INTRODUCTION
General Background
The Burmese proverb, “Rice offers survival for seven days; however, water gives life only a day,” emphasizes the vital importance of water for all living beings Despite being a renewable resource, effective management and strategic exploitation of water are crucial for maintaining a sustainable supply The United Nations Sustainable Development Goal No 6, specifically target indicator 6.4, underscores the essential role of water in promoting efficient usage and addressing water scarcity (UNHCR, 2017).
Vietnam is rich in water resources with dense river networks accompanied by
Vietnam features 2,360 rivers, which can be classified into three primary categories based on basin area and administrative boundaries: those spanning multiple provinces, two provinces, and a single province (Taylor & Wright, 2001) The country is home to nine major river basins—Red River, Thai Binh, Bang Giang – Ky Cung, Ma, Ca La, Thu Bon, Ba, Dong Nai, and Mekong—that collectively cover approximately 80% of Vietnam's total land area (ADB).
2009) The average annual surface water runoff is approximately 830 – 840 bn m 3 in which only 43% is available for sustainable exploitation (2030WRG, 2017).
Vietnam has a groundwater potential of 63 billion cubic meters, yet only 7% of this can be sustainably utilized (FAO AQUSTAT, 2011) The sustainability of Vietnam's water resources is heavily influenced by transboundary water management, as 63% of the country's total water resources originate from neighboring countries, including China, Cambodia, Lao PDR, and Thailand (MONRE, 2006 & 2030WRG, 2017).
The Mekong and Red-Thai Binh basins are significant contributors to Vietnam's economy, accounting for approximately 42% of the country's GDP To ensure sustainable economic growth, effective long-term water resource management strategies are essential for these basins, particularly due to their high dependency on trans-boundary water sources—95% for the Mekong and 40% for the Red-Thai Binh Basin.
Vietnam's agricultural sector significantly contributes 18% to the country's GDP while utilizing approximately 80% of its total water resources Agriculture occupies one-third of Vietnam's land area, with 20.6% dedicated to farming, 12% for permanent crops, and 2.1% as permanent pasture Key crops include paddy, maize, coffee, and sugarcane, with the Mekong Delta, Red River Delta, and Northern Highlands recognized as the nation's primary rice-producing regions The agricultural master plan aims to maintain paddy fields at no more than 3.8 million hectares, targeting a production increase to 41-43 million tons per year by 2020 and 44 million tons by 2030 This underscores the substantial water resources needed to support both current and future agricultural production.
Annual water demand in Vietnam is increasing year by year; 80.2 bn m 3 /year in
2009, 80.6 bn m 3 /year in 2015 and projected to reach 95 bn m 3 /year by 2030 (ADB,
In 2016, the agricultural sector's water demand was 76 billion cubic meters, with projections indicating an increase to 91 billion cubic meters by 2030 Aquaculture required 10 billion cubic meters per year in 2016, expected to rise to 12 billion cubic meters annually by 2030 The industrial sector's water demand was 6 billion cubic meters in 2016, forecasted to grow to 15.6 billion cubic meters by 2030 Municipal water requirements stood at 3.1 billion cubic meters per year in 2016, anticipated to increase to 5.7 billion cubic meters annually Additionally, the hydropower sector's demand was 57 billion cubic meters in 2016, projected to reach 63 billion cubic meters by 2030.
Vietnam's wastewater treatment system is still in development, currently able to treat only 12-13% of municipal wastewater and 10% of industrial wastewater This has led to significant pollution in three major river basins: the Cau River Basin, Nhue-Day River Basin, and Dong Nai River Basin In 2015, the Ministry of Natural Resources and Environment (MONRE) amended technical guidelines for surface water quality to enhance water resource management and quality control MONRE now conducts quarterly water quality monitoring in the Cau River Basin, with results publicly disclosed through an index score.
Climate change significantly impacts watershed hydrology, with projections indicating that by 2030, annual runoff could increase to 15 billion cubic meters per year This includes a predicted rise of 25 billion cubic meters during the wet season and a decrease of 10 billion cubic meters in the dry season, highlighting alterations in rainfall patterns that affect seasonal water availability Key climatic factors, particularly precipitation and temperature, play a crucial role in the sustainability of water resources.
Addressing water demand and resources is essential in light of projected population growth and the needs of various industries and agriculture Employing integrated approaches can enhance both the quantitative and qualitative management of water resources Effective management of existing water supplies is vital to ensure a safe and reliable water source that supports multi-sectoral development.
Research Motivation
The Red-Thai Binh region is vital to Vietnam's socio-economic development, contributing around 15% to the national GDP Its strategic geographical location, spanning from the northern highlands to the alluvial plains in the northeast, makes it an ideal business hub The basin is responsible for 15% of the country's total rice production, playing a crucial role in national food security and generating foreign income It also hosts a significant concentration of industrial clusters, craft villages, and intensive agricultural fields, accounting for 65% of the nation's total To enhance the quality of life and ensure sustainable water resource management, Mr Tran Hong Ha, Minister of the Ministry of Natural Resources and Environment, has called for a comprehensive water resource exploitation plan involving scientists and experts.
Target Basin
The Upper Thai Binh River Basin (UPTBRB) is located in Northern Vietnam and is part of the larger Red-Thai Binh River Basin, which spans nearly the entire northeastern region of the country This transboundary river basin is shared among three countries: China (48%), Laos PDR (1%), and Vietnam (51%) Notably, the basin contributes 17% of Vietnam's total annual surface water runoff.
Mekong River Basin with 59% thus play an essential role to provide the sufficient water resources for the capital Hanoi (2030WRG, 2017).
Figure 1.1 Location of Target Basin (Source: TA-7629, VIE, 2012)
Problem Statements
The UPTBRB faces a significant risk of water stress due to escalating demand from multiple sectors, with the 2030 Water Resources Group projecting a 42% increase in overall water demand in the Red-Thai Binh Basin by 2030 Previously categorized as low water stress in 2016, the basin is expected to transition into the water-stressed category by 2030, intensifying competition for consistent access to water resources in the near future.
The UPTBRB faces significant risks from climate change, with projections indicating an average surface temperature rise of 1.9 to 2.4°C by 2100 under RCP 4.5 scenarios Additionally, precipitation levels are anticipated to increase by 5 to 15%, with changes in intensity and patterns influenced by geographical factors (Ngu et al., 2016).
The Cau River is recognized as the most polluted watershed in Northern Vietnam, highlighting significant water quality challenges due to spatial and seasonal variations (Tran et al., 2017) A comprehensive basin-wide assessment is essential to determine the pollution status, particularly between the main rivers and their tributaries Additionally, the absence of a water quality database in the Thuong and Luc Nam River Sub-basins presents substantial obstacles for effective water resource management.
Objectives
The main objectives of this study are as below;
- To evaluate basin water stress through integration of hydrological model and social statistics
- To evaluate basin river quality by multivariate statistical analysis and the
- To contribute integrated water resource management through quantitative and qualitative approaches
Thesis Structure
This thesis adheres to the Master's thesis guidelines set forth by Vietnam Japan University, with a structured format that includes a summary of the general research information and corresponding chapters.
Chapter 1 provides an overview of Vietnam's water resources, detailing the current water demand across various sectors and the initiatives undertaken by relevant organizations in water resource management It highlights key challenges related to water and outlines research objectives driven by these issues Additionally, baseline data on the natural and hydro-climatic conditions of the target basin is presented.
Chapter 2 focuses on a literature review that outlines the research scope and content, summarizing previous studies and key findings relevant to the basin It extensively discusses comparative studies that utilize similar research methodologies in other basins, serving as valuable references for this research.
Chapter 3 describes about hydrological modelling in the basin Detailed step by step method to construct the watershed model, technical considerations and results are provided.
Chapter 4 provides a comprehensive analysis of climate projections using regional climate models to predict future conditions It examines historical, current, and future scenarios, offering valuable insights into climate change impacts within the basin.
Chapter 5 addresses water stress by estimating sectoral water demand and resources, utilizing findings from Chapters 3 and 4 This section integrates various assessment methods to evaluate both current and future water stress levels.
Chapter 6 provides river water quality evaluation Seasonal river water sampling and laboratory analysis are performed The results are driven by multivariate statistical analysis and the Vietnamese regulations.
Chapter 7 is a conclusion and recommendation session in which all the key findings and necessary solutions are provided including limitations and future research direction.
Baseline Information about the Study Basin
The Upper Thai Binh River Basin (UPTBRB) consists of three primary rivers: the Cau River, Thuong River, and Luc Nam River, which together create three distinct sub-basins The Cau River Sub-basin (CRSB) features seven tributaries, including Cho Chu, Thuong Nghinh, Du, Cong, and Calo rivers, while the Luc Nam River Sub-basin (LNRSB) has the Dinh Dem river as its sole upstream tributary The Thuong River Sub-basin (TRSB) also includes only one tributary, the Rang river The Cau River, the longest at 288 km, originates in Bac Kan Province, followed by the 157 km long Thuong River flowing from Lang Son to Bac Giang Province, and the 200 km Luc Nam River, which is vital to Bac Giang Province All three rivers converge at Pha Lai, contributing to a rich and complex river network entirely within the country's boundaries.
Geographic location of the basin extends from 21°48'54.81"N to 106°27'44.98"E.
The basin spans an area of 12,720 km², representing approximately 4% of the total land area of the country It is bordered by dense high mountain ranges in the North-East, with the highest peak being Tam Dao Mountain at 1,592 m and the lowest point at Pha Lai, where the Thai Binh River originates, at 3.8 m The central and downstream regions of the basin consist of fertile alluvial plains, which provide excellent conditions for agriculture and settlement This basin encompasses several administrative provinces, including Bac Kan, Thai Nguyen, Hanoi, Vin Phuc, Bac Ninh, Lang Son, and Bac Giang.
2015, the basin population is estimated at approximately 4.95 million The average basin population density is 389 persons per km 2 that is higher than the country’s 280 person per km 2
The basin experiences a tropical monsoon climate characterized by an annual maximum temperature of approximately 28 °C and a minimum of 21 °C, resulting in a temperature gap of about 6°C The long-term average annual precipitation is around 1550 mm, with the heaviest rainfall occurring in July and August, exceeding 300 mm per month The region has four distinct seasons: winter (December to February), spring (March to May), summer (June to August), and autumn (September to November) Notably, from 2005 to 2019, approximately 67% of total runoff was recorded during the rainy season from June to September.
Figure 1.2 Long Term Annual Maximum Temperature Trend
Figure 1.3 Long Term Annual Minimum Temperature Trend
Figure 1.4 Long Term Monthly Precipitation Trend
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Figure 1.5 Average Monthly Discharge at the Gia Bay Hydrological Station
Summary
The annual rise in water demand is significantly influenced by seasonal variations and climate change, while inadequate water treatment facilities jeopardize river water quality and socio-economic development To enable water stakeholders to implement essential practices and measures, it is crucial to assess water resources from both quantitative and qualitative perspectives Figure 1.6 illustrates the relevant strategies to meet these specific objectives.
Figure 1.6 Work Flow of the Research
LITERATURE REVIEW
Administrative Provinces
Bac Kan province, situated in the mountainous region of Northern Vietnam, has a population of approximately 303,100 as of 2013, with a density of 64 people per square kilometer The province's economy is primarily driven by agriculture, although forests cover a significant portion of its land Bac Kan experiences a warm and temperate climate, characterized by high annual precipitation exceeding 2,294 mm and an average annual temperature of 20.6 °C.
Thai Nguyen province is located in the center of North East Region The province area is 3526.64 km 2 that exists together with Bac Kan Province in the North In
2019, it has a population of about 1.29 million while population density is 365 persons per km 2 The average temperature is 25 °C and the recorded average precipitation varies
2000 to 2500 mm In 2010, industrial clusters were rapidly established thus the main economy of the province had shifted to industry-construction from agriculture (Thai Nguyen Overview, 2015).
Vinh Phuc Province covers an area of 1,237.52 km² and had an average population of 1.03 million as of 2013, resulting in a population density of 832 people per km² The region experiences an average temperature ranging from 24 to 25 °C and receives approximately 1,650 mm of precipitation annually Notably, agriculture dominates the local economy, utilizing around 70% of the total land area.
Hanoi province, situated at the downstream of the Cau River and at the confluence of the Red River and Da River, is a bustling metropolitan area with a population of 4.7 million Covering a total area of 319.56 km², it boasts a high population density of 14,639 people per km².
11 temperature is 29.2 °C and the average precipitation is 1800 mm (Hanoi Geography, 2014).
Bac Ninh, located just 30 km northeast of Hanoi, is the smallest province in Vietnam, covering an area of 823 km² As of 2015, it has a population of approximately 1.54 million, resulting in a population density of 1,676 people per km² The province experiences an average temperature of 23.6 °C and receives about 1,926 mm of precipitation annually Bac Ninh is renowned as an industrial hub, particularly for its foreign investment and trade opportunities.
Lang Son is a picturesque mountainous province in Vietnam, situated to the north of Quang Ninh Province, south of Bac Giang Province, and west of Bac Kan Province Covering an area of 8,310 square kilometers, Lang Son has a population of approximately 831,887, resulting in a population density of 102 individuals per square kilometer.
In 2009, the primary land covers in the province were forest and agriculture, reflecting its tropical monsoon climate The average temperature ranged from 17 to 22 °C, with annual precipitation levels between 1200 and 1600 mm, according to the Lang Son Statistics Office (2021).
Bac Giang Province, covering an area of 3,895.59 km², is bordered by Lang Son Province to the north and Bac Ninh Province to the west As of 2019, it has a population of approximately 1.8 million, resulting in a density of 483 individuals per km², with 72% of its area consisting of mountainous districts The province experiences an average annual temperature of 23.3°C and receives about 1,915 mm of rainfall each year The industry and construction sector is the largest contributor to the province's GDP, followed by agriculture (Wu A., 2021).
Hydrological Modelling
Hydrological modelling, as defined by Allaby & Allaby (1999), involves characterizing real hydrologic features and systems through the integration of physical models, mathematical equations, and computer simulations This process aids in understanding watershed hydrology by accurately depicting runoff over time, both on land and underground, and assessing the quantity of water stored in the soil.
Hydrological models can be categorized into lumped and distributed types, with lumped models ignoring spatial variability, while distributed models break the catchment into smaller units Additionally, models can be classified as static or dynamic based on time factors Prominent examples in hydrological studies include empirical models like Artificial Neural Networks (ANN), conceptual models such as the HBV model and TOPMODEL, and physically based models like MIKESHE and SWAT (Devia et al., 2015).
The Soil and Water Assessment Tool (SWAT) is widely recognized for its effectiveness in studying watershed hydrology in the context of climate change, with 36 significant hydrological studies conducted in Vietnam (Tan et al., 2019) Research by Tran et al (2017) highlighted the notable correlation between rainfall and daily total nitrogen load in the Cau River Basin, while Thai et al (2017) utilized SWAT to predict future soil loss and increased river discharge due to climate change impacts in the Cau watershed To address data scarcity, Bui et al (2019) integrated SWAT and QUAL2K for water quality modeling in the Cau watershed, and Chuong et al (2014) evaluated water quality using SWAT in the Ta Trach watershed of central Vietnam Additionally, S Shrestha et al (2018) combined SWAT with climate models in the Songkhram River Basin, Thailand, revealing that stream flow was influenced by future climate, and nitrate-nitrogen levels were closely linked to land cover Furthermore, S Shrestha et al (2016) explored water resource potential in the Indrawati River Basin, Nepal, finding that river discharge patterns varied significantly with future climate scenarios.
Climate Change
Vietnam has experienced notable climate change, with an annual temperature rise of 0.62 °C from 1958 to 2014, averaging an increase of 0.1 °C per decade Specifically, a significant increase of 0.42 °C was recorded in 1985.
– 2014 In the North-East region, the increased days of maximum temperature greater than 35 °C is observed during 1960 – 2014 (Ngu et al., 2016) The annual changes of temperature were provided in figure 2.1.
Figure 2.1 Annual Changes of Temperature (Ngu et al., 2016)
Over a 57-year period from 1958 to 2014, Northern Vietnam experienced a decline in annual precipitation by 5.8% to 12.5%, despite a slight increase in overall precipitation at the country level In contrast, the South Central and Southern regions saw significant increases in precipitation, rising by 19.8%, 6.9%, and 18.8%, respectively Seasonal variations indicate a decrease in autumn precipitation in the North, while spring receives substantial rainfall; conversely, the Southern regions show increased precipitation during winter and spring (Ngu et al., 2016).
Table 2.1 Variation of Precipitation (%) during 1958 – 2014
Climatic Regions (MAR – (JUN – (SEP – (JAN – Annual
Climate change projections for Vietnam were developed using five regional climate models, including the Weather Research and Forecasting (clWRF) and the Conformal Cubic Atmospheric Model (CCAM) The analysis considered a long-term baseline period from 1986 to 2005, forecasting future climate changes across three time frames: early (2016-2035), mid (2046-2065), and late (2080-2099) 21st century Under the RCP 4.5 scenario, it is predicted that surface temperatures will rise by 1.9 to 2.4 °C, with precipitation expected to increase by over 20% in northern regions By mid-century, annual temperatures are projected to increase by 1.6 to 1.7 °C, with seasonal variations showing increases of 1.2 to 1.6 °C in winter and 1.6 to 2.0 °C in summer Additionally, the average annual maximum and minimum temperatures are anticipated to rise by 1.7 to 2.7 °C and 1.4 to 1.6 °C, respectively Rainfall patterns indicate a decrease of up to 10% in winter, a 10% increase in spring, a 5-15% increase in summer, and a significant rise of 15-35% in autumn.
The latest challenge in addressing climate change is reflected in the impact projections outlined in the IPCC AR5 Given the ongoing development of the basin as both an agricultural area and a modern industrial zone, it is crucial to establish a sustainable water resources policy and strategy.
Water Stress Assessment
Various methods exist to assess basin vulnerability concerning water stress, including the Falkenmark indicator, Green-Blue Water Scarcity Index, Water Stress Index, Smakhin Water Supply Stress Index, Stream Flow-Based Index, Water Scarcity Index, Water Supply Sustainability Risk Index, Watershed Sustainability Index, Green Water Stress Index, and Green Water Scarcity Index A comprehensive technical review of these methods is provided in the report by Xu and Hu (2017).
Water stress occurs when individuals lack access to a dependable and adequate water supply for their daily needs Water scarcity is characterized by regions facing significant challenges in meeting water demands.
A significant number of individuals are facing inadequate water supply for extended periods (Rijsberman et al., 2006) The commonly used index for assessing water stress is based on the ratio of water withdrawal to available resources, categorized into four ranges: low (0.4) (Vorosmarty et al., 2005; Xu and Hu, 2017).
The IPCC AR4 projections indicate a rise in global average temperatures alongside a decrease in precipitation, which will disrupt the water supply-to-consumption ratio (Solomon et al., 2007) Bromand (2015) highlighted potential water stress in the Kabul River Basin, Afghanistan, using a coupled SWAT model and climate models Koiso et al (2019) assessed the water supply and demand balance in the Mekong River Basin under the RCP 8.5 scenario, revealing that climate change would exacerbate water stress intensity Research by Yamamua et al (2018) developed a global-scale water resource assessment tool, identifying the downstream regions of the Mekong River Basin as highly vulnerable to water stress, with over 80% of the Murray-Darling Basin already experiencing significant stress since the 2020s Future indicators suggest that water stress levels will increase Additionally, Sun et al (2008) utilized water supply and scarcity stress indices to predict water stress in the Southeastern United States, finding that population growth intensified water stress, while future land use changes had minimal effects on water resource availability.
In Vietnam, the 2030 Water Resource Group conducted a general water stress assessment across 16 major river basins using the Water Exploitation Index (WEI) in 2016 The WEI, which measures the ratio of annual freshwater withdrawal to long-term water resource potential, indicates the pressure on water availability from municipal, agricultural, and industrial sectors The index categorizes stress levels as follows: no stress (0-10%), low stress (10-20%), stressed (20-40%), and severe stress (over 40%) For instance, the dry season WEI for the Red-Thai Binh basin was recorded at 19% in 2016, indicating low stress, but is projected to rise to 27% by 2030, approaching a stressed level (2030WRG, 2017).
Figure 2.2 Water Stress Levels in Vietnam (2030WRG, 2017)
Water Quality Evaluation
Vietnam is divided into seven major regions: Northwest, Northeast, Red River Delta, North Central Coast, South Central Coast, Central Highlands, and the Mekong River Delta While upstream water quality is generally good, it deteriorates downstream as it flows through residential areas, industrial clusters, craft villages, and agricultural zones, primarily due to the direct discharge of untreated wastewater and improper use of pesticides and fertilizers.
Water quality in the Northeast region, Red River Delta, South Central Coast, and Northeast Mekong region has significantly deteriorated Figure 2.3 (2030WRG, 2017) illustrates the distribution of seven regions, highlighting the status of water quality and the pollution sources contributing to this decline.
The Water Quality Index (WQI) is a mathematical formula that condenses various pollution factors into a single score, facilitating easier interpretation of water quality Each parameter is assigned a subjective weight, and its significance is assessed to produce a numerical value (Garcia, 2017) WQI is utilized in both developed and developing nations to aid decision-makers while promoting public awareness and engagement in water quality issues.
The Water Quality Index (WQI) has proven to be an effective tool for sound water resource management, with widespread applications in countries like India, Chile, England, Wales, Taiwan, Australia, and Malaysia (Phu, 2019) Vietnam implemented its first WQI in 2011, and an enhanced version has been publicly available and in use since 2019.
On November 12, 2019, the Vietnamese Government introduced a new method for calculating the Vietnamese National Water Quality Index (VN_WQI) through Decision No 1460/QD-TCMT, aimed at promoting efforts to assess surface water quality effectively.
2019) This new VN_WQI is an improved version of the former WQI stipulated on July
The updated VN_WQI now incorporates subjective purposes, such as pesticide or heavy metal modules, enhancing its calculation method This new version includes a broader range of water quality parameters compared to the previous WQI (2011), which was limited to only nine parameters, including WT, pH, BOD5, COD, turbidity, and TSS.
NH 4 + - N, PO 4 - - P and coliform (Pham et al., 2017 & Son et al., 2020) But physical parameters like turbidity and total suspended solid were no longer involved in the new VN_WQI It was noteworthy that the new VN_WQI brought more detailed classification than the previous method of calculation with six range of classification while the former had only five categories However, the index value remains the same, ranging from 0 to 100 Generally, it can be easily understood that the higher WQI indicates the better water quality.
Water Quality Index (WQI) is extensively utilized in Vietnam, as demonstrated by Son et al (2020), who assessed the water quality of the Cau River Their study employed a combination of various indices, including the water quality index, comprehensive pollution index, organic pollution index, and trace metal pollution index, to provide a thorough evaluation of the river's health.
Seasonal and spatial distribution significantly influence river water quality, with organic pollution observed in various locations, particularly in downstream areas, increasing the risk of eutrophication Phu (2019) utilized the Water Quality Index (WQI) to evaluate the water quality status of the Luy River in Binh Thuan Province, demonstrating WQI's effectiveness in analyzing seasonal variations in river water quality concerning watershed spatial distribution Additionally, Hong (2018) conducted a WQI-based assessment on the Tien River, further highlighting the importance of this method in water quality studies.
River quality in Southern Vietnam to find out long term annual and seasonal changes of water quality.
Statistical tools and analysis are effective to treat water quality data Zeinalzadeh
Rezaei (2017) utilized principal component analysis (PCA) and the Water Quality Index (WQI) to assess the Shahr Chai agricultural watershed in Iran, revealing that reduced stream flow and increased pollution and agricultural discharge negatively impacted downstream water quality Similarly, Yamazaki et al (2017) conducted a study on the Rekifune and Satsunai River watersheds in Japan, employing multivariate statistical techniques like PCA and cluster analysis Their findings indicated that significant pollution loads from tributaries, influenced by varying landscapes, greatly affect the overall quality of the major river.
In a comprehensive evaluation of water quality, Ali (2016) employed a combination of PCA, Cluster Analysis, and WQI to assess the Tigris River in Iraq, revealing key insights into point sources, pollutant characteristics, and treatment needs Similarly, Mishra (2010) utilized PCA to distill multiple water quality parameters of the River Ganges in India, identifying significant parameters and discharge sources Gajbhiye et al (2015) highlighted the effectiveness of PCA in identifying the main variability in surface water quality in Jabalpur City, India Tripathi & Singal (2019) further applied PCA to determine critical water quality parameters for WQI calculation in the River Ganges Additionally, Kamble & Vijay (2011) analyzed the coastal water quality of Mumbai, discovering seasonal variations in pollution levels through cluster analysis.
(2017) used a variety of WQIs and PCA to evaluate nutrient and organic loads in Acude Macela Macela Reservoir in Brazil.
Recent studies on Water Quality Index (WQI) in the study basin utilized outdated calculation methods, while the latest version integrates various water quality parameters for a more accurate assessment Additionally, multivariate statistical analyses, including Principal Component Analysis (PCA) and Cluster Analysis (CA), have demonstrated effectiveness in evaluating the spatial and temporal variations of river water quality across three sub-basins.
Figure 2.3 River Water Quality in Vietnam (2030WRG, 2017)
Summary
There is a notable lack of research on water stress assessment using hydrological models in the UPTBRB, with previous studies primarily concentrating on stream flow and erosion The reliance on outdated spatial and time series data poses challenges in accurately reflecting recent changes in watershed hydrology and climate change To effectively assess the current water quality of the Cau River and establish a comprehensive water quality database for the Thuong and Luc Nam River Basins, the implementation of an updated Water Quality Index (WQI) and multivariate statistical analysis is essential, offering deeper insights into basin water quality evaluation.
HYDROLOGICAL SIMULATION
SWAT Hydrological Model
The Soil and Water Assessment Tool (SWAT) is a widely utilized hydrological model designed for constructing watershed models, such as the one for the UPTBRB Given that agriculture is the predominant land cover in the basin, SWAT is ideal for simulating both water quantity and quality Supported by the United States Department of Agricultural Research (USDA), SWAT is a continuous, physically-based, and semi-distributed model capable of assessing the effects of land management practices on surface flow, sedimentation, and chemical yields For comprehensive insights and methodologies related to SWAT modeling, refer to the SWAT Theoretical Documentation and User’s Manual available online.
The model begins by dividing the targeted basin into multiple sub-basins and generating hydrological response units (HRUs) based on similar geomorphological and hydrological features It requires input data such as topography, land cover, soil layers, and climatic conditions SWAT then simulates physical processes, including runoff, nutrient, and sediment transport over continuous time This tool is particularly effective in data-scarce basins, allowing for the assessment of water resource impacts under various factors, including land use and climate change This study utilized the Arc interface of SWAT 2012 as an extension within ArcGIS.
In SWAT, the water balance of hydrological cycle is calculated using the following equation 3.1 with regard to the hydrological cycle in the basin.
SW t = SW 0 + ∑ R day – Q uaff – E a – w sweep – Q gw )
Where SW t is the final calculated soil water content (mm H 2 O) in a given time (days).
SW 0 is the initial amount (mm H 2 O) of soil water content, t is the time (days),
R day represents the daily precipitation measured in millimeters of water (mm H2O), while Q suf indicates the surface runoff for that same day Additionally, E a refers to the daily evapotranspiration amount in mm H2O, w sweep denotes the volume of water percolating through the soil profile, and Q gw signifies the return flow occurring on that day.
The surface runoff of the basin is estimated using the following equation 3.2.
The initial abstractions, represented by "I," encompass surface storage, interception, and infiltration before surface runoff occurs (measured in mm H2O) The retention parameter, denoted as "S" (also in mm H2O), is determined using equation 3.3, as it is influenced by factors such as soil characteristics, land management practices, topography, and the quantity of SW t.
The curve number (CN) for each hydrological response unit (HRU) is influenced by soil permeability, land cover, and soil water content variations Changes in CN are closely linked to moisture conditions, categorized as 1-dry (wilting point), 2-average moisture, and 3-wet (field capacity) Detailed equations for calculating CN for each condition are outlined below.
In scenarios where CN 1, CN 2, and CN 3 represent varying moisture conditions, it is important to note that the slope in CN 2 is less than 5% The topographic factor, particularly slope, influences the Curve Number (CN), necessitating the use of Equation 3.6 for areas with steeper slopes.
The CN 2s represents the adjusted curve number that accounts for slope, while CN 2 and CN 3 are modified based on a default slope of 5% The variable slp indicates the average slope of the basin.
3.1.1 Preparation of In-put Data for Model Set-up
The SWAT model distinguishes itself from other hydrological models by necessitating various input parameters, including a Digital Elevation Model (DEM), land cover data, soil classification maps, and comprehensive weather information such as temperature, precipitation, solar radiation, relative humidity, and wind speed This study utilized the most accurate and relevant data available for the basin to ensure effective model setup.
The Digital Elevation Model (DEM) is a crucial data source for hydrologic modeling, as it aids in defining basin boundaries and classifying elevation classes and sub-basin divisions This research utilized the SRTM3 DEM (version 3), which features a spatial resolution of 30 meters and was developed by the United States Geological Survey (USGS) and the National Geospatial-Intelligence Agency (NGA) The data was obtained from the CGIAR-CSI SRTM data webpage Figure 3.1 illustrates the DEM of the Upper Pahokee-Taylor Basin River Basin (UPTBRB).
Figure 3.1 DEM of the UPTBRB ii High Resolution Land Use and Land Cover
High Resolution Land Use and Land Cover (HRLULC), developed by the Earth Observation Research Center/Japan Aerospace Exploration Agency (EORC/JAXA) for mainland Vietnam in 2015, is a valuable tool for classifying basin land use patterns With a high spatial resolution of 15 meters and detailed land cover classes, HRLULC is particularly suitable for SWAT hydrologic modeling, allowing for precise assignment of specific land use patterns (EORC/JAXA, 2016).
The HRLULC dataset was obtained from the JAXA EROC website and subsequently extracted to the basin scale It was analyzed across six primary categories—barren land, agricultural land, forest, urban/built-up areas, water bodies, and others—to offer a comprehensive overview of the current land cover status (refer to Figure 3.2).
The upstream region of the basin is predominantly covered by forest, while agricultural land is the main land use throughout most of the basin Although residential land is minimal in the upstream area, small clusters begin to appear as one moves toward the downstream regions; however, this residential land remains significantly less than both forest and cultivated areas.
A few proportions of water body and barren land were also observed The share of others type like grass/shrub land was also found.
The analysis of land cover reveals that agriculture is the predominant category, accounting for 42.33% of the area, followed closely by mixed forests at 40.14% Additionally, rangeland comprises 7.61%, while residential land, water bodies, and barren land represent 4.12%, 2.69%, and 3.11%, respectively.
Figure 3.2 High Resolution Land Use and Land Cover Map of the UPTBRB (2015)
Figure 3.3 Major Land Cover Statistics in the UPTBRB
The land cover classification layers were processed according to SWAT land cover codes, resulting in a dataset with high resolution and accuracy This detailed classification allowed for careful assignment of the corresponding SWAT land use and land cover codes, as illustrated in Table 3.1.
Table 3.1 SWAT Land Use Code and Statistics (2015)
Land Use Area (ha) Area (%)
AGRR Agricultural Land-Row Crops 242185.8499 19.04
The essential soil database for the UPTBRB was obtained from the FAO GeoNetwork website The soil map was created by extracting data according to the basin boundaries, and an analysis of the spatial distribution of soil classes was conducted, as illustrated in Figure 3.4.
The detailed results of analysis were illustrated in the figure 3.5.
Figure 3.4 Soil Map of the UPTBRB
Figure 3.5 The Proportion of Soil Classes observed in the UPTBRB iii Weather Data
SWAT Model Set-up
The SWAT model effectively extracted key hydrological features, including flow direction and flow accumulation, after defining the outlet of the entire basin Additionally, it delineated the basin boundary and produced a comprehensive report detailing the topographic profile of the basin, which encompasses 27 sub-basins.
Following the division of sub-basins, a detailed analysis of Hydrological Response Units (HRUs) was conducted Spatial data, including land cover, soil maps, and slope classifications, were incorporated, resulting in the generation of 347 HRUs by SWAT (refer to Figure 3.7).
The SWAT weather database was developed utilizing five key parameters: maximum temperature, minimum temperature, precipitation, relative humidity, and wind speed, sourced from 11 stations Additionally, solar radiation data was calculated for one station and incorporated into the SWAT model.
SWAT model for the UPTBRB was successfully simulated for the period of
(2005 – 2019) whereas the first three years (2005 – 2007) was used as a warming period
29 to maintain model performance SWAT Error Check Tool was used to examine the simulated hydrological parameter value.
SWAT Model Calibration and Validation
The output of SWAT model was calibrated and validated through the Soil and
Water Assessment Tool Calibration and Uncertainty Procedure (SWAT-CUP) with
The Sequential Uncertainty Fitting (SUFI-2) algorithm addresses uncertainty in model calibration, which aims to align model simulations with observed data by adjusting input parameters This calibration process optimizes the objective function, while validation involves applying the calibrated parameters to an independent dataset without modifications Additionally, sensitivity analysis identifies the most influential parameters affecting the hydrological processes of the basin and, consequently, the model output For further details, refer to the SWAT User’s Manual (Abbaspour, 2012).
Model Performance Evaluation
The performance of the SWAT model was assessed using various statistical parameters, with the Nash Sutcliffe Efficiency (NSE) serving as the primary objective function due to its widespread acceptance in hydrology Additionally, other metrics were also considered to provide a comprehensive evaluation.
Percentage Bias (PBIAS), and the ratio of root mean square error (RSR) were also considered.
The NSE (Nash-Sutcliffe Efficiency) assesses the fitness of a model by comparing simulated data to observed data, with values ranging from -∞ to 1 Meanwhile, PBIAS (Percent Bias) indicates whether the simulated values tend to be larger or smaller than the observed values.
The optimum value for the model performance indicator RSR is zero, where a positive value signifies underestimation and a negative value indicates overestimation RSR is calculated as the ratio of root mean square error (RMSE) to the standard deviation, with higher values reflecting better model performance, ideally reaching the optimum value of zero (Singh et al., 2004; S Shrestha et al., 2016).
The observed flow quantities, denoted as Q m, are compared with simulated flow quantities to assess model performance, with the average discharge calculated for both observed and simulated periods The total number of flow quantities is represented by n, and i corresponds to the specific time-step The model's fitness criteria, as outlined by Moriasi et al (2007), are detailed in Table 3.3.
Results of Simulation
3.5.1 SWAT Model Calibration and Validation Result
The SWAT model was effectively calibrated and validated using river flow data from the Gia Bay hydrological station, achieving a notable NSE of 0.77 Table 3.4 presents the goodness-of-fit statistics for the SWAT model, highlighting its accuracy in simulating river flow.
The SWAT model demonstrated strong performance, with a PBIAS of -9.6 and an RSR of 0.48 during initial assessments, while model validation yielded even better results: NSE = 0.84, PBIAS = -7.4, and RSR = 0.39 These statistical indicators confirm the model's effectiveness in simulating the entire basin using calibrated parameters, as outlined in table 3.5, with the manual calibration tool in ArcSWAT facilitating the process.
Table 3.4 Goodness-of-fit Statistics for Discharge Simulation
Process Period NSE PBIAS RSR Performance
Table 3.5 Calibrated Parameters and Fitted Values
Method_Parame Description and Sensitivity Lower Upper Fitted ter Unit Rank Bound Bound Value
A_GWQMN.gw Threshold depth of 1 25 3600 3554 water in the shallow aquifer required for return flow to occur (mmH2O)
A_GW REVAP Groundwater revap 2 0.02 0.14 0.116 coefficient
DELAY.gw from soil to channel (days)
V_CH K2.rte Hydraulic 4 5 22 19.13 conductivity of the main channel (mm/h)
().sol capacity of soil layer (mmH2O/mm soil)
A_ALPHA_BF Base flow alpha 6 0.002 0.033 0.003 gw factor (1/days)
R_CN2.mgt Initial SCS runoff 7 -0.21 -0.17 -0.21 curve number II
V_ESCO.bsn Soil evaporation 8 0.01 0.22 0.07 compensation factor
V_SLSUBBSN Average slope 9 15 20 19.57 hru length
Method R involves multiplying the existing parameter value by (1 + a specified value), while Method V entails replacing the current parameter value with a specified value In contrast, Method A adds a specified value to the existing parameter value.
M ea n M on th ly F lo w ( m 3 /s )
L95PPU (Lower Limit) U95PPU (Upper Limit) Monthly Precipitation
Figure 3.8 Calibration Result at the Gia Bay Hydrological Station
M ea n M on th ly F lo w ( m 3 /s )
L95PPU (Lower Limit) U95PPU (Upper Limit) Monthly Precipitation
Figure 3.9 Validation Result at the Gia Bay Hydrological Station
After SWAT model calibration and validation, the simulated basin’s hydrological components were further studied and the results were described in the table 3.6 and 3.7.
Table 3.6 Average Monthly Hydrological Components
Month Rain Surf Q Lat Q Water
(mm) (mm) (mm) (mm) (mm)
Table 3.7 Annual Water Balance Statistics
Year Rainfall Surf Q LAT GWQ
(mm) (mm) (mm) (mm) (mm)
2019 1657.07 253.36 187.46 424.12 546.40 151.28 710.06 784.11 Note: Surf Q= Surface runoff, LAT Q= lateral flow, GWQ = groundwater contribution to stream, SW = soil water content, ET = actual evapotranspiration, PET= potential evapotranspiration
Rainfall Surf Q LAT Q Percolate SW ET GWQ 6000
Figure 3.10 Annual Water Balance of the UPTBRB (2008 – 2019)
The SWAT model indicates that the basin experiences an annual precipitation of 1788.3 mm, with actual evapotranspiration accounting for 661.05 mm This leads to an annual water resource potential of 1126.98 mm, which will be considered the current annual water resource potential for the long-term period from 2005 to 2019.
Summary
The SWAT model was developed using the latest spatial data, including a 30 m resolution Digital Elevation Model (DEM) and high-resolution land use and land cover data at 15 m resolution, along with recent climatic data from 11 observation stations The model calibration and validation demonstrated excellent performance with Nash-Sutcliffe Efficiency (NSE) values of 0.77 and 0.84, marking the best simulation results for the basin from 2000 to 2021 The SWAT model indicates that the basin experiences an annual precipitation of 1788.3 mm and an actual evapotranspiration of 661.05 mm per year.
CLIMATE PROJECTION
Future Climate Scenario
The study examined the effects of future climate change on the basin's water resources It utilized the MPI-ESM-LR Regional Climate Model, focusing on projected maximum and minimum temperatures, as well as precipitation patterns The analysis was based on the RCP 4.5 carbon emission scenario for the years 2046 to 2060, using baseline data from 1991 to 2004 Relevant data for this model was sourced from the European Network for Earth System Modelling, with additional details provided in table 4.1.
Driving Model Product Resolution Frequency Study Period
MPI-ESM-LR ENES 0.5˚ × 0.5˚ Daily RCP 4.5 2046 – 2060
Bias correction was achieved using the linear scaling method, addressing systematic errors that arise during model development across various grid cells This method demonstrated significant efficiency, with performance comparable to more complex correction techniques Specifically, the multiplicative approach was applied for precipitation correction, while temperature adjustments were made using the additive method (Shrestha et al., 2017).
& Okwala et al., 2020) Table 4.2 provided period of study and projected climatic variables.
Table 4.2 Period of Study and Projected Climatic Variables
Description Period Projected Variable Remark
Historical/baseline 1991 - 2004 5 Observed data tasmin
Future (MPI) 2046 - 2060 5 pr, tasmax, Corrected RCM tasmin data
Note: pr means precipitation, tasmax indicates maximum temperature and tasmin is for minimum temperature.
In this article, we discuss the variables used in climate data analysis, where P represents precipitation measured in millimeters (mm) and T denotes temperature in degrees Celsius (°C) The mean value of time series data is indicated by "à m," while "his" refers to historical data derived from Regional Climate Model (RCM) outputs, and "fut" signifies simulated future data The dataset is categorized into monthly (m) and daily (d) formats, with "obs" indicating observed data and "*" representing bias-corrected data.
The following equations of 4.5 and 4.6 were used to find out the changes in climatic variables (i.e., temperature) with respect to historical period under different time frames. ΔPrecipPrecip current/future = ∗ × 100 (4.5) ΔPrecip /
Where ΔPrecipPrecip current/future and ΔPrecip / are the changes in precipitation (%) and temperature (°C) with respect to historical period of study.
In this study, climate projection was performed at 5 meteorological stations covering the whole basin namely; Bac Kan, Thai Nguyen, Bac Gaing, Huu Long and
Luc Ngan (See figure 4.1) The future climate trends were accessed under the projected
RCM data The spatial distribution of these stations was good enough to reflect the climate condition of the basin.
Performance Analysis of Bias Correction Method
Before proceeding with further applications, it's crucial to assess the performance of the dataset following bias correction This evaluation compares the grid time series of raw data with bias-corrected values against the baseline study period of 1991 to 2004 The Root Mean Square Error (RMSE) serves as the metric for determining prediction accuracy between observed and simulated values, with RMSE values ranging from 0 to +∞, where lower values indicate better model predictions.
1 The Mean Absolute Error (MAE) value indicates the status of unbiased simulation whereas the optimum value is 0 Both RMSE and MAE have been serving as the profound indicators in the evaluation of climate change and hydrological modelling fields (Mendez et al., 2020).
Where: Mi and Oi are the simulated and observed values.
Results of Climate Projection
The projected climate variables were described by maximum temperature, minimum temperature and precipitation extracted from the MPI-ESM-LR under RCP
4.5 scenario The future climate trend was estimated with a reference to the baseline period of study (1991 – 2004) Before finding out the patterns of future changes, the performance of linear scaling bias correction method was evaluated using two statistical parameters; RMSE and MAE The following table 4.3 described the results of the performance evaluation of the applied bias correction method Linear scaling method showed high applicability and the performance of the corrected dataset was improved except some parameters.
BK BG TN HL LN
RAW ABC RAW ABC RAW ABC RAW ABC RAW ABC
Note: BK: Bac Kan, BG: Bac Giang, TN: Thai Nguyen, HL: Huu Long, LN: Luc
Ngan, RAW: Raw, ABC: After Bias Correction
Using the bias corrected RCM data (hereafter mentioned as MPI/MPI-Corrected), the analysis on monthly changes were made between the baseline/historical period (1991
The analysis of rainfall trends from 2004 to 2060 indicates that the rainy season consistently begins in May and concludes in August, with these months experiencing the highest levels of precipitation However, a notable shift in peak rainfall patterns has been observed, as historical data from the 20th century showed the highest precipitation occurring in July, which has since declined steadily after August in current conditions.
However, in the 21 st century, August has become the highest precipitation received month The future status of precipitation also agreed that peak rainfall would be occurred in August.
A comprehensive analysis of peak precipitation events revealed that Thai Nguyen station is projected to experience the highest levels of precipitation among the studied meteorological stations Historical data from 1991 to 2004 recorded a peak daily precipitation of 287.4 mm at Thai Nguyen station, while the period from 2005 to 2019 showed a peak of 262.1 mm Furthermore, Regional Climate Model (RCM) predictions for the years 2046 to 2060 indicate that Thai Nguyen station will continue to record the highest daily precipitation levels.
Table 4.4 illustrates seasonal precipitation trends, indicating a current winter increase of 2.68%, while future projections suggest a decline of 13.79% Spring precipitation has seen a decrease of 12.93%, yet is expected to rise by 0.9% in the future Additionally, summer and autumn precipitation have increased by 0.56% and 26.85%, respectively, but are forecasted to decrease by 3.04% and 4.60% in the coming years.
Table 4.4 Changes in Seasonal Precipitation
Season Historical Current Current-Changes MPI MPI-Changes
Recent findings indicate a slight increase of 1% in current annual precipitation, rising to 1551 mm from the historical average of 1537 mm However, future projections suggest a potential decrease of 2%, bringing annual precipitation down to 1512 mm in the basin The corrected RCM-MPI model predicts changes in early monsoon characteristics, with irregular extreme rainfall patterns and a significant reduction in autumn rainfall, which are closely linked to increased flood risks Additionally, winter months are expected to be drier than recorded historically, heightening the risk of drought.
40 re ci pi ta ti on ( m m /m on th )
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 4.2 Average Monthly Precipitation Trends (5-stations average)
Figure 4.3 Annual Changes of Precipitation Trends (5-stations average)
4.3.3 Projected Maximum and Minimum Temperature
Future temperature projections indicate a consistent increase in both maximum and minimum temperatures throughout the year compared to baseline and current conditions This warming trend begins in January, peaks in June, remains stable for about three to four months, and then gradually declines.
The highest recorded temperature in the basin reached 41.30 °C at Huu Long station during the current period, surpassing the historical high of 40.7 °C at Thai Nguyen station Conversely, the minimum daily temperatures recorded were 30.70 °C at Huu Long station historically and 30.20 °C at Thai Nguyen station in the current period.
2060, the record-breaking maximum temperature is likely to occur in Huu Long station and the record-breaking minimum temperature will be in the Bac Kan station.
Seasonal temperature analysis indicates a current decrease in maximum winter temperatures by 0.32 °C, with projections suggesting an increase of 1.46 °C by 2060 Specifically, maximum temperatures are expected to rise by 6.99 °C to 7.29 °C in spring, 8.24 °C to 9.64 °C in summer, and 8.44 °C to 9.74 °C in autumn Meanwhile, winter's minimum temperatures have decreased by 0.01 °C but are anticipated to increase by 1.26 °C in the future Additionally, spring minimum temperatures may rise between 0.24 °C and 0.98 °C, summer between 0.68 °C and 2.01 °C, and autumn between 0.44 °C and 1.28 °C Overall, both maximum and minimum temperatures are projected to rise across all seasons.
Table 4.5 Changes in Seasonal Maximum Temperature
Season Historical Current Current-Changes MPI MPI-Changes
Table 4.6 Changes in Seasonal Minimum Temperature
Season Historical Current Current-Changes MPI MPI-Changes
Table 4.7 reveals that the current annual maximum temperature has risen by +0.14 °C, with projections indicating an increase of approximately +1.33 °C by the end of 2060 Additionally, significant changes have been observed in the annual minimum temperature as well.
42 as it already increased +0.34 °C and also tend to increase about +1.39 °C in the future.
Average changes on both monthly and annual basis were illustrated in figure 4.4 to 4.7.
Table 4.7 Changes in Long Term Annual Temperature
Parameter Historical Current Current-Changes MPI MPI-Changes
M A X IM U M T E M P E R A T U R E ( °C ) Historical Current MPI-Corrected
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV
Figure 4.4 Average Monthly Changes of Maximum Temperature (5-stations average)
M IN IM U M T E M P E R A T U R E ( °C ) Historical Current MPI-Corrected
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV
Figure 4.5 Average Monthly Changes of Minimum Temperature (5-stations average)
Figure 4.6 Annual Changes of Maximum Temperature (5-stations average)
Figure 4.7 Annual Changes of Minimum Temperature (5-stations average)
Summary
Climate projection was successfully carried out at five meteorological stations.
Climate change analysis were done for three decadal time frames, historical (1991 –
2004), current (2005 – 2019) and future (2046 – 2060) The performance of extracted
The application of the linear scaling bias correction method enhanced RCM data, indicating that winter precipitation is expected to significantly decrease, while summer may see a slight reduction Conversely, spring and autumn are projected to experience an increase in precipitation.
Both maximum and minimum temperature showed upward trends in all seasons The bias corrected future climate data will be inserted into SWAT model to determine future water resource potential.
WATER STRESS ASSESSMENT
Water Demand
Water demand estimation involved the integration of various factors, including municipal, agricultural, and industrial sectors Key country reports and statistics provided the basis for assessing water exploitation rates across these sectors The overall water demand for the basin was calculated using the specified equation.
= (Domestic× Basin Population) + (Industrial × Area) + (Agricultural × Area)
Water Resources
The water resources in the basin were calculated using Equation 5.2, which incorporates annual precipitation and actual evapotranspiration (AET) data generated by the SWAT hydrological model, as detailed in Section 3.53.
Water Resources = (Precipitation ˗ AET) × Basin Area × 10 3 (5.2)
Water Stress
The water stress assessment utilized equation 5.3 to evaluate both current and future conditions, offering a detailed analysis of water demand and the potential resources available in the basin.
The water stress levels on sub-basin scale were evaluated based the value of corresponding indicators; no stress (the value < 0.1), low stress (0.1 to 0.2), high stress
(0.2 to 0.4) and severely stress (≥ 0.4) (Vorosmarty et al., 2005 & Xu and Hu, 2017).
Future Water Stress
5.4.1 Future Water Demand a) Population Projection
The population in the basin was projected in order to calculate future water demand for municipal sector Geometrical increase method with the baseline year of
(2010 – 2020) was collected from https://www.worldpop.org and applied to forecast future basin population This method considered decadal population growth percentage as a constant for decadal projection.
Where, P rate is present population growth rate, P present is number of present populations,
Where; is decadal population growth percentage, P rate1,2,etc present population growth rate, N is number of P rate
The future population growth in the basin is determined using a geometrical increase method formula.
Where, P future is estimated future population, P present is number of present populations, and n is the projected year.
The availability of water resources is closely linked to the hydrological cycle and climate trends By utilizing bias-corrected Regional Climate Model (RCM) data, particularly focusing on maximum and minimum temperatures as well as precipitation, the SWAT model was employed to assess the future potential of water resources in the basin.
Result of Water Stress Assessment
The estimation of sectoral water demand in the basin involved integrating data from municipal, industrial, and agricultural sectors For municipal water withdrawal, capita-based consumption was utilized, while spatial-based exploitation was employed to assess the economic implications of the industrial and agricultural sectors.
Water consumption statistics indicate that the average daily water usage ranges from 90 to 120 liters per person, varying by region However, in urban areas, the daily water requirement per capita is projected to exceed 120 liters (2030WRG, 2017).
Therefore, 43.8 m 3 /year was applied as the reference annual water demand per capita.
As of 2015, the total basin population was approximately 4.95 million thus the annual domestic water demand was mounted to approximately 216 mn m 3 /year. b) Industrial Water Demand
The industrial clusters in the basin lead to a unique pattern of industrial water demand compared to other regions in Vietnam To enhance the management of industrial parks and economic zones, the Ministry of Finance has introduced circular No 43/2019/TT-BTC, aimed at promoting trade and attracting foreign investment.
According to Circular No.33 TCXD VN33/2006, the average water demand for one hectare of industrial area is 45 m³ per day In the basin, there are 10 major industrial parks covering around 3,300 hectares, but current water demand reflects a 50% reduction, accounting for 1,650 hectares still in development Consequently, the estimated annual water demand for the industrial sector is approximately 27 million m³.
Agriculture is the primary land use, accounting for 42.33% of land, with 47% dedicated to row crops, 35% to orchards, and 18% to paddy fields The irrigation water requirement varies seasonally, with rice fields needing about 10,000 to 12,000 m³/year in the dry season and a minimum of 5,000 m³/year during the rainy season Consequently, a uniform water requirement of 5,000 m³/ha/year is applied across all agricultural lands The total water demand for the agricultural sector is estimated to be around 2.7 billion m³/year.
After estimating the sectoral water demand, the proportion of water demand was depicted in figure 5.1.
Figure 5.1 Proportion of Sectoral Water Demand
The basin experiences an annual precipitation of 1,788.27 mm, with actual evapotranspiration measuring 674.94 mm Utilizing equation 5.3, the annual available water resource potential for the entire basin is estimated at 14 billion cubic meters per year.
Using equation 5.3, water stress assessment was carried out based on the individual sub-basins It was revealed that 15% (4 sub-basins) of the basin had low water stress, 33%
Among the sub-basins analyzed, 9 are categorized as having moderate water stress, while 48% (13 sub-basins) experience medium water stress Additionally, 4% (1 sub-basin) falls into the stressed category Notably, the only outlet basin is marked with high water stress, indicated by a red color, although it was not specified in the analysis.
49 appeared vividly due to its small area Detailed statistics about current water stress was provided in Appendix 2.A.
Figure 5.2 Distribution of Current Water Stress Levels
With the purpose to adopt necessary response plans and mitigation measures, future water stress was also predicted through multi-projection methods. a) Municipal Water Demand
The future population of the basin is projected to grow by approximately 6.51 million by 2060, utilizing the geometrical increase method for the period of 2046 to 2060 According to the 2030 Water Resources Group (2030WRG), the per capita water demand is expected to reach 150 liters per day by 2030 Based on the geometric population growth data presented in Appendix 1 and the calculations from equation 5.1, the annual municipal water demand is anticipated to reach around 354 million cubic meters per year.
By 2060, all industrial clusters are projected to operate at full capacity across a total area of 3,300 hectares, leading to an estimated increase in future water demand of approximately 54 million cubic meters per year.
The Agricultural Master Plan of 2020 aims to boost rice production to 41-43 million tons per year by 2020 and 44 million tons by 2030, alongside an anticipated rise in domestic consumption and exports of agricultural products To support this production growth, a projected 10% increase in agricultural land is necessary, which translates to a future water demand of 2.99 billion cubic meters per year to adequately supply these cultivation areas.
The future water resource potential was assessed using the SWAT hydrological model in conjunction with climate projections By integrating the SWAT model with bias-corrected RCM data, future annual precipitation is projected to be 1,596 mm/year, while annual actual evapotranspiration is expected to reach 685 mm/year Consequently, the estimated water resource potential for the basin is approximately 11.6 billion cubic meters per year.
Future water stress projections indicate that by 2060, 4% of the sub-basins will experience low water stress, while 30% will face moderate water stress Additionally, 41% are predicted to encounter medium water stress, and 26% are expected to undergo significant water stress For more detailed statistics on future water stress, refer to Appendix 2.B.
Figure 5.3 Distribution of Predicted Water Stress Levels
Figure 5.4 Comparison of Current and Future Water Statistics (Annual)
Figure 5.5 Changes in Water Demand by Each Sector
Figures 5.4 and 5.5 illustrate the projected increase in future water demand alongside a decline in available water resources This trend underscores the urgent need to focus on alternative water resources and the importance of ecological performance parameters.
Summary
The SWAT hydrology model outputs and climate projections were utilized to assess current and future water stress at the sub-basin level Currently, the basin experiences medium water stress (less than 0.4), indicating that natural processes can help replenish water levels However, projections suggest that by 2060, water stress may escalate to high levels (greater than 0.4) unless effective water management strategies are implemented This situation will intensify competition for water resources, particularly among users in the middle and downstream areas.