while web application is meant for the
public access. Figure 1 illustrates the system’s overall structure.
Tide forecast tables for Kelantan waters, which obtained from Getting station, Kelantan, was
provided by JUPEM. The tables were then stored in the database. Flood warnings can be generated
based on the status of current river water level and rainfall rate.

57
 Figure 1. Overall Structure of the System

3.0 Results and Discussion

A desktop application and its web-based application, called ‘Integrated River Basin Monitoring and Flood
Warning System’ were programmed. The desktop application is made accessible only for registered users
with different levels; administrators and operative employees. An additional privilege to add new
registered users, edit or remove existing users is given to the administrator. Figure 2 shows the login
page of the desktop application, where only registered users with legal username and password can log
into the system. Users whom are intended to have access to the system need to contact the administrator
to get the username and password. Information of users such as name, position, staff ID and email
address are required upon registration.

Figure 2. Desktop Application's Login Page

Figure 3(a) shows the main page, in which the locations of the hydrological telemetry stations are
shown on a map along with some information of each station in a pop-up, as in Figure 3(b). In this figure,
information for Kuala Krai Station such as its GPS coordinates, district in which it is located, as well as
related river and river basin are popped-up.

58
 Figure 3:(a) Stations' Location on Kelantan Map and (b) Kuala Krai’s Station Information

In Figure 4(a), all the data obtained from DID website are displayed in a table. River levels are
monitored according to mainly three levels; alert, warning and danger levels while rainfall rate is
monitored according to certain values as per determined by DID. The ‘Status’ column will show the status
for the latest data updated. Status of the river level will change to ‘alert’, ‘warning’ or ‘danger’ if the river
level value exceeds the predetermined values of alert, warning and danger levels respectively. ‘Normal’
status will be shown if the river level stays below the alert level. Otherwise, if the station is offline, it will
show ‘no data available’ status. Rainfall rate is displayed in the unit of millimeter per hour (mm/h) and it is
collected accumulatively starting at midnight for each day as shown in Figure 4(b). ‘No Rain’ status will be
shown if the station does not receive rainfall since the midnight of the day. Status ‘Light rain’, ‘Moderate
Rain’, and ‘Heavy Rain’ will be updated if the rainfall rate of the day is in the range of 1-10 mm/h, 11 – 30
mm/h, and 31 – 60 mm/h respectively. If the rainfall rate exceeds 60 mm/h, status ‘Extreme Rain’ will be
shown.

Figure 4.(a) River Water Level Data and (b). Rainfall Rate Data

Tidal prediction is shown for a period of one week. Hourly heights predictions for Geting Station
are displayed along with times and heights of low and high water in Figure 5(a). Figure 5(b) shows the
weather forecast for Kelantan waters for a week.

59
 Figure 5:(a) Tidal Prediction Table and (b) Weather Forecast for Kelantan Waters

Figure 6(a) shows the main page of the web application. Users will be shown their current
locations in order to identify whether they are nearby the monitored areas. When clicking on the ‘Kelantan
Station’ button, users will be redirected to water level stations on Kelantan Map, which is set as the
default display (Figure 6(b)). Users can choose to view waterlevel stations, rainfall stations or tidal
prediction by clicking the options in the tabs. Besides, Users can get information about each station along
with its current value and status by hovering the cursor on the marker. Users can also zoom in by clicking
in the marker to view detailed location of the station. Some links to important portals and websites are
also provided so that users can get access to any other information fast and easy.

Figure 6:(a) Default Main Page and (b) Default View after clicking 'Kelantan Station' Button

Figure 7(a) shows the water level stations after clicking on the Water Level Data tab and Figure
7(b) shows information of Kuala Krai by hovering the cursor on the marker. Figure 8(a) shows the rainfall
stations after the Rainfall Data tab is clicked and Figure 8(b) shows the information of Kusial station and
its current rainfall rate value and status. For tidal prediction, users can choose to view the hourly tidal
predictions or times and heights of low and high waters by clicking on the buttons respectively. Figure
9(a) shows the hourly tidal prediction after clicking on the Tidal Prediction tab while Figure 9(b) shows the
times and heights of high and low water by clicking on its button.

60
 Figure 7:(a) Water Level Station and (b) Pop-up information of Kuala Krai Station

Figure 8:(a) Rainfall Stations and (b) Pop-up Information of Kusial Station

Figure 9: (a) Hourly Tidal Prediction and (b) Times and Heights of High and Low Water

4.0 Conclusion

4.1 The most suitable tools, software and equipment were determined through a detailed
investigation and comparison.
4.2 The compatibility of rainfall rate and water level data as well as tidal forecast data with
the selected database was identified.
4.3 The multiple types of data were integrated into one fusion centre.
4.4 Desktop application for integrated river basin monitoring and flood warning system was
developed.
4.5 An interactive and informative web-based application for integrated river basin monitoring
and flood warning system was built.

References
Chang, N. B., & Guo, D. H. (2006, April). Urban flash flood monitoring, mapping and forecasting via a tailored sensor
network system. InNetworking, Sensing and Control, 2006. ICNSC'06. Proceedings of the 2006 IEEE
International Conference on (pp. 757-761). IEEE.
Chang, Y. C., & Chang, N. B. (2002). The design of a web-based decision support system for the sustainable
management of an urban river system. Water science and technology, 46(6-7), 131-139.
Creutin, J. D., Muste, M., Bradley, A. A., Kim, S. C., & Kruger, A. (2003). River gauging using PIV techniques: a proof
of concept experiment on the Iowa River. Journal of Hydrology, 277(3), 182-194.

61
De Roure, D., Hutton, C., Cruickshank, D., Kuan, E. L., Neal, J., Roddis, R., ... & Zhou, J. (2006). FloodNet–
Improving Flood Warning Times using Pervasive and Grid Computing. Neal, J, Atkinson, P and Hutton, C.
Hughes, D., Greenwood, P., Blair, G., Coulson, G., Pappenberger, F., Smith, P., & Beven, K. (2006, September). An
intelligent and adaptable grid-based flood monitoring and warning system. In Proceedings of the UK
eScience All Hands Meeting (p. 10).
Islam, M. A., Islam, T., Syrus, M. A., & Ahmed, N. (2014, May). Implementation of flash flood monitoring system
based on wireless sensor network in Bangladesh. In Informatics, Electronics & Vision (ICIEV), 2014
International Conference on (pp. 1-6). IEEE.
Sapphaisal, C. 2007. Forecasting and warning flood, Civil Engineering Magazine, Engineering Institute of Thailand,
Thailand, Vol. 2, pp.38-49.
Zhang, J., Zhou, C., Xu, K., & Watanabe, M. (2002). Flood disaster monitoring and evaluation in China. Global
Environmental Change Part B: Environmental Hazards, 4(2), 33-43.

62
 INSPiRE: INTERACTIVE DAM SAFETY DECISION SUPPORT SYSTEM FOR FLOOD DISASTER
REDUCTION

Project Information
Project Leader : Prof. Ir. Dr. Lariyah Mohd Sidek
University : Universiti Tenaga Nasional
Address : Jalan Ikram-Uniten, 43000 Kajang Selangor
Contact number : +60192780324
Email : lariyah@uniten.edu.my
Project Members : Dr. Sivadass Thiruchelvam
Dr. Gasim Hayder
Dr. Chow Ming Fai
Rahsidi Sabri Muda
Hamdan Basri

1.0 Introduction
Dams are often referred to as monolithic hydraulic structures that serve human needs and activities.
Despite, dams can also impose risks to the public. In line with current global awareness of water security,
the aspect of dam safety has drawn increasing attention from the public as it constitutes the element of a
country’s national security. Loss of human life is generally accepted as the most important consequences;
therefore it often dominates dam safety decisions (Othman, 2006; Myers, 2002). It was also being
outlined by the Chief Government Security Office (CGSO) Malaysia which is an important arm under
Prime Minister’s Office of Malaysia, dams are considered as one of the 15 national strategic targets
(CGSO, 2013).
Dam safety programs are vital to minimize the impacts of dam failure and to ensure that all
related personnel are ready and equipped with action plans in the event of dam failure. Lifesaving
missions could only be successful if the relevant SOPs have been tested and refined with proper sense of
coordination between the dam owner and all other critical agencies (Lariyah et al., 2014).
Based on statistical analysis of 534 dam failures from 43 countries before 1974, it was reported
that earth-rock dam failure accounted for the largest proportion of all failures which related to 49%
overtopping, 28% seepage in dam body and 29% seepage in foundation (Graham, 1999). In Malaysia,
there are almost 51 recorded dams (60% of the dams are of earthfill type) under different ownerships
(Othman, 2006). To-date we have not experienced any failures with the dams. Nevertheless, as dams in
Malaysia continue to age, it is time that we take notice of the conditions of these dams from the safety
perspective. Therefore, it is very important to ensure the dam in operated at its designed water level. Few
cases of flooding due to dam spill or release has been reported worldwide and summarized in Table 1.

TABLE 1: Historical flooding issues due to dam operational release

Year Area / Dam Cause of Flooding Impacts
January 2009 City of Pacific,  Heavy rain and melting  112 residential and 10
Washington snow commercial properties affected,
(Mud Mountain Dam)  11,700 cfs peak outflow with damages of $5 million and
released $10 million.
May 2010 Nashville & Davidson  Major precipitation  Floodwaters on the
County, Tennessee across the region Cumberland and its main
(U.S. Army Corps of  Limited flood storage tributaries, including the Harpeth,
Engineers’ Cordell Hull, capacitates all three dams Red, Stones, Caney Fork, and
Old Hickory, and Obey Rivers.
Cheatham Locks
&Dams)
2011 Minot, North Dakota  Heavy summer rainfall  12,000 residents were
(Rafferty and Alameda caused the spillway gates as evacuated
Dams) Rafferty Reservoir were fully  USD 600M damage to
opened and releasing a peak flow property and infrastructure
of 26,000cfs into Minot
October 2013 Bertam Valley  Heavy rains in the  3 dead and one missing
(Sultan Abu Bakar Cameron Highlands brought on  Muddy waters washed
Dam)

63
 flash flooding caused an overflow away nearly 100 cars and
into Lake Ringlet that bring destroyed some 80 homes
maximum levels
November Bertam Valley  Heavy rain caused  Five died and five injuries
2014 (Sultan Abu Bakar increasing amount of domestic &  100 damaged houses
Dam) agricultural waste lower the
capacity

In order to mitigate the hydro hazard due to dam break, UNITEN has developed a new software
known as INSPIRE (Interactive Dam Safety Decision Support System). INSPIRE as intelligent Dam
Safety is developed to address emergency situations which demand fast, decision making and effective
multi-agency collaboration due to dam break event. The fundamental role of INSPIRE is to provide an
integrated system that may be used by dam operators as an interactive emergency response plan to
mitigate the risk of dam failure. Therefore, INSPIRE is a decision support tools aimed to support the
decision processes regarding dam safety event. The DSS enables simple, efficient & practical early
warning system to alert and notify all stakeholders in the event of dam related disaster. INSPIRE is aimed
to help to provide timely warning to reduce catastrophic impact to the downstream area which means that
the emergency management would be more efficient and more lives will be saved in a dam break event.
In the long run, INSPIRE will help the multi-stakeholders capacity for effective and efficient management
in the event of a dam related disaster. In terms of commercialization value, a market study has been
done and comparison was made to similar system available in the market as shown in Table 2.

TABLE 2: Comparison with other similar software in market

Criteria
Dam Does Not Require Database of Emergency
Software
Break Interactive Deep Engineering Dam Break Response
Modeling Knowledge Study Plan
DAMSAFE √ √ √
DSS WISE √ √
SAGE B √ √
Dam Break Emergency
√ √ √ √
Response Information System
INSPiRE √ √ √ √ √

2.0 Methodology
The dam safety information is different from one dam to another. The development of INSPiRE system
requires planning of user interface which allows flexibility of dam properties, to suit the different
requirement from the dam owner. Therefore, the development of INSPiRE software is designed as
integrated system to mitigate the risk of dam failure (Figure 1a), and will be based on the different
module, as shown in Figure 1. In order to allow system flexibility with several level of security, the
software was also designed to be equipped with certain level of permission with different authentication
level.

64
 FIGURE 1(a) : INSPiRE Concept

FIGURE 1(b): INSPiRE Modules

i) Module 1 (Database): Development of Database for INSPIRE Decision Support System (DSS)
INSPIRE Dam Safety DSS has develop dam safety database database for different dams. The list of
dams are as shown in Table 3:

TABLE 3 : List of Available Dam Information

Dam Dam Operator

Sultan Abu Bakar Dam

Jor Dam
Mahang Dam
Temengor Dam
Tenaga Nasional Berhad
Bersia Dam
Kenering Dam
Chenderoh Dam
Pergau
Bukit Merah Dam Drainage and Irrigation Department, Malaysia
Pedu Dam
Muda Dam Muda Agriculture Development Authority
Ahning Dam

Among the data available inside the database are as follows:

i. Dam Emergency Response Plan
ii. Dam Standard Operating Procedure
iii. Flood Hazard Map
iv. Dam Monitoring Equipment
v. Dam Design
vi. Dam failure mode & triggering level
vii. Dam Emergency Personnel

ii) Module 2 Failure Mode Analysis

Failure Modes Analysis has been used extensively in performing risk assessments. For a risk assessment
the probability of failure and the consequences of failure are used to calculate the risk of loss of life or
property damage for a particular dam or project. In order to evaluate the probability of failure for a dam,
the failure modes must first be identified. Identifying the failure modes is typically performed by a team of
experts who brain storm on different ways that the dam could fail. This brain storming includes identifying
the events that would occur leading to failure. The results are usually presented as event trees. Currently,
dam faliure modes for all dams listed above has been identified. In general, three main dam failure modes
was used to run dam break simulation are follows Flood Loading, Earthquake Loading and Normal
Loading / Clear Day Failure. Figure 2 shows the critical reservoir level based on overtopping dam failure
mode.

65
 FIGURE 2 : Example of Dam Failure Mode

iii) Module 3: Formulation of Emergency Response Plan (ERP)

Development of ERP for specific dam are following USBR or Hydro Tasmania standard. Currently, we
have obtained information of available dam safety guideline around the world. This enable us to
understand the concept of dam safety regulations parcticed by dam owners and operators. Table 4 lists
the dam safety guidelines from different countries.

TABLE 4: List of Dam Safety Guidelines

Guideline Country
Federal Guidelines for Dam Safety Emergency Inter-Agency Committee on Dam Safety
Action Planning for Dam Owners (ICODS), USA, 1998
Guidelines on Dam Safety Management Australia National Committee on Large
Dam Incorporated (ANCOLD).
On-site Plan for Reservoir Dam Incidents – Department of Environment, Food, and
Guidance on Reservoir Emergencies Rural Affairs (DEFRA), United Kingdom
(UK)
Guidelines for Development and Implementation Government of India, Central Water
of Emergency Action Plan for Dams Commission Dam Safety Organization,
2006.
Federal Guidelines for Dam Safety Emergency USA
Action Planning for Dam Owners (ICODS)
Guidelines on Dam Safety Management by AUSTRALIA
Australia National Committee on Large Dam
Incorporated (ANCOLD)
On-site Plan for Reservoir Dam Incidents – United Kingdom
Guidance on Reservoir Emergencies
Guidelines for Development and Implementation INDIA
of Emergency Action Plan for Dams

We have also made comparison between USBR and HTC format as shown in Figure 3, and also identify
countries that has develop the ERP similar to the ICODS and ANCOLD guidelines as shown in Table 4.

FIGURE 3: Comparison of Emergency Identification and Classification between USBR & HTC Format

66
iv) Module 4: Communication Module
In order to provide timely warning to dam owners and agencies in the event of dam safety emergency,
INSPIRE was designed with its own email system integrated inside of the application. Among the
features in communication module incluse:
 Multiple email recipients
 User authentication to avoid false alarm
 The information of user, location, water level date and time is taken directly from theapplicatio to
prevent any information error / unwanted data redundancy

Figure 4 shows screen shot of INSPIRE system under communication module.

FIGURE 4 : Screen shot of Communication Module

v) Module 5: Education Module

It is important that all dam personnel are familiar with their task and responsibility during dam safety
emegencies. Thus, Eduction Module provides animation video as part of the INSPIRE system. It also
aims to create awareness on dam safety to dam personnel. Figure 5 shows example of animation on
dam safety tutorial for Bukit Merah Dam.

FIGURE 5 : Dam Safety Animation for various emergency level

67
vi) Module 6 : Documentation
Documentation module provides dam safety Emergency Response Plan manual for user references.
User are able to upload the latest revision of dam safety manual into the system. Apart from that, this
module also provides informations such as relevant reports and dam break flood hazard map. This
module allows user to rely on one stop information rercources during emergency events. Figure 6 shows
screen shot documentation module.

FIGURE 6: Documentation Module

3.0 Results and Discussion

INSPIRE Dam Safety Decision Support System (DSS) has develop dam safety database for different
dams. Currently, INSPiRE is being developed for two TNB Hydro Schemes, which are Cameron
Higlands-Batang Padang Hydroelectric Scheme and Sungai Perak Hydroelectric Scheme, which consist
a total of 8 dams. The system was also developed for Muda Agricultural Development Authority (MADA),
who owns and operated 3 dams and Drainage and Irrigation Department (DID). INSPiRE will contribute
towards the sustainability of dam’s owner as corporate reputations can be ruined through dam structural
failures that can affect the economy of the nation and enhance the quality of life of the people. The 4 key
elements of INSPIRE are shown in Figure 7:

FIGURE 7: Key Elements of INSPIRE Software

Currently, we have obtained information of avalibale dam safety guideline around the world. This enable
us to understand the concept of dam safety regulations parcticed by dam owners and operators. Table 5
lists the dam safety guidelines from different countries.

68
 TABLE 5: List of Dam Safety Guidelines
Guideline Country
Federal Guidelines for Dam Safety Inter Agency Committee on Dam Safety
Emergency Action Planning for Dam (ICODS), USA, 1998
Owners
Guidelines on Dam Safety Management Australia National Committee on Large Dam
Incorporated (ANCOLD).
On-site Plan for Reservoir Dam Incidents – Department of Environment, Food, and Rural
Guidance on Reservoir Emergencies Affairs (DEFRA), United Kingdom (UK)
Guidelines for Development and Government of India, Central Water
Implementation of Emergency Action Plan Commission Dam Safety Organization, 2006.
for Dams
Federal Guidelines for Dam Safety USA
Emergency Action Planning for Dam
Owners (ICODS)
Guidelines on Dam Safety Management by AUSTRALIA
Australia National Committee on Large Dam
Incorporated (ANCOLD)
On-site Plan for Reservoir Dam Incidents – United Kingdom
Guidance on Reservoir Emergencies
Guidelines for Development and INDIA
Implementation of Emergency Action Plan
for Dams

TABLE 6 : List of Countries which adopts ICODS & ANCOLD Guideline

ICODS ANCOLD

USA Australia

Canada Malaysia (TNB)

India Philippines (San Roque Dam)

Turkey Fiji (Monasavu Dam)

South Africa

Jordan

Argentina

i. Product Testing & Verification

In order to ensure the software meets the user requirement, product testing and verification was done
with various dam owner and agencies. Table 7 list the product and verification activities and Figure 8
shows the pictures during the emegency exercise with dam owner and government agencies.

TABLE 7 : Software Testing & Verification

TYPE OF EMERGENCY
DATE DAM OWNER
EXERCISE

Sultan Abu Bakar Dam

1 Nov. 2012 Drill Exercise (Internal) (Cameron Highlands TNB
Hydro Scheme)

Temengor Dam
21 – 22 Oct. 2013 Drill & Tabletop (Agencies) (Sg Perak Hydro TNB
Scheme

Sultan Abu Bakar Dam

15-16 April 2014 Tabletop Exercise (Agencies) TNB
(Cameron Highlands

69
 Hydro Scheme)

14 Mei 2015 Tabletop Exercise (Agencies) Bukit Merah Dam DID

2 Sept. 2015 Functional Exercise SAB Dam TNB

Sultan Abu Bakar Dam

Evacuation Drill (Agencies &
28 Sep. 2015 (Cameron Highlands TNB
Community)
Hydro Scheme)

8 Oct. 2015 Tabletop Exercise (Agencies) Kenyir Dam TNB

27 Oct. 2015 Drill Exercise (Internal) Pedu Dam MADA

1-2 Dec. 2015 Tabletop Exercise (Agencies) Pedu Dam MADA

ii. Product Benchmarking

50,000
FIGURE 9 : Product Benchmarking

TABLE 8: INSPiRE Benchmarking

Criteria

Dam Break Interactive Detail Database of Emergency

Software Modeling Engineering Dam Break Response Plan
Knowledge Study

DAMSAFE √ √ √

DSS WISE √ √

70
 SAGE B √ √

Dam Break ER √ √ √ √
System

INSPiRE √ √ √ √ √

iii. Commercialization Strategy

Short term commercialization strategy was planned for INSPIRE software as shown in Figure 9.

FIGURE 10: INSPIRE Commercialization Strategy

4.0 Conclusion
INSPiRE has been successfully developed to address emergency situations which demand fast, decision
making and effective multi-agency collaboration due to dam break event. INSPiRE helps to provide timely
warning to reduce catastrophic impact to the downstream area which means that the emergency
management would be more efficient and more lives will be saved in a dam break event. In the long run,
INSPiRE will harness as well as help the multi-stakeholders capacity for effective and efficient
management in the event of a dam related disaster. Apart from that, important dam break information
such as flood arrival time, time to max flood depth, max flood flow velocity, max flood inundation depth,
and flood extent also included in the software. INSPiRE would be a powerful tool for the decision makers,
as it can help to save time and enhance the effectiveness of decisions.

References
Chief Government Security Office (CGSO). (2013). Background. Retrieved 30 January 2013, from
http://www.cgso.gov.my/~cgso/portal/index.php/en/faq.html
Dave Carlton, DKCarlton & Associates and Heidi Kandathil, H.A.Kandathil Planning Services. Samuel Johnson
(2013). A Strategy to Reduce the Risks and Impacts of Dams on Floodplains, Federal Emergency
Management Agency,
Graham, W.J., (1999). A Procedure for Estimating Loss of Life Caused by Dam Failure, U.S. Department of Interior,
Bureau of Reclamation, Dam Safety Office, Denver, Colorado
Lariyah, M.S, Hidayah, B., Sivadass, T., Rahsidi S.M., Azwin zailti A.R., and Zuraidah, A., (2014). Implementation of
Dam Safety Management Program in Malaysia : From Theory to Practise, 2nd International Conference on
Civil, Offshore & Environmental Engineering (ICCOEE2014), 3-5 June 2014, Kuala Lumpur - Malaysia.
Othman, Z.I. (2006). Overview of dam safety in Malaysia. Jurutera. January 2006, pp. 22-23.
Thestar.com.my,. 'Three Killed In Mud Flood After Water Released From Cameron Highlands Dam - Nation | The
Star Online'. N.p., 2013. Web. 6 Sept. 2015.

71
 INTEGRATED MOBILE SOLUTION FOR ALERT, SEARCH & RESCUE (A-SAR)

Project Information
Project Leader : Hema Latha Krishna Nair
University : Asia Pacific University of Technology & Innovation
Address : Technology Park Malaysia
Contact number : 011 3375 8850
Email : hema.krishna@apu.edu.my
Project Members : Amad Arshad
Mohamad Firdaus Che Abdul Rani

1.0 Introduction
The project aim is to enhance preparedness and response by improving coordination and enhancing the
capability of those organizations involved in flood rescue via early detection and integrated
communication using ICT. The project will provide an organizational capability and structure to enable the
delivery of a coordinated national response to flooding incidents, such that organizations can work
together to minimize the loss of life and injury and to reduce the physical and financial effects of
consequential loss and collateral damage.
The desired outcome from the Project is that organizations involved in Flood Rescue will have
additional capabilities and capacity to respond to major flooding incidents so that they will be able to:
i. Deal with the consequences of major flooding, by having the appropriate operational
capability ready.
ii. Maintain the capability over the long term so that it is available whenever required.
iii. Deploy resources swiftly once a flooding incident occurs
iv. Command, control and communicate at major flooding incidents, from assessment and initial
response, through to the management of recovery and reestablishment of preparedness
v. Introduce national mutual aid and standards such that the organizations can communicate
and coordinate information via single channel.

2.0 Methodology
Data Acquisition: Water Overflow Detection & Impact Data Collection:
The integrated system to detect water catchments volume and forecast overflow would be the first phase
of trigger in this solution. Besides analyzing data such as rainfall, River & Dam water level Data, the
regional population census data has to be updated on timely manner. Population data needs to be
classified according to their zone of habitant or their address. Mobile contact no would be one of the
important attribute selections. The Land & Flood Plain data will support the system in identifying the
possible zone of flooding, water flow direction and water channeling routed. By gathering these spatial
data, system would be able to mark zonal impact and further be prepared for mitigation plans according
to the hazard level. One of the most important data feed is the Access Route & Transportation allocation
data. Information such as access roads, public transportations, diversion route and reachability to the
population can be identified. This data integration would map the population and logistics, indirectly
support plans for evacuation.

Data Preprocessing: Selection, Cleaning, Transformation & Integration:

This phase is highly important as any missing data or noisy data could lead to inaccurate prediction.
Domain experts and department authorities are required to share their knowledge in data selection and
ensure that accuracy values are set at the optimum standards. All data from various sources will be
integrated under one schema. A centralized data warehouse will be created and data will be ready for
analytics. The data comes from various sources (cloud data, batch processed, real time, etc.).

Infrastructure & Data Ware House:

Supporting various types of data (structured, unstructured, spatial, etc.) could be overloading during data
request. Data analytics and mobile communication needs fast and prompt query processing over secured
channel. To support a good architecture at an affordable cost, cloud computing is adapted. Infrastructure
as a Service (IaaS) is a form of cloud computing that provides visualized computing resources over the
Internet. IaaS is one of three main categories of cloud computing services, alongside Software as a

72
Service (SaaS) and Platform as a Service (PaaS). Infrastructure as a Service gives us an advantage of
managing our Applications, Run times, Security & Integration as well as the database.

Data Analysis Engine:

The engine development is the functional area of the integrated solution. Online Analytical Processing
(OLAP) is a technique to visualize consolidated data and drill down into the specific details or roll up to
view the generalized summation. Prediction / Forecasting are a technique under supervised learning
which has predefined class labels. This technique can be well embedded to manage Water Overflow,
River Overflow and Speed, Evacuation Plan & Navigation, Resource Procurement & Distribution,
Volunteer / Assistance Requirement & Placement.

Modeling:
Data Visualization Tool (for Mobile & Web Portal) is available via open source channels and paid
enterprise tools. The best modeling tool would be essential to demonstrate data, let it be hypothetical or
precise data. The modeling tools should be able to synchronize with mobile alert, customized to the type
of data sent and the hazard level.

Disaster Simulation Test

Testing a real-time scenario could lead to tremendous cost and unnecessary lose. The best option for
disaster management test is simulating the process. Mobile alert & Communication can be tested by
pumping into test data and evaluation of the reach. Test data could be past data or real-time data which is
used and directed following a standard operating procedure. With the simulated data, a rescue exercise
can be planned.

3.0 Results and Discussion

The solution is incomparable to any existing solutions as it is customized to manage Flood disaster in
Malaysia, focusing on specific region. Nevertheless, the solution is developed dynamically to manage all
key users, locations, and basic functional areas in all flood disaster management. Thus, it is capable of
expansion of use in other regions in Malaysia or can evolve as a benchmark for any flood SAR worldwide.

Figure 1. A-SaR Components

73
Government Agency Unit Core Functionality:
1. Monitoring & Water Channeling : A precaution solution to warn the water levels in catchment areas,
forecasting possible overflow (flood alert) and proposing water diversion or channeling to distribute
water to dry area within the region.
2. Failing to manage overflow, Flood warning can be predicted before occurrence based on historical
data analytics. This encourages early warning to the affected areas, preparation of logistics and
immediate aids, and budget allocation by government.
3. Mobile / handheld devices will be an added value for this solution as it can alert each key user,
creating awareness even though they are not online with the system. This rule out all parties blaming
each other for not being effective in handling a disaster.

Voluntary / Other Unit Core Functionalities:

th
1. As described my MERCY (9 January 2015) report, basic assistance during the disaster focused on
Health Care, Water, Sanitation and Hygiene, Food and Nonfood items.
2. A centralized information center for media coverage & public awareness
3. Donation request and distribution. Compilation of items needed, location and quantity would be
identified and updated in real-time to avoid wastage & unauthorized campaigns

Special Features:
1. Blending the mobile disaster system application with features such as customizing information in
order to adapt each victims need or (i.e.: sending a mass general SMS SARS panic alerts to the
population without considering the user’s location, wealth condition / age and other peculiar
information could just generate more panic).
2. For End User
 Flood Disaster Safety Tips (FAQ containing recommended safety tips)
 Identify & Navigate nearest reachable Rescue/Evacuation Center (To recommend navigation
based on accessibility conditions)
 Flood Disaster Reporter (To propagate any potential disaster information to nearest
Rescue/Evacuation Centre)
 Emergency SOS (To send an SMS containing GPS Coordinates to a list of contacts)
 Tourist Helper (To confirm safety of a route)
3. For Government Staff
 Track Help Requests (To know location of people stranded in their territory)
4. Social Media enabled (RSS feeds, Facebook, Twitter, etc.) for immediate awareness to the nation.

By doing that the mobile disaster management system applications will probably become more appealing
to the public and as a result increase people’s trust and contribution on timely manner.

4.0 Conclusion
We cannot prevent the flood disaster; however we make better use of the current web and mobile
technologies and promote future advances. As mobile devices become more common nowadays, the
government has the chance to provide first responders and citizens with the tools necessary to save lives
during this threatening event. Hence this prototype hopefully can be considered to be one of the major
solutions to manage the disaster before, during and after the event.

References
CityPopulation (2014) Malaysia, [Online] Available from: http://www.citypopulation.de/Malaysia- UA.html [Accessed
on: 13 Nov 2014].
Department of Statistics Malaysia (2014) Selangor @ a Glance, [Online], Available from:
http://www.statistics.gov.my/portal/index.php?option=com_content&view=article&id=533%3A selangor-a-
glance&catid=115%3A-glance&Itemid=1&lang=en [Accessed on: 9 Nov 2014].
GenerateData (2014) Generate, [Online], Available from: http://www.generatedata.com/ [Accessed on: 25th Nov
2014].
Hoong, Tan, and Mustafa, (2007) Contamination Levels of Selected Organochlorine and Organophosphate
Pesticides in The Selangor River, Malaysia between 2002 to 2003, Chemosphere, [Online], 66 (1) p.1153-
1159.

74
MapCustomizer (2014) MapCustomizer, [Online], Available from: http://www.mapcustomizer.com/ [Accessed on: 31
Oct 2014].
New Straits Times Online (2014) Mud Flood Tragedy Lesson to Residents in Bertam Valley, 6th November, [Online],
Available from: http://staging.nst.com.my/node/50234 [Accessed on: 9 Nov 2014].
Ministry of Natural Resources and Environment (2014a) Drought Monitoring by Dam Levels, [Online], Available from:
http://infokemarau.water.gov.my/drought_monitor.cfm [Accessed on: 1 Nov 2014].
Sandaysoft (2013) Rain Rate, [Online], Available from: http://wiki.sandaysoft.com/a/Rain_measurement [Accessed
on 31 Oct 2014].
World Health Organization (2014) Water Sanitation Health, [Online], Available from:
http://www.who.int/water_sanitation_health/emergencies/qa/emergencies_qa5/en/ [Accessed on: 13 Nov
2014].
Ministry of Natural Resources and Environment (2014b) Dam Station for Drought Monitoring, [Online], Available
from: http://infokemarau.water.gov.my/drought_front.cfm [Accessed on: 1 Nov 2014].

75
 MODELING THE EFFECT OF LAND USE/COVER CHANGES IN FLOOD SOURCE AREAS ON
DOWNSTREAM FLOOD PEAK OF KELANTAN RIVER BASIN

Project Information
Project Leader : Wan Nor Azmin Sulaiman
University : Universiti Putra Malaysia
Address : Department of Environmental Science, Faculty of Environmental
Studies, UPM, 43400 UPM serdang, Selangor
Contact number : 03 -89466751
Email : wannor@upm.edu.my
Project Members : Mohd Firuz Ramli
Nor Rohaizah Jamil

1.0 Introduction
Quantification of the effect of land use and land cover change on the runoff dynamics of a river basin has
been an area of interest for hydrologists in recent years. Little is known so far if there is a well-defined
quantitative relationship between the land use properties and the runoff generation mechanism. Different
methodologies have been implemented in attempts to fill in the deficiency of knowledge in the subject, but
no general and credible method has been established yet to predict the effect of land use changes on
hydrology in a watershed (Kokkonen and Jakeman, 2002).
Although the use traditional statistical trend analysis of recorded flood data has been reported to
register a great success in detecting non-stationary hydrologic response of a watershed, it is unable to
calculate the change. Another setback of the trend analyses is their inability to predict the effect of land
use and climate change. Hydrologic models are considered the most appropriate tools in evaluating and
predicting hydrologic response variation at catchment level (Saghafian et al. 2007).
The use of computer models for evaluating the effect of land use change on runoff has been in
use for over four decades. For example; Onstad and Jamieson, 1970; Hookey, 1987; Bultot et al., 1990;
Anderson, 2000; Toriman et al., 2009; Amini et al., 2011; Basaruddin et al., 2014; Khalid et al., 2015. In
recent years there is growing concern for studies regarding climate change at the global scale, where
models are considered the major back bone for predicting hydrological changes based on land use at the
catchment scale. For example; Mah et al., 2011; Alaghmand et al., 2012; Kabiri et al., 2013; Basarudin et
al., 2014.
During early December, 2014, heavy rainfall occurred for many days that resulted in catastrophic
flooding in several part of the east coast state of Peninsula Malaysia. Many claims that illegal logging and
unrestricted land cover conversion without consideration of environmental repercussions, has alters
natural hydrologic systems of the basin. But there is no data to support or refute this argument for the
basin. Following this flood event, a study was initiated to investigate whether past and present land use
changes in the watershed may have increased, or will change, the flood hazard of Kelantan river basin.
Specifically, relative increase/decrease of the flood peaks is of primary interest. The study also
attempts to identify flood source areas with respect to flood occurrence at the downstream reaches for
further flood control planning. Land use maps corresponding to a 30-year period are prepared and HEC-
HMS rainfall-runoff model is calibrated and applied to quantify the impact of past and present land use
change on downstream flood peaks. The model was later used to determine the contribution of various
sub-basins on downstream flood peaks.

2.0 Methodology
This study used rainfall data and streamflow data from Kelantan River basin. Kelantan river basin has an
annual rainfall of about 2383±120 mm, a large amount of which occurs during the North-East Monsoon
3 −1
between mid-October and mid-January.The estimated runoff for this area is 500 m s (DID, 2000).
Hourly river discharge data over 31 years was used to determine flood peaks and its corresponding
volume and duration
Mann–Kendall test was employed to detect annual trends in precipitation and AMF data.
Statistically non-significant increasing and statistically significant increasing trends were obtained for the
annual maximum series of 24-hour precipitation and AMF data respectively. Soil series map and three
land use maps corresponding to 1984, 2002 and 2012 were obtained from the Department of Irrigation

76
and Drainage, Malaysia (DID) for land use analysis. Each of the land use map was intersected with the
soil series map to calculate curve number (CN) using ArcCN script in ArcGIS Zhan and Huang (2004).
In this study, Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER)
Global Digital Elevation Model (GDEM) ASTER GDEM with 30 m resolution was used for hydrologic
model simulation. The HEC-HMS was used to simulate the hydrologic response of the watersheds in this
study. SCS curve number (CN) method was used to determine infiltration. The model was calibrated and
validated using extreme rainfall events selected for the year 2014 to simulate the effect of different land
uses (1984, 2002 and 2013) on the 2014 flood.

3.0 Results and Discussion

Land use changes were studied from 1984 to 2002 and from 2002 to 2013 in each of the four studied
catchments using GIS. The land use classes were categorized into forest, paddy, agriculture, grassland,
urbanization, cleared land, mangrove swamp, secondary forest, rivers, ponds and lakes and mining.
Tables (1-8) shows the result of the land use analysis. The results indicate that conversion of forest to
agriculture is the major land use change observed in all the four catchments. The total area of cultivated
lands has increased non-uniformly across the catchments. In Galas basin, there is 51% reduction in
forest from 1984 to 2002, while in Pergau forest was found to reduce from by 19.32% in the same year
period. While forest was observed to reduce in this year sequence, grassland was observed to increase
by 15.77% in Galas and 7.11 in Pergau. Little reduction in forest land use was observed from 2002 to
2013 for Galas basin (3.4%) and 7.11% for Pergau basin. While the major land use change observed was
decrease in grassland from 15.85% to 0.46% in Galas and from 12.33% to 0.53% in Pergau for the same
year under study. Other land use that that were observed to increase in all the studied years and
locations are urbanization and cleared land.

Table 1: Land use change in Galas basin from 1984 to 2002

1984 Area 2
Land Use type 1984 (%) 2002 (%) 2 2002 Area (km ) % change
(km )
Forest 98.39 47.38 27552.05 740.80 -51.01
Paddy 0.005 0.00 1.32 0.00 -0.005
Agriculture 1.18 35.07 329.96 567.57 33.89
Grassland 0.08 15.85 23.08 257.07 15.77
Urbanization 0.01 0.45 0.85 7.34 0.44
Cleared land 0.33 0.72 93.49 11.62 0.39
Secondary
0.00 0.35 0.00 0.56 0.35
Forest
Rivers, Ponds
0.00 0.02 0.00 0.26 0.02
and Lakes
Minning 0.000 0.16 0.00 2.54 0.16

Table 2: Land use change in Galas basin from 2002 to 2013

2002 Area 2 %
Land Use type 2002 (%) 2013 (%) 2 2013 Area (km )
(km ) change
Forest 47.38 44.04 740.80 714.99 -3.34
Paddy 0.00 0.00 0.00 0.00 0.00
Agriculture 35.07 43.19 567.57 701.14 8.12
Grassland 15.85 0.46 257.07 745.39 -15.39
Urbanization 0.45 1.67 7.34 27.18 1.22
Clearedland 0.72 0.13 11.62 2.11 -0.59
Secondary
0.35 9.03 0.56 146.55 8.680
Forest
Rivers, Ponds
0.02 0.46 0.26 7.52 0.44
and Lakes

77
 Minning 0.16 1.01 2.54 16.43 0.85

Table 3: Land use change in Pergau basin from 1984 to 2002

1984 Area 2
Land Use type 1984 (%) 2002 (%) 2 2002 Area (km ) % change
(km )
Forest 86.3 66.98 19725.24 1530.66 -19.32
Paddy 0.22 0.16 5.02 3.68 -0.06
Agriculture 7.2 25.72 164.49 587.73 18.52
Grassland 4.78 12.33 109.32 281.89 7.55
Urbanization 0.06 0.6 1.42 13.81 0.54
Clearedland 1.44 0.51 32.84 11.65 -0.93
Secondary
0.00 0.03 0.00 0.80 0.03
Forest
Rivers, Ponds
0.00 0.26 0.00 5.98 0.26
and Lakes
Minning 0.00 0.19 0.00 4.45 0.19

Table 4: Land use change in Pergau basin from 2002 to 2013

2002 Area 2 %
Land Use type 2002 (%) 2013 (%) 2 2013 Area (km )
(km ) change
Forest 66.98 59.87 1530.66 1368.13 -7.11
Paddy 0.16 0.03 3.68 0.75 -0.13
Agriculture 25.72 26.64 587.73 608.7 0.92
Grassland 12.33 0.53 281.89 12.18 -11.8
Urbanization 0.6 0.71 13.81 16.34 0.11
Cleared land 0.51 0.48 11.65 11.04 -0.03
Secondary
0.03 0.005 0.8 0.12 -0.025
Forest
Rivers, Ponds
0.00 11.19 0.00 255.78 11.19
and Lakes
Minning 0.26 0.32 5.98 7.39 0.06

Tables 5-8 shows the past and present land use changes in Lebir and Nenggiri catchments.
Reduction in forest is the major land use change observed in both Lebir and Nenggiri catchments from
1984 to 2002. In Lebir forested area reduced from 88.46% to 63.55% signifying a reduction of 24.91%,
while Nenggiri the percentage of forested areas reduce was 17.69% for the same year under study.
Unlike forest that was found to reduce, agriculture recorded an increase of 15.77% in Lebir and 2.49% in
Nenggiri. Grassland was also found to be on the increase in both Lebir and Nenggiri from 1984 to 2002
where an increase of 10.92% and 14.84% occurred respectively. From 2002 to 2013 forested recorded a
small decrease of 2.91% in Lebir and an increase of 6.98 in Nenggiri. The major land use change
observed is decrease in grassland for both Lebir and Nenggiri while urbanization and cleared land all
witnessed a small increase.

Table 5: Land use change in Lebir basin from 1984 to 2002

1984 Area 2
Land Use type 1984 (%) 2002 (%) 2 2002 Area (km ) % change
(km )
Forest 88.46 63.55 2919 2096.09 -24.91
Paddy 0.00 0.004 0.00 0.14 0.004
Agriculture 8.29 24.06 273.54 793.68 15.77
Grassland 1.06 11.98 34.99 395.11 10.92

78
Urbanization 0.02 0.45 0.61 14.73 0.43
Clearedland 2.18 0.04 71.82 1.45 -2.14
Secondary
0.00 0.06 0.00 0.20 0.06
Forest
Rivers, Ponds
0.00 0.0001 0.00 0.003 0.0001
and Lakes
Minning 0.00 0.001 0.00 0.05 0.001

Table 6: Land use change in Lebir basin from 2002 to 2013

2002 Area 2 %
Land Use type 2002 (%) 2013 (%) 2 2013 Area (km )
(km ) change
Forest 63.55 60.64 2096.09 2000.79 -2.91
Paddy 0.004 0.00 0.14 0.00 -0.004
Agriculture 24.06 30.44 793.68 1004.45 6.38
Grassland 11.98 0.32 395.11 10.42 -11.66
Urbanization 0.45 0.45 14.73 14.84 0
Clearedland 0.04 1.29 1.45 42.56 1.25
Mangrove
0.06 0.00 0.2 0.00 -0.06
swamp
Secondary
0.00 6.41 0.00 211.5 6.41
Forest
Rivers, Ponds
0.0001 0.33 0.003 10.89 0.3299
and Lakes
Minning 0.001 0.02 0.05 0.66 0.019

Table 7: Land use change in Nenggiri basin from 1984 to 2002

1984 Area 2 %
Land Use type 1984 (%) 2002 (%) 2 2002 Area (km )
(km ) change
Forest 98.22 80.53 3852.88 80.53 -17.69
Paddy 0.00 0.00 0.00 0.00 0.00
Agriculture 1.41 3.9 55.46 3.9 2.49
Grassland 0.37 15.21 14.51 15.21 14.84
Urbanization 0.00 0.09 0.00 0.09 0.09
Clearedland 0.00 0.15 0.00 0.15 0.15
Mangrove
0.00 0.00 0.00 0.00 0.00
swamp
Rivers, Ponds
0.00 0.001 0.00 0.001 0.001
and Lakes
Minning 0.00 0.004 0.00 0.004 0.004

Table 8: Land use change in Nenggiri basin from 2002 to 2013

2002 Area 2 %
Land Use type 2002 (%) 2013 (%) 2 2013 Area (km )
(km ) change
Forest 80.53 87.51 3148.76 3431.51 6.98
Paddy 0.00 0.00 0.00 0.00 0.00
Agriculture 3.90 5.99 152.4 234.69 2.09
Grassland 15.21 0.86 594.87 33.51 -14.35
Urbanization 0.09 0.22 3.62 8.66 0.13
Clearedland 0.15 0.51 5.99 19.86 0.36

79
 Mangrove
0.00 0.00 0.00 0.00 0.00
swamp
Secondary
0.00 3.96 0.00 155.17 3.96
Forest
Rivers, Ponds
0.001 0.45 5.99 17.74 0.449
and Lakes
Minning 0.004 0.51 0.19 19.99 0.506

3.1 Effect of Land Use on Flood Peak

The HEC-HMS model was used to predict land use effects on floods of the major catchments in Kelantan
river basin using the 2014 flood as the basis for this study. First the, the maximun daily rainfall were used
at the rainfall staations mentioned in section 2.5 above were statistically analyzed and the daily storm
depths corresponding to different return periods were estimated. Inverse distance squared method was
used to determine the spatial distribution of design rainfalls. The model was run for 1984, 2002 and 2013
land use conditions to see the extent of different land use changes on the 2014 flood. The effect of
different land use conditions on the outflow peak discharge is investigated for storms with return periods
from 5 to 100 years in the major catchments studied in this study.

4.0 Conclusion
The effect of different land use conditions on the outflow peak discharge is investigated for storms with
return periods from 5 to 100 years in the major catchments of Kelantan river basin and the results of the
findings are.

4.1 Comparison of the land uses in Kelantan river basin have shown that forest have been
coverted into cultivated areas and grasslands, that leads to increase in flood volume.
4.2 All the catchments experienced significant land use changes especially forest in the past
30 years. Galas recorded the highest decrease in forested areas with 54.35%, followed
by Lebir (27.82%), Pergau (26.47%) and Nenggirri (24.67%)Lebir catchment gives the
highest contribution of flow followed by Nenggiri, Galas and Pergau in that order.
4.3 Galas subbasin is the subbasin with the highest increase in urbanization with 1.66% for
the period under study.
4.4 Flood peak for the 5 year return period in Lebir subbasin corresponding to 1984, 2002
3 3 3
and 2013 is 5439.47 m /s, 7058.16 m /s and 7168.57 m /s respectively.
4.5 Since the effect of location and other factors, particularly the spatial distribution of rainfall,
were incorporated, it is evident that the flood peak generated by this catchments are
more pronounced under 2013 land use in almost all the studied catchments.

References
Alaghmand S., R. A., Abustan I., Said M. and Vosoogh B. (2012). GIS-based river basin flood modelling using HEC-
HMS and MIKE11-Kayu Ara river basin, Malaysia
Amini, A., Ghazali A.T., Aziz A. and Akib S., (2011). Impacts of land-use change on streamflows in the Damansara
Watershed, Malaysia. Arab. Journal of Science and Engineering 36 (5), 713–720.
http://dx.doi.org/10.1007/s13369-011-0075-3.
Anderson D.J. (2000) GIS-based hydrologic and hydraulic modeling for floodplain delineation at highway river
crossing, MSc diss, The University of Texas, Austin.
Basarudin Z., Adnan N. A., Latif A. R. A., Tahir W. and Syafiqah N. (2014). Event-based rainfall-runoff modelling of
the Kelantan River Basin. IOP Conf. Series: Earth and Environmental Science 18 (2014) 012084 IOP
Publishing doi:10.1088/1755-1315/18/1/012084.
Bultot, F., Dupriez, G.L. and Gellens, D., (1990). Simulation of land use changes and impacts on the water balance -
a case study for Belgium. Journal of Hydrology, 114: 327-348.
Hookey, G.R. (1987). Prediction of delays in groundwater response to catchment clearing. Journal of Hydrology,
94:181-198.
Kabiri R., Chan A. and Bai R. (2013). Comparison of SCS and Green-Ampt methods in surface runoff-flooding
simulation for Klang Watershed in Malaysia. Open Journal of Modern Hydrology, 3(2013)102-114
http://dx.doi.org/10.4236/ojmh.2013.33014 Published Online July 2013 (http://www.scirp.org/journal/ojmh)

80
 FLOOD WATER LEVEL PREDICTION MODELING USING NNARX STRUCTURE FOR SG PAHANG
BASIN

Project Information
Project Leader : Dr. Fazlina Ahmat Ruslan
University : Universiti Teknologi MARA
Address : Fakulti Kejuruteraan Elektrik, UiTM Shah Alam, 40450, Shah Alam, Selangor.
Contact number : 012-3086645
Email : fazlina419@salam.uitm.edu.my
Project Members : PM Dr Ramli Adnan, PM Sr Dr Abd Manan Samad
Dr Mazidah Tajjudin

1.0 Introduction
There were a total of 58 events of natural disaster in Malaysia for the period between years 1980 to 2010
that claiming a total of 1,239 lives of the 640,000 people affected. These data were based on statistics
provided by United Nation Officer for Disaster Risk Reduction (UNISDR). From all different categories of
natural disasters considered, flood accounted for over half the registered events. Floods contribute to 8
out of 10 disaster events with the highest human exposure and affect over 85 % of all the disaster-
stricken people. Floods are thus the primary hazard which affecting Malaysia, in particular the west coast
of Peninsular. Therefore, an accurate and reliable flood prediction model is very much needed to provide
early warning for residents nearby flood locations for evacuation purposes. However, current trends in
flood prediction only involve flood modeling because no prediction time was mentioned and discussed.
Furthermore, in Malaysia there is none of flood model or flood prediction model developed yet. An
existing system in the Department of Irrigation and Drainage Malaysia is only the alarming system which
alarms the users only when the water level exceeds the danger limit. Based on these scenarios, the
research objective is to obtain a flood water level prediction model for Pahang flood prone area using
Neural Network Autoregressive Model with Exogenous Input (NNARX) structure. The samples used for
model training, model validation and model testing were carefully selected. In order to obtain good flood
water level prediction model, all samples must be the data when flood events happened. All samples
were real-time data that were obtained from the Department of Irrigation and Drainage Malaysia upon
special request. From carefully selected samples, several optimal flood prediction times were suggested
for flood location in Pahang. Model validation and model testing were conducted to observe the prediction
performances. The optimal prediction time was selected based on the results of prediction performances.
Results show NNARX model successfully predicted flood water level ahead of time.

2.0 Methodology
The samples used in model development were divided into three sets namely: training; validation; and
testing samples. The model was first obtained using training samples and then validate using validation
samples. The new testing samples were then fed to the model to verify the performance of the proposed
model. Figure 1 shows the general block diagram of flood water level prediction model for Pahang flood
prone area using neural network model structure. Four inputs were fed to the Neural Network model to
predict flood water level ahead of time.. The prediction time, Tp can be set at any value that is significant
for evacuation purposes. The input water levels were normalized between +1 and -1 before fed into the
model to keep the water level within the same range. Later, the water levels were denormalized back to
obtain the actual predicted flood water level value at the output. The value of prediction time, T p can be
set at any normal water level condition at flood location before the water level started increasing. ST1,
ST2, ST3 and ST4 represent four upstream rivers and x represents current water level at flood location.
ŷ represents predicted water level at flood location.

81
 ST1
ST2
ST3
NNARX Structure yˆ Tp  k 
ST4
y
k
Figure 1 NNARX Structure for Flood Water Level Prediction

3.0 Results and Discussion

3.1 Modeling using 10 hours Prediction Time
Figure 2 shows 10 hours prediction result of NNARX model using Gradient Descent (traingd) as training
algorithm. The actual water level is under predicted at early time steps, k. However, the models still able
to predict the actual water level quite good at the rising stage till it reaches peak water level stage from
time steps 1000 till 1200. Nevertheless, NNARX model under predicted actual water level at low
frequency segment when the flood water level drops from peak to normal water level. At this segment, it
can be clearly seen that NNARX model only able to predict at linear section thus producing Best Fit of
80.1068% as given in Table 1.

25.8

Actual
25.6

25.4
water level(m)

25.2

Prediction
25

24.8

24.6

24.4
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Time steps

Figure 2. 10 hr Flood Water Level Prediction of NNARX Modeling Result

Table 1 : Performance Indices Result for Tp = 10 hr

Training Best Fit RMSE

V (m) FPE (m)
Algorithm (%) (m)

traingd 80.1068 0.0901 0.0081 0.0081

Another set of samples have been applied to test the developed NNARX model. This samples fulfilled the
same criterion as training and validation samples and the time range is between 14/11/2014 21:00:00 till
12/12/2014 15:30:00 with the total number of 4000 samples. It is important to select the testing samples
in the same year as training and validation samples. This is due to the fact that the longer the time
difference between those samples then more physical changes on river basin and thus will effect the
performance result. Figure 3 shows the flood water level prediction after testing samples were applied to
the model obtained from Figure 2. It can be observed that the NNARX model leading the actual flood
water level by 10 hours. However, the predicted flood water level is not in good agreement with the actual
water level, thus mapping for both graphs need to be done for RMSE analysis.

82
 25.2
Prediction Actual
25.1

25

24.9

water level(m)
24.8

24.7

24.6

24.5

24.4

24.3 Tp

0 500 1000 1500 2000 2500 3000 3500 4000

Time steps

Figure 3. 10 hr Flood Water Level Prediction of NNARX Model Test Result

Figure 4 shows the mapping for both the predicted and actual flood water level graphs for RMSE
analysis. As expected from the results from Figure 6, the comparison between the predicted and actual
water level did not show any significant results because NNARX model cannot predict the actual water
level at any segments thus producing RMSE value of 0.0695 meter.

25.2

Actual
25.1

25

24.9
water level(m)

24.8 Prediction

24.7

24.6

24.5

24.4

24.3

0 500 1000 1500 2000 2500 3000 3500 4000

Time steps

Figure 4. Mapping of the Predicted and Actual Water Level for RMSE Analysis

3.2 Modeling using 15 hours Prediction Time

Next, the prediction time, Tp was increased furthermore to 15 hours for performance analysis. Figure 5
shows flood water level prediction result using traingd as training algorithm. The results have shown good
agreement between predicted and actual water level at high frequency section however the NNARX
model under predicted the actual water level at low frequency section. The model is trained using Tan-
Sigmoid (tansig) as transfer function for input, hidden and output layers whereas the number of nodes for
hidden layer is set to 25 to meet the error goal of 0.00001m. Despite the error goal was not converged,
the NNARX model still manage to provide good Best Fit value of 84.82% as shown in Table 2.

83
 25.8

Actual
25.6

25.4

water level(m)
25.2
Prediction

25

24.8

24.6

24.4
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Time steps

Figure 5. 15 hr Flood Water Level Prediction of NNARX Modeling Result

Table 2 : Performance Indices Result for Tp = 15 hr

Training Best Fit RMSE
V (m) FPE (m)
Algorithm (%) (m)
traingd 84.82 0.0687 0.0047 0.0047

Using the same model structure as in Figure 5, the testing samples were applied to the NNARX model to
evaluate the performance result. From Figure 6, it can be seen that the performance of NNARX model is
degrading compared to Tp = 12 hours. The model shows slightly different pattern trend from the actual
water level. For detailed analysis on NNARX model performance, mapping for both predicted and actual
water level need to be done. Figure 7 shows mapping of both graphs for RMSE analysis. Generally,
NNARX model only manage to track actual water level at time steps 400 till 1490.

25.2

25.1
Actual
25
Prediction
24.9
water level(m)

24.8

24.7

24.6

24.5

24.4

24.3
Tp

0 500 1000 1500 2000 2500 3000 3500 4000

Time steps

Figure 6. 15 hr Flood Water Level Prediction of NNARX Model Test Result

25.2

25.1 Actual
25

24.9
water level(m)

24.8

24.7

24.6

24.5

24.4

24.3 Prediction

0 500 1000 1500 2000 2500 3000 3500 4000

Time steps

Figure 7. Mapping of the Predicted and Actual Water Level for RMSE Analysis

84
4.0 Conclusion

4.1 The flood water level prediction model using existing Neural Network Autoregressive with
Exogenous Input (NNARX) model has been successfully developed and tested for flood
prone area in Pahang.
4.2 The NNARX model had shown satisfactory performance in all assessments including
model validation and model testing. The performance result shows that 15 hours is the
best prediction time for flood water level prediction in Pahang.
4.3 However, 10 hours prediction time were still reliable for flood water level prediction
model.
4.4 Variations value of prediction time were tested starting from 3 hours till 24 hours, but the
optimal prediction time for Sungai Pahang basin is between 10 to 15 hours.

References
L.-H. Feng and J. Lu, "The practical research on flood forecasting based on artificial neural networks," Expert
Systems with Applications, vol. 37, pp. 2974-2977, 2010.
Center for Hazards and Risk Research at Columbia (CHRR). (2015, 30 Mac). Malaysia natural disaster profile.
Available: http://www.Ideo.columbia.edu/chrr/research/profiles/malaysia.html
W. Al-Sabhan, M. Mulligan, and G. A. Blackburn, "A real-time hydrological model for flood prediction using GIS and
the WWW," Computers, Environment and Urban Systems, vol. 27, pp. 9-32, 2003.
K. Jasper, J. Gurtz, and H. Lang, "Advanced flood forecasting in Alpine watersheds by coupling meteorological
observations and forecasts with a distributed hydrological model," Journal of Hydrology, vol. 267, pp. 40-52,
2002.
S. Jialan, L. Xiaohui, L. Weihong, J. Yunzhong, and W. Hao, "Development of a flood forecasting system and its
application to upper reaches of Zhangweihe River Basin," in Geomatics for Integrated Water Resources
Management (GIWRM), 2012 International Symposium on, 2012, pp. 1-4.
H. Moradkhani, S. Sorooshian, H. V. Gupta, and P. R. Houser, "Dual state–parameter estimation of hydrological
models using ensemble Kalman filter," Advances in Water Resources, vol. 28, pp. 135-147, 2005.
J. C. Neal, P. M. Atkinson, and C. W. Hutton, "Flood inundation model updating using an ensemble Kalman filter and
spatially distributed measurements," Journal of Hydrology, vol. 336, pp. 401-415, 2007.
C. W. Dawson, R. J. Abrahart, A. Y. Shamseldin, and R. L. Wilby, "Flood estimation at ungauged sites using artificial
neural networks," Journal of Hydrology, vol. 319, pp. 391-409, 2006.
T.-Y. Pan, Y.-T. Yang, H.-C. Kuo, Y.-C. Tan, J.-S. Lai, T.-J. Chang, C.-S. Lee, and K. H. Hsu, "Improvement of
watershed flood forecasting by typhoon rainfall climate model with an ANN-based southwest monsoon
rainfall enhancement," Journal of Hydrology,vol. 325, pp. 261-274, 2009.
P.-A. Chen, L.-C. Chang, and F.-J. Chang, "Reinforced recurrent neural networks for multi-step-ahead flood
forecasts," Journal of Hydrology, vol. 497, pp. 71-79, 2013.
G. Kim and A. P. Barros, "Quantitative flood forecasting using multisensor data and neural networks," Journal of
Hydrology, vol. 246, pp. 45-62, 2001.
B. Bazartseren, G. Hildebrandt, and K.-P. Holz, "Short-term water level prediction using neural networks and neuro-
fuzzy approach," Neurocomputing, vol. 55, pp. 439-450, 2003.
Y. Wei, W. Xu, Y. Fan, and H.-T. Tasi, "Artificial neural network based predictive method for flood disaster,"
Computers & Industrial Engineering, vol. 42, pp. 383-390, 2002.
K. Chaowanawatee and A. Heednacram, "Implementation of Cuckoo Search in RBF Neural Network for Flood
Forecasting," in Computational Intelligence, Communication Systems and Networks (CICSyN), 2012 Fourth
International Conference on, 2012, pp. 22-26.
P. Leahy, G. Kiely, and G. Corcoran, "Structural optimisation and input selection of an artificial neural network for
river level prediction," Journal of Hydrology, vol. 355, pp. 192-201, 2008.
T. Kerh and C. Lee, "Neural networks forecasting of flood discharge at an unmeasured station using river upstream
information," Advances in Engineering Software, vol. 37, pp. 533-543, 2006.
G. Corani and G. Guariso, "Coupling fuzzy modeling and neural networks for river flood prediction," Systems, Man,
and Cybernetics, Part C: Applications and Reviews, IEEE Transactions on, vol. 35, pp. 382-390, 2005.

85
 A COUPLED HYDROLOGIC-HYDRAULIC MODEL FOR SIMULATION OF FLOOD INUNDATION
EXTENT AND DEPTH IN KELANTAN RIVER BASIN
Project Information
Project Leader : Tarmizi Ismail
University : Universiti Teknologi Malaysia
Address : Faculty of Civil Enngineering, Univ Teknologi Malaysia, 81310 UTM Skudai,
Johor
Contact number : 07-5538709 / 019-7547747
Email : tarmiziismail@utm.my
Project Members : Shamsuddin Shahid, Sobri Harun, Zaitul Marlizawati Zainuddin,
Arien Heryansyah, Usman Ullah Sheikh

1.0 Introduction
The first half of northeast (NE) monsoon is one of the most spectacular climatic period of Malaysia when
east coastal region of Peninsular Malaysia receives 50%, often as much as 70% of its total annual rainfall
(Yik et al., 2014). The heavy rainfall events occurred 3 to 4 times in the seasons with each episode
usually lasts for 3 to 5 days (Moten et al., 2014). The large rainfall episodes often amount to several
hundreds to over a 1000 mm rainfall and cause devastating flood in the east coastal region (Yik et al.,
2014). The state of Kelantan, as one of the eastern coastal states is the most severely affected region
due to NE monsoonal flood. The Kelantan River regularly overspills its banks during the NE monsoon and
cause floods in the most parts of the state. The historical records show that the river basin which covers
85% of the state suffered from floods in almost every year, among those, the floods in 1926, 1967 and
2014 are the most severe. Some other records of flood experiences in the country as a whole available in
literature with the State of Kelantan inclusive are; 1931, 1947, 1954, 1957, 1963, 1965, 1969, 1971, 1973,
1983, 1988, 1993, 1998, 2001, 2006, 2007, 2008, 2009 and 2010 (Khan et al., 2014).
The recent monsoonal rain driven flood (December 2014) in Kelantan was the worst in the
county’s recent history. Two consecutive heavy rainfall episodes in the last half of December 2014
caused severe floods with Kelantan as the worst affected area. Most parts of the state were inundated
and hundred thousand people were forced to evacuate. The flood severely affected the food supplies,
electricity, clean water, sewerage, health care and other emergency services and caused an
unprecedented public outcry. The severe devastation appealed for urgent scientific solution to this
recurrent natural hazard.
Though heavy rainfall episodes are considered as the major cause of flood in the region,
unprecedented severity of recent flood drawn attention to some other likely factors such as changes in
land use, changes in climate, high tides, insufficient drainage and lack of flood protection system as the
possible causes of amplification of flood induced by monsoonal rainfall. It has been reported that
precipitation in the Kelantan River basin has increased in the wet season and decreased in the dry
season significantly in recent years (Adnan and Atkinson, 2011). This has caused increase in streamflow
in both the upstream and downstream sub-catchments of the basin (Adnan et al., 2014). Basarudin et al.
(2014) reported that recent activities involving land use changes from lowland forest to vegetation and
urban area have contributed to increase floods in recent decade in the river basin. Adnan et al. (2014)
reported that land use change, predominantly deforestation for agricultural purposes, has potentially
caused some increases in hydrological response over time in the upstream area of Kelantan River Basin.
The objectives of the present study are (i) to simulate the flood inundation extent and depth at the
upper sub-catchment of Kelantan river basin using a coupled hydrologic – hydraulic model; and (ii) to
assess the impacts of landuse change on flood peak and volume in the basin. Of the six catchments in
the Kelantan river basin, Sg. Galas was chosen for the research. The selection is guided by the basic
philosophy that peak discharge being generated at any point of the upstream sub-catchment propagates
to the lower reaches with consequence of occurrence of flooding at this region. In addition, records had
shown higher rainfall amounts in the stations within and around the catchments than the rainfall stations
in the neighbouring catchments (Nenggiri and Pergau), and more so, deforestation in the region is
generally believed to be more than the two other catchments. It is therefore believed that simulation of
runoff at this upstream sub-catchment will help to develop early warning system for emergency
preparedness.

86
2.0 Methodology
The study area Sg Galas is one of the six sub-catchments in the Kelantan river basin. The area lies
0 0 0 0
between Longitudes 101.923 E, 102.188 E and Latitudes 4.622 N, 5.534 N with elevations vary between
25 m and 1351 m above the mean sea level. The upper reaches of the basin consists of forested
mountains and the lower reaches with lowland forest as well as limestone hills (Basarudin et al., 2014).
Sg. Galas consists of nine sub-catchments namely: Sg. Asap, Cheweh, Chiku, Galas, Kelasa, Kerak,
Ketil, Kundor and Nireh as shown in Figure 1. There is no streamflow station in Sg. Galas catchment, and
therefore, two other catchments namely, Sg. Nenggiri and Pergau were included for calibration and
validation, since these two gauged catchments have common outlet with the studied catchments.
A field reconnaissance of the drainage basin was carried out to identify the physical features of
the basin, such as the drainage networks, soils and geologic conditions, landuse and vegetative cover.
The hydraulic roughness characteristics of the drainage basins in extreme flow conditions were also
examined, considering overbank flows. The channel can be described as incised in most cases, while the
banks are inhabited, cultivated or remained forested. The river bed is more than 60 m wide at the
upstream and about 120 m at the lower reaches, covered with coarse, medium to fine sand and silt. In
situ permeability test using permeameter to establish the hydraulic conductivity of the soils under different
land use conditions was also conducted.
For effective simulation and in-depth understanding of the scenarios, coupled hydrologic and
hydraulic model was developed using XPSWMM software. Preparation of inputs for the model is
described below.

2.1 Hydrologic data

The rainfall data for the study was acquired from the Department of Irrigation and Drainage (DID),
Malaysia. Continuous records of 5mm rainfall from automatic stations are available for the study area.
The data was converted to hourly data as the appropriate interval for the model. Hourly rainfall data for
selected events such as; 2008 flooding in at Sg. Galas di Dabong was used for the calibration while 2009
January episode for validation. Eighteen selected rainfall stations spread across the catchments with
consistent data were used (Figure 2). It is to be noted that the distributions of rainfall stations at the study
catchments are sparsely distributed and several stations contain missing data. These do not conform to
standard norms required for accurate simulation and thus remained a limitation.
In order to effectively simulate flow in Sg. Galas, streamflow data available at three locations in
the catchments were used for calibration and validation of the model. These are Sg. Pergau in Batu
Lembu, Sg. Nenggiri in Jam. Bertam and Sg. Galas in Dabong. It is of note that Sg. Galas in Dabong is
the common outlet of the catchments and thus records combined flow from the catchments (Figure 2).

87
 Figure 1: The study area map

Figure 2: Rainfall and streamflow gauging stations used for study

88
2.2 Geomorphologic characteristics
High resolution DEM data is required for better accuracy in hydrological modelling. However, as the high
resolution DEM data was not available, the best recourse available resources were used. The LiDAR data
of 3m resolution for the major rivers and 30m resolution USGS DEM were used to generate drainage
network and topographic features of the study area. The geomorphologic characteristics formed part of
the pertinent data required for hydrologic studies, which include catchment area, slope and width. These
are properties were derived from ASTER DEM data obtainable from US Geographical Survey (USGS)
website using geographical information system (ArcGIS 10.3).

2.3 Land use and land cover

The relationship between land use and land cover is known to be dynamic and rarely in a stable
equilibrium (Niehoff et al., 2002). The land use and land cover play major role in controlling the
catchment’s hydrologic response (Veldkamp and Fresco, 1996). Quantifying various flood parameters in
the catchment, requires in-depth study of land use in order to understand the influence of natural and
man-induced factors in recent flooding. Therefore, the land use maps for the study area were obtained
from the Department of Agriculture, Malaysia (DOA). The land use maps for the years from 1980’s to till
date were collected. From the available maps, growth in land uses in the catchment was detailed. Figure
3(a - e) shows changes in land uses in Sungai Galas catchment.

2.4 Soil data

The soil property of a catchment constitutes one of the major factors to its hydrologic response to runoff
volume during and after rainstorm events. Soil texture determines the water-holding characteristics of a
soil as well as the infiltration capacity of a soil layer. Therefore, studies to ascertain the characteristics of
the soils are important. There are basically three major soil series in the studied catchments. These are:
Charang Hungus/Lubok Kiat, Durian/Munchong/Bungor and Rengam/Jerangau. However, Charang
Hungus is the dominant soil type in the catchments. Description of soil series are given below:

Charang Hungus/Lubok Kiat: The soils are characterized in the field by light grey clay with clay content
ranging from 35 – 60%. The soils have moderate structure, plastic consistency having poor drainage
characteristics. The hydraulic conductivity of the soils under different land use conditions as determined
-5 -4
from the field test are between 3.8 x 10 to 4.0 x 10 cm/s.

Durian Munchong Bungor: The soils are in groups 2, 1 and 6 respectively according to 26 Malaysia soil
series. The textural classifications of the soils are silty clay loam and silt loam. These are typically B and
C hydrologic soil group (HSG).

Rengam Jerangau: These soils are in group 6 according to 26 Malaysia soil series. They are clay and
sandy clayey soils belonging to HSG C. Figure 5 shows the soil map of the study area.

89
Figure 2: Changes in land uses in Sungai Galas catchment over time

90
 Figure 3: Soil map for the study area

3.0 Results and Discussion

The summary of the pertinent data for the model calibration and validation as well as the results from the
simulation are hereby presented and discussed below.

Catchment characteristics
The characteristics of the study catchment are given in Table 1. The catchment consists of nine sub-
catchments with total area of 160,276 ha. The characteristics of each sub-catchment were extracted as
input parameter for the model. Width of the catchment is an important characteristic and should never
exceed 10000m. Based on this, the catchments were further sub-divided into smaller unit areas such that
the same catchment can have different rainfall data, and other catchment properties depending on the
size and requirements.
The land use maps were used to derive changes in forest areas in the study catchment. Figure 6
shows the variations in forest area in the region over time.

Table 0-1: Catchment characteristics

Sub-Catchment Area (ha) Slope (%) HSG
Sg. Asap 13,667 0.975 C
Sg. Cheweh 13,067 0.478 C
Sg. Chiku 46,767 0.262 B
Sg. Galas 29,419 1.357 C
Sg. Kelasa 10,604 0.248 B
Sg. Kerak 4,975 0.347 C
Sg. Ketil 18,449 2.260 C
Sg. Kundor 11,111 0.442 B
Sg. Nireh 12,217 0.143 B

91
 Figure 6: Variations in forest in the region over time

3.1 Model calibration

To calibrate the model, the flood event of 25-30 Nov., 2008 was used. For 1D coupled hydrologic – hydraulic model,
rainfall, catchment area, slope, catchment width, percentage of impervious area, CN, depression storage, Manning’s
roughness coefficient and evaporation data were used. The initial abrasion of 0.2 along with the CN and others
aforementioned parameters were used to represent the infiltration data for the model using SCS method. In
simulating the hydraulic model, channel cross sectional data, longitudinal length, slope as well as manning’s
coefficient was used. Figure 7 shows the calibrated hydrograph for Sg. Galas at Dabong Station.

2500
Sg. Galas at Dabong
2250 Calibrated
Observed
2000

1750
Flow (m3s-1)

1500

1250

1000

750

500

250

0
26 Wed

29 Sat
25 Tue

27 Thu

30 Sun
28 Fri

1 Mon

Nov. , 2008 Time

Dec.

Figure 7: Observed and calibrated hydrograph at Sg. Galas in Dabong in November, 2008

92
The parameters of the model were adjusted until calibrated flow matched excellently with the observed
th
flow. The hydrograph has two peaks as seen in the figure. The first peak occurred on Thursday, 27
3 -1
November 2008 with maximum of 1258.94 m s , which did not however resulted to any flood, while the
th
second one was on the Sunday, 30 November 2008, which caused flood. Peak flow simulated by
3 -1 3 -1
calibrated model was 2280.34m s as compared to the observed value of 1982.75m s . It was noted that
both the rising and recess limbs of the calibrated hydrograph was steeper than the observed and hence a
little higher value of peak discharge was simulated. The times to peak for observed and calibrated flows
th
were 124 hrs and 122 hrs, respectively starting from 25 November 2008.

3.2 Model Validation

The validation of the model was carried out using the January 2009 flood event (01-06 January, 2009)
having relatively similar characteristics in terms of the peak discharge as well as the span of the event.
Because of the closeness of the events, it was assumed that the land use characteristics remained
unchanged. Using the same rainfall stations as those for calibration, the hydrographs produced at
Dabong and Nenggiri stations are shown in Figures 8 and 9, respectively.

Validated
Sg. Galas at Dabong
Observed
3000

2500

2000
Flow (m3s-1)

1500

1000

500

0
4 Sun
1 Thu

5 Mon

6 Tue
2 Fri

3 Sat

Jan., 2009 Time

Figure 8: Observed and validated hydrograph at Sg. Galas in Dabong, January 2009

Figure 9: Observed and validated hydrograph at Sg. Galas outlet, January 2009

93
3.3 December, 2014 flood scenario
The essence of the current study is to quantify various flood parameters in Sg. Galas catchment during
th
the 2014 flood event. The torrential rainfall which caused the devastating flood began on the 16
December 2014. The extreme rainfall continued for the next ten days. This event gave persistent raise in
water levels in the upstream rivers among which is Sg. Galas at Dabong. The highest recorded level at
this station was 46.47 m which was far above the danger level of 38m. During the episode, the peak
3 -1 3 -1
discharge was as high as 11052 m s , where the flood magnitude of approximately 2000 m s always
resulted flooding. Figure 10 show the flows from Dabong station during flood period.

Figure 4: Simulated hydrograph for Sg. Galas in Dabong, December 2014

4.0 Conclusion
The elements of a flood cause-consequence chain are interrelated in the downstream and upstream parts
of a river network. There is no doubt that changes in an upstream part of a river have a strong impact on
downstream river sections. Therefore, to understand the cause and mechanism of flood generation, it is
required to pay more attention on hydrological changes in upstream sub-catchments of a river basin.
Kelantan River basin has six major sub-catchments. Considering the availability of long-term data, a sub-
catchment located in the upstream of the river basin known as Sg. Galas catchment was selected for the
present research. To effectively simulate the flow from Sg. Galas catchment; two more catchments,
namely, Sg. Nenggiri and Sg. Pergau were included as those two catchments have common outlet with
Sg. Galas at Dabong station and flow records are available at gauging stations in these two catchments.
Therefore, discharge data at these two catchments were also used for the calibration and validation of
model.
A coupled hydrologic-hydraulic model was used in the present study to fulfill the objectives. The
calibration and validation of the model suggested its suitability for use. The model input parameters were
rainfall, catchment area, slope, catchment width, percentage of impervious area, CN, depression storage,
Manning’s roughness coefficient and evaporation data. Furthermore, channel cross sectional data,
longitudinal length, slope as well as manning’s coefficient were used in simulating the hydraulic model.
The model was run with different land use data available for series of years in order to quantify the
impacts of land use change on flow peak and volume. The results show that flow peak and volume
reduced 0.74%, 3.12%, 5.71% and 6.71% for the years 1984, 1990, 2000 and 2008, respectively
compared to that in year 2014 when same rainfall episodes were considered. These are small reductions
3 -1
considering a peak discharge of over 3000 m s from a catchment having an area of approximately
3 -1
160,276 ha, where contributions of peak flow of 600 m s in the catchment results flood. This means that
impact of land use change in flood peak and volume in the study area is negligible. The present study
also analyzed the effects of floods from the catchment. The study identified the far reaching effect of
flooding from the catchments due to backwater effect occasioned by repulsive force from the fast moving
main channel flow to its tributaries which have less volumes as well as slower velocities. Overall, the
model is found to simulate the inundation extent and depth appropriately, thereby providing a measure
which can be used to suggest mitigation measures.

94
References
Adnan, N.A. and Atkinson, P.M. (2011). Exploring the impact of climate and land use changes on streamflow trends
in a monsoon catchment. International Journal of Climatology. 31, 815–831.
Adnan, N.A., Atkinson, P.M., Yusoff, Z.M. and Rasam, A.R.A. (2014). Climate Variability and Anthropogenic Impacts
on a Semi-Distributed Monsoon Catchment Runoff Simulations. 1-6.
Basarudin, Z., Adnan, N.A., Latif, A.R.A., Tahir, W. and Syafiqah, N. (2014). Event-based rainfall-runoff modelling of
the Kelantan River Basin. IOP Conference Series: Earth and Environmental Science. 18, 012084.
Khan, M.M.A., Shaari, N.a.B., Bahar, A.M.A., Baten, M.A. and Nazaruddin, D.a.B. (2014). Flood Impact Assessment
in Kota Bharu, Malaysia: A Statistical Analysis. World Applied Sciences Journal. 32 (4), 626-634.
Moten, S., Yunus, F., Ariffin, M., Burham, N., Yik, D.J., Adam, M.K.M. and Sang, Y.W. (2014). Statistics of Northeast
Monsoon Onset, Withdrawal and Cold Surges in Malaysia. In: Department, M. M. (ed.). Kuala Lumpur,
Malaysia.
Niehoff, D., Fritsch, U. and Bronstert, A. (2002). Land-use impacts on storm-runoff generation: scenarios of land-use
change and simulation of hydrological response in a meso-scale catchment in SW-Germany. Journal of
Hydrology. 267, 80–93.
Veldkamp, A. and Fresco, L.O. (1996). CLUE: a conceptual model to study the Conversion of Land Use and its
Effects.pdf>. Ecological Modelling. 85, 253-270.
Yik, D.J., Adam, M.K.M., Sang, Y.W. and Moten, S. (2014). Anomalous Winter Monsoon Season of 2012/2013 Over
the Malaysian Region. In: Department, M. M. (ed.). Kuala Lumpur, Malaysia.

95
 FORMULATION OF A TRUST EVALUATION ALGORITHM USING TRUST MODEL FOR PEER
COMMUNICATIONS TO IMPROVE INFORMATION RELIABILITY DURING THE DISASTER
ENVIRONMENT
Project Information
Project Leader : Dr. Yunus Yusoff
University : Universiti Tenaga Nasional
Address : Jalan IKRAM-UNITEN, 43000 Kajang, Selangor.
Contact number : 03-8928 7502
Email : Yunusy@uniten.edu.my
Project Members : Assoc. Prof. Dr. Roslan Ismail, Dr. Asmidar Abu Bakar
Ramona Ramli, Dr. Norshakirah Ab Aziz (Univ. Teknologi Petronas)
Muhammad Harith Rosli

1.0 Introduction
During the event of a disaster, dissemination and sharing of information such as location and situation of
the disaster are very much sought after. With the presence of ICT applications and smartphone devices,
dissemination and sharing of information can be carried out with ease. Everybody can post and share
information about the disaster they are currently experiencing. Although this freely sharing of information
is good, it does give rise to one critical issue. Can we trust and verify the information obtained from these
technologies? How do we ensure that the information being transmitted is genuine? It was discovered
that in order for help to be sent immediately, people may resort to fabricated information so as to make
the authority believe that they are in great danger.
In gathering data from users, crowdsourcing is considered as the most suitable approach. This
term was coined by Jeff Howe [1] and further refined by other researchers [2]. This approach allows users
to give their feedback easily as in data collected in Haiti earthquake in 2010 [3, 4]. This approach was
also adopted in other well-known disasters such as the Queensland & Australian Flood, the Christchurch
Earthquake and the Japan Earthquake [5].
Based on the abovementioned scenarios and our interviews with the victims of Kelantan’s flood,
the rescue centres were lacking the necessary methods to check for the trustworthiness of the data
captured from the victims. As such, all incoming data (raw data) are always considered as trustable. It is
always possible that there exists unreliable, untrusted or fake data, among the huge amount of data being
captured form the public. Should the authority acted on this untrusted data, their effort to save the real
victims may be hindered.

2.0 Methodology

a) Reviewing of the trust evaluation requirements and constraints for information reliability
In this phase, information reliability characteristics and disaster information systems will be
identified and analyzed. In addition, special requirements and constraints related to flood scenario
will also be studied.
b) Reviewing of the current trust models.
In this phase, various trust models be studied and analyzed. The focus is to find the models that
would be suitable in the disaster environment.
c) Designing and developing the lightweight trust evaluation algorithm
Inputs from the earlier phases will be used to design and develop the algorithm Some
modifications to the selected trust model will be carried out to suit our aim to propose lightweight
algorithm.
d) Evaluation of Algorithm
The developed algorithm is analyzed and evaluated to ensure of its usability.

3.0 Results and Discussion

Taking into consideration of the Kelantan big flood, it was discovered that the GPS location and user
information are needed in order to ascertain the trust of the transmitted information.

96
 Fig. 2: Available users

In Figure 2, the possible users available during the flooding scenario are displayed. These users are
categorized based on their physical location and whether they have registered with the rescue center.
Local users are those located within the disaster area and Outside users are those outside the disaster
area. Description of each category of users is provided in Table 2. Each type of user is assigned with a
trust value based on their level of reliability. The more reliable the user, the higher the trust value will be.
In Figure 2 & Table 2, the most reliable user is Ar (Registered Authority) and highest trust value of 1.0 is
assigned to them.
The registered authority (Ar) has a full trust value since they are the authorized personnel to
handle the rescue operation. As for the local user, a high trust value (0.7 & above) [6] is assigned to them
since they are currently experiencing the disaster.

Table 2: Types of users.

User Type Explanation Trust
Value
Lr Local Registered User 0.8
Lx Local not Registered User 0.7
Or Outside(non-local) Registered User 0.6
Ox Outside(non-local) not Registered 0.5
User
Ar Authority Registered User 1.0

We propose the following formula for calculating the trust of information received from various individuals
affected by flood situation. An initial value of 0.5 is initially assigned to all areas. Whenever a user reports
a flooding in a respective area, the trust value (meaning the possibility of flood in that area) will be
increased accordingly. Likewise, if a user report that an area is not flooded, the trust value of that area will
be reduced accordingly.

NATV = CATV + (UTV * AHTV) or NATV = CATV - (UTV * AHTV)

Where,
NATV = New Area Trust Value; CATV= Current Area Trust Value
UTV = User Trust Value; AHTV = Area History Trust Value
+, if reported flooding, -, if reported not flooding

NATV refer to the new area trust value. CATV refers to the current trust value of the respective area. UTV
refers to the trust value of the user. Different user is assigned with different trust value. Local registered
users are given a high trust value of 0.8, since they are registered and currently in the disaster area.
Local users, but not registered is given a slightly lower trust value of 0.7, since the authority would not be
able to immediately ascertain their identities. The same concept is used for outside users whereby
Outside but registered user is given 0.6 and Outside but did not register is given 0.5 trust value. AHTV
refers to the historical flood situation for the disaster area. There are three historical flood situations
namely, frequently flooded, seldom flooded and rarely flooded of which is assigned 0.03, 0.02 and 0.01
trust value respectively. The value for each situation is assigned to ensure that the overall trust value
would increase linearly instead of exponentially.

97
 Fig. 3: Threshold values

Based on literature reviews [7,8,9], we have adopted a value of 0.7 as a threshold for high trust value.
The distance between 0.7 to 1.0 is 0.3. Using the similar distance from 0, we have opted to define 0.3 as
the threshold for not flooded. Trust values between 0.3 and 0.7 indicate an Alert level for the particular
area. Areas with 0.7 & above trust values, indicates that they are experiencing serious flooding and need
immediate attention. The rescue centers can focus their rescue efforts on those critical areas (trust values
exceeding the Upper Treshold).

4.0 Conclusion
Important findings of the research works are summarized as follows:

4.1 A huge amount of information is communicated to the rescue centers during a disaster
and information via mobile devices is significantly large. The rescue centers do not have
sufficient mechanism to ascertain the truthfulness of the information received.
4.2 The method to calculate the trust value for each information received was developed
using the GPS coordinate, user data and historical data. Information received can be
ranked based on calculated trust values. Rescue centers can focus their rescue efforts
on the critical areas.
4.3 A prototype was developed to demonstrate the workability of the developed trust
algorithm.

References
J. Howe, Crowdsourcing: Why the power of the crowd is driving the future of business. New York: Crown Publishing
Group, 2008.
E. Estelles-Arolas and F. Gonzalez-Ladron-de-Guevara, Towards an integrated crowdsourcing definition, Journal of
Information Science, 32(2):189-200, 2012.
Zook, M., Graham, M., Shelton, T., Gorman, S (2010). Volunteered Geographic Information and Crowdsourcing
Disaster Relief: A Case Study of the Haitian Earthquake. World Medical & Health Policy, Vol. 2 [2010], Iss.
2, Art. 2
J. Heinzelman and C. Waters, Crowdsourcing Crisis Information in DisasterAffected Haiti, United States Institute of
Peace (USIP), Washington DC, Special Report 252, Oct. 2010. [Online]. Available: www.usip.org. Accessed:
Feb. 03, 2016.
A. Rajabifard and D. Coleman, Spatially enabling government, industry and citizens: Research and development
perspectives. Needham, Ma.: GSDI Association Press, 2012, pp 201-214
A. Bruns, J. Burgess, K. Crawford, and F. Shaw, Crisis Communication on Twitter in the 2011 South East
Queensland Floods, ARC Centre of Excellence for Creative Industries and Innovation, Brisbane, 2012.
M.Haque, M., & I.Ahamad, S. An Omnipresent Formal Trust Model (FTM) for Pervasive Computing Environment.
Proceeding of the 31st Annual International Computer Software and Applications Conference (COMPSAC
2007), IEEE Press, Beijing, China, July 2007, pp 49-56.

98
 A COMMUNITY-BASED STUDY ON THE EFFECTIVENESS OF FLOOD EMERGENCY WARNING
SYSTEM IN MALAYSIA

Project Information
Project Leader : Nor Eliza Binti Alias
University : Universiti Teknologi Malaysia
Address : Fakulti Kejuruteraan Awam, Universiti Teknologi Malaysia, 81310
Skudai, Johor
Contact number : 018-9742075
Email : noreliza@utm.my
Project Members : Mohd Badruddin Bin Mohd Yusof, Radzuan Bin Sa'ari
Kamarul Azlan Bin Mohd Nasir, Tarmizi Ismail
Shazwin Binti Mat Taib, Ilya Khairanis Binti Othman

1.0 Introduction
Floods contribute to the highest percentage of natural disaster occurrence in Malaysia. It also produces
the highest impact to the people and economy. The total number of people affected by major floods in
Malaysia reached millions and the damage costs estimated to billions of Ringgit Malaysia. One of the
ways to reduce the risk from flood disaster is to implement an effective flood early warning system. An
effective warning system will help significantly in reducing numbers of loss lives and assets. In order to do
this, assessing current effectiveness of the flood early warning system is crucial. Questions arise whether
the disseminations used are effective in reaching the community. By conducting surveys and
questionnaires, the effectiveness of current early warning system and the preparedness of the Malaysian
people towards flood can be assessed.

2.0 Methodology
Subsequent to the big Yellow flood of the December 2014 Kelantan, a set of critical questions were
developed and put into survey to obtain preliminary assessment right after the flood. Through the
preliminary assessment, a more detailed and revised survey questionnaire were developed particularly
for a larger-scale public survey. Tests were conducted to screen out difficult and unsuitable technical
terms as well as assessing the time taken by respondent. The final set of questionnaire was then
developed. Convenience samplings were conducted randomly across 9 districts in Kelantan. A total of
567 respondents were obtained.

3.0 Results and Discussion

Three questions related to flood awareness were asked. One important question include whether the
people were aware of the coming flood season. Slightly more than half answered ‘No’ (Figure 1). Even if
the results are compared in terms of age and education levels, similar assessment can be observed. This
shows that generally the awareness of the people on flood season in Kelantan is not significantly biased
by age and education levels. Results however showed that slightly majority of the people aged less than
18 were aware of the flood season. This is contributed by the sample taken from one of the primary
school in Kelantan (Sekolah Kebangsaan Cenderung Batu, Wakaf Baru, Tumpat Kelantan). Information
gathered from the school teacher suggest that the reason majority of the students answered ‘yes’ is due
to regular flood occurrence in their villages; whether it be minor or major. The North-east monsoon is
experienced every year from November to March. The season overlaps with the school final vacations in
Malaysia which usually starts in December. The associations of the monsoon season to the school final
vacation indirectly force the students to remember and be aware of the flood season (‘Musim Bah’ or
‘Musim Tengkujuh’ among the locals). Moreover, many of the parents are fisherman. Hence, most of the
students are exposed to information related to the monsoon. A different perspective can be seen by
viewing the results based on the locations. Majority of the people located in Kuala Krai and Machang
were aware of the flood season as compared to other locations. The least percentage of awareness is
from Tanah Merah and Pasir Mas.

99
 Figure 1: Awareness on coming flood season.

3.1 Flood Early Warning System

Two main important questions related to flood early warning system are; 1) Were people aware of the
existence of a flood warning system by the authorities; and 2) Did they received any warnings on extreme
rainfall or flood during the December 2014 flood event. Subsequent to those questions important details
gathered from the questionnaire also include which medium of the early warning system they knew, and;
from which medium and who they received the warnings from. Analyses from the question on whether the
people were aware on the flood early warning system by the authority show majority with 67% were not
aware of it (Figure 2). From the 33% of people answered ‘yes’, majority of them knew it from the media
social of Facebook (38%), followed by website (34%), the rest is from the sirens situated near rivers and
others. Exposure of the flood warning system through social media was expected to be high for young
generations (Under 30 year of age). Results however showed even majority of the respondents with age
up to 49 years old selected Facebook. This could signify the power of media social as an information
bridge besides websites.

100
 In order to assess how effective available warnings were prior to the December 2014 flood, the
people were asked on whether any warnings reached them. Despite the lack of awareness on available
flood early warning system, warnings did reached 56% of the people (Figure 3a). Majority of the people
who received the warnings received it through the medium of television (52%), followed by 18.6% through
rumors. By associating the medium the warnings received with whom they received it from (Figure 3b),
majority of the respondents who received the warning through TV had chosen they received it from the
meteorological agency, followed by from the public. The fact that majority of the respondents also chose
public were not expected. Another point concerned is warnings received through rumors were initially
expected to be received from the public; however majority was from the AJKK. This shows the importance
of local community association and could signify the importance of giving early warnings to the local
village committee to further spread the information. Another interesting outcome is on SMS as the
medium of the warnings. The percentage of people received the warning through SMS is higher
compared to from the radio. Majority of the respondents choosing SMS as the medium it received chose
the public as whom they received it from.

(a) (b)

(c)
Figure 2: Awareness on existing early warning system

101
 (a) (b)
Figure 3: Effectiveness of existing early warning system

4.0 Conclusion

4.1 Despite the lack of awareness on available flood early warning system, warnings did
reached 56% of the people. Among the 56%, the effectiveness of the early warning
system highly depends on the dissemination method.
4.2 Warnings received through television and rumors were the highest.
4.3 Rumors or warnings received among the people are majority received through the local
village committee
4.4 Importance of considering giving early warnings directly to the local village committee to
further spread the information.

102
FLOOD OCCURRENCE REDUCTION MEASURES BY RUNOFF PREDICTION BASED ON LAND USE
SCENARIO ANALYSIS USING SCS-CN METHOD AND GIS FOR KELANTAN RIVER BASIN

Project Information
Project Leader : Dr. Siti Nurhidayu Abu Bakar
University : Universiti Putra Malaysia
Address : Department of Forest Management, Faculty of Forestry, Universiti
Putra Malaysia, 43400 UPM Serdang, Selangor
Contact number : +6011-33502559
Email : sitinurhidayu@gmail.com
Project Members : Prof. Madya Dr. Shamsuddin Ibrahim, Dr. Norizah Kamarudin
Mr. Ismail Adnan Abdul Malek, Dr. Khalid Rehman Hakeem
Mr Mohammad Faizalhakim Ahmad Shafuan

1.0 Introduction
Runoff potential is fundamental information in water resource planning and management. The estimation
of runoff potential for large river basins could be complicated due to variability of land cover, soil
properties and rainfall pattern. Integrated river basin management approach is critical to ensure the
conversion of area with high infiltration capacity such as forest into other land use types (low infiltration
capacity) should not reach the hydrological function limits of the basin. This study was aimed to estimate
runoff potential using SCS-CN and GIS by simulating land use and rainfall intensity for Kelantan River
Basin (12940km2). 30 years (1984-2014) climate and hydrological data of the basin obtained from DID
(57 stations) and MMD (23 stations) were analyzed. The land use was classified into eight classes; forest
(9007km2), oil palm (1352km2), rubber (1626km2), paddy (178km2), other agriculture (386km2), urban
(108km2), open area (151km2) and water bodies (135km2). SCS-CN for each land use was obtained
from the previous studies in tropics e.g. Thailand, Malaysia, China and India. The land use change from
1994 to 2014 was analysed based on land use map, Landsat-5 and SPOT-5 satellite imageries. Forest
area in KRB was declining by 14% whereas oil palm and rubber/Timber Latex Clone (TLC) were
increased by 150% and 22%, respectively. The land use conversion of whole basin to forest, rubber, oil
palm and urban has changed the runoff by -30, +42, +107 and +204%, respectively. The eastern and
middle part of Gua Musang and Jeli, the middle part of Kuala Krai and some part of Kota Bharu, Machang
and Tanah Merah were identified as the sensitive areas that significantly increase excess runoff at the
river basin scale. Future simulation for sensitive areas identification will consider different scenarios (e.g.
land use conversion on higher elevations, steeper slopes and shallow soil areas) with different rainfall
intensities based on economic projection until 2050. The identified sensitive areas should be protected
and taken into account for the integrated land use planning and management at river basin scale.

2.0 Methodology
This study estimated runoff for 1994, 2004 and 2014 used the SCS-CN formula as the rainfall, P is the
input and CN value is the abstraction and an output is excessed rainfall after abstraction that become
runoff, Q. The higher rainfall intensity (61mm/hour) was used for estimating runoff considering the highest
th
rainfall intensity in KRB recorded at Kg. Tandak station (71.1mm/hour) on 24 December 2014. CN were
determined based on NEH-4 as follow (McCuen, 1982 and 1989): (a) identify the land use, treatment
class, and soil type in the basin. A soil can be classified based on the minimum infiltration rates; (b)
identify the antecedent moisture condition based on 5-day antecedent rainfall (SCS, 1985); (c) determine
the CN-value for each land use classes from Table TR-55 (SCS, 1986). CN values for KRB were obtained
from the previous tropical research by referring to the soil hydrological index in the area, land use type
and antecedent moisture condition in KRB.
Initially, the land use of the basin was classified into 8 land use types and was adopted from the
previous tropical research by referring to National Engineering Handbook Section 4, Hydrology (NEH-4)
i.e. urban, open area, oil palm, rubber, paddy, other agriculture, forest and water due to limited
information of curve number in tropical. The land use classification for 10 years interval (1994, 2004, and
2014) were processed and digitized used SPOT-5 satellite imagery (2004 and 2014) and Landsat-5
satellite imagery (1994), supporting with land use map obtained from DOE (1984, 1990, 2002, 2004 and
2014) as reference. The ground truthing was conducted to clarify and calibrate the land use classification
processing, there are 135 ground truthing points was selected and observed scattered around KRB and

103
the rate of the accurate point is almost 95% from the processed land use classification with the actual
land use. Then, the soil classification of the KRB categorized into hydrologic soil group (Fig. 3), consist of
Group A, B, C, and D which A (highest infiltration rate) and D (lowest infiltrate rate). As CN is also varied
with antecedent soil moisture conditions, the antecedent moisture condition (AMC II) were used as it
represents moderate moisture conditions and recommended for most hydrologic analysis (Clopper,
1980). CN map with integration of the land use, hydrologic soil group and AMC has been mapped for
each year in 1994, 2004 and 2014.
The runoff potential simulation was simulated whole land use in 2014 into a single land use i.e. forest,
rubber, oil palm and urban based on important land use classes in KRB using different rainfall intensities
classes (light: 1, moderate: 11, heavy: 31, very heavy: 61mm/hour) by DID (2015).

3.0 Results and Discussion

Land use reclassification using Landsat-5 satellite imagery for 1994 showed KRB consists of forest
(81%), rubber (9%), oil palm (4%), paddy (2%), other agriculture (2%), water (1%), urban (0.4%) and
open area (0.6%). In 2004, there was declining of forest area (72%), in inverse other land uses were
increasing rubber (11%), oil palm (10%), paddy (2%), other agriculture (5%), water (1%), urban (0.8%)
and open area (1.2%). Forest area continuing to decline in 2014 to 69%, there was slight decreasing of
paddy (2%) and open area (1%) in 2014 (SPOT-5). Similar trends of growing trends particularly for oil
palm (10%), other land use e.g. rubber (13%), other agriculture (4%), and water (1%) and urban (1%)
showed slight increment. In brief, the land use change (1994-2014) analysis (Fig. 1) indicates that forest
was declining by 14% and paddy by 22% since 1994. On the other hand, oil palm and rubber plantations
were increased by 150 and 22%, respectively. Growing population and economics have boosted the land
conversion to other land uses, there were also an increment observed for other agriculture land (90%),
open area (414%) and urban area (228%).

Figure 1: KRB land use changes based on SPOT5 and Landsat 5 satellite imagery for year 1994, 2004 and 2014
shown an extension of oil palm and rubber area over time

The CN in KRB (Fig. 2) ranging from 39 to 93 (Faizalhakim et al., In preparation; SCS, 1986). The higher
CN values indicates higher runoff potentials. The larger area with higher CN values within KRB was
observed over time (1994-2014) due to deforestation and forest conversion to other land uses (e.g. oil
palm, rubber, other agriculture, open area and urban) (Fig. 1). In 1994, most of KRB area (65%) covered
with moderate high CN (yellow) and 30% of KRB covered with high CN value area (red) (Fig. 2) which is
located in the area which covered with urban, paddy, oil palm and rubber with the HSG C and D. There
are small extension of the high CN value area in 2004 due to increasing of oil palm and rubber area.
Interestingly in 2014, the ratio of high CN and moderate CN area is same due to decreasing of forest area
leads to expansion of agriculture land i.e. oil palm and rubber. If this trend continues, KRB will mostly
covered with high CN which indicates high runoff.

104
 Source: (Faizalhakim et al., In preparation)
Figure 2: CN generated in KRB for 1994, 2004 and 2014, showed the extension area of high CN (red) and the
diminishing of moderate high CN area (yellow) over time (1994 to 2014)

Runoff potential estimation using very heavy rainfall intensity by DID Malaysia (2015) I60=61mm/hour for
1994, 2004 and 2014. The runoff depth in KRB for 1994, 2004 and 2014 are ranging from 0.00–45.21mm
distributed over the KRB. As results, the larger area with very high runoff, high runoff and moderate runoff
potential was observed over time (1994-2014) (Fig. 3). This is related to oil palm and rubber expansion,
and increment of urban and open area with the higher CN values that increase the area with high and
very high runoff potential particularly in the middle and eastern part of Gua Musang, small part in Jeli and
Kuala Krai. There are also high runoff potential observed in floodplain part of KRB i.e. Kota Bharu and
Tumpat. The area consists of urban and paddy fields on clay soil with permanently high water table (HSG
3
D). Overall, the runoff volume (m ) (Table 1) shown an increment of 14% over time from 1994 to 2014 for
whole KRB.

Figure 3: Estimated runoff potential using hourly I60=61mm/hour showed the increasing trends of area with high runoff
potential particularly in east part of Gua Musang, Kuala Krai, Machang, Tanah Merah, Pasir Mas and Kota Bharu
area.

Table 1: Runoff depth (mm) for KRB ranging from 0.00 – 45.21mm, runoff volume (mil m3) increased by 14% for ten
years interval: 1994, 2004 and 2014)
Year 1994 2004 2014
Runoff Depth (mm) 0.00–33.27 0.00–45.21 0.00–41.91
3
Runoff Volume (mil m ) 123.40 135.19 144.42

105
4.0 Conclusion
Important findings of the study are summarized as follows:

4.1 The land use of the KRB changed intensely in the 1994 and 2014 with the decreasing of
forest area from 10,481 to 9007km2 (-14%) and expansion of oil palm plantations and
rubber estates from 540 to 1352km2 (+150%) and 1330 to 1626km2 (+22%), respectively.
4.2 The results shown that the impact of land use changes was significantly increased runoff
potentials and extension area of high runoff potentials within two decades (1994-2014)
due to land use changes.
4.3 The clear effects were observed if the whole KRB converted to: (i) oil palm plantation, (ii)
rubber estates and (iii) urban area, lead to high runoff potential.
4.4 If the whole basin covered by forest, it potentially reduces the runoff potential, but in
certain areas with HSG D produced higher runoff potential.
4.5 The major influencing factors of runoff including rainfall intensity, land use types, and soil
hydrologic conditions and characteristics.
4.6 The eastern and middle part of Gua Musang and Jeli, the middle part of Kuala Krai and
some part of Kota Bharu, Machang and Tanah Merah were identified as the sensitive
areas that significantly increase excess runoff at the river basin scale.

References
Clopper, P.E. (1980). Antecedent moisture consideration in the SCS curve number rainfall-runoff method. MS thesis,
Colorado State University, Fort Collins, Colo.
Department of Irrigation and Drainage Malaysia. (2015). Online rainfall data (mm). Assessed February 25, 2016.
http://infobanjir.water.gov.my/rainfall_page.cfm?state=KEL
McCuen, R.H. (1982). A Guide to Hydrologic Analysis Using SCS Methods. Prentice-Hall, Inc., Englewood Cliffs,
New Jersey.
McCuen, R.H. (1989). Hydrologic Analysis and Design. Prentice-Hall, Inc. Englewood Cliffs, New Jersey, 867pp.
NRCS. (1986). Urban Hydrology for Small Watersheds TR-55. USDA Natural Resource Conservation Service
Conservation Engineering Division Technical Release 55, 164. http://doi.org/Technical Release 55
Soil Conservation Services, SCS. (1956, 1964, 1971, 1985, 1993). Hydrology, National Engineering Handbook,
Supplement A, Section 4, Chapter 10. Soil Conservation Service, USDA, Washington

106
 LONG TERM ANALYSIS OF TIDAL RANGE, HIGH TIDES AND SELECTED ESTUARINE WATER
LEVEL IN EAST COAST OF MALAYSIA

Project Information
Project Leader : Dr Ilya Khairanis Othman
University : Universiti Teknologi Malaysia
Address : Fakulti Kejuruteraan Awam, UTM Skudai, 81310, Skudai Johor.
Contact number : 0132665021
Email : ilya@utm.my
Project Members : Radzuan Bin Sa'ari, Mohamad Hidayat Bin Jamal
Mohd Ridza Bin Mohd Haniffah, Ahmad Shahlan Bin Mardi
Nor Eliza Binti Alias

1.0 Introduction
Summary of actual tide range measurements should be important in verifying cyclical changes in tide
potential that may alter Earth’s climate on short and long timescales (Keeling et al., 1997, Keeling et al.,
2000). The tidal duration and magnitude of the high and low tides can also signify the effect of climate
change. Hence, a summary of long-term tidal information, in general, is useful for assessing climate
change impact to the hydrodynamic behaviour of beaches and estuaries, particularly in the East Coast of
Malaysia that are subjected seasonal flooding.
The fundamental question here is how the spring tidal range (STR) and mean sea level (MSL)
changes over decades? The present study focuses on the trends in STR and MSL surrounding the East
Coast of Malaysia and compares with previous local studies. Additionally, to further understand the tidal
intrusion in Kelantan River Estuary, a field study was conducted to collect the water level and salinity data
along a 12km stretch of Kelantan River Estuary.

2.0 Methodology
2.1 Secondary data collection
The data were collected from Jabatan Ukur dan Pemetaan Malaysia (JUPEM). The present study
compiled more than 25 years worth of hourly water level/ tidal observation records for four tidal stations in
the East Coast namely; Geting (GET), Kelantan, Cendering (CHD), Terengganu, Tanjung Gelang (NKP),
Pahang and Tanjung Sedili (SED), Johor.
The observed water levels are used to calculate the tidal range on a monthly basis. Since there
are two spring tides and two neap tides in every lunar month, only maximum monthly tidal range is taken
for the yearly averaging. The tidal ranges discussed in this section are; yearly averaged monthly spring
tidal range (STR), spring tidal range during the wet season (WTR) and spring tidal range during the dry
season (DTR). The selected months for WTR are November until March, which is during the North East
Monsoon Seasons that bring heavy rainfall to the East Coast of Malaysia. The selected months for DTR
are May to September, which are dry periods with relatively low rainfall (South West Monsoon).

2.2 Primary data collection and fieldwork

The December 2014 flood in Kelantan was associated with the coincidence of high tides. The tidal
records did show the presence of spring tides a few days before the big flood and contribute to the rising
limb of the flood. However, there are no abnormalities in the tidal range during the heavy rainfall, but
there is a small increase in the local mean sea level at that time. Hence, there is a need to understand
the tidal intrusion in Sg Kelantan and the characteristics of Sg Kelantan River estuary ideally prior, during
and after the Northeast monsoon. In this study a field work was conducted prior to the monsoon season
focusing on the last12km stretch of Kelantan River, starting at the inlet of river mouth to upstream of the
river.
The data were limited to secondary water level records at Geting (from JUPEM) and Jeti Kastam
(from DID) and primary water level measurements at three stations namely; Jabatan Laut, Medan Ikan
Bakar and Tambatan Diraja. Local water level and salinity data were obtained by using Water Lever
Logger and YSI Water Quality instruments respectively. During fieldwork, the water level was recorded
from 23 October 2015 till 22 November 2015 at the three stations, shown in Fig. 1. The salinity was
measured on 24 October 2015 and 25 October 2015 at 7 cross sections. The distance from every cross
section was approximately 2km apart starting from the river mouth until 12km upstream (Fig. 1).

107
 Figure 1: Study Location- Kelantan River. (a) Water Level Stations
(b) Salinity measurement sections.

1.0 Results and Discussion

3.1 Mean Sea Level (MSL)
MSL is critical in the determination of the susceptibility of flooding and or inundation at the river mouth.
The increase in MSL increases chances of flooding and saltwater intrusion into a nearby river. In this
study, the hourly water levels data are averaged over the record durations to get the local mean sea level
(LMSL). A unique trend is identified along the East Coast whereby the LMSL is highest around Tanjung
Gelang, 2.8m, and decreases towards northerly and southerly stations (Table 1). The LMSL at northerly
stations decreases by 21% and 18% for Chendering (LMSL=2.2m) and Geting (LMSL=2.3m)
respectively. The southerly stations, Tanjung Sedili (LMSL=2.4m) indicate a lesser decrement of 14%
from Tanjung Gelang. Chendering exhibits the lowest LMSL, which is 60cm lower than Tanjung Gelang.
The yearly mean sea level (YMSL) variations in Figure 2 illustrates an increasing trend similar to
the previous studies (NAHRIM, 2010, Jeofry & Rozainah, 2013). The fitted linear trend analysis suggests
that the rate of sea level rise (SLR) in Chendering is the highest, 0.36cm/year. However, the differences
are insignificant with the other stations (Table 1) The root mean square error (RMSE) between the
computed YMSL using the linear trend and actual data are in the order of 2cm, indicating good
predictions using linear trend. The rate of SLR predicted by the present study is about 1mm/year higher
than Jeofry & Rozainah (2013), but the majority of NAHRIM's prediction (2010) are larger than the current
and previous study.

300
y = 0.3245x - 367.95
280 R² = 0.6269
YMSL, (cm)

y = 0.2792x - 316.76
260
R² = 0.5222
240 y = 0.3469x - 462.91
R² = 0.507
220 y = 0.3596x - 496.43

200 R² = 0.6397
1980 1985 1990 1995 2000 2005 2010 2015

Geting Year Chendering

Tanjung Gelang Tanjung Sedili
Linear (Geting) Linear (Chendering)
Figure 2: Trend of yearly mean sea level.

108
 Table 1: LMSL and estimated rate of SLR based on linear trend analysis
LMSL Estimated rate of RMSE between predicted
Station Duration
(cm) SLR (cm/year) and measured YMSL (cm)
1987-
Geting, Kelantan 231.01 0.347 2.76
2014
Chendering, 1985-
222.64 0.360 2.26
Terengganu 2013
Tanjung Gelang, 1986-
281.09 0.325 2.17
Pahang 2014
Tanjung Sedili, 1987-
241.85 0.279 2.16
Johor 2014
Average 244.15 0.328

3.2 Tidal Range

It is known that the water level observations include non-tidal effects from winds, atmospheric pressure
and river flow on the water level. Hence, to detect the seasonal influence, the tidal range are calculated
during wet and dry seasons. The overall concept is that the during high runoff periods (wet season), the
river flow raises the mean sea level, reduces the tide range, and distorts the tide curve asymmetrically,
which affect the average tidal range in comparison to dry season. As oppose to YMSL, the STR, WTR
and DTR does not illustrate any increasing or decreasing trend, instead fluctuates about their mean value
but with insignificant (Fig. 3). The insignificant % difference are found between STR, WTR and DTR
(Table 2), where Tanjung Gelang indicate/exhibit the highest average STR, WTR and DTR, about 3m and
decreases southeasterly and northeasterly. The STR, WTR and DTR decrease by about 15-18% at
Tanjung Sedili (≈2.5m), 23-24% at Chendering (≈2.3m) and further reduce by 55-56% at Geting (≈1.3m).
The insignificant differences between the STR, WTR and DTR signify that the seasonal variations (limited
to dry and wet seasons) have a negligible impact on the ocean tides or there be a time shift in the
seasonal months and duration.
Among the east coast stations, tidal ranges in Geting indicate the lowest followed by Chendering,
suggesting that flooding in Kelantan and Terengganu may not be strongly due to the tidal intrusion but
perhaps due to wave setup and or wave pumping action. However, the speculations deserve extensive
further analysis that requires combined record investigations to quantify the contribution of different
forcing conditions (i.e. waves, ocean currents, the wind) to the previous flooding in the East Coast. The
local estuary morphodynamics may also influence the flooding.

350

300
Tidal range, (cm)

250

200

150

100
1980 1985 1990 1995 2000 2005 2010 2015
Year

STR GET STR CHD STR NKP STR SED WTR GET WTR CHD
WTR NKP WTR SED DTR GET DTR CHD DTR NKP DTR SED

Figure 3: Tidal ranges trend

109
 Table 2: Tidal ranges values
%Difference %Difference %Difference
%Differe
relative to relative to relative to
STR DTR WTR nce
Station STR DTR WTR
(cm) (cm) (cm) relative
Tanjung Tanjung Tanjung
to DTR
Gelang Gelang Gelang
Geting 133.45 -55.4 132.14 -56.3 137.17 54.6 3.8
Chendering 228.22 -23.7 232.20 -23.2 231.63 23.3 -0.24
Tanjung
299.37 0.0 302.14 0.0 301.81 0.0 -0.21
Gelang
Tanjung
251.15 -16.1 247.39 -18.2 255.87 15.2 3.43
Sedili
- ve sign indicate decreament and + ve sign indicate increment

3.3 Tidal intrusion in Kelantan River Estuary

The December 2014 flood in Kelantan was associated with the coincidence of high tides. The tidal
records did show the presence of spring tides a few days before the big flood and contribute to the rising
limb of the flood in Kelantan. However, there are no abnormalities in the tidal ranges during the heavy
rainfall, but there is a small increase in the local mean sea level at that time. Hence, there is a need to
understand the tidal intrusion in Sg Kelantan and the characteristics of Sg Kelantan River estuary ideally
prior, during and after the Northeast monsoon.
Salinity measurements conducted before the North East demonstrate the longitudinal pattern of
average surface salinity along the 12km stretch of Sungai Kelantan (Fig. 4). The salinity decreases from
river mouth towards the upstream of the river and reaches 0.0 ppt at 11km upstream. The trend of the
salinity is asymmetrical on both right and left side of the river, partly due to the different flow depth, river
bed profiles and freshwater discharge from the riverside housing. Therefore, under heavy rainfall, higher
river discharge may suppress the length of tidal intrusion even more.

15 15 15
Salinity(ppt)
Salinity(ppt)

Salinity(ppt)

10 10 10

5 5 5

0 0 0
0 Distance
2 4from6river8 mouth(km)
10 12 0 2 4 6 8 10 12 0 2 4 6 8 10 12
Distance from river mouth(km) Distance from river mouth(km)
Figure 4: Salinity along the downstream of Kelantan River.

The tidal range calculated from secondary data (Table 3) and recorded from field measurement (Table 3)
indicate below 1.5m. According to Davies (1964), for estuary with tidal range <2m, it can be classified as
micro-tidal estuary where the tides are too small to influence the estuary water level.

Table 3: Tidal range at 5 stations from field record

Tidal Range (m) during spring tides
Station
3 Nov 2015 15 Nov 2015
Genting 0.9 1.2
Jab. Laut 1.1 1.1
Medan Ikan Bakar 1.2 1.1
Jeti Kastam 1.2 1.2
Tambatan Di Raja 1.3 1.1

Geomorphology of Kelantan River Estuary can be classified as a bar-built estuary. The formation of sand
bar or spit between the coast and the ocean can be seen at profile section of A-A’ (Fig. 5). The profile
section is obtained from a part of field survey (Sa'ari, personal communication, December 2015) between
coordinates 688 150m N to 688 300m N. Bar-built estuaries occur when sediment are deposited from
cross-shore transport by ocean waves and currents. The sand bar acts to reduce the wave and tidal

110
actions at the river mouth. However, higher river discharge during the wet season is likely to wash the
sand bar away. The wave setup or wave pumping action could contribute to coastal flooding when waves
overtopped the sand bar. This condition may occur during Northeast Monsoon if the low-pressure system
develops offshore and drive greater wind and larger wave than the usual.

Sand Barrier

Figure 5: Estuary of Kelantan River. Top: Location of section A-A’. Bottom: Profile view of a section of A-A’ using
plotting range scale of the 50m interval.

4.0 Conclusion
A long-term summary of more than 25years of hourly tidal records for four East Coast tidal stations are
presented and analysed. The findings are listed below:

4.1 the local mean sea level is highest around Tanjung Gelang, 2.8m, and decreases
towards northerly stations by 18-21% and southerly stations by 14%
4.2 The predicted average rate of sea level rise from linear regression for the East Coast is
0.328cm/year 1mm/year higher than Jeofry & Rozainah (2013), but the majority of
NAHRIM's (2010) prediction. The largest rate of sea level rise is in Chendering,
0.36cm/year.
4.3 Insignificant % differences are found between values of spring tidal range, wet tidal range
and dry tidal range, signifying that the seasonal variations have negligible impact to the
ocean tides.
4.4 Tanjung Gelang exhibits the highest average spring tidal range, wet tidal range and dry
tidal range and the tidal ranges decrease towards northeasterly stations by 55-56% and
southeasterly stations by 15-24%.
4.5 Salinity measurements before the wet season in Kelantan River Estuary suggest that the
tidal intrusion under normal conditions is up to about 11km upstream.

References
Davies, J. L., 1964. A Morphogenetic Approach to World Shorelines. Zeitschrift fur Geomorphologie, 8, 127pp.
Jeofry, M.H. & Rozainah M.Z., 2013. General Observations about Rising Sea Levels in Peninsular Malaysia.
Malaysian Journal of Science 32 (SCS Sp Issue): 363-370
NAHRIM, 2010. The study of the impact of climate change on the sea level rise in Malaysia (Final Report). National
Hydraulic Research Institute Malaysia, 172pp
Keeling, C.D. and T.P. Whorf, 1997. Possible forcing of global temperature by the oceanic tides. Proceedings of the
National Academy of Sciences, 1997. 94(16): 8321-8328.

111
 FLOOD HAZARD MAP UTILIZING PUBLIC DOMAIN INUNDATION HYDROLOGICAL (RRI) AND
HYDRAULIC (HECRAS) MODELS AND GIS

Project Information
Project Leader : Prof Ismail Abustan
University : Universiti Sains Malaysia
Address : School of Civil Engineering, Engineering Campus, USM, 14300
Nibong Tebal, Pulau Pinang
Contact number : 0124113183
Email : ceismail@usm.my
Project Members : Mohd Remy Rozainy Mohd Arif Zainol, Noorhazlinda Abd Rahman
Choong Wee Kang, Muhammad Salleh Bin Haji Abustan
Nabsiah Abdul Wahid, Sina Alaghmand
Muhammad Azraei Abdul Kadir, Chong Kai Lin
Kaoru Takara

1.0 Introduction
Sungai Kelantan catchment is one of the major catchment in Malaysia which is located at the North
Eastern part of Peninsular Malaysia. The maximum length and breadth of the catchment are 150 km and
140 km respectively. The length of Sungai Kelantan is about 248 km long starts at Bnajaran Titiwangsa
and endup in the South China Sea. It drains an area of 13,088 km2 and occupies more than 88 % of the
State of Kelantan. There are six sub-catchments in Sungai Kelantan namely Galas, Nenggiri, Pergau,
Guillemard Bridge, Kuala Krai and Lebir. The entire catchment contains large areas of tropical forested
mountains, lowland forest and limestone hills. In 2014, two extrem precipitation events were hit the
Sungai Kelantan catchment between 15th and 21st December 2014 with daily rainfall between 100-
300mm while on 22nd and 24th December 2014 the daily rainfall up to 500 mm. The consequent of these
extrem rainfall events, a prolong and extrem flooding occurred during these periods that cause 25 death
and around RM2.81 billion losses.

2.0 Methodology
This study applied a 2D Rainfall-Runoff-Inundation (RRI) model(Sayama et.al. 2010) to simulate the flood
in December 2014 in the Sungai Kelantan catchment. This model has been used intensively in Bangkok,
Upper Citarum watershed, Kabul and Pakistan flood events (Sayama et al, 2012; Sayama et al, 2015;
Nastiti et al 2015; Ruangrassamee et al, 2015). The ground gauge rainfall and field survey of rivers cros-
sectionals data were used as input to the model. The RRI model interface allowed preparation of
geometric data for import into RRI and generation of GIS data from GRASS GIS. A digital elevation model
(DTM) represented by a triangulated irregular network (TIN) then generated by RRI. The automated
procedures for extracting geometric data proved consistent and efficient for the development of floodplain
scenarios. GRASS GIS data generated is used to identify and visualize potential impacts to induced
flooding to the adjacent floodplain. The model performance was investigated compared with a remote
sensing flood extent map (Edlic Sathiamurthy, 2015). The overall objectives of this study as follows:

1) Conduct RRI model simulation over the Kelantan River basin to investigate the model
ability to show flood inundation areas detected by satellite remote sensing;
2) Identify flooded areas that are not detected by remote sensing but indicated by model
simulation;
3) Quantify the effect of flood inundation on streamflow discharge and its peak arrival time.

Finally, all the simulation results could be generated into a flood hazard map and further, delineation of
flood plains could be generated as well. The final output is a floodplain delineation that considered flood
inundation due to downstream river contractions. In view of this, research into flooding represents a
pressing concern and should be seen as one of the most important applied roles of the hydrological
sciences and as a tools towards future sustainability development.

112
2.1 Flood Mapping of Kuala Kerai Extream Flood
Field survey on the flood depth had been conducted from 2nd until 4th March 2015. It was conducted with
several researchers from Disaster Prevention Research Institute, Kyoto University, Japan.

Table 1: Flood Mark Depths in Kuala Kerai, Tangga Kerai and Tualang
Location Reading1 Reading2 Average
Kuala Depan Pasar 3.851 3.940 3.896
Krai SMO bookstore 5.020 5.042 5.031
Pasar Besar 3.615 3.588 3.602
Pondok 3.964 4.150 4.057
Kedai Singer 4.997 5.110 5.054
Bank Islam 5.484 5.164 5.324
Avon Dealer 4.861 5.032 4.947
Oppo Bank Islam 5.062 5.049 5.056
Caltex Petrol Station 4.406 4.206 4.306
Tangga Lan's Jean Repair 4.706 4.668 4.687
Krai Rumah Tepi Sungai Galas 2.947 2.787 2.867
Tualang Rumah MERCY Malaysia 5.318 5.295 5.307
Kampung Hujung Tualang 5.689 5.702 5.696

Table 1 indicates that the maximum water level in Kuala Kerai area reached more than 5 m from the
existing ground level. The worst water level was found in the Kampung Hujung Tualang area with
maximum height of 5.696 m. These values will be used to check inundation level of RRI model in the
model calibation purposes.Paragraphs immediately following their headings are to be justified on both
sides with 0.5cm indentation for first lines. Insert single line spacing throughout the entire document.

2.2 Measurement of River Morphology and Velocity using ADCP

The initial task for research activities is collected available data especially meteorological and catchment
characteristics. The meteorological data were available from Bahagian Pengurusan Sumber Air &
Hidrologi, JPS Malaysia for rainfall and stream flow. However, the cross-sectionals of Sungai Galas and
Sungai Lebir are not available. Therefore, a field survey was conducted several times on the 10-12 of
June 2015, 27-29 of July and 3-5 of August 2015.

2.3 Modelling of Rainfall Runoff Inundation

Rainfall distribution is an important input for RRI Model. Two main input data need in order to used RRI
model namely rainfall and topography data. This study used ground gauge rainfall for the rainfall data.
This data was interpolated over the Sungai Kelantan catchment by the Theissen polygon method.
Topographic data used in this study were obtained from HydroSHEDS (30-s resolusion) (Lehner et al.
2008). The model also requires information on river channel locations and cross-sections. The flow
accumulation data sets included in HydroSHEDS. For the river cross-section, a field survey using ADCP
was conducted. The regression parameter for the river width was estimated at several locations, based
on Google Earth images, while the depth parameters were estimated by ADCP survey depth. The
observed rainfalls from seventeen rain gauges were spatially interpolated over the Sungai Kelantan
catchment using the Inverse Distance Square (IDWS) Method. With the collabration of Kyoto University,
a monthly simulation was conducted for the period 1st December 2014 (00:00) to 1st January 2015
(00:00).

3.0 Results and Discussion

The initial results are presented from RRI are shown in Figures 4 a, b, and c. Digital Elevation Model
(DEM), flow accumulation (ACC), and flow direction (DIR) were the topography input for RRI Model. DEM,
ACC, and DIR were obtained from Hydrological data and maps based on Shuttle elevation Derivatives at
multiple scale (HydroSHEDS).

113
 Figure 4: a) Delineation of Sungai Kelantan and its tributaries, b) Digital Elevation Map (DEM) for Sungai Kelantan
and its tributaries and c) Calculated Flow Direction for Sungai Kelantan Catchment

3.1 Findings of Rainfall Runoff Inundation

The remote sensing image in Figure 5 indicates a large-scale inundation in Kuala Krai, Dabong, Kuala
Balah and Manik Urai areas. Meanwhile, Figure 6 is the simulated og the maximum flood inundation by
RRI for a monthly simulation. Further, Figure 7 shows the superimposed of inundation areas and
simulated RRI inundation results. The good agreement of remote sensing observed data and RRI
simulation indicates that the majority of the inundation flooded area could be well identified and simulated.
However, a small different of flooded areas at Kuala Balah and Dabong could not well simulated possibly
due to the underestimated discharge or overestimated width and depth or meshing size of the simulation
scale give a low visual resolution.

Figure 5: Flood extension inundation mapping on 26th December 2014 from remote sensing (Edlic Sathiamurthy,
2015)

114
 Figure 6: RRI simulation for maximum flood inundation for Kuala Kerai area

Figure 7: Comparison of inundation areas by superimposed observed (remote sensing) and RRI simulated flood
inundation event in Kuala Kerai areas

The two peak discharges and inundations were occured as shown in Figure 8. Two peak discharges and
inundations were occurred approximately on the 19th and 25th December 2014. The results are similar
magnitude and timing with the flood report by DID. The second peak discharge and inundation height are
higher than the first peaks. However, the peak observed hydrograph value was not recorded by the
gauging station because all the equipment in the gauging station was swept away by the extreme flood.

115
 Figure 8: Water depth and discharge over time for Sungai Kelantan at Kuala KraiBrief the results and discussion

4.0 Conclusion
Kelantan suffered from extremely flood which occurred in December 2014. More than 19,544 people were
affected by this flood disaster in the Kelantan River basin in Kelantan. This study, a 2 dimensional model
was applied to simulate rainfall-runoff and inundation simultaneously. The main objective is to discuss
whether or not the simulation model could provide useful information for the extreme flood phenomenon.
The study is particularly focused on how well the simulated flood inundation areas agree with those in an
inundation map prepared with satellite remote sensing.
The RRI simulation results and the remote sensing observed data showed good agreement in the
extended maximum inundation flooded area along Sungai Kelantan particularly in Kuala Kerai areas. A
local DID report confirmed severe flood damage in these areas. The analysis indicated that the model
simulation may be capable of providing additional information to remote sensing. There is no doubt that
model simulation involves large uncertainty. Therefore, we do not argue that such simulation alone can
provide sufficient information for emergency response. Nevertheless, the model simulation performed
here can provide additional information to help identify where flood damage may occur during emergency
situation based on limited local information.

References
Sayama, Takahiro, Tatebe Yuya and Shigenobu Tanaka (2010), ‘Rainfall-runoff-inundation analysis for flood risk
assessment at the regional scale’, Proceedings of the Fifth Conference of Asia Pacific Association of
Hydrology and Water Resources (APHW), 568-576.
Lehner, B., Verdin, K. and Jarvis, A., 2008. New global hydrography derived from spaceborne elevation data. Eos,
Transactions of the American Geogphysical Union, 89 (10), 93-94.
Sayama, Takahiro, Tatebe Yuya and Shigenobu Tanaka (2015),’ An Emergency Response-Type Rainfall-Runoff-
Inundation Simulation for 2011 Thailand Floods, JOURNAL OF FLOOD RISK MANAGEMENT, John Wiley
& Sons Ltd (in press)
Nastiti, Kania Dewi, Yeonsu Kima, Kwansue Junga, Hyunuk Anb (2015),’The application of Rainfall-Runoff-
Inundation (RRI) model for inundation case in upper Citarum Watershed, West Java-Indonesia’, Procedia
Engineering 125, The 5th International Conference of Euro Asia Civil Engineering Forum (EACEF-5) pg 166
– 172
Takahiro Sayama, Go Ozawa, Takahiro Kawakami, Seishi Nabesaka and Kazuhiko Fukami (2012), ‘Rainfall–runoff–
inundation analysis of the 2010 Pakistan flood in the Kabul River basin’, Hydrological Sciences Journal,
57:2, pg 298-312
Piyatida Ruangrassamee, Teerawat Ram-Indra and Patinya Hanittinan (2015), ‘Uncertainty in Flood Forecasting
under Climate Change: Case Study of Yom River Basin, Thailand’, ASCE World Environmental and Water
Resources Congress 2015: Floods, Droughts, and Ecosystems, pg 1155-1162

116
 MODELING FLOOD HYDROGRAPH IN SG. KELANTAN RIVER BASIN USING FUNCTIONAL
CONCEPT

Project Information
Project Leader : Shariffah Suhaila Binti Syed Jamaludin
University : Universiti Teknologi Malaysia
Address : Department of Mathematical Sciences, Faculty of Science, UTM Johor
Contact Number : 07 5534317
Email : suhailasj@utm.my
Project Members : Fadhilah Yusof
Norazlina Ismail

1.0 Introduction
In Malaysia, the phenomenon of extreme rainfall events particularly floods which are highly unpredictable,
contributed to the lost of millions of ringgits and the worst cases, risk lives. Locally, the magnitude of
recent floods seems to be increasing and occurs more frequently. To overcome the flood risks and having
effective planning and management of water resources, river flows must be continuously measured. In
practice, a river may have various shapes of flood hydrographs. The shape of a hydrograph varies in
each river basin and each individual storm event. The objective of this study is to propose a new
framework in hydrological application using the hydrographs as functional data.
Entire hydrograph as a curve with respect to time can be considered as a single observation
within the functional context. There were some efforts to study the hydrograph as a function such as in
the study of the design flood hydrograph (e.g. Yue et al. 2002) and in the flow duration curve study (e.g.
Castellarian et al.2004). However, their studies are remaining limited. Recent study by Chebana et al.
(2012) demonstrated the need to introduce a functional framework to study the whole hydrograph as a
functional observation. Hence, a new statistical framework known as functional data analysis (FDA) is
employed in this study by analysing the whole hydrograph as a functional observation.
The main input in this study is daily streamflow series from Sg. Kelantan River Basin which
constitutes a hydrograph throughout the year. Entire hydrograph as a curve with respect to time can be
considered as a single observation within the functional context. Functional descriptive statistics and
functional principal component are the functional data analysis tools which are introduced in this study. It
is concluded that the method of functional data analysis which treats the whole hydrograph as a function
is more representative of the real phenomena and makes better use of available data.

2.0 Methodology
Suppose we have a data set such as Yi   yi  t1  ,..., yi  tT   , i  1,..., n , j  1,..., T , with T=365 days, n is the

number of years with n = 32, and yi t j   is the flow measured at the day tj of the i-th year. The goal of
FDA is to transform discrete observed data at discrete time intervals to smooth curves xi  t  as temporal
functions through the basis function. A linear combination of basis function is used for representing the
functions, given as
K
xi  t   β ψ
k 1
k k t 
(1)
where βk refers to the basis coefficient, ψk is the known basis function while K is the size of the maximum
basis required. Spline basis is commonly used in FDA for those non periodic data. Since the flow data
shows periodicity, the Fourier basis is employed in the analysis. Fourier basis is written in the form of sine
and cosine functions. The coefficients of the expansion βk are determined by minimizing a least square
criterion. It is essential to choose the number of basis functions that can reflect the characteristics of data.
If a large number of basis functions are used, a penalty term can be added to ensure the regularity of the
smooth function.

117
 In the context of FDA, smoothing location curves can be used to characterize a given river basin
and for comparison or grouping a set of basins visually. Suppose that a sample composed of curves
xi  t  , i  1, 2,..., n .The sample mean and variance functions are defined as
n n

   x t   x t  .
1 1
x t   xi  t  var  t  
2
i
n i 1
n 1 i 1 (2)

The covariance function summarizes the dependence structure between curve values at times s and t,
respectively and can be written as
n

  x  s   x  s    x t   x t   .
1
Cov  s, t  i i
n 1 i 1 (3)

The surface of covariance and the contour map are used to plot the variability of the data set.
On the other hand, the functional principal component analysis (FPCA) can be employed to find
new functions that reveal the most important type of variation in the curve data. Based on the scores of
two main FPCA, we could then classify the data into several clusters to examine the curve patterns. In
order to explore, visualize and examine certain features such as outliers that might not be captured via
summary statistics, three graphical methods are introduced. The first method refers as the rainbow plot is
a simple plot of all the data with the only added feature being a color palette based on an ordering of the
data. The bivariate bagplot is based on Tukey half-space depth function while the functional bagplot is a
mapping of the bagplot of the first two robust principal component scores to the functional curves. The
last method, the functional HDR boxplot is based on the bivariate HDR boxplot applied to the first two
principal component scores. The bivariate HDR boxplot is constructed using a bivariate kernel density
estimate. A detail review of these three methods can be found in (Hyndman and Shang 2010).

3.0 Results and Discussion

A functional hydrograph constitutes a year starting from 1 January through 31 December of the year.
Using a suitable number of basis functions, the discrete flow data are converted to smooth curves. The
smooth representation of flow data is done with a 365 day base period and selected k basis functions.
Large values of k allow us to reach almost all the daily flow points including the peaks. However, it was
found to be less smooth and the variance in the first few components cannot be captured. On the other
hand, small values of k gave a bad quality of smoothing. Hence, it is recommended that to choose a
number k which combines the quality of smoothing and a high percentage of explained variance by PCA
analysis. Hence, the number k =183 is chosen based on the quality of smoothing and a high percentage
of explained variance. Fig.1 shows all the smoothing curves for the studied period with 183 basis
functions. It shows that these curves have several different shapes.

Figure 1: Smoothing curves of flow data with 183 basis functions.

For a certain flood event, they might be a case that there exist a multipeak of flood hydrograph in
a given river basis. Compared to univariate and bivariate approaches, the functional consider the whole

118
hydrograph as functional observation. Instead of several univariate or multivariate analysis, one analysis
can be conducted for the whole data using FDA. Information on first and second derivative could also be
obtained.
The smooth location curves showing the mean, median and modal curves are presented in Fig. 2.
The maximum flow is normally observed in the middle of November up to early January which can be
considered as the Northeast Monsoon flow. Due to the contribution of heavy rainfall during the Northeast
Monsoon, it is often found that heavy floods occurred during this period. This kind of flood is exhibited by
the median curve which is higher than the mean and the mode.

Figure 2: Fourier smoothing location curves.

Fig.3 shows the temporal bivariate variance-covariance surface and its corresponding contour map. As
shown in the figure, the highest variability occurs at the end of the year and early January in which this
period is also corresponds to the highest flow.

Figure 3: Estimated variance-covariance surface of the flow curves for years 1980 to 2014.

The most important type of variation in the curve data is explained by the functional PCA. The
scores of the first two principal components were mapped onto Fig.4. Several clusters can be classified
according to these scores. The curves for 2014, 1988 and 1993 may be considered having a unique
cluster of their own while for certain curves, they are high possibilities that they can be grouped together.
Based on these findings, it could be said that the shape of the hydrograph curves are different for a
certain year based on the behaviour of the flow data. Hence, a classification of the hydrographs based on
their shape is very important.

119
 Figure 4: Mapping of the scores of the first two principal components.

In order to justify the above unusual years, the outliers’ detection methods were employed. The functional
HDR boxplot is a mapping of the bivariate HDR boxplot (Hyndman 1996) of the first two robust principal
component scores to the functional curves. The functional HDR boxplot displays the modal curve, the
inner and the outer regions. The inner region is defined as the region bounded by all curves
corresponding to points inside the 50% bivariate HDR and usually 99% outer highest density region.
Fig. 5 displays the bivariate HDR and the associated functional HDR boxplots of the smooth flow
curves for 99% of probability coverage. With the 99% coverage probability, the outliers detected in flow
series are 1988 and 2014.

Figure 5: (a) Bivariate score HDR box-plot and the corresponding (b) functional HDR
box-plots with 99% of probability coverage.

4.0 Conclusion

4.1 The location curves which represent the mean, median and modal curves are much
better in providing more information, specifically in adding temporal aspects concerning
the hydrological regime in the basin than classical statistical approach. Our findings
indicate that the highest flow at Sg. Kelantan is observed between November and
January. It is suggested that the flood which occurred during this period is exhibited by
the median curve. The highest variability was again observed between November and
January (northeast monsoon season) as shown by the bivariate (temporal) variance-
covariance surfaces as well as the first two functional PCA.
4.2 It is found that FDA approaches gave an additional insight to the hydrological regime
variability than the real value or matrix in the univariate and multivariate contexts. In
addition, the visualization methods via functional bagplot and functional HDR boplot were
used in this study to provide information that might not have been apparent using

120
 mathematical models and summary statistics. Based on 99% of probability coverage,
1988 and 2014 are considered as outliers since they lie outside the region.
4.3 The functional framework is more general and more flexible and can represent a large
variety of hydrographs and able to make use the full information contained of the
hydrograph.

References
Yue, S., Ouarda, T.B.M.J., Bobée, B., Legendre, P., Bruneau, P., (2002). Approach for describing statistical
properties of flood hydrograph. Journal of Hydrologic Engineering, 7,2, 147-153.
Castellarin, A., Vogel, R.M., Brath, A., (2004). A stochastic index flow model of flow duration curves, Water
Resources Research, 40, W03104, doi: 10.1029/2003WR002524.
Chebana, F., Dabo-Niang, S., Ouarda, T.B.M.J., (2012). Exploratory functional flood frequency analysis and outlier
detection, Water Resources Research, 48, W04514. doi: 10.1029/2011WR011040.
Hyndman, R.J., Shang, H.L., (2010). Rainbow plots, bagplots, and boxplots for functional data, Journal of
Computational and Graphical Statistics, 19,1, 29-45.
Hyndman,R.J. 1996. Computing and Graphing Highest Density Regions." The American Statistician.
50(2): 120-126.

121
 HYDRODYNAMIC SIMULATION OF KELANTAN RIVER BANKEROSION AND CHANNEL CHANGE
DURING HEAVY FLOODS

Project Information
Project Leader : Zainab binti Mohamed Yusof
University : Universiti Teknologi Malaysia
Address : Department of Hydraulics and Hydrology, Faculty of Civil Enginering, UTM
Skudai, Johor Bahru.
Contact number : 013-7709717
Email : zainabyusof@utm.my
Project Members : Noor Baharim Bin Hashim, Shahabuddin Bin Amerudin
Muhammad Zulkarnain Bin Abdul Rahman, Noraliani Binti Alias
Zulkiflee Bin Ibrahim, Zulhilmi Bin Ismail
Radzuan Bin Sa'ari, Mushairry Bin Mustaffar

1.0 Introduction
River bank erosion is a natural process that over time has resulted in the formation of the productive
floodplains and alluvial terraces. Massive flood events like flooding in Kelantan, December 2014 can
trigger dramatic and sudden changes in rivers and streams. However, land use and stream management
can also trigger erosion responses. The responses can be complex, often resulting in accelerated rates of
erosion and sometimes affecting stability for decades. The erosion process starts when raindrops
dislodging soil particles and runoff water carries the dislodge particles to the river. Erosion by water has
appears to be one of the chronic phenomenon whereby sediment budget involved estimating the
sediment contributed from three main processes of erosion; hillslope erosion, gully erosion and bank
erosion. The Department of Irrigation and Drainage (DID) reported that annual flooding of Sungai
Kelantan is due to river’s bank overflows, which occurs at least once a year. It becomes worst when
associated with a northeast monsoon climate experiences from November to February, in which most of
th th
the areas in Kelantan hit by heavy rainfall from 17 - 24 December 2014. The 2014 Kelantan flood
event was said as the worst flood after floods in 1984. In this study, the objectives are (1) to measure the
stage-discharge, flow characteristics and flow resistance for inbank and overbank flows of Kelantan River
and its tributaries, (2) to model the turbulence flow due to the secondary flow as a result of bed shearing
stress leading to bank erosion, and (3) to map bank erosion for flood risk assessment.

2.0 Methodology
2.1 Phase 1: Determination of Hydraulics and sediment characteristics
This phase involves site investigation on flow parameters such as velocity, river cross-sections, water
level, rainfall and soil types. Hydraulics and hydrology data can be obtained from the DID, whilst, for soil
type characteristics were taken from the Minerals and Geoscience Department. Extreme rainfall data was
gathered between the year 1990 and 2014, particularly rainfall from December to January. Similarly on
stream flow data, which was taken directly form DID for the same monthly data.

2.2 Phase 2: Modelling of river bank erosion

The simulation combined a model of fluvial bank erosion with bank stability analyses to account for the
influence of hydraulic erosion on mass failure processes, based on parameterized outputs from detailed
in phase 1. The results were to identify two mechanisms that explain how most bank retreat usually
occurs during flood peaks. This includes the investigation of maximum flow velocity during the flood that
migrates away from the outer bank as flow discharge increases, reducing sidewall boundary shear stress.
In this study, a three-dimensional flow-sediment model of the Kelantan River is developed using the
EFDC (Environmental Fluid Dynamic Code) in order to study the changing riverbed. The model will be
also calibrated and validated by daily measuring of the water surface elevation and using suspended
sediment concentration data secured from flood data event between December 2013 and December
2014. Based on this model, the mobile sediment on the riverbed was simulated for the high water period
of the Kelantan flood event 2014. The EFDC model's hydrodynamic component is based on three-
dimensional hydrostatic equations formulated in curvilinear-orthogonal horizontal coordinates and a
sigma vertical coordinate (Hamrick, 1992). Thus, in this application, the EFDC model was configured to

122
simulate time varying surface water elevation, velocity and cohesive and non-cohesive sediment
transport.

2.3 Model Setup

EFDC is a general purpose modeling package for simulating three-dimensional flow, transport and
biogeochemical processes in surface water systems including rivers, lakes, estuaries, reservoirs, wetland
and coastal regions. The EFDC is an advanced public domain model developed by Dr. John Hamrick,
hydrodynamic components which based on three-dimensional hydrostatic equations formulated in
curvilinear-orthogonal horizontal coordinates and a sigma vertical coordinate. The model is calibrated and
validated by daily measuring of the water surface elevation and using suspended sediment concentration
data secured from flood data event in December 2014. Based on this model, the mobile sediment on the
riverbed changes was simulated for the high water period of the Kelantan flood event 2014. The Digital
Terrain Model (DTM) uses in the model was generated by combining different data sources, as follow,
1. Airborne LiDAR,
2. Interferometry Synthetic Aperture Radar (IfSAR) and
3. Shuttle Radar Topography Mission (SRTM).
All data was resampled and combined to produce a single DTM with a spatial resolution of 10 m. As for
model’s cross-sections, it divided into several reaches, which was 500 m interval. Meanwhile, for initial
and boundary conditions setup, the domain was assigned the rainfall event from January 2014 to
December 2014 data. The model’s initial and boundary setup as follow,
1. Downstream boundary condition: Water level data at Kota Bahru (Guillemard Bridge station)
2. Upstream boundary conditions: Streamflow data at Lebir river (Tualang station) and Galas
river (Dabong station).

3.0 Results and Discussion

3.1 Model
General verification th th

DecemberAras
Aras
Taburan
Figure 1 shows Air
Hujan-
the Laporan
observed Catatan Aras
data Air
of Tertinggi
maximum Mengikut
waterTempat
stage at selected stations in Kelantan on 24 - 26
Air Sungaiwhilst, in Figure 2 indicates the model verification of water stage at Kuala Krai station.
2014,
The resultLaluofLintas
model’s verification of water stage at Kuala Krai station gave a good agreement with
Kemalangan
observed data Tempohfrom: 09- 09-
the2015DID. Meanwhile,
- 09- 09- 2015 Figure Sungai : Semuathe
3 shows Sungaieroded areas (red circles)
Cari due to maximum
Cetak
shear stressGerakan Menyelamat/Bantuan
happened in a December flood event. A detailed of erosion model near Dabong have shown
Pemindahan
a critical bed shear Mangsastress happened near Kuala Krai at Lebir’s river (a circle red), as shown in Figure 4.
Therefore,Agihan
theBekalan Banjir of maximum shear stress predicted indicates the occurrence of high velocities
result
Bantuan Kesihatan
leading to a river bank failure at Dabong and Kuala Krai. This can be seen in Figure 5, where the model
Bantuan PBSMpredicted the high potential area of erosion happened during the peak flow. The result
has successfully Laporan Catatan Aras Air Tertinggi Mengikut Tempat
was then compared with the actual image taken from the google images; it showed that the massive
erosion happened at the same location as the model predicted.
Kemaskini Pada : 09-09-2015 06:02:46
Text
Aras Normal Aras Berjaga Aras Amaran Aras Bahaya Catatan Tertinggi
Bil Sungai Tempat
(M) (M) (M) (M) Tarikh & Waktu Sukatan (M)
1 Sungai Galas Dabong 28.00 32.00 35.00 38.00 24-12-2014 16:00 46.47
2 Sungai Lebir Tualang 23.00 27.00 31.00 35.00 24-12-2014 06:00 42.17
3 Sungai Kelantan Tangga Krai 17.00 20.00 22.50 25.00 25-12-2014 15:00 34.17
4 Sungai Kelantan Jambatan Guillemard 10.00 12.00 14.00 16.00 26-12-2014 00:00 22.74
5 Sungai Kelantan Tambatan DiRaja 1.00 3.00 4.00 5.00 26-12-2014 00:00 6.89
6 Figure
Sungai Golok1: Maximum
Jenob water stage recorded
19.00 for21.50
selected 22.50 23.50
stations in 12-01-2014
Kelantan 00:00 2014).
(DID, 25.10
7 Sungai Golok Rantau Panjang 5.00 7.00 8.00 9.00 18-12-2014 11:00 10.84
8 Sungai Semerak Pasir Putih 0.40 2.00 2.30 3.00 18-12-2014 07:00 2.67

123
 DIDdata Model
40

35

30

Water Level (m)

25

20

15

10

0
335 340 345 350 355 360 365
st st
Figure 2: Model verification of water stage between observed and simulated data from 1 – 31 December 2014 at
Kuala Krai station.

th
Figure 3: Simulated bottom shear stress of the riverbed on 26 December 2014.

(a) Before flood 2014 (b) After flood 2014

Figure 4: Google Earth Images; (a) before flood 2014, (b) eroded area after flood 2014.

124
 Figure 5: Model erosion against actual erosion; (a) simulated of maximum shear stress, (b) actual eroded area after
December flood event.

4.0 Conclusion
The study can be concluded as follows:

4.1 Flood episode happened in Kelantan, in December 2014 has triggered dramatic and
sudden changes in rivers and streams - due to extreme rainfall that creates rapid
changes in flood velocities, leading to have potential of having river channel erosion.
4.2 The preliminary results of water column and bed shear stress indicate that the critical
erosion was occurred at the river’s banks near Kuala Krai and at the upstream at Lebir
River, respectively.
4.3 As for velocities analysis, it has found that the velocities increase along the river of Sg.
Kelantan. For example, the critical velocity was expected to occur at the junction between
Lebir River and Galas River, near Kuala Krai. Thus, it marked as potential areas of river
bank erosion.

References
Hamrick, J. M. (1992). A three-dimensinal environmental fluid dynamics computer code: theorical and computational
aspect. Special report No. 317, in Applied Marine Science and Ocean Engineering.
Department of Irrigation and Drainage, Malaysi (DID), (2014). http://ebanjir.kelantan.gov.my/p_parpt01.php.
Retreived on 9 July 2015.

125
 Lampiran 1

END OF REPORT

Project Title : DETERMINATION OF FLOODPLAIN CHANGE AND ITS IMPACT

ON FLOOD MODELLING USING MULTI-TEMPORAL DEM AND LAND
USE/LAND COVER FROM INSAR/SAR TECHNIQUE
A. Project Information
Start Date : 01/04/2015
End Date : 31/12/2015
Extension Date : 31/03/2016
Project Status : Completed
Project Leader : DR. MD LATIFUR RAHMAN SARKER
I/C Number : 12159
University : UTM, Malaysia
Address : FAKULTI GEOINFORMASI DAN HARTA TANAH
Contact number : 01115158825
Project Members : Prof. ZULKIFLI BIN YUSOP, PEJABAT TNC(P&I), DR. Ami Hassan Bin Md

Din, FAKULTI GEOINFORMASI DAN HARTA TANAH

B. Project Achievement
Project Progress : 100%
Research Output : (_2_), Conference Proceedings

Talent : RA (4_), PhD student (_0_), Master student (2_)

C. Expenditure
Budget Approved : RM 96000
Amount Spent : RM 95832.14
Balance : RM 167.86
% of Amount Spent : 99.83
Summary of Research Findings
Floods are one of the most severe natural disasters around the world which cause huge damage
of property and loss of life every year. There are several causes for flooding (Merkuryeva et al.,
2014) but the most common cause of flooding is the overflow of streams and rivers and
abnormally high tides resulting from severe storms (Larry, 2011). The most significant natural
disaster in Malaysia is floods that have been recorded since 1920s. Recently, the worst flood in
decades was occurred from 25 December 2014 to 05 January 2015 that affected 200,000 people
in several states in Malaysia such as Kelantan, Terengganu, Pahang, Perak, Sabah and Sarawak
due to heavy monsoons rainfall which claimed the 21 lives and cost nearly million dollars of
property. The natural disaster such as floods can’t be stopped quickly but damage of property
and loss of life can be reduced or minimized by the appropriate flood monitoring or modelling
approach.
In floods modeling, two important parameters of floodplain are necessary; land use/land cover
(LULC) and topography (Chen et al., 2009; Amini et al., 2011; Fox et al., 2012). Although
several studies have pointed out that changes in LULC and topography have enormous impacts
on the flood modelling, but unfortunately most of the studies related to flood modelling have
ignored the importance of these two parameters probably due to the unavailability of appropriate
LULC change information as well as Digital Elevation Model (DEM) resulting in uncertainties
in flood modelling results. Therefore, this study is going to take an opportunity to examine the
effects of these two parameters especially by generating DEMs from the multi-sensor SAR data.
Considering the research goals, this study produced several outputs before going to develop
flood model which includes i) updating the LUCL maps and generating new LULC maps based
on several available data, ii) examining the existing DEM and exploring the possibility the use of
existing DEM for the generation of flood model, iii) generation of DEM from the SAR data, iii)
generating flood models using different types of LULC maps and DEM generated from the
different sources. Although, the initial result is promising, however, it is important to mention
that more investigation is required to validate the generated DEM as well as to generate multi-
temporal DEM for this study area. Flood modeling results using LiDAR DEM indicates a
successful determination of flooding extent and agrees with field data. Nevertheless, the quality
of DEM needs to be improved and flood model using InSAR data needs to be investigated along
with different types of LU/LC information in future endeavors since it was not possible to
investigate all the aspect rigorously due to constraint, man-power and financial ability.

1.0 Introduction
Floods are natural events that have always been an integral part of the history of earth. Flooding
occurs i) along rivers, streams, and lakes, ii) in coastal areas, iii) on alluvial fans, iv) in ground-
failure areas such as subsidence, v) in areas that flood due to surface runoff and locally
inadequate drainage, and so on. Floods damages are usually reported as direct damages of
property and loss of life but there are several indirect damages results from the adverse effects of
flooding. Floods are no doubt a major hazard and the risks they pose are increasing due to shifts
in meteorological forcing, population pressure, as well as anthropogenic change to riverine
landscapes (Schumann et al., 2015). There are plenty of examples of flooding around the world
indicated that severity and the intensity of flooding is increasing, and one of the recent examples
is the 2014-2015 flooding in Malaysia from 15 December 2014 – 3 January 2015 which affected
more than 200,000 people while 21 killed persons (AsiaOne, 2013).
Indeed, the necessity of monitoring and modelling of flood events is enormous for various
reasons especially for the planning of flood mitigation and floods control is always a huge
challenge for governments and local authorities (Chiang et al., 2010) because of the complexity
of the processing and requirements of precise measurement of several parameters. In the
developed world, dense river gauging networks, high quality channel survey data, as well as fine
resolution ground elevation data over floodplains are available that allow long-term monitoring
and modeling of flood events effectively. In the developing world, however, the situation is very
different where river-gauging stations are often sparse and are only operated in very large basins,
channel survey data hardly exist and if so, they are many restrictions to use or share this data
(Merkuryeva et al., 2014). However, several hydraulic models are available for the modeling of
flooding such as HEC-RAS, LISFLOOD-FP, TELEMAC-2D, and so on (Bates and De Roo,
2000) that required simple information. Moreover, in recent years, substantial efforts are being
made to improve this complex situation of observing, mapping, and modeling flood processes,
both in terms of flood model development and remote sensing, particularly satellite platforms
(Merkuryeva et al., 2014).

Nevertheless, efforts have already been made for the modelling of flooding by the integration of
hydraulic models and remote sensing data (Merkuryeva et al., 2014). The key purpose of all
these flood modeling approaches is to accurately predict inundation extent and risk (Di
Baldassarre et al., 2010; Pappenberger et al., 2007), however, most of the modeling technique at
least requires few measurements such as soil, underlying geology, land use/land cover (LULC)
of the watershed, channel topography, and initial and boundary conditions (Anmala et al., 2000,
Sanyal et al., 2014). While geology, and soil types of watersheds generally remain the same, the
initial condition and boundary condition can be estimated by interpolating the observations from
available gauges, and by specifying the upstream and downstream ends of the system
respectively. However, several human activities cause changes in LULC and to some extent
topography in watersheds over time (van Dijk et al., 2011; Nagy et al., 2011), therefore, an up-
to-date LULC as well as an accurate DEM are necessary but unfortunately this requirement is
mostly ignored and in most of the cases, researchers use LULC and DEM from the secondary
sources (such as SRTM and ASTER GDEM) without considering the effects of these two
parameters on flood modeling approach. Therefore, this proposed study investigates the effect of
DEM and LU/LC on flood modeling.

2.0 Methodology
In this study, two methodologies were developed: one for DEM generation and the other for
flood modeling. Firstly, a DEM was generated for the study area using ALOS PALSAR data
following the several steps (Figure 1) but main data processing steps include; i) multi-looking, ii)
image co-registration, iii) raw interferogram calculation, iv) interferogram phase flattening, v)
filtering, vi) phase unwrapping, vii) phase to height conversion, and viii) geocoding. Secondly,
another methodology was also developed using 4-steps of main processing for doing flood
modeling using LiDAR data (Figure 2).
 Figure 1 Overall Methodology for DEM generation Figure 2 Overall Methodology for Flood Modeling

3.0 Results and Discussion

The results of DEM generation using ALOS PALSAR data are presented in Figure 3, 4 and 5 for
flat view, shaded relief view and contour view of DEM respectively. However, the results of
flood modeling are presented in Figure 6, 7 & 8. It is clear from the DEM results (Figure 3, 4, 5)
that a DEM can be generated from the SAR data using InSAR technique effectively.
Nevertheless, more investigation is needed to do the validation and to improve the accuracy of
DEM using better processing software in order to make this DEM usable for flood modeling. On
the other hand, it is obvious from the results of flood modeling (Figure 7 & 8) that modeling of
flooding can be done appropriately if necessary data especially accurate DEM and LU/LC maps
are available. This study found that flood depth using original LiDAR data was not well-
delineated (Figure 7) due to some unwanted points/errors in the DEM. However, a better
delineation of flood depth was found after correcting the original LiDAR data based on TIN
(Figure 8).

Figure 3: Flat view of PALSAR DEM Figure 4: Shaded view of PALSAR DEM Figure 5: Contour view of PALSAR DEM
 Figure 6: River cross-section with LU/LC Figure 7: Flood depth using original Figure 8: Flood depth using modified
maps LIDAR DEM LIDAR DEM
4.0 Conclusion
It is possible to detect flood depth and velocity accurately using accurate DEM and
LU/LC data. However, getting an accurate DEM and LU/LC is not easy and needs lots of
effort and time especially time series information.
This study found that several factors are actually involved for creating the devastating
flooding in Kelantan, however, major factor of this flooding is excessive rainfall within a
short period. Other factors such as LU/LC change is important and need to investigate
further as we were unable to detect the effect of this factor due to time limitation.
It is important to note that a simulation can be done in future using different levels
of rainfall in different watershed in order to understand the relationship between
rainfall and flooding depth. If this process is done successfully in future, then this
information can be incorporated with the real time flood warning system.
This study tried to generate DEM from ALOS PALSAR data using InSAR technique. A
promising result was obtained but further investigation is needed for the improvement of
the PALSAR DEM using robust processing software and necessary field data.
There are several limitations in this study especially constraint of time, man-power,
financial ability to obtain good data to procure robust processing software. All these
problems make this study difficult to complete but we have tried our best to finish this
work successfully. Hope, all these difficulties make us more confident in future to do a
better research.

References
Amini, A., Ali, T., Ghazali, A., Aziz, A., Akib, S., (2011). Impacts of land-use change on streamflows in
the Damansara Watershed, Malaysia, Arabian Journal for Science and Engineering, 36, 5, 713–720.
Chen, Y., Xu, Y., Yin, Y., (2009). Impacts of land use change scenarios on storm-runoff generation in
Xitiaoxi basin, China, Quaternary International, 208, 121–128.
Fox, D.M., Witz, E., Blanc, V., Soulié, C., Penalver-Navarro, M., Dervieux, A., (2012). A case study of
land cover change (1950–2003) and runoff in a Mediterranean catchment. Applied Geography, 32, 2, 810–
821.
Larry W. M. (2011). Ground and Surface Water Hydrology, John Wiley and Sons, Inc.
Merkuryeva, G., Merkuryev, Y., Sokolov, B. V., Potryasaev, S., Zelentsov, V. A., Lektauers, A. (2014).
Advanced river flood monitoring, modelling and forecasting, Journal of Computational Science, 10, 77-85.
 Lampiran 2

END OF REPORT
(maximum 5 pages of end report)

Project Title : SIGNIFICANCE OF SEASONAL RIVER MOUTH SANDBAR

FORMATION TO FLOOD LEVELS IN THE KELANTAN RIVER
BASIN

A. Project Information
Start Date : 01/04/2015
End Date : 31/12/2015
Extension Date : 31/03/2016
Project Status : Completed
Project Leader : AHMAD KHAIRI BIN ABD WAHAB
I/C Number : 611001-08-6487
University : UNIVERSITI TEKNOLOGI MALAYSIA
Address : PEJABAT TNC(P&I)
Contact number : 012-7132300
Project Members : MUSHAIRRY BIN MUSTAFFAR
RADZUAN BIN SA’ARI
ZULHILMI BIN ISMAIL
FARIDAH BINTI JAFFAR SIDEK
ILYA KHAIRANIS BINTI OTHMAN

B. Project Achievement
Project Progress : 90%
Research Output : Indexed Journal (-), Non-indexed Journal (-), Conference Proceedings
(1), Book Chapter (__), Others (1)
Talent : RA (1), PhD student (-), Master student (__)

C. Expenditure
Budget Approved : RM 64,100.00
Amount Spent : RM 39,432.43
Balance : RM 24,667.57
% of Amount Spent : 61.57%
Summary of Research Findings

1.0 Introduction
The mouth of Sg. Kelantan is subject to the persistent dynamic forces of the sea and seasonal
river flows. The monsoon regimes, dominated by the North-east monsoon (between the months of
November to March) bring strong winds and waves from the northeast direction while during the
South-west monsoon (between May to September), a relatively calmer period prevails along the
eastern coastlines of Peninsular Malaysia.
There are strong interdependency between these forces, which creates a state of dynamic
equilibrium between the coastal processes of littoral and onshore-offshore sediment transport
leading towards an evolution of the shorelines towards erosion, accretion and stability.
At river mouths, another important element that needs to be included is the river discharge that
brings along its own load of inland sediment out to the sea. Thus, the mouth of any major river
would be subject to an interaction between the various factors, which, inevitably, creates a pattern
of sedimentation, erosion and mass movements around the river mouth area that is complex and
seasonal. These are normally realized through as the formation of sand bars and offshore islands
that partially blocks the river outflow, causing further sedimentation towards inland and may
affect the efficiency of the river outflow especially during the months of heavy river discharges.
The research attempts to investigate the significance of this phenomenon on the outflow of Sg.
Kelantan, especially during the monsoon season when the precipitation rates and the magnitudes
of winds and waves from the sea are at their highest. This would be a major contributor towards
increasing the flood levels and the inundation periods in the coastal areas.
A critical element of this research is the inclusion of seasonal sandbar migration obtained through
field monitoring and model simulation of the flow hydrodynamics through the sandbar
formations. An understanding of these two interlinked processes would lead to a better strategy of
flood mitigation at the coastal plains and river mouth area.

2.0 Methodology
The flow hydrodynamics within the estuary and the general area of the surrounding waters around
the Kelantan river mouth were simulated using the software package TELEMAC-2D (LNH-EDF,
2002). The TELEMAC system is a fully vectorised finite element software for the solution of the
shallow water equations. The flow field is depth-averaged over a two-dimensional horizontal
domain. The finite element formulation requires discretisation of the model domain into non-
structured spatial meshes or triangular or quadrilateral elements in Cartesian or spherical
coordinates. For any numerical modelling executions, the most important element is the
availability of good quality data to set up the model domain and to calibrate and validate the
dynamic processes being simulated. The mouth of Sungai Kelantan, also known as Muara Kuala
Besar has a very dynamic morphological system. Past detailed hydrography and topography
surveys are very crucial to understanding and quantifying the dynamics involved. However, for
this site, such information are not readily available except for a 2009 survey conducted by the
Marines Department Malaysia. Apart from the river and near shore processes, the research also
investigated shoreline evolution of the nearby coastlines to further enhance the littoral sediment
transport investigation. This was done by analysing Landsat 4-5 TM satellite images for the years
 1990 to 2015 at 5-year intervals. The analyses were conducted using the Digital Shoreline
Analysis System (DSAS), an extension of ArcGIS software, developed by the USGS.
Two sets of primary hydrographic data consisting of water level monitoring, sediment sampling,
current measurements and river and nearshore hydrography surveys of Sungai Kelantan and
Muara Kuala Besar were conducted by the project team from 30th May to 5th Jun 2015 and from
6th to 12th March 2016. These were used for the hydrodynamic modelling work and they are also
can be made available through the Pengurusan Data Tebatan Banjir 2015 data repository.
The coastal area survey covered an area with a rectangular size of 3.6km (alongshore) by 2.2km
(offshore). River survey coverage was 13km inland to the Sultan Yahya Petra Bridge in Kota
Bharu.
Secondary data were sourced from the Department of Drainage and Irrigation, JUPEM, Royal
Malaysia Navy, MetMalaysia and Agensi Remote Sensing Malaysia. Additional survey records of
Kuala Besar (2009) were also obtained courtesy of the Malaysia Marine Department.

3.0 Results and Discussion

Results of the modelling outputs were analysed in terms of flow capacity for the river, water level
profiles, flow pattern and magnitudes, sediment transport and shoreline changes. The satellite
images purchased through this project were used to obtain shoreline evolution and longshore
sediment drift.
A notable profile change is from the 2010 and 2015 records, depicting pre- and post-2014 flood
event as shown below (Figure 1) where transects close to discharge outlets record increased
sedimentation after the flood event.
3
4
5

26
25
24
6

27
23

34
7

28

33

35
22

29
8

36
32

37
30
9

21
10
11
12

20

38
14
13

15

39
16
17

40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55

LEGEND
Transects

/
LEGEND

/
Baseline
Shoreline_2010 Shoreline_2010
0 1 2 4 0 1 2 4
Shoreline_2015
Kilometers 1:72,000 Shoreline_2015 Kilometers 1:83,000

Shoreline Change Rates (m/yr) 2010-2015

280

260

240

220

200

180
Shoreline Change Rates (m/yr)

160

140

120

100

80

60

40

20

-20

-40

-60

2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64
Transect

Figure 1: Shoreline evolution around Muara Kuala Besar from Landsat data for 2010-2015 period
Hydrodynamic modelling of 2015 and 2016 bathymetry were conducted using normal and flood
discharges values. A flood discharge value of 900m3/s were adapted for the simulation. The time
variation of water levels at the Medan Ikan Bakar station reach are shown below in Figure 2. The
blue line indicated measured level during normal condition and the red line are simulated flood
levels with discharge of 900 m3/s. Water level elevations along the river were charted up to the
13km mark from the river mouth. Figure 3 shows the variation along the river reach from Km-0
at the river mouth to Km-13 at Sultan Yahya Petra Bridge for the flood tidal phase. Comparison
with other stages of the tide depicted similar patterns. Figure 4 shows the ebb and flood current
vectors at Muara Kuala Besar.
2.500

2.000
Measure
1.500 900q 20r

1.000
9/11/2015
9/11/2015
0:00 10/11/2015
12:00 10/11/2015
0:00 11/11/2015
12:0011/11/2015
0:00 12/11/2015
12:0012/11/2015
0:00 13/11/2015
12:00 0:00

Figure 2: Comparison of flood and normal flow water levels at Medan Ikan Bakar Station

Figure 3: Comparison of flood and normal flow water levels along Sungai Kelantan from river
mouth to the Sultan Yahya Petra Bridge during flood tidal phase.

Figure 4: Ebb (left) and flood (right) current vectors around Muara Kuala Besar.
4.0 Conclusions

4.1 Two detailed hydrography surveys of Sg Kelantan (up to the Sultan Yahya Petra Bridge)
and a 3.6 by 2.2km area of the river mouth were conducted in May-Jun 2015 and March
2016.
4.2 The shoreline evolution at 5 year intervals from 1990 to 2015 was conducted using
DSAS.
4.3 The flood event accumulated vast quantities of sediment to the Sg Kelantan river mouth
through its multiple discharge outlets and littoral sources. This was ascertained from the
shoreline evolution analyses.
4.4 Simulation for flood flow using Telemac 2D indicated increase water levels and velocity
during flood events.
4.5 Complete analyses involving quantitative determination of sandbar movements and
refined current dan discharge values are being prepared and will be included in the
upcoming publication.

References

1. J.K. Raj, Y. Ismail and W.H. Abdullah (2007), Past and present-day coastal changes between
Kuala Sungai Besar and Kuala Besar, Kelantan Darul Naim, Geological Society of Malaysia,
Bull. 53, pp.15-20.
2. DID (1994), The National River Mouth Study in Malaysia (Vol. IV), report submitted to Drainage
and Irrigation Department (DID) Malaysia by Japan International Cooperation Agency (JICA).
3. USACE (1999), Sediment Budget Analysis System (SBAS). Coastal Engineering Technical Note
IV-20. United States Army Corps of Engineers, Washington D.C.
4. EPU (1985), National Coastal Erosion Study, report submitted to the Economic Planning Unit
(EPU), Prime Minister’s Department Malaysia by Stanley Consultants.
Lampiran 3
 Lampiran 4

END OF REPORT
(maximum 5 pages of end report)

Porject Title : The potential of parameter estimation through regionalization for flood
simulations in ungauged mesoscale catchments

A. Project Information
Start Date : 01/04/2015
End Date : 31/12/2015
Extension Date : 31/03/2016
Project Status : Completed
Project Leader : Dr. Chow Ming Fai
I/C Number : 831212-06-5365
University : Universiti Tenaga Nasional (UNITEN)
Address : Jalan IKRAM-UNITEN, 43000 Kajang, Selangor.
Contact number : +60389212256
Project Members : Assoc. Prof. Ir. Dr. Marlinda Abdul Malek

B. Project Achievement
Project Progress : 100%
Research Output :Indexed Journal(_1_), Non-indexed Journal (__), Conference Proceedings
(_1_), Book Chapter (__),….
Talent : RA (__), PhD student (__), Master student (_1_)

C. Expenditure
Budget Approved : RM 50,500.00
Amount Spent : RM 49,443.00
Balance : RM 1057
% of Amount Spent : 97.9%
Summary of Research Findings
1.0 Introduction
Reliable estimates for the occurrence of flood event are important for planning measures which
reduce or even prevent flood damage. Particularly on the mesoscale catchment (drainage area of
roughly 10-1000 km2 in the present case), there is a great need for such estimates, as was e.g.
shown in the aftermath of the 2014 serious flood event in Kelantan. Usually, flood with various
recurrence intervals in catchments with long gauge records can be estimated by using extreme
value statistics. However, the results for flood estimation are noticely influenced by many
factors, such as the choices of theoretical extreme value distribution function, parameter
estimation method and ignorance of processes governing individual flood events (Klemes, 2000).
Other than that, short record or absent of local runoff data that used for calibrating model
parameters would become the main challenge on flood estimation in ungauged catchments. For
solving this problem, parameter regionalization method is widely used to estimate the parameters
for calibrating the hydrological models (Fernandez, 2000; Szolgay et al. 2003; Hundecha, Y.,
and A. Bardossy, 2004; Lamb, 2000; Wagener and Wheather, 2006). In parameter
regionalization, parameter values in ungauged catchments are normally extrapolated using
calibrated values obtained from gauged catchments by applying different regionalization
techniques. The successful of regional model calibration is mainly depends on the robustness of
model parameters obtained from gauged catchments that are used in parameter estimation. The
more robust parameters are contributing to the reduction of runoff simulation uncertainty when
transposing them to ungauged catchments. Typically, the regression models are used to relate the
model parameters with catchment attributes and climatic characteristics to determine the
regionalized parameter values (Jarboe, 1974; Karlinger, 1988; Merz, 2004). Other commonly
used methods are including global average (Merz and Bloschl, 2004; Kokkonen et al. 2003);
average based on expected similarities in watershed hydrologic responses (Schmidt et al. 2000)
and kriging (Vogel, 2006).

2.0 Methodology
Study Site
Kelantan River is divided into the Galas and Lebir Rivers near Kuala Krai, about 100 km from
the river mouth. The Galas River is formed by the junction of the Nenggiri and Pergau Rivers.
The Kelantan River system flows northward passing through such major towns as Kuala Krai,
Tanah Merah, Pasir Mas and Kota Bharu, before finally discharging into the South China Sea.
The basin has an annual rainfall of about 2,500 mm much of which occurs during the North-East
Monsoon between mid-October and mid- January. The mean flow of the Kelantan River
measured at Guillemard Bridge is 557.5 m3/s.

Integrated Flood Analysis System (IFAS)

IFAS is a deterministic model which developed by ICHARM (International Center for Water
Hazard and Risk Management) of Japan for flood modeling on river basin. The aim of IFAS is to
create flood forecasting for river basins with insufficient hydrological data. The Public Works
Research Institute Distributed Hydrological Model (PDHM) is employed in IFAS as its runoff
simulation model. The structure of PDHM is consists of surface tank model, subsurface tank
model, aquifer tank model and river tank model. The IFAS software is capable to import
satellite-based rainfall data such as GSMaP and 3B42RT for insufficient ground observation
river basin.

Regionalization methods
Regionalisation 1: Global average-based parameter regionalization
The global average parameters were determined by computing the mean of each of the
parameters from selected gauged catchments as listed in Table 1. The parameter value will be
reset to the maximum value of range if any mean value exceeding the reasonable range for the
parameter. These parameter values were then inserted into their respective model input files for
the tested river basin. The IFAS model was then run using the global average parameter values
and the corresponding stream hydrographs were obtained.

Regionalisation 2: Regression
In the Regression approach, model parameters are related directly to selected catchment
attributes. For each of the m model parameters (parami), a linear regression model containing n
attributes (attribj) is set up:
A specific regression model is built for each tuned-able parameter, containing the several
attributes which show highest correlation with the respective parameter. The resulting parameter
set for the ungauged catchment was then checked for plausibility: Values which exceed or fall
short of the range of parameter values realized in calibration are set to the respective threshold.
This avoids unreasonable parameter values in catchments with exceptional conditions.

Regionalisation 3: Nearest Neighbours

The Nearest Neighbour approach consists in finding a calibrated donor catchment which is as
similar as possible to the ungauged target basin. All tuneable model parameters are then
transferred from the donor to the target as a complete, unchanged set.

Model efficiency
The model performance is evaluated by using the Nash-Sutcliffe coefficient (NS) for peak flow,
runoff volume and wave shape. The NS coefficient (as show in equation 1) is a measure of
model efficiency that compares the simulated results to the corresponding measured results:
∑
(∑ ̅
) (1)

Where Qi is the measured value (stream discharge), Qi’ is the simulated value, Q is the average
measured value, and n is the number of data points.

3.0 Results and Discussion

Performance analysis of parameter regionalization methods
The calibrated IFAS parameter values are obtained from gauged catchments in Dungun
catchment (Terengganu), Upper Indus basin (Pakistan), Kabul river basin (Pakistan), Cagayan
river basin (Philippines) and Solo river basin (Indonesia). The global average values for each
parameter in IFAS model were calculated and summarized in Table 3. Slightly difference was
observed for surface roughness coefficient with a percentage of different of 21.4%. However,
global average values for vertical hydraulic conductivity, maximum water height, rapid
intermediate flow regulation coefficient and runoff coefficient of unconfined aquifer are
exhibiting higher magnitude of different with percentage of 200%, 60%, -40% and 104%,
respectively. The global average parameter values in Table 3 were inserted into the IFAS model
for simulating the historical flood event (December 2016 – January 2017) in Kelantan river
basin. The simulation performance was evaluated using Nash-Sutcliffe coefficient for parameters
obtained using global averaging method. The model performance statistics were then compared
to default and calibrated models for the Kelantan River basin for the period December 2006 –
Jan 2007. Generally, the model performance was not comparable to calibrated model when
global average parameters were used. For the flood event during period December 2006 –
January 2007, the NS value obtained from global averaging was 0.4 compared to 0.8 obtained
through calibration. The IFAS model performance statistics for parameters obtained using
regression based method was also shown in Table 4. The NS values range between 0.68 and 0.73
which are considered as good. In general, the model performance obtained using regression-
based parameters was comparable to that obtained through calibration. The regression-based
method for parameter estimation for IFAS simulation is consistently outperformed the default
model outputs in the tested flood event.

Table 1. Comparison of model performance statistics

Nash-Sutcliffe Coefficient
Time period Items Default Calibration Global Regression
average
Dec. 2006 – Jan 2007 Peak flow 0.45 0.80 0.40 0.70
Runoff 0.50 0.77 0.56 0.68
volume
Wave shape 0.60 0.74 0.60 0.73

4.0 Conclusion
This study had evaluated the two parameter regionalization methods: global average and
regression-based as a means of obtaining IFAS model parameters for use in ungauged
catchments.
 The model performance results obtained using regression-based parameters was comparable
to that obtained through calibration. The Nash-Sutcliffe efficiencies for predicting the peak
flow using global averaged and regression-based parameters are 0.40 and 0.70, respectively.
 The result of regression-based technique is comparable with value of 0.80 obtained through
calibration. In general, the results suggest that it is possible to estimate the IFAS parameter
using regression-based techniques.
References
Fernandez, W., R.M. Vogel., A. Sankarasubramanian. (2000). Regional calibration of a watershed model.
Hydrological Science Journal. 45: 689-707.
Hundecha, Y., and A. Bardossy. (2004). Modeling of the effect of land use changes on the runoff generation of a
river basin through parameter regionalisation of a watershed model. Journal of Hydrology. 292: 281-295.
Jacomino, V.M.F.; Fields, D.E. (1997). A critical approach to the calibration of a watershed model. J. Am.Water
Resour. Assoc. 33: 143-154.
Jarboe, J.E.; Hann, C.T. (1974). Calibrating a water yield model for small ungauged watersheds. Water Resour.
Res. 10: 256-262.
Karlinger, M.R.; Guertin, D.P.; Troutman, B.M. (1988). Regression estimates for topological—hydrograph
input. J. Water Resour. Plan. Man. 114: 446-456.
Klemes, V., 2000. Tall tales about tails of hydrological distributions. Part I and II. Journal of Hydrologic
Engineering. 5(3): 227-239.
Kokkonen, T.S.; Jakeman, A.J.; Young, P.C.; Koivusalo, H.J. (2003). Predicting daily flows in ungauged
catchments: Model regionalization from catchment descriptors at the Coweeta Laboratory, North Carolina.
Hydrol. Process. 17: 2219-2238.
Lamb, R., J. Crewett, and A. Calver. (2000). Relating hydrological model parameters and catchment properties
to estimate flood frequencies from simulated river flows. Paper presented at 7th National Hydrology
Symposium, Br. Hydrol. Serv., Newcastle-upon-Tyne, U.K.
Merz, R.; Bloschl, G. (2004). Regionalisation of catchment model parameters. J. Hydrol. 287: 95-123.
Popov, E.G. (1979). Gidrologicheskie Prognozy (Hydrological Forecasts). Gidrometeoizdat.: Leningrad, Russia.
Schmidt, J.; Hennrich, K.; Dikau, R. (2000). Scales and similarities in runoff processes with respect to
geomorphometry. Hydrol. Process. 14, 1963-1979.
Sorooshian, S.; Gupta, V.K.(1995). Model Calibration. In Computer Models of Watershed Hydrology; Singh,
V.P., Ed.; Water Resources Publications: Highlands Ranch, CO, USA.
Szolgay, J., K. Hlavcova, S. Kohnova, R. Danihlik. (2003). Regional estimation of parameters of a monthly
water balance model. Journal of Hydrology and Hydromechanics. 51: 256-273.
Vogel, R.M. (2006). Regional Calibration of Watershed Models. In Watershed Models; Singh, V.P.,Frevert,
D.K., Eds.; CRC Press: Boca Raton, FL, USA. Chapter 3, pp. 47-71.
Wagener, T., and H.S. Wheather. (2006). Parameter estimation and regionalization for continuous rainfall-
runoff models including uncertainty. Journal of Hydrology. 320: 132-154.