Influences of groundwater extraction on flow dynamics and arsenic levels in the western Hetao Basin, Inner Mongolia, China
Data on spatiotemporal variations in groundwater levels are crucial for understanding arsenic (As) behavior and dynamics in groundwater systems. Little is known about the influences of groundwater extraction on the transport and mobilization of As in the Hetao Basin, Inner Mongolia (China), so groundwater levels were recorded in five monitoring wells from 2011 to 2016 and in 57 irrigation wells and two multilevel wells in 2016. Results showed that groundwater level in the groundwater irrigation area had two troughs each year, induced by extensive groundwater extraction, while groundwater levels in the river-diverted (Yellow River) water irrigation area had two peaks each year, resulting from surface-water irrigation. From 2011 to 2016, groundwater levels in the groundwater irrigation area presented a decreasing trend due to the overextraction. Groundwater samples were taken for geochemical analysis each year in July from 2011 to 2016. Increasing trends were observed in groundwater total dissolved solids (TDS) and As. Owing to the reverse groundwater flow direction, the Shahai Lake acts as a new groundwater recharge source. Lake water had flushed the near-surface sediments, which contain abundant soluble components, and increased groundwater salinity. In addition, groundwater extraction induced strong downward hydraulic gradients, which led to leakage recharge from shallow high-TDS groundwater to the deep semiconfined aquifer. The most plausible explanation for similar variations among As, Fe(II) and total organic carbon (TOC) concentrations is the expected dissimilatory reduction of Fe(III) oxyhydroxides.
KeywordsArsenic Groundwater extraction Spatiotemporal trends Salinization China
Influence de l’exploitation des eaux souterraines sur la dynamique des écoulements et les concentrations en arsenic dans l’ouest du bassin de Hetao, Mongolie intérieure, Chine
Les données spatio-temporelles des variations des niveaux piézométriques sont cruciales pour la compréhension du comportement de l’arsenic (As) et la dynamique des systèmes aquifères. On dispose de peu de connaissance sur l’influence de l’exploitation des eaux souterraines sur le transport et la mobilisation de l’As dans le bassin de l’Hetao, Mongolie intérieure (Chine). Ainsi les niveaux d’eau souterraines ont été enregistrés sur 5 piézomètres de 2011 à 2017 et sur 57 puits d’irrigation et deux forages multiniveaux en 2016. Les résultats montrent que les niveaux piézométriques dans la zone irriguée à partir des eaux souterraines présentent deux baisses chaque année du fait de prélèvements très importants d’eau souterraine alors que les niveaux piézométriques du secteur irrigué par dérivation de la rivière (Fleuve Jaune) présentent deux pics chaque année résultant de l’irrigation par les eaux de surface. De 2011 à 2016, les niveaux piézométriques du secteur irrigué par les eaux souterraines présentent une tendance à la baisse du fait d’une surexploitation. Des échantillons d’eau souterraine ont permis une analyse géochimique chaque année de juillet 2011 à 2016. Une augmentation des tendances d’évolution des concentrations en éléments totaux dissous et As dans les eaux souterraines a été observée. Du fait d’une inversion de la direction des écoulements des eaux souterraines, le lac Shakai devient une nouvelle source de recharge des aquifères. Les eaux du lac renferment des sédiments en surface du fonds du lac, caractérisés par une forte composante d’éléments solubles qui entraine une augmentation de la salinité des eaux souterraines. De plus, l’exploitation des eaux souterraines induit un fort gradient hydraulique à la baisse, ce qui amène un drainage des eaux souterraines superficielles de forte teneur en éléments dissous (TDS) vers les parties profondes de l’aquifère semi-captif. L’explication la plus plausible à des variations similaires des concentrations en As, Fe(II) et organique total dissous (TOC) est. la réduction dissimilatrice attendue des oxyhydroxydes de Fe(III).
Influencias de la extracción de agua subterránea en la dinámica del flujo y en los niveles de arsénico en la cuenca occidental de Hetao, Mongolia Interior, China
Los datos sobre las variaciones espaciotemporales en los niveles de agua subterránea son cruciales para comprender el comportamiento y la dinámica del arsénico (As) en los sistemas de agua subterránea. Poco se sabe sobre las influencias de la extracción de agua subterránea en el transporte y la movilización de As en la cuenca de Hetao, Mongolia Interior (China), por lo que se registraron niveles de agua subterránea en cinco pozos de monitoreo desde 2011 a 2016 y en 57 pozos de riego y dos pozos multinivel en 2016. Los resultados mostraron que el nivel freático en el área de riego de agua subterránea tenía dos depresiones cada año, inducido por la extracción de agua subterránea, mientras que los niveles de agua subterránea en el área de riego de agua derivada del río tenían dos picos cada año, como resultado del riego de agua superficial. De 2011 a 2016, los niveles de agua subterránea en el área de riego con aguas subterráneas presentaron una tendencia decreciente debido a la extracción excesiva. Las muestras de agua subterránea se tomaron para análisis geoquímicos cada año en julio desde 2011 a 2016. Se observaron tendencias crecientes en sólidos totales disueltos (TDS) y As. Debido a la dirección inversa del flujo del agua subterránea, el lago Shahai actúa como una nueva fuente de recarga de agua subterránea. El agua del lago había inundado los sedimentos cercanos a la superficie, que contienen abundantes componentes solubles y una mayor salinidad del agua subterránea. Además, la extracción de agua subterránea indujo a fuertes gradientes hidráulicos descendentes, que llevaron a la reposición de filtraciones desde aguas subterráneas poco profundas de alta TDS hasta el acuífero semi confinado profundo. La explicación más plausible para variaciones similares entre las concentraciones de As, Fe (II) y carbono orgánico total (TOC) es la reducción disimilatoria esperada de los oxihidróxidos de Fe (III).
Influências da extração de águas subterrâneas na dinâmica do fluxo e níveis de arsênio na Bacia do oeste de Hetao, Mongólia Interior, China
Dados de variações espaçotemporais nos níveis das águas subterrâneas são cruciais para a compreensão do comportamento e dinâmica do arsênio (As) em sistemas de águas subterrâneas. Pouco se sabe a respeito das influências da extração de águas subterrâneas sobre o transporte e mobilização de As na Bacia de Hetao, Mongólia Interior (China), assim, níveis de águas subterrâneas foram registrados em cinco poços de monitoramento de 2011 a 2016 e em 57 poços de irrigação e dois poços multiníveis em 2016. Os resultados mostraram que o nível das águas subterrâneas na área irrigada com águas subterrâneas teve duas recessões a cada ano, induzidas pela extração extensiva de águas subterrâneas, enquanto os níveis das águas subterrâneas na área irrigada com água desviada do rio (Rio Amarelo) tinham dois picos por ano, resultantes da irrigação com águas superficiais. De 2011 a 2016, os níveis de águas subterrâneas na área de irrigação com águas subterrâneas apresentaram tendência decrescente devido à superextração. Realizou-se amostragens de águas subterrâneas para análise geoquímica anualmente, em julho, de 2011 a 2016. Foram observadas tendências crescentes sólidos solúveis totais (SST) e As. Devido à direção inversa do fluxo de águas subterrâneas, o Lago Shahai atua como uma nova fonte de recarga de águas subterrâneas. A água do lago carreou sedimentos próximos da superfície, com abundantes componentes solúveis, e aumentou a salinidade das águas subterrâneas. Além disso, a extração de águas subterrâneas induziu fortes gradientes hidráulicos descendentes, encaminhando a recarga de vazamento de águas subterrâneas rasas de alto SST para o aquífero semiconfinado profundo. A explicação mais plausível para variações semelhantes entre as concentrações de As, Fe (II) e carbono orgânico total (COT) é a redução dissimilatória esperada de oxihidróxidos de Fe (III).
High concentrations of arsenic (As) have been found in groundwater worldwide (Ravenscroft et al. 2009), and occur under both oxic conditions and reducing conditions (Smedley and Kinniburgh 2002). Reducing aquifers that host high-As groundwater have been reported in various river deltas (including deltas of the Ganges, Mekong, and Red River in Asia, and Pearl River, Yangtze River and Yellow River in China; Polizzotto et al. 2008; Eiche et al. 2008; Norrman et al. 2008; Fendorf et al. 2010; Radloff et al. 2011; Wang et al. 2012; Guo et al. 2014a; Erban et al. 2014; Stuckey et al. 2016) and inland basins (including the Datong basin, Hetao basin, west Songnen basin, Yinchuan basin, and Zhunger basin in China, and the Pannonian basin and Danube basin in Europe; Luo et al. 2012; Guo et al. 2003, 2014a, b, c; Rowland et al. 2011; Ujević et al. 2010). It is well known that reductive dissolution of Fe(III) oxyhydroxides leads to As release from aquifer sediments under anoxic conditions (Islam et al. 2004; Mladenov et al. 2009; Fendorf et al. 2010; Guo et al. 2013a, b), which depends in turn on the quality and quantity of available organic matter, local geomorphologic characteristics and superimposing anthropogenic effects (Harvey et al. 2002, 2006; Neumann et al. 2010; Lawson et al. 2013; Neidhardt et al. 2013; Desbarats et al. 2014; Schaefer et al. 2016; Postma et al. 2016).
Many studies have been done concerning the effects of groundwater extraction on groundwater As in river deltas. In the Bengal Basin (India and Bangladesh), around 10 million irrigation and potable-water wells had been installed during the past four decades (Harvey et al. 2005). At a site in Bangladesh, the recharge from surface water, induced by intensive groundwater extraction, which can be rich in labile organic matter, was thought to cause the release of As by fueling reductive dissolution of Fe(III) oxyhydroxides (Harvey et al. 2002); however, in the Yangtze River Basin, owing to extensive groundwater extraction in the dry season, the surface water seasonally supplied oxidizing water to the anoxic aquifer, which promoted a transient drop in As concentrations (Schaefer et al. 2016). Massive extraction of deep groundwater also created local depression cones in the aquifers in the Bengal Basin, causing subsequent drawdown of As-rich shallow groundwater into deep pumping wells (Michael et al. 2009; Burgess et al. 2010). In addition, Stahl et al. (2016) showed that the Hanoi (Vietnam) aquifers adjacent to the Red River were susceptible to further As contamination where riverine recharge is drawn into aquifers by extensive groundwater pumping, with the water flowing through recently deposited river sediments before entering the aquifer. Hence, diverse temporal changes in As concentrations have been attributed to variations in local groundwater flow patterns caused by excessive water extraction in river deltas; however, few studies have been carried to investigate the response of As concentrations to variations in groundwater flow patterns in inland basins.
The Hetao basin is a typical inland basin in northwest China, where the prevalence of endemic arsenicosis was up to 25% of the population in 2002 (Jin et al. 2003). There are more than 5,000 deep wells that extract groundwater for irrigation and drinking at the mountain fronts along the margin of the basin (Guo et al. 2016a). Among these deep wells, groundwater As concentration has been found up to 390 μg/L (Jia et al. 2017). Dissimilatory and bacterial sulfate reduction (BSR)-induced reduction of Fe(III) oxides were demonstrated to be the important mechanisms for As mobilization based on Fe and S isotope signatures (Guo et al. 2013a, b, 2016b), and hydrogeological and biogeochemical investigation (Guo et al. 2011; Jia et al. 2017).
Nevertheless, little is known regarding the influences of groundwater extraction on groundwater flow dynamics and As distribution in the Hetao basin, although reverse variation patterns of shallow groundwater levels are observed in the groundwater irrigation area (GIA) and diverted Yellow River water irrigation area (YIA; Guo et al. 2013a, b). Owing to the flat topography, extremely low flow rates provoke a high vulnerability toward the effects of groundwater extraction, which can lead to severe disturbances of the naturally established hydrochemical conditions and possibly As distribution. Therefore, it is crucial to characterize the relationship between temporal variations in groundwater flow pattern and groundwater chemistry and to understand the effects of groundwater extraction on groundwater As dynamics for developing effective strategies for sustainable usage of low-As groundwater in As-affected areas.
The objectives of this study are to (1) investigate the spatiotemporal variations of groundwater levels at the mountain fronts that are subject to intensive groundwater extraction, (2) characterize temporal trends in groundwater chemistry and As concentration on an interannual time scale, and (3) evaluate the influences of groundwater extraction on groundwater As distribution.
Materials and methods
The study area
The Hetao basin, as one of the Cenozoic rift basins, is located between Yinshan uplift and Erdos platform in western Inner Mongolia. Active faults, inactive faults and insidious faults bound the northwest, the east, and the south of the basin, respectively. The Langshan Mountains, to the north of the basin, are mainly composed of Mesoproterozoic deeply metamorphic rocks and intrusive rocks and Jurassic to Cretaceous metamorphic rocks. The basin is a SW–NE tilting flat plain, with increasing sediment thickness from 500 to 1,500 m in the east to 4,000–8,000 m in the west. The basin, located in an arid-semiarid climate zone, has an average annual precipitation of 130–220 mm (mainly during July to September) and annual evaporation rate of about 2,000–2,500 mm (Guo et al. 2008a).
The groundwater level generally rises from around 20 m below land surface (bls) in the alluvial fans to around 2.0 m bls in the flat plain for both shallow groundwater and deep groundwater. Groundwater is mainly recharged by fracture water along the mountain front in the alluvial fans, and by vertically infiltrating precipitation, ditch water (irrigation channels), and irrigation water in the flat plain, and discharged mainly via evapotranspiration, drainage, and extraction. As a natural boundary between the two geomorphic units (the piedmont alluvial plain in the north and the alluvial lacustrine plain in the south), the drainage channel is regarded as the discharge route of shallow groundwater, where groundwater flows from the piedmont area in the north and from the flat plain in the south (Zhang et al. 2013). The general direction of groundwater flow is from the alluvial fans, through the transition area, to the flat plain, but the flow rate has been recorded as relatively slow, with the range between 0.002 and 0.2 m/day due to the gentle topography and low permeability of the aquifer sediments (Inner Mongolia Institute of Hydrogeology 1982).
For more than 50 years, the diverted water from the Yellow River has been predominately used for agricultural irrigation in the basin (Guo et al. 2011). Since the diverted Yellow River water is unavailable near the mountains due to the high elevation and overuse of the water in the upstream of the irrigation channels, groundwater has recently been used for both irrigation and drinking to the north of the drainage channel. Reverse variation patterns of groundwater levels are observed in GIA and YIA (Guo et al. 2013a, b). During irrigation seasons, high groundwater levels occur in YIA, while low levels occur in GIA; therefore, irrigation activities, including the overextraction of groundwater and the diverted Yellow River water irrigation, may change groundwater flow patterns and potentially affect the chemistry of groundwater.
According to monitoring data of groundwater levels/heads in GIA in this study, the GIA may be divided into two zones: the northwestern groundwater recharge area (zone I) near the piedmont, and the southeastern surface water recharge area (zone II) close to the Shahai Lake. During irrigation seasons, groundwater flows from northwest to southeast in zone I, while it flows from southeast to northwest in zone II.
Monitoring of groundwater levels
Five monitoring wells were installed in Shahai town; two of them were located in GIA (Nos. 1-3 and 5-2) and three in YIA (Nos. 2-4, 3-4, and 4-5; Fig. 1b). Groundwater levels were monitored every 30 min from 2011 to 2016 using water level data loggers (HOBO U20, Onset) and local atmospheric pressure was also monitored in the same way. Water levels of the Shahai Lake were monitored every 30 min from May 2016 to May 2017. Groundwater heads of 45 irrigation wells (Fig. 1b) were measured using an electronic water sensor (Model 101B, Solinst) in July 2016 (irrigation season), while water heads of 57 irrigation wells were measured in October 2016 and March 2017 which are in the nonirrigation seasons; additionally, two multilevel wells (Nos. K1 and K2) were installed in zone II and zone I, respectively. Each multilevel well had seven piezometers at seven different depths. Piezometers No. K2-3 (41 m) and No. K2-6 (65 m) in zone I, and No. K1-3 (38 m) and No. K1-6 (74 m) in zone II, were selected to regularly measure groundwater levels/heads of shallow groundwater and deep groundwater, in April 2016, May 2016, August 2016, November 2016 and March 2017. The real-time kinematic difference global positioning system (RTK-GPS) was used to measure the elevations of all well heads and to calibrate water level/head depth measurements to water level/head elevations above mean sea level (amsl).
Sample collection and analysis
Groundwater samples were collected in 2011 (n = 80), 2012 (n = 16), 2013 (n = 13), 2014 (n = 25), 2015 (n = 29), and 2016 (n = 18) from irrigation wells (in GIA), and yearly from two monitoring wells (in YIA) from 2012 to 2015 (Fig. 1b). In particular, from the piedmont area to the Shahai Lake, groundwater samples of 10 typical irrigation wells (five wells: Nos. I22, I21, I20, I5 and I9 in zone I; five wells: Nos. I10, I3, I1, I14 and I15 in zone II, in GIA) were collected yearly from 2011 to 2016. The depths of these irrigation wells mostly ranged between 60 and 110 m. The monitoring wells have depths of around 20 m bls. In addition, spring water, Shahai Lake water and diverted Yellow River water were collected in July 2016 (Fig. 1b).
Before groundwater sampling, wells were pumped at least 20 min until water temperature, EC, pH, and Eh kept stable. All samples were filtered through 0.22-μm membrane filters in the field. Water samples for major cation and trace element analysis were acidified with ultrapure 6 M HNO3 to pH <2.0. Those for analysis of As species were preserved with 0.25 M EDTA in amber bottles. Samples for anion analysis were filtered but unacidified. All samples were kept and transported to the laboratory at 4 °C, and kept in a refrigerator at 4 °C.
At the time of sampling, parameters including water temperature, EC, pH, and Eh were measured using a multiparameter portable meter (HI9828, HANNA), while Fe(II) concentration was determined by using a portable spectrophotometer (DR2800, HACH) and alkalinity using a Model 16,900 digital titrator (HACH) using bromocresol green-methyl red indicator. Eh readings were normalized with reference to a hydrogen electrode.
Concentrations of major cations and trace elements were determined by ICP-AES and ICP-MS, respectively. The analytical precision of ICP-AES and ICP-MS was 3.0%, whereas the detection limit for As was 0.01 μg/L. Unacidified aliquots were analyzed for major anions by ion chromatography (DX-120, Dionex), with the analytical precision less than 5.0%; however, for most samples, ion charge imbalances were less than 5%. Arsenic species in groundwater samples were analyzed by HPLC-HG-AFS and detection limits of As(III) and As(V) were 2 and 4 μg/L, respectively.
Variations in groundwater levels/heads
Although water levels/heads varied, higher water levels/heads were generally observed in zone II than those in zone I. Groundwater levels/heads of well No. K1 in zone II were higher than those of well No. K2 in zone I of both shallow groundwater and deep groundwater during April 2016 and March 2017 (Fig. 3). It indicated that groundwater would flow from zone II to zone I, which is opposite to the groundwater flow direction reported before by Zhang et al. (2013).
In GIA, hydraulic connection between the shallow aquifer and deep aquifer may exist. Groundwater levels/heads were similar in both the shallow aquifer and deep aquifer in nonirrigation seasons (Fig. 3); however, groundwater heads of the deep groundwater (Nos. K1-6 and K2-6) declined faster than those of the shallow groundwater (Nos. K1-3 and K2-3; Fig. 3) in irrigation seasons, indicating that the hydraulic connection would be weak. Accordingly, the lower water heads in the deep aquifer increased hydraulic gradients between the shallow groundwater and deep groundwater, and therefore recharged from the shallow groundwater into the deep groundwater in irrigation seasons.
In YIA, the shallow groundwater levels increased during irrigation seasons. For well Nos. 2-4, 3-4, and 4-5, water level elevations had two peaks in each year. One peak was observed in May–August and the other in November, both of which were caused by the extensive irrigation using the diverted Yellow River water. A slight decline was observed from August to October because of the decrease in the recharge of irrigation water and evaporation. During winter, the shallow groundwater levels decreased by 2 m, possibly due to strong evaporation and freezing. There was no obvious interannual change of water levels in those three wells from 2011 to 2016 (Fig. 2b). A similar trend in water level was observed in the lake water and well No. 2-4, both of which had two water level peaks within 1 year. Generally, the lake-water levels were higher than those of well Nos. 1-3 and 5-2 and lower than those of well Nos. 3-4 and 4-5 (Fig. 2). The groundwater levels of well No. 4-5 were generally higher than the other two wells (Nos. 2-4 and 3-4), indicating that groundwater flow direction was from the flat plain region to the drainage channel, which is consistent with the observation of Zhang et al. (2013) in YIA.
Interannual variations in groundwater chemistry
Groundwater in YIA was mainly of Na-Cl-HCO3 type, which witnessed a slight decrease in TDS within 6 years (Fig. 5). There were statistically long-term declining trends of TDS for well Nos. 2-4 and 3-4 from 2012 to 2015 (r = 0.69 and 0.96, respectively). Especially, the decreasing trend in TDS of well No. 3-4 (12%) was more evident than that of well No. 2-4 (2.1%). The decrease in groundwater TDS was undoubtedly accompanied by decreases in some major ions, mainly Cl−, SO42− and Na+ (12, 0.8 and 2.7% for well No. 2-4; 19.6, 37.3 and 9.7% for well No. 2-4, respectively; Fig. 7; Table S1 of the ESM). In addition, the spring water was of Ca-HCO3-SO4 type, while both the lake water and the diverted Yellow River water were of Na-Cl-SO4 type. Accordingly, the former had a lower TDS (438 mg/L) than both of the latter (1,050 and 657 mg/L, respectively).
Interannual variations in groundwater as and redox-sensitive components
Moderate groundwater As concentrations were observed in YIA. Two monitoring wells (Nos. 2-4 and 3-4) had As concentrations between 48.2 and 106 μg/L, whereby well No. 3-4, which had As concentrations lower than 50 μg/L between 2012 and 2014, especially did not meet the standard in 2015, while well No. 2-4 remained relatively stable during 2012–2015. The temporal variations in As(III) were comparable to those of total As, whereby arsenic concentrations of spring water, lake water and the diverted Yellow River water were lower than 10 μg/L.
From the piedmont to the flat plain, groundwater Eh presented decreasing trends. Concentrations of redox-sensitive components Fe(II) in groundwater displayed rising trends. In GIA, the groundwater Fe(II) concentrations of five irrigation wells in zone I ranged from <0.01 to 0.66 mg/L and from 0.02 to 0.55 mg/L, which were lower than those of another five irrigation wells in zone II, varying from 0.59 to 2.48 mg/L and 0.90 to 6.40 mg/L, respectively (Table S1 of the ESM).
From 2011 to 2016, the fluctuation (% RSD, relative standard deviation) of Fe(II) concentration of the five irrigation wells in zone I remained <20.6%, excluding the wells with Fe(II) concentrations <0.05 mg/L (Table S1 of the ESM). In addition, slight rising trends in Fe(II) concentrations were observed in well Nos. I5 and I9 of zone I, increasing by 0.04 and 0.03 mg/L/year respectively. In zone II, five irrigation wells showed obvious long-term increases in Fe(II) concentrations between 0.03 and 0.19 mg/L/year (r > 0.7; Table S1 of the ESM); however, in YIA, no temporal change was observed concerning Fe(II) concentrations in well Nos. 2-4 and 3-4.
Influences of extraction on groundwater flow fields
Irrigation wells have been gradually installed in GIA owing to the improvement of electric power facilities and difficulties in diverting the Yellow River water to the piedmont since 2006. A total of 127 irrigation wells, each with a flow rate of 0.027 m3/s have been used for irrigation in the studied area until 2016. Around 15.5 million m3 of groundwater has been extracted for irrigation in the spring of each year. Although the groundwater levels/heads fluctuated due to the irrigation activities, the Shahai Lake generally kept higher water levels than the northwestern groundwater levels/heads, which acted as a recharge area throughout the year (Fig. 11b). Therefore, groundwater flowed from the Shahai Lake to the northwest in zone II, which is opposite to the observation by Zhang et al. (2013), although it flowed from the alluvial fans to the southeast in zone I. During the spring irrigation, pumping caused sharp drops in the water levels of well Nos. 1-3 and 5-2 (Fig. 2b) and also induced a depression cone in GIA (Fig. 4). As the pumping continued, the water level difference between the center of the depression cone and the Shahai Lake gradually increased, which resulted in the increase of hydraulic gradient and the flow rate of groundwater recharge from the Shahai Lake (around twice as fast as normal). Accordingly, the lake water level presented a decreasing trend. When the spring irrigation ended, the groundwater levels/heads rose again and the depression cone subsequently shrunk (Figs. 2b and 4b). In the period of winter irrigation, relatively less groundwater (2.1 million m3) was needed for irrigation, which caused a slight drop in groundwater levels. Before the drop, a slight peak was also observed due to the influence of the diverted Yellow River water irrigation which occurred earlier than the groundwater irrigation. During winter irrigation, the diverted Yellow River water irrigation affected groundwater levels near well No. 5-2. As soon as the winter irrigation ceased, the water levels recovered. Before the next irrigation season, the area of the depression cone decreased to the minimum. The lake-water recharge continuously introduced chemicals into the groundwater between the depression cone and the Shahai Lake in zone II, which may affect groundwater As variations.
For the whole year, well No. 5-2 had higher water levels than well No. 1-3, showing that the flow direction was from the west to the east of the depression cone. Additionally, from 2012 to 2016, significant downward trends of water levels were observed in well Nos. 1-3 and 5-2, with the decrease rates of 0.25 and 0.2 m/year, respectively (Fig. 2b). It indicated that groundwater extraction led to the continuous decrease in groundwater levels/heads, and the previous groundwater flow conditions had been broken in the study area (Fig. 11).
Nevertheless, in YIA, the diverted Yellow River water irrigation was still the predominant irrigation method. Groundwater levels had two peaks each year, which corresponded to the spring irrigation season and the winter irrigation season, indicating that the diverted Yellow River water irrigation was the main recharge source for groundwater. No obvious interannual variations in groundwater levels were observed, suggesting that the recharge and discharge of groundwater maintain a long-term stable state. Therefore, the groundwater flow conditions remained unchanged in YIA, which is consistent with the observation by Zhang et al. (2013).
Spatiotemporal variations of groundwater chemistry
Although the groundwater flow field in zone I was disturbed by the depression cone due to the intensive extraction (Fig. 4), the groundwater recharge source did not change. Owing to the deep water table, ranging between 10 and 20 m bls (Fig. 4), groundwater evaporation was relatively weak; consequently, groundwater TDS or major ions concentrations kept relatively stable from 2011 to 2016 (Fig. 8).
However, in zone II, groundwater TDS increased significantly over time, which was accompanied by increases in the major ions. This phenomenon was governed by the change of groundwater flow conditions. The shift of the groundwater recharge source from the low-TDS piedmont groundwater to the high-TDS lake water, which was induced by the depression cone, increased groundwater salinity in zone II (Figs. 6 and 7). Since zone II was the local discharge area of shallow groundwater before 2006, near-surface sediments in this zone contained a lot of soluble components (Yuan et al. 2017). The lake-water recharge flushed those sediments and carried a large amount of soluble components into the groundwater. Moreover, intensive extraction of deep groundwater increased the hydraulic gradient between the shallow groundwater and deep groundwater (Fig. 3b) and might induce recharge of shallow highly saline groundwater into the deep low-salinity groundwater. Hence, the TDS of the deep semiconfined groundwater tended to gradually increase over time with the intensive extraction. Although shallow groundwater levels in YIA would lead to intensive evaporation (Fig. 2a), the dilution by the diverted Yellow River water (with low TDS 657 mg/L) counteracted the evaporation, resulting in the relatively constant TDS (Fig. 7).
Influences of groundwater flow changes on groundwater As
High-As groundwater (> 50 μg/L; Figs. 8 and 9) was observed in zone II and the YIA with high concentrations of Fe(II) and TOC and lower Eh (Fig. 10). Obvious increases in groundwater As concentrations were observed in zone II from 2011 to 2016 (Fig. 10). Temporal variations in As concentrations were well correlated to temporal variations in Fe(II) (r > 0.7) for five irrigation wells (Fig. 10). Increases in groundwater TOC were also obvious in zone II (Table S1 of the ESM). The increases in groundwater As may be associated with the increases in groundwater TOC and may indicate that reductive dissolution of Fe(III) oxides played a dominant role in As mobilization. This process was usually fueled by natural organic matter (Islam et al. 2004; McArthur et al. 2004); however, in addition, higher As groundwater generally had lower concentrations of SO42−/Cl− in the study area (Fig. 10). The bacterial sulfate reduction (BSR) was verified by S and O isotopes of groundwater SO42− and chemical analysis in the study area, and by the abundant distribution of sulfate-reducing bacteria (Guo et al. 2008b; Li et al. 2014a, b; Guo et al. 2016b), which would induce reduction of As-bearing Fe(III)-oxyhydroxides and lead to the release of As into groundwater.
The shift of the groundwater recharge source from the piedmont groundwater to the lake water would be related to the increase in groundwater As concentrations in zone II. The Shahai Lake water had higher TOC contents than the piedmont groundwater. The shift of the groundwater recharge source introduced more dissolved organic matter (DOM) into the aquifers, and the recharge source DOM was believed to be more labile in comparison with the resident groundwater DOM. The labile DOM readily fueled dissimilatory reduction of Fe(III) oxides and BSR-induced reduction of Fe(III) oxyhydroxides. In addition, pumping-induced fluctuations of groundwater levels/heads could physically disturb the sediment matrix, which would destabilize the mineral-associated sedimentary organic matter (Eusterhues et al. 2003). The destabilized sedimentary organic matter is believed to be bioavailable for microbes (Neumann et al. 2014) and may also contribute to the increase in groundwater As concentrations in zone II; however, in YIA, no obvious changes of groundwater As were observed, which is consistent with the previous observation (Guo et al. 2013a, b). No significant variations in As concentrations are attributed to the relatively steady groundwater flow direction and recharge source (Fig. 10).
It is noteworthy that groundwater in zone I has the potential to be contaminated by As, although relatively low As concentrations are currently observed. With the intensive groundwater extraction, the depression cone will be moved northwestwards, and high-As groundwater will flow northwestward. In this case, groundwater As in zone I is likely to increase (Fig. 9); therefore, the extent of the groundwater depression cone must be restrained or decreased to protect the low-As groundwater in zone I via regulation of groundwater extraction in zones I and II.
Groundwater levels temporally fluctuated during the 6 years of observation. The fluctuation patterns of groundwater level were different between the GIA and YIA, as they were affected by groundwater extraction, the diverted Yellow River water irrigation and evaporation. In GIA, groundwater was extracted for irrigation during April–August and in November, resulting in two troughs in groundwater levels each year. From 2011 to 2016, groundwater level presented a decreasing trend due to the overextraction. However, in YIA, surface water was utilized for spring irrigation and winter irrigation, resulting in two peaks in groundwater levels each year. In GIA, the intensive groundwater extraction caused a depression cone, which reversed the groundwater flow direction and changed the recharge source in zone II. From 2011 to 2016, groundwater TDS of zone II showed an obvious increase, which could be explained by the fact that the recharge from the lake water flushed near-surface sediments, which contained soluble components, and finally carried shallow high-TDS groundwater into the deep aquifer. Furthermore, a rising trend of groundwater As was also observed, which was accompanied by the increases in groundwater Fe(II) and TOC and the decrease in SO42−/Cl−, implying that the biologically degradable organic carbon introduced by the lake water promoted dissimilatory reduction of Fe(III) oxides and BSR-induced reduction of Fe(III) oxyhydroxides; besides, the low-As groundwater in zone I would be potentially affected by the high-As groundwater from zone II due to the reverse groundwater flow direction. Slight decreases in groundwater TDS occurred in YIA due to the dilution effect of surface water, and temporal variation in groundwater As was slight. The groundwater extraction, altering recharge sources of groundwater and consequently providing labile organic matter for biogeochemical processes of As release, should be restrained in terms of effective protection of groundwater quality.
The study was financially supported by the National Natural Science Foundation of China (grant Nos. 41672225 and 41222020), the program of China Geology Survey (grant No. 12120113103700), the Fundamental Research Funds for the Central Universities (grant No. 2652013028), and the Fok Ying-Tung Education Foundation, China (grant No. 131017).
- Eiche E, Neumann T, Berg M, Weinman B, van Geen A, Norra S, Berner Z, Kim Trang PT, Viet PH, Stüben D (2008) Geochemical processes underlying a sharp contrast in groundwater arsenic concentrations in a village on the Red River Delta, Vietnam. Appl Geochem 23:3143–3154. https://doi.org/10.1016/j.apgeochem.2008.06.023 CrossRefGoogle Scholar
- Guo HM, Zhang D, Wen DG, Wu Y, Ni P, Jiang YX, Guo Q, Li FL, Zheng H, Zhou YZ (2014b) Arsenic mobilization in aquifers of the southwest Songnen basin, PR China: evidences from chemical and isotopic characteristics. Sci Total Environ 490:590–602. https://doi.org/10.1016/j.scitotenv.2014.05.050 CrossRefGoogle Scholar
- Guo Q, Guo HM, Yang YC, Han SB, Zhang FC (2014c) Hydrogeochemical contrasts between low and high arsenic groundwater and its implications for arsenic mobilization in shallow aquifers of the North Yinchuan basin, PR China. J Hydrol 518:464–476. https://doi.org/10.1016/j.jhydrol.2014.06.026 CrossRefGoogle Scholar
- Guo HM, Jia YF, Wanty R, Jiang YX, Zhao WG, Xiu W (2016a) Contrasting distributions of groundwater arsenic and uranium in the Hetao basin, Inner Mongolia: implication for origins and fate controls. Sci Total Environ 541:1172–1190. https://doi.org/10.1016/j.scitotenv.2015.10.018 CrossRefGoogle Scholar
- Harvey CF, Swartz CH, Badruzzaman ABM, Keon-Blute N, Yu W, Ali MA (2005) Groundwater arsenic contamination on the Ganges Delta: biogeochemistry, hydrology, human perturbations, and human suffering on a large scale. C R Geosci 337:285–296. https://doi.org/10.1016/j.crte.2004.10.015 CrossRefGoogle Scholar
- Inner Mongolia Institute of Hydrogeology (1982) Hydrogeological setting and remediation: approaches of soil salinity in the Hetao Basin, Inner Mongolia (in Chinese). Scientific report, Inner Mongolia Institute of Hydrogeology, Hohhot City, Inner MongoliaGoogle Scholar
- Jin YL, Liang CH, He GL, Cao JX, Ma F, Wang HZ, Ying B, Ji RD (2003) Study on distribution of endemic arsenism in China (In Chinese with English abstract). J Hygiene Res 23(6):519–540Google Scholar
- Li B, Shi B (2011) Hydrochemical characteristics of groundwater before and after water-saving reform in Hetao Irrigation District, Inner Mongolia (in Chinese with English abstract). Trans Chin Soc Agric Eng 30(21):99–110Google Scholar
- Li SY, Guo HM, Huang SF, Ma JW, Liu HY, Sun YY (2016) Variation in Hetao Basin wetlands from 1973 to 2014 (in Chinese with English abstract). Resour Sci 38(1):0019–0029Google Scholar
- McArthur JM, Banerjee DM, Hudson-Edwards KA, Mishra R, Purohit R, Ravenscroft P, Lowry D (2004) Natural organic matter in sedimentary basins and its relation to arsenic in anoxic ground water: the example of West Bengal and its worldwide implications. Appl Geochem 19(8):1255–1293. https://doi.org/10.1016/j.apgeochem.2004.02.001 CrossRefGoogle Scholar
- Ravenscroft P, Brammer H, Richards KS (2009) Arsenic pollution: a global synthesis. Wiley-Blackwell, Singapore, 114 pp Google Scholar
- Yuan R, Guo H, Zhang D, Li Y, Zhang Y, Cao W (2017) Soluble components of sediments and their relation with dissolved arsenic in aquifers from the Hetao Basin, Inner Mongolia. J Soil Sediment 17(12):2899–2911. https://doi.org/10.1007/s11368-017-1770-9