Beruflich Dokumente
Kultur Dokumente
Address:
Research Division on Applied Geology
Faculty of Earth Science and Technology
Institut Teknologi Bandung
Basic Science building 3th floor
Jl. Ganesa no 10, Bandung - 40132
Title of paper:
Outlining Hydrogeological System using Multivariate Analysis on Groundwater Quality at Mt.
Ciremai, West Java, Indonesia
Volcanic slopes are important sources of water. Groundwater observation in east slope of Ciremai was
conducted in period of dry season of May until June 2006 on 119 springs. Fourteen variables was
measured: elevation, spring discharge, TDS, EC, pH, T water, T air, calcium, magnesium, chloride,
sodium, sulphate, potassium, bicarbonate.
Cluster analysis has successfully extracted 3 clusters: cluster 1 (112 obs), cluster 2 (5 obs), cluster 3
(2 obs). The hydrogeological schematization has been constructed based on interpretations of 3
clusters. There are 3 hydrogeological systems: 1) Hydrogeological system 1 is characterized by:
cluster 1, heterogeneous data, normal water temperature, TDS, EC, and major elements
concentrations, and many chemical influences. 2) Hydrogeological system 2 is characterized by:
cluster 2, homogeneous data, high water temperature, TDS, EC, the result of interaction between
normal meteoric water and hot water of volcanic origin. 3) Hydrogeological system 3 is characterized
by: cluster 3, high water temperature, TDS, and EC with homogeneous characters, deep flow system,
and interaction with sedimentary layers of Fm. Kaliwungu.
Key words:
1 Strato volcanoes
2 Hydrogeological Tracer
3 Multivariate analysis
Outlining Hydrogeological System using Multivariate Analysis on Groundwater Quality
at Mt. Ciremai, West Java, Indonesia
Abstract
Volcanic slopes are important sources of water. Groundwater observation in east slope of
Ciremai was conducted in period of dry season of May until June 2006 on 119 springs.
Fourteen variables was measured: elevation, spring discharge, TDS, EC, pH, T water, T air,
calcium, magnesium, chloride, sodium, sulphate, potassium, bicarbonate.
Cluster analysis has successfully extracted 3 clusters: cluster 1 (112 obs), cluster 2 (5 obs),
cluster 3 (2 obs). The hydrogeological schematization has been constructed based on
interpretations of 3 clusters. There are 3 hydrogeological systems: 1) Hydrogeological system
1 is characterized by: cluster 1, heterogeneous data, normal water temperature, TDS, EC, and
major elements concentrations, and many chemical influences. 2) Hydrogeological system 2
is characterized by: cluster 2, homogeneous data, high water temperature, TDS, EC, the result
of interaction between normal meteoric water and hot water of volcanic origin. 3)
Hydrogeological system 3 is characterized by: cluster 3, high water temperature, TDS, and
EC with homogeneous characters, deep flow system, and interaction with sedimentary layers
of Fm. Kaliwungu.
1. Introduction
Indonesia is a part of ring of fire, consisting of almost 128 volcanoes. In other figure, 13 –
17% of world’s volcanoes are located in Indonesia. Such large number of volcanoes makes
Indonesia one of important country to solve volcanoes problems. Subduction zone lies across
the country forming volcanic belt, most of it are strato volcano. Hundreds of volcanoes
produce volcanic deposit which covers 33,000 km 2 or one sixth of Indonesia’s land (Dept. of
Mining and Energy, 1979).
At this volcano, volcanic deposit plays role as productive aquifer, as shown by the emergence
of spring belt with enormous discharge and excellent quality. The aquifer comes as porous
system as well as fracture system. For an example on Mt. Ciremai, there are at least 116
springs with variable discharge, from 10 liter/seconds to 1000 liter/seconds.
This paper describes the hydrological assessment performed on the Mt. Ciremai (Figure 1).
Here we analysed 116 springs around the volcano on water quality and quantity. The results
show radial flow patterns, a dependency on slope aspect and altitude and lithology. This paper
will elaborate on the relationship between volcanic geomorphology and hydrology that was
found and discuss how this information could be used for assessing the spatial patterns of
local groundwater systems on volcanic slopes. Lastly, we make a spatial water balance to
illustrate the regional differences in water availability.
2. Problem Statement
On the Java island Indonesia, the water demand increases due to population growth and rising
water consumption per capita (Puradimaja et.al, 2004). Although Indonesian islands receive
abundant precipitation (2000 - 4000 mm/yr), it is not well distributed both spatially and
temporarily. For example the Java Island shows large spatial differences of rainfall between
the coastal areas (less than 250 mm/yr) and volcanic areas (more than 2500 mm/yr). The
monsoon climate divides the year in a clear dry and wet period.
Water scarcity, especially for agriculture is a well-known problem in Indonesia. However, the
water resources are not managed yet in this respect. By example, upslope movement of
habitation and agriculture changes the water budget of that particular region and influences
the amount of stream water available downstream.
3. Literature Review
3.1 Hydrogeological Setting
Mount Ciremai is a solitaire-strato volcano with elevation of 3072 masl, situated at
Majalengka (west flank) and Kuningan Regency (East flank) (Figure 2), 20 km south of
Cirebon, Indonesia. It lies at 6° 53’ 30‖ latitude and 108° 24’ 00‖ longitude. The diameter of
this volcano, from the peak to the foot slope is about 10 km. Mt. Ciremai has recorded 5
eruptions in 1698, 1772, 1775, 1805, and 1937, with 3 -112 years interval, historically. Those
eruptions produced 22 volcanic deposits, consist of: 11 lava flows, 9 pyroclastic materials,
and 2 laharic breccias. Many studies have been conducted in the area.
Situmorang (1995) has published the volcanic geological map of Mt. Ciremai. According to
the author, the exposed volcanic deposit was produced by 4 generation of eruptions, which
generated 3 types of deposits. Lava flow consists of andesite rock, black to brownish in
colour. It has fractures, cooling joints and columnar joints due to mass unloading and cooling
processes. Pyroclastic materials consist of andesite fragments planted in tuff lithic and tuff
crystal. It comes in flowing and falling mechanisms. Laharic breccias consist of andesite
fragments planted in volcanic sands, tuff lithic, and tuff crystal. It comes in water dominant
flowing mechanisms. The first regional hydrogeology condition was introduced by IWACO –
WASECO (1989). Regional aquifer system at Mt. Ciremai area is divided in to 3 systems:
Surficial Alluvium, Quaternary Volcanic (Young Volcanic), and Tertiary Sediment system.
More details study was conducted by Puradimaja et.al (2003). According to the author, there
are 3 main aquifer units: pyroclastic breccias, lava, and laharic breccias. All of the observed-
aquifers are unconfined aquifers. The aquifer feeds water to spring zone encircling Mt.
Ciremai. The spring zone is interpreted to be following slope morphology which controlled by
change of rock distribution. Such condition forms 2 slope breaks at 750 masl (4 o difference)
and 1350 masl (19o difference).
Irawan (2006) stated that there 3 factors which control the spring emergence. First factor is
the change of rock distribution from lava to laharic breccias. Morphological features in form
of ridges and valleys also contribute to control groundwater flow pattern. Second factor is
fracture and continuous voids zone controls the level of spring discharge in volcanic terrain.
Third factor is the weathering processes in the study area is very intensive, resulting in thick
residual soil and high final infiltration rate, which is very potential to store and to be
infiltrated by rain water and surface water.
There were 3 thermal groups of spring, based on 23 spring observations, consists of:
hypothermic, mesothermic, and hyperthermic (Figure 3). Based on the thermal classification,
it can be interpreted that thermal differences were triggered by interaction between
groundwater temperature and environmental temperature. The hypothermic group shows the
closed system of groundwater. Mesothermic group shows the interaction between
groundwater and surface temperature. Hyperthermic groups were characterized by the
interaction of groundwater with specific subsurface heat source. (Hem, 1970 and Matthess,
1982). Piper plot is presented in Figure 4, to show the differentiation of chemical characters.
The intensity of weathering processes in the study area is very high, indicated by 2 m to
nearly 10 m of soil thickness. Such thick residual soil is very potential to store and to infiltrate
rain and surface water in to the aquifer. Infiltration tests (according to Chow et.al. 1964;
Miyazaki, 1993) was carried out to verify the final infiltration rate of residual soils. Residual
soil from lahars shows the largest values of 1.26 – 2.53 cm/min, followed by residual soil
from pyroclastic breccias 1.5 cm/min, and from lava flow 0.5 – 1.2 cm/min. High final
infiltration rate (Linsley & Franzini, 1978) indicates the high capacity of residual soil to be
infiltrated by rain water and surface water.
Subsequently, in 2002, Bapeda Kuningan Regency has mapped 161 springs with total of
8285.2 l/sec. The result is five classes of spring discharge magnitude (Meinzer 1923, op.cit
Todd, 1980): Six springs of Class II (4%), 44 springs of Class III (27%), 15 springs of Class
IV (9%), 40 springs of Class V (25%), 56 springs of Class VI (35%). Both preceding studies
have not analysed the control of geological setting to groundwater springs.
Step 1: Select cluster variables and distance measures. How many and which variables are to
be selected will affect the analysis results. In cluster analysis, it is implicitly assumed
that every variable is equally important.
Step 2: Select cluster algorithm. Cluster algorithm is the procedure to determine ―clusters,‖
or ―groups.‖ There are two categories of cluster algorithms, hierarchical and non-
hierarchical. In this paper, we are going to use hierarchical algorithms.
Step 3: Perform cluster analysis. Cluster analysis will determine the cluster structure—
specifically, which objects form a cluster, how many clusters, the features of clusters,
etc.
Step 4: Interpretation. We need to explain what these clusters mean and how should we
name and make sense of these clusters. The interpretation is based on geological
facts.
N
d ik (x
j 1
ij xkj ) 2 Equation 1
In cluster analysis, it is desirable that the distances between objects within a cluster (group)
are small and the distances between different clusters are large, as illustrated in Figure 6. The
definition of the distance between clusters depends on the methods to determined relationship
between clusters, called linkage. There are several different linkage methods which we will
discuss as follows. In single linkage method, the distance between two clusters is defined to
be the distance of the nearest neighbours (Figure 6). Specifically, assume that we have two
clusters, cluster R and cluster S. Let r represent any element in cluster R, and let s represent
any element in cluster S. Then the distance between cluster R and cluster S in single linkage
method is defined as
d ( R )( S ) min d rs r R, s S Equation 2
In a dendogram, the distances between clusters and the joining process are described very
well. We usually want to form more than a cluster in further analysis. As we discussed earlier,
a good clustering should be as follows:
1. The objects within a cluster should be similar one another, in other words, the distances
between the objects within a cluster should be small.
2. The objects from different clusters should be dissimilar, significantly, or the distances
between them should be large.
4. Methodology
The delineation of groundwater systems aims: 1). the recognition of the hydrogeological
boundaries enclosing the system; 2) the mechanisms of recharge, and discharge, along with
the flow paths of groundwater from recharge areas to discharge areas Mandel & Shiftan
(1981). In order to map the hydrogeological boundaries and recharge – discharge mechanism,
this research used 3 main approaches: desk study, hydrogeological mapping, hydrochemical
sampling and analysis. Detailed work describes as follows.
1) Desk study: The desk study consisted of studying the topographical map, geological maps,
hydrogeological maps, and re-analyses of previous studies;
2) Hydrogeological mapping: The hydrogeological mapping is based spring observations;
consist of: coordinates, local geological observations, measurement of the spring
discharges and the water qualities. The discharge was measured using the area-velocity
method. The water velocity was measured using current meter. For small discharge (less
then 1 l/s), the measurements were taken using volumetric method with a 10 liter bucket
and stopwatch. Duplets measurements were taken at each observation.
3) Hydrochemical sampling: At each location, physical parameters measurements were taken
consist of: air temperature, water temperature, Electro-Conductivity (EC), Total Dissolved
Solids (TDS), and pH. The air temperature was measured only during sampling using a
standard thermometer, while other above parameters were measured using Lutron portable
equipment. For laboratory chemical analysis, the spring water was sampled using 1 litre
plastic bottles. The duplets laboratory analysis comprises: the calculation of major
elements concentrations using titration method. Chemical test results have to be validated
using ion balance equation (see equation 3), before further analyses. We determined 20%
error balances as permit-able limit. Samples have higher than 20% of error balance will be
re-tested while samples have lower than 20% error will be analyzed.
[(Σ cations - Σ anions) / (Σ cations + Σ anions)] x 100% equation 3
4) Statistical analysis: The hydrochemical parameters and the result from field observations
were analyzed using basic statistical analysis and cluster analysis to assist the
hydrogeological analysis, by using Minitab (trial version).
5) Interpretation: The interpretation aims to schematization of hydrogeological system based
on each spring clustering.
Acknowledgement
The authors would like to thank The Board for Regional Planning Kab. Kuningan for
facilitating data and Director of PDAM Kab. Kuningan for access to spring abstraction site.
We also would like to thank our team of undergraduate students that gave their time and effort
to help us surveying spring data and gathering the first level analysis.
References
Bapeda Kuningan, 2002, Inventarisasi Mataair Kab. Kuningan, Bapeda Kuningan.
Chow, 1964, Soil Water, Prentice-Hall.
Costello, A.B. and Osborne, J.W., 2005, Best Practices in Exploratory Factor Analysis: Four
Recommendations for Getting the Most From Your Analysis, Practical
Assessment Research & Evaluation, Vol 10, No 7.
Dept. Pertambangan dan Energi, 1979, Data Dasar Gunungapi Indonesia, Dept. Pertambangan
dan Energi.
Hem. J.D., 1973, Study and Interpretation of Natural Water, USGS Water Supply Paper.
Irawan, D.E. and Puradimaja, D.J., 2006, The Differentiation of Hyperthermal Groundwater
Origin by using Multivariate Statistics on Water Chemistry, Journal Geoaplika,
vol 1, no 2.
Irawan, D.E.and Puradimaja, D.J., 2006, The Hydrogeology of The Volcanic Spring Belt,
East Slope of Gunung Ciremai, West Java, Indonesia, Proceeding of IAEG
Conference.
IWACO – WASECO, 1989, Kuningan Regency Provincial Water Supply Report, Dept. of
Public Works.
Kitano, Y. (ed), 1975, Geochemistry of Water, Benchmark Paper in Geology, Hutchinson &
Ross, Pennsylvania, pp. 273 – 296.
Linsley, R.K., Franzini, J.B., Freyberg, D.L., Tchobanoglus, G., 1992, Water resources
engineering, McGraw-Hill, New York.
Mandel and Shiftan, 1981, Groundwater resources: Investigation and Development,
Academic Press.
Matthess, G., 1982, Properties of Groundwater, McGraw-Hill.
Miyazaki, T., 1993, Groundwater Basin Management, Tokai University Press
Multivariate Statistical Methods and Quality, Downloaded from Digital Engineering Library
@ McGraw-Hill (www.digitalengineeringlibrary.com), McGraw-Hill.
Puradimaja, D.E., Sukarno, I., Abidin, Z., Irawan, D.E., 2002, Sistem Pengembangan dan
Pengusahaan Air Bersih di Jawa Barat. Potensi dan Pola Bisnis Air Bersih serta
Air Minum, Dipresentasikan pada acara Seminar ―Pemanfaatan dan Pengelolaan
Air Bersih Guna Meningkatkan Kesehatan Masyarakat Jawa Barat Menuju Era
Globalisasi‖, Aula Barat ITB, 22 Nopember 2002.
Puradimaja, D.J, Irawan, D.E., Hutasoit, L.M., 2003, The Influence of Hydrogeological
Factors on Variations of Volcanic Spring Distribution, Spring Discharge, and
Groundwater Flow Pattern, Bulletin of Geology, Vol 35, No 1/2003, pp: 15 – 23,
ISSN: 0126-3498.
Situmorang, 1995, Peta Geologi Gunung Ciremai, Direktorat Vulkanologi.
Smith, L.I., 2002, A Tutorial on Cluster Analysis, downloaded from
http://www.cs.montana.edu.
Figures and tables
Scale 1 :
A B
C D
Locati
on of
Figure 1 The location map of Mt. Ciremai. Box marks the study area. The area is more – less
20 km south from Cirebon, West Java.
Figure 2 The geological map of Mt. Ciremai (Situmorang, 1995) and groundwater spring
location
KLASIFIKASI TERM AL M ATA IR
y=4093.409-152.69 *x
R2=0,81
140
21 k e lo m p o k h ip Transition
erterm al zone
21
120
10 0 k e lo m p o k m e s o t e r m a l
32
102 1 2 53
43
80
143
Elevation
E
(masl) 60 153 1 76 3
243 23 123
93
83 76 3
253
k e lo m p o k h ip o t e r m a l k e lo m p o k h ip e r t e r m a l
40
21 08 3 193
20
1
4
TA
Figure 3 Chart of thermal gradient of groundwater at Ciremai (Puradimaja, D.J., et.al. 2003).
100
10
0
Ca
+M
l
+C
g
4
SO
7
0
86
2
9 1
4
0
100 100
3
CO
Na
+H
+
SO 4
Mg
3
CO
2
98
61
7 4
0
10
10
6
0
0 2
7
0
491 8
10
0
0
0
10
0
Ca Cl
GEOLOGICAL HYDROGEOLOGICAL
MAP MAP
HYDROGEOLOGICAL
SURVEY
PHYSICO-
SPRING
CHEMICAL
OBSERVATION
MEASUREMENTS
CLUSTER
ANALYSIS
INTERPRETATION
Figure 6 Illustration of what is Euclidean distance (upper left), cluster distance and between
cluster distance (upper right), and dendogram as the final results of cluster analysis (lower).
Table 3 Descriptive analysis of the variables. Large deviations as shown by several variables:
TDS, EC, hardness, chloride, sodium, and bicarbonate should affect the clustering processes.
Observations with maximum value of those variables should separated relatively from the
other observations with normal value.
Table 4 Cluster analysis results (Minitab trial version). Centroid is the focal point of cluster.
Maximum distance of observation from centroid is measured on each cluster. Cluster 1 shows
more variation as shown by the large maximum distance from centroid (9.23), relatively to
cluster 2 (3.05).
Average Maximum
distance distance
Cluster basics description Number of from from
observations centroid centroid
Cluster1 112 1.99 9.23
Cluster2 5 1.55 3.05
Cluster3 2 0 0
Centroid distance
Cluster1 Cluster2 Cluster3
Cluster1 0 13.97 15.68
Cluster2 13.97 0 9.28
Cluster3 15.68 9.28 0
V3
V20
V27
V60
V80
V103
V51
V90
V69
V99
V72
V4
V18
V6
V17
V97
V91
V131
V102 Similarity
V55
V15
V53
V92
100,00
75,46
50,92
26,38
V34
V13 V16
Observations
V35 V49
V95 V43
V73 V8
V96 V50
V81 V23
V59 V12
V14 V42
V100 V1
V31 V2
V109 V24
V52 V33
V74 V63
V71 V25
(112 obs)
V58
V21 V89
V10 V61
V93 V129
V130 V112
Cluster 1 (continuation)
V134 V22
V62 V110
V128 V132
V98
(5 obs)
V3
Cluster 2
V77 V20
V94 V27
(112 obs)
V105 V60
(2 obs)
Cluster 3
V38 V80
V67 V103
V76 V51
V9
V90
Cluster 1 (continuation)
V108 V69
V106 V99
V65 V72
V56 V4
V85 V18
V84 V6
V86
V17
V29 V97
V48 V91
V131
V102
V55
V15
V53
V92
V34
V13
Observations
V35
V95
V73
V96
V81
V59
V14
V100
V31
V109
V52
Figure 7 Dendogram of cluster analysis (Minitab trial version). The lower dendogram is the
continuation of the upper dendogram. There are 3 clusters: cluster 1 (112 observations. cluster
V74
V71
Cluster analysis mataair Kab. Kuningan
V64
V70
V68
V40
V88
V37
V39
V19
V26
3500
3000
2500
2000
Elevasi
1500
Hgl 1
Normal water
1000 descent
500
0 Tertiary sediments
108.15 108.2 108.25 108.3 108.35 108.4 108.45 108.5 108.55 108.6 108.65
3500
3000
2500
2000
Elevasi
1500
Normal water
1000 descent
500
Hgl 2
Hot water rising
0 Tertiary sediments
108.15 108.2 108.25 108.3 108.35 108.4 108.45 108.5 108.55 108.6 108.65
3500
3000
2500
2000
Elevasi
1500
Normal water
1000 descent
500