Abstract
Water demands and stresses in the Vietnam Mekong Delta (VMD) are of high concern for the coming decade. System dynamics modeling (SDM) can be used to understand the impacts of the changes on water resources at a regional scale. SDM framework was applied to assess the impacts of land use changes and climate on surface water resource in the VMD. The developed model was modified from a global-scale model and added in-situ specific variables to model the conditions of the VMD using historical data during 1986–2005, and future projections to 2050 under two Representative Concentration Pathways scenarios (RCP4.5 and RCP8.5). Projected changes in land use and population were based on the most recent governmental resolution to support adaptation and the national plans. Temperature and rainfall changes cause the increase in water stress; however, it could be mitigated by shifting from three to two crops of rice each year. Water stress differs from place to place, such as in freshwater and brackish coastal zones. Water stress is most strongly affected by changed temperature, rather than rainfall. The developed system dynamics model and framework can be effectively applied in the VMD to support decision-making about sustainable water use for paddy rice.
INTRODUCTION
Water resource assessment is a key step to understand and adapt to climate change, urbanization, and industrialization. A system dynamics modeling (SDM) approach (Sahin et al. 2016) can be used for assessing the cumulative impacts of climatic changes and human activities. For example, two main components in a water assessment model are the water use sector (domestic, industrial, and agricultural water use) and hydrological sector (rainfall, surface runoff, and groundwater) (Alcamo et al. 2003). In addition, the interactions among available water supplies and socio-economic and environmental factors were based on the combination of various sectors, including climate, carbon cycle, land use, population, economic, water demand, water quality, and hydrological cycle sectors (Davies & Simonovic 2011; Akhtar et al. 2013). For regional-scale modeling, the amount of water demand and supply was simulated based on the changes of population and water availability until 2050 to compare alternative policy scenarios in Las Vegas, Nevada, USA (Stave 2003; Qaiser et al. 2011). In another study, SDM was used as a useful management tool for assessing the water scarcity in a coastal area, including the relation between water demands of different socio-economic sectors and local hydrology settings in Kairouan, Tunisia (Sušnik et al. 2012). SDM was used for water management in Bear River Basin in Utah and Wyoming, USA (Sehlke & Jacobson 2005), Red River Basin and British Columbia, Canada (Simonovic & Ahmad 2005; Langsdale et al. 2009), Yulin and Bayingolin, China (Wang et al. 2011; Sahin et al. 2015) and Volta River Basin, Ghana (Kotir et al. 2016). These have demonstrated the efficiency of the SDM approach for water resources assessment and allocation in specific places.
The Vietnamese Mekong Delta (VMD), located in the lowest downstream area of the Mekong River, receives a high surface flow from the upstream countries (Food and Agriculture Organization (FAO 2011)) (Figure 1). The VMD has two distinct seasons, with the dry season from November to April and the wet season from May to October, respectively (Sona et al. 2012; Shrestha et al. 2016; Pham-Duc et al. 2017). The VMD contributed more than 56% of paddy rice production in all of Vietnam in 2015 (General Statistics Office of Vietnam (GSO 2016)). Climate change strongly affects agriculture, the greatest economy in the VMD, through the changes in freshwater quantity (Sebesvari et al. 2012). The upstream areas are flooded during the annual wet season. The extreme floods damage the infrastructure, buildings, human life, and crops. During a 20-year period (1995–2015), there were three extreme flood events, which occurred in 2000, 2001, and 2011 (Hanington et al. 2017). In addition, the coastal areas of the VMD face salinity intrusion in freshwater under the impact of climate change, affecting the agricultural and domestic water supply (Renaud et al. 2015). For example, there was extreme drought and salinity intrusion in the dry season during 2015–2016, affecting nine of 13 provinces (Nguyen 2017; Trinh et al. 2018). In addition, the lack of surface freshwater has led to the increase of groundwater pumping, causing land subsidence in the coastal districts (Erban et al. 2014; Minderhoud et al. 2017).
The Mekong River Basin (left), and the Vietnamese Mekong Delta and main hydrologic stations (right). The Mekong catchment includes six countries and is 795,000 km2. The average annual flow of the Mekong is 475 km3/year. The Delta comprises 39,000 km2, and is roughly the size of the Netherlands or Switzerland.
The Mekong River Basin (left), and the Vietnamese Mekong Delta and main hydrologic stations (right). The Mekong catchment includes six countries and is 795,000 km2. The average annual flow of the Mekong is 475 km3/year. The Delta comprises 39,000 km2, and is roughly the size of the Netherlands or Switzerland.
Previous hydraulic and hydrological modeling has examined the impacts of climate change on the livelihoods of local residents living in the VMD through influences on water resources. The impacts of various human activities on the hydrological system have been less studied. For example, hydraulic and hydrological modeling has focused on the impacts of climate change on dyke systems and flood events (Tri et al. 2012), and on flood hazards (Van et al. 2012; Hanington et al. 2017). In addition, the salinity intrusion under the impacts of different scenarios of sea level rise and upstream discharge changes was studied using hydraulic models in the VMD (Smajgl et al. 2015; Trinh et al. 2018).
The policies of Vietnam focus on sustainable agricultural development in the Mekong Delta, and the development of industry and services is also directed towards protecting the ecosystem. In 2014, the Government of Vietnam issued a plan for developing the VMD (Decision 245), including socio-economic development, saving natural resources, protecting ecology environment, mitigating and adapting to climate change and sea level rise. The objective of this plan is to develop the VMD as a key economic region of Vietnam. In addition, the current Resolution 120 for Development of the Vietnamese Mekong Delta to adapt to climate change, issued by the Government of Vietnam in 2017, determined that cultivation systems which lower effective commercial benefit should be reduced to protect the ecosystem. This study developed a surface water assessment model; namely, Surface Water Resource-Vietnamese Mekong Delta (SWR-VMD), for paddy rice production in the VMD using SDM. The objectives of this study are: (1) to understand the changes of dynamics of surface water in several future scenarios, including projected climatic changes (temperature and rainfall changes) and human activities (domestic, industrial, and irrigation activities) for the VMD; and (2) to evaluate whether the proposed solution (land use change) can solve identified problems.
METHODS: MODELING DESCRIPTION
System dynamics approach
Figure 2 illustrates the loop diagram of the SWR-VMD model indicating the climate, land use, and socio-economic variables on surface water resources in the VMD. The increases in rainfall affect surface flow positively while temperature causes a loss due to water consumption. Water stress would decrease due to high surface water flow but increase due to high water withdrawal. The strong impacts on surface water resource include climate and land use in the present model. The future climate, irrigated and industrial land area were based on (1) the climate change scenario report issued by the Ministry of Natural Resources and Environment (MONRE 2016), (2) Resolution 120, and (3) Decision 1581 (issued by the Government of Vietnam in 2009) of the plan for developing the VMD until 2020 and vision for 2050.
Loop diagram of the SWR-VMD model; (+) and (−) illustrate the positive and negative relationships between the variables in the model, respectively.
Loop diagram of the SWR-VMD model; (+) and (−) illustrate the positive and negative relationships between the variables in the model, respectively.
Structural model description
The terrestrial water hydrology was modeled, including average rainfall, evapotranspiration (ET), and surface flow at Tan Chau and Chau Doc (Figure 1). Surface water storage is lost due to ET, infiltration from the surface to subsurface, and surface flow. The loss of water in subsurface storage is the flow from the subsurface into rivers and the ocean (Simonovic 2012). The model structure of the hydrological components is shown schematically in Figure 3. The term ‘Local Surface Flow’ is the surface water flow in the VMD, not including the contribution from upstream countries. The controlling variables include ‘Average Future Rainfall’ and ‘Sectoral Water Consumption’.
Structure of hydrological component in the SWR-VMD model. The terms in the circle present the strong affecting variables.
Structure of hydrological component in the SWR-VMD model. The terms in the circle present the strong affecting variables.
The water stress is mainly affected by ‘Temperature’, ‘Irrigated Land Use Area’, and ‘Total Renewable Flow’ (Figure 4). The terms ‘ET ref’ and ‘ET crop’ present the reference and crop ET from irrigation activity. The ‘Sectoral Water Consumption’ and ‘Sectoral Water Withdrawal’ are the water consumption and withdrawals in domestic, industrial, and irrigation activities, respectively. The sectoral water withdrawal depends on population, industrial land area, and irrigated land use area of paddy rice.
Structure of water demand component in the SWR-VMD model. The terms in the circles present the strong affecting variables.
Structure of water demand component in the SWR-VMD model. The terms in the circles present the strong affecting variables.
Mathematical description
Table 1 shows the mathematical description of the SWR-VMD model, explaining Figures 3 and 4. Based on the global integrated assessment models (Davies 2007; Davies & Simonovic 2008; Akhtar 2011), the SWR-VMD model was modified for local conditions of climate, land use types, and population in the VMD (in 2000–2015). The modifications include: (1) the water requirement of irrigation water per hectare which was calculated from crop water need referenced from Food and Agriculture Organization (FAO) (www.fao.org/docrep/S2022E/s2022e07.htm); (2) estimation of base domestic water consumption (water demand per capita); and (3) the industrial water consumption (industrial water demand per hectare), based on the water requirement according to the National Standard of Vietnam 33:2006 issued by Ministry of Construction of Vietnam (MOC 2006). In addition, this model was simulated in a small time step (month) compared to the original model (year). By using a small time step, the results of the model can reflect the dynamics of surface water resource in more detail. In Table 1, is current rainwater flow (at time t) over the surface (m3/month);
is average rainfall per month during 1986–2005 (mm/month−1);
is the area of the VMD basin (km2);
is current ET from surface to the atmosphere, respectively (m3/month);
and
are initial (at t= 0) and current reference ET (mm/day−1);
is the crop water requirement (mm/day−1);
is number of dry days in the initial month (at t = 0), which was calculated as the number of days minus the number of rainy days in a month; d is the number of days in a month;
and
are initial and current surface water flows from surface to oceans, respectively (m3/month);
and
are initial and current infiltration flows from surface to subsurface, respectively (m3/month);
and
are initial and current groundwater flows from subsurface to rivers and oceans, respectively (m3/month);
and
are initial and current surface water storages, respectively (m3);
and
are initial and current subsurface water storages, respectively (m3);
,
,
and
are water consumption added to the atmosphere, subsurface, surface and water loss, respectively, through human activities (m3/month) (referenced from the model of Davies (2007));
is groundwater withdrawal (m3/month);
,
, and
are domestic, industrial, and irrigation water withdrawals, respectively (m3/month);
,
, and
are domestic, industrial, and irrigation water consumption, respectively (m3/month);
and
are initial and current population sizes (people);
is population increase per month (people/month);
and
are initial and current population growth rate (1/month);
is decline in the population growth rate (1/month2); b is a coefficient that matches simulated values with historical population;
is monthly water stress;
and
are industrial area and irrigated land use area for paddy rice, respectively (ha);
is domestic water requirement per capita (m3/person/month);
is industrial water requirement per ha (m3/ha/month);
is irrigation water requirement per ha (m3/ha/month);
,
, and
are fractions of domestic, industrial, and irrigation water consumption per withdrawal, respectively;
is total polluted water (m3/month) (referenced from the model of Davies (2007)); p is the average daily percentage of annual daytime hours depending on the latitude of the area;
is the mean temperature (°C); and
is a crop coefficient.
Main equations in the SWR-VMD model
Equation number . | Equation . | New point . |
---|---|---|
(1) | ![]() | Time step (t) was modified from ‘year’ to ‘month’ |
(2) | ![]() | Time step (t) was modified from ‘year’ to ‘month’ |
(3) | ![]() | ![]() |
(4) | ![]() | Time step (t) was modified from ‘year’ to ‘month’ |
(5) | ![]() | Time step (t) was modified from ‘year’ to ‘month’ |
(6) | ![]() | Time step (t) was modified from ‘year’ to ‘month’ |
(7) | ![]() | It was modified based on average rainfall per month and the VMD area (39,000 km2) (Quang 2002) |
(8) | ![]() | Time step (t) was modified from ‘year’ to ‘month’ |
(9) | ![]() | It was modified based on the reference ET per day (ww.fao.org/docrep/S2022E/s2022e07.htm) |
(10) | ![]() | ![]() |
(11) | ![]() | ![]() |
(12) | ![]() | ![]() |
(13) | ![]() | ![]() |
(14) | ![]() | The irrigated water requirement was estimated for agricultural activity based on the FAO Blaney–Criddle function where weather data is not sufficient (www.fao.org/docrep/S2022E/s2022e07.htm) |
(15) | ![]() | It was referenced from www.fao.org/docrep/S2022E/s2022e07.htm |
(16) | ![]() | Time step (t) was modified from ‘year’ to ‘month’ |
(17) | ![]() | Time step (t) was modified from ‘year’ to ‘month’ |
(18) | ![]() | Time step (t) was modified from ‘year’ to ‘month’ |
(19) | ![]() | Coefficient b was adjusted to calibrate the population |
(20) | ![]() | The estimated water availability is the usable freshwater (that is, the difference between total surface water and polluted water) |
Equation number . | Equation . | New point . |
---|---|---|
(1) | ![]() | Time step (t) was modified from ‘year’ to ‘month’ |
(2) | ![]() | Time step (t) was modified from ‘year’ to ‘month’ |
(3) | ![]() | ![]() |
(4) | ![]() | Time step (t) was modified from ‘year’ to ‘month’ |
(5) | ![]() | Time step (t) was modified from ‘year’ to ‘month’ |
(6) | ![]() | Time step (t) was modified from ‘year’ to ‘month’ |
(7) | ![]() | It was modified based on average rainfall per month and the VMD area (39,000 km2) (Quang 2002) |
(8) | ![]() | Time step (t) was modified from ‘year’ to ‘month’ |
(9) | ![]() | It was modified based on the reference ET per day (ww.fao.org/docrep/S2022E/s2022e07.htm) |
(10) | ![]() | ![]() |
(11) | ![]() | ![]() |
(12) | ![]() | ![]() |
(13) | ![]() | ![]() |
(14) | ![]() | The irrigated water requirement was estimated for agricultural activity based on the FAO Blaney–Criddle function where weather data is not sufficient (www.fao.org/docrep/S2022E/s2022e07.htm) |
(15) | ![]() | It was referenced from www.fao.org/docrep/S2022E/s2022e07.htm |
(16) | ![]() | Time step (t) was modified from ‘year’ to ‘month’ |
(17) | ![]() | Time step (t) was modified from ‘year’ to ‘month’ |
(18) | ![]() | Time step (t) was modified from ‘year’ to ‘month’ |
(19) | ![]() | Coefficient b was adjusted to calibrate the population |
(20) | ![]() | The estimated water availability is the usable freshwater (that is, the difference between total surface water and polluted water) |
MATERIALS AND ASSUMPTIONS
Upstream flow
The upstream flow was calculated by the total flow per month of the Chau Doc and Tan Chau stations during 1977–1999 (Figure 1 and Table 2) estimated from the Mekong River Commission (Renaud & Kuenzer 2012). The lowest surface flow is 6.1 km3/month in April, and the highest flow is 69.2 km3/month in October. Although the data were estimated from the monthly average surface flow (m3/second), it could be used to assess the changes in available water in the future. The model focuses on studying the dynamics in the future of surface water resources within the VMD under climate change impact in the case of unchanged surface flow in Tan Chau and Chau Doc.
Total upstream surface flow in Tan Chau and Chau Doc during 1977–1999 (km3/month)
Station . | Jan . | Feb . | Mar . | Apr . | May . | Jun . | Jul . | Aug . | Sep . | Oct . | Nov . | Dec . |
---|---|---|---|---|---|---|---|---|---|---|---|---|
Tan Chau | 16.7 | 9.0 | 7.0 | 5.2 | 7.1 | 18.6 | 30.2 | 43.9 | 54.8 | 54.5 | 39.6 | 27.3 |
Chau Doc | 3.6 | 1.7 | 1.1 | 0.9 | 1.2 | 3.8 | 6.4 | 10.6 | 13.5 | 14.7 | 12.2 | 7.3 |
Total flow | 20.3 | 10.7 | 8.1 | 6.1 | 8.3 | 22.4 | 36.6 | 54.5 | 68.3 | 69.2 | 51.8 | 34.6 |
Station . | Jan . | Feb . | Mar . | Apr . | May . | Jun . | Jul . | Aug . | Sep . | Oct . | Nov . | Dec . |
---|---|---|---|---|---|---|---|---|---|---|---|---|
Tan Chau | 16.7 | 9.0 | 7.0 | 5.2 | 7.1 | 18.6 | 30.2 | 43.9 | 54.8 | 54.5 | 39.6 | 27.3 |
Chau Doc | 3.6 | 1.7 | 1.1 | 0.9 | 1.2 | 3.8 | 6.4 | 10.6 | 13.5 | 14.7 | 12.2 | 7.3 |
Total flow | 20.3 | 10.7 | 8.1 | 6.1 | 8.3 | 22.4 | 36.6 | 54.5 | 68.3 | 69.2 | 51.8 | 34.6 |
Climate and climate change
According to IPCC (2013), four Representative Concentration Pathway scenarios (RCP2.6, RCP4.5, RCP6.0, and RCP8.5) are based on radiative forcing values (2.6, 4.5, 6.0, and 8.5 W/m2, respectively) in the atmosphere in 2100, in which the lowest and highest greenhouse gas emission scenarios are RCP2.6 and RCP8.5, respectively. In Vietnam, the most possible scenario (RCP4.5) and the worst case (RCP8.5) were chosen to assess the future changes of climate. MONRE (2016) published a report of temperature and rainfall changes under RCP4.5 and RCP8.5 for each region of Vietnam, including the VMD. The report was based on the results of five regional models, including the Meteorological Research Institute Atmospheric General Circulation Model (AGCM/MRI) (for example, Murakami et al. 2012); Providing Regional Climates for Impacts Studies (for example, de Carvalho et al. 2011); Conformal Cubic Atmospheric Model (for example, Chiew et al. 2010); Regional Climate Model (RegCM) (for example, Bhatla et al. 2016); and Climate-Weather Research and Forecast (for example, Zittis et al. 2013).
The average rainfall values of 11 main hydrologic stations, including Moc Hoa, My Tho, Ba Tri, Chau Doc, Cao Lanh, Can Tho, Cang Long, Soc Trang, Bac Lieu, Rach Gia, and Ca Mau in the VMD during 1986–2005 (http://econ.worldbank.org) are shown in Figure 5(a). Figure 5(b) shows the average rainfall changes by seasons of 11 hydrologic stations during 2046–2065 under the two climate change scenarios (RCP4.5 and RCP8.5) compared to the base period (1986–2005). Average rainfall by season changes in both positive and negative directions, including the largest decrease in spring in Ca Mau (−9.4% under RCP4.5) and the highest increase in winter in Bac Lieu (158.0% under RCP8.5) (MONRE 2016). The climate changes are different in the four seasons, including winter (between December and February), spring (between March and May), summer (between June and August), and autumn (between September and November).
(a) Average rainfall of 11 main hydrologic stations during 1986–2005 (mm/month), and (b) average rainfall changes by seasons (%) of 11 hydrologic stations during 2046–2065, under RCP4.5 and RCP8.5, compared to the base period in the VMD (reference: MONRE (2016)).
(a) Average rainfall of 11 main hydrologic stations during 1986–2005 (mm/month), and (b) average rainfall changes by seasons (%) of 11 hydrologic stations during 2046–2065, under RCP4.5 and RCP8.5, compared to the base period in the VMD (reference: MONRE (2016)).
The average values of monthly temperature in the VMD by stations during 1982–2012 were collected from https://en.climate-data.org (Figure 6(a)). The seasonal average temperature of 11 stations of the VMD was projected to increase from 1.3 °C (in autumn under RCP4.5) to 2.0 °C (in winter under RCP8.5) in the period 2046–2065 (Figure 6(b)). The lowest increase occurs in Rach Gia (Kien Giang), and the highest increase occurs in My Tho (Tien Giang) (MONRE 2016).
(a) Average temperature of 11 stations during 1982–2012 (°C) and (b) average temperature changes by seasons (°C) during 2046–2065, under RCP4.5 and RCP8.5, compared to the base period in the VMD (reference: MONRE (2016)).
(a) Average temperature of 11 stations during 1982–2012 (°C) and (b) average temperature changes by seasons (°C) during 2046–2065, under RCP4.5 and RCP8.5, compared to the base period in the VMD (reference: MONRE (2016)).
Domestic and industrial activities
The total area of industrial zones in the VMD also increased with time. According to the survey, the area of industrial zones was 24,000 ha in 2005. Decision 1581 projected that the industrial land area will increase to 30,000 and 50,000 ha in 2020 and 2050, respectively. According to National Standard of Vietnam 33:2006, the base industrial water withdrawal determined from the standard daily industrial water requirement is 45 m3/ha/day; and the average standard daily domestic water use was estimated to be 120 and 150 liters per capita per day in 2010 and 2020, respectively (MOC 2006). The amount of monthly domestic water withdrawal per capita was estimated by extrapolation (that is, the annual increase is approximately 3 liters per capita per day). Thus, the domestic water use was projected to be 240 liters per capita per day in 2050.
The groundwater withdrawal is mainly used for domestic and industrial activities (FAO 2011). In some cases, paddy rice flood irrigation can recharge groundwater, but this is site-specific and is not modeled. The total groundwater withdrawal is approximately 2.5 million m3/day in 2015 (Minderhoud et al. 2017). In the coastal areas of the VMD, the over-exploitation of groundwater can cause land subsidence, floods, and salinity intrusion; therefore, the amount of groundwater withdrawal should not be increased (FAO 2011; Erban et al. 2014; Shrestha et al. 2016). Therefore, the groundwater withdrawal in 2015 was assumed to be unchanged in the future in this model.
Scenarios of climate and land use changes
Nine scenarios of land use changes to adapt to climate change and sea level rise effects on freshwater resource were simulated for the year 2050 (Figure 7 and Table 3). Based on the land use map conducted by College of Environment and Natural Resources of Can Tho University, Vietnam in 2010, there are three types of paddy rice crops, including WS: winter and spring (from November to March); SA: summer and autumn (from April to August); and AW: autumn and winter (from August to December). In the case of land use in the VMD, Figure 7(a) indicates five types of land use for paddy rice, including one kind of triple rice (three rice crops per year, WS-SA-AW), two kinds of double rice (WS-SA and SA-AW), and two kinds of single rice (WS and AW) (Bouvet et al. 2009; Chen et al. 2011). According to Resolution 120, the irrigated land use types with low commercial benefit would be reduced in the future. The paddy rice system is projected to face salinity intrusion due to projected sea level rise of 30 cm in 2050 (the isoline of salinity of 4 g/L – the threshold at which rice can grow stably) (Smajgl et al. 2015). Thus, there are four considerable cases of land use changes in the future, including unchanged land use in 2010 (Figure 7(a)), land use is changed from the triple rice system to WS-SA in the upper zone due to flood events in the wet season (Figure 7(b)), SA-AW in the coastal zone due to the risks of salinity intrusion and drought in the dry season (Figure 7(c)), and land use is changed in both the upper zone in the wet season and coastal zone in the dry season (Figure 7(d)).
Description of climate change scenarios and land use changes on surface water use in 2050 and base scenario in 2015
Scenario . | Climate . | Land use . | Simulated year . |
---|---|---|---|
1 | Base period | Base land use system in 2010 | 2015 |
2 | RCP4.5 | Unchanged land use in 2010 | 2050 |
3 | RCP8.5 | ||
4 | RCP4.5 | Land use change in the upper zone: Only triple rice in the upper zone is shifted to WS-SA | 2050 |
5 | RCP8.5 | ||
6 | RCP4.5 | Land use change in the coastal zone: Only triple rice in the coastal zone is shifted to SA-AW | 2050 |
7 | RCP8.5 | ||
8 | RCP4.5 | Land use change in the whole VMD: Triple rice systems in both upper and coastal zones are shifted to WS-SA and SA-AW, respectively | 2050 |
9 | RCP8.5 |
Scenario . | Climate . | Land use . | Simulated year . |
---|---|---|---|
1 | Base period | Base land use system in 2010 | 2015 |
2 | RCP4.5 | Unchanged land use in 2010 | 2050 |
3 | RCP8.5 | ||
4 | RCP4.5 | Land use change in the upper zone: Only triple rice in the upper zone is shifted to WS-SA | 2050 |
5 | RCP8.5 | ||
6 | RCP4.5 | Land use change in the coastal zone: Only triple rice in the coastal zone is shifted to SA-AW | 2050 |
7 | RCP8.5 | ||
8 | RCP4.5 | Land use change in the whole VMD: Triple rice systems in both upper and coastal zones are shifted to WS-SA and SA-AW, respectively | 2050 |
9 | RCP8.5 |
Land use change projection for 2050, including (a) unchanged land use, (b) only triple rice system in the upper zone is shifted to WS-SA (winter spring–summer autumn), (c) only triple rice system in the coastal zone is shifted to SA-AW (summer autumn–autumn winter), and (d) triple rice system is shifted to WS-SA in the upper zone and SA-AW in the coastal zone; the upper zone (from the black bold line to upstream of the main rivers) receives freshwater and the coastal zone (from the black bold line to the oceans) faces salinity intrusion in the future (the isoline of salinity of 4 g/L was referenced from Smajgl et al. (2015)). Please refer to the online version of this paper to see this figure in colour: http://dx.doi.org/10.2166/wcc.2019.176.
Land use change projection for 2050, including (a) unchanged land use, (b) only triple rice system in the upper zone is shifted to WS-SA (winter spring–summer autumn), (c) only triple rice system in the coastal zone is shifted to SA-AW (summer autumn–autumn winter), and (d) triple rice system is shifted to WS-SA in the upper zone and SA-AW in the coastal zone; the upper zone (from the black bold line to upstream of the main rivers) receives freshwater and the coastal zone (from the black bold line to the oceans) faces salinity intrusion in the future (the isoline of salinity of 4 g/L was referenced from Smajgl et al. (2015)). Please refer to the online version of this paper to see this figure in colour: http://dx.doi.org/10.2166/wcc.2019.176.
The impact of surface water resource on human activities is presented through water stress. Water stress reflects the relation between water demand and available water for supply. Large water stress means lack of water leading to drought and salinity intrusion. High water amount causes flooding; however, the annual flood events come from the upstream countries in the upper zone. The SWR-VMD focuses on studying the water stress in the VMD in the condition of unchanged surface flow from upstream countries.
RESULTS
Projected population
The water stress multiplier (coefficient b in Equation (19)) was adjusted to calibrate the simulated population to meet the historical population from 2001 to 2015. The coefficient b is 0.02 and the coefficient of determination is R2 = 0.996 (Figure 8). The simulated population is approximately 18,650 thousand people in 2050, increasing by 990 thousand people compared to 2015.
Historical population from 2000 to 2015, and simulated population until 2050.
There are two population scenarios for the year 2050 according to the Mekong Delta Plan issued by the Government of Vietnam (GV 2013); that is, the size of the population will be about 15 million and 30 million people under the migration and rapid urbanization scenario, respectively. In the present study, the population was projected to be the average value of those two scenarios. That means the population size would continue increasing but the population growth rate would decline in the future due to migration related to water stress.
Projected rainfall and temperature
The average rainfall per month and temperature of the VMD, estimated based on the average rainfall data of 11 stations, are shown in Figure 9. Under the base scenario, the highest and lowest values of average rainfall are 309 mm/month (±37 mm/month) and 6 mm/month (±5 mm/month) in October and February, respectively; large changes occur from September to December under climate change scenarios (Figure 9(a)). Under the base scenario, the highest and lowest values of temperature are 28.6 °C (±0.5 °C) and 25.7 °C (±0.4 °C) in October and February, respectively; most of the temperature increases are significant under climate change scenarios (Figure 9(b)). In the dry season, the temperature values significantly increase but the rainfall values do not much change; this may cause the risk of water shortage, especially in March and April, which is the period of frequent salinity intrusion.
Historical and projected rainfall (a) and temperature (b) in the VMD in 2050. The error bars indicate the standard deviations of average values of rainfall at 11 hydrologic stations.
Historical and projected rainfall (a) and temperature (b) in the VMD in 2050. The error bars indicate the standard deviations of average values of rainfall at 11 hydrologic stations.
Impacts of climate change
Figure 10 illustrates model results, simulated monthly water stress in the VMD with climate change and human activities and the land use change. Water stresses vary based on the different temperature and rainfall values in different months in a year. Under the base scenario, water stresses range from 2.6% (±0.1%) to 32.3% (±3.2%) during the wet season (from May to October), and from 10.8% (±0.3%) to 69.3% (±7.9%) during the dry season (from November to April). Climate change and human impacts (Scenarios 2 and 3) lead to increases in water stress compared to Scenario 1; large increases occur during the dry season, such as 2.4%, 3.9%, and 6.2% in February, March, and April, respectively. Despite the large increase in April (6.2%), it is uncertain due to the high error (+7.9%).
Simulated monthly water stress under the nine scenarios of climate and land use changes in 2050. In the VMD, water stress increases under climate change (Scenarios 2 and 3); however, it can be solved by shifting from triple to double rice (from Scenario 4 to Scenario 9); and water stress differs depending on seasons and hydrological zones.
Simulated monthly water stress under the nine scenarios of climate and land use changes in 2050. In the VMD, water stress increases under climate change (Scenarios 2 and 3); however, it can be solved by shifting from triple to double rice (from Scenario 4 to Scenario 9); and water stress differs depending on seasons and hydrological zones.
Responses of land use changes to climate change
By shifting from triple to double rice systems in the upper zone, water stress would decrease by 1.1% in September and October under Scenarios 4 and 5 compared to Scenarios 2 and 3, respectively. For the land use change in the coastal zone (Scenarios 6 and 7), the significant decreases are 1.8%, 4.0%, and 5.7% in January, February, and March, respectively, compared to Scenarios 2 and 3 (Figure 10). Thus, the change from triple to double rice systems in the upper zone would mainly reduce water stress in the wet season (that is, in September and October); however, this reduction is less than the land use change in the coastal zone in the dry season. Considering the changes occur in both upper and coastal zones (Scenarios 8 and 9), water stress is projected to decrease in both the wet season (that is, in September and October) and dry season (that is, from January to March). Land use change strategy throughout the VMD is effective for reducing water stress in the dry season; however, it is not effective in the wet season because the base water stresses are low (Scenario 1).
Comparison between the two climate change scenarios
According to the SWR-VMD model, temperature increases lead to the increase in water stress on crops and increased water demand; however, the increase in rainfall results in a larger amount of available surface water, and a decrease in water stress. The RCP8.5 indicates greater increases in temperature compared to RCP4.5. Large water stresses under RCP8.5 compared to RCP4.5 demonstrate a strong effect of temperature (Figure 10).
DISCUSSION
The present research has used the average values of different sub-areas in the VMD as representative values; therefore, it does not account for small-scale differences. The assumptions and basis of this modeling are reasonably reliable, including the historical data (World Bank, Can Tho University), global projections (climate-data.org), previous studies (Vermeulen et al. 2013; Renaud et al. 2015; Shrestha et al. 2016; Minderhoud et al. 2017), national reports (GV 2013; MONRE 2016), standard (MOC 2006), Decision 1581 in 2009 and Resolution 120 in 2017. Although the Blaney–Criddle function could give high errors (up to 60–40%), the model could reflect the impact of temperature change on water stress for a large area in the future. This study presented a system dynamics approach by focusing on the future surface water resource for paddy rice production.
In the VMD, the upper zone is annually flooded due to high water flow from upstream countries (approximately 90% of water in the VMD comes from foreign countries) while the coastal zone usually faces a shortage of fresh water and salinity intrusion from the ocean (Tuan et al. 2007). The VMD is a flat land region with many canals and is bordered by the East and West Sea. The coastal zone was projected to be flooded due to sea level rise (Van et al. 2012). Therefore, the coastal zone is unlikely to be flooded by high local rainfall in the case of water flows in upstream countries being unchanged.
The large projected increase of rainfall will increase the water availability in the wet season; however, under the impacts of human water use and temperature increases, the resulting increased water stress will cause water shortages in the dry season. The model shows that irrigation water demand (ranging from 1.11 to 4.00 km3/month) is significantly greater than domestic and industrial water consumption, (0.14 and 0.07 km3/month, respectively) in 2050. Optimization of water use in irrigation is widely recommended and clearly helpful both to crops and overall water availability for multiple uses. Several effective solutions to optimize the water consumption have been recommended to farmers and managers to adapt to climate change in the VMD, including the further development of irrigation systems (such as canals, sluice gates, dams, and so on.). This requires coordination between farmers and local governments. Irrigated systems have been sharply expanded to improve paddy rice production in the VMD (Manh et al. 2014; Dang et al. 2016; Hanington et al. 2017). Therefore, there should be further adaptation efforts for surface water conservation in the VMD, including land use changes for paddy rice systems.
The reduction of rice crops (from three to two crops per year) is one of the reasons leading to higher prices of rice. This can help farmers stabilize profits; however, the continuous increase in rice prices may lead to hardship in competition with other countries. In order to increase farmers' income, strategies of replacing the third crop with other livelihoods (for example, aquaculture, services, tourist business, etc.) should be considered. In addition, the focus on growing rice of higher quality would also help farmers increase financial benefits and meet the requirements for export.
CONCLUSIONS
The model was examined under nine scenarios, including the changes of climate and land use for paddy rice systems until 2050. It indicates that SDM can be applied at regional scales such as the VMD, and can be a useful tool for decision-making. The SWR-VMD model indicates the changes in surface water status in the future under changing temperature and rainfall. Temperature changes will have a greater impact than rainfall changes on irrigation water demand during the dry season. It also indicates that the change of land use for paddy rice systems is a viable adaptation to climate change, especially for the crops in the dry season. However, the changes must be matched to each hydrological zone (that is, the upper and coastal zones). The shift of land use in the coastal zone to reduce water stress is significant in the dry season and less significant in the upper zone, considering climate change, land subsidence, and sea level rise.
The present model provides a general assessment of surface water resource, especially for paddy rice production. In the future, the studies using the SDM approach for different specific zones in the VMD (for example, provinces or districts) and other sectors need to be considered.
ACKNOWLEDGEMENTS
This project is supported by the R&D Center for Reduction of Non-CO2 Greenhouse Gases (2016001690005) funded by Korea Ministry of Environment (MOE) as Global Top Environment R&D Program. We are also grateful to Professor Slobodan P. Simonovic from the University of Western Ontario (Canada) for providing the materials of the integrated assessment model (IAM). Many thanks to Mr. Michael Furniss (Lecturer of Humboldt State University, USA) for reviewing this manuscript. Nguyen Thanh Tuu and Jeejae Lim are co-first authors.