View
242
Download
0
Tags:
Embed Size (px)
DESCRIPTION
Flood
Citation preview
1
Quantitative risk measures of flood damage under the impacts of climate change
Quan Van Dau, Kittiwet Kuntiyawichai *
Water Resources Engineering, Department of Civil Engineering, Faculty of Engineering,
Khon Kaen University, Khon Kaen, 4002, Thailand
Abstract:
Flood is one of the most frequent hazardous natural disasters, which causes economic
depression, damages to livelihood and properties. Quantitative of the flood damage is
considered to be significant in minimizing potential losses from the natural hazard. In this
attempt, the paper aims to investigate the feasibility of assessing flood simulations and flood
damage that corresponds to the impacts of climate change in the central region of Vietnam.
The system was developed using the outputs of the Hadley Centre Coupled Model, version 3
(HadCM3) for A2 and B2 scenarios, a coupling of hydrological model (HEC – HMS) and
hydrodynamic model (HEC – RAS). In addition, the flood damage reduction analysis model
(HEC – FDA) was also considered for estimating flood damage. The findings indicated that
the weather will become hotter in the future with the temperature increasing from 0.4°C to
2.2°C and 0.19°C to 0.6°C correspond to A2 and B2 scenarios, respectively. Consequently,
the annual rainfall will also increase from 3.3% to 14.5% and 3.6% to 6.8% correspond to A2
and B2 scenarios, respectively. The results also demonstrated that runoff and water level can
potentially increase in the future and it might play a large role in destruction of crops,
transportation, water supply, and communities in the basin. The uncertainty of flood damage
costs was also estimated approximately $18,512.91 million to $19,148.59 million from the
decades of 2020s to 2080s. Additionally, the outcomes of this study can be applied for the
entire Huong River Basin in order to mitigate the flood problems in the future.
Keywords: Food damage, Climate change, Downscaling, HEC – HMS, HEC – RAS. * Corresponding author. Tel: +66 817 184 330; Fax: +66 4336 2142
Email address: [email protected]
1. Introduction
2
Flood damage is an essential issue in water engineering science and economic
development. It can be considered in term of directly flood damage such as resulting from the
physical contact with the flood water, and indirectly flood damage occurred on time and after
the event which is related to cost of alternative accommodation and loss of income .
According to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change
(IPCC), the global temperatures are likely to increase by 0.3°C to 4.8°C under the
Representative Concentration Pathway (RCP) 2.6 to 8.5 scenarios by the end of the century .
Flood damage might be occurred seriously when associated with high amount of carbon-
dioxide (CO2) concentration. In this sense, determining the flood damage would be very
significant in order to quantify the influences of flood that may threat in the future. By taking
the impacts of climate change in this paper, the General Circulation Model (GCM) is
considered as the most common source to evaluate the hydrological impact studies. Since,
GCMs remain limited as these coarse resolution and uncertainty, to capture many important
regional parameters. Thus, the statistical downscaling model is subsequently applied in this
study to transfer the coarse spatial outputs to the finer resolution of regional level.
Recently, there are several studies carried out regarding flood damage such as: Chen
(2007) used the hydrodynamic model HEC - RAS to develop an inundation mapping, while
integrated the depth-damage function to estimate the flood damage, where
is damage factor, is maximum damage in USD/m2; Mohamma di et a l. (2014)
estimated flood damage using flood damage reduction analysis HEC – FDA model in the
Neka River; Van Ooteg em et a l. (2015) constructed four multivariate damage models for
pluvial flood, as well as the Tobit-estimation technique on identifying victims; McGra th et a l.
( 2015) studied the sensitivity analysis of flood damage in Fredericton based on the U.S.
FEMA’s standardized methodology, etc.
In this paper, the potential impacts of climate changes in the central region of
Vietnam are investigated using the Statistical Downscaling Model (SDSM). In addition, to
evaluate the hydrological processes, a coupling of the hydrological (HEC – HMS) and
hydrodynamic models (HEC – RAS) are implemented. As the expected outcomes, flood
damage is also analyzed using the HEC - FDA model to quantify flood-risks in the future.
2. Study area
3
The Huong River Basin located in Thua Thien Hue Province, Vietnam was selected as
the case of study area. Geographically, it contains a total area of 2,830 km2 with a length of
104 km from the upstream of Thuong Nhat station to Thuan An outlet. There are three major
reservoirs: Huong Dien, Binh Dien and Ta Trach, which are situated along the tributaries of
the Bo, Huu Trach, and Ta Trach rivers, respectively. The river basin accounts almost 80% of
the entire area, it is mainly mountainous areas with its height varies from 200 m to 1,850 m in
comparison to the mean sea level (MSL). The river basin also accommodated the Tam Giang
– Cau Hai lagoon with 68 km of length, and 22,000 hectares of water surface areas.
The Huong River Basin is considered to have an abundant climate characteristic, with
the mean annual rainfall approximately from 2,500 mm to 3,500 mm vary from the coastal
areas to the hilly areas. The annual average temperature is measured approximately from
21°C to 26°C with the highest recorded temperature of 41.3°C at Hue City. The Huong River
Basin is known to be influenced by major of extreme natural disasters include of tropical
cyclones and flood as well as inundation at the downstream of the basin.
Fig. 1 The Huong River Basin, Vietnam (Thua Thien Hue Hydro-Meteorology and
Forecasting Center, Vietnam, 2014)
3. Data and Methods
3.1 Data collection
4
The meteorological data consist of daily rainfall, maximum and minimum
temperatures during 1976 to 2012 obtained from Thua Thien Hue Hydro-Meteorology and
Forecasting Center (TTH – HMFC). In additional, the observed discharge and water level
from 2009 to 2014 were also collected from the TTH – HMFC for hydrology - hydraulic
modeling simulations. As representative the topography for environmental modeling, the
spatial geo-referenced data set of the Digital Elevation Model (DEM) with 10 m x 10 m
resolution generated from 1:25,000 topographic contour was implemented. The Curve
Number (CN) values were also determined based on land use and soil maps generated in
2014 from the Department of Resources and Environment of Thua Thien Hue Province.
For global climate data, the daily predictors were derived from the National Centers
for Environmental Prediction (NCEP) re-analysis during the period of 1961 to 2001.
Moreover, the Hadley Centre Coupled Model, version3 (HadCM3) daily outputs for the
period 1961 to 2099 were obtained from the Met Office Hadley Centre, England. HadCM3
was chosen since there is unavailable any specific GCM models at present in Vietnam, the
selection based on the previous studies including the central region of Vietnam and the
Southeast Asia . Moreover, according to the World Bank, Vietnam population will increase
over 100 million people in the middle of the century and it tends to be higher in the end of the
21st century. Therefore, the high scenario (A2) and medium scenario (B2) were selected for
this study area. The A2 scenario is described a very heterogeneous world, the increasing of
the population and slow to economic; whereas, the B2 scenario emphasized on local solutions
to economic and increasing of the population .
3.2 Statistical Downscaling Model
The SDSM model is defined as multiple regression-based method, which was
introduced by Wil by et a l. (2002) . The model was developed to convert the resolution from
large scale predictor variables into local climate variables. Initially, the NCEP reanalysis
predictors were selected to identify sensible relationships between the predictor variables and
observed data. A correlation and partial correlation analyses were also applied to investigate
inter-variable correlations of the potential utilization of predictor and predictand (rainfall,
temperature) relationships. Regarding calibration, rainfall was modeled as a conditional
process because the amounts of rainfall depend on wet-day occurrence while the temperature
was modeled as an unconditional process. Alternatively, the transformation of the “fourth
roots” was used to account for the skewed nature of the rainfall distribution. It suggested that
the climatological baseline period for 30-year such as the period of 1961 to 1990. However,
because of the lack of observation data, the period of 1976 to 1990 was used as a baseline
5
period in this study. The model was simulated for daily rainfall, maximum and minimum
temperatures, i.e. the period of 1976 to 1990 for calibration and 1991 to 2001 period for
validation.
3.3 Hydrology model HEC – HMS
The Hydrologic Engineering Center – Hydrologic Modeling System (HEC – HMS)
model was developed by the U.S. Army Corps of Engineer. It was used to utilize the rainfall
– runoff processes of river basins and it was widely applied in many hydrological studies. To
represent the behavior of the river basin, HEC – HMS contains runoff volume, direct runoff,
baseflow and channel flow .
There are several loss methods that are designed in the HEC – HMS model to
calculate actual infiltration from sub-basins. Among the remaining loss methods, the Soil
Conservation Service - Curve Number (SCS – CN) method was used for the event-based
simulation due to its simplicity manner. In addition, the SCS – CN method provides better
results in comparison to the other methods in this study area. The Snyder Unit Hydrograph
(SUH) is well known as a common method used to define how excess rainfall is transformed
into runoff in sub-basins. SUH calculated runoff by utilizing a standard unit hydrograph,
which comprises of five characteristics, i.e. peak discharge, basin lag, basin time, and the
widths of the unit hydrograph at 50% and 70% of peak discharge . This method has recently
been proposed by in the same river basin, and it was selected in this study area. For
calculating the actual subsurface flow from sub-basins, the recession baseflow method was
considered due to its capability to automatically reset after each flood event; and this method
does not conserve mass for each sub-basin. Recession baseflow is broadly practiced in many
hydrological studies and it can be applied for both event and continuous simulations.
Alternatively, the Muskingum routing method was also applied to represent a segment of a
river. This method is examined as one of the most popular methods at present, since it used a
simple conservation mass method to route flow along the channel.
Moreover, there were three stations selected to represent the hydrological processes
applied to model evaluation, i.e. Phu Oc station, Binh Dien and Ta Trach reservoirs. At Phu
Oc station, due to lacking of observed flows in the Bo River, the observed daily water level
data measured at Phu Oc station for the period of 2009 to 2010 was used for calibration, and
2011 to 2012 period was used for validation. The observed daily inflow to Binh Dien
reservoir during the period of 2010 to 2011 was used for calibration and the year 2012 was
used for validation. For Ta Trach reservoir which was just operated in May 2014, the
6
available observed inflow data for the period of September to October 2014 was used for
calibration, while the period of November to December 2014 was used for validation.
3.4 Hydrodynamic model HEC – RAS
The Hydrologic Engineering Center – River Analysis System (HEC – RAS) is a
hydrodynamic model developed by the U.S. Army Corps of Engineers. It was designed for
flood prone determination and hydraulic analysis of river channel. The model includes four
river analysis components such as the sediment transport, water quality, steady flow and
unsteady flow .
To carry out a model performance, the unsteady flow analysis was implemented in
HEC – RAS model. Simulation of unsteady flow uses dynamic Saint-Venant equation to
simulated one dimensional flow through the channel. The Manning’s n values were adjusted
vary from 0.12 to 0.15 for the left and right of the river banks, and from 0.032 to 0.04 for the
main of channel. In addition, the discharges at the outlets of Huong Dien, Binh Dien and Ta
Trach reservoirs were utilized as the upstream boundary conditions, and tidal water level at
Thuan An outlet was employed as the downstream boundary condition. Two streamflow
stations located at the downstream of the river basin include of Kim Long and Phu Oc
stations, were chosen to evaluate the model performances. The period of 2009 to 2010 was
used for calibration and 2011 to 2012 period was used for validation.
3.5 Risk-based analysis of flooding
To carry out the preliminary risk-based analysis of flooding in the Huong River Basin,
the Hydrologic Engineering Center - Flood Damage Reduction Analysis (HEC - FDA) model
developed by the U.S. Army Corps of Engineering was considered. HEC - FDA provided
state-of-the-art analysis to formulate and evaluate flood damage reduction, it used the risk-
based analysis to compute expected annual damage (EAD) . EAD was defined as the mean
damage derived from the damage-exceedance probability function as shown in Fig.2. To
compute EAD, the uncertainty of discharge-exceedance probability, state-discharge and
damage-stage functions were quantified and associated with economic performance analysis .
Mathematically, EAD can be calculated using Equation 1.
Discharge (m3/s)
Stage (m) Probability
Damage ($)
III
III IV
EAD
7
Fig 2 Expected annual damage computation sensitivity analysis
(1)
Where: D is the damage in year; f(D) is the probability density function
Furthermore, to perform the interaction between the hydrology - hydraulic and
economic relationships, the Monte Carlo simulation was applied. The Monte Carlo
simulation method was defined as the numerical integration approach, it used the input
probability and function uncertainties to determine a value of flood damage for many
possible scenarios. Based on the hydrodynamic model and flood inundation mapping, water
surface profiles (WSP) for the river basin was developed, which consists of eight profiles (i.e.
0.50-, 0.20-, 0.10-, 0.04-, 0.02-, 0.01-, 0.004-, and 0.002-exceedance probability flood
events). In this study, WSP was derived from the hydrodynamic model HEC - RAS include
of discharges and stages-based simulated with the unsteady flow analysis.
A discharge - probability function was developed using a graphical exceedance
probability method from WSP. The graphical method was selected since it is widely used for
managed flow conditions, particularly well-suited for the unsteady flow analysis. This
method was relied on the “equivalent record length and error limit curves” using the “less
simple” approximation to the Order Statistics method, random sampling of the graphical
relationship can be determined. A stage-discharge function was performed to transform the
flow into water level for each exceedance probability. A normal-distribution type of the
probability density function (PDF) and the standard deviation (SD) were applied. A damage-
stage function represents the relationship between depths of water to the cost of damages
were developed by classifying into five categories (e.g. agriculture, industry, infrastructure,
irrigation and resident). The classification and structure inventory development based on the
observed data taken from Hinh (2009) . In addition, the flood damage in 2009 was assigned
8
as the base structure module for model simulation. The model was operated separately for the
difference periods during the decade of 2020s, 2050s, and 2080s.
3.6 Model evaluation criteria
For the assessment of model performances, there are four statistical measures for
judging the goodness of fit of model simulations were used in this study include
determination coefficient (R2), Nash – Sutcliffe coefficient (ENS), Root Mean Square Error
(RMSE), and Percent Error in Peak (PEP).
(2)
(3)
(4)
(5)
Where: is observed data; is mean of observed data; is simulated data; is mean of
simulated data; n is total number of variable; is simulated peak; is observed peak
4. Results and discussions
4.1 Evaluation of the statistical downscaling model
In statistical downscaling, the finding of sensible predictors is an essential part and is
the most time consuming step. According to Wil by et a l. (2007) , the predictors contain higher
correlation value implies a higher degree of association. In fact, rainfall downscaling is
necessarily more problematic compared with the temperature due to its coarse spatial
resolutions, poor resolved by regional-scale predictors, and containing many regional driving
factors. For instance, the results expressed that relative correlation for the mean temperature
at 2 m was given of 0.7, whereas, it indicated an underestimated for daily rainfall with
correlation values donated of 0.1. The obtained findings can also be collated to Shrest ha et a l.
(2014) , and Hu et a l. (2013) , where downscaling of rainfall is an argumentative but still fairly
representative. Nevertheless, the results demonstrated a good outcome for mean monthly
9
rainfall and temperatures as exhibited in Fig. 3. The R2 and RMSE values for rainfall during
the calibration and validation also ranged between 0.88 to 0.98, and 3.72 to 1.92,
respectively. In view of maximum and minimum temperature, the R2 and RMSE values
ranged between 0.88 to 0.99, and 1.06 to 0.05, respectively. Based on the statistical indices, it
can be reasoned that SDSM was performed well in this study.
Fig 3 The calibration results for (a) mean monthly rainfall and (b) mean monthly
temperature during 1976 – 1990 at Hue station.
4.2 Projection of future climate for rainfall and temperature
The projection for future mean annual rainfall indicated increase trend under
consideration with the A2 and B2 scenarios in the Huong River Basin (Fig 4). The results
implied that rainfall tends to increase during the summer season from January to August,
particularly, it showed higher in February with 60% increasing for A2 scenario and 37%
increasing for B2 scenario; meanwhile, mean monthly rainfall seem to be decrease about 7%
– 12% in April for the both scenarios. During the monsoon season, rainfall is expected to
decrease by an approximately 2.4% to 8.5% in October, and possibly increase by 3% to
16.8% in November to December. As the investigation indicates, the overall findings
suggested that the mean annual rainfall will increase between 3.3% to 14.5% for A2 scenario
and 3.6% to 6.8% for B2 scenario from the decades of 2020s, 2050s and 2080s in comparison
to the baseline period.
a) b)
a) b)
10
Fig. 4 Mean monthly rainfall changes for the decades of 2020s, 2050s and 2080s in
comparison to the baseline period under (a) A2 scenario and (b) B2 scenario
The results also interpreted that temperatures will increase and the weather will be
hotter in the future. As illustrated in Fig. 4, it can be observed that mean monthly temperature
changes for the A2 scenario is higher compared with the B2 scenario. Particularly,
temperature suggested to increase in January and May between 0.5°C to 2.2°C, and slightly
decrease in March and July between 0.2°C to 0.4°C. The study also exposed an increment of
temperature during the monsoon season by 5.3°C and 1.68°C for A2 and B2 scenarios,
respectively. It can be concluded that the mean annual temperature will increase in the future
with a range of 0.4°C to 2.2°C and 0.19°C to 0.6°C under the emissions of A2 and B2
scenarios, respectively.
Fig. 4 Mean monthly temperature changes for the decade of 2020s, 2050s and 2080s in
relation to the baseline period under (a) A2 scenario and (b) B2 scenario
4.3 Hydrologic HEC –HMS model evaluation
The calibration and validation processes were applied for the three stream stations
(i.e. Ta Trach, Binh Dien and Phu Oc stations) to prove the reliability and stability of the
model. As showed in Fig.5, the results performed a good achievement between observed and
simulated in terms of trends and magnitudes, except a certain period of November 18-19,
2011 at Ta Trach station, which illustrate a clear difference due to the effect of tropical
northeast monsoon as reported by . The outcomes also gave the PEP value ranges from 1% to
8.5%, 1% to 10% and 14% to 20% during the calibration and validation periods for Ta Trach,
Binh Dien and Phu Oc stations, respectively. Based on the ENS and R2 statistical values as
exhibited in Table 1, it can be evinced that HEC – HMS performed well during the
simulation.
a) b)
11
c) d)
a) b)
12
Fig. 5 The HEC – HMS model calibration results (a, c, e) and validation results (b, d, f) at Ta
Trach, Binh Dien, and Phu Oc stations, respectively
Table 1 Statistical indices of HEC – HMS model performance
StationCalibration Validation
R2 ENS R2 ENS
Phu Oc 0.85 0.84 0.85 0.78Binh Dien 0.85 0.83 0.84 0.77Ta Trach 0.94 0.91 0.91 0.89
4.5 HEC – RAS model for flood mapping
A total 79 cross sections were imported into HEC – RAS model include observed data
and DEM with 10m x 10m resolutions. The results clear exposed that HEC – RAS model
performed well in simulating water levels at both Phu Oc and Kim Long stations during the
calibration and validation processes (Fig.6). Despite the correlation values for validation
process at Phu Oc station was rather low due to the effect of unprecedented levels of
overflow from the Bo River which influenced its flow pattern. However, based on the
statistical indices summarized in Table 2, it was considered that HEC – RAS model is
suitable to simulate flood regimes in the Huong River Basin.
e) f)
13
Fig. 6 The HEC – RAS model calibration results (a, c) and validation results (b, d) at Kim
Long and Phu Oc stations, respectively
Table 2 Statistical indices of HEC – RAS model performance
StationCalibration Validation
R2 ENS R2 ENS
Phu Oc 0.74 0.73 0.67 0.64Kim Long 0.78 0.77 0.79 0.74
4.6 Flood simulation for future emission scenarios
As representative worst case projected high of CO2 concentration, the A2 scenario
was selected for flood simulations. The results manifested trends in increasing water level
and flow magnitude at Phu Oc and Kim Long stations under the impacts of climate change
(Table 4). Increasing water level and discharge can logically lead to increase flood
occurrences and risk damages in the Huong River Basin in the future. Furthermore, in order
to identify the flood prone areas for the basin, the flood inundation mapping in the decades of
2020s, 2050s, and 2080s were delineated as presented in Fig 7.
Table 4 Maximum water levels and discharges under A2 scenario
Period Phu Oc Kim Long
a) b)
c) d)
14
(m+MSL) (m3/s) (m+MSL) (m3/s)
2020s 1.39 224.53 0.83 434.522050s 1.58 267.67 0.87 462.992080s 2.01 380.88 1.15 604.03
Where: is the maximum flood level (m+MSL); is the maximum discharge at
selected cross section (m3/s)
Fig. 7 Flood inundation extent of the Huong River Basin based on (a) satellite imagery
taken on 11th June 1999 (b) HEC – RAS delineated flood inundation extent for the
decade of 2020s (c) HEC – RAS delineated flood inundation extent for the decade of
2050s (d) HEC – RAS delineated flood inundation extent for the decade of 2080s
under A2 emission scenario
4.7 Flood damages estimation
From the flood inundation mapping, the flooded areas were identified and taken into
account to quantify the flood damage costs. Consequently, the sensitivity analysis of the
relationship between discharge-exceedance probability, stage-discharge, damage-stage, and
mean damage-exceedance probability were generated using the Monte Carlo simulation as
presented in Fig. 8 below.
a) b)
c) d)
15
Fig. 8 EAD computation sensitivity analysis for the relationships of (a) discharge-
exceedance probability function (b) stage-discharge function (c) damage-stage
function (d) mean damage-exceedance probability at Tuan-Kim Long reach
Based on the outcome of study, the results demonstrated an agreement of the flood
damage costs between observed and modeled in 2009 (Table 5). It also can be realized that
the flood damages for the infrastructure type was ranked as the highest damages compared
with the other types of damages. Furthermore, the findings expressed an increment trend of
the flood damage costs approximately from $18,512.91 million to $19,148.59 million
corresponds to the decades of 2020s to 2080s, respectively.
Table 5 Flood damage estimation for the Huong River Basin under future emission
scenarios
Damage categories
Observed($1,000)
Modeled($1,000)
2020s($1,000)
2050s($1,000)
2080s($1,000)
Resident 3,252.53 3,255.38 3,127.33 3,211.12 3,311.88
Agriculture 4,149.77 4,146.78 3,977.73 4,003.60 4,093.45Irrigation 4,261.93 4,252.97 4,100.86 4,136.86 4,243.04Infrastructure 6,168.58 6,155.14 5,960.83 5,990.97 6,091.23
Industry 1,401.95 1,406.78 1,346.16 1,355.20 1,408.99
a) b)
c) d)
16
Total 19,234.76 19,217.05 18,512.91 18,697.75 19,148.59
Note: 1 USD = 17,800 VND in 2009
5. Conclusions
In this paper, an assessment of flood simulation as well as flood damage was carried
out in the Huong River Basin. The results explored that HadCM3 model provided a realistic
finding in estimating the potential impacts of climate change in this study area. From the
performed research, it is possible to conclude that the application of hydrological model HEC
– HMS and hydraulic model HEC – RAS are suitable for the Huong River Basin.
Alternatively, the flood damage reduction analysis HEC – FDA was also demonstrated good
performances in quantitative flood damage costs. From the main finding of this study, the
results suggested that the weather will be hotter in the future, which ranges between 0.4°C to
2.2°C and 0.19°C to 0.6°C for A2 and B2 scenarios, respectively. Moreover, the annual
rainfall will also increase from 3.3% to 14.5% and 3.6% to 6.8% correspond to A2 and B2
scenarios, respectively. Increasing of temperature and rainfall will significantly influence
discharges and water levels in the basin, which might cause negative impacts, including loss
of human life, damage to property and destruction of crops, etc. As the results suggest, the
quantitative of flood damage costs were estimated approximately from $18,512.91 million to
$19,148.59 million in the future. Finally, the outcomes of this research can be used as a
guideline for adaptation measures corresponding to climate changes in the other parts of
Vietnam.
Acknowledgements
The author would like to acknowledge the generous financial support of the Khon
Kaen University Scholarship for ASEAN and GMS Countries’ Personnel program. The
author’s thanks are also extended to the Thua Thien Hue Hydro-meteorology Forecasting
Center for its support in providing the meteorological data.
References
Chen J. 2007. Flood Damage Map for the Huong River Basin. In: University of Twente, The Netherlands Faculty of Constructing Sciences, Civil Engineering, Hanoi University of Science - VNU, pp: 46.
Chow VT, Maidment DR, Mays LW. 1998. Applied Hydrology. In: Civil Engineering Series, McGraw-Hill International Editions, Assaye Melaku, pp: 733-742.
17
Dinh N. 2014. Research on changes of some hydraulic and hydrological factors of the Huong River system’s downstream under the impact of irrigation and hydropower constructions and climate change. In: Science Technology and Enviromental pp: 17.
Hinh VN. 2009. Typhoon Ketsana (09) damaged to Thua Thien Hue Provinces approximately 343 billion VND
Hoa P. 2013. Flood forecasting in Huong River Basin using the HEC-HMS model. Thua Thien Hue-Hydro-meteorology and Forecasting Vietnam.
Hu Y, Maskey S, Uhlenbrook S. 2013. Downscaling daily precipitation over the Yellow River source region in China. In: Theor Appl Climatol.
IPCC. 2000. Methodological and Technological Issues in Technology Transfer. Bert Metz OD, Jan-Willem Martens, Sascha Van Rooijen and Laura Van Wie Mcgrory (ed.) Cambridge University Press, UK, pp: 432.
IPCC. 2013. Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. pp: 1535.
Khoi D, Hang P. 2015. Uncertainty Assessment of Climate Change Impacts on Hydrology: A Case Study for the Central Highlands of Vietnam. In: Managing Water Resources under Climate Uncertainty, Shrestha S, Anal AK, Salam PA, van der Valk M (eds.) Springer International Publishing, pp: 31-44.
Mai DT. 2009. Flood prediction model for Huong River and Vu Gia-ThuBon River basins in Central Vietnam. Vrije Universiteit Brussel, Belgium, pp: 210.
McGrath H, Stefanakis E, Nastev M. 2015. Sensitivity analysis of flood damage estimates: A case study in Fredericton, New Brunswick. In: International Journal of Disaster Risk Reduction.
Mina F, Yin TP. 2015. GIS Techniques for Flood Modeling and Flood Inundation Mapping. In: EJGE, pp: 6795-6805.
Mohammadi SA, Nazariha M, Mehrdadi N. 2014. Flood Damage Estimate (Quantity), Using HEC-FDA Model. Case Study: The Neka River. In: Procedia Engineering, pp: 1173-1182.
Shrestha S, Deb P, Bui T. 2014. Adaptation strategies for rice cultivation under climate change in Central Vietnam. In: Mitig Adapt Strateg Glob Change, Springer Netherlands, pp: 1-23.
USACE-HEC. 2000. Hydrologic Modeling System HEC-HMS. In: Technical Refencence Mannual Feldman AD (ed.) U.S. Army Corps of Engineers, pp: 155.
USACE-HEC. 2010. HEC-RAS, River Annalysis System Hydraulic Reference Mannual. Gary WB (ed.) US Army Corps of Engineers Hydraulogic Engineering Center, pp: 411.
USACE-HEC. 2014. HEC-FDA Flood Damage Reduction Analysis. In: User's Mannual, CEIWR-HEC (ed.) US Army Corps of Engineers Institute for Water Resources Hydrologic Engineering Center (HEC), pp: 378.
Van Ootegem L, Verhofstadt E, Van Herck K, Creten T. 2015. Multivariate pluvial flood damage models. In: Environmental Impact Assessment Review, pp: 91-100.
VTC14. 2014. Weather forecasting on November 17, 2014: Rainfall in the Central region Vietnam.
Wilby RL, Dawson CW, Barrow EM. 2002. sdsm — a decision support tool for the assessment of regional climate change impacts. In: Environmental Modelling & Software, pp: 145-157.
Wilby RL, Dawson CW, Barrow EM. 2007. Using SDSM version 4.2 – A decision support tool for the assessment of regional climate change impacts. .
18