GEOLOGIJA 31, 32, 555-576 (1988/89), Ljubljana
UDK 556.334.34.36.551.44(495) = 20
Development of the brackish karstic spring
Almyros in Greece
Sanacija zaslanjenega kraškega izvira
Almyros V Grčiji
Marko Breznik
University of Ljubljana, Hajdrihova 28, 61000 Ljubljana
Abstract
In karstic massifs of anisotropic permeability groundwater circulates practi-
cally, only through zones of high permeability called veins. Coastal springs are
contaminated by sea-vi^ater either at the mouth of a submarine spring or inside the
karstic massif in a branching of veins. The methods to desalinate brackish karstic
spring are:
- isolation of the karstic aquifer from sea-water intrusion by a dam at the spring
or by a grout-curtain that seals the lower vein with sea-water;
- interception of fresh-water by boreholes, wells or drainage galleries within the
karstic massif, inland of the sea-water influence;
- rise of the spring level by a dam. In that case, fresh-water accumulates in the
karstic massif and prevents the intrusion of sea-water.
There are already numerous cases of successful development of brackish
springs. A more detailed analysis discusses the development possibilities of the
important brackish spring Almyros with a mean discharge of 7m^/s on Crete
island in Greece.
Kratka vsebina
Podzemna voda se pretaka v kraških masivih neenakomerne prepustnosti
v pomembnih količinah samo skozi močno prepustne cone, imenovane žile. Ob-
morski izviri se zaslanijo ali v ustju podmorskega izvira ali v razcepu žil znotraj
kraškega masiva. Zaslanjene izvire lahko saniramo na naslednje načine:
- z izolacijo kraškega masiva od morske vode s pregrado pri izviru ali z injekcij-
sko zaveso, ki zatesni spodnjo žilo z morsko vodo;
- z zajetjem sladke vode z vrtinami, vodnjaki ali drenažnimi rovi v notranjosti
kraškega masiva izven vpliva morske vode;
- z dvigom gladine izvira s pregrado. V tem primeru se sladka voda akumulira
v kraškem masivu in prepreči vdor morske vode.
Imamo že mnogo primerov uspešne sanacije zaslanjenih kraških izvirov. Bolj
podrobno razpravljamo o možnosti sanacije pomembnega zaslanjenega izvira
Almyros s srednjim pretokom 7 m^/s na otoku Kreti v Grčiji.
556 Marko Breznik
Contamination in the branching of veins
Karst aquifers of anisotropic permeability are characterized by zones both of very
low and of high permeability. Groundwater circulates practically only through zones
of high permeability, called veins. The term "vein" does not determine its shape
which can be a solution canal, a cave, a fractured or cavernous zone, etc. The vein in
which fresh-water circulates is a primary vein. The place where the primary vein
branches (Fig. 1) into the lower vein connected with the sea and into the upper vein
leading to the spring is called a branching of veins (or vein-branching). Lower veins
were formed by fresh-water when their carbonatic rocks massif was above sea level
during previous geological periods. Later the elevation of the sea level increased in
relation to the massif with veins either due to the tectonic subsidence of the massif
Fig. 1. Scheme of karstic veins of a brackish spring (after Breznik 1973)
mo - Sea; gm - Sea level; i - Brackish spring; v - Primary vein with fresh-water; m - Lower vein
with sea-water; s - Upper vein with brackish water; r - Vein-branching; u - Mouth of lower
vein; m, min - Lowest point of lower vein; h and gf - Height above reference level; i - Mean sea
level; 2 - Reference level; 3 - Piezometric head line of the primary and upper veins; 4 - Piezome-
tric head line of the lower vein; 5 - Energy head line of the primary and upper veins; 6 - Energy
head line of the lower vein; 3 to 6 - All the heads expressed through the head of fresh water
SI. 1. Položaj kraških »žil« zaslanjenega izvira (po Breznik 1973)
mo - morje; gm - gladina morja; i - zaslanjen izvir; v - dovodna žila s sladko vodo; m - spodnja
žila z morsko vodo; s - zgornja žila s somornico; r - razcep žil; u - ustje spodnje žile; m, min
- najgloblji del spodnje žile; hing- višina nad primerjalno ravnino; 1 - srednja morska gladina;
2 - primerjalna ravnina; 3 - piezometrična višina dovodne in zgornje žile; 4 - piezometrična
višina spodnje žile; 5 - energetska višina dovodne in zgornje žile; 6 - energetska višina spodnje
žile; 3 do 6 - višine so podane s stebrom sladke vode
Development of the brackish karstic spring Almyros in Greece 557
(Tertiary to Holocene periods), or due to the rise (Holocene) of the sea level (Herak,
1975, 1977).
The mechanism of contamination inside the karstic massif was explained by
Gjurašin (1943), Kuščer (1950) and Breznik (1973). This type of contamination
of coastal springs is the most frequent one along the limestone coasts of Yugoslavia
and Greece, and there are some such springs along the coasts of France, Italy, Turkey,
Syria, Lebanon and other countries.
Supposed disposition of karstic aquifers
The Almyros spring on the Crete island in Greece is a typical brackish spring
contamined in the branching of veins inside a karstic massif of anisotropic permea-
bility (Bur don & Papakis, 1964; Kuščer I. & Kuščer D., 1962; Plataki,
1968).
Fig. 2. Schematic block diagram of veins in the karstic massif of the Almyros spring (after
Breznik 1978)
1 - Neogene deposits (sand, clay, organic limestone); 2 - Tripolitza series deposits (limestone
mostly); 3 - Metamorphic schists (quartzite, phyllite, chlorite, marble); 4 - Direction of flow^ in
veins during dry period; i - Almyros spring; v - Primary vein; r - Branching of veins; s - Upper
vein; m - Lower vein; m, min - Lowest point of lower vein; u - Mouth of lower vein; gm - Sea
level; / - Fault
SI. 2. Shematičen blok diagram žil v kraškem masivu izvira Almyros (po Breznik 1978)
1 - neogenski sedimenti (pesek, glina, organski apnenec); 2 - sedimenti serije Tripolitza
(pretežno apnenec); 3 - metamorfni skrilavci (kvarcitni filit, klorit, marmor); 4 - smer pretoka
v žilah v sušnem obdobju; i - izvir Almyros; v - dovodna žila; r - razcep žil; s - zgornja žila;
m - spodnja žila; m, min - najnižji del spodnje žile; u - ustje spodnje žile; gm gladina morja;
/ - prelom
558 Marko Breznik
The probable disposition of the lower karstic aquifer is explained in Fig. 2. The
lower main karstic aquifer which drains the Psiloritis massif is a deep, confined
aquifer contaminated by sea-water in several 500 to 1000 meter deep vein-branch-
ings (mixing points). There are some vein-branchings of different depths at
a supposed distance of 2 to 5 km from the spring.
The upper secondary karstic aquifer which drains the NE part of the Psiloritis
massif (Keri plateau) is a semi-unconfined aquifer. The main outflow of both aquifers
is the Almyros spring. The circulation between veins of the two aquifers is probably
weak, except in the spring area.
The very deep position of the primary and lower veins is explained by a gradual
subsidence of the eastern part of the Psiloritis massif (Papadopoulos & Scan-
vic, 1968) and by the existence of a Mesozoic limestone stratum below the Neogene
deposits of the Iraklion graben (Fig. 2). In former geologic periods the lower vein of
the lower aquifer was the primary vein and the main outflow into the sea probably to
the NE of Iraklion. During the Pleistocene period, with an 80 meter lower sea level,
the direct outflow into the sea in the area of the Almyros spring was blocked by a belt
of metamorphic schists of Roghdia. The present upper vein and the spring were
formed or reactivated afterwards. The present upper vein has namely a better
hydraulic gradient in comparison with the present lower vein which is supposedly
over 20 km long. Such a long vein is not an exception. For instance the famous sea-
water mills with a maximum inflow of 1.7 m^/s of sea-water on the western shore of
the Kefallinia island feed the lower vein of the Sami Springs with a discharge of
lOm^/s of brackish water situated on the eastern shore at a distance of 15 km.
Observations of Almyros spring
Measurements of discharge, salinity and elevation of water level of brackish
springs are the most important observations necessary to determine their mechanism.
Discharge versus salinity relation
The salinity starts and stops suddenly in correlation with the spring discharge in
the spring Blaž in Yugoslavia. The Almyros spring, however, has a gradual increase
of the salinity during the decrease of its discharge.
The discharge-salinity relations of the Almyros spring (Fig. 3) cleary show that
the salinity of the spring depends mainly on discharge and only sligtly on spring
level. In natural conditions, losses of head due to friction in the upper vein predomi-
nantly influence the important piezometric surface of fresh-water in the branching of
veins. The discharge of about 12-13 m^/s is the critical discharge that regulates the
inflow of sea-water at a spring level of 3 metres above M.S.L.
Summer test 1977 with rise of spring level
A dam was constructed downstream of the Almyros pool and the level of the
spring artificialy raised with the aim to reduce its salinity. There was not any rain
during the test. The observation data are presented in Table 1.
The influence of the rise in spring level was either small - 5 percent decrease in
discharge and salinity - or nil as the measurement accuracy was about 10 percent.
Percolation of water below the dam and springs on the downstream slope of the dam
were observed.
Development of the brackish karstic spring Almyros in Greece 559
Boreholes between the spring and the sea
From 1968 to 1971 many boreholes were drilled between the spring and the sea
with the aim to find the area with sea-water intrusion into the karstic massif, and to
explore the geology of the area. The deepest boreholes reached below 400 m. The
limestone massif was found to be underlain by metamorphic schists at a depth of
about 300 metres below the spring. The schists crop out between torrent Keri and the
sea. Limestone strata extend in the depth for about 0.5 km from the spring towards
the sea where they are cut of by a fault. The level and the salinity at different depths
were measured. Sea-water was not detected and the highest salinity of water in
boreholes was only one half of the Almyros spring salinity (United Nations,
1968, 1971, 1972).
Piezometric boreholes
Several new boreholes were drilled in the karstic massif behind the Almyros
spring. Levels of karstic groundwater and salinity were observed in these boreholes
and in the old ones. The results are presented in the following Table 2.
In most of the boreholes the levels were measured on the 11. 02. and 6. 10. 1983
and the salinity on the 17. 02. and 21. 10. 1983.
Mehanism of groundwater flow
Mechanism of contamination
An about 40 meter high column of sea-water exerts due to its higher density the
same hydrostatic pressure as an about 41 meter high column of fresh water. That is
known as the Ghyben-Herzberg law.
In coastal aquifers in sand and gravel of isotropic permeability the surface of
equal fresh and sea water pressures is called interface or zone of mixing. This zone is
about 2 m thick and can be detected in every borehole in the coastal area (Cooper,
1959).
Consequenty we recommended (Breznik, 1977, 1978) a reconstruction of the
dam spillway and a similar test to be performed in winter when the discharge is
greater.
Winter test 1987 with rise of spring level
The stability of the dam was increased with relief wells and the spillway recon-
structed. The level of the spring was raised to 10 metres above M.S.L. Discharge and
salinity were observed. The former and new results are presented in Fig. 3. The
salinity is reduced below 50 mg Cl/1 one or two days after the discharge rose over
about 12-13 m^/s. The brackish water, namely has to be washed out of the under-
ground storage after the inflow of sea-water has been blocked. Intrusion of sea-
water, perceptible by a rise in salinity over 100 mg Cl/1 started when the discharge
was reduced either below 12.5 mVs (29 March 1987) or below 9.2 mVs (14 May 1987).
A discharge of about 10-11 m^/s was a critical discharge which regulated the sea
water inflow at the spring level of 10 metres above M.S.L. during the 1987 test.
Fig. 3. Salinity in dependence of discharge and pool level elevation in the Almyros spring
CI - salinity; Q - discharge; (hrhrJ - elevation of spring pool surface
SI. 3. Slanost v odvisnosti od pretoka in višine gladine jezerca izvira Almyros
Cl - slanost; Q - pretok; (hrhm) - višina gladine jezerca izvira
562 Marko Breznik
Remark:   * discharge measured at the spring
*♦ discharge measured at DEI and sea canals
In coastal karstic aquifers of anisotropic permeability such an interface does not
exist because there the main quantity of water flows through fissure or canal veins.
A discontinuous plane of equal fresh and seawater pressures in veins is called the
equilibrium plane (Breznik, 1973). The difference between an interface and an
equilibrium plane is like the difference between the phreatic surface of an uncon-
fined aquifer and the piezometric surface of a confined karstic aquifer. The first
exists in the whole mass of gravel and could be detected in every borehole. The
second exists in veins only and in boreholes which have penetrated into the karstic
veins.
The position of the equilibrium plane determines the mixing process of fresh and
sea water in the vein-branchings. In the spring-autumn period the piezometric
surface of fresh-water in the primary vein is low and the equilibrium plane crosses
the vein-braching what enables the sea-water to intrude into the karstic aquifer. The
energy for such a flow pattern is furnished by fresh-water flowing through the vein-
branching.
Spring-Autumn Period in Almyros spring
In that period the Almyros is contaminated by sea water. The contamination
occurs in those vein-branchings where sea water has an equal pressure than the fresh
water. The head and energy lines explain the relevant flow conditions. Some vein-
branchings at different depths produce the gradual salination of the spring.
Table 1. Observations during summer test 1977
Tabela 1. Meritve med poletnim poizkusom 1977
Development of the brackish karstic spring Almyros in Greece 563
♦ perhaps blocked
Winter period in Almyros spring
In the v^inter period the piezometric surface of fresh-water is high and the
corresponding equilibrium plane in a low position. The Almyros spring discharges
fresh-water.
The lowest point of the lower vein - in limestone somewhere below the impervious
Neogene deposits of the Iraklion-Festos graben - is certainly deeper than the vein-
branching. A high pressure of sea-water in a lower vein of such a shape prevents
losses of fresh-water out of submarine springs during the fresh-water flow of the
Almyros spring in winter. Submarine springs were never observed in the Iraklion sea,
although they could be easily detected by their wheels, i.e. characteristic surface
surges. Evidently they do not exists in this area.
Possibilities of Development
Methods of Development
The methods to desalinate brackish karstic springs are:
- isolation of the karstic aquifer from sea-water intrusion by a dam at the spring or
by a grout-curtain that seals the lower vein;
- interception of fresh-water by boreholes, wells or drainage galleries within the
karstic massif, inland of the sea-water influence;
- rise of the spring level by a dam. In that case, fresh-water accumulates in the
karstic massif. The fresh-water piezometric surface is raised to a higher level and
the equilibrium plane is pushed down so that sea-water intrusion is blocked also
during dry periods.
Table 2. Water level and salinity in boreholes (Breznik, 1984a)
Tabela 2. Gladina in slanost vode v vrtinah (Breznik, 1984 a)
^. ^ Depth Groundwater o т•  -t
Distance ^^^^ Salinity
Benevole «Cs T'"il-'-        ^
^P'-inS       '„Tl. «lSan»S«n,     Winter Autu™
km m mm mm
564 Marko Breznik
Development of the Almyros Iraklion Spring
Extensive geological and geophysical investigations as well as drilling of deep
boreholes in 1968 to 1971 in the area between the spring and the sea have failed to
detect the lower vein. Now we are sure that it is not there. The lower vein is
presumably very long and at a great depth at an unknown position, and probably
connected with the sea to the NE of Iraklion. As its precise location is not known, it is
impossible to isolate the spring by grouting (Breznik, 1958, 1962, 1976, 1978).
Otherwise, it is possible to intercept the fresh-water in the area of the Gonies
gorge or in the Tylissos area by deep boreholes connected by a drainage gallery. The
technical and economic disadvantage of the interception method are the very great
number of boreholes wanted to intercept every important karstic vein, and that is
still aggravated by the high ground elevation. In the Gonies gorge about 450 meter
deep tube-wells with an about 320 meter pumping head and in the Tylissos area
about 300 meter deep tube-wells with an about 200 meter pumping head would be
needed.
Observations of the natural conditions indicate that the Almyros spring offers
favourable possibilities for development by means of a rise in spring level. The most
important observation results to consider are:
- Discharge - spring level - salinity relation
During winter the spring delivers fresh-water at a discharge of about 12-13 m^/s
and level of 3 m above M.S.L. During smaller discharges and lower levels the
spring water is brackish.
- Flood discharge and loss of water
As submarine springs have never been observed in the sea of Iraklion (Breznik,
1984 b), there are no losses of water through the lower vein, even during the flood
periods with discharges of about 50 m^/s and a very high piezometric head of fresh-
water in vein-branching and lower vein.
There are already numerous cases of successful development of brackish springs
(Biondić & Vulić, 1974; Fritz & Pavlin, 1978). The important Bačvice
springs in Yugoslavia were developed by grout-curtains that have blocked their
lower veins and isolated the aquifer from sea-water intrusion. Two regional drinking
water supply systems were constructed with desalinated water of this springs. The
Žrnovica spring in Yugoslavia was developed by means of a grout-curtain and by the
rise of the spring level. The period of fresh-water flow was extended by a short grout-
curtain in the area of the Tabačina spring (Pavlin, 1982).
The Dobrica, Dobra and Golubinka springs in Yugoslavia and Brojnica spring in
Italy were isolated by small diaphragm walls from seawater influence. The Kiveri
springs in Greece were developed by isolation of the spring area from the sea by
a dam and by the rise of the spring level (Pavlin & Biondić, 1971; Ständer,
1971; Pavlin, 1973, 1982; Pavlin & Fritz, 1978; Breznik, 1973).
Fresh-water of the springs Gustirna, Opačica, Bakar, Kovča-Zaton, Stari grad.
Korita, Jelša, Dubrava, Zvir and others in Yugoslavia was intercepted in galleries or
shafts. Alike the success was achieved with many deep boreholes in Greece and Italy
(Tadolini & Zanframundo, 1988) at distances of 5 to 10 km from the sea. It has
been possible to intercept only about 20% of the discharge of the correspondent
springs (Fritz, 1978, 1979; Mijatović, 1976, 1986, Breznik 1973).
Development of the brackish karstic spring Almyros in Greece 565
Fig. 4. Flow directions in the lower aquifer of the Psiloritis-Iraklion region in the winter period
during fresh-water flow of Almyros spring (after Breznik, 1984 a). See legend of Fig. 1
7 - Piezometric head line of the lower vein; 8 - Energy head line of the lower vein; 7-8 all the
head expressed through the head of sea-water; 9 - Equilibrium plane in winter period; 10
- Equilibrium plane during winter flood discharge; 11 - Pressure line of the lower vein
expressed through the head of fresh-water
S1. 4. Smeri pretoka v spodnjem vodonosniku področja Psiloritis-Iraklion pozimi med
sladkovodnim izlivom izvira Almyros (po Breznik, 1984 a). Glej legendo slike 1
7 - Piezometrična višina spodnje žile podana s stebrom morske vode; 8 - Energetska višina
spodnje žile podana s stebrom morske vode; 9 - Ravnotežna ploskev v obdobju zime; 10
- Ravnotežna ploskev v obdobju zimskih poplavnih pretokov; 11 - Tlačna črta spodnje žile
podana s stebrom sladke vode
- Discharge - raised spring level - salinity relation
During the 1977 summer test the level was raised to 10 m above M.S.L., and with
a discharge of about 4 rn^/s the decrease in discharge and salinity of 4500 mg Cl/1
were either very small - about 5 % - or possibly nil.
- During the 1987 winter test the level was raised to 10 m above M.S.L., and with
a discharge of about 10-11 ^/s the spring delivered fresh water.
The explanation of these results is as follows:
The piezometric head in the vein-brancing regulates the position of the equilib-
rium plane and of the inflow of sea-water into the spring.
- During winter the karstic massif is filled with water, so the piezometric surface
rises. The fresh-water head and pressure at the vein-branching and in the lower
vein are high enough to block the inflow of sea-water. The equilibrium plane is
below the vein-branching and sea-water cannot infiltrate into the vein at dis-
charges above 12-13 m'Vs (Fig. 4).
- During flood discharges the head and pressure at the vein-branching are higher
due to very high losses of head in the upper vein. The equilibrium plane is lowered
566 Marko Breznik
Fig. 5. Flow direction in the lower aquifer of the Psiloritis-Iraklion region in the spring-autumn
period during brackish water flow of Almyros spring (after Breznik, 1984b). See legend of Figs.
1 and 4
12 - Equilibrium plane in the spring-autumn period
SI. 5. Smeri pretoka v spodnjem vodonosniku področja Psiloritis-Iraklion od spomladi do jeseni
med zaslanjenim pretokom izvira Almyros (po Breznik, 1984 b). Glej legendo si. 1 in 4
12- Ravnotežna ploskev v obdobju pomlad-jesen
and interesects the lower vein at a lower elevation. But not all the lower vein is
above equilibrium plane, as this would induce losses of fresh-water through the
lower vein. There is no flow of either fresh- or sea-water through the lower vein
(Fig. 4).
- From spring to autumn, the karstic massif is drained, the spring level lowered and
the fresh-water head at the vein-branching subsides as losses of head in the upper
vein are smaller due to smaller discharges. The ascending equilibrium plane
crosses the vein-branching and with discharges below 12-13 m^/s some sea water
infiltrates into the upper vein (Fig. 5). In the case of more vein-branchings sea
water infiltrates first throught the lower ones.
- During the 1977 test the fresh-water head and pressure at the vein-branching were
equal to the sea-water head and pressure, expressed through the head of fresh-
water, so sea water countinued infiltrating into the upper vein. This indicates that
the fresh-water head and pressure at the vein-branching were at that time smaller
than during the winter when the discharge had amounted to 12-13 m^/s and the
spring water was fresh. The artificial rise in spring level in 1977 did not counterba-
lance the difference in head losses between the discharges of 4 m^/s and 12-13 m^/s.
The artificial rise in spring level was too small. Evidently, during the summer
a higher artificial rise in spring level is needed.
Development of the brackish karstic spring Almyros in Greece 567
Supposition on the Necessary Artificial Rise in Sping Level
The unknown value is the piezometric head of fresh-water in the vein-branching
(hr - hm) during the fresh-water flow of the spring. It is the sum of the spring
elevation (h, - h^) and of losses of head yQ^ in the upper vein reduced for the velocity
head avr^/2g in the vein-branching.
The direction of flow, largely determined by the orientation of the veins, may
influence the mixing process during high discharges but cannot be accounted for
with certainty. This influence is very small in the spring-autumn period as the
velocity of flow is small due to small discharges.
The elevation of the spring can be measured. Losses of head depend on length,
cross section and roughness of the upper vein and the square of discharge. On that
last quantity also depends the velocity head. The natural condition of the vein does
not change.
Thus, the changes in loss of head and velocity head can be expressed as functions
of discharge:
Ahn_i : ^K : Ah,+ i = Q^.j^ : Q„2 ; Q^^^2 and
avn_i^  avn^  aVn+i^
2g      2g      2g
where
Ah - loss of head m
ctv^
--head of velocity m
2g
g   - acceleration of gravity m/s^
V   - velocity of flow m/s
Q - discharge mVs
The thus derived ratios of loss of head for characteristic discharges are given in
the following Table 3.
The ratios between the losses of head and the discharge clearly indicate the
predominance of the influence of discharge on the fresh-water head and pressure in
the vein-branching.
The characteristics of the upper vein that determine the losses of head in vein-
branching can be calculated if the discharge at which the fresh-water flow starts is
measured at two different spring levels.
- During the 1987 winter test the spring level was raised to 10 m above M.S.L. and
the spring delivered fresh water at discharges of about 10-11 m^/s. This indicates
a clear influence of the level rise on the spings salinity.
568 Marko Breznik
Table 3. Ratios between losses of head as function of discharges
Tabela 3. Razmerje med tlačnimi izgubami in pretokom
During the winter test 1987 the spring level was 10 m. The discharge increased
very quickly and considering the volume of underground storage it is difficult to
assess the critical discharge at which the fresh-water flow initiated. It was easier to
evaluate the discharge at which fresh-water (salinity below 50 m Cl/1) began to
become slightly salty (100-200 mg Cl/1).
This discharge was 10 to llm^/s during the winter test 1987 and 12 to 13m^/s
during the years 1971-1977 when the spring level was 3 m above M.S.L.
hr-h„ = (h,i-hJ+yQi2
hr - h^ = (hi.2 - h„) + yQs^ (B r e z n i k 1978)
hi - hm - spring level m above M. S. L.
hr - hm — piezometric head in vein-branching during the fresh-water flow of
the spring m above M.S.L.
y - hydraulic characteristics of upper vein
Q        - discharge
For Qi = 12.5 mVs, (hji - hm) = 3 m, Q2 = 10.5 mVs
and (hi.2 - hm) = 10 m is y = 0.1522 and h^-hm = 26.78 m
For measurement data of the piezometric levels in boreholes in the karstic marssif
behind the Almyros spring see Table 2. The boreholes II (ALD 3), Tl and G5 (ALD 1)
with winter levels of 45.8, 44.6 and 42.6m above M.S.L. are too far inland. The
boreholes G3/68 and G4 (FAO 9) with winter levels of 7.3 and 8.9 m are apparently in
the secondary upper karstic aquifer and the borehole G2 (FAO 6) with a winter level
of 2.9m in the outflow area of both aquifers.
However, the boreholes Argiry-Tylissos, T2 and G6 (AGR 1) with winter levels of
22-30 m, 31.1m and 28.0 above M.S.L., all situated in the Tylissos-Koubedes area,
could give an indication on the winter levels either in the area of the vein-braching,
or in an area more inland of that area, or in an area drained by the upper aquifer.
We can surmise therefore that the needed autumn elevation of spring level should
be from 20 to 30 m above M.S.L. This elevation has to be determined by new tests in
the rise of the spring level.
Development of the brackish karstic spring Almyros in Greece 569
Underground dam
An underground dam similar to many constructed below surface dams and in
embankments of reservoirs in karstic areas of Yugoslavia and Spain offer the
advantage of a step by step construction and of a smaller financial risk.
The first stage is the installation of 2000 mm steel pipes into the lower and upper
springs with the assistance of divers and a crane on the surface. The pipes are fixed
and sealed with concrete plugs and grouting by pumping the concrete and the grout
from the surface. An outlet valve and a gate or 2 valves with driving mechanism on
the pool shore and on the road are installed later. Piezometric boreholes are needed
too. The first test of raising the piezometric surface in the rock massif by partial
closure of the valve and the gate could be made in summer or autumn during small
discharges of the spring. There is a fair chance to achieve a désalinisation of the
spring, because the rise in piezometric level can be only small due to permeability
and a too small stability of the rock above the upper spring. That could be sufficiently
increases by prestressed anchors, however. The main task of these constructions will
be implemented in the second stage (Fig. 6).
The second stage consists in the drilling of boreholes and interception wells and
excavation of a collecting gallery. The wells have to be drilled either out of the
collecting gallery by a technique used in mining (e. g. diameter 350 mm, depth 400 m,
Bleiberg mine, Austria, works performed by Geološki zavod Ljubljana) or from the
surface by the normal drilling technique and by 30 m deeper wells. The aim is to open
new ways for underground circulation. A test of closing the lower and upper springs
by the valves installed in the first stage and observation of the piezometric surface
Rockfill dam
The construction of an appropriate rockfill dam might be confronted with serious
problems. The requested height (20 to 30 m) of the dam is not exactly known in
advance because it depends on the rise of the spring level achievable by the dam. That
requests a dam and a spillway that can be constructed in stages. A concrete face
rockfill dam would be an appropriate solution.
Makropolos (1985/86) concluded that there is a 90-percent probability for
a 0.21-0.24 g horizontal acceleration earthquake at Knossos near Iraklion in the next
50-100 years. We have to expect a bigger earthquake at the Almyros rockfill dam site
due to the presence of the main fault and sandy materials below the dam.
Fine to medium sand saturated with water is not a good foundation ground in
seismic areas as the overturning of a structure in the 1961 Niigata earthquake in
Japan and the collapse of the San Fernando dam in the 1971 earthquake in USA have
demonstrated. Very gentle slope of dam faces will be required.
The stability of the present dam was endangered by springs on the downstream
face in the 1977 test and improvement measures were undertaken afterwards.
Construction of a deeper diaphragm wall and other measures will be necessary in
order to obtain the Lane's ratio (length of flow: head of water) of 5-8 needed for the
hydraulic stability of sand layers against underground erosion. Construction of the
present diaphragm wall was difficult already because of blocks of older rocks in sand
layers. A semicircular dam is needed as the space around the spring is limited and the
rocky abutments are short.
570 Marko Breznik
should show the flow capacity of the new water passage. Afterwards, a second test in
raising the piezometric surface in the rock massif can be performed also in spring
period. A cupola-shaped concrete plug in the access tunnels would make such a test
possible. During a to high spring discharge the plugs could be blasted because a to
high water pressure inside the rock massif could be dangerous. The concrete plug in
a spring at Fatničko polje in Yugoslavia had to be blasted after the piezometric head
in the rock massif rose over 100 m and endangered a village and a road. There are
better chances to succeed in this test than in the first stage because the rise can be
done in spring period also.
The third stage entails closure of canals of the lower and upper springs with
concrete plugs cast by pumping concrete from the surface. The underground dam is
constructed by sealing the main canals by a grout-curtain of cement-bentonite-sand
grout (Borelli & Pavlin, 1965, Pavlin 1970, 1982) and by sealing the cavities
either by pumping cement mortar or by constructing prepact concrete through 15 mm
Fig. 6. Cross-section of Almyros springs with structures needed for the first and second stages of
explorations
SI. 6. Prerez izvirov Almyros z objekti potrebnimi za prvo in drugo fazo raziskav
Development of the brackish karstic spring Almyros in Greece 571
boreholes - a technique used at Nikšič dams in Yugoslavia (Breznik, 1979, 1985) or
by pumping concrete through 300 mm boreholes - a technique used at Keban dam in
Turkey (Uzbek 1975).
In the outer part of access tunnels steel pipes of 2000 mm have to be fixed and
sealed by concrete plugs. The valve and the gate installed in the first stage into the
springs have to be dismantled and fixed on steel pipes at the outflow of access tunnels
(Fig. 7). By partial closing of the valve and of the gate the piezometric surface will be
gradually raised to 20 or 30 m. Desalination of the spring has to succeed also in
autumn during the lowest discharges. The stages from one to three are still explora-
tory, and a final positive result is not assured in advance.
The forth stage is to be implemented only if a complete desalination has been
achieved in the third stage. The construction works will include the extension and
Fig. 7. Cross-section of Almyros springs with structures needed for the third stage of
explorations
SI. 7. Prerez izvirov Almyros z objekti potrebnimi za tretjo fazo raziskav
572 Marko Breznik
Evaluation of development success
With the plan of raising the spring level the two following circumstances are to be
taken into consideration as possibilities of an unsuccessful development.
Escape of fresh-water through the lower vein
Due to the increased spring level fresh-water could escape through a lower vein
that rises from the vein-branching toward the sea (Breznik 1973). The equilibrium
plane in such a case is below the lower vein. If a lower vein is falling from the vein-
branching, the equilibrium plane might cross the lower vein and fresh-water would
not be lost. It seems that the Almyros spring indeed has a falling lower vein, as no
losses of fresh-water into the sea have been observed also during the flood periods
(Fig. 2).
Losses of fresh-water around the dam or into the upper karstic aquifer
In such cases the spring level could not be raised to the needed elevation in the
spring pool and not high enough in the vein-branching, and so the spring water
would remain brackish during the lowest discharges. We estimate that there is
a possibility of such water losses. In the spring and perhaps also in the summer month
the losses would be less than the fresh-water discharge, and the levels in the spring
and vein-branching could be raised to the necessary elevation; the spring water
would be fresh. Yet in the autumn period the levels in the spring and vein-branching
might decline and admit the instrusion of seawater.
During the summer test in 1977 with a spring-level elevation of 10 m above M.S.L.
the losses around the dam were small, and only an about 5-percent decrease in spring
discharge was estimated. The rise of water level in boreholes around the spring was
small, too. This indicates small losses of water outside the main spring during the
1977 rise of spring level. On the base of these observationes we estimate that the
losses into the upper karstic aquifer will not deplete too much the spring fresh-water
flow. The losses around the dam can be blocked by a larger grouting screen.
We have to allow for a 20-percent risk of failure in the autumn period. But also in
such a case of small probability the water would persist fresh during the winter,
spring and early summer periods and could be used for irrigation or stored in
a reservoir in the Neogene area for use in late summer and autumn.
condensing of the grout-curtain, completion of the bottom outlets, construction of
a shaft with spillway, chute and skijump, of the water supply tunnel, of the pipeline
with control valve and of the monitoring and driving devices. A small hydropower
station could be installed, too.
It cannot be known in advance how many interception wells will have to be
drilled, but the underground canals area is not wide. The proposed underground dam
is very small in comparison with the already constructed grout-curtains of Yugoslav
dams in karstic areas (Breznik, 1979, 1983, 1985, 1988, 1989; Pavlin, 1961, 1970).
Underground dams are safe in special conditions, e. g. wars, terroristic attacks or
earthquakes.
Sanacija zaslanjenega kraškega izvira Almyros v Grčiji 573
Rehabilitation of the natural scenery
The Almyros spring was a natural beauty with its depth, size, shape, flora,
quantity of water and a historical monument with old mills.
A surface dam would destroy all this forever. An underground dam offers all the
possibilities to restore the previous natural scenery and the old mills. This and
a museum explaining the mechanism of the sea-water intrusion, of many exploration
campaigns, of the engineering works performed and of the mechanism of an unique
desalination achieved could made Almyros spring world famous and a touristic
attraction similar to the Kefallinia island sea-water mills.
In the case of a development failure a rockfill would remain as a „monument" of
an unsuccessful! development. On the other hand all the structures of the under-
ground dam development method would not be visible.
Author's rights
All the rights to use parts of this article as explanation of the mechanism of
contamination and desalination, programme of the step-by-step explorations and
design of the development structures after the underground dam method are with the
author.
Acknowledgements
Thanks are due to the Government of Greece and to many of my Yugoslav
colleagues who gave their unpublished data at the disposal.
Sanacija zaslanjenega kraškega izvira
Almyros V Grčiji
Povzetek
V kraškem vodonosniku neenakomerne prepustnosti se pretaka kraška podzemna
voda skozi žile, ki so lahko rovi, kaverne, razpokane ali skrasele cone. Sladka voda, ki
doteka po dovodni žili, se zaslani v razcepu žil s težjo morsko vodo, ki doteka skozi
spodnjo žilo. Mehanizem takega zaslanjenja so pojasnili Gjurašin (1943), Kuščer
(1950) in Breznik (1973).
In any case we have to expect some losses of fresh-water into the upper karstic
aquifer or around the dam. The dam alone cannot capture the whole fresh-water flow
of the spring. Yet the lost water will be probably fresh and could be captured by
wells, by the existing dam and smaller dams.
The Almyros spring offers very good natural conditions and technical possibilities
for a successful desalination but the progress of explorations has been surprisingly
slow. We estimate the distrust in the proposed rise-in spring-level with underground
dam method being the main reason.
574 Marko Breznik
Izvir Almyros na otoku Kreti v Grčiji s srednjim pretokom okrog 7 m^/s in gladino
3 m nad morjem, je pretežni del leta zaslanjen z do eno četrtino morske vode. Ob
večjih pretokih pa ima krajši čas sladko vodo. Zadnjih 20 let so bile izvršene obsežne
raziskave, ki naj bi razjasnile mehanizem zaslanjevanja in ugotovile možnost sana-
cije. Slanost je odvisna predvsem od pretoka in le delno od gladine izvira.
Z obsežnimi raziskavami smo ugotovili, da morska voda ne doteka v razcep žil iz
okrog 1 km oddaljenega morja. Spodnja žila mora biti zelo dolga, podobno kakor na
otoku Kefallinija pri znanih mlinih na morsko vodo. Točen položaj spodnje žile ni
poznan, zato izvira ni možno sanirati z izolacijo morskega vpliva z injekcijsko zaveso,
ki bi zatesnila spodnjo žilo.
Manjši del vode izvira bi bilo možno prestreči v razdalji okrog 4 km od morja
s 300 m globokimi vrtanimi vodnjaki z višino črpanja 200 m.
V področju Irakliona niso nikoli opazili podmorskih izvirov, tudi ne ob poplav-
nem pretoku Almyrosa z okrog 50 m^/s. Zato sklepamo, da se spodnja žila od razcepa
žil proti morju spušča v večjo globino, kar je posledica pogrezanja tektonskega jarka
Iraklion-Festos. Spuščajoča se spodnja žila omogoča sanacijo z dvigom gladine
izvira. Posledica je splošen dvig gladine sladke vode v kraškem masivu. Povečan
pritisk sladke vode v razcepu žil pa naj bi onemogočil vdor morske vode vanj. Analize
so pokazale, da je v času, ko je izvir sladek, pri pretokih nad okrog 12-13 m^/s,
piezometrična gladina v razcepu žil predvsem pod vplivom tlačnih izgub v zgornji
žili. Poleti 1977 so s pregrado dvignili gladino izvira na 10 m nad morje. Poizkus ni bil
pozitiven, ker se slanost ni zmanjšala in tudi ne negativen, ker se pretok ni zmanjšal.
Očitno je bila pri tem poizkusu piezometrična gladina v razcepu žil nižja, kakor je
pozimi ob sladkovodnem izviru. Umetni dvig gladine izvira ni nadomestil manjših
tlačnih izgub v zgornji žili pri pretoku okrog 4 m^/s poleti 1977. V sušnih obdobjih bi
bil potreben mnogo večji dvig gladine izvira. Na osnovi analize razlike tlačnih izgub
pri različnih pretokih in gladine podzemne vode v oddaljenosti okrog 4 km v notra-
njost kraškega masiva cenimo, da bi bil potreben dvig gladine izvira na 20 do 30 m.
Vendar bo možno ugotoviti končen pozitiven rezultat samo s terenskim poizkusom.
Ker so takšne raziskave izredno drage predlagamo postopne raziskave in tudi
izgradnjo injekcijske zavese, namesto nasute pregrade. Ocenjujemo da je 20 % mož-
nost za neuspeh, ker zaradi prepustnosti morda ne bi mogli dovolj dvigniti gladine
v razcepu žil ob suši. Zato bi bila gradnja pregrade ali injekcijske zavese še vedno
poizkus, končen uspeh ni zagotovljen v naprej, vendar je velika verjetnost, da bo
uspel. Preseneča počasnost raziskav tako pomembnega izvira, ki je predvsem posle-
dica nezaupanja v predlagan način sanacije.
Development of the brackish karstic spring Almyros in Greece 575
References
Biondić, B. & Vulić, Ž. 1974, Hidrogeološke prilike slivnih područja izvora sjevernog
dijela Hrvatskog Primorja (Hydrogeologie Conditions of the Springs in the Northen Part of the
Croatian Coast). Zbor. rad. 3. jug. simp, o hidrogeol. i inž. geoL, 1-11. Opatija.
Borelli, M. & Pavlin, B. 1965, Karst Storages Buško Blato, Peruča and Kruščica.
Unesco, I. H. D. Symp. on Hydrology of Fracturated Rocks. No 30, 32-62. Dubrovnik.
Breznik, M. 1958, Probno injekciono polje Plašnica (Plašnica-Grouting test plot). 4. jug.
kongr. o visokim branama, 294-301, Skopje.
Breznik, M. 1962, Akumulacija na Cerkniškem in Planinskem polju (Water Accumulation
in the Cerknica and Planina Poljen). Geologija, 7, 119-149. Ljubljana.
Breznik, M. 1973, Nastanek zaslanjenih kraških izvirov in njihova sanacija (The Origin of
Brackish Karstic Springs and Their Development). Geologija, 16, 83-186. Ljubljana.
Breznik, M. 1976, Mogućnost saniranja zaslanjenih kraških izvora sa injektiranjem
(Possibility of Development of Brackish Karstic Springs By Grouting). Prvi jug. simp, za kons,
tla, 293-296. Zagreb.
Breznik, M. 1977, Test to Raise the Water-level of the Almyros Irakliou Spring - Evalu-
ation of Results of the 1977 Summer Test. Not published report presented to the Greek
Government, 1-11, Iraklion.
Breznik, M. 1978, Mechanism and Development of the Brackish Karstic Spring Almyros
Irakliou. Ann. Geol. des Pays Hell, 29-46. Athènes.
Breznik, M. 1979, The Reliability of and Damage to Underground Dams and Other Cutt
Off Structures In Karstic Regions. 13'^ Inter, Cong, of Large Dams. New Delhi, C. 4, 57-79,
Paris.
Breznik, M. 1980, Safe Yield of Wells in Arid Areas. lARH-Unesco Sem. Hydr. Res. and
River Basin Devel., Nairobi, 130-133, Beograd.
Breznik, M. 1983, Večnamenska akumulacija Cerkniško jezero (Multipurpose Accumula-
tion Basin Cerknica). Gradb. vest. št. 1, 3-15, Ljubljana.
Breznik, M. 1984 a. Development of the Almyros Irakliou brackish spring. Not published
report presented to Greek Government. 1-34, Ljubljana.
Breznik, M. 1984b, Exploration of the Bali brackish springs. Not published report
presented to Greek Government. 1-5. Malia.
Breznik, M. 1985, Exploration, design and construction of cut offs in Karstic regions, 15'^
Inter. Cong, of Large Dams, Lausanne. Q. 58, Vol. Ill, 1111-1129, Paris.
Breznik, M. 1988, Influence of the Mavčiče reservoir on the environment. 16'^ Int. Cong, of
Large Dams, San Francisco, Q. 60, Vol. I, 1547-1568, Paris.
Breznik, M. 1989, Explorations, mechanism and development of brackish karst spring
Almyros toy Irakleiou. Unpublished report presented to Greek Government and Universities, 59
p., 21 figs., 3 tabs., Ljubljana.
Burdon, D.J. & Papakis, N.J. 1964, Preliminary Note on the Hydrogeology of the
Almyros Spring Iraklion. Inst, for Geol. and Subs. Res., Athènes.
Cooper, H. H. Jr. 1959, A hypothesis concerning the dynamic balance of fresh water and
salt water in coastal aquifer. J. Geophys. Res., 64, no. 4, U.S.A.
Fritz, F. 1978, Hidrogeologija Ravnih kotara i Bukvice (Hydrogeology of Ravni Kotari and
Bukvica). Jug. akad. znan. i umj., Carsus lug. 10/1, 1-43. Zagreb.
Fritz, F. 1979, Hidrogeološke rajonizacije priobalnog krša Hrvatske u svetlu novijih
saznanja (Hydrogeologie Zoning of the Coastal Karst of Croatia in the View of Recent Studies).
Geološki vjesnik 31, 327-336. Zagreb.
Fritz, F. & Pavlin, B. 1978, Vodonosno područje kod Žmana na Dugom otoku (Ground-
water Conditions at Žman on Dugi Otok Island). 5'^ Jug. simp, za hidrogeol. i inž. geol., 57-64.
Beograd.
Gjurašin, K. 1943. Prilog hidrografiji primorskog krša. Tehnički vjesnik, 1-2, 1-17,
Zagreb.
Herak, М. 1975, Geotectonics and Karst, Acta geologica, VIII/2Í, (Prir. istr. JAZU, 41).
387-395, Zagreb.
Herak, М. 1977, Tecto-genetic approach to the classification of Karst terrains, Jug. akad.
znan. i umj., Carsus lug., 9/4, 227-238, Zagreb.
Kuščer, I. 1950, Kraški izviri ob morski obali (Karst Sources at the Sea Coast). Dissertati-
ones. Academia Scientiarum et Artium Slovenica, Classis III, I, 97-147, Ljubljana.
576 Marko Breznik
Kuščer, I. & Kuščer, D. 1962. Observation on Brackish Karst Sources and Swallow-
Holes in the Yugoslav Coast. Mem. de D'Association Inter, des Hydrogeologues. V. Reunion
d'Athenes.
Makropoulos, K. 1985/1986, Seismic Hazard of Knossos Hystorical Site. Ann. Geol. des
Pays Hell., 169-189, Athens.
Mijatović, B. 1976, Kaptaža Mojdež kot Hercegnovog (Captage Mojdež près de Herceg-
novi). 4"^ Jug. simp, za hidrog. i inž. geol., 235-249. Skopje.
Mijatović, В. 1986, Problèmes de captage en régions Karstiques littorales. Univ. de
Neuchatel. Bull, du centre d'hydrogéoL, 1986, 65-106. Neuchatel.
Ozbek, E. 1975, Cavity at Keban Dam. Bull. Inter. Assos. Eng. Geol., No 12, 45-48.
Krefeld.
Papadopoulos, N. &i Scanvic, Y. 1968, Esquisse géologique de l'île de Crète. Raport
inédit. Paris.
Pavlin, B. 1961, Réalisation du bassin d'accumulation de Peruča dans le Karst Dinarique.
Vll'h Inter. Congr. of Large Dams. Q. 25, 1-28. Rome.
Pavlin, B. 1970, Kruščica Storage Basin in the Cavernous Karst Area. X'*' Inter. Cong, of
Large Dams. Montreal. Q 37, R 13, 209-224. Montreal.
Pavlin, B. 1973, Estabilishment of Surbsurface Dams and Utilization of Natural Subsur-
face Barriers for Relization of Underground Storages in the Coastal Karst Spring Zones and
Their Protection Against Sea-Water Intrusion. ХГ*> Int. Con. of Large Dams. Q 40, R 33,
487-501. Madrid.
Pavlin, B. 1982. Izvor Tabačina. Tehničke mjere za zaštitu protiv upliva mora. Zbor. ref.
sav. Obezb. i zašt. voda. Sisak. Sav. inž. i tehn. Jug. 58-77. Beograd.
Pavlin, В. & Biondić, В. 1971, Dopunska kaptaža kraškog izvora Opačica za vodovod
Herceg Novog (Aditional Catchment Works in the Karst Spring Opačica Feeding Herceg Novi
Municipal Water Supply System). Zbor. rad. I. jug. simp, o hidrogeol. i inž. geol., 181-192.
Hercegnovi.
Pavlin, B.& Fritz, F. 1978, La protection du système des sources karstiques de Golu-
binka contre la contamination par la mer. SIAMOS-78, 227-236. Granada.
Plataki, E. 1968, O Almyros toy Irakleioy. Kritika Hronika, 55-104, Irakhon.
Ständer, W. 1971, written communication.
Tadolini, T. & Zanframundo, P. 1988, A Water Balance of the Karstic Aquifer in
Apulia - Southern Italy in Viewof a Rational Exploitation of Ground Water Resources. Proc.
Karst Hydrol. Karst Envir. Prot., 1983-1984, Guilin China.
United Nations UNDP-FAO (Ré, Breznik 1968), The Problems of the Almyros Spring of
Iraklion. Tech. note. Nr. 2, 1-114. Iraklion.
United Nations UNDP-FAO (Breznik 1971), Geology and Hydrogeology of the Almyros
Spring Area. Tech. note Nr. 103, 1-94, 22 fig., Irakhon.
United Nations UNDP-FAO (Dietrich, Ré 1972), Study of the Almyros Spring of
Iraklion. Tech. Rep. 3, 1-177. Irakhon.
United Nations UNDP-OTC (Breznik 1977), Exploration and Development of Coastal
and Submarine Brackish Springs in Turkey. Not published report presented to Turkish Govern-
ment. 1-15, Ankara.