GU Wei-zu¹, SEILER Klaus-Peter², STICHLER Willibald², LU Jia-ju³

¹Hohai University, Nanjing, P.R. China

²GSF, National Research Centre, Neuherberg, Germany

³Nanjing Institute of Hydrology and Water Resources, Nanjing, P.R. China

Abstract:  Inner Mongolia is going to develop the Badain Jaran Shamo and the adjacent grass lands of the Gobi. In this effort it is essential to quantify natural as compared to man made desertification, especially groundwater gains and losses. With this information appropriate groundwater management strategies can be developed. - Groundwater of the Badain Jaran Shamo flows predominantly in fissured rocks from south to the north and has low flow velocities. - From meteorological stations around the Badain Jaran Shamo mean precipitation has been extrapolated to 70-100mm/year; in contrast pan evaporation accounts in the warm season for about 4000mm/year and declines significantly during the cold season; therefore, as an average it is assumed that yearly actual evaporation from lakes and wetlands exceeds about 30 times precipitation. Since the high evaporation areas of the Badain Jaran Shamo are in the south-east and groundwater slopes from south to north, a simple water balance shows that local groundwater recharge cannot cover all evaporative water losses. - The evaluation of stable isotopes of the water molecule shows that most groundwater has an evaporation signature from lakes and the small wetlands; however, there are also – albeit scarce - recharge contributions without evaporation signature. - Lake levels vary according to seasonal variations of evaporation; any groundwater level variation related to the very small groundwater recharge is considered negligible, because of the high porosity of aeolian sands. Hence, in summer strong evaporation losses exceed groundwater inflow to the lakes, decline the lake and groundwater table and lowers by isotope fractionation the deuterium excess of lake water; in contrast, during winter the groundwater/lake level is rising, because of a stronger non-evaporated groundwater inflow than evaporative water consumption from lakes, and therefore the deuterium excess of lake water increases in winter time. - Stable isotope variations at constant deuterium excess occur in springs and dug wells close to lakes; these variations have been evaluated as compared to the respective variations in precipitation to account a mean residence time of 6years. - All lakes have tritiated water (maximum 20TU), because of a locally high tritium level in atmospheric moisture, and have high DOC contents, >>2mg/L C; in contrast groundwater, which is not influenced by lake water has low tritium (1<TU<5) and DOC (<<2mg/L C) contents; this should be considered in applying the Cl-groundwater-balance method for groundwater recharge studies. Groundwater recharge according to the Cl-input (about 2 mg/L) and output amounts to 1.5 to 1 mm/year. ¹⁴C dating leads to water ages from present to about 30,000years at shallow depth; hence groundwater flow has an actual and fossil recharge component; actual recharge forms a thin groundwater cover above historic and fossil recharge. The whole system belongs basically to a transient discharge system since historic and geologic time scales. Taking old shore lines of lakes above the actual sea levels and ¹⁴C dating of it results in a decline of about 1mm/year over the last 10,000years; over the run of 30,000years, this decline can be approximated as exponential and thus allows an extrapolation to any near future development.

Key words: Groundwater recharge, environmental tracers, groundwater dynamics, groundwater dating

 

1       The task

Today’s China has a desert area of about 370,000km², accounting for about 8.5% of its territory, and extending during historical and geological times by desertification. As in many other areas of the world, desertification has a geologic/historic, non-man-made background and became accelerated by men’s activities since the mid of the 19^th century. Desertification is documented by land drying, vegetation changes, wind erosion, sand accumulation, and soil and water salinization. All these phenomena threaten people’s local economy, health and safety and make a region still more vulnerable to draughts, floods, wind and sand storm disasters as it already is.

In between the many aspects of desertification, our project concentrates

l         To the Badain Jaran Shamo area and the adjacent grasslands (Fig. 1), which already has been studied by Chinese and international research groups (e.g. YU et al. 1962, ZHOU et al. 1986, DAVIS 1986, WALKER et al. 1987, GHEY et al. 1996, JÄKEL 1996, HOFMAN 1999), hence has already a good, but still non-sufficient data base to develop groundwater management strategies, and

l         The natural desertification processes in these two areas during the past millennia and the actual and future consequences for the water resources.

Fig. 1 The Badain Jaran Shamo in Inner Mongolia with the Gurinai and Ongt Gol grasslands in the north-west respectively north. The red circle indicates the research area of the years 2004/2005, the white arrow a general groundwater flow direction beneath the sand sea.

The present stage of the project intends to

l         To contribute to the actual groundwater balance,

l         To account for any groundwater inflow to the desert and grassland areas from neighbouring regions,

l         To study the decline of the water resources in the Badain Jaran Shamo during the past millennia and

l         To extrapolate such natural developments to the near future.

Special instruments to be used are space and time integrating methods in combination with traditional hydro-geologic investigations.

l         On the small scale field works provide mostly instantaneous results;

l         On the large scale satellite imaging delivers space and environmental tracer methods space and time integrating results.

Satellite imaging provides data over the last 50 years, environmental tracer methods data over the past millennia till recent years. In this research it is a special favour that groundwater chemistry in the Badain Jaran Shamo is not yet influenced by human contamination sources, except some radioactive environmental isotopes, which may have the signature of an actual air contamination, because the Badain Jaran Shamo and adjacent areas are a nuclear exercise area of the PR China. In between these radioactive isotopes Tritium with a half life of only 12.3years may be considered a valuable tracer, because it allows under special boundary conditions to qualitatively recognize actual groundwater recharge; the disadvantage of the radioactive air pollution is that ³H can’t be used properly to quantified groundwater recharge (SEILER & GAT 2007), because the input function is unknown, hence any comparison with the output function is impossible.

2       Geologic and Hydrogeological Settings

2.1  Geologic settings

The Badain Jaran Shamo is situated between E 100° to E 106° and N 40° to N 42° (Fig. 1) and reaches altitudes of 1,200 and 2,000 m a.s.l.. In its S and mostly SE part it has many small lakes. The regional geologic and tectonic setting of this area was detailed by HOFMANN (1999). The characteristics of the Badain Jaran Shamo are low and huge sand mountains and exceptionally scarce outcrops of any bed rocks (JÄCKEL 1996). Therefore, little is known about the pre-existing morphology and the bed rock geology of the Badain Jaran Shamo, covered both by aeolian sands. However, bathymetry of lakes, microbiolite islands and springs in lakes make sure that many lakes are in close contact with consolidated fissured rocks.

1.       Recent bathymetric measurements in the Lake Noertu (Fig. 2) and other deep lakes suggests that the lake underground is build up by consolidated rocks, because the slopes of the lake floors are often too steep (>>36°) for unconsolidated rocks.

2.       In some lakes carbonate islands (microbiolites) occur, which have been described by ARP et al. (1996). On top of these islands springs may discharge (Fig. 3), indicating a hydraulic head of these springs above the lake level; such springs can only refer to groundwater in wide fissures of consolidated rocks (SEILER 1968).

3.        In other lakes, springs discharge like little fountains into the shallow part of the lake; this also indicates a hydraulic head above the lake level, hence, groundwater flow in fissured rocks (Fig. 4).

Mother materials of the sand mountains are preferentially silicate rocks because the sand consist basically from quartz and silicates with very few or even missing carbonate rocks. The large grain sizes of Aeolian sands (0.2 to 0.6mm) in the sand mountains further suggest that the source and sink areas of sands cannot be distant.

Since the geology of the Badain Jaran Shamo is not very well known (WALKER et al. 1987, ZHENG et al. 1982), no precise answers can be given on the question if the sands are covering a mountain or are sitting as huge dunes on a flat and perhaps weathered old landscape. It was suggested that the sands are positioned as dunes on clays and that these clays act as an outflow horizon for dune infiltration (JÄCKEL 1996, HOFMANN 1999); as compared to the water balance, however, this interpretation seems to be questionable.

2.2  Groundwater table and subsurface water flow in the Badain Jaran Shamo

Groundwater levels have been measured during the 2002, 2004 and 2005 excursions using GPS; although GPS measurements of altitudes have no high precision and the groundwater surface is always uneven in hard rock aquifers, these results allow nevertheless some basic statements: Yearly repetition of measurements differ by less than 2m, groundwater flow is directed from south to north (Fig. 5) or south-east to north-west and in the investigated south-east of the Badain Jaran Shamo, the border area, the water table slopes steeper than in the central part of the desert. The latter observation could be interpreted in terms of individual inflow areas of fresh groundwater into the desert, which widen like a delta, thus producing the flat groundwater table in the inner part of the Badain Jaran Shamo; in other words groundwater flows from small to huge cross-sections on the way from south to the north. If this interpretation can be approved by more precise and extended measurements, it resulted in two statements:

l         Groundwater flow velocities in the south-east must be higher than in the central Badain Jaran Shamo and

l         Groundwater residence times in lakes and, hence, mineralization by evaporation are higher respectively stronger in the inner part of the desert than in the south-east.

Figure 2   Bathymetry of the Noertu lake in the Badain Jaran Shamo (September 2005;

measurements by EarthWatch volunteers)

       

Fig. 3 A spring on top of a carbonate island in the Lake Noertu, Badain Jaran Shamo. The hydraulic head of this spring lies above the lake level. Left photo: the red arrow indicates the position of the spring; right photo the artesian outflow of the spring on top of the island

 

Fig. 4 Three springs discharging like a fountain into the shallow part of the Lake Indertu, Badain Jaran Shamo. The hydraulic head of this spring lies above the lake level

Fig.5 Groundwater counter map from GPS water level measurement during the years 2002, 2004 and 2005

The expeditions to the Badain Jaran Shamo made clear that lake levels obviously undergo variations as indicated during summer by the missing vegetation along the actual littoral zone of the lakes. Since lake water is uncovered groundwater, such variations can be attributed either to seasonal variations of the intensity of evaporation (E) or groundwater recharge variations. Precipitation accounts to about 70-100 mm/year, summer evaporation amounts 4000mm/year and the sand cover of the fissured aquifer has high porosities; therefore the little groundwater recharge (70-Eactuel or 100-Eactuel) cannot produce significant groundwater table fluctuations. These variations are most probably due to seasonal changes of the intensity of lake evaporation as compared to the small amounts of groundwater inflow to the lakes. Lakes act in the warm season as a complete and in the cold season as a partial sink for groundwater fluxes; therefore, monthly observations of stable isotopes of the water in the lake Cahachulutu (Fig. 6) show that the deuterium excess declines in the summer and increases in the winter season, because in summer evaporative water losses exceed groundwater inflow and in winter groundwater inflow exceeds evaporation losses and thus dilute the summer deuterium excess.

Fig. 6 Stable isotope variations in the Cahuchulutu lake over the run of a year

Groundwater level fluctuations have not yet been investigated in the study area. In contrast stable isotope variations have been observed during one year in the groundwater close to the Shuqui Lake (Fig. 7). The deuterium excess is constant all over the year but the stable isotope content varies. This variation as compared to the respective stable isotope variations in precipitation can be used to approximate the mean residence time (t’):

 

For the Shuqui groundwater a mean residence time of 6.7years results, which is quite short as compared to the relief energy, characterizing the study area; this can, therefore, only be interpreted as a high portion of preferential fluxes contributing in the unsaturated zone of sands to groundwater recharge (SEILER et al. 2002).

2.3  Chemical and isotope characteristics of Badain Jaran Shamo Groundwater

Groundwater in the Badain Jaran Shamo belongs either to fresh or saline water:

 

 

Fig.7 Stable isotope variations during the year in the groundwater close to the Shuqui Lake.

l         If it did not undergo evaporation at open water surfaces, it is fresh and has a pH close to neutral (pH 7.5) or

l         If it underwent serious evaporation in lakes pH rises to 9 or 10 and lake water appears saline.

Main chemical constituents of waters from the Badain Jaran Shamo are sodium, chloride and sulphate Tab. 1. The most surprising result, however, are high DOC concentrations in groundwater although there exist no soil cover. The high DOC concentrations (>2mg/L C) are attributed to the high number of organisms in all Badain Jaran Shamo lakes; this allows to distinguish between groundwater in contact with lakes (DOC>>2mg/L C) and groundwater without (DOC<<2 mg/L C) and thus to apply also the chloride groundwater balance method to calculate groundwater recharge from areal infiltration to groundwater. As shown by the short mean residence time of groundwater from infiltration to detection time, preferential flow seams to be an important factor to protect infiltrates from evaporation. This can be confirmed by stable isotopes (Fig. 9) and by observations of the water content/saturation in sand dunes, which is even high close to the land surface (Fig. 8).

With the actual analytic results on groundwater chlorides and supposing a Cl-input concentration with precipitation of 1mg/L and a mean precipitation amount of 70-100mm/year resulted in 1 to 1.5mm/year of groundwater recharge, which is close to the experience of many desert areas the world around (SEILER & GAT 2007).

Nearly all groundwater have either an evaporation signature (Fig. 9) from the land surface or represent a mixture between evaporated lake and non-evaporated infiltration water. As compared to precipitation, groundwater recharge occurs in all seasons, however, winter seams to be the main recharge season; this recharge is probably linked to snow melt, which contributes more efficient to ground water recharge than rain infiltration.

Only few samples have a measurable tritium content (Fig. 9), because either present groundwater recharge is focussed to a thin layer close to the groundwater surface or most samples belong to historic (<10,000years) or fossil (>10,000years) groundwater recharge.

Table 1 Results of chemical analysis from lakes (L), springs (S), shallow drills (sd) and dug wells (dw) in the    

           Badain Jaran Shamo. n.d. = not detected. Brown = exceptionally high values, green = normal values

Place/ date	As µg/L	Na+ mg/L	K+ mg/L	Mg++ mg/L	Ca++ mg/L	NH4+ mg/L	NO2- mg/L	NO3- mg/L	F- mg/L	Cl- mg/L	SO4-- mg/L	DOC mgC/L
Detection limit	5 µg/L	0.5 mg/L	0.25 mg/L	1 mg/L	5 mg/L	0.05 mg/L	0.5 mg/L	1 mg/L	0.1 mg/L	2.5 mg/L	5 mg/L	0.5 mg/L
NRT(S) 02-09-25	911	84451	4100	133	3040	n.d.	n.d.	n.d.	n.d.	73978	11405	6.8
NRT 02-09-25	n.d.	266	59.0	22.5	33.0	n.d.	n.d.	14.2	6.7	137	164	2.4
HGN 01-09-25	n.d.	137	20.3	67.8	26.6	0.39	0.50	n.d.	3.6	104	21.1	2.8
ZEGT(S) 00-09-26	n.d.	608	16.9	15.2	7.28	n.d.	n.d.	18.8	n.d.	337	243	2.8
HBT 02-09-29	n.d.	427	48.1	133	91.2	n.d.	1.70	n.d.	n.d.	245	407	19.2
Lutu (sd) 02-09-27	n.d.	574	43.2	115	78.4	n.d.	n.d.	n.d.	n.d.	n.d.	377	3.1
Lutu  (IR) 02-09-27	n.d.	150	1.9	129	34.2	n.d.	n.d.	n.d.	n.d.	100	59.0	3.3
Lutu  (S) 02-09-27	n.d.	400	26.4	89.9	59.6	n.d.	n.d.	n.d.	n.d.	265	280	5.4
Lutu  (L) 02-09-27	n.d.	196	18.1	53.8	51.8	n.d.	n.d.	n.d.	n.d.	149	123	2.8
Dw A												15.8
SGIL(S) 00-09-26	n.d.	276	12.2	n.d.	12.5	n.d.	n.d.	8.40	1.2	88.3	153	0.7
SGIL(DW)00-09-26	n.d.	102	12.3	n.d.	n.d.	n.d.	n.d.	6.49	1.7	69.8	111	1.2
Indertu (holy Spring) 22-10-04												<0.2
Luerto 27-09-04												<0.1
Spring 29-09-04												<0.2

 

¹⁴C analysis on water samples from springs and dug wells gave water ages from present to 30,000years. The very old, shallow groundwater is not supposed to belong to sand dune recharge, but to long and shallow flow paths from outside the desert.

2.4  Long term groundwater level decline

In the Badain Jaran Shamo old shore lines have been found, occurring about 40m and less above the actual lake levels (HOFMANN 1999). These shore lines contain organic matter, carbonate shells and carbonate precipitates, which have been dated with 14C.  Snails from these high positioned shore lines (e.g. Planorbis) indicate further former lake salinities, which were lower than actual. The correlation of these former altitudes and the respective ages of shore lines indicates in a first approximation an exponential decline of groundwater/lake water tables. This correlation must be extended to investigate if this decline was continuous and refers to a hydraulically buffered system (transient system) or if the known climate changes at the end of the last ice age and the climate optimum about 5,000 years ago makes this decline discontinuous.

 

 

Fig. 8 Water saturation in aeolian sands of the Badain Jaran Shamo as a function of depth beneath ground

 

 

Fig. 9 ²H/¹⁸O diagram of water samples from the Badain Jaran Shamo. Some springs group along the local meteoric water line (blue), some along an evaporation line (red), starting from δ¹⁸O=-11.9%o and δ²H=-83.2%o on the LMWL

3       Conclusion

Main aquifers in the Badain Jaran Shamo are fissured, most probably crystalline rocks; hard rock outcrops in this area are scarce. The large grain sizes of dune sands and the conservation of many stone-age tools stand for short aeolian transport distances of sands from the source and the sink area.

Groundwater flows from adjacent areas in the south to the Badain Jaran Shamo; it is actually also recharge in the desert by about 1 to 1.5mm/year. 14C-data as compared to 3H-data indicate that actual groundwater recharge forms a very thin cover over old groundwater (1,000 to 30,000years). Both cross in north and north-west direction the Badain Jaran Shamo to reach the grassland. The infiltration process of actual groundwater recharge seems to be mostly produced by preferential flow, because mean residence time calculated by comparing stable isotope variations in groundwater as compared to precipitation yields residence times of only 6 years in an area with high relief energy.

Groundwater in contact with lakes is traced by high DOC contents from the huge number of organisms in the lakes; this DOC signature allows distinguishing between groundwater with and without lake contact, being important in applying the Cl groundwater balance method. Stable isotopes indicate that most groundwater is evaporated, following one individual evaporation line, which starts from winter precipitation; however, there are also summer infiltration events of non-evaporated water.

Acknowledgement

The authors wish to express their high appreciation to the EarthWatch Institute in Boston that supported this research by sending volonteers for our field campaigns.

 

References

[1]      ARP, G., HOFMANN, J. & REITNER, J. (1998): Microbial fabric formation in spring mounts (Microbialites) of alkaline salt lakes, Badain Jaran Sand Sea, PR China. – Palaios 13, 581-592

[2]      GEYH M A, GU WEI-ZU, JAECKEL D. (1996): Groundwater recharge study in the Gobi desert. – China J. Geosciences 7/8, 279-280

[3]      HOFMANN, J. (1999): Geoökologische Untersuchungen der Gewässer im Südosten der Badain Jaran Wüste (Aut. Region Innere Mongolei/VR China) – Status und spätquartäre Gewässerentwicklung. – Berl.Geogr. Abh. 64, 5-247

[4]      JÄKEL, D. (1996): The Badain Jaran desert: Its origin and development. – Geowissensch. 14, 272-274

[5]      SEILER, K.-P. (1968): Kluft- und Porenwasser im Mittleren Buntsandstein des Südlichen Saarlands. – Geol. Mitt. 9, 75-96

[6]      SEILER, K.-P. & LINDNER, W. (1995): Near surface and deep groundwater. - J. Hydrol. 165, 33-44

[7]      SEILER, K.-P. & GAT, J.R. (in preparation): Groundwater recharge from run-off, infiltration and percolation. - Springer

[8]      SEILER, K.-P., LOEWENSTERN v., S., SCHNEIDER, S. (2002): Matrix and bypass-flow in quaternary and tertiary sediments of agricultural areas in south Germany. - Geoderma 105, 299-306

[9]      WALKER, A.S., ALSEN, J.W. & BAGEN (1987): The Badain Jaran desert: remotte sensing investigations. – The Geogr. J. 153, 205-210

[10]  YU, S., LI, B., CAI, W.Q., & TAN, J.N. (1962): A survey of the Gobi of Western Inner Mongolia and Badain Jaran Desert. – Nation. Techn. Information Service Pub. JPRS 19993, 244-276

[11]  ZHENG, Z. et al. (1982): Stratigraphischer Überblick über das Badain Jaran Gebiet, Innere Mongolei. – J. Stratigr. 6, 225-230

[12]  ZHU, Z., LIU, S., WU, Z. & Di, X. (1986): Deserts in China. – Inst. Desert Research, Lanzhou, 1-131