Impact of climate change on groundwater recharge and salinity distribution in the Vientiane basin, Lao PDR

Groundwater is the world's most important source of freshwater (Todd & Mays 2005) and it is safer from pollution than surface water resources (Döll & Fiedler 2008). Groundwater is used freely in several areas and is not managed according to academic principles. In the past, humans believed that there was enough water to meet demand, so the water was widely used. At the moment, they are thinking about protecting long-term yield. The Vientiane basin (VTB) is one of Central Laos’s most important socioeconomic areas. Laos’s population growth and natural resource development included hydropower, mining, industrial processing plants, and groundwater extraction projects. Groundwater resources may also be extensively used by factories and households. The trend of groundwater use for these purposes is increased. However, the use of the mentioned groundwater is without proper study and monitoring for both quantity and quality.

The sustainable use of groundwater is recharge, which is less or comparable to renewable groundwater resources (Hahn et al. 1997; Döll & Fiedler 2008). Groundwater recharge at the global scale was simulated by Döll & Fiedler (2008). On a regional scale, it is found that some studies have been conducted in order to estimate groundwater recharge (Scanlon et al. 2002; Saraphirom et al. 2013a; Lacombe et al. 2017; Pholkern et al. 2018). Groundwater recharge is a small portion of runoff, which makes the Asian Monsoon region highly susceptible to rainfall variability and pollution (Döll & Fiedler 2008).

Many areas around the world that were salt-affected as assessed by the FAO in 2000 were greater than 8 million km² (Martinez-Beltran & Manzur 2005) and spread across continents (Green et al. 2011). Salinity has an impact on both surface and groundwater sources, agricultural products, and the environment. The processing of salinity is essential in processing nearly a link to the landscaped processed formations. However, the activity of humanity could be accelerated by processed salinization and contribute to the long-standing degradation of the environment (Pholkern et al. 2018). Salinization of soils and waters is an inevitability and associated with problems around the world in irrigation, agriculture, and a difficult problem for sustainable harvest output (Lin & Garcia 2008; Gates et al. 2009).

Climate change is a global phenomenon, but its consequences are felt at the regional and local levels. Generally, these climate change models are created based on the CMIP5 and as per the 5th IPCC Assessment Report (AR5) (IPCC 2013). Although groundwater is the most important freshwater source, particularly in rural areas far from rivers, and is the best alternative, research on the effects of climatic variables on groundwater resources is limited (IPCC 2019). In the Mekong region, a few studies have examined the impact of climate change on groundwater resources (Saraphirom et al. 2013a, 2013b; Pholkern et al. 2018; Petpongpan et al. 2020). Climate change factors (CFs), especially variability in temperature and precipitation, have a significant impact on various hydro-meteorological variables as well as the stability of water resources (Arab Amiri & Gocić 2021).

General Circulation Models (GCMs) play an important role in climate projection for many centuries. Because each GCM has a different assumption and approach, the historical outputs of six GCMs are initially screened (MRCS 2014; Petpongpan et al. 2020) by comparing them to observation data from 1976 to 2005. Precipitation is the variable under consideration since it has a significant impact on the hydrological process. The average monthly precipitation collected from GCMs and rainfall stations is compared to identify three GCMs that have the highest concordance with measured data.

The precipitation, temperature, and evapotranspiration variables that are usually used in various models to predict the future are potentially affected by climate change but need to be either analyzed or bias-corrected before being used (Amiri & Gocić 2021; Arab Amiri & Gocić 2021; Gocić & Arab Amiri 2021). However, precipitation and temperature are two variables that are still mostly used to predict climatic scenarios (Hanasaki et al. 2013; Saraphirom et al. 2013a; IWMI 2016; Pholkern et al. 2018; Petpongpan et al. 2020). In this study, CFs were used for computing by considering differences and ratios of GCMs in the future and historical periods, then adding or multiplying with observed temperature and precipitation data (Petpongpan et al. 2020).

This study aims to predict the impact of climate change on groundwater recharge and salinity distribution in Laos using numerical simulations. The pilot study area is the VTB in the central part of Laos. The information from this study can be used for future management planning and implementation of groundwater management and governance in Laos and also for the study of the transboundary aquifer of the GMS region.

The station's most comprehensive meteorological records at the VTB (DMH 2020) and eight rainfall stations during the years 1976–2020 indicate a yearly rainfall average of 2,037 mm/year with an increasing trend over the recorded period. The climate in the research area is distinguished by two different seasons: rainy and dry. The average temperatures for the minimum, mean, and maximum were 14.8, 26, and 34.6 °C, respectively. The study area comprises the Mekong River, which is the main river that flows through the study area from south to southeast with a width of 247–267 m, a depth of 18–20 m, with the following three stations: Pakmong, Vientiane, Nong Khai. Besides, the middle part is the Nam Ngum River, which is a Mekong River branch, that is, flow begins from the northwest of the study area, flows through the middle part, and then flows to the Mekong River in the southeast of the VTB.

In this work, the HELP3 model was employed to evaluate the recharge rates of groundwater. Moreover, it was used to investigate the effect of climatic changes on groundwater recharge rates in several areas around the world such as New Jersey, the Grand River Watershed, and the Huai Khamrian subwatershed in Thailand (Jyrkama et al. 2002; Jyrkama & Sykes 2007; Saraphirom et al. 2013a). For a comprehensive explanation of the HELP model, see Schroeder et al. (1994). The HELP3 model is a semi-2D, deterministic water routing model that is used to compute the balances of water. It replicates the daily water transport on the earth and takes into consideration rainfall, waterbody, runoff, evapotranspiration, vegetal interception, a flow that is not saturated, and temperature impacts as factors to consider (Saraphirom et al. 2013a; Pholkern et al. 2018).

The parameters required for the HELP3 model are provided by Jyrkama & Sykes (2007) and Saraphirom et al. (2013a). Groundwater recharge in 2010–2020 simulations was estimated using meteorological stations in Vientiane and Phonhong. The six future climatic scenario output data from the MPI-ESM-MR, CNRM-CM5, and MIROC5 under RCP 4.5 and 8.5 scenarios were used as inputs to evaluate the recharge during the next 30 years.

The MODFLOW model classified recharge zones based on land use, soil type, and ground surface slope (FIPD 2015; Pholkern et al. 2018; NAFRI 2020). The soil profile information was conducted to design the soil column in the HELP3 model (Saraphirom et al. 2013a). In the groundwater model, 13 recharge zones were given, as illustrated in Table 1 and Figure 4(a). The HELP3 results were used as inputs for each zone in MODFLOW and MT3D. The calibrated model suggests that recharge rates range from 0.75 to 15% of rainfall in discharge and recharge areas, respectively.

The model's domain is 95 × 112 km in size and 900 m in depth. It has 32,864 active cells, and a uniform square grid cell has a resolution of 1,000 m divided into 102 rows, 95 columns horizontally, and divided into eight layers of varying depths. The thickness of the model in each layer ranged from 10 to 40 m, with elevations varying from 0 to 900 m amsl. Layer thickness in the model varies with the topography; layers were assigned to be 20–60 m thick, as shown in Figure 4(c). The uppermost layer is considered an unconfined aquifer, whereas the remaining layers are considered confined aquifers according to the equated geologic unit between the VTB and the Sakon Nakhon basin. So, flow and mass transport parameters of the Central Huai Luang Basin, and Huai Khamrian subwatershed, Northeast Thailand, can be used in the VTB (Saraphirom et al. 2013a; Pholkern et al. 2018). The important parameters are properly used in MODFLOW and MT3D model simulations as shown in Table 2.

The river network was used to establish river boundaries in MODFLOW (Figure 4(b)). The characterized river input parameters, such as stage, width, and conductivity from previous study information and field investigation results were used. The river stages were considered to be unchanged from the current condition in the projection simulations. The boundaries of the watershed from the west to the northeast of the groundwater divide ware assigned as a no-flow boundary, as well as the lateral boundaries of the subwatershed. The initial head was created by 41 observed data in September 2014 and the initial concentration was created by 25 observed data. The rock salt (K2tn) under the Vientiane and Saysomboun units was designated as a constant boundary with a concentrate of 100,000 mg/l in layers 5–8 in the lower part of the VTB (Srisuk et al. 1999; Saraphirom et al. 2013a; Pholkern et al. 2018, 2019).

The model was verified using 41 observations well data from 2014 to 2016 (DWR 2015; NRERI 2016). The findings reveal that the absolute residual mean is 1.12 m, the RMS error is 1.36 m, and the normalized residual mean is 4.03%. The flow model performs well when compared to the observation data.

To determine which parameters were most influencing the models, sensitivity analyses of the MODFLOW, MT3D, and HELP3 models’ parameters and boundary conditions were performed.

The effects of the changes on groundwater recharge and salinity distribution in the VTB were simulated by using verified groundwater flow and saline transport models. The recharge rates in the future condition depend on changes in climate from three GCMs (MPI-ESM-MR, MIROC5, and CNRM-CM5) with RCP 4.5 and 8.5 scenarios. The rainfall and temperature data from GCMs were retrieved and used in HELP3. The recharge model estimation results by HELP3 were used in the MODFLOW and MT3D models to predict the groundwater flow and salinity distribution boundary.

GCMs projected that the climate of the research area would have much higher rainfall and temperature than the baseline period, 2011–2020, when the average annual rainfall is roughly 1,438 mm. For the MPI-ESM-MR, MIROC5, and CNRM-CM5 climate models, the average annual rainfall was projected to be significantly higher than the baseline condition by about 230, 250, and 700 mm/year, respectively, from 2021 to 2050, and the average annual rainfall increased to 1,668, 1,690, and 2,156 mm, respectively. The average annual temperatures are projected to increase by 0.6, 0.85, and 0.54 °C for the MPI-ESM-MR, CNRM-CM5, and MIROC5 under the RCP4.5 scenario, respectively, and increased by 0.9, 1.15, and 0.83 °C for the MPI-ESM-MR, CNRM-CM5, and MIROC5 under the RCP8.5 scenario, respectively in the period from 2021 to 2050.

There are several methods for recharge estimation on a global and regional scale. Lacombe et al. (2017) calculated recharge rates for a regional study in Laos and found that in the VTB, annual recharge rates range from 200 to 500 mm/year. But, Döll & Fiedler (2008) calculated the annual recharge of groundwater on a global scale and found it varied between 20 and 300 mm/year across the VTB, which is similar to the range of recharge rates that we estimated at 10 to 250 mm/year (0.75–15% of rainfall) in the zones of discharge to recharge, respectively (Table 1). The comparison results confirm that recharge rates on a local, regional, and global scale in the previous research are positively correlated with the result of the HELP3 model. Table 3 shows that the change in groundwater recharge results can be summarized in that the baseline annual groundwater recharge (272 MCM/year) can be increased by varying from 334 to 401 MCM/year or about 22.7 to 47.5%. While the present situation of groundwater use is only 20 MCM/year. Table 4 shows that changes in salinity distribution project that the area with the water with TDS between 500 and 1,500 mg/l will tend to decrease, while the freshwater (TDS <500 mg/l) area will tend to increase as shown in Figures 8 and 9. This simulated groundwater salinity distribution is consistent with previous studies about groundwater quality studies (Perttu et al. 2011b; Brindha et al. 2019). The TSD monitoring is similar to our studies such as in Thoulakhom and Hadxayfong districts.

The influence of climate change on groundwater recharge and salinity dispersion was studied using a series of numerical models, including the groundwater recharge model (HELP3), groundwater flow model (MODFLOW), and salt transport model (MT3D). The models were formulated using hydrogeological data and relevant information. The baseline period (2011–2020) groundwater condition of the VTB was modeled. The calibration and validation period were drawn from the monitoring data during 2018–2020 and 2014–2016, respectively. The calibrated model was used to project recharge rates and the spread of saline from 2021 to 2050.

Three GCMs (MPI-ESM-MR, CNRM-CM5, and MICROC5) under two RCPs (4.5 and 8.5) were selected from several GCMs and downscaled by the Change Factor Method (CFM) to project future rainfall and temperature near the future (2021–2050) periods were used for projection in the VTB. The annual groundwater recharge and storage are projected to increase from the baseline for all scenarios of future climate. The simulation of groundwater model results indicates that areas of saline groundwater will gradually increase year by year until the year 2050 in every scenario. The area with the water with the TDS between 500 and 1,500 mg/l will tend to decrease, while the freshwater (TDS < 500 mg/l) area will tend to increase. The current average annual groundwater recharge (272 MCM/year) can be increased by varying from 334 to 401 MCM/year or about 22.7–47.5%, whereas the groundwater used currently is 20 MCM/year. Although groundwater resources are abundant, the groundwater with good water quality should be used at a depth less than 40 m from the ground surface as the VTB is underlain by rock salts close to a shallow aquifer, causing groundwater salinity to expand.

Influence of climate change on recharge rates and the spread of salinity by applying a combination of HELP3, MODFLOW, and MT3D models is a substantially effective tool. The limitation of the numerical model is that more long-term monitoring data in the smaller basin are required for their accuracy. The groundwater recharge should be improved with an optimization model, which is not available for the MODFLOW and MT3D models. The integration of surface water models to simulate the effect of changing surface water flows is also recommended to increase the confidence of the predictions and to evaluate potential downstream impacts and tradeoffs associated with alternative options.

This research study was supported by a Khon Kaen University Scholarship, Khon Kaen, Thailand.

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.