Climate change impacts under CMIP5 RCP scenarios on water resources of
the Kelantan River Basin, Malaysia

Mou Leong Tan, Ab Latif Ibrahim, Zulkifli Yusop, Vivien P. Chua, Ngai Weng
Chan

PII: S0169-8095(17)30083-2
DOI: doi:10.1016/j.atmosres.2017.01.008
Reference: ATMOS 3864

To appear in: Atmospheric Research

Received date: 4 August 2016

Revised date: 9 January 2017
Accepted date: 18 January 2017

Please cite this article as: Tan, Mou Leong, Ibrahim, Ab Latif, Yusop, Zulkifli, Chua,
Vivien P., Chan, Ngai Weng, Climate change impacts under CMIP5 RCP scenarios on
water resources of the Kelantan River Basin, Malaysia, Atmospheric Research (2017),
doi:10.1016/j.atmosres.2017.01.008

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Climate change impacts under CMIP5 RCP scenarios on water resources of the

Kelantan River Basin, Malaysia

T
Mou Leong Tan1,2*, Ab Latif Ibrahim2, Zulkifli Yusop3, Vivien P. Chua1, Ngai Weng Chan4

R IP
1

SC
Department of Civil and Environmental Engineering, National University of Singapore, 1

Engineering Drive 2, Singapore.

2
NU
Geoscience and Digital Earth Centre, Research Institute for Sustainable Environment,
MA
Universiti Teknologi Malaysia, Johor Bahru, Malaysia.
D

3
Centre for Environmental Sustainability and Water Security, Research Institute for
TE

Sustainable Environment, Universiti Teknologi Malaysia, Johor Bahru, Malaysia.

P
CE

4
Geography Section, School of Humanities, Universiti Sains Malaysia, Penang, Malaysia.
AC
 ACCEPTED MANUSCRIPT

Abstract

This study aims to evaluate the potential impacts of climate change on water

T
resources of the Kelantan River Basin in north-eastern Peninsular Malaysia using the Soil and

IP
Water Assessment Tool (SWAT) model. Thirty-six downscaled climate projections from five

R
General Circulation Models (GCMs) under the three Representative Concentration Pathways

SC
(RCPs) 2.6, 4.5 and 8.5 scenarios for the periods of 2015-2044 and 2045-2074 were

NU
incorporated into the calibrated SWAT model. Differences of these scenarios were calculated

by comparing to the 1975-2004 baseline period. The results showed that the SWAT model
MA
performed well in monthly streamflow simulation, with the Nash-Sutcliffe efficiency values

of 0.75 and 0.63 for calibration and validation, respectively. Based on the ensemble of five
D

GCMs, the annual rainfall and maximum temperature are projected to increase by 1.2-8.8%
TE

and 0.6-2.1oC, respectively. This corresponds to the increases in annual streamflow (14.6-
P

27.2%), evapotranspiration (0.3-2.7%), surface runoff (46.8-90.2%) and water yield (14.2-
CE

26.5%) components. The study shows an increase of monthly rainfall during the wet season,

and decrease during the dry season. Therefore, the monthly streamflow and surface runoff are
AC

likely to increase significantly in November, December and January. In addition, slight

decreases in monthly water yield are found between June and October (1.9-8.9%) during the

2015-2044 period. These findings could act as a scientific reference to develop better climate

adaptation strategies.

Keywords Climate change, CMIP5, RCP, SWAT, Kelantan, Malaysia.

 ACCEPTED MANUSCRIPT

1 Introduction

T
IP
Water resources are among the most vital natural resources, as humans depend

R
heavily on water for survival and a wide spectrum of usage (Chan, 2012). The

SC
Intergovernmental Panel on Climate Change (IPCC) reported increase in water-related risks

NU
(e.g. droughts and floods) that may be attributed to climate changes such as in local

precipitation and temperature. Roughly two thousand natural drought and flood cases were
MA
reported between 2005 and 2015 in the Emergency Disaster Database

(http://www.emdat.be/disaster_profiles/index.html). These events have affected more than

D

1.2 billion people, caused damaged around USD$ 403 billion, and resulted in about 82,000
TE

deaths. Therefore, evaluation of future water resources under climate change is important to
P

develop better water management systems and climate adaptation strategies for achieving the
CE

United Nations’ Sustainable Development Goals (SDGs).

AC

Incorporation of future climate projections from General Circulation Models (GCMs)

in simulations of a hydrological basin is regarded as one of the most reliable methods to

evaluate water resources changes (Xu et al., 1999). Generally, an ensemble of various GCMs

from different groups around the world could provide a better water resources assessment

than just a single GCM (Pierce et al., 2009). For example, Sellami et al. (2016) used an

ensemble of four GCMs to investigate changes of water balance components in the Chiba and

Thau catchments located in Tunisia and France, respectively. Elsewhere, Tan et al. (2014)

evaluated future streamflow of the Johor River Basin, Malaysia using an ensemble of six
 ACCEPTED MANUSCRIPT

GCMs selected from 18 GCMs. They concluded that the application of an ensemble of

GCMs could reduce GCM structure uncertainty in hydro-climatic studies.

T
IP
The IPCC has released four new greenhouse gas scenarios for the fifth Assessment

Report (AR5) in 2014, known as Representative Concentration Pathways (RCPs) 2.5, 4.5, 6

R
SC
and 8.5. These scenarios were named based on their possible range of radiative forcing values

(Wm-2) by the end of the 21st century compared to the pre-industrial values (van Vuuren et al.,

NU
2011). Many hydro-climatic studies have been carried out using the GCMs and RCP

scenarios as future climate scenarios (e.g. Tan et al., 2014, Zhang et al., 2016). Ouyang et al.
MA
(2015) found a decrease of future streamflow in the Huangnizhuang Catchment in China,

using an ensemble of six GCMs under three RCP 2.6, 4.5 and 8.5 scenarios. Ho et al. (2016)
D

used the climate projections under the RCP 4.5 to study streamflow changes of the Tocantins-
TE

Araguaia (Brazil), and found large declines of streamflow in the annual and dry season
P

periods scenario for the period 2071-2100.

CE

The Soil and Water Assessment Tool (SWAT) model has been widely applied in
AC

evaluating the impacts of climate change on water resources (Gassman et al., 2007;

Krysanova and White, 2015). The model is freely available and can simulate water resources

changes under different environmental conditions and management practices. However, a

direct application of the SWAT model is still a challenging issue in tropical regions due to

diversity of soil types, tree species and climate systems (Krysanova and Mike 2015). Yesuf et

al. (2016) evaluated the capability of the SWAT model in monthly streamflow simulations in

a small tropical watershed (~1.14 km2) in Ethiopia. Similarly, the SWAT model’s capability

assessment was also conducted by Fukunaga et al. (2015) in the upper Itapemirim River
 ACCEPTED MANUSCRIPT

Basin, Brazil. Both studies showed that the SWAT model can adequately simulate

streamflow in tropical regions, but more research is required to improve the model. To date,

only limited applications of the SWAT model in Malaysia were reported (e.g. Memarian et

T
IP
al., 2014; Tan et al., 2014, 2015a).

R
SC
The main objective of this study is to evaluate the future changes of water resources

in the Kelantan River Basin (KRB) under climate change impacts using an ensemble of five

NU
GCMs and the SWAT model. The notable aspects of this study are: (1) to assess the

capability of the SWAT model in the streamflow simulation in the KRB, and (2) to identify
MA
the future rainfall and temperature changes in the early-21st century (2015-2044) and mid-21st

century (2045-2074) against the baseline period (1975-2004) under the RCP 2.6, 4.5 and 8.5
D

scenarios.
P TE

There is little work done in the assessment of future climate changes and their impacts
CE

on water resources in the KRB using the Coupled Model Intercomparison Project Phase 5

(CMIP5) GCMs. Yet, the findings from this study could be used by water managers to
AC

develop a comprehensive water resources management plan in the KRB. In addition, the

projected water cycle components could also be used as input data for other applications such

as dam development planning, flood management, modelling ecological changes and aquatic

modelling.
 ACCEPTED MANUSCRIPT

2 Materials and Methods

2.1 Study Area

T
IP
The KRB (12,134km2) is one of the major basins in Malaysia (Figure 1a). It is

R
occupies more than 80% of the Kelantan state in north-eastern Peninsular Malaysia. The

SC
basin lies between latitudes 4° N to 6°N and longitudes E to E. The

NU
elevation varies from 8 to 2174m above mean sea level, with mountains in the west and

southwest regions. The Kelantan River is about 248km long and originates from the
MA
Titiwangsa and Tahan mountain ranges. In 2015, the population of the Kelantan state is about

1.7 million.
D
TE

The KRB is a tropical basin, with mean annual rainfall more than 2500mm and mean
P

annual temperature about 27.5°C (Figure 1c). The KRB is dominated by tropical rainforest,
CE

followed by rubber and oil palm plantations. Granite soil is found in the mountainous ranges

that located in the eastern and western regions. Elsewhere, peat and silt soil types are mainly
AC

found in the northern region. The mean annual streamflow at the Jambatan Guillemard station

(outlet of the KRB) is about 500 m3s-1.

In Kelantan, surface water and groundwater are the main freshwater source for

domestic use and irrigation purposes. However, this region is frequently affected by monsoon

floods during the wet season (November to January) and droughts in the dry season (March

to May). These water-related disasters are expected to be more severe in the future, as Tan et

al. (2016) reported that the maximum 1-day (Rx1d) and 5-day (Rx5d) precipitation amount
 ACCEPTED MANUSCRIPT

indices had increased significantly from 1985 to 2014. For instance, a massive flood occurred

in 2004 resulted in the evacuation of more than 10,000 people, caused losses of about

USD$ 370 million and 12 deaths.

T
R IP
2.2 SWAT Model

SC
The SWAT model is a physically-based, spatially semi-distributed and continuous

NU
hydrological model developed by the United Stated Department of Agriculture (USDA) and
MA
Texas A&M University (Arnold et al., 1998). It was developed to analyse the impacts of land

management practices on quantity and quality of water resources in un-gauged river basins.

The model can be used to study future hydro-climatic changes by modifying the climate
D
TE

parameters based on future climate projections.

P

In the SWAT model, a basin is divided into various sub-basins, which are then further
CE

divided into the hydrologic response units (HRUs) that are comprised of unique land use,
AC

slope and soil characteristics. Initial simulation of hydrological cycle occurs at the HRU level,

and excess discharge is then aggregated across the HRUs. At each HRU, the SWAT model

calculates the hydrology cycle using the water balance equation as follows:

t
SWt  SWo   ( Rday  Qsurf  Ea  wseep  Qgw )
i 1
 ACCEPTED MANUSCRIPT

Where SWt is the final soil water content (mm H2O), SWo is the initial soil water

content on day i (mm H2O), t is the time (days), Rday is the amount of precipitation on day i

T
(mm H2O), Qsurf is the amount of surface runoff on day i (mm H2O), Ea is the amount of

IP
evapotranspiration on day i (mm H2O), wseep is the amount of water entering the vadose zone

R
from the soil profile on day i (mm H2O), and Qgw is the amount of return flow on day i (mm

SC
H2O). More theoretical information of the SWAT model is described in Neitsch et al. (2011).

NU
MA
2.3 SWAT Model Input Data

Figure 2 shows a schematic diagram of the research framework. The main inputs for
D
TE

the SWAT model are digital elevation model (DEM), land use, soil, rainfall and temperature

data. The Shuttle Radar Topography Mission (SRTM) DEM with 90m resolution, which had
P

a better performance in SWAT modelling in Malaysia was used (Tan et al., 2015b). The
CE

DEM data is downloaded from http://srtm.csi.cgiar.org/. The land use map (1990) and soil
AC

map (2002) were obtained from the Ministry of Agriculture and Agro-based Industry

Malaysia (MOA). Soil properties information (e.g. soil texture and depth) was extracted from

a soil report prepared by Paramananthan (2000).

Daily climate data from 1974 to 2004, including rainfall, maximum and minimum

temperature were collected from twenty-eight climate stations managed by the Malaysian

Meteorological Department (MMD) and the Department of Irrigation and Drainage Malaysia

(DID). Only two climate stations located in the downstream region (Figure 1) contain long-

term maximum and minimum temperature data. Missing climate data were filled with climate
 ACCEPTED MANUSCRIPT

data of the nearest station. Monthly observed streamflow data at the Jambatan Guillemard

(outlet of the KRB) was used to calibrate and validate the SWAT model. The river network

data was digitized from the topography map prepared by the Department of Survey and

T
IP
Mapping Malaysia (JUPEM), in order to improve river network extraction and basin

delineation in the low land region.

R
SC
2.4 Model Setup, Calibration and Validation

NU
MA
In this study, the ArcSWAT 2012, an interface of the SWAT model (version 2012) in

the ArcGIS 10.2 system, was used to develop the SWAT model for the KRB. The ArcSWAT

2012 can manage and handle multiple spatial data sets easily. There are six main steps in
D
TE

SWAT modelling: (1) basin delineation and river network extraction; (2) HRU definition; (3)

climate station formation; (4) parameter sensitivity analysis; (5) calibration; and (6)
P

validation.
CE
AC

The basin is delineated into 36 sub-basins, with the digitized river network merged

into the DEM using the “burn-in” method. The sub-basins were further divided into 177

HRUs. The HRUs definition threshold values of the land use, soil and slope were set as 20%,

10% and 20%, respectively, which are recommended by Winchell et al. (2013). These

percentage values could ignore minor land uses, soils and slopes in each sub-basin for

controlling the number of HRUs (Tan et al., 2015b). This could reduce model errors in HRUs

aggregation. Moreover, large threshold values (e.g. 20%) are generally less sensitive in the

water quantity simulations (e.g. streamflow and water yield) compared to the water quality

simulations (Her et al., 2015). The first year (1974) was considered as a warm up period to
 ACCEPTED MANUSCRIPT

initiate the hydrological parameters. In this study, the Hargreaves method was selected for the

evapotranspiration (ET) computation, as this method requires only precipitation and

temperature data. Moreover, surface runoff (SURQ) and streamflow (SF) routing were

T
IP
measured using the curve number and variable storage methods, respectively.

R
SC
Model parameter sensitivity analysis, calibration and validation were performed with

the SWAT-CUP tool (http://swat.tamu.edu/software/swat-cup/). The global sensitivity

NU
analysis method was applied to evaluate the most important parameters for monthly

streamflow simulations in the KRB. Then, the SWAT model was calibrated using the
MA
sequential uncertainty fitting algorithm (SUFI-2), with 500 different parameters combinations

(one iteration) for the period 1975-1989. The SUFI-2 was selected due to its capability in
D

analysing many parameters in the model runs. In an iteration, the SUFI-2 measures the
TE

goodness of fit and the 95% prediction uncertainty (95PPU) between simulated and observed
P

streamflow (Abbaspour et al., 2015). In addition, new parameters ranges were produced
CE

which can be used in the next iteration, to re-calibrate the model until the best parameters

ranges were obtained. These best parameters ranges were then applied to validate the monthly
AC

streamflow from 1990 to 1999.

The SUFI-2 allows application of numerous objective functions to evaluate the

quality of the SWAT simulations. Five objective functions including the Nash-Sutcliffe

efficiency (NSE), coefficient of determination (R2), adjusted R2 (bR2), sum of the squares of

residuals (SSQR) and percentage bias (PB) (Abbaspour, 2015) were used to calibrate the

SWAT model from 1975 to 1989 with the same parameters ranges, in order to identify the

most suitable objective function in this study. The mean square error (MSE) was used to
 ACCEPTED MANUSCRIPT

identify the optimal simulations, where the value closer to zero is better. Finally, the NSE

with the best performance, was then selected as the optimal objective function in the SWAT

calibration and validation. The NSE is one of the most widely applied objective functions in

T
hydrological modelling (Willmott et al., 2015). The NSE values range from -∞ to , with a

IP
negative NSE value being regarded as unacceptable performance. Moreover, collinearity

R
between simulated and observed streamflow was analysed using the R2, and its value varies

SC
from -1 to 1 (ideal). The performance of the SWAT model can be characterised as good and

NU
satisfactory with NSE values from 0.65-0.75 and 0.5-0.65, respectively (Moriasi et al. 2007).
MA
2.5 Future Climate Scenarios
D
TE

The climate projections from five CMIP5 GCMs that showed good performance in

historical rainfall simulation in Malaysia region (Siew et al., 2014; Tan et al., 2014) were
P

used in this study. These GCMs were produced by various international institutions including
CE

(1) the National Center for Atmospheric Research, United States (CESM1-CAM5); (2)
AC

Centre National de Recherches Meteorologiques, France (CNRM-CM5); (3) Atmosphere and

Ocean Research Institute (The University of Tokyo), Japan (MIROC-ESM); (4)

Meteorological Research Institute, Japan (MRI-CGCM3); and (5) Norwegian Climate Centre,

Norway (NorESM1-M).

Five GCM outputs (daily rainfall, minimum and maximum temperature) under the

RCP 2.6, 4.5 and 8.5 scenarios were utilized to project the early-21st (2015-2044) and mid-

21st (2045-2074) centuries climate scenarios of the basin. The RCP 2.6 scenario indicates a

very low forcing level, with a peak of 3.1 Wm-2 in the mid-21st century, subsequently
 ACCEPTED MANUSCRIPT

declining to 2.6 Wm-2 by the end of the 21st century (van Vuuren et al., 2011). Greenhouse

gas emissions should be reduced significantly to match the RCP 2.6 scenario. While, forcing

level of the RCP 4.5 scenario is stabilized with 4.5 Wm-2 before 2100 by applying a range of

T
IP
strategies and technologies to reduce the emission of greenhouse gases. In contrast, the RCP

8.5 scenario is a very high greenhouse gases emission scenario, with a rising radiative forcing

R
pathway leading to 8.5 Wm-2 by 2100.

SC
NU
Generally, the spatial resolution of raw GCMs is too coarse (~200km) to study

regional or local hydro-climatic processes, therefore downscaling must be conducted before

MA
applying future projections of GCMs into the calibrated SWAT. Two main approaches to

downscale GCMs to finer scale are (1) dynamic downscaling - nesting of a regional climate
D

model within a GCM; and (2) statistical downscaling - developing empirical relationship of
TE

interested variables between GCMs and observed data, and applying these relationships to
P

downscale the GCM. The popular change factor approach (Ouyang et al., 2015; Basheer et al.,
CE

2016) under statistical downscaling method was used in this study.

AC

Thirty-seven climate projections under various GCMs, RCP scenarios and study

periods are listed in Table 1. Climate change impacts on water cycle components (streamflow,

evapotranspiration, surface runoff and water yield) were assessed based on annual and

monthly changes between each climate scenario (ID 1-36) compared to the baseline scenarios

(ID 37). Besides that, climate projections of a mean ensemble of five GCMs are known as

“Ensemble_ ”. Many researchers have applied the climate projection of an ensemble of three

to six GCMs in their hydro-climatic studies, to obtain reliable future projections (e.g. Tan et

al., 2014; Ouyang et al., 2015; Sellami et al., 2016; Zhang et al., 2016).
 ACCEPTED MANUSCRIPT

3 Results

T
3.1 Parameter Sensitivity Analysis

R IP
Table 2 ranks the parameters for the SWAT model in the KRB in terms of sensitivity.

SC
The results show that the most sensitive parameters were groundwater ‘revap’ coefficient

NU
(GW_REVAP), followed by the channel effective hydraulic conductivity (CH_K2), baseflow

alpha factor (ALPHA_BF), initial SCS CN II value (CN2), groundwater delay

MA
(GW_DELAY), soil evaporation compensation factor (ESCO), threshold water depth in the

shallow aquifer for flow (GWQMN), manning’s value for main channel (CH_N2), available
D

water capacity (SOL_AWC), surface runoff lag time (SURLAG), threshold depth of water in
TE

the shallow aquifer for “revap” to occur (REVAPMN) and deep aquifer percolation faction
P

(RCHRG_DP). Generally, the most sensitivity parameters of the basin are groundwater-
CE

related parameters (GW_REVAP, ALPHA_BF and GW_DELAY) and surface runoff

parameter (CN2), which are also found to be sensitive in other studies in Malaysia
AC

(Memarian et al., 2014; Tan et al., 2014). This could be attributed to the intensive

groundwater-stream water interactions in the basin due to the shallow groundwater.

3.2 Calibration and Validation

Figure 3 presents the calibration and validation results using monthly streamflow at

the Jambatan Guillemard station for the periods 1975-1989 and 1990-1999, respectively. The

initial and final parameters ranges as well as a set of the best parameters are listed in Table 2.
 ACCEPTED MANUSCRIPT

Statistical analysis of the calibration (NSE=0.75; R2=0.8) and validation (NSE=0.63; R2=0.64)

indicates the model performed well in monthly streamflow simulation. The SWAT model

underestimated the peak flows during the periods of 1975-1985 and 1991-1997, but

T
IP
overestimated the peak flows for the 1988-1990 and 1998-1999 periods . Similarly, the

SWAT model was also unable to match the peak flows in southern Peninsular Malaysia (Tan

R
SC
et al., 2014). This could be due to sparse distribution of the rainfall stations.

NU
Incompleteness of the observed hydro-climatic data could also influence the model

simulation results. For example, only a climate station located within the basin contains
MA
temperature data, but it was excluded from this study due to an inappropriate operation period.

Although there are several streamflow stations in the basin, however most of them had
D

missing data and suffer from a short operation period. The model also appears to
TE

underestimate the evapotranspiration processes of the basin, which is commonly observed in

P

applications of the SWAT model in tropical regions (Nyeko, 2014; Krysanova and White,
CE

2015). To circumvent these problems, the GW_REVAP parameter that controls the water

movement from shallow aquifer into unsaturated zone was modified by increasing its default
AC

value, to allow more water for evapotranspiration process in the basin. Besides that, the

SWAT model also overestimated the baseflow and underestimated the peak flow. As

recommended by Abbaspour et al. (2015), the CN2 parameter was increased from the

original value, to increase the surface runoff and reduce the infiltration of the basin.
 ACCEPTED MANUSCRIPT

3.3 Future Precipitation & Temperature Changes

In this section, the Haiwan Machang and Gua Musang stations were used to represent

T
IP
climate changes of the lower and upper basins, respectively (Figure 1a). Future changes in

annual rainfall and maximum temperature at the Haiwan Machang and Gua Musang stations

R
SC
(Figure 1a) are shown in Figure 4. For the Ensemble_5 under the three RCP scenarios, the

annual rainfall and maximum temperature could increase from 1.2 to 3.6% (2.5 to 8.7%) and

NU
0.6 to 1oC (0.8 to 2.1oC) for the 2015-2044 (2045-2074) period, respectively. The greatest

annual rainfall (8.7%) and maximum temperature (2.1oC) changes are observed at the Gua
MA
Musang station during the 2045-2075 period under the RCP 8.5 scenario, showing that

climate changes in the upper basin are greater compared to the lower basin during the mid-
D

21st century.
P TE

Figure 5 shows the monthly changes in rainfall and maximum temperature of the
CE

KRB from the Ensemble_5 for the periods of 2015-2044 and 2045-2074 under the three RCP

scenarios. We observed that the monthly rainfall increases dramatically in January and
AC

December, ranging from 20.2 to 47% and 40.4 to 71.3%, respectively. Moreover, monthly

rainfall is projected to decrease in April (-2.7 to -22%), June (2.9 to -7.5%) and July (3.2 to -

9.2%). The results indicate monthly rainfall may increase during the wet season, and decrease

during the dry season. These findings are consistent with the historical annual and monthly

rainfall trend analysis of the basin reported by Adnan and Atkinson (2011).
 ACCEPTED MANUSCRIPT

Under the three RCP scenarios, the average maximum temperature for the early-21st

and mid-21st century have a rising trend in all months, and varies from 0.1 to 2.2 oC and 0.3

to 2.3 oC at the Haiwan Machang and Gua Musang stations, respectively. The monthly

T
IP
maximum temperature changes are higher in April, May and June, which could lead to water-

scarcity problem in the KRB. In addition, high temperature and low rainfall during these

R
SC
months might reduce productivity of paddy, oil palm and rubber plantations within and

surrounding of the basin (Paterson et al., 2015).

NU
MA
3.4 Impact of Climate Change on Streamflow

Table 3 presents the results of the SWAT-simulated annual streamflow changes under
D
TE

thirty-six climate scenarios as listed in Table 1. The annual streamflow at the Jambatan

Guillemard station is projected to increase by 14.6 to 27.2% (91.3 to 170.3 m3s-1) for the
P

2015-2044 and 2045-2075 periods under the three RCP scenarios, which could be attributed
CE

by an increase in annual rainfall and temperature over the basin. Our results show that the
AC

increases under the RCP8.5 scenario is larger compared to the RCP 2.6 and 4.5 scenarios.

The future monthly streamflow changes of the Ensemble_5 at the Jambatan

Guillemard station for the 2015-2044 and 2045-2074 periods under the three RCP scenarios

are shown in Figure 6(a). We observe that the monthly streamflow tends to increase largely

from November to February, with the largest increase in December by 115.1% under the RCP

8.5 scenario during the 2015-2044 period. Similarly, Shaaban et al. (2011) reported increment

of average and maximum monthly streamflow at the same station for the future periods of

2025-2034 and 2041-2050. The monthly streamflow is projected to decrease from June to
 ACCEPTED MANUSCRIPT

October during the 2015-2044 period, varying from -1.1 to -9.1%. The results imply that the

KRB will experience less freshwater supply during the dry season due to climate changes.

T
IP
3.5 Impact of Climate Change on Evapotranspiration

R
SC
The average annual evapotranspiration over the entire basin are expected to increase

by 0.3-2.7 % (0.7-6.5 mm) for the 2015-2044 and 2045-2074 periods, as shown in Table 3.

NU
The highest average annual evapotranspiration (2.7%) was found during the 2045-2074
MA
period for the RCP 8.5 scenario. Figure 6(b) presents the monthly evapotranspiration will

increase during the wet season (September to February) by 0.6-3% and 1.4-5.4% in the 2015-

2044 and 2045-2074 periods, respectively. In contrast, decreases of monthly precipitation in

D
TE

April will cause reduction of monthly evapotranspiration from -1.1 to -5%. Overall,

evapotranspiration changes are less significant compared to other water cycle components,
P

which is consistent with Al-Mukhtar et al. (2014).

CE
AC

Similarly, Mishra and Lilhare (2016) reported that increase in rainfall and temperature

will lead to an increase in evapotranspiration and vice-versa. Therefore, rainfall also plays an

important role that causes a significant change in evapotranspiration. As future land use

changes are not considered in this study, we expect that the evapotranspiration changes of the

basin will be amplified, if unbridled expansion of rubber and oil palm plantations continues

to occur in the future. This is because evapotranspiration rates are directly influenced by

deforestation (Bosch and Hewlett 1982; Hirano et al. 2015). For example, Giambelluca et al.

(2016) demonstrated that higher evapotranspiration rates occur in rubber plantations in

Thailand and Cambodia compared to tropical forest regions.

 ACCEPTED MANUSCRIPT

3.6 Impact of Climate Change on Surface Runoff

T
IP
The response of the basin to Ensemble_5 climate scenarios in terms of both annual

R
and monthly surface runoff is shown in Table 3 and Figure 6(c), respectively. Our results

SC
show that the increase in the average annual rainfall and temperature causes a significant

increase of annual surface runoff ranging from 41.5 to 90.2% (176.6 to 384.2 mm) in the

NU
2015-2044 and 2045-2074 periods. The greatest surface runoff change (90.2%) was observed
MA
under the RCP 8.5 scenario during the period 2015-2044. Human activities such as

deforestation, urbanization, agriculture expansion and industrialization would increase

surface runoff. Therefore, the surface runoff changes of the basin will be higher in the near
D
TE

future.
P

On the monthly scale, the ensemble mean surface runoff of the basin is projected to
CE

increase dramatically in November (23.2 to 191.4%), December (89.1 to 285.3%) and

AC

January (48 to 98.2%) under the three RCP scenarios. As shown in Figure 6(c), the increment

of surface runoff in December and January is larger during the 2015-2044 period than the

2045-2074 period. Larger increases in surface runoff could be attributed to changes of rainfall

during the northeast monsoon season. Suhaila et al. (2010) reported that the magnitude and

frequency of total rainfall and extreme rainfall events increased significantly during northeast

monsoon season in Peninsular Malaysia. Besides that, Adnan and Atkinson (2011) also found

an increasing rainfall trend during the northeast monsoon over Kelantan, and a decreasing

trend in the southwest monsoon. Surface runoff is likely to rise further, if there are more

extreme rainfall events (Trenberth, 2011).

 ACCEPTED MANUSCRIPT

3.7 Impact of Climate Change on Water Yield

T
IP
Water yield is also considered in this study, as it can be represented water resources

R
availability of a basin (Sun et al., 2006). Table 3 shows annual water yield of the basin is

SC
expected to rise across the three RCP scenarios, ranging from 14.2 to 26.5% (238.7 to 444.4

mm), indicating water resources are sufficient to support population and environment of the

NU
KRB until the mid-21st century. However, an effective and appropriate water resources
MA
management system should be put in place in the basin, in order to prevent decline of water

resources in term of quantity and quality, especially during the dry season.
D
TE

Figure 6(d) displays the monthly water yield changes over the basin under the three

RCP scenarios. During the 2015-2044 period, the large increases in water yield can be found
P

in December (83.5 to 118.5%) and January (30.4 to 36.2%), while, the decreases were
CE

recorded in August (-5.4 to -8%), September (-3.9 to -7.9%) and October (-6.5 to -8.9%). By
AC

the 2045-2074 period, monthly water yield is projected to increase across all the months by

4.2 to 42.5% under the RCP 4.5 scenario, showing that application of greenhouse gases

reduction strategies and technologies could help to increase water yield of the basin during

the dry season. These findings show water yield of the KRB will be increased significantly

during the wet season and slightly decrease during the dry season. Therefore, more efficient

irrigation systems for oil palm and rubber plantations as well as paddy crops should be

developed and implemented to minimize losses during both drought and flood seasons.
 ACCEPTED MANUSCRIPT

4 Discussion

Monsoon floods are the main disasters in Kelantan, frequently occurring during the

T
northeast monsoon season (November to January). For example, one of the worst floods in

IP
the recent decades occurred between 14th December 2014 and 3nd January 2015, causing the

R
deaths of more than 21 people and devastated the states Kelantan, Terengganu, Pahang and

SC
Perak, with damage to infrastructure estimated at more than USD$ 560 million. The basin is

NU
also affected by severe drought events, especially during the El Niño period. For example, the

2014 drought affected more than 8,000 paddy farmers in Kelantan, and caused about
MA
USD$ 22 million crop losses. These disasters caused massive impacts on agriculture,

aquaculture, freshwater supply and industrial sectors in the KRB. As such, a systematic
D

assessment of future hydro-climatic study is essential for this basin.

P TE

The SWAT model performed satisfactorily in simulating streamflow in the KRB, with
CE

validation outputs similar as a study conducted in southern Peninsular Malaysia (Tan et al.,

2014). In this study, the accuracy of streamflow simulations in other sub-basins was not fully
AC

studied, as only observed streamflow from a stream gauge was considered for calibration and

validation purposes due to incompleteness of streamflow data. In addition, the SWAT model

tends to overestimate low flow conditions of the Kelantan River. This could be due to the

plant growth and land use models within the SWAT model were developed for temperate

regions, and these should be improved for application of SWAT model in tropical regions

(Strauch and Volk, 2013).

 ACCEPTED MANUSCRIPT

SWAT calibration outputs are largely influenced by different objective functions

(Willmott et al., 2015), as well as on different hydrological components simulation. There

are, however, some drawbacks with these objective functions. For example, the NSE and R2

T
IP
are more sensitive on high flow than low flow (Legates and McCabe, 1999), resulting in the

model calibration producing better performance in the wet period compared to the dry period

R
SC
(Zhang et al., 2015). Despite this, assessment of various objective functions on model

calibration has received less attention, particularly the low flow component (Pushpalatha et

NU
al., 2012). Hence, selection of an appropriate objective function remains a challenging task to

modellers.
MA

The relationships between atmosphere and water cycle in the basin are unable to be
D

fully captured, as there are large uncertainties in GCMs structure and RCP emission scenarios.
TE

These uncertainties in future precipitation and temperature values could be propagated into
P

streamflow modelling, which influence future water cycle simulations. For example, different
CE

climate scenarios as projected by different GCMs (ID 1-30) lead to varying annual

streamflow changes (-4.9 to 111.2%), indicating high uncertainty may occur in the CMIP5
AC

GCM projections which is similar to findings of other studies (Kingston et al., 2011; Tan et

al., 2014). In addition, projection of future precipitation is more complex and difficult than

temperature due to their high spatial and temporal variability.

Wilby et al. (2012) found the change factor downscaling approach to be useful for

application of multiple GCMs in long-term hydro-climatic assessment due to its simplicity

and low computation cost. However, this approach could also lead to uncertainty in hydro-

climatic studies because it does not address local topography, micro-climate system, future
 ACCEPTED MANUSCRIPT

precipitation occurrence and distribution. In addition, it is less accurate in predicting short-

term extreme precipitation events (Ouyang et al., 2015). Therefore, impacts of different

downscaling approaches on future hydro-climatic studies should be further explored, to better

T
IP
quantify the future environment changes.

R
SC
5 Conclusion

NU
To develop better global, regional and local climate adaptation strategies, a robust

assessment of future hydro-climatic changes in a specific region is necessary. This study

MA
provides a comprehensive framework and up-to-date assessment of climate change impacts

on water resources of the Kelantan River Basin (KRB), Malaysia. Future climate changes of
D

the basin were projected from thirty-seven climate scenarios from five GCMs under the RCP
TE

2.6, 4.5 and 8.5 scenarios for the periods of 2015-2044 and 2045-2074. These climate
P

scenarios were then used as inputs into a calibrated SWAT model, and the relative changes of
CE

future water cycle components were compared to the baseline period (1975-2004) condition.
AC

The SWAT model performed well in monthly streamflow simulation in the KRB. The

GW_REVAP parameter is found to be the most sensitive parameter during the calibration

process, indicating that the SWAT model might be underestimating evapotranspiration

process in tropical regions. Moreover, SWAT model tends to overestimate low flows and

underestimate peak flows in this tropical basin. Therefore, more studies on SWAT model

modification for application in tropical regions should be conducted.

 ACCEPTED MANUSCRIPT

The future annual rainfall and temperature are projected to increase for all time

periods under the three RCP scenarios, which results in similar trends for annual streamflow,

evapotranspiration, surface runoff and water yield. The basin is most sensitive towards the

T
IP
change in annual surface runoff (up to 90.2%), followed by streamflow, water yield and

evapotranspiration. On the monthly scale, significant increases of surface runoff and

R
SC
streamflow were observed in November, December and January. Decreases of monthly water

yield between June and October would lead to water scarcity problems in the future. Hence,

NU
climate adaptation strategies such as reforestation, rainwater harvesting, water reuse and

recycling, and employment of effective irrigation should be introduced, to reduce climate

MA
change impacts on water resources in KRB.
D

Future studies should be evaluated with other dynamically downscaling climate

TE

scenarios such as the Coordinated Regional Climate Downscaling Experiment (CORDEX)

P

Southeast Asia. In addition, extreme events assessment of future hydro-climatic conditions is

CE

essential in future, to develop better flood and drought adaptation strategies. Lastly,

application of the SWAT model is still relatively limited in Malaysia, and more hydro-
AC

climatic studies should be performed in other regions.

 ACCEPTED MANUSCRIPT

Acknowledgment

This research was supported by the Ministry of Higher Education Malaysia and Universiti

T
Teknologi Malaysia under the Transdisciplinary Research Grant Scheme

IP
(R.J130000.7809.4L835). The authors would like to thank to the MMD, DID and DOA for

R
supplying hydro-climatic data, land use and soil maps for this study. The authors also wish to

SC
thank the CMIP5 project for providing future climate data. The authors acknowledge the

NU
2016 Summer Institute for Disaster and Risk Research fellowship for the opportunity to work

for two-weeks in the Beijing Normal University, China. Special thanks to the reviewers for
MA
their helpful comments and suggestions.
D
P TE
CE
AC
 ACCEPTED MANUSCRIPT

References

Abbaspour, K.C., Rouholahnejad, E., Vaghefi, S., Srinivasan, R., Yang, H., Kløve, B., 2015.

T
A continental-scale hydrology and water quality model for Europe: Calibration and

IP
uncertainty of a high-resolution large-scale SWAT model. J. Hydrol. 524, 733-752.

R
SC
Abbaspour, K.C., 2015. SWAT Calibration and Uncertainty Programs—A User Manual.

NU
Swiss Federal Institute of Aquatic Science and Technology (Eawag).
MA

Adnan, N.A., Atkinson, P.M., 2011. Exploring the impact of climate and land use changes on
D

streamflow trends in a monsoon catchment. Int. J. Climatol. 31(6), 815-831.

P TE
CE

Al-Mukhtar, M., Dunger, V., Merkel, B., 2014. Assessing the Impacts of Climate Change on

Hydrology of the Upper Reach of the Spree River: Germany. Water Resour. Manag.
AC

28(10), 2731-2749.

Arnold, J.G., Srinivasan, R., Muttiah, R.S., Williams, J.R., 1998. Large area hydrologic

modeling and assessment - Part 1: Model development. J. Am. Water Resour. Assoc.

34(1), 73-89.
 ACCEPTED MANUSCRIPT

Basheer, A.K., Lu, H., Omer, A., Ali, A.B., Abdelgader, A.M.S., 2016. Impacts of climate

change under CMIP5 RCP scenarios on the streamflow in the Dinder River and

ecosystem habitats in Dinder National Park, Sudan. Hydrol. Earth Syst. Sci. 20(4),

T
IP
1331-1353.

R
SC
Bosch, J.M., Hewlett, J.D., 1982. A review of catchment experiments to determine the effect

NU
of vegetation changes on water yield and evapotranspiration. J. Hydrol. 55(1), 3-23.
MA
Chan, N.W., 2012. Managing urban rivers and water quality in Malaysia for sustainable water

resources. Int. J. Water Resour. D. 28 (2), 343-354.

D
TE

Fukunaga, D.C., Cecílio, R.A., Zanetti, S.S., Oliveira, L.T., Caiado, M.A.C., 2015.
P
CE

Application of the SWAT hydrologic model to a tropical watershed at Brazil. Catena

125, 206-213.
AC

Gassman, P.W., Reyes, M.R., Green, C.H., Arnold, J.G., 2007. The soil and water assessment

tool: Historical development, applications, and future research directions. T. ASABE

50 (4), 1211-1250.
 ACCEPTED MANUSCRIPT

Giambelluca, T.W., Mudd, R.G., Liu, W., Ziegler, A.D., Kobayashi, N., Kumagai, T.,

Miyazawa, Y., Lim, T.K., Huang, M.Y., Fox, J., Yin, S., Mak, S.V., Kasemsap, P.,

2016. Evapotranspiration of rubber (Hevea brasiliensis) cultivated at two plantation

T
IP
sites in Southeast Asia. Water Resour. Res. 52(2), 660-679.

R
SC
Her, Y., Frankenberger, J., Chaubey, I., Srinivasan, R. 2015. Threshold effects in HRU

NU
definition of the Soil and Water Assessment Tool. T. ASABE 58(2), 367-378.
MA
Hirano, T., Kusin, K., Limin, S., Osaki, M., 2015. Evapotranspiration of tropical peat swamp

forests. Glob. Change Biol. 21(5), 1914-1927.

D
TE

Ho, J.T., Thompson, J.R., Brierley, C., 2016. Projections of hydrology in the Tocantins-
P
CE

Araguaia Basin, Brazil: uncertainty assessment using the CMIP5 ensemble.

Hydrolog. Sci. J. 61(3), 551-567.

AC

Kingston, D.G., Thompson, J.R., Kite. G., 2011. Uncertainty in climate change projections of

discharge for the Mekong River Basin. Hydrol. Earth Syst. Sci. 15(5), 1459-1471.

Krysanova, V., White, M., 2015. Advances in water resources assessment with SWAT—an

overview. Hydrolog. Sci. J. 60 (5), 771-783.

 ACCEPTED MANUSCRIPT

Legates, D.R., McCabe, G.J., 1999, Evaluating the use of "goodness-of-fit" measures in

hydrologic and hydroclimatic model validation. Water Resour. Res. 35(1), 233-241.

T
IP
Memarian, H., Balasundram, S.K., Abbaspour, K.C., Talib, J.B., Teh, B.S.C., Sood, A.M.,

R
2014. SWAT-based hydrological modelling of tropical land use scenarios. Hydrolog.

SC
Sci. J. 59(10), 1808-1829.

NU
Mishra, V., Lilhare, R., 2016. Hydrologic sensitivity of Indian sub-continental river basins to
MA
climate change. Global Planet. Change 139, 78-96.
D
TE

Moriasi, D.N., Arnold, J.G., Van Liew, M.W., Binger, R.L., Harmel, R.D., Veith, T., 2007.

Model evaluation guidelines for systematic quantification of accuracy in watershed

P
CE

simulations. T. ASABE 50(3), 885-900.

AC

Neitsch, S.L., Arnold, J.G., Kiniry, J.R., Grassland, J.R.W., 2011. Soil and Water Assessment

Tool Theoretical Documentation Version 2009. Texas Water Resources Institute

Technical Report No. 406. Texas A&M University, Texas.

Nyeko, M., 2014. Hydrologic modelling of data scarce basin with SWAT model: capabilities

and limitations. Water Resour. Manag. 29(1), 81-94.

 ACCEPTED MANUSCRIPT

Ouyang, F., Zhu, Y.H., Fu, G.B., Lu, H.S., Zhang, A.J., Yu, Z.B., Chen, X., 2015. Impacts of

climate change under CMIP5 RCP scenarios on streamflow in the Huangnizhuang

catchment. Stoch. Environ. Res. Risk Assess. 29 (7), 1781-1795.

T
R IP
Paramananthan, S., 2000. Soil of Malaysia: Their Characteristic & Identification, first ed. Vol

SC
1. S. Paramananthan, Kuala Lumpur.

NU
Paterson, R.R.M., Kumar, L., Taylor, S., Lima, N., 2015. Future climate effects on suitability
MA
for growth of oil palms in Malaysia and Indonesia. Sci. Rep. 5, 14457.
D
TE

Pierce, D.W., Barnett, T.P., Santer, B.D., Gleckler, P.J., 2009. Selecting global climate

models for regional climate change studies. P. Natl. Acad. Sci. 106(21), 8441-8446.
P
CE
AC

Pushpalatha, R., Perrin, C., Moine, N.L., Andréassian, V., 2012. A review of efficiency

criteria suitable for evaluating low-flow simulations. J. Hydrol. 420–421, 171-182.

Sellami, H., Benabdallah, S., La Jeunesse, I., Vanclooster, M., 2016. Quantifying

hydrological responses of small Mediterranean catchments under climate change

projections. Sci. Total Environ. 543, 924-936.

 ACCEPTED MANUSCRIPT

Shaaban, A.J., Amin, M.Z.M., Chen, Z.Q., Ohara, N., 2011. Regional modeling of climate

change impact on Peninsular Malaysia water resources. J. Hydrol. Eng. 16(12),

1040-1049.

T
R IP
Siew, J.H., Tangang, F.T., Juneng, L., 2014. Evaluation of CMIP5 coupled atmosphere–

SC
ocean general circulation models and projection of the Southeast Asian winter

NU
monsoon in the 21st century. Int. J. Climatol. 34(9), 2872-2884.
MA
Strauch, M., Volk, M., 2013. SWAT plant growth modification for improved modeling of

perennial vegetation in the tropics. Ecol. Model. 269, 98-112.

D
TE

Suhaila, J., Deni, S.M., Zin, W.Z.W., Jemain, A.A., 2010 Spatial patterns and trends of daily
P
CE

rainfall regime in Peninsular Malaysia during the southwest and northeast monsoons:

1975-2004. Meteorol. Atmos. Phys. 110 (1-2), 1-18.

AC

Sun, G., Zhou, G.Y., Zhang, Z.Q., Wei, X.H., McNulty, S.G., Vose, J.M., 2006. Potential

water yield reduction due to forestation across China. J. Hydrol. 328(3-4), 548-558.

Tan, M.L., Ficklin, D.L., Dixon, B., Ibrahim, A.L., Yusop, Z., Chaplot, V. 2015b. Impacts of

DEM resolution, source, and resampling technique on SWAT-simulated streamflow.

Appl. Geogr. 63, 357-368.

 ACCEPTED MANUSCRIPT

Tan, M.L., Ficklin, D.L., Ibrahim, A.L., Yusop, Z., 2014. Impacts and uncertainties of

climate change on streamflow of the Johor River Basin, Malaysia using a CMIP5

General Circulation Model ensemble. J. Water Clim. Change 5(4), 676–695.

T
R IP
Tan, M.L., Ibrahim, A.L., Cracknell, A.P., Yusop, Z., 2016. Changes in precipitation

SC
extremes over the Kelantan River Basin, Malaysia, Int. J. Climatol. In press, doi:

NU
10.1002/joc.4952. MA
Tan, M.L., Ibrahim, A.L., Yusop, Z., Duan, Z., Ling, L., 2015a. Impacts of land-use and

climate variability on hydrological components in the Johor River basin, Malaysia.

D

Hydrolog. Sci. J. 60(5), 873-889.

P TE
CE

Trenberth, K., 2011. Changes in precipitation with climate change. Clim. Res. 47(1-2), 123-

138.
AC

van Vuuren, D.P., Edmonds, J., Kainuma, M., Riahi, K., Thomson, A., Hibbard, K., Hurtt,

G.C., Kram, T., Krey, V., Lamarque, J.F., Masui, T., Meinshausen, M., Nakicenovic,

N., Smith, S.J., Rose, S.K. 2011. The representative concentration pathways: an

overview. Clim. Change 109(1-2), 5-31.

 ACCEPTED MANUSCRIPT

Wilby, R.L., Dawson, C.W., Barrow, E.M., 2002. SDSM - a decision support tool for the

assessment of regional climate change impacts. Environ. Modell. Softw. 17(2), 147-

159.

T
R IP
Willmott, C.J., Robeson, S.M., Matsuura, K., Ficklin, D.L., 2015. Assessment of three

SC
dimensionless measures of model performance. Environ. Modell. Softw. 73, 167-

NU
174. MA
Winchell, M., Srinivasan, R., Luzio, M.D., Arnold, J. 2013. ARCSWAT interface for

SWAT2012: user's guide. Blackland Research and Extension Center Texas Agrilife
D

Research, 720 East Blackland Road, Temple, Texas 76502.

P TE
CE

Xu, C.Y., 1999. Climate change and hydrologic models: A review of existing gaps and recent

research developments. Water Resour. Manag. 13(5), 369-382.

AC

Yesuf, H.M., Melesse, A.M., Zeleke, G., Alamirew, T., 2016. Streamflow prediction

uncertainty analysis and verification of SWAT model in a tropical watershed.

Environ. Earth Sci. 75(9), 1-16.

Zhang, D., Chen, X., Yao, H., Lin, B., 2015. Improved calibration scheme of swat by

separating wet and dry seasons. Ecol. Model. 301, 54-61.

 ACCEPTED MANUSCRIPT

Zhang, Y., You, Q., Chen, C., Ge, J., 2016. Impacts of climate change on streamflows under

RCP scenarios: A case study in Xin River Basin, China. Atmos. Res. 178-179, 521-

534.

T
R IP
SC
NU
MA
D
P TE
CE
AC
 ACCEPTED MANUSCRIPT

List of Figures

Figure 1: (a) Kelantan River Basin (KRB), (b) Peninsular Malaysia, and (c) mean monthly

T
IP
rainfall and temperature at the Haiwan Machang station from 1975 to 2004 (The station with

black dotted is the Gua Musang station).

R
SC
Figure 2: Schematic diagram of this study.

NU
Figure 3: Comparison of observed and simulated monthly streamflow during calibration and
MA
validation.
D

Figure 4: Annual rainfall and maximum temperature changes at the (a) Haiwan Machang and
TE

(b) Gua Musang stations under three RCP scenarios for the 2015-2044 and 2045-2074
P

periods.
CE

Figure 5: Relative changes in monthly rainfall and maximum temperature at the (a) Haiwan
AC

Machang and (b) Gua Musang stations under the RCP scenarios for the 2015-2044 and 2045-

2074 periods.

Figure 6: Relative changes of monthly (a) streamflow, (b) evapotranspiration, (c) surface

runoff and (d) water yield for the periods of 2015-2044 and 2045-2074 under the three RCP

scenarios.
 ACCEPTED MANUSCRIPT

T
IP
R
SC
NU
MA
D
PTE

Figure 1
CE
AC
 ACCEPTED MANUSCRIPT

T
IP
R
SC
NU
MA
D
PTE
CE
AC

Figure 2
 ACCEPTED MANUSCRIPT

T
IP
R
SC
NU
MA

Figure 3
D
PTE
CE
AC
 ACCEPTED MANUSCRIPT

T
IP
R
SC
NU
MA
D
PTE
CE

Figure 4
AC
 ACCEPTED MANUSCRIPT

T
IP
R
SC
NU
MA
D
PTE

Figure 5
CE
AC
 ACCEPTED MANUSCRIPT

T
IP
R
SC
NU
MA
D
PTE
CE
AC

Figure 6
 ACCEPTED MANUSCRIPT

List of tables

T
IP
Table 1: Thirty-seven climate change scenarios applied in this study.

R
ID Model Scenario Period Symbol

SC
1 CESM1-CAM5 RCP 2.6 2015-2044 CESM2.6_15-44

NU
2 CESM1-CAM5 RCP 2.6 2045-2074 CESM2.6_45-74

3 CESM1-CAM5 RCP 4.5 2015-2044 CESM4.5_15-44

MA
4 CESM1-CAM5 RCP 4.5 2045-2074 CESM4.5_45-74

5 CESM1-CAM5 RCP 8.5 2015-2044 CESM8.5_15-44

D

6 CESM1-CAM5 RCP 8.5 2045-2074 CESM8.5_45-74

TE

7 CNRM-CM5 RCP 2.6 2015-2044 CNRM2.6_15-44

8 CNRM-CM5 RCP 2.6 2045-2074 CNRM2.6_45-74

P
CE

9 CNRM-CM5 RCP 4.5 2015-2044 CNRM4.5_15-44

10 CNRM-CM5 RCP 4.5 2045-2074 CNRM4.5_45-74

AC

11 CNRM-CM5 RCP 8.5 2015-2044 CNRM8.5_15-44

12 CNRM-CM5 RCP 8.5 2045-2074 CNRM8.5_45-74

13 MIROC-ESM RCP 2.6 2015-2044 MIROC2.6_15-44

14 MIROC-ESM RCP 2.6 2045-2074 MIROC2.6_45-74

15 MIROC-ESM RCP 4.5 2015-2044 MIROC4.5_15-44

16 MIROC-ESM RCP 4.5 2045-2074 MIROC4.5_45-74

17 MIROC-ESM RCP 8.5 2015-2044 MIROC8.5_15-44

18 MIROC-ESM RCP 8.5 2045-2074 MIROC8.5_45-74

 ACCEPTED MANUSCRIPT

19 MRI-CGCM RCP 2.6 2015-2044 MRI2.6_15-44

20 MRI-CGCM RCP 2.6 2045-2074 MRI2.6_45-74

21 MRI-CGCM RCP 4.5 2015-2044 MRI4.5_15-44

T
IP
22 MRI-CGCM RCP 4.5 2045-2074 MRI4.5_45-74

23 MRI-CGCM RCP 8.5 2015-2044 MRI8.5_15-44

R
SC
24 MRI-CGCM RCP 8.5 2045-2074 MRI8.5_45-74

25 Nor-ESM1M RCP 2.6 2015-2044 Nor2.6_15-44

NU
26 Nor-ESM1M RCP 2.6 2045-2074 Nor2.6_45-74

27 Nor-ESM1M RCP 4.5 2015-2044 Nor4.5_15-44

MA
28 Nor-ESM1M RCP 4.5 2045-2074 Nor4.5_45-74

29 Nor-ESM1M RCP 8.5 2015-2044 Nor8.5_15-44

D

30 Nor-ESM1M RCP 8.5 2045-2074 Nor8.5_45-74

TE

31 Ensemble RCP 2.6 2015-2044 ENS2.6_15-44

P

32 Ensemble RCP 2.6 2045-2074 ENS2.6_45-74

CE

33 Ensemble RCP 4.5 2015-2044 ENS4.5_15-44

34 Ensemble RCP 4.5 2045-2074 ENS4.5_45-74

AC

35 Ensemble RCP 8.5 2015-2044 ENS8.5_15-44

36 Ensemble RCP 8.5 2045-2074 ENS8.5_45-74

37 Observed dataset Baseline 1975-2004 Baseline

 ACCEPTED MANUSCRIPT

Table 2: Calibrated parameters range for the SWAT model (1: most sensitive)

No Parameter Initial range Final range Fitted

T
Min Max Min Max value

IP
1 V__GW_REVAP.gw 0.02 0.4 0.1 0.4 0.38

R
2 V__CH_K2.rte 0 500 0 80 7.28

SC
3 V__ALPHA_BF.gw 0 1 0.1 0.5 0.34

4 R__CN2.mgt -0.5 0.5 0 0.35 0.14

5 V__GW_DELAY.gw 0 500
NU
0 130 15.21
MA
6 V__ESCO.hru 0 1 0.4 0.9 0.40

7 V__GWQMN.gw 0 5000 2200 4000 2723.80

8 V__CH_N2.rte 0 0.3 0.2 0.3 0.29

D
TE

9 R__SOL_AWC.sol 0 0.5 0 0.5 0.30

10 V__SURLAG.bsn 0.05 24 8 19 14.79

P

11 V__REVAPMN.gw 0 500 70 320 112.75

CE

12 V__RCHRG_DP.gw 0 1 0.4 0.6 0.53

AC

* V__: the parameter value is replaced with the given value, R__: the parameter value is

multiplied with the (1 + given value).

 ACCEPTED MANUSCRIPT

Table 3: Changes of annual climate and water cycle components for the periods of 2015-2044

and 2045-2074. (SF = streamflow; ET = evapotranspiration; SURQ = surface runoff; WYLD

= water yield).

T
R IP
ID Symbol SF ET SURQ WYLD

SC
% m3s-1 % mm % mm % mm

NU
1 CESM2.6_15-44 3.3 20.8 0.0 -0.1 19.5 82.8 3.2 54.2

2 CESM2.6_45-74 5.8 36.5 1.6 3.9 20.0 85.3 5.7 96.5

MA
3 CESM4.5_15-44 5.9 36.7 0.7 1.6 25.8 109.8 5.7 95.9

4 CESM4.5_45-74 9.8 61.3 2.3 5.6 27.4 116.5 9.7 162.3

D

5 CESM8.5_15-44 8.9 55.4 1.6 3.8 28.0 119.4 8.7 145.8

TE

6 CESM8.5_45-74 4.2 26.2 2.3 5.6 20.7 88.2 4.2 70.8

7 CNRM2.6_15-44 4.3 27.2 0.5 1.2 10.6 45.3 4.3 71.6

P
CE

8 CNRM2.6_45-74 3.1 19.4 0.8 2.0 10.8 46.2 3.1 51.4

9 CNRM4.5_15-44 0.9 5.6 0.7 1.7 5.9 25.0 0.9 15.2

AC

10 CNRM4.5_45-74 5.2 32.8 1.8 4.3 12.2 52.1 5.2 87.3

11 CNRM8.5_15-44 1.9 11.9 0.7 1.7 7.4 31.5 1.9 31.8

12 CNRM8.5_45-74 6.7 41.9 1.7 4.1 35.5 151.2 6.6 110.9

13 MIROC2.6_15-44 1.8 11.6 1.4 3.4 3.5 14.8 1.9 31.4

14 MIROC2.6_45-74 4.2 26.1 1.7 4.2 7.5 31.9 4.2 69.9

15 MIROC4.5_15-44 3.2 20.2 1.7 4.1 6.6 27.9 3.2 54.2

16 MIROC4.5_45-74 10.8 67.6 4.1 10.1 15.6 66.4 10.8 180.6

17 MIROC8.5_15-44 3.6 22.4 2.3 5.6 6.4 27.1 3.6 60.6

 ACCEPTED MANUSCRIPT

18 MIROC8.5_45-74 27.2 169.9 6.0 14.6 43.3 184.4 26.8 449.9

19 MRI2.6_15-44 0.5 3.4 0.2 0.6 3.9 16.5 0.6 9.4

20 MRI2.6_45-74 -4.9 -30.6 -1.4 -3.3 0.7 2.8 -4.8 -80.5

T
IP
21 MRI4.5_15-44 -3.6 -22.4 -0.9 -2.2 -0.2 -0.6 -3.5 -58.8

22 MRI4.5_45-74 7.6 47.7 1.5 3.6 15.5 65.8 7.5 126.3

R
SC
23 MRI8.5_15-44 10.6 66.5 -1.0 -2.5 48.1 204.8 10.3 173.0

24 MRI8.5_45-74 27.6 172.4 0.1 0.3 91.0 387.3 26.9 451.4

NU
25 Nor2.6_15-44 75.4 471.6 -0.7 -1.7 252.5 1075.0 73.1 1228.2

26 Nor2.6_45-74 64.7 404.9 0.9 2.2 195.2 831.0 62.9 1056.5

MA
27 Nor4.5_15-44 84.4 527.7 0.6 1.5 266.7 1135.7 81.8 1374.2

28 Nor4.5_45-74 49.9 312.2 2.9 7.1 136.7 582.1 48.6 816.6

D

29 Nor8.5_15-44 111.2 695.4 0.4 1.0 361.3 1538.2 107.8 1810.6

TE

30 Nor8.5_45-74 58.8 367.9 3.3 8.0 165.5 704.5 57.3 962.6

P

31 ENS2.6_15-44 17.1 106.9 0.3 0.7 58.0 246.9 16.6 279.0

CE

32 ENS2.6_45-74 14.6 91.3 0.7 1.8 46.8 199.4 14.2 238.7

33 ENS4.5_15-44 18.2 113.5 0.6 1.3 61.0 259.5 17.6 296.1

AC

34 ENS4.5_45-74 16.7 104.3 2.5 6.1 41.5 176.6 16.4 274.6

35 ENS8.5_15-44 27.2 170.3 0.8 1.9 90.2 384.2 26.5 444.4

36 ENS8.5_45-74 24.9 155.7 2.7 6.5 71.2 303.1 24.4 409.1

 ACCEPTED MANUSCRIPT

T
R IP
SC
NU
MA
D
PTE
CE
AC

Graphical abstract
 ACCEPTED MANUSCRIPT

Highlights

- Future hydro-climatic changes in the Kelantan River Basin are investigated.

T
IP
- The SWAT model performed well in a tropical basin.

- Annual climate, temperature and water cycle components are projected to increase in

R
SC
future.

- Rainfall, Streamflow and surface runoff will increase significantly from November to

NU
January.

- Water yield is expected to decrease between June and October during the period
MA
2015-2045.
D
P TE
CE
AC