Professional Documents
Culture Documents
Abstract
A mathematical model is proposed to simulate brackish karstic springs. Rainfall data constitutes model input
information while output information is the discharge and the chloride concentration of the water versus time. The
model was constructed by considering the mass and mechanical energy balance on the hydrodynamic analog, which
includes three reservoirs outflowing in a tube that lies adjacent to the spring. Two reservoirs emulate the karstic system, and the third one emulates the sea. The discharge of the spring is given by the sum of the discharge of the reservoirs, and the chloride concentration by the solution of the mixing problem between the fresh and the salty water,
which exists in the tube leading to the spring. The model is applied to the spring of Almiros at Heraklion, Crete,
Greece. The agreement between model values and field measurements is very good for depletion periods and satisfactory for recharge periods.
Introduction
The term karstic spring denotes the point of water
outflow of karstic hydrogeologic reservoirs. The latter are
systems of ground water storage and transfer that comprise
a great number of tubes of various shapes and sizes. Karstic
reservoirs occur in water-soluble rocks, such as limestone.
The tubes are formed by the solvent action of water, usually
on pre-existing discontinuities. In the preceding years,
many studies with computer models have been conducted
on the evolution of a karstic system during the geologic
time (Dreybrodt 1996; Gabrovsek and Dreybrodt 2000;
Palmer 2000; Clemens et al. 1999; Ford et al. 2000). From
the hydraulic point of view, the characteristic of these reservoirs is preferential flow through the tubes and, in many
cases, turbulent flow (Dreybrodt 1988), which makes
Darcys law invalid.
608
case it gets brackish, the variation of the chloride concentration of its water against time. The model, which is fed by the
rainfall measures of the recharge area, uses parameters that
represent the properties of the karstic system, which are estimated by fitting the model to the available field measurements. Actually, the model transforms the rainfall data of the
recharge area to the hydrographs of the spring (Figure 1). It
has been applied to the Almiros spring at Heraklion, Crete,
Greece.
Hydrodynamic Model of
Brackish Karstic Springs
During the initial phase of karstic aquifer depletion,
two emptying rates can be identified: a fast one corresponding to a section of karst that is emptied quickly, and a slow
one corresponding to another section of karst that is emptied rather slowly. The first part of karst consists of a few
tubes of large diameter, and the second consists of many
tubes of small diameter. Shoeller (1967) was the first to
introduce such an idea. Recently, Worthington (1999) and
Worthington and Ford (1997) postulated for a mature karst
aquifer a low hydraulic conductivity related to the rock
matrix and a high one related to large conduits draining the
system. Eisenlohr et al. (1997), on the other hand, speak of
conduits that are immersed in a low hydraulic conductivity
medium with slow flow velocities and discharge at one or
more discrete springs. Thus, it can be argued that two different karstic subsystems, according to hydrogeologic
behavior, feed the karstic springs. If the later happen to be
brackish, they are also fed by the sea. In Figure 2, a hydrodynamic model is presented that is constructed according to
these ideas. The reservoirs karst 1 and karst 2 represent
the karstic subsystems. while the other one represents the
sea. The sea reservoir is of infinite size. The discharge of
the spring at any time is given by the sum of the discharge
of the reservoirs. The chloride concentration of the spring
water is a result of the mixing of fresh water of the karstic
reservoirs with the salty water of the sea. This happens in
the tube, which brings the water to the spring.
Mathematical Model
Spring Discharge Calculation
In order to develop a mathematical model based on the
previously mentioned karstic spring hydrodynamic model,
the mass and energy balances were considered in the reservoirs, which emulate the two karstic subsystems and the
sea. Each reservoir is considered as a whole in which the
mass and energy conservation principle is applied. The pro-
609
Table 1
System Equations
Karstic Subsystems Equations
sS1
sS1H
.
dH
1 Q2 2 PS1 5 0
dt
.
dH
Q 32
Q2 2
1
2
HP
S
1
L
c
d Q2 5 0
1
dt
2gS22
sS1
P
Q 23
1 3 2 H 5 0
g
2gS23
s
p3 5 wgaH 1 h 1 Hb 2 La Q 22
Variables
Q2
Q3
H
P3
Parameters
P
s
S1
g
S2
L
S3
H
La
3.
4.
Table 2
Almiros Spring Model Equations
Tripolis Limestone
SS
.
1 Q 2 (1 2 f)PS 5 0
t
sSHt
Ht 1t 2 Ht
t
Q 3
2gS2
Q 2
.
2 Ht 1 t (1 2 f)PS 1 L a
b Q 5 0
sS
Ionian Zone Limestone
sS
Ht 1 t 2 Ht
t
.
.
1 Q 2 ftPS 2 PS 5 0
sSHt
Ht 1 t 2 Ht
Ht 1 t 2 Ht
t
Q 3
2gS2
Q 2
.
.
2 Ht 1 t afPS 1 PS 1 L a
b Q 5 0
sS
Sea
Q2
2gS2
P3
g
2 H 5 0
P3 5 wgaH 1 h 1
Ht 1 t b 2 LaQ2
611
was not taken into account either. Time interval for the
model calculations was one day (t = 1 day).
The model fitting has been done with the least squares
method using commercially available optimization software. According to this method, the optimization program
was tasked to estimate the parameter values so that the
mathematical expression
2
2
a 3 1 QMt 2 QFt 2 4 1 a 3 1 CMt 2 CFt 2 4
D
t51
t51
Table 3
Parameters, Variables, and Output Information of Almiros Spring Model
Parameters Symbol Notation
Indices Notation
f
S,
S, ,
H
h
s
L
L
w
g
Daily rainfall
Infiltation coefficient
Percentage of Tripolis water that feeds Ionian limestone directly
Recharge area
Time Interval
Qt+t
Qt+t
Q
Ht+t
Ht+t
p3
t = 1 day
Q 5 Qt 1 t 1 Qt 1 t 1 Q*
ppmCl 2 5
CQ 1 w Cw (Qt 1 t 1 Qt 1 t )
Q 1 w (Qt 1 t 1 Qt 1 t )
*Seawater discharge is accepted positive if the sea feeds spring reservoir or negative if the sea is fed by spring reservoir.
612
Table 4
Rainfall 19951997 (mm)
Sep Oct Nov Dec Jan Feb Mar Apr May June July Aug
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
0
1.4
0
3.4
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
0
3
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1 12.6
0
11
0
0
0
0
8.4
0
4
0.5
0
0
0
2.7
1.4
0
6.4
4
3
7
0
1
0
0
0
0
0
0
0
0
0
0
1
43.6
26
4.2
0
0
0
0
0
0
0
0
0
0
21
0
0
0.5
20
0
9
9.4
0
2
0
0
3
0
1.4
2
0.4
0
0
0
0
0
0
0
0
2
3.6
0
15.6 38
4.2
0
20.7 0
47.2 0
12
2
0
0
3
15
4
30
14.4 22
5.4 13
0
3
0
0
18.2 9
0
0
35.4 1
5.3
0
20.4 1
21.1 0
0
3.4
0
0
0
0
1
0
16 1.6
0
0
0
0
0
0
0
0
0
10
0
21
0
20
34.5
0
0
0
12.3
41.4
0
0
8.3
5
0
0
0
25.8
0
1
0
0
6.4
6.2
4
5
1
0
0
0
0
0
0
0
1
0
0
0
0
0
0
6.6
0
0
0
4.5
17
8.4
0
0
0
0
1
0
11.6
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
5
6
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Sep Oct Nov Dec Jan Feb Mar Apr May June July Aug
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
43.2
0
0
0
0
0
0
0
0
0
0
0
0
17.3
13.3 12
0
0
0
0
5.7 14 3.6
0
0 63.3
2
0
11
0
5
16
0
0 60.4 0
5.6
0
0
0 31.6 0
4
0
0
0
0
0
0
0
0
0
0
9
0
0
0
0
0
0
44
0
0
0
0
0
6.4
0
0
0
0
0
7
3
12.4 0
8
0
0 79.7
0
0
10
3
0 77.7
0
0
1
50
0
24
0
0
0
0
0
0
0
0
0
38
0
0
0
0
0
27
0
0
0
0
0
0
0
0
0
00
0
0
14
0
0
0
0
0
43
0
0
0
0
0
14
0
3.8
0
0
0 16.4 6
1.5
0
0
0
0
19
4.8
0
0
0
0
3
0
0
0
0
0
3
20.6 0
0
0
0
2
58.4 0
0
0
0
0
9
6
8.6
1
0
11
0
13
2
9.1
0
0
0
0
0
0
0
0
6.8
0
0
0
0
0
0
7
0
0
16
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
3
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Figure 6. Comparison between Model calculated discharge values and field measurements. AB, DE: Recharge periods, CD,
EF: Depletion periods.
613
2.
Conclusions
With the proposed model it has been possible to simulate a karstic spring even if it gets brackish. The advantages
of the model are:
1.
2.
3.
4.
5.
Acknowledgments
The authors would like to thank Editor Mary P. Anderson, who handled the review, Ken Hardcastle, and two
anonymous reviewers for their valuable suggestions.
References
Angelini, P., and W. Dragoni. 1997. Problem of modeling limestone springs: The case of Bagnara (North Apennines, Italy).
Ground Water 35, no. 5: 612618.
Bezes, C. 1976. Contribution a la modelisation des systemes
aquiferes karstiques; establissement du modele BEMER;
son application a 4 systemes karstiques du Midi de la
France. These 3e cycle. Universite de Montpellier, Memoires du CERGH.
Breznic, M. 1973. The origin of brackish karstic springs and their
development. Col. Razpr. In Por. 16 Knjig, 83186.
Castany, G. 1967. Traite Pratique des Eaux Souterraines. Paris:
Dunod.
Figure 7. Comparison between Model calculated water chloride concentration values and field measurements. AB, EF:
Recharge periods, CD, GH: Depletion periods.
614
Table 5
Discharge 19951997 (m3/sec)
Sep Oct Nov Dec Jan Feb Mar Apr May June July Aug
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
4.0
3.7
3.5
4.2
4.5
3.7
3.5
5.6
5.3
4.7
5.0
5.0
5.7
7.6
7.1
6.4
5.8
7.8
12.8
10.0
7.6
6.1
5.0
4.5
4.2
4.0
3.7
3.6
3.5
5.4
15.1
14.0
17.4
22.4
20.9
15.9
14.0
11.9
14.8
13.7
12.9
12.3
14.4
16.8
19.9
21.2
21.1
20.6
18.0
15.5
14.6
15.7
16.6
15.1
16.6
20.7
20.5
14.6
18.0
18.4
24.4
25.0
28.3
23.4
20.9
20.7
19.7
20.3
22.1
24.2
24.8
23.0
19.9
19.3
17.2
16.1
15.1
14.2
13.5
12.7
11.9
11.7
11.2
10.8
10.3
8.8
8.3
8.1
8.6
22.5
21.1
17.2
17.1
17.0
23.4
22.3
18.4
16.8
16.5
15.9
14.6
16.1
18.0
20.3
19.9
17.8
15.9
14.0
13.1
12.9
12.9
12.6
12.3
11.7
11.0
10.6
10.6
10.5
9.8
9.3
9.1
8.8
8.7
8.5
8.3
8.1
7.8
7.6
7.4
7.4
7.4
7.5
7.6
7.8
7.8
7.8
7.8
7.6
7.4
7.4
7.4
7.4
7.2
7.2
7.2
7.1
7.1
7.1
7.1
7.1
7.1
7.1
7.0
6.9
6.9
6.9
6.8
6.8
6.8
6.8
6.8
6.8
6.8
6.8
6.7
6.6
6.5
6.4
6.4
6.4
6.4
6.4
6.4
6.4
6.4
6.4
6.3
6.2
6.1
6.0
5.6
5.6
5.0
4.5
Sep Oct Nov Dec Jan Feb Mar Apr May June July Aug
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
4.0
5.2
7.9
5.6
5.0
4.5
4.1
4.0
4.5
6.6
6.1
5.3
4.8
4.5
4.2
4.0
4.0
4.0
3.8
3.8
3.8
3.8
3.8
3.8
3.8
3.8
3.8
3.8
3.8
3.8
3.8
3.8
3.8
3.8
3.8
3.8
3.8
3.8
3.8
7.8
15.7
18.5
27.8
21.7
20.9
14.8
10.3
9.5
8.8
8.6
8.6
8.3
8.2
7.6
6.9
6.6
6.6
6.1
5.8
5.6
5.3
5.0
5.0
5.0
5.0
19.5
32.4
24.4
17.6
15.1
11.6
11.3
10.0
9.3
8.8
7.9
7.9
7.9
7.8
7.6
8.8
14.0
15.1
25.6
32.4
25.9
21.7
15.4
14.0
12.7
11.7
10.5
9.5
9.0
8.6
8.3
7.9
7.6
7.4
7.1
6.9
6.9
6.9
6.9
6.9
6.8
6.4
13.5
16.5
18.4
16.3
13.7
12.3
11.7
11.5
11.3
11.0
13.3
19.5
19.5
20.7
18.4
16.1
14.4
13.5
13.1
12.8
12.8
Jan
13.1
23.2
40.3
34.3
30.6
22.4
18.4
16.1
14.2
13.7
25.6
50.2
52.1
34.4
35.6
30.4
31.6
28.0
23.0
20.3
18.6
17.6
16.6
15.8
15.0
13.8
13.6
13.5
13.5
13.3
13.3
13.3
12.9
12.2
11.7
11.7
12.3
13.2
13.0
12.7
12.5
12.5
12.4
12.2
12.0
11.7
11.0
10.5
9.9
9.7
9.4
9.2
9.2
9.1
8.9
8.6
8.4
8.2
8.0
7.8
7.7
7.6
7.6
7.5
7.3
7.1
6.9
6.8
6.7
6.6
6.6
6.6
6.5
6.5
6.4
6.4
6.3
6.3
6.3
6.2
6.2
6.2
6.2
6.2
6.2
6.2
6.2
6.2
6.2
6.2
6.2
6.2
6.1
5.9
5.3
Table 6
Chloride Concentration 19951997 (mm)
Sep
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Oct
Nov Dec
Jan
791
159
4792
852
4940
2130
5076
4757
71
3603
2378
142
4573
71
1100
4704
3248
1917
5083
722
4792
603
2875
Sep
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Oct
Nov Dec
3585
3658
2502
576
3674
1331
284
3639
4250
124
4045
781
4508
35
3816
3514
2511
2982
2236
497
4440
142
923
1065
615
Table 7
Parameter Estimates by Model Fitting
Tripolis Zone Limestone
Recharge area St (m2)
Percentage of water that feeds
Ionian limestone directly f
Cross section of reservoir outlet tube S (m2)
Energy loss coefficient Lt (m-1 sec-2)
Infiltration coefficient of recharge area
Coefficient of storage s
2.2 3 108
0.1
1
1
0.6
0.0075
3.1 3 108
1
1
0.6
1.25 3 105
0.0195
Sea
Cross section of reservoir outlet tube S (m2)
Pressure loss coefficient of the sea tube
outlet region La (kgr m-7)
Depth of seawater outflow tube outlet point
H (m)
0.06
10
850
(A1)
EE
Table 8
Standard Deviation Between Model Calculated Values and Field Measurements
Recharge
Period
19951996
Depletion
Period
19961997
Depletion
Period
19951996
Depletion
Period
19961997
Total
19951997
Discharge (m3/sec)
3.37
6.24
0.38
0.51
3.57
Chloride
concentration (ppm)
1035
1356
200
309
1160
616
sS1
or
sS1
d
d
d
rdV 5 9 rd 1 sVs 2 5 3 srwS1dH (A2)
dt 9
dt
dt
OE
OE
where s is the storage coefficient of the aquifer, Vs the saturated volume of the rock, w the fresh water density, S1 the
transverse inlet or outlet surface of the reservoir, and H the
level of water in the reservoir (Figure 3). If the aquifer is
unconfined, level (H) corresponds to the mean level of
water in the entire aquifer. In the case of confined aquifers,
the variable H does not have a direct physical meaning, but
if this variable is multiplied with the coefficient s/, where
is the specific yield, it corresponds to the mean water
level of an equivalent unconfined aquifer, which has the
same volume of water.
The second term of Equation A1 becomes
.
dH
1 Q2 2 gPS1 5 0
dt
(A4)
u2
u2
p
d
a
a
1
gzbrdV
1
" 2 1 gz 1 b
dt 9 2
EE
OE
.
(u n)dS 1 Sw L 5 0
(A5)
S1
EE
1 6 w (u2 ? n 2 )dS2
S2
dH
1 u2S2 2 u1S1 5 0
dt
u2
d
a 1 gzbdV
9
dt
2
(u n) dS 52 w U1dS1 1 rwu2dS2
6
6
S1
EE
OE
S2
d
d
5
gzdV 5
sgHwS1dH
9
dt
dt 3
0
OE
H
d
csgwS1 3 HdHd
dt
0
(A3)
5 sgwS1
1 dH
dH
5 sgwS1H
2 dt
dt
(A6)
d
sr S dH 1 rwu2S2 2 rwu1S1 5 0
dt 3 w 1
0
u2
p
a
" 2 1 gz 1 b(u n)dS
EE
617
u21
p
5 6a
gz1 1 1 b w (u1 n1 )dS1
2
w
sS1H
dH
u2
w
1 2 Q2 2 HQ1 1 L Q2 5 0
g
dt
2g
s1
u2
P2
1 6 a 2 1 gz2 1 bw (u2 n2 )dS2
w
2
S2
u2
P
u2
5 6 a 2 1 gz2 1 2 u2 dS2 26 a 1 1 gz1
w
2
2
S2
S1
sS1H
dH
u2
1 2 Q2 2 HQ1 1 Lu2Q2 5 0
dt
2g
p
u22
u2
" a 2 1 gz 1 b(un)dA 5 6 2 w u2dS2
s2
sS1H
dH
u2
Q 2
1 2 Q2 2 HQ1 1 L c 2 d Q2 5 0
dt
2g
sS1
EE
or
2 6 gz1wu1dS1
sS1H
s2
#
dH
Q 32
Q2 2
1
d Q2 5 0 (A9)
2 2 HPS1 1 L c
dt
2gS2
sS1
#
Q2
,Q1 5 PS1. The parameters S2 and L are
S2
estimated by the fitting of the model.
where u2 5
EE
u2
u2
5 w 2 u2S2 2 gz1wu1S1 5 w 2 Q2 2 gz1wQ1
2
2
(A7)
OE
p
u2
.
1 " a 1 gz 1 b(u n)dS 1 wL 5 0
dH u2
#
#
sgwS1H +w 2 Q2 2 gz1wQ1+wL 1 wS 5 0 (A8)
dt
2
#
z1 5 H,wL 5 wLQ2
EE
dH u22
+ Q 2 gHwQ1+wLwQ2 = 0
dt w 2 2
(A10)
OE
EE
618
(A5)
S3
P
u2
P
1 3 b(u3 n3 )dS3 1 6 a 4 1 gz4 1 4 b
a
S4 2
Q 23
p
1 3 2 H 5 0
g
2gS23
(A12)
a (u4 n4 )dS4
u4, p4 negligible
z3 = 0
Uniform flow distributions on the inlet-outlet surfaces.
p
u2
" a 2 1 gz 1 b(u n)dA
EE
5a
u23 p3
1 bQ3 1 gHaQ4
2
(A11)
u23
p
#
#
1 3 bQ3 1 gHaQ4 1 wL 1 wS 5 0
2
#
It has been assumed that wL 5 0. Therefore,
a
u23
p
1 3 bQ3 1 gHaQ4 5 0
2
u23
p
1 3 b 2 gH 5 0
2
(A13)
(A14)
p
u23
1 3 b Q3 1 gH aQ4 5 0
u23
p
1 3 b 2 gHq 5 0
619