DSpace at VNU: Impact of climate and land-use changes on hydrological processes and sediment yield-a case study of the B...
Trang 1On: 03 June 2014, At: 22:40
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Hydrological Sciences Journal
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/thsj20
Impact of climate and land-use changes on hydrological processes and sediment yield—a case study of the Be River catchment, Vietnam
Dao Nguyen Khoia & Tadashi Suetsugib
a
Faculty of Environmental Science, University of Science, Vietnam National University, Ho Chi Minh City, Vietnam
b
Interdisciplinary Graduate School of Medicine and Engineering, University of Yamanashi, Kofu, Yamanashi 400-8511, Japan
Accepted author version posted online: 04 Jul 2013.Published online: 29 Apr 2014
To cite this article: Dao Nguyen Khoi & Tadashi Suetsugi (2014) Impact of climate and land-use changes on hydrological
processes and sediment yield—a case study of the Be River catchment, Vietnam, Hydrological Sciences Journal, 59:5,
1095-1108, DOI: 10.1080/02626667.2013.819433
To link to this article: http://dx.doi.org/10.1080/02626667.2013.819433
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content
This article may be used for research, teaching, and private study purposes Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
Trang 2Impact of climate and land-use changes on hydrological processes and
Dao Nguyen Khoi1 and Tadashi Suetsugi2
1
Faculty of Environmental Science, University of Science, Vietnam National University, Ho Chi Minh City, Vietnam
dnkhoi86@gmail.com
2
Interdisciplinary Graduate School of Medicine and Engineering, University of Yamanashi, Kofu, Yamanashi 400-8511, Japan Received 9 June 2012; accepted 8 May 2013; open for discussion until 1 November 2014
Editor Z.W Kundzewicz; Associate editor Q Zhang
Citation Khoi, D.N and Suetsugi, T., 2014 Impact of climate and land-use changes on hydrological processes and sediment yield—a case study of the Be River catchment, Vietnam Hydrological Sciences Journal, 59 (5), 1095 –1108.
Abstract The impact of climate and land-use changes on hydrological processes and sediment yield is investi-gated in the Be River catchment, Vietnam, using the Soil and Water Assessment Tool (SWAT) hydrological model The sensitivity analysis, model calibration and validation indicated that the SWAT model could reasonably simulate the hydrology and sediment yield in the catchment From this, the responses of the hydrology and sediment to climate change and land-use changes were considered The results indicate that deforestation had increased the annual flow (by 1.2%) and sediment load (by 11.3%), and that climate change had also significantly increased the annual streamflow (by 26.3%) and sediment load (by 31.7%) Under the impact of coupled climate and land-use changes, the annual streamflow and sediment load increased by 28.0% and 46.4%, respectively In general, during the 1978 –2000 period, climate change influenced the hydrological processes in the Be River catchment more strongly than the land-use change.
Key words climate change; hydrology; land-use change; sediment yield; SWAT model; Be River catchment, Vietnam
Impact des changements climatiques et de l’utilisation des terres sur les processus hydrologiques
et la production de sédiments—étude de cas du bassin versant de la rivière Be, Vietnam
Résumé L ’impact des changements du climat et de l’utilisation des terres sur les processus hydrologiques et l’apport de sédiments dans le bassin versant de la rivière Be (Vietnam) a été étudié en utilisant le modèle hydrologique SWAT L’analyse de sensibilité, l’étalonnage et la validation des modèles indique que le modèle SWAT peut raisonnablement simuler l’hydrologie et la charge sédimentaire dans le bassin versant C’est donc avec cet outil que les réponses de l’hydrologie et des sédiments au changement climatique et au changement d’utilisation des terres ont été étudiées Les résultats indiquent que la déforestation a augmenté l’écoulement annuel (1,2%) et la charge sédimentaire (11,3%), et que le changement climatique a également augmenté de manière significative le débit annuel (26,3%) et la charge sédimentaire (31,7%) Sous l’impact couplé du changement climatique et du changement d’utilisation des terres, l’écoulement annuel et la charge de sédiments ont respectivement augmenté de 28% et 46,4% En général, le changement climatique a eu une influence plus importante sur les processus hydrologiques que le changement d ’utilisation des terres dans le bassin versant de la rivière Be durant la période 1978 –2000.
Mots clefs changement climatique ; hydrologie ; changement d’utilisation des terres ; production de sédiments ; modèle SWAT ; bassin versant de la rivière Be, Vietnam
1 INTRODUCTION
The principal influences on hydrological processes and
soil erosion include not only climate change but also
land-use/land-cover change Climate change is likely to
affect the hydrological cycle with changes in
tempera-ture and precipitation, and this may lead to changes in
water availability, as well as the transformation and transport characteristics of pollutants (Tu 2009) Changes in land use as a result of deforestation, agricultural expansion and urbanization have altered surface runoff generation, and have then affected the hydrological processes and the transport of pollutants
http://dx.doi.org/10.1080/02626667.2013.819433
Trang 3As a result, climate and land use are identified as key
factors controlling the hydrological and sediment
beha-viours of catchments (Elfert and Bormann2010) It is
important to understand the hydrological and sediment
responses to these changes in order to develop
strate-gies for land-use planning and water resource
manage-ment Studies of the hydrological and water quality
impacts of climate change and land-use change are
desirable (Tong et al.2012)
Many studies have considered the impact of cli-mate change and land-use change on hydrology (Li
et al.2009,2012, Ma et al.2009,2010, Mango et al
2011, Zhang et al.2011) However, few studies have
investigated changes in hydrological processes and
water quality as well as sediment yield under the impact
of climate and land-use changes on a basin scale (Ward
et al.2009, Tong et al.2012) To assess the hydrological
and sediment impacts of environmental change, the
common methods used are the paired catchment
approach, statistical analysis and hydrological
model-ling (Li et al.2009,2012) Among these approaches,
the hydrological method is an appealing option, because
it is most suitable to be used as a part of scenario studies
There are numerous hydrological models, such as the
Water Erosion Prediction Project (WEPP), Hydrologic
Simulation Program Fortran (HSPF), the Soil and Water
Assessment Tool (SWAT) and the physically-based
distributed hydrological model Système Hydrologique
Européen TRANsport (SHETRAN), that could be used
in simulating the runoff and transport of sediment and
pollutants in the catchment The SWAT model has been
selected for the current study because it is widely used
to assess hydrology and water quality in agricultural
catchments around the world (see the SWAT literature
database: https://www.card.iastate.edu/swat_articles/)
Another reason for its selection is its availability and
user-friendliness in terms of handling input data
(Arnold et al.1998)
Vietnam has experienced climate changes, including rising air temperature and more variable
precipitation (MONRE2009) In addition, rapid
agri-cultural and industrial development, as well as
popu-lation growth have occurred in recent decades (Trinh
2007) These changes have affected soil erosion and
the availability of water resources in Vietnam
However, no studies have investigated the effects of
climate change and human activities on hydrological
cycles and sediment yield in Vietnam Moreover,
Wang et al (2012) emphasized that the local impacts
of climate change and human activities on hydrology
and sediment yield vary from place to place and need
to be investigated on a regional scale
The overall objective of this study was to quan-tify the impacts of past land-use change and climate change on hydrological processes and sediment yield
in a case study of a catchment in Vietnam The specific objectives were: (a) to calibrate and validate the SWAT model in terms of streamflow and sedi-ment load in the Be River catchsedi-ment; (b) to evaluate the separate impacts of climate and land-use changes
on hydrology and sediment yield; and (c) to assess the impacts of combined climate change and land-use change on hydrological processes and sediment yield The results achieved through this study provide deci-sion-makers with a comprehensive understanding of the interactions among hydrological processes, land-use change and climate change, which are required to assist with water resource planning efforts and sus-tainable development
2 STUDY AREA The catchment selected for study lies in the Dong Nai River basin in south Vietnam between latitudes 11°10′– 12°16′N and longitudes 106°36′–107°30′E (Fig 1) It is located in Dak Nong, Binh Phuoc, Binh Duong and Dong Nai provinces, and has a catchment area of about 7500 km2 The altitude varies from 1000 m a.m s.l in the highland area to 100 m a.m.s.l in the plain area, in a northeast to southwest and south direction The origin of the branched-tree drainage system of the Be River lies in Tuy Duc on the international border between Vietnam and Cambodia, in Dak Nong province The study area is located in the steep area The degree of slope can be divided into three levels: slopes of 0% to 7% account for 45% of the total area, slopes of 8–15% account for 33% of the area, and slopes greater than 15% account for 22% of the area The climate is tropical monsoon The annual rainfall varies between 1800 and 2800 mm, with an average of
2400 mm year-1 The area has two seasons: the rainy season and the dry season The rainy season lasts from May to November and accounts for 85–90% of the total annual precipitation The average temperature is about 25.9°C, the maximum temperature is 36.6°C and the minimum temperature is 17.3°C The area has relatively fertile land (75% basalt soil), consistent with agricultural development The main land-use types in this catchment are forest and agricultural lands The total population in
2010 was approximately one million inhabitants The mean annual flow of the catchment is about 7.51 × 109 m3 Similar to the distribution of rainfall, the flow is distinguished by two distinct seasons: the flood season (accounting for 67% of the total annual
Trang 4flow) and the low-flow season (accounting for 33% of
the total annual flow) The Be River catchment has been
assessed as having the most abundant water resources in
the Dong Nai River basin and significant hydropower
potential
3 METHODOLOGY
3.1 Change detection in hydro-meteorological
data The Mann-Kendall test (Mann 1945, Kendall 1975)
is a non-parametric test for identifying trends in
hydro-meteorological time series The
Mann-Kendall test statistic is calculated as follows:
Zc ¼
Sffiffiffiffiffiffiffiffiffi1 var S ð Þ
p S > 0
Sffiffiffiffiffiffiffiffiffiþ1 var S ð Þ
p S < 0
8
>
where
S ¼X
n 1
i ¼1
Xn
j ¼iþ1
sgn xj xi
(2)
sgn xj xi
¼ þ1 xj xi
> 0
0 xj xi ¼ 0
1 xj xi < 0
8
<
varðSÞ ¼
n nð 1Þ 2n þ 5ð Þ Pm
i ¼1tiðti 1Þ 2tð iþ 5Þ
18
(4)
where n is the length of the data set, xiand xjare the sequential data values, m is the number of tied groups (a tied group is a set of sample data with the same value), and t is the number of data points in the mth group The null hypothesis H0 (there is no trend) is accepted if –Z1 –α/2 ≤ Zc ≤ Z1 –α/2, where α is the significant level A positive value of Zc indicates an increasing trend, and a negative value indicates a decreasing trend
In the Mann-Kendall test, the Kendall slope is another very useful index that estimates the magni-tude of the monotonic trend and is given by:
β ¼ Median xj xi
j i
"i < j (5)
Fig 1 Location map of the Be River catchment.
Trang 5where 1 < i < j < n The estimator β is calculated as
the median of all slopes between data pairs for the
entire data set
The Pettitt test (Pettitt 1979) is a non-parametric approach used for detecting the change point There
are two samples (x1, x2,…, xt) and (xt+1, xt+2,…, xN)
that come from the same population (x1, x2, …, xN)
The test statistic Ut,N is given by:
Ut ;N ¼X
t
i ¼1
XN
j ¼tþ1
sgn xi xj
(6)
The null hypothesis of the Pettitt test is the absence of a change point Its statistic Kt and
asso-ciated probabilities are given as:
Kt ¼ max U t ;N (7)
p¼ exp 6Kt2
N3þ N2
(8)
When p is smaller than the specific significance level, the null hypothesis is not accepted The time t
when Kt occurs is the change point time
These methods have been commonly used to detect changes in hydro-meteorological data (Ma
et al 2008, Zhang et al.2009, Zhang et al 2011)
3.2 SWAT model
The SWAT model is a physically based, distributed,
continuous time model that is designed to predict the
effects of land management on the hydrology, sediment
and agricultural chemical yields in agricultural
water-sheds with varying soils, land-use and management
conditions (Arnold et al 1998) In the SWAT model,
a catchment is divided into a number of sub-watersheds
or sub-basins Sub-basins are further partitioned into
hydrological response units (HRUs) based on soil
types, land-use and slope classes that allow a high
level of spatial detail simulation The model predicts
the hydrology at each HRU using the water balance
equation, comprising precipitation, surface runoff,
eva-potranspiration, infiltration and subsurface flow
The SWAT model provides two methods for esti-mating surface runoff: the SCS curve number procedure
(USDA-SCS1972) and the Green and Ampt infiltration
method (Green and Ampt1911) SWAT calculates the
peak runoff rate using a modified rational method The
potential evapotranspiration is estimated in the SWAT
model using three methods: the Penman-Monteith
method (Monteith 1965), the Priestley-Taylor method (Priestley and Taylor1972) and the Hargreaves method (Hargreaves et al.1985) Channel routing is simulated using the variable storage coefficient method (William
1969) and the Muskingum method (Chow1959) The SWAT model uses the Modified Universal Soil Loss Equation (MUSLE) to simulate the sedi-ment yield for each HRU The MUSLE (William
1995) is given as:
sed¼ 11:8 Qsurf qpeak areaHRU
KUSLE CUSLE PUSLE LSUSLE
CFRG
(9)
where sed is the sediment yield on a given day (t),
Qsurfis the surface runoff volume (mm ha-1), qpeakis the peak runoff rate (m3 s-1), areaHRU is the area of the HRU (ha), KUSLE is the USLE soil erodibility factor, CUSLE is the USLE cover and management factor, PUSLE is the USLE support practice factor,
LSUSLE is the USLE topographic factor and CFRG
is the coarse fragment factor
The channel sediment-routing model consists of deposition and degradation, which operate simulta-neously In the channel, deposition or degradation can occur, depending on the sediment loads from upland areas and the transport capacity of the channel network If the sediment entering a channel is larger than its sediment transport capacity, channel deposi-tion will occur Otherwise, channel degradadeposi-tion will
be the dominant process
Further details of hydrological and sediment transport processes can be found in the SWAT Theoretical Documentation (Neitsch et al.2011)
3.3 SWAT model set-up The input data required for the SWAT model include weather data, a land-use map, a soil map and a digital elevation map (DEM), as listed inTable 1 The land-use data were generated from Landsat satellite images— Landsat Thematic Mapper (TM) image in 1990 and Landsat Enhanced Thematic Mapper Plus (ETM +) image in 2001—obtained from the US Geological Survey Earth Resources Observation and Science Center Land-use maps were generated using supervised classification based on the maximum likelihood algo-rithm in the ENVI Version 4.4 image processing soft-ware Overall accuracy and kappa statistic (κ) were used
to assess classification accuracy based on 256 ground control points selected from the referenced land-use map
Trang 6for 2001, which was obtained from the Southern
Institute for Water Resources Planning (SIWRP2002)
Land-use types were classified into the following
cate-gories: forest, rangeland, agricultural land, urban area
and water
Daily river flow data measured at Phuoc Long (1981–1993) and Phuoc Hoa (1981–2000) gauging
stat-ions (Fig 1) were used for the model calibration and
validation of flow simulation Monthly sediment load
data measured at Phuoc Hoa station (07/1999–2004)
were used for the calibration and validation of sediment
simulation Streamflow and sediment load data were
provided by the Hydro-Meteorological Data Center of
Vietnam
The model set-up consists of five steps: (a) data preparation, (b) sub-basin discretization, (c) HRU
defi-nition, (d) parameter sensitivity analysis, and (e)
cali-bration and validation Sensitivity analysis was carried
out to identify the most sensitive parameters for the
model calibration using Latin hypercube and
one-fac-tor-at-a-time (LH-OAT), an automatic sensitivity
ana-lysis tool implemented in SWAT (Van Griensven et al
2006) Those sensitive parameters were calibrated
using the auto-calibration tool that is currently available
in the SWAT interface (Van Liew et al.2005)
3.4 Performance evaluation of the SWAT model
The model performance was evaluated using
statisti-cal analysis to compare the quality and reliability of
the simulated discharge with the observed data In
this study, the model evaluation methods used
included: the Nash-Sutcliffe (1970) efficiency
criter-ion, NSE; per cent bias, PBIAS; and the ratio of the
root mean square error (RMSE) to the standard
deviation (STDEV) of measured data, RSR The
RSR is calculated as (Moriasi et al.2007):
RSR¼ RMSE
STDEVobs
¼
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
PN
i ¼1ðOi PiÞ2
s
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
PN
i ¼1ðOi OÞ2
where Oi is the observed value Pi is the simulated value, O is the mean of the observed data, and N is the total number of observations
According to Moriasi et al (2007) and Rossi
et al (2008), model simulation can be judged as satisfactory if NSE > 0.5, RSR ≤ 0.70 and PBIAS = ±25% for streamflow simulation, and NSE > 0.5, RSR ≤ 0.70 and PBIAS = ±55% for sediment simulation
4 RESULTS AND DISCUSSION 4.1 Land-use changes
Based on the Landsat images, land-use maps were generated for 1990 and 2001, as illustrated inFig 2
An accuracy assessment of land-cover classification, obtained by computing the confusion matrix in ENVI 4.4 software, showed an overall accuracy value of 98.2% for 1990 and 98.1% for 2001 The κ coeffi-cients for 1990 and 2001 were 0.96 and 0.97, respec-tively The dominant land-use types in the Be River catchment were agricultural land and forest (Table 2), which accounted, respectively, for 40.28% and 50.69% in 1990 and 55.18% and 36.62% in 2001 Range land, urban and water covered about 8.89%, 0.03% and 0.11%, respectively, of the total catchment area for 1990, and 6.85%, 0.13% and 1.22%, respec-tively, of the total area for 2001 In general, there were two main trends of land-use change: a decrease
in the forest (deforestation) and an increase in agri-cultural land (agriagri-cultural expansion) Compared with
1990, the forest decreased by 14.07% and cropland increased by 14.89% of the catchment area Aside from this, there were slight changes in the range land (–2.03%), water (1.11%), and urban area (0.11%) These changes were likely caused by a population increase, which led to the expansion of settlements and agricultural land, and ineffective forest manage-ment that led to excessive forest exploitation (SIWRP
2002, 2008) The population of the Be River catch-ment was about 680 000 in 2000 compared to
400 000 in 1990 (SIWRP 2002) This represents a population increase of about 170%
Table 1 Spatial model input data for the Be River catchment.
Data type Description Resolution Source
Topographic map Digital elevation map (DEM) 90 m SRTM
Land-use map Land-use classification 1 km Landsat TM, ETM+ (USGS/GLOVIS) Soil map Soil types 10 km FAO
Weather Daily precipitation, minimum and maximum temperature 9 stations Hydro-Meteorological Data Center (HMDC)
Trang 74.2 Change detection for hydro-meteorological
data Annual temperature, precipitation and streamflow
were tested using the Mann-Kendall and Pettitt
meth-ods, as reported in Table 3 and illustrated in Fig 3
The results showed rises in annual temperature,
pre-cipitation and streamflow (by 0.035°C year-1,
20.613 mm year-1 and 3.142 m3 s-1 year-1, respec-tively) at the 5% significance level In other words, the null hypothesis H0 was not accepted for the annual temperature, rainfall and streamflow time ser-ies Change points in the annual rainfall and stream-flow were detected as occurring around 1989, with a significance level of 5%, while the change point in annual temperature was statistically significant
in 1986
The Mann-Kendall test was also applied to the data series for monthly precipitation and temperature,
as summarized in Table 4 There are no significant trends in most of the monthly precipitation time series, except for October and December The pre-cipitation in October and December showed signifi-cant increasing trends of 2.24 and 1.73 mm year-1, respectively In the case of the monthly temperature, significant increasing trends were detected for most
of the monthly temperature time series, except for February, March, April and May
Fig 2 Land-use maps of the Be River catchment.
Table 2 Statistics for land-use changes in the Be River catchment for the period of 1978 –2007.
(km 2 ) (%) (km 2 ) (%) (km 2 ) (%) Agricultural land 3015 40.28 5129 55.18 1114 14.89
Forest 3794 50.69 2741 36.62 –1053 –14.07
Table 3 Summary of Mann-Kendall trend test and Pettitt
test statistics for annual rainfall, temperature and
stream-flow in the Be River catchment.
Mann-Kendall test Pettitt test for change point
Z c β p KT t p Precipitation 2.27 20.613 * 88 1989 *
Temperature 3.16 0.035 * 102 1986 *
Streamflow 1.74 3.142 * 78 1989 *
*indicates significant at p < 0.05.
Trang 84.3 Hydrological and sediment responses to
land-use and climate changes
To investigate the impacts of climate change and
land-use change on hydrological processes and
sedi-ment yield, the approach of one factor at a time was
used (Li et al 2009) The change point of
precipita-tion is selected as the change point in climate data,
because rainfall plays a key role in hydrology and is
the most fundamental meteorological variable on the
catchment scale The meteorological data were divided into two periods, 1978–1989 and 1990–
2000, based on the change point analysis, and each period included one land-use map The land-use map for 1990 was used to represent the 1978–1989 per-iod, and that for 2001 was used to represent the 1990–2000 period The following four scenarios were investigated:
– Scenario 1 (Baseline): Land-use in 1990 and cli-mate data for the 1978–1989 period
– Scenario 2 (Climate change): Land-use in 1990 and climate data for the 1990–2000 period – Scenario 3 (Land-use change): Land-use in 2001 and climate data for the 1978–1989 period – Scenario 4 (Climate and land-use changes): Land-use in 2001 and climate data for the 1990–2000 period
4.4 Model calibration and validation The LH-OAT parameter sensitivity analysis procedure showed that the most sensitive parameters for flow simulation were curve number (CN2), soil evaporation compensation factor (ESCO), threshold water depth in the shallow aquifer for flow (GQWMN), baseflow alpha factor (ALPHA_BF), soil depth (SOL_Z), available water capacity (SOL_AWC), channel effective hydrau-lic conductivity (CH_K2), groundwater‘revap’ coeffi-cient (GW_REVAP), Manning’s value for the main channel (CH_N2), and saturated hydraulic conductivity (SOL_K) The most sensitive parameters for sediment simulation were the linear re-entrainment parameter for channel sediment routing (SPCON), the exponent of
re-Table 4 Summary of Mann-Kendall trend test statistics for monthly rainfall and temperature in the Be River catchment.
Month Precipitation Temperature
Zc β P Zc β p January 0.66 0.208 2.82 0.108 * February 1.58 0.589 1.50 0.041 March 0.63 0.762 0.37 0.010 April 0.58 0.741 –0.62 –0.021 May 1.85 3.313 0.06 0.001 June –0.48 –2.740 2.51 0.051 * July 1.40 2.783 2.37 0.024 * August –1.11 –3.676 3.10 0.043 * September –0.16 –0.233 3.21 0.042 * October 2.27 2.244 * 2.51 0.048 * November 0.53 1.325 2.51 0.050 * December 2.09 1.733 * 3.59 0.085 *
*indicates significant at p < 0.05.
Fig 3 Variations of mean values in (a) annual
precipita-tion, (b) annual temperature and (c) annual discharge in the
Be River catchment (1978–2000).
Trang 9entrainment parameter for channel sediment routing
(SPEXP), and the USLE support practice factor
(USLE_P) These sensitive parameters were optimized
using the auto-calibration extension of ArcSWAT 2009
to calibrate the model The daily streamflow for 1981–
1989 at the Phuoc Long and Phuoc Hoa stations and the
land-use map for 1990 were used for model calibration
The daily streamflow for 1990–1993 for the Phuoc Long
station and 1990–2000 for the Phuoc Hoa station and
land-use map for 2001 were used for the model
valida-tion of flow simulavalida-tion This approach was used in the
study undertaken by Li et al (2009) Because of the lack
of observed sediment load, these data were only
avail-able from 07/1999 to 2004 at monthly levels They were
divided into two periods for calibration (07/1999–2001)
and validation (2002–2004) using the land-use map for
2001 The flow calibration and validation was conducted
first, and then the sediment calibration and validation
As a result of the calibration, the most sensitive
flow-related parameter of CN2 was adjusted to have values of
–0.29 for Phuoc Long and –0.36 for Phuoc Hoa, and the
most sensitive sediment-related parameter SPCON was
adjusted to have a value of 0.001 This value of SPCON
found here was similar to that in the study conducted by
Phan et al (2011) in the Cau River watershed in northern
Vietnam The details of the calibrated parameters are
presented inTable 5
The SWAT flow simulations were calibrated against the daily flow from 1981 to 1989 and validated
from 1990 to 1993 at the Phuoc Long gauging station,
as shown in Fig 4 The simulated daily flow fit the
observed data for the calibrated period well, with NSE, PBIAS and RSR values of 0.77, 1.60% and 0.47, respectively For the validation period, the values of NSE = 0.79, PBIAS = 3.30% and RSR = 0.45 suggest that there was good agreement between the simulated and observed streamflow during this period, based on
Table 5 SWAT sensitivity parameters and calibrated values.
Simulation Parameter Description of parameter Range Calibrated value
Phuoc Long Phuoc Hoa Flow CN2 Initial SCS CN II value*** ±0.5 –0.29 –0.36
ESCO Soil evaporation compensation factor* 0 – 1 0.95 0.42 GQWMN Threshold water depth in the shallow aquifer for flow** 0 – 5000 456 2356 ALPHA_BF Baseflow alpha factor* 0 – 1 0.11 0.61
SOL_AWC Available water capacity*** ±0.5 0.23 0.40 CH_K2 Channel effective hydraulic conductivity** –0.01 – 500 184 184 GW_REVAP Groundwater ‘revap’ coefficient** 0.02 – 0.2 0.17 0.17 CH_N2 Manning’s value for main channel* –0.01 – 0.3 0.04 0.04 SOL_K Saturated hydraulic conductivity*** ±0.5 –0.06 0.07 Sediment SPCON Linear re-entrainment parameter for channel sediment
routing*
0.0001 – 0.01 0.001 SPEXP Exponent of re-entrainment parameter for channel sediment
routing*
1 – 1.5 1.01 USLE_P USLE support practice factor* 0 – 1 0.42
*Parameter value is replaced by given value.
**Parameter value is added by given value.
***Parameter value is multiplied by (1 + a given value).
Fig 4 Observed and simulated daily flow hydrograph at the Phuoc Long station: (a) calibration and (b) validation.
Trang 10the performance criteria given by Moriasi et al (2007).
The aggregated monthly average flow values from the
daily flow values improved the fit between the model
predictions and observed flows More detail can be seen
in Table 6 Figure 5shows a hydrograph of the
simu-lated and observed daily flow for the calibration and
validation periods at the Phuoc Hoa station The
statis-tical evaluations shown in Table 6 also suggest that
there was good agreement between the daily measured
and simulated streamflow during these periods,
accord-ing to Moriasi et al (2007) This agreement is shown by
values of NSE = 0.86, RSR = 0.37 and
PBIAS = –1.90% for the calibration period and
NSE = 0.71, RSR = 0.54 and PBIAS = –6.20% for
the validation period In the case of the aggregated
monthly average flow, the match between the simulated
flow values and the observed values was improved
This match is shown inTable 6 Although the simulated and observed streamflow followed the same trend, the peak flow was overestimated for Phuoc Long station and underestimated for Phuoc Hoa station This may have resulted from the uneven spatial distribution of the rain gauges In the study area, eight rain gauges are located in the lower area of the catchment; however, only one rain gauge located in the upper area of the catchment has long-term records (Fig 1) A further reason can be attributed to the CN2, which is used to simulate the surface runoff The CN2 method assumes a unique relationship between cumulative rainfall and cumulative runoff for the same antecedent moisture conditions (Betrie et al 2011) Generally speaking, these results reveal that the hydrological processes in SWAT are modelled realistically for the Be River catch-ment, which is important for the simulation of sediment The simulated sediment load values were cali-brated against monthly observed data from 07/1999 to
2001 and validated from 2002 to 2004 at the Phuoc Hoa station, as presented in Fig 6 The fit between the simulated and observed sediment loads was acceptable, according to Moriasi et al (2007) The fit was indicated
by the values of NSE = 0.74, RSR = 0.51 and PBIAS = –1.10% for the calibration period and NSE = 0.55, RSR = 0.66 and PBIAS = 33.77% for the validation period (Table 7) Although an underesti-mation of the monthly sediment yield by the model for the validation period was within the satisfactory level of acceptance, it can generally be said that the simulated result was relatively satisfactory
From the results of the calibration and valida-tion, it is reasonable to conclude that the SWAT model could simulate the hydrology and sediment yield in this catchment well The calibrated para-meters were accepted for the scenario simulations
4.5 Response to climate change
In order to investigate the impact of climate change
on hydrological processes and sediment yield, the simulation was carried out using the land-use
Table 6 Model performance for the simulation of runoff.
Period Phuoc Long station Phuoc Hoa station
Time step NSE PBIAS RSR Period Time step NSE PBIAS RSR Calibration (1981 –1989) Daily 0.77 1.60% 0.48 Calibration (1981 –1989) Daily 0.86 –1.90% 0.37
Monthly 0.87 1.60% 0.36 Monthly 0.94 –1.90% 0.25 Validation (1990 –1993) Daily 0.79 3.30% 0.45 Validation (1990 –2000) Daily 0.71 –6.20% 0.54
Monthly 0.91 3.30% 0.30 Monthly 0.79 –6.20% 0.46
Fig 5 Observed and simulated daily flow hydrograph at
the Phuoc Hoa station: (a) calibration and (b) validation.