Identifying the recharge sources and age of groundwater in the Songnen Plain (Northeast China) using environmental isotopes

Abstract

The recharge sources and groundwater age in the Songnen Plain, Northeast China, were confirmed using environmental isotopes. The isotopic signatures of the unconfined aquifers in the southeast elevated plain and the north and west piedmont, cluster along local meteoric water lines (LMWLs) with a slope of about 5. The signature of source water was obtained by the intersection of these LMWLs with the regional meteoric water line (RMWL). This finding provides evidence that the recharge water for these areas originate from the Changbai Mountains and the Low and High Hingan Mountains, respectively. Groundwater in the unconfined aquifer in the low plain yields a LMWL with a slope of 4.4; its nitrate concentration indicates the admixture of irrigation return flow. The δ-values of the unconfined aquifer in the east elevated plain plot along the RMWL, reflecting recharge by local precipitation. The mean residence time of groundwater in these aquifers is less than 50 years. However, the ¹⁴C age of the groundwater in the confined Quaternary aquifer ranges from modern to 19,500 years, and in the Tertiary confined aquifer from 3,100 to 24,900 years. Modern groundwater is mainly recharged to the Quaternary confined aquifer on the piedmont by local precipitation and lateral subsurface flow.

Résumé

Les aires de recharge et l’âge des eaux souterraines de la plaine de Songnen, Chine du Nord-Est, ont été confirmées au moyen d’isotopes environnementaux. Les signatures isotopiques des aquifères libres dans la plaine soulevée Sud-Est et dans les piedmonts Nord et Ouest, sont groupées le long de droites météorique locales (LMWLs) de pente environ 5. La signature de l’eau d’origine a été obtenue par l’intersection de ces droites météoriques locales avec la droite météorique régionale (RMWL). Ce résultat montre que la recharge de ces zones provient respectivement des montagnes Changbai et des basses et hautes montagnes Hingan. L’eau souterraine de l’aquifère libre des basses plaines présente une LMWL de pente 4.4; sa concentration en nitrates indique la contribution d’eau d’irrigation infiltrée. Les valeurs de δ pour l’aquifère libre de la plaine Est soulevée s’alignent le long de la RMWL, ce qui reflète une recharge par les précipitations locales. Le temps moyen de résidence de l’eau souterraine dans ces aquifères est de moins de 50 ans. Cependant, l’âge ¹⁴C de l’eau souterraine de l’aquifère quaternaire captif est compris entre des âges récents et 19 500 ans, celui l’aquifère tertiaire captif entre 3 100 et 24 900 ans. L’eau souterraine récente recharge principalement l’aquifère quaternaire captif sur les piedmonts du fait des précipitations locales et des écoulements latéraux de subsurface.

Resumen

Se verificaron las fuentes de recarga y edad de agua subterránea en Songnen Plain, Noreste de China, usando isótopos ambientales. Las señales isotópicas del acuífero no confinado en la planicie elevada del sudeste y en el piedemonte norte y oeste, se agrupan a lo largo de la línea de agua meteórica (LMWLs) con una pendiente de alrededor de 5. La señal de la fuente de agua fue obtenida por la intersección de estas LMWLs con la línea de agua meteórica regional (RMWL). Este hallazgo provee evidencias que la recarga de agua para estas areas se origina a partir de las montañas Changbai y las montañas de la Baja y Alta Hingan, respectivamente. El agua subterránea en el acuífero no confinado en la planicie baja exhibe un LMWL con una pendiente de 4.4; su concentración de nitrato indica la mezcla del flujo de retorno de la irrigación. Los valores de δ- en el acuífero no confinado en la planicie elevada este graficados a lo largo del RMWL, refleja la recarga por precipitación local. El tiempo medio de residencia del agua subterránea en estos acuíferos es menor que 50 años. Sin embargo, la edad ¹⁴C del agua subterránea en acuífero confinado cuaternario varía desde lo moderno hasta 19,500 años, y en el acuífero confinado Terciario de 3,100 a 24,900 años. El agua subterránea actual es principalmente recargada al acuífero confinado Cuaternario en el piedemonte por precipitación local y por flujo subsuperficial lateral.

摘要

本文应用环境同位素对松嫩平原地下水的补给来源与年龄进行了研究. 东南部高平原和西, 北部山前平原潜水含水层地下水样品呈现出沿着斜率约为5的当地降水线 (LMWLs) 分布的稳定同位素特征, 该当地降水线与区域降水线 (RMWL) 交点的稳定同位素特征表明上述区域的补给水分别来自长白山区和大小兴安岭山区. 中部低平原潜水含水层地下水样品沿着斜率约为4.4 的当地降水线分布, 其硝酸盐含量表明相当部分的补给来自灌溉回归. 东部高平原潜水含水层地下水样品沿着区域降水线分布, 说明补给来自当地降水. 这些潜水含水层地下水滞留时间小于50年, 而第四系承压含水层地下水14C年龄范围为从现代至距今19,500年, 第三系含水层地下水年龄为距今3,100年至24,900年, 现代补给水通过降水入渗和地下侧向径流在山前平原补给第四系承压含水层.

Resumo

As fontes de recarga e a idade das águas subterrâneas na planície de Songnen, nordeste da China, foram confirmadas utilizando isótopos ambientais. As assinaturas isotópicas dos aquíferos livres da planície elevada a sudeste e do norte e oeste do Piemonte agrupam-se ao longo das linhas de águas meteóricas locais (LMWLs) com uma inclinação de cerca de 5. A assinatura da origem da água foi obtida pela intersecção daquelas linhas de águas meteóricas locais (LMWLs) com a linha de água meteórica regional (RMWL). Esta descoberta prova que as águas de recarga para estas áreas têm origem nas montanhas Changbai e nas baixas e altas montanhas Hingan, respectivamente. As águas subterrâneas do aquífero livre da planície baixa fornecem uma linha de água meteórica local (LMWL) com uma inclinação de 4.4; o seu teor em nitratos indica mistura com escoamentos de retorno da irrigação. Os valores de δ do aquífero livre da planície elevada a leste dispõem-se ao longo da RMWL, reflectindo a recarga pela precipitação local. O tempo médio de residência das águas subterrâneas nestes aquíferos é inferior a 50 anos. No entanto, a idade de ¹⁴C das águas subterrâneas do aquífero confinado do Quaternário varia do moderno até 19,500 anos, e no aquífero confinado do Terciário a idade varia entre os 3,100 e os 24,900 anos. As águas subterrâneas recentes são principalmente recarregadas para o aquífero confinado do Quaternário no Piemonte, por precipitação local e fluxo lateral sub-superficial.

Introduction

The Songnen Plain is one of the largest agricultural production areas in China. Groundwater provides about 40% of the water supply, for irrigation and industrial and municipal use. In the past two decades, increased pumping rate has resulted in groundwater level decline and reduced groundwater discharge to streams. To understand the effects of a further increase in groundwater abstraction, and the long-term potential of the aquifer system, a good approach is to identify the sources of recharge and to estimate the age of the pumping groundwater. Although the hydrogeological setting has been well investigated in previous studies (Chen and Ma 2002; Zhao et al. 2009), the source of water to aquifers has been an unanswered question because multiple sources of recharge to the aquifer system are present across the plain (Zhao et al. 2009). The isotopic characteristics of groundwater samples are likely to broadly reflect the characteristics of the source.

Environmental isotopes (¹⁸O, ²H, ³H and ¹⁴C) of groundwater have been widely and extensively used over the past 60 years in groundwater investigations. The stable isotope compositions of oxygen and hydrogen, expressed as δ¹⁸O and δ²H values, respectively, are commonly used in regional groundwater studies to identify flow regimes and sources of recharge (Fontes 1980; Gonfiantini 1986; Clark and Fritz 1997). Radioactive isotopes (³H and ¹⁴C) can be used to estimate groundwater ages. The presence of detectable ³H is evidence for groundwater recharge after the period of nuclear bomb tests (1952–1964; Clark and Fritz 1997). ¹⁴C was used as a dating method to estimate the age of water from several hundred to about 35,000 years (Tamers 1975; Ingerson and Pearson 1964; Mook 1976; Fontes and Garnier 1979; Clark and Fritz 1997; Geyh 2000). A few isotope hydrological studies were done on the groundwater in the Taoer River catchment, which is a small part of the Songnen Plain (Wang 1995; Jia et al. 2004; Zhang et al. 2007). Jia et al. (2004) presented a study of the isotopic data available for the recharge sources in the aquifer of the Taoer River alluvium fan. Zhang et al. (2007) successfully employed chemical and isotopic data to understand the change of the groundwater quality in the Taoer River catchment.

In this study, stable isotopes (δ¹⁸O and δ²H) were for the first time used to investigate the regional groundwater flow system in the Songnen Plain. Radiogenic isotopes (tritium, ³H; and carbon-14, ¹⁴C) were also used to estimate groundwater ages. Using well established isotope hydrogeological models, groundwater residence times were estimated and recharge sources were identified in the Songnen Plain.

Study area

The Songnen Plain (121°21′–128°18′ E and 43°36′–49°26′ N), lies between the High Hinggan Mountains in the west of the study area and the Low Hinggan and the Changbai Mountains in the east (Fig. 1). It is mainly covered by alluvial deposits (Fig. 2) of the Songhuajiang and Nenjiang rivers. The plain extends over ~178,000 km² and is populated by about 35 million people.

Fig. 1

Study area (the Songnen Plain) and sampling sites. Areas I–V are divisions of the unconfined aquifer system (see text)

Fig. 2

Hydrogeologic cross sections along the transect A–A' in Fig. 1

Geographically, the plain consists of the piedmont area in the west part, the central low plain, and the elevated plain in the east part (Fig. 1). Most of the Songnen Plain is less than 200 meters above sea level (m asl). The climate is continental to semi-arid. As a result of large topographic elevation differences and the surrounding mountains, the precipitation varies widely. The mean annual precipitation ranges from 500–600 mm in the east elevated plain to 300–450 mm in the west piedmont area. As affected by the East Asia monsoon, about 70% of annual rainfall occurs between the middle of June and mid-August. The mean annual temperature ranges from –4 to +10°C and gradually decreases from south to north. Accordingly, the mean potential evaporation rate increases from 1,200–1,600 mm/a in the northeast to 1,500–1,900 mm/a in the southwest. The entire plain is drained by the Songhuajiang River which emerges from two main sources, the second Songhuajiang River and the Nenjiang River. The second Songhuajiang River originates from the Heavenly Lake in the Changbai Mountains with an elevation of 2,744 m asl. The headwater area of the Nenjiang River is the High Hinggan Mountains, ranging from 700 to 2,000 m asl, with a mean elevation of 1,200 m asl. Both rivers converge and become the Songhuajiang River which discharges into the Sanjiang Plain (Fig. 1). Smaller rivers such as Taoer, Huolin, Wuyuer, Shuangyang, Hulan, Tongken, Lalin, Yalu, Chaoer and Nuomin, are entirely supplied by upstream precipitation and are dry during the dry season.

The Quaternary aquifers (Q₁, Q_2–3 and Q₃) are the main target for the water supply (Fig. 2). The fractured Cretaceous aquifer (K) has a thickness of less than 50 m and has a low permeability. It is only exploited in the water scarce area in the east high plain for domestic use. The confined Tertiary aquifer (N) is mainly exploited in the central plain. The local unconfined shallow aquifer system is divided into five systems: Nemoer-Keluo (I), Nuomin-Taoer (II), Wuyuer-Shuangyang (III), Tongken-Hulan (IV), the second Songhuajiang (V) (Fig. 1).

The Nemoer-Keluo system (I) is located at the piedmont of the Low Hinggan Mountains with an area of 14,366 km². The Nenjiang River is its west boundary and the Low Hinggan Mountains is the north and east boundary. A local groundwater divide along the Wuyuer River is the south boundary which separates this system from the adjacent Wuyuer-Shuangyang (III). The unconfined aquifer consists of moraine and glacio-fluvial deposits with thickness of ~60 m. Groundwater recharge in this system involves infiltration of precipitation, and seepage of Nemoer and Keluo rivers. Groundwater flows from the north and northeast towards the south and southwest and is discharged finally into the Nenjiang River.

The Nuomin-Taoer System (II) is located in the west piedmont along the Nenjiang River with an area of 50,457 km². The High Hinggan Mountains is its west boundary and the Nenjiang River is its east boundary. A local groundwater divide along the second Songhuajiang River separates it from the second Songhuajiang System (V). The west piedmont plain consists of a series of alluvial fans. The aquifer of these fans is unconfined with thickness ranging from 50 to 100 m, and consisting of gravel and coarse-grained sand. Depth to the water table for this aquifer is 10–20 m. However, the unconfined aquifers in the central low plain, the catchments of Taoer and Huolin River, consists of fine sand and silt with thickness ranging from 20 to 60 m. Depth to the water table in this area is 5–8 m. Groundwater is mainly recharged by infiltration of precipitation, seepage of rivers and lateral flow from mountains. Groundwater flows towards the east and discharges into the Nenjiang River.

The Wuyuer-Shuangyang System (III) is located in the central plain with an area of 47,430 km². It extends across the elevated plain and low plain. The local groundwater divide is along the Wuyuer River and Tongken River, which are its north boundary and east boundary. The Low Hinggan Mountains is its northeast boundary. The Nenjiang River and Songhuajiang River are its west and south boundary, respectively. The porous unconfined aquifer consists of fine sand and silt with thickness of ~50 m. Depth to the water table is less than 5 m. Groundwater is recharged by infiltration of local precipitation and subsurface lateral flow from the Low Hinggan Mountains. Groundwater flows form northeast towards southwest and is discharged finally into the Nenjiang River and Songhuajiang River.

The Tongken-Hulan System (IV) is located in the elevated plain with an area of 34,714 km². The local groundwater divide along the Tongken River is its west and north boundary. The Low Hinggan Mountains is its east boundary. The Songhuajiang River is its south discharge boundary. The unconfined aquifer consists of loessal silt loam and silt clay with thickness of 10~30 m. Depth to the water table is less than 5 m. Groundwater is recharged by infiltration of local precipitation and rivers. Water flows towards the central low plain and discharged into the Songhuajiang River.

The second Songhuajiang System (V) is located in the southern part of the elevated plain with an area of 35,830 km². The Changbai Mountains is its southeast boundary. The Songhuajiang River is its northern boundary. The local groundwater divide along the second Songhuajiang River is its western boundary. The unconfined aquifer consists of loessal silt loam and silt clay with thickness of 10~30 m. Depth to the water table is less than 5 m. Groundwater is recharged by infiltration of precipitation and seepage of rivers from mountains. Groundwater flows from the south boundary towards the central low plain and discharges into the Songhuajiang River.

The Quaternary confined aquifer (Q₁) which consists of gravel and sand mainly distributes in the central low plain and lies below a depth of 20–60 m. Its thickness is 30–70 m in the northwest part and 5–30 m in the east and south part of the plain. The biggest thickness of about 106 m is found in Qiqihaer. There is a regional clay layer as an aquitard between this aquifer and the upper unconfined aquifer. However, there is no continuous aquitard between the unconfined upper aquifer and the lower Tertiary aquifer. Naturally, the hydraulic head of the Quaternary confined aquifer is lower than that of the unconfined aquifer near the piedmont and groundwater of the upper aquifer flows into the confined aquifer by leakage. The hydraulic head of the confined aquifer is higher than that of the unconfined aquifer in the discharge area at the convergence of the Nenjiang and Songhuajiang Rivers, and groundwater in the confined aquifer discharges into the unconfined aquifer. However, groundwater abstraction from the confined aquifer causes its hydraulic head to be lower than that of the unconfined aquifer in some areas (e.g. Daqing), resulting in leakage from the upper aquifer.

The confined Tertiary aquifer (N) lies below a depth of 80–140 m and consists of sandstone and argillaceous sandstone with a thickness of 10–100 m. Recharged water flows through the upper Quaternary aquifer in the piedmont and enters the Tertiary aquifer by lateral flow in its marginal zone. But its hydraulic head is higher than that of the upper Quaternary confined aquifer in the central plain and this results in water discharge to the upper aquifer by leakage.

The undisturbed groundwater level (June 1988) of the unconfined aquifer shown in Fig. 1 indicates that groundwater moved from the plain margins toward the centre. Due to the steady increase of groundwater pumping the water level has considerably declined and a pumping cone has developed in the vicinity of Daqing and Qiqihaer, resulting in changes of the groundwater-flow direction towards its centre.

Multiple sources of groundwater recharge are present within the Songnen Basin (Zhao et al. 2009). The most important one is the mountain-river infiltration, which occurs parallel to mountainous areas.

Samples and methods

Samples were collected for isotope analyses in 2003 and 2004. Shallow groundwater was sampled from different aquifers (Fig. 1). Surface water samples were collected along the Nenjiang River. Samples from the deep confined aquifers were mainly taken in the low central plain along the flow path. The Cretaceous aquifers were not sampled due to the lack of suitable wells.

Prior to sampling, the wells were pumped until the electrical conductance (EC) remained constant. Samples for stable isotope analysis (δ¹⁸O and δ²H) and for tritium measurement (³H) were collected in 50 and 500-ml glass bottles, respectively, and closed with airtight caps. The dissolved inorganic carbon (DIC) for ¹⁴C determination was extracted as BaCO₃ from 120-L groundwater by adding BaCl₂ and CO₂-free NaOH until pH 12. Sample analyses using standard methods were performed at the Institute of Hydrogeology and Environmental Geology (IHEG), Chinese Academy of Geological Sciences. The water for stable isotope ratio measurements of oxygen and hydrogen were treated by the CO₂ equilibration method (Epstein and Mayeda 1953) and the zinc-reduction method (Coleman et al. 1982), respectively. The mass spectrometric analysis was carried out with a Finnigan MAT 253. The isotope compositions were reported in standard δ notation represented as per mil deviation from the V-SMOW (Vienna Standard Mean Ocean Water) standard. The precision of the δ²H and δ¹⁸O values is ±1.0 and ±0.2‰, respectively.

Tritium values were determined by liquid scintillation counting (1220 Quantulus) on electrolytically enriched water samples and given in tritium units (TU). The 1 σ analytical precision is about ±0.5 TU corresponding to a detection limit of about ≥ 1 TU. ¹⁴C of DIC was determined radiometrically by liquid scintillation counting (1220 Quantulus) after conversion to benzene. The specific ¹⁴C activity was reported in the unit of percent modern carbon (pMC). The results are compiled in Tables 1 and 2.

Table 1 Sample information, and analytical results of isotopes for waters

Table 2 Groundwater ages determined by application of ¹⁴C-DIC models

Results

Stable isotopes in precipitation

Precipitation for isotope analysis was sampled in Changchun (1999–2001; 43.9° N, 125.22° E, elevation 237 m asl) , Haerbin (1986–1990; 45.68° N, 126.62° E, elevation 172 m  asl) and Qiqihaer (1988–1992; 47.23° N, 123.55° E, elevation 147 m  asl ) by IHEG. The weighted mean values of δ¹⁸O and δ²H values of the three sites were –9.3 and –67‰, –9.6 and –70‰, –10.6 and –77‰, respectively. Several researchers (Zhang et al. 1991; Wang 1995; Liu et al. 1997) obtained deviating δ¹⁸O/δ²H relationships for each site due to a too short and often discontinuous monitoring period. Hence, the published meteoric water lines may not reflect the actual isotope composition of the regional precipitation. The least-squares fitted (LSF) reliable regional meteoric water line (RMWL) from the study data of the above three stations is δ²H = 7.6 × δ¹⁸O + 4.5‰ and represents the precipitation of the three stations (IAEA/WMO 2006).

Isotopes in river water

Seven samples of surface water were collected along the perennial Nenjiang River in order to localize the headwater area by means of the isotopic signature. Samples in the central plain were not taken because of mixing of discharged local groundwater.

The δ¹⁸O and δ²H values ranged from –14.5 to –12.5‰ and –103 to –92‰, respectively. The mean values are –13.6 ± 0.8‰ and –98 ± 5‰, respectively, which are more negative than the mean value of precipitation in Qiqihaer area (Fig. 3). The depletion of ¹⁸O varies between about –0.15 and –0.5‰ per 100-m rise in elevation, and an altitude effect of –0.25‰ per 100-m rise between 250 and 3,250 m in elevation was provided by Clark and Fritz (1997). The mean elevation of the catchments of the rivers was estimated to be 1,100 m asl above the elevation of Qiqihaer (147 m asl) using the isotopic altitude effect of –0.25‰/100 m. The tritium values range from 11 to 29 TU. They are representative of the precipitation in the High and Low Hinggan Mountains.

Fig. 3

Plot of δ²H (δD) versus δ¹⁸O for shallow groundwater systems: a the Nemoer-Keluo and Nuomin-Taoer system, b the Wuyuer-Shuangyang system, and c the Tongken-Hulan and the second Songhuajiang system

Isotopes in groundwater

δ¹⁸O and δ²H values of groundwater in the unconfined aquifer increase along the flow path from the piedmont area and elevated plain towards the central low plain. Those from the piedmont area have ranges of –11.8 < δ¹⁸O < –8.6‰ (mean value of –10.1 ± 1.4‰) and –91 < δ²H < –68‰ (mean value of –78.0 ± 7‰). Groundwater in the elevated plain has similar stable isotope compositions as that in the piedmont plain with δ¹⁸O and δ²H values of –11.0 to –8.7‰ (–9.9 ± 0.7‰) and –79 to –66‰ (mean value –73 ± 1‰), respectively. The tritium value of groundwater in the Songnen Plain ranges from 1 to 60 TU. Tritium values >10 TU are representative of groundwater emerging along the piedmont area.

Groundwater in the assigned five shallow groundwater systems each has its specific isotopic characterization. The δ¹⁸O and δ²H values in the Nemoer-Keluo system (I) ranges from –11.8 to –10.5‰ (mean – 11.4 ± 0.5‰) and from –91 to –81‰ (mean –86 ± 3‰), respectively. Tritium values are between 2 and 40 TU within 50 m below ground. These ranges correspond to those of the mountain rivers which have relatively low δ¹⁸O and δ²H values. The groundwater is somewhat more depleted in heavy isotopes than precipitation. The delta values plot on a local meteoric water line (LMWL) with a slope of 5.1 (Fig. 3a). The δ¹⁸O and δ²H values in the Nuomin-Taoer System (II) ranges from –10.6 to –8.1‰ (mean –9.4 ± 0.5‰) and –81 to –68‰ (mean –74 ± 3‰), respectively. Tritium (3–60 TU) is present down to 110 m below ground. The delta values plot along the local meteoric water line δ²H = 5.2 × δ¹⁸O –23.5 and are less negative than those of the Nemoer-Keluo system (I) (Fig. 3a). The δ¹⁸O and δ²H values in the Wuyuer-Shuangyang System (III) range from –10.8 to –9.2‰ (mean –10.0 ± 0.5‰) and –83 to –68‰ (mean –76 ± 5‰), respectively. Tritium (3–30 TU) is found within 50 m below ground. Two distinct groups of groundwater are shown in Fig. 3b. One group represented the samples collected near the Wuyuer River in the north part of elevated plain. Their δ¹⁸O and δ²H values plot along the fitted LMWL, δ²H = 4.3 × δ¹⁸O –37.8. The other group was sampled in the south-west part of the central low plain. The delta values fit a LMWL, δ²H = 4.4 × δ¹⁸O –30.1. The δ¹⁸O and δ²H values in the Tongken-Hulan System (IV) range from –11.0 to –9.4‰ (mean –10.3 ± 0.5‰) and from –81 to –68‰ (mean value –76 ± 3‰), respectively. The delta values cluster along the RMWL within the range between the Changchun and Qiqihaer precipitation (Fig. 3c). The tritium values range from 3 to 60 TU. The δ¹⁸O and δ²H values in the second Songhuajiang System (V) increase along the flow path towards the centre of the plain. They range from –9.7 to –8.7‰ (mean –9.2 ± 0.4‰) and –76 to –66‰ (mean –71.0 ± 3.0‰), respectively. The tritium value ranges from 3 to 20 TU.

Groundwater samples of the Quaternary and Tertiary confined aquifers were mainly collected from the central low plain. The δ¹⁸O and δ²H values of the groundwater in the Quaternary confined aquifer range from –11.3 to –8.7‰ (mean value –9.8 ± 0.6‰) and from –81 to –74‰ (mean value –77 ± 4‰), respectively. Tritium content is less than 5 TU and ¹⁴C activity ranges from 83.6 to 6 pMC. The Tertiary groundwater samples were mainly collected from the central low plain. These samples do not have detectable tritium and have a specific ¹⁴C activity less than 60 pMC. The δ¹⁸O and δ²H values of the groundwater in the confined Tertiary aquifer range from –11.9 to –9.1‰ (mean –10.2 ± 0.9‰) and –93 to –71‰ (mean –81 ± 1‰), respectively.

Discussion

Recharge to aquifers

Stable isotope data can be used to evaluate possible sources of water to the Quaternary and Tertiary confined aquifers. The δ²H–δ¹⁸O plots of groundwater in Tertiary and Quaternary confined aquifers is shown in Fig. 4. The samples of Quaternary groundwater plot on a LMWL of δ²H = 4.4 × δ¹⁸O –33.4, while the samples of Tertiary groundwater plot on a LMWL of δ² H = 7.2 × δ¹⁸O –7.1, which is parallel to the regional meteoric water line of δ²H = 7.6 × δ¹⁸O + 4.5. There is no distinction between these two aquifers, suggesting that the indirect groundwater recharge by precipitation is the same for both aquifers. All the samples of confined aquifers shown in Fig. 4 deviate from the RMWL and fit a line of δ² H = 6.2 × δ¹⁸O –16.3 which is analogous to that of the Tertiary aquifer (Fig. 4). This fit line is tentatively termed the isotope recharge line (IRL). Groundwater with such isotopic composition has been found in semi-arid regions of Gurinai (China), Canada and Chile, and the delta values plot to the left of the IRL in the case of deep groundwater levels (Geyh and Gu 1999). The typical hydrological conditions of the Songnen Plain are a sediment cover and a shallow water table. The delta values of samples from shallow groundwater systems I–V scatter between the RMWL and the IRL (Fig. 5). Two possible processes, admixture of river water and admixture of irrigation return flow, may be responsible for this feature.

Fig. 4

Plot of δ²H versus δ¹⁸O for groundwater in Quaternary and Tertiary aquifer

Fig. 5

Plot of all δ²H versus δ¹⁸O values for all samples from Systems I–V

The groundwater in system I is somewhat more depleted in heavy isotopes than groundwater in other shallow systems (Fig. 5). In system I nitrate concentration ranges from 2.7 to 8.5 mg/l with a mean value 6.2 mg/l. Background levels for nitrate in groundwater in the piedmont are always less than 10 mg/l (Zhao et al. 2009). Nitrogen fertilizers have been applied to agricultural land in the plain during the last two decades. The residual fertilizer may contribute high concentrations of nitrate to the underlying aquifer by water percolating through the soil. Where irrigation occurs, the return water will carry nitrate into shallow groundwater. As a result, an elevated nitrate concentration (>10 mg/l) can be expected. Thus, an admixture of irrigation return in system I can be excluded and it suggests that the groundwater contains some recharge water from rivers emerging in the high mountains. The intersection of its LMWL with the RMWL has δ¹⁸O and δ²H values of –13.2 and –96‰, respectively, similar to those of mountains rivers (Fig. 3a). Adopting the δ¹⁸O-altitude gradient of –0.25‰/100 m (Clark and Fritz 1997), the average elevation of the catchment of the river is estimated about 1,000 m asl which is within the range of the High and Low Hinggan Mountains. Hence, the groundwater of this system is mainly recharged by mountains river seepage along the mountain front.

The delta values of groundwater in system II are less negative than those of system I (Fig. 3a). Samples collected in the southwest low plain (s17-1 and s22-1) plot closely to RMWL, while samples from the west piedmont plain (Nos. s11, s24, s19, s25, s18-1.s23-1 and s26) plot closely to IRL. This may indicate that both local precipitation and seepage of the high mountain rivers contribute to groundwater recharge. The nitrate concentration of groundwater ranges from 1.0 to 90.0 mg/l with a mean value 27.2 mg/l. This elevated nitrate concentration is consequently associated with agricultural practices. However, the irrigation return should not be a dominant recharge process because of a slightly elevated nitrate concentration in groundwater. The δ¹⁸O and δ²H values of the intersection point of the LMWL with the RMWL are –12.1‰ and –87‰ (Fig. 3a), respectively. This indicates that the origin of the recharged groundwater may be different from system I. But these delta values are more negative than those of the Qiqihaer precipitation. Hence, the groundwater seems to be a mixture of seeped local precipitation in the piedmont area and stream seepage of the high mountain rivers along the mountain front. Using mean δ¹⁸O value for Qiqihaer precipitation (–10.6‰) and Nenjiang River (–13.6‰) as two end members, the relative contributions of recharge from river and precipitation can be calculated using a mass balance method. This gives estimates of 55 and 45% for the relative contributions of river recharge and precipitation to the shallow groundwater, respectively.

The nitrate concentration of groundwater in system III ranges from 1.4 to 594.0 mg/l with a mean value 173.8 mg/l; thus, the irrigation return could be a contribution of recharge. The feature, that the delta values of samples (Nos. s28–s33) from the elevated plain are lower than those from the central low plain (Nos. s34–s40), may indicate the different sources of irrigation water. The first group mainly distributes along the Wuyuer River. The δ¹⁸O and δ²H values of the intersection point with the LMWL are –12.8 and –93‰, respectively. This means that water for irrigation is from the rivers emerging in the Low Hinggan Mountains. Another group distributes in the central low plain. The isotope composition at the intersection of its LMWL corresponds to the weighted-mean isotope composition of the Qiqihaer precipitation. Therefore, the source of irrigation is from local groundwater.

The delta values of groundwater in system IV cluster along the RMWL within the range between the Changchun and Qiqihaer precipitation (Fig. 3c). The nitrate concentration of groundwater ranges from 1.1 to 53.2 mg/l with a mean value 21.8 mg/l. This indicates indirect groundwater recharge by local precipitation in the elevated plain. This is reasonable since the mean annual precipitation is 500–600 mm in the elevated plain.

The delta values of groundwater in system V plot on the fitted LMWL δ²H = 5.4 × δ¹⁸O –21.5. The δ¹⁸O and δ²H values of the intersection point with LMWL are –12.1 and –87.0‰, respectively, which are more negative than those of the weighted mean of the Changchun precipitation (Fig. 3c).The average elevation of groundwater recharge seems to be 1,000 m asl above Changchun (elevation 237 m asl.). The Changbai Mountains, as the headwater area of the second Songhuajiang River, is located in the southeast with an elevation of 300–2,600 m asl. The δ¹⁸O and δ²H values of the precipitation in the Changbai Mountains range from –10.0 to –14.4‰ and from –70 to –100‰, respectively (Wang 1997). The arithmetical mean δ¹⁸O and δ²H values of eight sites (Wang 1997) are 12.2 ± 0.47‰ and 86 ± 4.0‰, respectively. The initial isotope signature of recharged water is close to that of precipitation in the Changbai Mountains. Therefore, the source of the groundwater in system V is infiltrated river water from the Changbai Mountains.

Groundwater dating by ¹⁴C

The groundwater in the unconfined aquifer has tritium values ≥3 TU and is modern with an isotopic mean residence time of ≤50 years. However, groundwater in confined aquifers in the central low plain has no detectable tritium or contains less than 3 TU. This groundwater has been recharged pre-1950s. The age of this groundwater was determined by the ¹⁴C dating method.

Radiocarbon dating of DIC in groundwater is based on the radioactive decay of ¹⁴C. Rainwater infiltrates in the soil and dissolves biogenic CO₂ containing much ¹⁴C during groundwater recharge. On its way further down to the aquifer, the formed carbonic acid dissolves old soil carbonate without ¹⁴C (dissolved inorganic carbon, DIC). Other physical and chemical processes in the aquifer may affect the ¹⁴C activity of DIC. This has to be taken into account if the radiocarbon age of DIC is converted into actual ages. The first step of ¹⁴C groundwater dating is the determination of the initial ¹⁴C activity (A ₀); the second step involves the correction of the ¹⁴C value controlled by geochemical reactions in the aquifer.

Groundwater with tritium contains, at least partly, water recharged after 1950 and may also have a ¹⁴C value exceeding 100 pMC. Sample No. s1 from the north piedmont area has ³H and ¹⁴C values of 23 TU and 114.1 pMC, respectively. It, therefore, contains at least a fraction of recharge of post-bomb age. Young groundwater samples that are tritium-free from the recharge area may be used to estimate the initial ¹⁴C activity. Sample No. s10-2, collected in the west piedmont, has a tritium value of ≤1 TU and a ¹⁴C activity of 92.3 pMC. This seems to indicate a value of about 100 pMC for A₀. However, these two samples have the lowest δ¹³C values (–18.6‰ for s1 and –13.8‰ for s10-2) and bicarbonate concentrations (55.79 mg/l for s1 and 103.73 mg/l for s10-2). This finding indicates that they are less affected by carbonate dissolution than all the other samples. Hence it is not appropriate to estimate A₀ from these samples.

Generally, the initial concentration of dissolved inorganic carbon (DIC) is controlled by the CO₂ concentration in the topsoil (Fontes 1983; Kalin 1999). Secondary geochemical reactions in aquifers change it and lower the ¹⁴C activity of DIC. This effect has to be quantified in order to correct the ¹⁴C activity of DIC. In the Songnen Plain, several processes have to be considered:

• 1. Dissolution of carbonate: The initial δ¹³C value of –18.6‰ (No. s1) and –13.8‰ (No. s10-2) of the young samples increase to –7.6‰ (No. s77) and –7.9‰ (No. s17-3) in the central plain along the flow path. This implies secondary calcite dissolution in the aquifers. The linear trends between δ¹³C value, HCO₃ and Ca²⁺ support this inference (Fig. 6)

• 2. Cation exchange lowers the calcite saturation and enables additional dissolution of calcite in the aquifers. However, the absence of a linear relationship between Na⁺ and Ca²⁺ excludes this process (Fig. 6).

Fig. 6

Plots of a Ca²⁺ versus HCO ^–₃ , b δ¹³C versus HCO ^–₃ , and c Ca²⁺ versus Na⁺ for groundwater samples from confined aquifers

Several methods have been proposed to estimate A ₀ (Vogel 1970; Tamers 1975; Ingerson and Pearson 1964; Mook 1976; Fontes and Garnier 1979 and Kalin 1999). These methods consider either only chemical mixing between the carbon compounds and/or isotopic fractionation. The use of these models requires information on the δ¹³C and ¹⁴C values of the soil CO₂ and carbonate minerals in the recharge area. A δ¹³C values of –19‰ for soil CO₂ gas and 0‰ for soil carbonate were adopted to calculate groundwater age in the southwest part of the plain by Wang (1995). The same values were used for ¹⁴C dating of groundwater in the North China Plain (Kreuzer et al. 2009). These values were adopted for this study for the calculation of ¹⁴C ages of groundwater in the Songnen Plain. Six methods were used: (1) Vogel, (2) Tamers, (3) mass balance, (4) Ingerson and Pearson, (5) Fontes and Garnier and (6) IAEA.

The simplest method is the Vogel approach (Vogel and Ehhalt 1963; Vogel 1967; Vogel 1970). A ₀ = 85 pMC is assumed. The default ¹⁴C value of all models is 100 pMC for soil CO₂ gas and 0 pMC for soil carbonate. The Tamers model takes the bicarbonate and total DIC into account. Ingerson and Pearson (1964) estimate A ₀ using the δ¹³C values of DIC and soil CO₂ gas. Both the Tamers and Ingerson and Pearson models assume calcite dissolution under closed system conditions without exchange with soil CO₂ gas. The chemical mass balance model (CMB; Clark and Fritz 1997) also takes into account the involvement of dolomite, gypsum and CO₂ besides carbonate.

Fontes and Garnier (1979) consider isotope exchange of DIC with either the gas phase or the solid phase. This model describes a two-stage evolution of recharged groundwater accounting for dissolution and isotopic exchange of carbonate minerals with CO₂ in the unsaturated zone and isotopic exchange with carbonate in the saturated zone. Finally, the IAEA model, which was first published by Salem et al. (1980), considers mixing and some isotope exchange. The Tamers and Ingerson and Pearson models yielded similar mean A ₀ values of 53.6 ± 3.3 pMC and 59.8 ± 9.1 pMC, respectively. The Fontes and Garnier model gave a slightly elevated mean A ₀ of 69.4 ± 15.0 pMC.

The model ages calculated by these three A ₀ were not far from those of the Vogel method. However, the CMB model and IAEA model yielded considerably lower mean A ₀ values of 29.5 ± 11 pMC and 40.5 ± 7.3 pMC, respectively. In spite of the empirical result of A ₀ ≈ 100 pMC derived from ³H and ¹⁴C results of young groundwater, the mean value of the models by Vogel, Tamers, Ingerson and Pearson as well as Fontes and Garnier of around 70 pMC is most likely the actual A ₀ (Table 2). Accordingly, the ¹⁴C age of the groundwaters in the confined Quaternary and Tertiary aquifers range from 0 to 19,500 years, and from 3,100 to 24,900 years, respectively (Table 2). Groundwater age of the confined Quaternary aquifer increases along the direction of the groundwater flow from the piedmont and the elevated plain towards the central low plain (Fig. 7). Modern water ages were found along the west and east margins of the Quaternary aquifer. This result confirms present-day groundwater recharge for the confined aquifer at the west piedmont area and the east elevated plain by local precipitation and lateral subsurface flow.

Fig. 7

Spatial distribution of the ¹⁴C model ages of groundwater in the confined aquifers. The first number is the sample number (as for Fig 1) and the second value is the age (a = years)

The ¹⁴C ages can be used to delineate groundwater flow paths. Samples of relatively young age are located along the west piedmont (Nos. s71, s75 and s14-2) and east elevated plain (Nos. s78 and s70-1), while samples of old age occur in the central low plain. Generally, their distribution is consistent with the direction of groundwater flow. Groundwater age increases from 5,800 years (No. s79) to 10,100 years (No. s38-2) to the north of Daqing, which indicates that groundwater moves from SE to NW. This is consistent with the distribution of the piezometric surface (Fig. 7). However, ¹⁴C age of sample No. s38-2 (a well in Daqing) is 4,900 years which is contrary to the previously mentioned flow direction. This probably reflects the mixing of young water because of the cone of depression around Daqing. The elevated bicarbonate concentration (400–600 mg/l) may support this explanation. A similar case is also found in Qiqihaer. The ¹⁴C age inversion indicated another pumping cone near Qiqihaer. One sample from a 79 m-deep well (No. s23-2) in the central low plain has a relative young age (3,400 years). Its stable isotope ratio (δ¹⁸O: –9.4‰, δ²H: –75‰) lies close to the shallow samples (No. s23-2 with depth 24 m, δ¹⁸O: –8.7‰, δ²H: –71‰). Hydrogeological investigation found there are “windows” in the upper confining bed in this area (Zhao et al. 2009). Groundwater abstraction from the confined aquifer results in leakage from the shallow aquifer. The ¹⁴C age of the Quaternary well No. s83 (19,500 years) fits into the ¹⁴C age distribution of the groundwater in the Tertiary aquifer. Compared with the Tertiary well No. s22, their water chemistry and stable isotopic compositions are almost the same. Similarly, well No. s82-1 in the Quaternary aquifer and No. s82-2 in Tertiary aquifer have the same ¹⁴C age. These imply that the Quaternary and Tertiary aquifers are hydraulically connected.

Groundwater recharge rate to aquifers can be evaluated from ¹⁴C ages. The method was summarized by Vogel (1967) and Cook and Herczeg (2000). Assume an aquifer of constant thickness receives uniform recharge in the unconfined portion and receives no recharge (no vertical leakage) in the confined portion of the aquifer, the recharge rate can be estimated from the groundwater ages by using the Vogel (1967) model. The equation for the recharge estimation is:

$$ R = \frac{{H\varepsilon }}{{xt}}\left[ {x\ln (\frac{H}{{H - z}}) + x * } \right] $$

Where H is the total thickness of the aquifer and ε is mean aquifer porosity. x and x ^* are the horizontal length of recharge area and confined portion of the aquifer, respectively. z is sample depth below the water table. t is the age of a groundwater parcel at depth z.

Groundwater ages of nine samples which were regarded as being recharged from the piedmont were used to estimate recharge rate for confined aquifers (Table 3). The mean thickness of aquifer used in the calculation was 50 m for the Quaternary confined aquifer. Porosity of 0.2 was assigned to this aquifer as a representative value (Zhao et al. 2009). Application of the above equation to these data sets yields recharge rates that range from about 2 to 11 mm/a (mean 6.5 mm/a) for the Quaternary confined aquifer.

Table 3 Recharge rate estimated from ¹⁴C ages

Conclusions

Environmental isotopes in groundwater of the Songnen Plain were used to identify different sources of water contributing to the groundwater recharge. The δ¹⁸O and δ²H values confirm that the unconfined aquifers in the north and west piedmont are mainly recharged by drainage of the mountain rivers emerging in the High and Low Hinggan Mountains. The southeast part of the unconfined aquifer of the elevated plain is recharged from rivers emerging in the Changbai Mountains. Irrigation return and local precipitation are the main recharge sources of the shallow aquifer in the central low plain and the east elevated plain, respectively. Evidence is provided by detectable tritium in the groundwater down to a depth of ~110 m in the piedmont, ~80 m in the central plain, and less than 50 m in the elevated plain, and associated mean residence times younger than 50 years.

The isotopic signature of groundwater in the confined Quaternary aquifer provides evidence for groundwater recharge by infiltration of local precipitation and lateral subsurface inflow from the mountains. The ¹⁴C model age of this groundwater ranges from modern to 19,500 years. Modern groundwater only occurs along the margins of the piedmont and the elevated plain. The groundwater in the Tertiary aquifer (N) is free of detectable tritium and the ¹⁴C model age ranges from 3,100 to 24,900 years, indicating recharge during the late Pleistocene to middle Holocene.