Achievements

THE GROUNDWATER RECHARGE AND GEOCHEMICAL EVOLUTION FROM HEXI CORRIDOR TOINNER MONGOLIA PLATEAU, THE THE ARID ZONES OF NOTHWEST CHINA

Updated :10,12,2012

J.Z. Maa, ,  Z. Y. Ding T. M. Huanga, Y. H. Sub

a Key Laboratory of Western China’s Environmental System(Ministry of Education), Lanzhou University, Lanzhou 730000, China

b Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences, Lanzhou 730000, China

 

Abstract:The long term recharge in Hexi corridor and Inner Mongolia Plateau was estimated to be 0.96m yr-1 by using the chloride mass balance method from one unsaturated zone profile, which shows that no effective modern recharge is taking place. Overall the groundwater quality is good for drinking and irrigation in the Quaternary aquifer of Heihe River and Shiyang river basins. The salinities and the major ions in groundwater increases sharply down stream from the debouchure, more than 1 g l-1 TDS in the lower reaches. The mineralization of groundwater decreases from the shallow aquifer to the deep aquifer. The lakes in the Badain Jaran desert are generally of fresh water, although with variable compositions which suggest some in-basin biogeochemical modification. Stable isotopes (δ18O and δ2H) provide clear distinction between modern and ancient waters. A good rainfall database from nearby Zhangye provides definition of the composition of modern rainfall. The signature of groundwater from the late Pleistocene differs markedly from that of the Holocene, shown clearly by the stable isotopic compositions of -10.5‰ δ18O as compared with values of -7‰ at the present day. it is apparent that the groundwaters in the Minqin Basin, Ejina basin and feeding the lake system of the Badain Jaran are part of a regional flow network related to a wetter past climate as source of recharge. This system is largely decoupled from the river network in this area. The recharge source in the past and to a limited extent in the more arid conditions of the present day included the foothills of the mountains of the Tibetan plateau.The tritium age determinations accurate to the year are impossible and of no meaning to groundwater studies in the region due to tritium measurements in the shallow unconfined groundwater environment could either be influenced by vapor exchange with the soil atmosphere or an expression very localised modern recharge. A tritium value in the groundwater means multiple recharge ages in this region.

Key words: Groundwater, Recharge, Geochemistry, Isotope, Hexi corridor, Inner Mongolia Plateau



 



1. Introduction

Over the past decades, the Chinese government and scientists have carried out much research on the assessment, development, utilization and conservation of groundwater resources in northwest. However, the previous methods of groundwater research are mainly based on water balance, which are inadequate for low rainfall areas with high evapotranspiration(Allison et al.,1994). Research interests have been focused on aspects such as utilization survey, evolutionary prediction, reserve estimation and systematic assessment of natural water resources. Some reasonable allocation and utilization plans have been proposed. Study of the hydrochemistry (especially isotopic geochemistry) of water resources is rather sparse and past work on the arid area has dealt largely with the chemical properties of the water, and efforts to use the geochemistry data available to solve particular problems are even fewer or non-existent, although there are several investigations on the interaction and mechanism between surface water, groundwater and rock in particular areas in recent years (Shi et al., 2001; Feng et al, 2004; Zhang et al., 2005). Shi et al. (2001) have reported stable isotope data showing values of –6.6 to –9.7 δ18O which they claim to represent modern recharge, modified by evaporation and that the groundwater is sourced from the Shiyang River. These authors also report helium isotope data which shows some enrichment and also that the 3He/4He ratios denote a component of mantle helium. Feng et al. (2004) assessed the surface- vs. ground-water chemistry in the Heihe River basin and then Zhang et al. (2005) reported the hydrochemistry characteristic and groundwater recharge area using d18O and T in the Ejina basin, the lower reach of Heihe River. They concluded that the water recharge of the area is from the mountain area of the China–Mongolia boundary; and the main source of the groundwater in Gurinai oasis is from the leakage of the ancient river, and not from the Badain Jaran Desert.

However, it is short of the regional understand of the groundwater recharge and chemical evolution, the volume of water available in the aquifer is not known and its rate of recharge, if any, is very uncertain. However, there can be little doubt that water is being extracted much more rapidly than it is being replaced, that surface water supply is threatened by global warming, that the quality of water may change as volume declines, and that local farmers have little understanding of the problem(Ma et al., 2005(a)). This is an issue that faces the whole regions of arid China and other arid regions throughout the world. The present study focus on the issue as a part of an interdisciplinary study to look at the palaeo-ecology, climatic change and social consequences in the context from Hexi corridor to Inner Mongolia Plateau as the typical areas of arid NW China.

The scope of the present study and the main objective is to estimation of natural groundwater recharge to aquifers, and understand the groundwater evolution and identify chemical changes taking place along the original direction of groundwater flow towards the terminal lake of Inland river basins, and the specific targets are to

l      use the geochemical and isotopic technology to determine the present day recharge rates from unsaturated zone of Desert,

l      determine which are the predominant geochemical process taking place along the inferred horizontal groundwater flow line,

l      use stable isotopes and tritium to determine the evolution and the age of the groundwater under the natural conditions during the recent geological past.

This approach has been used successfully in other semi-arid regions underlain by large freshwater resources (Edmunds et al., 2003). Similar comprehensive approaches to groundwater understanding combining chemistry and isotopic indicators are relatively few in China but palaeoclimatic studies combined with chemistry are beginning to show that basins in Northern China were recharged under wetter conditions in the past and therefore non-renewable (Chen Z et al., 2003). Improved understanding of the resource therefore is fundamental to many important decisions that will need to be made in China (as in other arid regions), in relation to demographic change and the reform of agricultural practices (Ma et al., 2005(a)).

2. Geology and Hydrogeology

The study area includes the Shiyang river basin, Heihe basin and Baddain Jaran desert. From southwest to northeast, landforms can be divided into the Qilian Mountains, Hexi Corridor basin and LongshouMountains, gobi desert and Badain Jaran desert. The Qilian Mountains, trending WNW, are over 4500 m above sea level. The Hexi Corridor basins are clinoplains with altitudes of around 2000 m in the southwest and1500 m above sea level in the northeast. From southeast to northwest, three uplifts, the Dahuang Mountains, Yumu Mountains and Wenshu Mountains, separate the Hexi Corridor into the Wuwei, Zhangye, Jiuquan and Yumen basins.

The Hexi Corridor, located on the northeast margin of the Qinghai–Tibet Plateau, is a Cenozoic foreland basin system. The subduction of the Indian Plate under the Eurasian Plate caused the Qinghai–Tibet Plateau to become uplifted and pushed to the north (Harrison et al.,. 1992). The fault zone at the northeast front of the Qilian Mountain, trending WNW and dipping to SSW, is an active reverse fault zone. The active faults in the Hexi Corridor are mainly reverse faults. The faults trending WNW are dominated by sinistral slip and the faults trending NNW with dextral slip. This indicates a NE–SW compression in the Hexi Corridor. The uplift of the Qilian geosyncline occurred from the end of the Palaeozoic throughout the whole Mesozoic era, and created the embryonic form of the Hexi Corridor. This was followed by a complex tectonic stage when the structural zonesLongshou-hongya Mountain, Heli-Mazong Mountain were formed in the middle of the Hexi corridor; cutting the whole Corridor into two parts, that is the southern Basins and the northern one. From east to west in the southern part, three Tertiary uplifts, the Dahuang Mountains, Yumu Mountains and Wenshu Mountains, separate the Hexi Corridor into the Wuwei, Zhangye, Jiuquan and Yumen basins.

From the late Tertiary, especially from the end of the Pliocene and the beginning of the early Pleistocene, the surrounding Tibetan mountains began to rise up rapidly (Li et al., 1979) and the basins subsided further. At the same time the intensive denudation and erosion from the mountains led to significant transfer of clastic material to the basin depressions. This formation of thick Quaternary diluvial and alluvial sediments, and some aeolian and lacustrine deposits, led to the formation of the main aquifers.

Huge loosen Quaternary sediments with maximum over 1000 m had been deposited during geological history in the fore depression, which provides a large space for groundwater reserve and transport. The Quaternary basins can be divide into geomorphologic units such as piedmont alluvial plain, alluvial lacustrine plain and desert; and the sedimental compositions gradually turn from large pore gravel to medium and fine sand and silt respectively.

The active reverse faults zone and derivative Cenozoic fold along the North piedmont Qilian Mt. result in very complex groundwater hydraulic contact between the Mountain and Plain. In the eastern part such as in the Wuwei of Shiyang river and Minle in the Heihe river there occurs blind imbricate fault zone of 2-6 km wider and there is very thin Quaternary deposit on the fault bench which is recharged by the rainfall and infiltration water. There is a water table difference of 4-20m between upper wall and lying wall of the fault bench and there is a water difference of 80-200m between fault bench and Quaternary aquifer in the Plain (see A in Fig.1). In the western Heihe river there developed a fault mylonite zone of 100-400m wider, which is impermeable and lead to a subsurface linn of 120-180m difference between Mountain and plain aquifer(See B in Fig.1). There partially occurs no fault mylonite zone and the Quaternary aquifer contact with Mountain directly through fault such as in Longwangmiao in Zhangye and Maying in Shandan. The groundwater flow from the mountain area is absolutely obstruct and distribute a lot of springs along the upper wall of the fault, and the groundwater table difference amount to 200m between the Mountain and Plain(see C in Fig.1).

However, the Quaternary aquifers are all closely connected with streams originated in the Qilian mountains which discharge the tectonic fracture groundwater. These streams pass through two or three basins and flow into the terminal lakes once the runoff formed by melting ice-snow, precipitation and the springs in the mountains. There is a complex transformation relation between the surface and groundwater. More than 80%of surface water in the piedmont fan seeping down to the aquifer and then in the fine earthy plain belt, about 40% of groundwater discharge to feed the spring and river. Such transformation process of surface water to groundwater and then to surface water again in the quaternary basins maybe repeated many times because of the existence of tectonic uplifted zone in the middle of the Hexi corridor such as Longshou-Hongya Mountain, Heli-Mazong Mountain and some small hills, which led to groundwater barely crosses the fracture from south basin to recharge the north basins.

3 Methods

Field work took place from the summer of 1999 to September of 2003 during which several profiles are drilled with augurs and about 100 groundwater samples were collected. All groundwater samples were filtered (0.45um membrane filters) and an aliquot acidified with 1% HNO3 for analysis of trace elements. Unacidified samples were collected for anion analysis as well as for stable isotope analysis. On site analysis included temperature, specific electrical conductance (SEC), total alkalinity (as HCO3-) by titration and pH. Chloride, NO3-N, Br and F were analysed by automated colorimetry. Filtered and acidified samples were analysed for major cations, SO4, and trace elements either by ICP-OES (inductively coupled plasma optical emission spectroscopy) or ICP-MS (inductively coupled plasma mass spectrometry). Calibrations for cation analyses were performed using appropriately diluted standards and both laboratory and international reference materials were used as checks for accuracy. Instrumental drift during ICP-MS analysis was corrected using In and Pt internal standards. Samples for stable isotope analysis (18O, 2H) were measured by isotope ratio mass spectrometry. Determining the ionic balance provided an internal check on the quality of the data; the balance lay within ±6% except for four samples. Precision of measurement for stable isotopes was ± 0.1‰ for δ18O, and ±2‰ for δ2H.

Samples of sands from the unsaturated zone from two areas some 400m apart in the southern Badain Jaran desert, near to the Sayinwusu lake were taken by hollow stemhand augur (Edmunds et al. 1988) for investigation of the moisture




Fig.1 The simple hydrological sections showing the hydraulic contact relationship between the Qilian Mountain and Quaternary aquifer in Hexi Corridor.

 

content and unsaturated zone chemistry to obtain data on modern recharge rates and to determine recharge history. These samples were taken to compare with the isotopic and chemical compositions of the lakes and springs, as well as to compare the recharge record as a proxy for climate, with that of the shallow lake sediments. Water was extracted by elutriation. Samples were homogenised over sample intervals of 12cm. To a 50gsand sample 30ml of distilled deionised water was added to elute the solutes (for Cl and NO3). Supernatant solutions were recovered after 1 hour of agitation, and then after centrifugation. Sands were dried at 110o C and moisture contents reported on a wet weight basis. Deuterium was measured by mass spectrometry following direct reduction of the moist sand over zinc shot (Darling et al. 1989).

Estimations of groundwater recharge were carried out using the chloride mass balance (CMB) technique (Allison and Hughes 1978; Edmunds et al.1988). According to this simplified CMB approach the transport process for chloride is assumed to be 1-D, downward piston transport, and the Cl input is assumed to be constant at the surface. The age or residence time represented by Cl can be evaluated by dividing the cumulative total mass of chloride from the surface to that depth by the annual Cl input (Tyler et al.1996).

4 Results and Discussion

4.1. Recharge Estimation from Unsaturated Zone Profiles

The chloride results obtained by elutriation for each hole are presented in Fig. 2 as a function of depth together with the moisture content. The oscillations of chloride show that the recharge rate has not been constant. Chloride concentration has a mean value of 165 mg l-1 in the BB1 profile, with ranging from 59 mg l-1 to 581 mg l-1. The average chloride in SH2 profile is of 127 mg l-1, ranging from 26 mg l-1 to 643 mg l-1. Besides the high chloride value in the upper 1.5 m, there is also show the very high chloride concentration between depths of 6.7-9 m with a mean value of 225 mg l-1. At top 1.5 m, each of the profiles quickly reaches a maximum concentration of between 300 and 600 mg l-1. Such a peak concentration is typical in many chloride profiles in arid regions and indicates either a net infiltration rate that has changed over time, significant nonpiston flow, or a combination of these two processes.

 


Fig.2 Moisture content and chloride concentration for the unsaturated profile at the Badainmiao and Suhate sites

 

Values of Chloride in rainfall used are 1.5 mg l-1 and the mean annual rainfall of 89 mm. The use of constant chloride flux of 133.5 mg m-2 yr-1 in the chloride age calculations is a simplification of the total deposition corresponding to likely changes in environmental and climate conditions over the last 2000 years. The higher chloride value for BB1 (168 mg l-1) corresponds to a lower mean annual recharge of 0.81 mm yr-1, the record representing a residence time of 1025 years. For the area of Suhate the recharge rates are 1.05 mm yr-1, the records representing a residence time of 1660 years. The profile SH2 as archive of climatic change for periods up to 1660 years is illustrated in Fig. 3 after calibrated using recharge rate and moisture content according the model developed by Cook et al. (1992), subdividing the data according to intervals of high and low chloride. The record is 1660 years and climatic events of 10-20 years duration were preserved well in the profile. There are five markedly recharge stages occurred in500-550, 650-750, 850-980, 1280-1400, and 1500-1600 AD. The highest values of recharge rate of 5.0 mm occurred in the 690s AD, and then in the 530, 960 and 1350 AD occurred the higher recharge of 3.16mm, 2.47mm and 3.86 mm respectively, which obviously indicated the great climatic change. Afterward, the recharge rate decreases dramatically and had been lower than average value except for a few ages since the 19th century. The long-term low recharge rates between 980-1280 AD indicated the very drought episodes occurring in the “Medieval Warm period”.

4.2. Groundwater geochemical evolution

The TDS contents in the groundwater samples collected in the Heihe river ranged from 293 mg l-1 in piedmont fans of Zhangya to 6576 mg l-1 in the terminal lake of Ejina Basinwhile ranged from 317 mg l-1 inpiedmont fans of Wuwei to


Fig.3 The recharge history and the climatic record information from unsaturated zone profiles

 

8097 mg l-1 in the terminal lake of Minqin Basin in the Shiyang river. It is clear that the groundwater salinity increase from south to north in the direction of groundwater flow in both catchments. In Heihe river basin, the mean value of TDS from debouchure to Gaotai is 561 mg l-1, while 1235 mg l-1 TDS from Gaotai to Zhengyixia where is the boundary between middle reach and lower reach, in which the mean TDS is about 1960 mg l-1. In Shiyang river basin, the mean TDS from debouchure to Hongyashan Mountain is 601 mg l-1, while 1731 mg l-1 TDS in the Minqin Basin. Salinities higher than 1000 mg l-1 are found near the northern end of the Minqin basin, the influence of mixing with evaporated waters from the original lake basin. Higher than normal salinities in the central and upper parts of the basin are considered to be the result of some irrigation return waters. According to Todd’s classification of groundwater with TDS contents (1980), the groundwater in the middle reach belong to fresh water, while the lower reach is Brackish water both in Heihe River and ShiyangRiver basin.

The groundwaters feeding Baoriletigai lakes have a neutral pH and a total mineralisation in the range 720-1180 mg l-1 which is only slightly less than the lakes (1340-1790 mg l-1). This suggests that little evaporation is occurring and that these lakes are kept fresh by water continuously moving out of the system towards the north (implied regional flow direction). The total mineralisation at the nearby Lexiketu Lake is also similar but the absence of NO3 (and the presence of Fe of 1.94 mg l-1) suggests anoxic conditions. There is a dramatic contrast however between the two lakes less than 500 meters apart in the Badain Haizi region. The easterly lake is brackish with a pH value of 8.6 and TDS of 1800 mg l-1, while the westerly lake is highly alkaline (pH value of 10.6) and with total mineralisation of 400 g l-1. From other studies, the lakes in the southern Badain Jaran desert show a full range of salinities (Hofmann, 1996). The larger and deeper lakes belong to the hyposaline type (3-20 g l-1), whilst the small and shallow ones maybe subsaline (<3 g l-1) such as Baoritalegai and east Badanhaizi lakes. Small travertine islands may form where the artesian springs of freshwater exist. A sharp decrease of conductivity may be found near the lake bottom in some locations of a lake indicating groundwater inflows of lower salinity (Hofmann, 1996).

Hydrochemical coefficients, expressed in epm, show the relative concentration of various ions and can be used to indicate the predominance of a particular ion and to define locations of salt water intrusion. The mNa/Cl ratio in sea water is less than the unity(0.85), while groundwater has Na/Cl ratios great than the unity. There for, it is used for indicating areas suffering from salt water intrusion by fresh water. One significant characteristic of the groundwaters is the very low mK/Na ratio of 0.017, which would denote the dominance of albite over K-feldspars in the catchment area.

The mNa/Cl ratio in the aquifer of Heihe basin are variety between 0.64 to 3.85 and most of them are over 2, in the aquifer of Shiyang basin are variety between 0.59 to 3.3 and most of them are over 1.5. The mNa/Ca has a increasing trend along the groundwater flow from 1 to more than 8 in Heihe river and from 0.2 to more than 24 in Shiyang river. Both ratios indicating the reaction of silicate minerals and or some cation exchange is occurring at the expense of some cation. Saturation indices (Ma et al, 2005b) are just below saturation with calcite which underlines the mainly continental character of the aquifer and suggests the virtual absence of carbonate and with silicate weathering being the dominant process. The groundwaters are also close to dolomite saturation (mean value SI dolomite -0.38, range -2.32 to 0.40) and this reflects the mMg/Ca ratios close to 1, with Mg2+ dominant over Ca2+. This may be explained by the weathering of fresh, reactive mafic minerals in the sediments derived from diorites in the Qilian Shan source rocks.

4.3. Recharge environment and groundwater age

4.3.1. Stable isotopes

Precipitation data are available for several meteorological stations around the study area, including Lanzhou, Yinchuan, Lhasa and Zhangye, although chemical data are not generally available. Data taken from the IAEA network (IAEA GNIP database) from 1985 to 1996, were used and are compared to the global meteoric water line (GMWL). Values of deuterium (δ2H) and oxygen-18 (δ18O) in precipitation vary over a large range from –174‰ to 5‰ and 24.7‰ to 0‰ respectively, but the local line is quite similar to the GMWL with an equation of δ2H =7.38δ18O + 3.19 and an r2 of 0.97. The local line reflects only slight enrichment in moisture due to evaporation of the monsoon air mass. The weighted mean rainfall values at Zhangye station for seven non-consecutive years between 1986 and 1996 and all except one set of data plot on or close to the meteoric line with an overall mean composition of δ18O –6.51‰ and δ2H –43.9‰. (Fig. 4). This value, indicative of the heaviest rains, is most likely to be representative of present day local recharge.


Fig. 4 Stable isotope rations in the groundwaters of the Shiyang river, Heihe River and Baddain Jaran desert and surrounding area together with mean annual rainfall for Zhangye station

 

The groundwaters in Zhangye basin and Ejina of the Heihe river show a range from -40.6‰ to -62.‰ δ2H and –1.7‰ to –10.2‰ δ18O.The deep groundwaters in Zhangye basin has the same isotopic composition with that in Minqin. The δ2H and δ18O values of the deep groundwaters is significantly lighter than the shallow waters both in the Zhangye Basin and Ejina Basin, which were slightly enriched in the heavy isotopes and has an intercept on the Global Meteoric Water Line (WMWL) in the domain of modern rainfall (–7‰ δ18O). It is concluded that the deeper waters in both rivers are palaeowaters and that the shallower water with values between –7 and –11‰ δ18O represent mainly palaowaters mixed with limited modern recharge. The groundwaters in Minqin basin of the Shiyang river show a range from -31.6‰ to -73.8‰ δ2H and –7.14‰ to –11.54‰ δ18O. The deeper groundwaters (from depths of 200-320m) found mainly in the north of the basin generally have the lightest isotopic compositions of -9.38‰ to -10.65‰δ18O, but this is not always the case; lighter waters are found at two wells at 60-120m depth near the south of the basin. In addition one shallow well has a strongly depleted composition and another shows evaporation from an isotopically light source. The initial interpretation of these data is that there is likely to be a component of palaeowater-derived irrigation water. Some, but not all, of the heavier intermediate depth groundwaters (-7 to -8‰ δ18O) lie close to the main bifurcation of the Shiyang River around Minqin implying that the river source may also have been an important source of recharge in the past.

The isotopic composition of the lakes and the shallow groundwaters feeding the Baoriletigai and Lexiketu lakes in the Baddain Jaran desert show strong evaporation and are interrelated along a line that intercepts the local meteoric line at around -12‰ δ18O (Fig. 4). This implies that the groundwaters and the lakes in the desert are genetically related and that these are unrelated to modern recharge which has a weighted mean value lighter by some 5‰ δ18O. The lakes and inflowing groundwatres therefore originate as palaeowaters formed under a cooler climate and/or to a remote source rather than modern recharge. In this context it is noted that the weighted mean composition of modern high altitude precipitation at Lhasa is 13.7‰ δ18O. Whist the unsaturated zone moisture is strongly enriched isotopic signatures, intercepting the GWML around -6‰ δ18O and has an indicative of modern recharge.

4.3.2. Tritium

Tritium has been used in many groundwater studies to determine the residence time of groundwaters. Once 3H enters the subsurface as meteoric water, it becomes isolated from the influences of variable atmospheric 3H concentrations. Concentrations of 3H in the groundwater system depend primarily on the initial atmospheric concentration at the time of recharge and the radiogenic decay that has occurred since infiltration. In some cases, investigators have interpreted approximate groundwater ages from 3H concentrations. This requires that the initial precipitation 3H input record be known in order to interpreted quantitatively the groundwater age from the pattern of 3H concentrations along the groundwater flow path. In may regions of the world however, 3H precipitation data is not locally available. It is necessary in these case to extrapolate data from more distant locations.

The IAEA’s network of pluviometer in Zhangye is the only station in the north China that can provide the monthly precipitation 3H concentrations from 1983. However, detailed data are not available in these region before 1983. In order to infer what the historic disitribution of 3H has been in precipitation in the region, it is necessary to reconstruct the data. IAEA 3H data are abundant in the HongKong Station and Wei et al. (1980) has developed the detail models for reconstructing 3H concentration in precipitation according to the linear relationship in 3H with both increasing latitude and increasing distance from HongKong. In different latitude the3H has the different relationship with annual precipitation. Guan et al.(1986) has test the model with 3H records in several pluviometers. According to the model, the 3H data has been reconstructed from 1953 to 1982 inthe Zhangye station and used as the local 3H data(Fig.5). The reconstructed maximum peak was reaching 2100 TU in 1963which is lower than the tritium peak value of 3280TU in Vienna, but quite higher than the average tritium value of 600 TU in HongKong. The reconstructed 3H Value is about 110-150TU during the 1970s, Which has good relation with 3H record precipitation value of 110-180 TU observated by Wei et al. (1980) in the Mongolia Plateau. In order to make use of 3H as a groundwater tracer in the region to estimate approximate groundwater age, the precipitation 3H record must be modified to reflect the mass loss through radioactive decay. This can be done using a basic exponential decay equation using a half life for 3H of 12.43 years. The decayed precipitation record, which would represent 3H concentration in groundwater that had infiltrated between 1957 and 1996 is shown in Fig. 6. This groundwater 3H record now can be used assist in the interpretation of the regional groundwater flow system in the Mongolia Plateau.

However, age determinations accurate to the year are impossible and of no meaning to groundwater studies in the region according to the Fig.6. For example, a groundwater with a 3H value of 30-35 TU was possibly recharged by the precipitation in any of the 8 years: 1968-1971, 198619891992 and 1996. In addition, complicated mixing takes place in each aquifer, and the mode and extent of mixing of each year’s recharge with that of previous years recharge is unknown. However, semi quantitative dating is possible and very informative. Groundwaters with no more than 0.5 TU tritium has a pre-1952 age. Groundwater with significant tritium concentration (more than 10 TU) is of post 1957 age. In this case, only 5 years the left tritium may less than 20 TU. From the overall trend it is obviously concluded that groundwater with a post 1957 age should have the tritium concentration more than 20 TU. It is obvious that the majority groundwater collected in the baddain Jaran dasert and Heihei rever has the 3H values less than 10TU, and only 5 spring water sample in the Lonshoushan Mountain and western part of the Baddain Jaran Desert has the tritium value of 20-30TU, indicative of the pre-1957 ages. It is challenged that Chen JS et al (2003). have given an incomplete and misleading interpretation of the isotopic data combined with an absence of hydrogeological conceptualisation. Hence, the conclusion of modern deep groundwater flowing from the Qilian Mts to the badain Jaran desert is unsound. The spring flow in the Baddain Jaran desert recorded by Chen JS et al(2003) is likely to be maintained by palaeowater rather than modern snomelt. Radiocarbon data are required to confirm the hydrogeological model implied by isotopic data, but analogous results from a series study in Arid Northern China (Chen JS et al., 2003), the Minqin Basin(Ma et al., 2005a), the Ejina area(Wu et al.,2001) and Zhangye Basin(Zhang et al., 2005) confirm that the deeper groundwaters have late Pleistocene ages except where recharged by modern river runoff.

5 Conclusion

Chloride has the advantage of simple analysis and of being conserved during the recharge process so that a mass balance approach can be used. A mean value of


Fig.5 The reconstructed tritium value from 1953 to 1982 in the Zhangye station with a plot of observation data from 1983.

 

1.2 mm yr-1 for the three profiles based on chloride mass balance was estimated. Overall the groundwater quality is good for drinking and irrigation in both rivers. The TDS and the major ions in groundwater increases sharply down stream from the debouchure , more than 1 000 mg l-1 TDS in the lower reaches both in the Heihe and Shiyang rivers. Groundwater in the mountainous regions and the piedoment alluvial - diluvial fans is in HCO3- type with a low mineralization. The mineralization of groundwater in the basins in the lower reaches increases further and the type


Fig.6 The decayed precipitation record of 3H concentration in groundwater that had infiltrated between 1957 and 1996.

 

of Cl--SO42 - is dominant. The lakes in the Badain jaran desert are generally of fresh water, although with variable compositions which suggest some in-basin biogeochemical modification.

Stable isotopes (δ18O and δ2H) provide clear distinction between modern and ancient waters. A good rainfall database from nearby Zhangye provides definition of the composition of modern rainfall.  The signature of groundwater from the late Pleistocene differs markedly from that of the Holocene, shown clearly by the stable isotopic compositions (around -10.5‰ δ18O as compared with values of -7‰ at the present day). This primary difference is similar to that observed by Chen Z et al. 2003 for the North China Plain near Beijing. These authors also show a progressive enrichment in δ18O values from about 7 ka BP and an optimum around 6 to 4 ka BP equivalent to the global warming and a strengthened monsoon.

It is clear that tritium age determinations accurate to the year are impossible and of no meaning to groundwater studies in the region due to tritium measurements in the shallow unconfined groundwater environment could either be influenced by vapour exchange with the soil atmosphere or an expression very localised modern recharge. In addition, complicated mixing takes place in each aquifer, and the mode and extent of mixing of each year’s recharge with that of previous years recharge is unknown. A tritium value in the groundwater means multiple recharge ages in this region. It is challenged that Chen JS et al. (2004) given an incomplete and misleading interpretation of the tritium age.

Our results clearly show that modern direct recharge in this area is negligible. In addition it is apparent that the groundwaters being abstracted in the Minqin Basin, Ejina basin and feeding the lake system of the Badain Jaran are part of a regional flow network related to a wetter past climate as source of recharge. This system is largely decoupled from the river network in this area. The recharge source in the past and to a limited extent in the more arid conditions of the present day included the foothills of the mountains of the Tibetan plateau.

Acknowledgments  

The research is supported by the National Cooperation project (Grant No. 2002CB714004), the National Science Foundation of China (No. 40302031) and NSFC Innovation Team Project (No.40421101). This work also forms part of a wider UK-China collaboration and we acknowledge the support of the Royal Society through its link scheme (PEK/0992/306).