Baharia, Kharga, Dakhla, and Farafra (Siwa seems hitherto generally to
have been left out of consideration) has been a much-discussed question.
Some geologists, myself amongst them, have always regarded the water as
being derived from rainfall in the western Sudan, flowing underground in
permeable beds towards the Mediterranean. Others have held the view that
the oasis waters are merely Nile water which has penetrated more or less
laterally into the adjoining deserts. The arguments that have been urged
in support of the former view are, firstly, the high temperature of the
water in many of the oasis wells; and, secondly, that the levels of the
springs and wells are often much higher than those of the Nile in the
same latitudes. To these arguments it has been justly replied that
neither of them is conclusive; the high temperature of the outflowing
water merely testifies that it has descended to considerable depths at
some part of its underground path not very remote from the point of
outflow, but really tells us nothing as to its place of origin; and the
high level of the springs in Baharia, for instance, as compared with
that of the Nile in the same latitude, might be accounted for by the
seepage from the Nile taking place fairly high up in the river’s course.
There the question remained until 1925, when I was able to visit and
determine the positions and levels of a number of water-sources farther
in the interior than any of those on which the “Nile” argument was
based. Amongst other level-determinations, I ascertained that the Sheb
well is 228 metres above sea-level, and that Merga Lake, lying far to
the south-west (in lat. 19° 3′, long. 26° 18′), is at an altitude of no
less than 509 metres above the sea. Shortly before my tour in the Sudan,
Hassanein Bey had confirmed Rohlfs’ level of 400 metres for the Kufra
water-sources, and I had found that of Abu Mungar (north-west of Dakhla
Oasis) to be 117 metres. At all these places the water-supplies are
derived from underground sources in the same rock—namely, the Nubian
sandstone, which covers such vast areas in the Sudan and Egypt.
I had thus four well-determined natural water-levels at the corners of a
great quadrilateral whose sides averaged over 500 kilometres in length,
and embraced more than 20 square degrees of the Earth’s surface. Now
just as in solid geometry the levels of any three points determine the
inclination of an oblique plane to the horizontal, so on the Earth any
three levels will determine the inclination of a surface to the geoid
(of course assuming both geoid and surface to have the same curvature).
And on making the calculation, I found that I obtained practically the
same degree and direction of inclination for the natural water-surface,
whichever three of the four known points I utilized for the calculation.
In other words, I found that if I took, say, the levels of Kufra, Abu
Mungar, and Sheb, and deduced from them the inclination of the water-
surface to the horizontal, I could _calculate_ the level of Merga pretty
exactly. Extending the trial, I found that I could do the same with a
fairly close approximation for the other wells in the Sheb
neighbourhood, and also for wells in the oases of Dakhla and Kharga. The
conclusion seemed irresistible that all the wells considered were fed
from a continuous sheet of underground water; and it was evident that
this water did not come from the Nile, firstly because of the high level
of Merga, which is above that of the swamps of the Bahr el Ghazal and
other western feeders of the White Nile, and secondly because of the
direction of the downward slope of the underground static water-surface,
which is from the south-west, instead of from the south as it would have
been had the water been derived from the Nile in the Bahr el Ghazal
region. The true source of the water must be somewhere more or less
nearly on a line drawn south-west from Dakhla, for this is the direction
of upward slope of the underground static water-surface; and if such a
line be drawn on a map of Africa, it will be found to lead towards the
Erdi and Ennedi region, on the borders of the Chad basin. It is in the
highlands of Eastern Erdi and Ennedi, therefore, that we must look for
the source of the artesian water of the Egyptian oases. What is known of
this region from the recent explorations of Colonel Tilho lends good
support to our conclusion.[17] It is a bare and rugged sandstone
country, where, in spite of a rainfall which is by no means negligible,
permanent water-sources are scanty, and where, in consequence, there
must be a considerable absorption of moisture by the rocks; and it lies
at so high an altitude as to give sufficient “head” for the absorbed
water to percolate through the porous sandstones and thus to reach
Egypt.
Being convinced that I had at last arrived at the true origin of the
artesian water, I next began to entertain the idea of attempting to make
a map which would show the contours of the underground water-sheet, and
from which, in conjunction with the contour-map of the surface which I
had already prepared, I might be able to predict the depth of boring
required to tap the underground water at any point in the desert. But a
little consideration showed that this idea was an impracticable one, by
reason of our ignorance of the underground geological structure over the
greater part of the desert. The underground water would naturally pass
along permeable sandstone beds, often confined between impermeable clays
above and below. And although the general structure of the Libyan Desert
is doubtless one of simplicity as compared with that of other parts of
Egypt, yet we know, from observations in the oases and in the Owenat
region, that the beds are in some places folded and faulted, and that in
others they have been uplifted and entirely removed by denudation, with
the exposure of large areas of the underlying ancient crystalline rocks.
The only parts of the desert for which the boring-depths could safely be
predicted would be certain small areas within the oases, where wells
have been sunk in sufficient numbers to give us definite information as
to the local underground structure; and underground-water maps of these
small areas, though they might usefully systematize our knowledge
concerning them, would not be of any use for predictions at points
situated elsewhere in the deserts.
But while it was thus impracticable to prepare maps showing everywhere
the depth at which underground water actually exists, I conceived that
it might be quite possible to prepare maps showing _static water-
levels_; that is, the levels to which the water would anywhere rise
hydrostatically when once it _was_ tapped by borings. For the slope of
the static water-surface between known points will be largely
independent of the underground structure of the intervening country.
Apart from any physical changes which may still be going on in the
underground rocks themselves through geological agencies—changes which,
if taking place at all, must be so slow as to be negligible except in
the course of centuries—the only factors which can affect the slope of
the static water-surface, once it has been established, are variations
either in the rate of supply of water to the beds, or in the rate at
which it is removed from them. As regards variations in the rate of
supply, it is obvious that variations in the annual rainfall of the Erdi
and Ennedi region must cause very considerable variations from year to
year in the amount of water received by the underground beds. But the
resistance of friction to the flow of water through the pores of the
sandstones is so great, that the annual oscillations of pressure must be
rapidly damped out as the distance from the place of influx increases;
consequently the levels of the water in the wells of the Egyptian oases
(and even, so far as is known, that of the lake at Merga) show little or
no annual variation. And with regard to variations in the rate of
removal of water from the beds (by outflow to the Nile, or to the sea,
or into lakes wherein it evaporates, or by the exploitation of wells and
springs for irrigation purposes), these changes, though possibly in some
cases they may be progressive, and in restricted localities very
sensible, can exercise but little influence from year to year on the
general distribution of water-pressure within the underground strata. We
may therefore conclude that the gradient of the static water-surface
will everywhere have assumed practically a steady state. Unlike the
actual water bearing beds themselves, which may be much folded, the
static water-surface will in general have simple gentle slopes
everywhere in the open desert. In the oases, of course, where numbers of
wells yielding large outputs have been bored in proximity to each other,
the static water surface will be wrinkled; but over the vast bulk of the
desert the contours may be expected to be smooth curves. The
diagrammatic section below will, I think, make clear this point about
the general non-dependence of the shape of the static water-surface on
the geological structure:
[Illustration: _Diagrammatic section showing that the static water-level
is largely independent of the underground structure_]
In the diagram, FEKHG represents a water-conveying stratum, folded
throughout its course and faulted at HK, but having a general downward
slope from F to G. A and B represent points at which the water just
rises to the ground-level, either through natural fissures or in
artificial borings. The straight line drawn through A and B represents
very approximately the static level at any point between A and B, that
is, the level to which the water would rise in bores carried down to the
water-bearing bed. A boring at C, for instance, would have to go down to
E to tap the water, but once the bed was tapped the water would rise in
the bore as far as D. At the fault HK, the pressure of the water at K
will cause it to rise through the crushed rock at the fault-plane and
re-enter the porous stratum at H. If there is a considerable outflow at
B, and the fault-plane is a very narrow fissure, we may expect some drop
in the line AB over the fault, by reason of the extra frictional
absorption of head at this place. But unless the thickness and degree of
permeability of the fault-rock are markedly different from those of the
sandstone bed itself, the drop of pressure will not greatly disturb the
general slope AB. In any case it is apparent that the level of the
static water-surface at any place between A and B is capable of being
estimated with a far closer degree of approximation than is the level of
the water-sheet itself. We may therefore justifiably assume a uniform
gradient for the static level between points at which that level is
known, disregarding folds in the strata; and though we cannot entirely
allow in detail for unknown faults and variations in permeability, it
must be borne in mind that the _total_ effect of all the unknown factors
between any two known points is already automatically allowed for in our
data. It is only the variations from uniformity, due to the unknown
distribution of the faults and of the departures from the average
permeability, which can affect us; and these variations and departures
are probably but small in most of the great unexplored areas of the
south-west of Egypt, where the geological structure, from all we know of
it, appears generally to be remarkably uniform.
The first requisite for the construction of a map showing the contours
of the static water-surface was, of course, a sufficiency of well-
determined positions of points where the static water-level of the
artesian supply was fairly exactly known. Such points are the springs
and wells of the various oases, the surfaces of lakes occupying
depressions and presumably fed by underground supplies, and any places
on the Nile where the river taps artesian beds.
In regard to the wells and springs, it was obvious that only those known
to derive their supplies from artesian sources could be utilized as
giving points on the static water-surface. This consideration cut out
from the discussion the springs of Owenat and Arkenu, which are known to
be fed by local rainfall, and also the small water-sources of Kurkur,
Dungul, Nakheil, and Ain Amur, which occur in situations where
percolation from occasional local rainfall seems to be the only possible
source of supply. And for reasons which will appear presently, none of
the wells and springs situated to the north of the Siwa-Qattara-Faiyum
chain of depressions could be considered as entering into the problem.
With these exceptions, every water-source situated within the area of
the Libyan Desert covered by the map, and whose level was known, was
utilized; but in the greater oases and in the Wadi Natrun the wells and
springs are so numerous and so close together that in these localities
it was necessary to select one or two wells as representatives of a
group. I had no hesitation in including the wells of Siwa and the Wadi
Natrun, because the temperature of some of the wells and springs of
Siwa, and the quantity of the output of water at both places, seem to me
to afford conclusive evidence of the artesian character of their
supply.[18] The wells of the little oasis of Lageita, to the east of the
Nile near Qena, were included, for although they are not in the Libyan
Desert, they most probably derive their supplies from the same
underground flow which feeds the western oases. I have included only one
well in which the water does not rise nearly to the ground-level. That
well is one which was bored by the British Army during the Great War, at
a place called B6, some 40 kilometres to the east of Baharia Oasis. The
level of the ground at this point is 112 metres above sea-level, and as
the water was stationary in the bore at 78 metres below the ground, the
static level here is 34 metres above sea.[19] The well is said to have
yielded some 800 gallons per hour without the water-level in the bore
being sensibly changed.
The level of the well at Sarra has recently been determined by Prince
Kemal el Din; but I have not included it in my data, because he informs
me that the water-level fluctuates by 20 metres or more in different
years, while the level of the artesian water of the Egyptian oases and
Merga is very nearly constant. The inference I draw from the great
fluctuations in the water-level at the Sarra well is that it is
dependent on percolation from a more or less local rainfall rather than
on the same flow which feeds the Egyptian oases.
5. _Permanence of Lakes._
In regard to lakes and salt-marshes, the permanence of those occupying
the depressions of Areg, Bahrein, Sittra, and Qattara can only, I think,
be adequately explained by regarding them as fed, at least in part, by
underground supplies coming into them from the south. The total area of
the lakes of Bahrein, Nuemisa, Sittra, and Moghara is nearly 20 square
kilometres, and that of the salt-marshes (_sabakha_) is not less than
5000 square kilometres. The depressions are situated in a region which
is nearly rainless; in Siwa the mean annual rainfall is only about a
quarter of an inch, and that in the depressions farther south, such as
Bahrein and Sittra, is doubtless even smaller. The mean daily
evaporation from the lake-surfaces cannot well be less than some 4 mm.,
which would mean a lowering of the lake-levels by evaporation of 1½
metres each year unless there was some inflow to make up for the loss.
And though the rate of evaporation from the salt-marshes, area for area,
is doubtless very much smaller than that of the lakes, the 250-fold
greater extent of the marshes makes it certain that the total quantity
of water annually evaporated from them must far exceed that from the
lakes.
It appears unlikely that the loss by evaporation in the lakes and
marshes can be entirely made up from local rainfall and by seepages from
the northern slopes. The rocks forming the surface of the great Miocene
plateau, 200 metres high, which separates the depressions from the sea,
are chiefly limestones and clays; the beds are nearly horizontal, but
such slight dips as exist are believed to be towards the sea. The
average annual rainfall on the coastal portion of the plateau is about 6
inches; but it falls off rapidly inland, till it is only about a quarter
of an inch near Siwa. The heaviest rainfall on the plateau thus occurs
along a strip parallel to the coast, where it is largely drained off
towards the sea by the gullies which indent the plateau-edge. Of that
which falls on the plateau-surface farther inland comparatively little
is absorbed, owing to the generally non-porous nature of the uppermost
rocks; after a heavy shower, water lies on the surface in shallow pools
for a few days and is soon evaporated. So impervious to water is the
limestone in this region, that the Romans excavated chambers in it to
form reservoirs, of which many hundreds still exist. Nor can we think
that much surface drainage-water from the country to the south ever
finds its way into the depressions; for there is an almost complete
absence of drainage-lines entering them. At the feet of the northern
scarps of the Qattara depression, and along the north-eastern shores of
the lakes in the Wadi Natrun, there are, indeed, small springs which
show that some of the rain falling on the plateau does actually
penetrate the rocks and escape by seepage into the depressions. But the
amount of this seepage appears to be insignificant compared with the
volume of water which must annually disappear from the lakes and marshes
by evaporation. A further consideration bearing on this point is that
whatever may have been the agency by which the depression of Siwa was
formed, that same agency almost certainly operated to produce the other
depressions of the northern chain; and it seems most unlikely that a
connection should have been opened up with the underground water-bearing
beds in Siwa, and not also in the larger and much deeper depression of
Qattara.
It would be a difficult matter to estimate the relative proportion of
the water entering the depressions by underground flow from the south,
to that contributed by local rainfall and seepage from the northern
slopes. But that is not necessary for our immediate purpose. It is
sufficient to show that there must be _some_ influx into the depressions
from the same source as that which supplies the wells of the greater
oases, to establish the existence of that underground water-connection
which is all that we need to justify us in regarding the levels of the
lakes and salt-marshes as furnishing us with points on the static water-
surface; and from the considerations mentioned above it seems to me
certain that some influx of underground water really does take place.
I have also thought it justifiable to include the Birket el Qarun in my
collection of static water-level data, because although that lake was
probably first formed by an overflow of the Nile into the Faiyum, and is
even now being fed by Nile water through the Faiyum drains at the rate
of some 350 million tons a year, there is a certain amount of evidence
suggesting that it has some underground water-connection with the
Qattara depression. That evidence, to which attention was first drawn by
Professor Schweinfurth,[20] consists in the relatively low salinity (1·3
per cent.) of the lake, notwithstanding the long period through which it
has been subject to evaporation and the fact of its having shrunk to
dimensions very much smaller than it possessed in ancient times. Unless
there has been a large underground efflux of salt water from the lake,
it appears impossible to account for its present degree of freshness. In
Professor Schweinfurth’s day, of course, the existence of the Qattara
depression was unknown, and it was puzzling to suggest where the salt
water had gone to.[21] An underground leakage from the Birket el Qarun
into the Qattara depression is quite conceivable, for although the two
places are separated by some 200 kilometres, there is a very
considerable fall between them. Thus the salt in the marshes of the
Qattara depression may possibly have come in part from the Birket el
Qarun. The present rate of discharge of the Faiyum drains into the lake
is, however, just sufficient to make up for an average daily evaporation
from the lake-surface of a little over 4 mm., which is about the rate we
might expect; and although the level of the lake-surface has fallen some
5 metres since observations of it were first made in 1886, it is now
nearly stationary; hence it does not appear likely that there is much
underground leakage at present. If the former leakage from the lake took
place by lateral flow into porous strata near its surface, of course the
leakage may have been arrested by the lowering of the lake-level
uncovering the porous beds into which it took place; but I think a more
likely explanation is that the leakage occurred at or near the bed of
the lake, and has gradually been reduced by the continued deposition of
Nile mud on the lake-bottom, and by the diminution of head due to the
fall in the water-level.