Self-purification capability of underground water courses in the humid

tropics: Results of a tracing experiment at the Gran Caverna de Santo

Tomás, Cuba
L.F. Molerio León
CESIGMA, S.A.,
P.O. Box 6219, CP 10600,
Habana 6, Ciudad de La Habana, Cuba
Email: leslie@cesigma.com.cu NH8.03. Natural and anthropogenic hazards in karst areas
EGU2007-A-01841
H. Farfán González
Escuela Nacional de Espeleología,
El Moncada, Viñales, Pinar del Río, Cuba
E-mail: artibeus@geo-uh.cu

M. Parise
National Research Council, IRPI, Bari, Italy
Gruppo Puglia Grotte, Castellana-Grotte (BA), Italy
E.mail: m.parise@ba.irpi.cnr.it

C. Aldana Vilas
Escuela Nacional de Espeleología,
El Moncada, Viñales, Pinar del Río, Cuba
E-mail: artibeus@geo-uh.cu

Abstract. The Gran Caverna de Santo Tomás, 210 Km West of the City of Havana, capital of Cuba, is the biggest
cave system in the country, with almost 47 km of communicated underground galleries and several dozens of non
connected caves. The Gran Caverna de Santo Tomás is developed in the Sierra de Quemado where five small
superficial basins converge (Santo Tomás, Bolo, Peñate, Arroyo de La Tierra y Los Cerritos) entering the hills
through individual caves in the Eastern slope of the calcareous mountains. The town of El Moncada is located
upstream these small basins and discharges its untreated wastewaters to the rivers before they reach the
underground cave system. Water retention and groundwater velocity and mixing were evaluated after a tracer
test conducted by one of the authors of this contribution. Water sampling in several input-output stations allowed
the determination of the self –purification degree and the coefficient of oxygen consumption in this unconfined
conduit flow system. The ecohydrologycal vulnerability and the anthropogenic hazards to human health of the
communities living at the discharge slope of the Sierra de Quemado have been evaluated after this experiment.
The hydrological behavior of a system where flash floods linked to heavy and/ or hurricane rains can modify
groundwater flow and hence the retention and self – purification capability of the system is also discussed in this
paper.

INTRODUCTION
One of the major concerns in the sustainable management of the karst territory of the Sierra de
Quemado (17 km West Viñales y a 140 West of La Habana, Capital of Cuba, Fig. 1) is the water
quality at the outlets. This is mainly because upstream the Sierra de Quemado the village El Moncada
(population: 1600, Fig. 2) discharges its untreated domestic and agricultural waste waters to the surface
streams that enter the karst mountain. This water is also untreated used by the farmers living at the
discharge slope of the Sierra for domestic and agricultural consumption.

In 1994 a tracer experiment (MOLERIO et al., 1995a, 1995b; MOLERIO 2004, Fig. 3) showed
unsuspected hydraulic connections among the rivers entering the Sierra and several springs at the outlet
improving the general knowledge of the local hydrology (Fig. 4). Speleological knowledge of the
system is still uncompleted but has allowed the discovery of the biggest cave system of Cuba: the
Gran Caverna de Santo Tomás, with 47 km of connected passages (Fig. 5). The cave system has
been systematically explored since 1954 but few hydrological and environmental studies have been
carried in it and its surrounding basins (Fig. 6).

Fig. 1. Location Map of the Sierra de Quemado and the Gran Caverna de Santo Tomás.

Fig. 2. El Moncada village (population 1600). Partial view from the Sierra de Quemado (Photo by
M. Condis).

2
Fig. 3. Hydraulic connections established during the 1994 tracer test (shown by black arrows).

Fig. 4. Hydraulic relations of surface and goundwaters in the part of Sierra de Quemado
occupied by the Gran Caverna de Santo Tomás and its associated basins.

3
Fig. 5. General map of the Gran Caverna de Santo Tomás.

Fig. 6. The hydrologic structure of the Sierra de Quemado.

4
After a more detailed study of the chemical composition and water quality of the surface and
underground waters linked to the Sierra de Quemado (MOLERIO, 1995) a systematic monitoring was
established in the inlets and a detailed inventory of the contamination sources as well as of the water
distribution lines was carried out. The poor quality of the waters entering the Sierra via the streams
Santo Tomás and Arroyo de La Tierra was significantly highlighted because its eventual effects on
people’s health and on the underground biota. Nevertheless it was also considered that huge cave
development, where several superimposed dry and well ventilated cave levels exists, the fast flow
circuits that links the five entering streams with small or no retention in most part of the underground
passages and the differentiated development of the river and cave biota should be studied in a more
detailed way. Special attention was then focused to the eventual self purification mechanisms linked
with oxygen exchange that could take place in the cave system.

The experiment design was based on data derived from the 1994 tracer test, particularly for the mean
residence time of waters. pH, temperature, dissolved oxygen was measured in the field with calibrated
portable instruments. River discharges were measured during water sampling that also account for
major constituents. Water samples were processed for major constituents at the Central Mineral Lab of
the Ministry of Basic Industry in La Habana.

CONCEPTUAL MODEL
The conceptual model of this study was based on the spatial variation of Dissolved Oxygen (DO) of
water samples taken at several points within the cave system whose links with contaminant sources
were already known (Fig. 7). This model accounts for DO as indirect indicator for water quality
(PARSONS, J.E., D.L. THOMAS, R. L. HUFFMAN, 2004; THOUVENOT, M., T.KARVONEN,
2004; LEE, G. FRED, 2003).

Fig. 7. Conceptual model of Dissolved Oxygen depletion reactions (after LEE & JONES-LEE,
2003).

The theoretical basis for this approach has been early given by MCNEELY et al., (1979). It should be
stressed, according to them, that Oxygen is one of the gases that is found dissolved in natural surface
waters. It is moderately soluble in water. The amount of dissolved oxygen depends upon temperature,
salinity, turbulence –that allows mixing- and atmospheric pressure. The concentration of DO is subject

5
to diurnal and seasonal fluctuations due, in part, to variations of river discharge, photosynthetic activity
and temperature. Biodepletion and re-aeration processes also control the DO concentrations. The
decomposition of organic wastes and oxidation of inorganic wastes may reduce the DO levels to
concentrations of zero or approaching zero. It is known, despite most standards do not establish a
critical level for consumption waters that concentrations of less 4 mg/L are not appropriate for most
aquatic organisms. The general methodology followed was that described by PREKA & PREKA-
LIPOLD (1976). The formulations of this author and of MANCZAK (1966) were used in this study for
the degree of self purification (SA) and the Coefficient of Oxygen Consumption (k1(r)) for DO:

100(ODa − ODb )
SA =
ODa

1 ODa
k1 (r ) = log
t ODb

Where ODa and ODb are the dissolved oxygen concentrations at the inlet (a) and the outlet (b) or any
terminal station; t, holds from the time (in days) between measurements and was obtained from the
tracer test.

Seven stations were selected (Fig. 8), those in italic are referred to cave stations (see Fig. 5):
• Sumidero Santo Tomás
• Arroyo La Tierra
• Resolladero Santo Tomas
• Descarga Río Frío
• Represas Hoyo Fanía
• Lago Charco Hondo
• Lago permanente 2do cauce

Fig. 8. Surface locations of OD measurements described in this paper

6
MODEL FORMULATION
The water quality parameters for each of the inputs along with the upstream river flow are input at the
head of each reach and mixed according to the mass balance equation (ADEM, 2001a, 2001b, 2002):

Q1C1 + Q2C2 = (Q1 + Q2)C3

where:
Q1 = upstream flow (m3 /s)
Q2 = waste input flow (m3 /s)
C1 = upstream concentration (mg/L)
C2 = waste input concentration (mg/L)
C3 = instream mixed concentration (mg/L)

More complex equations describing variations in downstream constituents can be simplified to a more
mathematically convenient form using some basic assumptions. If we assume that:
• convection (flow in the river) is unidirectional, that is significant only in the X direction;
• diffusion effects are negligible; and
• there is no change in streamflow, temperature, waste loads and stream processes (i.e. steady
state prevails) then the following simplified model applies:

Where:
c = concentration of a substance, eg. dissolved oxygen - mg/L
V = Velocity in the x direction – m/s

This equation was developed by Streeter and Phelps (1925) and its solution known as the Streeter-Phelps
model. The Streeter-Phelps model accounts for only one sink of oxygen - decomposition of organic matter
(BOD) and one source - reaeration- and is applicable to rivers where these are the predominant processes.
Author like O'Connor and DiToro in 1970 added terms to the model to account for increases in dissolved
oxygen through the process of photosynthesis and decreases in dissolved oxygen through aquatic plant
respiration, sludge respiration and carbonaceous and nitrogenous biochemical oxygen demand. The
following processes are considered:
• decomposition of carbonaceous oxygen demand - CARBOD and nitrogenous oxygen demand -
NOD
• sediment oxygen demand, SOD
• algal and plant respiration, R
• photosynthetic oxygen production by plants and algae, P
• atmospheric reaeration, Ka

These processes are expressed in the following equation, which routes dissolved oxygen through the
reach (and therefore through time):

7
where
-
Kr = first order rate constant of CARBOD – day-1
L = concentration of CARBOD according to equation (3) - mg/L
Kn = first order decay rate constant of NOD – day-1
N = concentration of NOD according to equation (4) - mg/L
S = average rate of sediment oxygen demand - mg/L
R = average rate of algal and plant respiration
P = photosynthetic oxygen production rate- mg/L
-1
Ka = first order atmospheric reaeration rate constant - day
D(x,t) = dissolved oxygen deficit = (Cs - C) - mg/L
Cs = oxygen saturation - mg/L
C = oxygen concentration - mg/L

O'Connor and DiToro solved equation (6) using a Fourier series expansion for the P and R terms. Their
solution is the basis of the DOMOD3 steady state model used in the Kam River 1988 study with the P
and R terms assumed to be zero. The photosynthesis and respiration terms (P and R) have been ignored
in the equation below which is applied in the Kam R. spreadsheet version.

where:
-1
Kd = Deoxygenation rate constant for CARBOD - day
x/V = distance over velocity, or time of travel to location x . Note that the equation does not vary with
time only x (or x/V or travel time).

Temperature corrections for the de-oxygenation terms, re-aeration term and SOD will be calculated by
the following general equation:

Where:
KT = generic reaction rate (1/day) for temperature T
K20 = specific reaction rate at 20ºC (1/day)
θ = Arrhenius equation temperature constant for each parameter.
T = temperature (C)

8
The following hydraulic relationships are used in the model:

Where:
V = velocity, m/s
Q = streamflow, m3/s
D = river depth,
m a, b, c, d, e, f = Leopold-Maddock coefficients with: a + c + e = 1 and b x d x f =1.

The units for areal rate terms need some clarification. The S, P and R terms by definition are channel
bottom processes conventionally presented in units of gm/m2/day. The relationship to the volumetric
term used in the model is as follows for the SOD term, with units shown.

S (mg/L/day) = s (gm/m2/day) / Depth (m)

The implication of this is that while the SOD rate would remain constant as flow varies, the volumetric
S rate varies with depth and must be adjusted for predicting the impact of lower flows as follows:

S2 = S1* D1/D2

With subscript 1 referring to the base data and 2 referring to the prediction scenario and D referring to
depth (m). From this calculation it can be seen that lowering the flow rate (and thus the depth)
magnifies the impact of SOD by the ratio of the depth. The same unit conversions apply to the P and R
terms.

As stated in the literature, the major consumption of DO in streams occurs through the aerobic
chemical and microbial breakdown of long-chained organic molecules into simpler, stable end
products.

carbohydrate → CO2 + H2O

proteins → amino acids → ammonia → nitrite → nitrate→ sulfate and phosphate

If O2 is exhausted, aerobic decomposition ceases and further breakdown must be accomplished by

anaerobic bacteria. A moderately high DO is necessary for the maintenance of healthy aquatic

9
ecosystems. When a large amount of industrial or municipal waste enters the stream, the breakdown of
the waste may depletes O2 in the stream.

When a waste enters a stream, it becomes completely mixed with stream water after a short distance.
The assimilation of oxidizable materials begins to consume O2 and increase O2 deficit. O2 is
replenished by reaeration from the atmosphere, the rate of which depends mainly on the O2 deficit, the
width of the stream, turbulence, and water temperature.

The balance between O2 consumption and replenishment leads to a profile of net O2 deficit, which
shows characteristic dissolved-oxygen sag (Figs. 9-11).

The profile of O2 deficit is estimated by the Streeter-Phelps equation (Fig. 12);

Where
D: O2 deficit of the stream [mg/L]
Di: initial O2 deficit of the waste mixture [mg/L]
k : rate constant of O2 consumption [day-1]
L : initial BOD of the waste mixture [mg/L]
r : reaeration rate constant [day-1]
t : time from initial mixing [day]

Fig. 9. Disolved oxygen sag.

10
Fig. 10. Typical DO behavior.

Fig. 11. Evolution of DO and nutrients.

Fig. 12. Streeter and Phelps equation terms.

11
The word “initial” indicates the time when the waste is completely mixed with stream water. The initial
waste mixture may have significant O2 deficit. The rate of O2 consumption per unit BOD is dependent
on the type of waste. Many agricultural processes release wastes that can be rapidly oxidized. In
contrast, pulp wastes are assimilated slowly. Therefore, the rate constant k depends on the type of
materials in the waste. The constant r represents how fast O2 is being replenished. Both k and r are
dependent on water temperature (BENOIT, 2001).

Assuming complete and instantaneous mixing,

Lo = initial BOD of the mixture of river and wastewater (mg/L)

Lr = ultimate BOD of the river just upstream of the point of discharge (mg/L)
Lw = ultimate BOD of the wastewater (mg/L)
Qr = volumetric flow rate of the river just upstream of the discharge point (m3/s)
Qw = volumetric flow rate of wastewater (m3/s)
DO = initial oxygen deficit of the mixture of river and wastewater
DOs = saturated value of DO in water at the temperature of the river
DOw = DO in the wastewater
DOr = DO in the river just upstream of the discharge point

According to USEPA (2001) the simulation of DO dynamics is based on the oxygen deficit, which is
the difference between saturation and actual conditons. The following equation was used to calculate
DO at saturation:

Where T is temperature in degrees Kelvin, or 273.15 + T °C.

The following equation was used to simulate DO dynamics:

Where:
D = DO deficit leaving a given segment
D0 = DO deficit entering a given segment
Ka = reaeration rate
L0 = iron concentration entering a given stream segment
Kr = Kd (decay) + Ks (settling) rates
x = length of stream segment
u = flow velocity

Similarly to the dynamics of the iron loss equation, a longer distance, slower flow velocity,

12
increased iron floc settling rate, or increased iron decay rate would increase the oxygen deficit, as
would a lower reaeration rate.

Reaction rates were adjusted for temperature using the following equation:

Where 2 = 1.024 for reaeration and 1.047 for decay and settling.

RESULTS AND DISCUSSION

Table 1 shows part of the data input used in this paper. Fig. 13 shows the Schoeller diagram for these
major constituents. For location references please refer to Figs. 4 and 8.

Table 1. Average major constituents, nutrients and dissolved oxygen (in mg/L).

Hydrological
role Locations HCO3 Cl SO4 Ca Mg Na K NH4 NO2 NO3 TDS DO
INLETS Sumidero Santo Tomás 108 8 13 28.2 2.6 6.5 1.6 0.35 <0.03 <10 152 4.89
Arroyo La Tierra 221 12 10 30.0 5.5 8.9 13.1 15.5 <0.03 <10 267 0
OUTLETS Descarga Río Frio 167 8 12 46.0 3.9 6.4 1.4 0.38 <0.03 <10 229 7.9
Resolladero Santo Tomás 108 8 10 24.9 3.2 6.0 1.3 0.31 <0.03 <10 146 7.9
CAVE Lago permanente 2do cauce
STATIONS 114 6 10 32.0 5.2 2.5 1.0 <0.25 <0.03 10.9 161 1.43
Represas Hoyo Fanía
152 8 10 42.1 4.2 4.1 2.3 <0.25 <0.03 <10 208 3.8
Lago Charco Hondo
36 7 10 5.7 2.0 1.7 4.0 <0.25 0.125 <10 53.7 3.85

Fig. 13. Schoeller diagram for major constituents of waters tested in this paper.
2do Cauce
Gráfico de Schoeller
Represa- Fanía
1000 Charco Hondo
Sumidero S.Tomas
Río Frio
Resolladero
100
La Tierra
mg/L

10

1
HCO3 Cl- SO42- Ca Mg Na K

13
Fig. 14. Schoeller diagram for surface waters entering Sierra de Quemado.

Hidroquimica de las aguas de ingreso

Cerritos
10
S.Peñate
A.LaTier
Bolo
SumS.Tom
1
HCO3 Cl SO4 Ca Mg Na K
mEq/L

0.1

0.01

Point hydrogeochemistry does not differ from the generalization of MOLERIO (1995) (Fig. 14) except
for the case of Charco Hondo, not sampled before. Charco Hondo represents a point basically linked to
waters entering via infiltration and partially mixed with flooded surface waters of the Santo Tomás
river. Na and K concentrations in Arroyo la Tierra were a little higher than historically.

Water entering Sierra de Quemado via Arroyo de La Tierra receives the contribution of a a is non
persistent, non cumulative toxic substance major pollution source coming from the untreated
wastewaters of a Chicken Farm. This point source supplies the highest mineralized surface waters to
the system coming primarily from the carbonated products used for cleaning and feeding of the
chickens. In addition, the higher NH4 concentrations are recorded in these waters. This dissociated
ammonium ion NH4 represents a toxic compound entering the system that threatens freshwater aquatic
life and, because of the general alkaline environment it should be more toxic.

Arroyo de La Tierra shows the most depleted Dissolved Oxygen values of the sampled stations and, in
turn, this allows for the general picture described above. The way and intensity that this behavior affect
the cave biota has not been already studied but in the surroundings of the inlet and along the surface
water course before entering the cave no aquatic life has been recognized.

All streams entering the Sierra the Quemado and stored in the cave lakes like Represas, Charco Hondo
and Segundo Cauce shows depleted values of DO. All values are below the standard limit of 4 mg/L
DO adopted for the preservation of aquatic life.

Charco Hondo and Represas Hoyo de Fanía are basically linked with the Santo Tomás River except in
heavy rains were some back flow could be produced due to the interaction of this stream with the
runoff of El Bolo stream. But in the conditions described no back flow nor heavy rains took place.
Therefore, waters from Santo Tomás river in the line Charco Hondo-Represas became more depleted in
DO, decreasing in about 1 mg/L. In the large flow path the decomposition of organic matter continues
to proceed, depleting Oxygen in the system.

14
In the permanent lake of 2do Cauce, the situation is even worst. Waters mainly stagnant of this lake
allowed for particular table conditions of temperature and flow, increasing oxidation of the organic
matter and high consumption of oxygen. In fact the lowest values recorded within the cave system
were measured at this point. Therefore, this place constitutes a particular ecological notch of depleted
high residence time waters. No biota has been ever reported from this lake except after heavy rains
producing high floods in the cave.

DO values measured at the outlets of Rio Frio and Santo Tomás have shown unexpected results. In fact
a recovery in DO values was recorded in both monitoring stations together with practically no changes
in the concentration of NH4. Therefore, it seems that re-aeration processes are particularly important
close to the outlets of the underground streams. This process should be theoretically controlled by the
turbulence of the streams close to the discharge slope of the Sierra de Quemado and the influence of
the wide and well-ventilated underground passages in the interface atmosphere-water. The rate of re-
aeration in the first model outputs is less than 1 day from the short distance separating Charco Hondo
to the Santo Tomás resurgence.

CONCLUDING REMARKS
• Two of the five streams entering the Sierra de Quemado conveying into the Gran Caverna de
Santo Tomás, the largest cave system in Cuba, are depleted or close to depletion dissolved
oxygen.
• Depletion continues to proceed underground for most part of the underground path of te Santo
Tomás river and loose close to 1 mg/L . Underground lakes permanent or temporarily linked
with this fluvial stream are particularly depleted.
• The Arroyo de La Tierra enters the system with zero dissolved oxygen and high NH4
concentration. Nevertheless, close to the outlet, this stream and the Santo Tomas river recovers
enough oxygen to be enriched close to 7 mg/L of DO.

REFERENCES
ADEM. ALABAMA DEPARTMENT OF ENVIRONMENTAL MANAGEMENT (2001a): The
ADEM Spreadsheet Water Quality Model. Water Division – Water Quality Branch, Alabama
Department of Environmental Management, 24:

ADEM. ALABAMA DEPARTMENT OF ENVIRONMENTAL MANAGEMENT (2001b):

Spreadsheets for Water Quality-Based Npdes Permit Calculations. Updated July 2000 by Greg
Pelletier, Water Division – Water Quality Branch, Alabama Department of Environmental
Management, 19:

ADEM. ALABAMA DEPARTMENT OF ENVIRONMENTAL MANAGEMENT (2002): Water

Quality Appendix K Final TMDL Development for Chase Creek / AL06030002-190_01. Alabama
Department of Environmental Management Low Dissolved Oxygen/Organic Loading, 28:

BENOIT, A.G. (2001): The Streeter and Phelps Equation. ENV 2101 Principles of Environmental
Engineering. 2:

LEE, G.F. & A. JONES-LEE (2003): Synthesis and Discussion of Findings on the Causes and
Factors influencing Low DO in the San Joaquin River Deep Water Ship Channel near Stockton,
CA: Including 2002 Data. G. Fred Lee & Associates. Conceptual Model of DO Depletion
Reactions in the SJR DWSC. Report Submitted to SJR DO TMDL Steering Committee/Technical
Advisory Committee and CALFED Bay-Delta Program, March 2003, Californa, 284:

15
MANCZAK, H. (1966): Ocena przelegu procesu samooczisczkoi rzek skanalizowanych na
potstame kryterium tlenowego i wynikow bagad rzek Odry. Secz. Nauk Politechn, Wroclaw,
Poland (144)

MCNEELY, R. N., V. P. NEIMANIS, L. DWYER (1979): Water Quality Sourcebook. A Guide to

Water Quality Parameters. Inland Water Dir. , Ottawa, 88 p.

MOLERIO LEÓN, L.F.(1995): Regionalización Hidrogeoquímica de las Aguas Subterráneas en la

Sierra de Quemado, Pinar del Río, Cuba. Congr. Internac. LV Aniv. Soc. Espel. Cuba y Primera
Reunión Iberoamericana, La Habana,:92-93

MOLERIO LEÓN, L.F. (2004): El enlace absorción-descarga de la Gran Caverna de Santo

Tomás: evidencias derivadas de un ensayo con trazadores artificiales. Ing. Hidr. y Ambiental, La
Habana, XXV (3): 22-26

MOLERIO LEÓN, L.F.; A. MENÉNDEZ; E. FLORES, C. BUSTAMANTE & M. GUERRA (1995a):

Hidrodinámica de los Grandes Sistemas Cavernarios de Cuba Occidental. Congr. Internac. LV
Aniv. Soc. Espel. Cuba y Primera Reunión Iberoamericana, La Habana,:88-89

MOLERIO LEÓN, L.F.; C. ALDANA VILAS; E. FLORES VALDÉS; E. ROCAMORA & ANA M.
SARDIÑAS (1995b): Resultados de un Ensayo con Trazadores Artificiales en la Gran Caverna de
Santo Tomás, Pinar del Río, Cuba. Congr. Internac. LV Aniv. Soc. Espel. Cuba y Primera Reunión
Iberoamericana, La Habana,:95

PARSONS, J.E., D.L. THOMAS, R. L. HUFFMAN (2004): Agricultural Non-Point Source Water
Quality Models: Their Use And Application, Southern Cooperative Series Bulletin #398, July, 2001
(Updated July, 2004), http://www3.bae.ncsu.edu/Regional-Bulletins/Modeling-Bulletin/modeling-
bulletin.pdf.

PREKA, N., N. PREKA-LIPOLD (1976): A contribution to study of self – purification capability of

underground watercourses. In V. Yevjevich (Ed.): Karst Hydrology and Water Resources. Proc.
U.S.-Yugoslavia Symp. Dubrovnik, Vol. 2, Fort Collins, Colo, USA:719-729.

THOUVENOT, M., T. KARVONEN (2004): Description of an in-stream water quality model with
particular interest in nitrogen cycling. Water Resources Engineering, Helsinki University of
Technology, 23:

16