International Symposium on Advances in Isotope Hydrology and its role in

sustainable Water Resources Management,

Vienna, Austria, 21-25 May, 2007
1
IAEA-CN-151/125

Tritium as an indicator of groundwater overexploitation in a tropical

karst aquifer

L. F. Molerio León

CESIGMA, S.A.,
PO Box 6219, CP 10600, Habana 6,
La Habana, Cuba

Abstract. Overexploitation of two huge Cuban karst aquifers is derived from Tritium sampling and isotope balance
modeling. During the dry season no 3H activity was measured in the ground waters discharged at the springs or
either at some of the observation wells. Occasionally very high values for 3H were recorded. These values are
strongly correlated with rainfall that took place at the 1980 decade or a little older. On the other hand, 3H activity
linked with present precipitation has been also recorded. Therefore a good mixture of water of different origin and
residence time moves through the aquifer. The change in the isotopic composition therefore, has been interpreted as
an indicator of overexploitation as far as waters that are not replenished during the present hydrological cycle are
pumped out of the aquifer. Residence time modeling applying Maloszewski and Zuber, [1,2] and Zuber and
Maloszewski, [3] lumped model approach fits these no Tritium spring waters with waters with around 100 years of
residence time. A more general conclusion is derived from the results obtained in the last years in both poljes
relative to the mixing hydrodynamics in karst aquifers were flow stratification is associated with cave levels thus
allowing that under certain conditions, old ground waters prevail in the mixing allowing important losses of ground
water reserves that are not replenished during the present hydrological cycle.

1. Introduction

During the last decade Cuba has suffered a large hydrological drought. In the particular case of the Vento
Basin, a karstic polje of about 325 km2, ground water levels has shown a sustained decline, several wells
became completely dry, and the yield of the Vento springs averaging 3 m3/s in the dry season has reduced
to about 1 m3/s. The isotopic composition of groundwater has dramatically changed and very old waters
(with no Tritium) began to appear systematically during the last months of each dry season (February-
April) at the springs that constitute the natural discharge of the aquifer.

The Vento Basin (Fig. 1) is one of the four aquifers supplying fresh water to La Habana, the capital of
Cuba (2,1 million inhabitants) since the last 100 years. The recorded change in the isotopic composition
therefore, has been interpreted as an indicator of overexploitation as far as waters that are not replenished
during the present hydrological cycle are pumped out of the aquifer. Residence time modeling [1-3]
lumped model approach fits these no Tritium spring waters with waters with around 100 years of
residence time. The continuous use of waters with so slow natural replenishment has largely contributed
to the depletion of the aquifer.

During the IAEA´s project ARCAL XIII a similar result was previously obtained by the author and his
colleagues [4,5] in the contiguous polje of Jaruco in 1993, were the hydrogeological structure is very
similar to that of the Vento Basin. In that case environmental isotope techniques provided a useful tool to
the decision-makers to sustain the exploitation of the El Gato well field, an important abstraction system
with a summary yield of about 3 m3/s.

1
After these results the application of isotope techniques to determine the mean residence time of ground
waters involved in the supply of the Capital city of Cuba was extended to the involved aquifers. The main
objective of those studies [6-12] was to determine if the same problem detected in Jaruco-Aguacate basin
was eventually a more generalized hydrogeological problem and, in that case, to identify the corrective
measures to sustain the exploitation of those aquifers. The contiguous Vento basin was selected because
of its similar geologic, hydrogeologic and geomorphologic structure.

FIG. 1. Vento Basin showing sampling points and reference locations (after [10]).

However, a more general conclusion is derived from the results obtained in the last years in both poljes
relative to the mixing hydrodynamics in karst aquifers were flow stratification is associated with cave
levels thus allowing that under certain conditions, old ground waters prevail in the mixing allowing
important losses of ground water reserves that are not replenished during the present hydrological cycle.

Overexploitation of groundwater is a primary focus of attention for decision makers. The effect of
groundwater resources depletion on food security and health is of particular importance in islands and
small islands dependent primarily of rain to recharge the aquifers. In karst regions were surface waters
does not exist the problem becomes strong dependent both from the adequate managing of the
groundwater resource and of the availability of recharge water.

The effect of the sustained droughts experienced by the Humid Tropics in the last years has largely
affected the availability of drinking water of adequate quality. Some discussions around the problem
focused their attention whether or not those droughts are random phenomena or are linked with climate
change. Despite theoretical considerations, the onset of old waters not linked with the present
hydrological cycle clearly shows that those aquifers are becoming more and more stressed by
overexploitation. Records show that in several aquifers of Western Cuba, old waters become
systematically prevalent in the water mixture. When they dominate the mixture, groundwater mining is an
irreversible result. A recent review of the capability of isotope techniques for groundwater exploitation
studies is due to Seiler [13].

2. The differentiated response of karst aquifers to recharge and overexploitation

It is supposed that when more water than that entering the aquifer is extracted, the available resources
become systematically exhausted or its quality deteriorated, but the concept is incomplete because it is
known that overexploitation can occur even when recharge is greater than abstraction because of bad
management practices. A more general agreement states that overexploitation of groundwater takes place
when a limit in volume, yield or water quality is surpassed. Therefore beyond those limits, some side
effects -usually ambiguously defined as “undesirable affects”- due to overexploitation appear, e.g:
systematic groundwater level decline that in turn causes the exhaustion of springs and the abandon of
water wells; loss of water quality and, in turn, contamination of the productive aquifer and subsidence.

2
For water supply purposes, this limit is often called “safe yield” or “groundwater resource”.
Conventionally overexploitation should take place when this boundary is surpassed, but experience show
that sometimes water is abstracted beyond this limit and no side “undesirable” effects become apparent.
Therefore the concept of overexploitation which is important to know with the kind of problem managers
are dealing with, falls into a category of ambiguous definitions, v.gr.: the limit established by the “safe
yield” and the safe yield itself; the limit established by the managing practices and the limit established by
the water quality that is particularly needed.

The safe yield of the aquifers is defined as the quantity of water that can be extracted annually of an
aquifer without producing undesirable effects [14]. As Todd points out, this intuitive definition is more
complex of that than at first sight it seems, since “there can be more than an undesired result effect … the
safe may be limited to an amount less than the net amount of water supplied to the basin and that the safe
yield can vary as the conditions governing it vary” [14:201]. The concept is completed when he points
out that "the safe yield cannot exceed the long time mean annual water supply to the basin…Extractions
exceeding this supply must come from storage within the aquifer”. And continues “… in any one year the
draft can exceed the recharge without causing permanent depletion. But on a long-term basis, when
series of wet and dry years would tend to average out, the draft becomes an overdraft if the mean supply
is exceeded” [14:201].

But as Adams and MacDonald [15] have pointed out, “certain aquifers are more susceptible to
overexploitation than others –equally certain managing practices are conducive to overexploitation”.
Karst aquifers, particularly those of the Humid Tropics fall in this category because the particular way in
which groundwater flow is organized in them.

In a regional karst aquifer system groundwater converging from different local and subregional flow
systems do not necessarily contributes continuously to ground water resources (Fig. 2). This singularity
means that different parts of the aquifer behave differently in time and space. In fact, according to the
degree of karst development, its distribution within the aquifer, its relation with local erosion base levels
and the way they module the recharge inputs, karst aquifers show very complex hydrodynamic responses.
Under certain boundary conditions some of the local flow systems and even its associated epikarst could
seasonally or inter annually become saturated by water or be completely dry. This changing behavior is
not necessarily reflected in the yield of springs or in the water level decline. The generally big
fluctuations recorded in groundwater levels -sometimes associated with the effect of hurricanes or heavy
rains- of large regional karst flow systems masks the actual behavior of the aquifer. Some cave levels or
local flow systems become hydrological active by several months or years leading to an erroneous
assessment of groundwater reserves and therefore to the establishment of an abstraction plan based on an
erroneous safe yield estimation that eventually could lead to overexploitation.

FIG. 2. Recharge patterns of a karst aquifer (slightly modified after [16]).

3
On the other hand, groundwater abstraction could isolate local flow systems interrupting its contribution
to major systems that could become dried. Sometimes an active epikarst could be completely drained
producing a local and sometimes abrupt lack of water supply. River runoff could be shortened and even
interrupted when feed by karst springs (concentrate or diffuse) that become dry when their local flow
systems are drained. The inverse is also true and some episodic rivers use to flow only when their local
flow systems become hydrological active.

The typical heterogeneity of the system of collectors and conducts of ground waters and the differences in
the natural recharge patterns in karst systems is the main cause of the differentiated space and time
distribution of the replenishment of ground water resources. In fact, two extreme recharge patterns are
recognized in karst systems. A rapid, fast concentrated recharge along vertical shafts and open cracks in
bare karst areas and a slow, diffuse, recharge trough joints and soil cover in buried karst systems. This
means that different recharge rates exist for a same recharge event. Even more, different arrival times to
the outlet are then recognized depending on the degree of water mixing. In turn this allows to different
isotopic behavior. This difference is the cause that in particular flow domain water with different age
coexists.

The sustained extraction of ground waters belonging to the current hydrological cycle causes, inexorably,
the exhaustion of the available resources and this relation could be rigorously determined with the use of
the proper isotopic tracers. In general, the variable "residence time" is not taken into account in the water
balance and in the assessment of groundwater resources. But the abstraction of “old” ground waters or the
exploitation of water of very slow replenishment leads to the sometimes very fast exhaustion of the
resources, -particularly in karst aquifers-.

Remediation works like artificial recharge, the protection against the contamination of the water supply
wells and springs or the deep injection of waste liquids are other aspects in which the residence time is
fundamental and, neither, regrettably, it is usually considered. As a basic principle it is highly encouraged
that waters not participating of the current hydrological cycle should not be exploited in a way to
guarantee its replenishment. However, the term “current hydrological cycle” is ambiguous by nature and
has to be more precisely defined. Anyway, the isotopic techniques provide appropriate resources to know
the residence time of ground waters. Quantifying this variable provides criteria to adopt appropriate
management practices without damaging the aquifer by the exhaustion of its resources or the artificial
recharge of contaminated waters [17-19].

Final remarks

The Tritium based isotopic balance for the different karst flow systems evidenced the overexploitation of
the ground waters. As an indicator of groundwater overexploitation in two Cuban karst aquifers the
following behavior could be observed:
1. The isotopic balance of the ground water showed that in the discharge area and in some points of
the aquifer converge waters with different residence time, indicating a stratification of the aquifer
system associated to the development of different cave levels (Figs. 3-4).
2. During the dry season in certain observation boreholes and at the system’s outlet conveys waters
with no Tritium activity. Is highly probable that these waters are not linked with the natural
replenishment associated to the current hydrological cycle or at least that takes place in the last
fifty years. Model shows the best fit for waters of 100 years of residence time.
3. During the dry season the exploitation of the eventual volumes that might be recharged associated
to the “cold fronts” rains does not reach the aquifer, limiting or impeding its the natural regulation
and, in turn, contributing to the exhaustion of the aquifer because of the systematical lack of
replenishment.

4
 4. In some cases, the difference among transit times varies from three months in the upper part of
the flow system to 100 years in the lowest levels. The exploitation of these waters during the dry
season represents a constant reduction of water resources.
5. Because of the nature of the integration of the cavernous system a good mixing of waters exists.
The fact that successive events of recharge are not consecutively identified is a good evidence of
it. This that recharge waters arrive to the outlet and to other sampling points in a differentiated
way, some first one than others, but without following a strict chronological order, in what seems
to be a typical feature of the Cuban karst systems.
6. Most recharge events that possibly can be identified with a well-known Tritium input in rain
waters basically takes place during the rainy period (May to October). This fact confirms that
these karst system receives some fresh recharge annually but not in the whole extension of the
aquifer.
7. The isotopic results show that, at least during a part of the year losses by evaporation take place
in several parts of the aquifer, mainly close to the outlet or in karst depressions where
groundwater outcrops. Therefore some of these karst features behaves as points of loss of
groundwater and not of recharge as intuitively can be supposed.

FIG. 3. Cave levels identified at the Vento Basin

FIG. 4. Wells drilled at different depths intersect different cave levels and provide differentiated
information on groundwater residence time.

Acknowledgements. The author is particularly indebted to his former Professors at the 1991 IAEA´s Training
Group in Isotope Hydrology: R. Gonfiantini. D. Louvat, K. Rozanski, L. Araguas-Araguas, A. Plata-Bedmar, P.
Maloszewski, W. Stchler, D.T. Dubinchuk, Y. Yurtsever, J.Ch. Fontes and K. Frohlich. P. Maloszewski also
provided updated versions of the modeling software. Special thanks to those colleagues who shared the field work
and data interpretation in the studied aquifers: C. Dapeña and H. Panarello (INGEIS-Argentina), J.L. Peralta, J.
Carrazana, R. Gil, M. Peña, D. Leyva and I. Fernández (CPHR-Cuba), M. Guerra and M.Pin (formerly at INRH-
Cuba). IAEA´s support for this presentation is also acknowledged as well as the friendly cooperation of its staff
members C. Devia-Torres and L. Araguas-Araguas. Cooperation of my wife Ana, was permanent in field sessions
and data processing.

5
 REFERENCES

[1] MALOSZEWSKI, P., A. ZUBER (1996): Lumped parameter models for interpretation of
environmental tracer data. Manual on Mathematical Models in Isotope Hydrology. IAEA-
TECDOC 910, IAEA, Vienna, : 9-58
[2] MALOSZEWSKI, P., A. ZUBER (s/f):Manual on lumped parameter models used for the
interpretation of environmental tracer data in groundwater.
[3] ZUBER, A., P.MALOSZEWSKI (2000): Lumped parameter models. Environmental Isotopes in
the Hydrological Cycle. Vol. VI, Modelling. IAEA-UNESCO, Paris, Tech. Doc.:5-36
[4] MOLERIO LEÓN, LESLIE F.; P. MALOSZEWSKI; M.G. GUERRA OLIVA; O. A.
REGALADO; D. M. ARELLANO ACOSTA; C. MARCH DELGADO & K. DEL ROSARIO
(1993): Dinámica del Flujo Regional en el Sistema Cársico Jaruco-Aguacate, Cuba. in/ Estudios
de Hidrología Isotópica en América Latina 1994, IAEA TECDOC-835, Viena, :139-1
[5] MOLERIO LEÓN, L.F., P. MALOSZEWSKI, M.G. GUERRA OLIVA, D.M. ARELLANO, K.
DEL ROSARIO (2002): Hidrodinámica isotópica de los sistemas acuíferos Jaruco y Aguacate,
Cuba. Ing. Hidr. y Ambiental, La Habana, XXIII (2):3-9
[6] SEILER, K.-P. (2001): Volume V: Man´s impact on groundwater systems. In W.G. Mook (Ed.):
Environmental isotopes in the hydrological cycle. Principles and applications. IHP-V. Tech. Doc.
Hydrology, 39, UNESCO, Paris, 104:
[7] MOLERIO LEÓN, L.F. (2006): Isotopic evidence of the overexploitation of karst aquifers. in:
Demuth, S., A. Gustard, E. Planos, F. Scatena, E. Servat (Eds.) Climate Variability and Change:
Hydrological Impacts. IAHS Publ. 308, Wallingford.Climate: 629-634
[8] PERALTA VITAL, J.L., R.G. CASTILLO, D.L. BOMBUSE, L. MOLERIO, M. PIN (2005):
Caracterización físico química del agua subterránea de la Cuenca Almendares-Vento.
Contribución a la Educación y la Protección Ambiental. La Habana, Vol. 6. :Q-16-Q23
[9] PERALTA VITAL, J.L., L.F. MOLERIO, R. GIL, M. PIN, D. LEYVA (2005): Empleo de
herramientas geomatemáticas para el diseño de una red de monitoreo radiosotópico en el
modelaje hidrodinámico de un recurso hídrico complejo. Contribución a la Educación y la
Protección Ambiental. La Habana, Vol. 6. :29G-39G
[10] MOLERIO LEÓN, L.F. (2005): Balance isotópico de tritio y rendimiento seguro de acuíferos
cársicos. Resumen. RH-035. Convención Internacional de Medio Ambiente.
[11] DAPEÑA, C., J.L. PERALTA VITAL, R. GIL CASTILLO, D. LEYVA BOMBUSE, H.O.
PANARELLO, I. M. FERNÁNDEZ GÓMEZ, L.F. MOLERIO LEÓN, M. PIN (2006):
Caracterización isotópica de la Cuenca kárstica Almendares-Vento, Cuba. Resultados
Preliminares. XI Congr. Geol.. Chileno, Antofagasta, Actas. Vol. 2, Simp. Hidrogeol,: 607-610
[12] PERALTA VITAL, J.L., R. GIL CASTILLO, L.F. MOLERIO LEÓN, D. LEYVA BOMBUSE,
C. DAPEÑA, H. O. PANARELLO, J. CARRAZANA GONZÁLEZ, M. PIN (2006): Uso de la
hidrología isotópica en la evaluación de una importante cuenca cárstica cubana, para la gestión
sostenible de sus recursos hídricos. VIII Congr. Latinoamericano de Hidrología Subterránea,
Asunción, Paraguay, Septiembre,: 1-22
[13] MOLERIO LEÓN, L.F., J.C. TORRES RODRÍGUEZ (2002): Métodos Geomatemáticos de
diseño y optimización de redes de monitoreo de aguas subterráneas. Ing. Hidr. y Ambiental, La
Habana, XXIII (2):38-45
[14] TODD, D.K. (1970): Ground Water Hydrology. Edic. Revolucionaria, La Habana, Cuba, 336:
[15] ADAMS, B & A. MACDONALD (1995): Overexploitation of Aquifers. Final Report. Tech.
Rep. WC/95/3, British Geol..Surv, Keyworth, Nottingham, 53:
[16] MANGIN, A. (1975): Contribution a létude hydrodynamique des aquifers karstiques. Thése.
CNRS: Moulis: 356p
[17] MALOSZEWSKI, P., A. ZUBER (1990): Mathematical modelling of tracer behaviour in short
term experiments in fissured rocks. Water Resourc.Res. 26(7):1517-152

6
[18] MALOSZEWSKI, P., A. ZUBER (1992): On the calibration and validation of mathematical
models for the interpretation of tracer analysis in groundwater. Adv. Water Resourc. 15:47-62
[19] MALOSZEWSKI, P., A. ZUBER (2004): Manual of lumped parameter models used for the
interpretation of environmental tracer data in groundwaters. Unpublished Report. GSF-Institute of
Hydrology, Nehuerberg.