1000 m

V 1_


35 0


Q = 35 1

Fig. 3. Flow (arrows) and age distribution (broken lines) in an aquifer system with hydraulic conductivities and relative discharges (b, left and right), with (a) and without (c) groundwater abstraction from deep groundwater.

Badain Jaran Shamo doubtless indicates that the lake waters represent uncovered groundwater, flowing through the lakes with low flow velocities. The present lakes in the Badain Jaran Shamo are residuals of more extended lakes in the historic and geological past. This is deduced from shores 40 m to 50 m above the actual lake levels; the shores are traced by Palaeolithic artefacts, carbonate precipitates from plants growing in shallow water, fossil freshwater snails and organic lake sediments. Carbonates have been dated by 14C to 15 000 years (Hofmann 1999). These relicts and ages show the decay of a groundwater resource beneath the Badain Jaran Shamo, emptying

Fig. 4. Response of deep groundwater on exploitation of 20% of groundwater recharge over a run of 6.3 (left) and of 5% of groundwater recharge until reaching steady-state conditions (right). Lines represent water ages in years.

continuously or discontinuously over millennia. Any discontinuous decay can be attributed to climate changes as well as to changes in the size of the subsurface catchment, feeding the discharge; the latter are typical when the subsurface flow system changes, e.g. from exorheic to endorheic.

It is interesting to note that the lakes in the Badain Jaran Shamo have old groundwater without tritium, but most shallow waters from dug wells contain significant amounts of tritium (Tables 3 and 4). This finding may be considered as a qualitative confirmation of results of scenario modelling. Here present groundwater recharge exists, but is of patchy pattern and appears only as a thin cover over groundwater without tritium. As indicated by the high water contents (>2 vol%) at shallow depth (50 cm) below the ground surface found everywhere, and because of the often steep dune slopes, this groundwater recharge occurs through the unsaturated zone, reaches local freshwater springs and discharges to the lakes.

A first estimate of present groundwater recharge in the Badain Jaran Shamo, applying chloride balances, results in 0.5 to 1 mm/a. This low recharge

Table 3. Stable isotope measurements and tritium content in groundwater samples of the Badain Jaran Shamo


818O (%<>)

82H (%<,)


3H (TU)

Free of tritium





0.7 + 0.7




- 25.1

1.1 + 1.1

Containing tritium



- 20.1


1.5 + 0.7

LT (S)



- 76.1

1.6 + 0.7




- 18.6

2.9 + 0.7



- 49.5

- 25.4

4.6 + 0.7



- 2.0

- 66.0

4.7 + 0.7



- 62.2

- 20.5

11.4 + 0.9



- 23.6


16.1 + 1.2

Sampling carried out in September 2002. DW, dug well; IR, irrigation dug well; S, spring.

Sampling carried out in September 2002. DW, dug well; IR, irrigation dug well; S, spring.

Table 4. Carbon-14, d13C and tritium in groundwater samples of the Badain Jaran Shamo


C14 Activity pMC (corrected)

D14C (%<>)

14C Age yrBP


3H (TU)


98.76 + 0.33

-18.3 + 3.3

100 + 25

-10.64 + 0.19



88.35 + 0.22

-121.2 + 2.2

995 + 20

- 11.21 + 0.07

0.7 + 0.7


80.54 + 0.37

-199.6 + 3.7

1740 + 35

-12.82 + 0.24

0.9 + 0.7


31.25 + 0.27

-689.3 + 2.7

9340 + 70

-10.36 + 0.25



23.17 + 0.17

-769.7 + 1.7

11 750 + 60

-4.08 + 0.13

0.8 + 0.7

Sampling carried out in September 2002. DW, dug well; S, spring.

Sampling carried out in September 2002. DW, dug well; S, spring.

rate together with slow shallow groundwater fluxes from remote areas, however, cannot compensate the evaporation losses from lakes over the last 15 000 years, hence a groundwater decline of about 1 mm/a results.

It is well known that subsurface reservoirs with low groundwater recharge rates have a long transient response through the decay of groundwater mounds that have been built up by past groundwater recharge. As the outflow of a reservoir is also associated with enlargement of the size of the catchment area, by pulling down former subsurface water divides and thus changing groundwater flow directions, transient outflow lasts longer and even becomes discontinuous. Hence groundwater flow or the decline of groundwater heads is oscillating on the one hand, but follows a general trend of decay on the other hand.

Description of transient groundwater flow

In a first approximation, the outflow (Qt) of a subsurface reservoir with fixed boundary conditions and missing groundwater recharge can be characterized by the Maillet function:

where Q0 = initial discharge, T = mean turn-over time, t = time variable. The MTT (T) in this equation can be determined by means of 14C in dissolved inorganic or dissolved organic carbon. As can be seen from Figure 5, a 100% discharge approaches slowly or quickly to 37% according to a high (20 000 years) or low MTT (5000 years). In terms of shallow and deep groundwater this result shows that although both belong to a hydraulic continuum, young groundwater reaches steady-state conditions much more quickly than deep groundwater, when hydrodynamic boundary conditions change. With regard to sustainable groundwater exploitation, i.e. beneficial for several generations, Figure 6 further shows that under natural boundary conditions hydraulic heads or discharge declined according to low or high

MTTs within 200 years of observation and without further groundwater recharge by 85% (MTT = 100 years), 30% (MTT = 500 years) or by only 3% (MTT = 5000 years). This transient behaviour of groundwater systems in dry lands or deep aquifers is mostly neglected when developing management strategies for groundwater extraction under changing input or emergency conditions.

Typical examples of such transient groundwater behaviour are known from the Arabian peninsula (Verhagen et al. 1987) and north Africa (Sonntag 1985) and play a role in groundwater resources adjacent to the Jordan graben (Dead Sea) (Salameh & Udluft 1984) as well as in the Molasse basin in south Germany (Lemcke 1976). In the two latter cases the decline of pressure heads were remobilized during geological time periods by entrapped hydrocarbons, which now appear as asphalt in terminal lakes (e.g. Dead Sea), or as methane or high sulphur contents in springs along the regional discharge base in waters from two deep wells in the Molasse basin, south Germany (Lemcke 1981).

Water balance studies mostly refer to the meteorological water balance:

Years before present

Fig. 5. Outflow of a spring according to an exponential age distribution in groundwater. Mean turn-over-times of 20 000, 10 000 and 5000 years.

Years before present

Fig. 5. Outflow of a spring according to an exponential age distribution in groundwater. Mean turn-over-times of 20 000, 10 000 and 5000 years.

Fig. 6. Discharge or hydraulic head changes in groundwater with different mean turn-over-times.


Fig. 6. Discharge or hydraulic head changes in groundwater with different mean turn-over-times.

Precipitation (P) and discharge (D) are measured over many years and discharge is further analysed by hydrograph or chemical/ isotope separation methods. Potential evapotranspiration (ET) is estimated using meteorological data sets and reducing it to actual ET according to empirical transformations. Stored water (AS) is generally considered as belonging to the percolation zone; it includes, however, water stored over long periods of time in emptying parts of aquifers, and in transient systems contributes much more to the water balance than the small water in the percolation zone. Hence water balances over-estimate groundwater recharge in decaying, and underestimate it in filling up systems.

As a consequence, from water balances based on Equation 2 it becomes evident that not only is a representative number of years necessary to neglect any storage in the subsurface, but it should also be ascertained that the discharge has short MTTs, to reliably calculate actual ground-water recharge. If the turn-over time is high (>100 years), the present discharge does not represent actual groundwater recharge and vice versa.

As an example, the Azraq springs in the Jordan Badia region discharged about 1.2 m3/s before systematic groundwater abstraction by tube-wells started close to the spring area; today the springs do not exist any more. The springs drained an oro-graphic catchment area of 12 700 km2; for this catchment, with a discharge of 1.2 m3/s, a recharge rate of 0.1 l/(s km2) or about 3.2 mm/a is calculated. However, this calculation is actually incorrect because the water age is about 5000 years and the contribution of actual recharge was calculated by other methods to be less than 1 mm/a. This interpretation is supported by tritium analysis of water samples from shallow wells. Only a few samples out of dug wells close to wadis yielded detectable 3H concentrations or water ages of less than 100 years; all other wells were free of tritium except recently-recharged groundwater.

Considering that discharge from the Azraq springs was three times higher 5000 years ago

(Fig. 4), the same size of catchment resulted in a recharge rate of 0.3 l/(s km2) or 10.6 mm/a for the past. A smaller size of the catchment would shift the groundwater recharge rate to an even higher number, close to the groundwater recharge under present semi-arid climate conditions, that perhaps prevailed 5000 years ago in this area.

All these examples make clear that it is quite difficult to recognize any reduction of groundwater recharge rates in semi-arid or arid areas by immediate changes of discharges, hydraulic heads or environmental isotope methods. Such changes are therefore mostly disregarded in developing exploration and exploitation strategies for these areas.

Since it is often difficult to gain reliable values on safe yield in drylands, spring discharges or well yields are mostly taken as a base for developing management strategies. However, this always leads to an over-estimate of real water availability, because hydraulic aquifer properties have no direct link to groundwater regeneration and transient systems give misleading information.

Transient flow and groundwater management

Transient groundwater flow is characteristic of low-recharge, mostly dry areas, which underwent in the past millennia more serious climate changes than other areas. Because of the small natural recharge rates, small input changes over long periods of time are finally expressed in strong responses. A detailed consideration of this transient flow, however, leads to a special aspect of groundwater management for small consumer units per square kilometre.

Transient groundwater flow in dry lands ends in oceans or terminal lakes or appears in through-flow lakes and springs. There, it is completely or partially lost by evaporation, or it becomes saline and creates groundwater contamination downstream of the lakes; hence such water losses or degradations are unproductive.

Therefore, apart from water balance studies, to better management of these unproductive, natural water losses for human or ecosystem use could also be considered. In this exercise, however, one principle must be followed: avoid groundwater over-exploitation in excess of natural losses. To reach this goal, the groundwater surface should either be removed from the influence of evaporation, thus be kept a few metres below the ground surface, or groundwater should be used before it becomes exposed to evaporation. However, ground-water exploitation should always take account of the exponential decline of groundwater discharges or hydraulic heads as characterized by the analysis of discharge or hydraulic head recession.

Approximating, for example, the lake surfaces of the Badain Jaran Shamo to about 27 km2 and the potential evaporation from lakes as 3000 mm/a leads to unproductive water losses of 81 000 000 m3/year; this is a minimum number because it does not include evaporation losses from groundwater resources close to the land surface.

Avoiding and using unproductive evaporation losses, indeed, is insignificant for water supply; however, it is sufficient for small communities in widespread, water-scarce areas.

0 0

Post a comment