Impacts of dike systems on hydrological regime in Vietnamese Mekong Delta Open Access

The Vietnamese Mekong Delta (VMD) as a key production landscape plays a critical role in the nation's food security and economic development. In recent years, the Government of Vietnam has introduced policies and promoted solutions to develop the region to foster socio-economic and sustainable development. Essential to the national economy, the region has undergone remarkable growth with significant recent achievements in agricultural and rural development sectors (Mekong Delta Plan (MDP) 2013). The VMD's location renders it highly vulnerable to the impacts of climate change, including rising sea levels, flood events during the rainy season and water shortages during the dry season (Wassmann et al. 2004; Khang et al. 2008). Changes to the flow regime, reduced sedimentation riverbank erosion and salinity intrusion caused by anthropogenic interventions or as a result of climate change remain grave challenges.

Since the 1990s, extensive dike systems and irrigation infrastructures have been incrementally constructed to improve crop yields and reduce the impact of annual flood on the region (Nguyen et al. 2017; Minh et al. 2019a, 2019b). Concurrently, alongside economic development and transformation, new and expanded residential areas, transport infrastructure and services have gradually reduced the area of forests, agricultural land, and water surfaces. Furthermore, the expansion of cultivation areas, industrial zones, and hydropower dam construction in upstream countries has reduced water and sediment sources downstream, while increasing erosion, surface water pollution, and loss of fertile sediments (Dat et al. 2011). Previous studies in the VMD's upstream provinces have found surface water pollution directly caused by agricultural runoff (Chau et al. 2015; Thu Minh et al. 2020). The surface water quality monitoring system in the VMD is primarily concentrated along the main rivers and canals. However, the secondary and tertiary water bodies, which act as the primary source of water supply for adjacent households, often suffer from the highest levels of pollution.

Anthropogenic activities have impacted discharge volumes greatly. Hydropower dams, reservoirs, sluice-gates, dikes, and embankments have all been developed to retain water and protect and serve lives and livelihoods. The negative effects of hydropower and irrigation projects are well documented and increasing in significance, magnitude and frequency (Fujihara et al. 2016; Nguyen et al. 2017; Minh et al. 2019a, 2019b). Several studies have investigated the cumulative effects of Chinese dams in the Mekong River Basin. Lu et al. (2014) investigated the water flow regime at Chiang Saen in the lower Mekong, whilst Hecht et al. (2019) studied a range of hydrological impacts from hydropower dams. Gunawardana et al. (2021) went further and considered the multiple drivers of hydrological alteration in the transboundary Srepok River Basin of the Lower Mekong Region.

The fine sediment and geological structure of the riverbanks are changing the regime of the water–soil pressure balance (Liu et al. 2017; Van Tho 2020). Many studies have also shown that sand-mining activities and the presence of clay in riverbank soils have increased both the frequency and severity of erosion events (Liu et al. 2017; Marchesiello et al. 2019; Van Tho 2020). Some studies have used modelling such as a study on an ANFIS-based approach for predicting sediment transport (Azamathulla et al. 2012) and have used linear genetic programming to scour below a submerged pipeline (Azamathulla et al. 2011). Other studies suggest that vertical force transmissions from riverbank construction works may also contribute to riverbank collapse (Coumou 2017; Van Tho 2020). However, very few or none of these studies have comprehensively addressed the impacts of the development of dike systems on the hydrological regime in the region with the exception of Binh et al. (2020). Several studies have used the Mann–Kendall test to analyse long-term trends in climate variables for different climatic regions (Mehta & Yadav 2022). Perera et al. (2020) used the Mann–Kendall test to detect the magnitude of trends for well-known climate gauges in Trinidad and Tobago. Monthly climatic data including cumulative rainfall and average of the minimum and maximum atmospheric temperatures were all processed to identify the trend analysis using the above-stated non-parametric tests.

The objective of this study is to examine the impacts of the construction of irrigation-related infrastructures on river flow in the VMD. The study utilised a hydrological change index method and the Mann–Kendall test to assess temporal changes in both water levels and discharge along the main river channels of the delta. The study is expected to reveal evidence on the extent to which the dike system has altered the hydrological regime.

The northeast and southwest monsoons both have an impact on the climate in the Mekong Delta in Vietnam. Typically, the rainy season lasts from May to November, while the dry season lasts from December to April. The flood season begins one to two months later than the rainy season and ends nearly simultaneously with it. During the flood season, the delta facilitates water transport and drainage. Conversely, the delta frequently suffers from saline intrusion during the dry season due to the typically low flow rates, which are approximately only 2% of that of the rainy season, and which is insufficient to prevent large-scale and long-term saline intrusion (Nguyen 2016). In the dry season, the water levels rise and recede slowly, with an average amplitude of 5–7 cm/day, yet during the rainy season, flood waters can rise rapidly by 20–30 cm/day. The hydrological regime in the VMD is directly influenced by the upstream flow, rainfall, and the East and West Seas tidal regimes. Heavy upstream rainfall, hydroelectric dam construction, deforestation, irrigation canals, dikes and urban development are the primary causes of floods in the VMD.

An Giang Province (10°11′ to 10°58′ N, 104°46′ to 105°35′ E) is located in the upstream section of the VMD with a natural area of 354 × 10³ ha, of which 283 × 10³ ha are designated as agricultural areas (79.9%). The province had a total population of 1.9 × 10⁶ in 2019, distributed over 11 administrative units; two cities (Long Xuyen and Chau Doc), one town (Tan Chau) and eight districts (An Phu, Phu Tan, Cho Moi, Chau Phu, Chau Thanh, Thoai Son, Tinh Bien and Tri Ton). The study area is bordered by Cambodia to the north, Can Tho City to the south, Dong Thap Province to the east and southeast, and Kien Giang Province to the west and southeast. An Giang province exhibits an interlaced system of rivers and canals, with two main rivers, Mekong (Tien) and Bassac (Hau), which during the rainy season provide abundant surface water for irrigation-fed agricultural production. The hydrological regime divides into two distinct seasons: the flood season (from September to November) and the dry season (from February to May). The maximum discharge to the Mekong Delta is about 35,000 m³/s in the middle flood years and up to 44,000 m³/s in the large flood years, in which at Tan Chau station on the Mekong River it is 29,000 m³/s (accounting for 66%), on the Hau River at Chau Doc it is 8,200 m³/s (accounting for 18%) and the total overflows across the border to the wetland Plain of Reeds (Đ̀ông Tháp Mu'ò'i) and Long Xuyen Quadrangles are about 7,000 m³/s (16%). The benefits of floods for agricultural production in recent years are such as bringing in sediment, flushing fields, improving soil quality, water quality, replenishing groundwater sources, bringing aquaculture resources and creating jobs for farmers in the flood season. However, floods also affect socio-economic activities such as causing loss of life, damaging property, and resulting in increasing infrastructure investment and maintenance costs.

This study used indicators of hydrologic alteration (IHA) and the Mann–Kendall test to assess the changes in water level and discharge under dike system development in the upper section of the VMD. Because it does not require normally distributed datasets and has a low sensitivity to sudden breakdowns owing to inhomogeneous time series, the nonparametric Mann–Kendall technique has become highly utilised for trend analyses in recent years (Yue & Wang 2004; Shahid 2010). Data analysis was undertaken using XLSTAT software (Addinsoft, Paris, France) version 2019.

The Nature Conservancy's IHA software, which Richter et al. (1996) developed, was used to assess flow-regime change by calculating 33 hydrologic variability parameters as summarised in Table 1 (all parameters were based on daily measurements). This method has been widely used in assessing the impacts of climate change and anthropogenic interventions on flow regimes (Ty et al. 2011; Liu et al. 2018; Zhang et al. 2019; Wang et al. 2021, 2022; Guo et al. 2022).

In this study, flow data was collected from the two hydro-meteorological stations of Tan Chau on the Mekong River and Chau Doc on Bassac River for the period of 1986–2019, and Vam Nao station for the period of 1997–2019. Data was then divided into three phases, consisting of phase 1: 1986–1996, phase 2: 1997–2010, phase 3: 2010–2019 (Tan Chau and Chau Doc stations); and phase 1: 1997–2004, phase 2: 2005–2010, phase 3: 2011–2019 (Vam Nao station). These phases correspond to: (i) before the construction of the dike system (1986–1996); (ii) during and after the construction of the dike system (1997–2010); and (iii) after completion of the dike system (2011–2019).

When the observed frequency of the post-impact annual values that fall within the range of variability approach (RVA) target range equals the expected frequency, HA equals zero, whereas a positive deviation value indicates that the annual parameter values fall within the RVA target range more frequently than expected; and a negative value indicates that the annual values fall within the RVA target less frequently than expected (Jiang et al. 2014).

The RVA was used to investigate flow variability by determining the frequency of natural flows as the basis for considering variation. The value of the IHA parameters at the post-impact stage should be as close as possible to the initial value, i.e., the RVA should be as close to zero as possible. The RVA is divided into three grades ranging from 17% to median values: (i) the lowest category, which includes all values less than or equal to 33%; (ii) the mean category, which includes all values between 34% and 67%; and (iii) the highest category, which includes all values greater than 67%. Although the RVA level boundaries can be adjusted, using 33% and 67% ensures that the pre-impact value falls into each category in the majority of cases, making the results easy to understand and analyse.

Dike development in An Giang province commenced in the 1970s and by 1987 was already strongly developed. During this period, dikes were built to protect the two-crop rice protection areas (summer–autumn and winter–spring rice crops). By 1996, the first 800 ha of land in Cho Moi district were thoroughly protected by full dikes. Thereafter the area of agricultural lands protected by full dikes continued to increase.

The results showed that the change in flow regime (discharge) at Tan Chau in phases 1 and 2 was 68.2% and 76.6% respectively, whereas at Chau Doc, the change in flow for phases 1 and 2 was 71.2% and 66.6%, respectively. The least was at Vam Nao, where the changes in flow for phases 1 and 2 were 49.8% and 60.7%, respectively. In general, Chau Doc station showed the greatest flow alteration (71.2%) during the dike system construction (1997–2010), followed by Tan Chau (68.2%) and Vam Nao (49.8%). However, when considering phase 2 (2011–2019), when the dike system was relatively complete and no more dikes were constructed, the change in flow regime at both Tan Chau and Vam Nao stations was seen to increase significantly to 76.6% and 60.7%, respectively.

The change in hydrological indicators at Tan Chau and Chau Doc stations may be directly attributed to the Mekong River's upstream flow. Due to a lack of upstream data, this study only considered the assumption that the flow alteration seen at the stations was the result of dike development.

When the three indicators of group 5 (31 – flow rate increase, 32 – flow rate decrease, 33 – amount of reverse flow) are examined in detail, all three indications highlight significant changes (Figure 6). The highest change was seen in flow rate increase and was seen at all three stations. Meanwhile, indicators 32 and 33 at Vam Nao station experienced changes over both phases. According to The Nature Conservancy (2009), changes of indicators in group 5 (rate and frequency of flow variation) directly influence ecosystem functioning and may cause drought stress in plants and desiccation stress in low-mobility stream-edge organisms.

The dike system development to aid intensive rice production has significantly contributed to water infrastructure development and water supply. However, dike development in the VMD has also resulted in environmental pollution as the flushing of highly contaminated water from protected fields releases water with high pollutant loads that harm downstream ecosystems. This in combination with upstream hydropower dam development leads to a series of cumulative impacts on fish numbers and sediment supply. Therefore, the benefits (e.g., ecosystem services) and the costs (such as the sediment reduction, biodiversity losses, and land degradation) of the system should be taken into consideration (Binh et al. 2022). Hecht et al. (2019) reviewed the hydrological impacts due to hydropower dams in the Mekong River Basin and strongly recommended that dam impact studies should consider hydrological alteration in conjunction with fish-passage barriers, geomorphical changes and other contemporaneous stressors.

The flow-regime alteration was assessed in the upstream province of the VMD by calculating 33 hydrologic variability parameters in five groups including magnitude, duration time, timing of extreme flow, and frequency which impact on sediment and nutrition transport, and thus effect the diversity of downstream flora and fauna. The results from this study would provide an assessment of the scale of change and further research on the impacts of flow-regime changes on aquatic diversity going forward is needed. Naturally, such large-scale alteration to the hydrological regime can create significant threats to both plant and animal species and lead to undesirable environmental impacts. The above-mentioned issue has been recently discussed in detail by several studies such as Li et al. (2017), Hecht et al. (2019) and Gunawardana et al. (2021).

According to Li et al. (2017), human activities, such as dam construction, significantly altered the flow regimes in the Mekong River, particularly after the completion of two large dams, namely Xiaowan and Nuozhadu in the years 2010 and 2014, respectively. The construction and operation of dams clearly has significant impacts on low-flow duration. It has been observed that climate change dictated the changes in the annual streamflow during the transition period 1992–2009, whereas human activities contributed more in the post-impact period 2010–2014.

Hecht et al. (2019) reviewed the hydrological impacts due to hydropower dams in the Mekong River Basin and concluded that the basin's hydropower reservoir storage, which may rise from ∼2% of its mean annual flow in 2008 to ∼20% in 2025, is attenuating seasonal flow variability downstream of many dams with integral powerhouses and large storage reservoirs. In addition, tributary diversions for off-stream energy production are reducing downstream flows and augmenting them in recipient tributaries. According to Gunawardana et al. (2021), among the IHA parameters, the fall rate has the most significant effect on hydrological alteration of all drivers. Hydropower development in the basin mostly affects the fall rate and reversal. Identifying the connection between these multiple drivers and hydrological alteration could help decision-makers to design more efficient and sustainable water management policies.

From all the above discussion, therefore, the change in hydrological indicators at Tan Chau and Chau Doc stations may be due not only to the province's dike system and climate change, but also to upstream changes in flow. As a result more research is needed to fully understand the factors that have led to the observed changes in flow regime.

Figure 9 shows the downward trend in the wet season and upward trend in the dry season between 2000 and 2018 as observed in the annual discharge at Can Tho station. This suggests that there was no significant trend for both seasons. More observational data is required in order to establish a scientific basis for the flow variance at Can Tho station as a result of natural and/or human influences.

Upstream An Giang province in the VMD has strongly developed protection dikes, especially during the periods of 1997–2004 and 2007–2010. By 2011, the dike system regulated 69% of the province's total area, of which semi-dikes covered 15% and full dikes 54%. The results of the flow-regime-change assessment showed that the flow alteration at both Chau Doc and Tan Chau stations in periods 1 and 2 was at a high level (over 67.0%); and in Vam Nao during periods 1 and 2 it was 49.8% and 60.7%, respectively. In general, during the construction period of the dike system (1997–2010), Chan Doc station was the most affected (71.2%), followed by Tan Chau (68.2%) and Vam Nao was the least (49.8%). However, when considering period 2 (2011–2019) after the relative completion of dike system construction, the change in flow regime at both Tan Chau and Vam Nao stations was seen still to increase significantly, by 76.6% and 60.7%, respectively. Of the five IHA groups, group 5 showed the biggest change at all three stations. In particular, indicator 31 changed at a very high rate in both Chau Doc and Tan Chau; meanwhile, indicators 32 and 33 of group 5 at Vam Nao station had a significant change in both phases.

The change in hydrological indicators at Tan Chau and Chau Doc stations were due not only to the province dike system and climate change, but also to the change in upstream flow of the Mekong River. More research is needed to consider the various drivers of flow-regime alteration associated with natural and human activities both inside and external to the study area.

The research received no funds.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.