An integration of GIS and remote sensing in groundwater investigations:

A case study in Burdur, Turkey

Erhan Sener · Aysen Davraz · Mehmet Ozcelik

Abstract The importance of groundwater is growing become very handy tools in accessing, monitoring and
based on an increasing need and decreasing spring dis- conserving groundwater resources. It also has been found
charges in the Burdur area. Remote Sensing and the Ge- that remote sensing, besides helping in targeting potential
ographic Information System (GIS) have been used for zones for groundwater exploration, provides input to-
investigation of springs, which are an important ground- wards estimation of the total groundwater resources in an
water source. The chemical composition of groundwater area.
is not of drinking water quality in Burdur city and water It was the purpose of this investigation to discover new
in the Burdur residential area is being obtained from the water resources in Burdur, located in the southwest of
Cine plain. Turkey, using remote sensing and geographical informa-
The purpose of this study was to investigate new water tion system (GIS) methods (Fig. 1). The research was
sources by using remote sensing and GIS methods. Ge- done in an area of approximately 437 km2. Drinking
ology, lineament and land use maps of the research area water in Burdur has been obtained from wells and springs
were prepared using the Landsat TM satellite image about 10 km south of Burdur in the Cine plain . Due to the
composed of different analyses on the TM 7–4-1 band. In over development of wells in the Cine plain a decrease of
addition, contours, creeks, roads and springs were digi- groundwater levels has been observed. Separately, local
tized using a topographic map of 1/100,000 scale to management plan the needs for water supply to settle-
produce a drainage density map. A groundwater potential ments on the Cine plain. Therefore, it is necessary to find
map was produced which integrated thematic maps, such new water sources for Burdur.
as annual rainfall, geology, lineament density, land use,
topography, slope and drainage density. According to this
investigation, the surrounding villages of Askeriye, Bug- Methodology
duz, Gelincik, Taskapı and Kayaaltı were determined to
be important from the point of view of groundwater po- This research was accomplished by using the Erdas Image
tential in the research area. 8.4 process software, Arc GIS 8.2 and Arc View 3.2 (3D
analyst and spatial analyst extensions) and SPSS 11 sta-
Keywords Burdur · Geographical Information System · tistical programs at the Suleyman Demirel University
Groundwater Exploration · Remote Sensing Remote Sensing Centre in Turkey. In this study, Landsat
ETM+ (178–34 scenes) from the year 2000 images and a
1/100,000 scaled topographical map of Isparta- M24 have
Introduction been used. In addition, annual rainfall data for the sur-
rounding area was used for the annual rainfall map. The
Remote Sensing and the Geographical Information Sys- topographical map was digitized using the Arc Info 7.2.1
tem (GIS) with their advantages of spatial, spectral and version software program. Topological changes and co-
temporal availability and manipulation of data covering ordinate transform were corrected and the data converted
large and inaccessible areas within a short time have to Arc View software program with the 3.2 version. In
this study, four stages were applied: collection, satellite
image processing, digitizing—building a database and
Received: 31 July 2003 / Accepted: 21 July 2004 spatial analyses (Fig. 2).
Published online: 6 November 2004

Springer-Verlag 2004
Satellite Data Analysis
E. Sener · A. Davraz ()) · M. Ozcelik Satellite data provides quick and useful baseline infor-
Department of Geological Engineering, mation on the parameters controlling the occurrence and
Suleyman Demirel University, movement of groundwater, e.g. geology, geomorphology,
Isparta, Turkey
e-mail: adavraz@mmf.sdu.edu.tr
soils, landuse, lineaments etc.

Hydrogeology Journal (2005) 13:826–834 DOI 10.1007/s10040-004-0378-5

 827
Fig. 1 Location map of the
study area

In this study, basic maps such as lineaments and ge- It was deposited as a flysch facies of Upper Paleocene-
ologic maps in a digital form were produced to analyze Medium Eocene age (Gormus and Ozkul 1995). The
and to interpret the satellite data. The satellite data reg- Kayikoy Formation, representing the flysch properties,
istration, correction (geometric and atmospheric) and was described as an impermeable unit.
other image processing (such as enhancement, filtering, The Gokcebag ophiolitic complex is formed of lime-
classifications, resolution merge), together with field stone, diabase, serpentine, peridotite, gabbro, radiolarite
checking of the relevant area were applied. All the and chert series. The emplacement age of the complex is
available spatial data were assembled in a digital form, Late Lutetian (Ertunc et al. 2001). The Gokcebag Com-
and properly registered to make sure the spatial compo- plex was also described as an impermeable unit. The
nent overlapped correctly. This produced mapped layers, Akdag Limestone is located as an olistolith within the
including annual rainfall, geology, lineament density, Gokcebag Complex. The unit, which shows aquifer
topographic elevation, slope, drainage density and lan- properties, is widespread in the research area. Kocapınar
duse. Various analyses, such as table analysis and weight and Taskapı springs, which discharge from the Akdag
calculation, were made for extracting this data. Suitable Limestone, provide some of the drinking water needs of
weights were assigned for different parameters to assess Burdur city (Tasdelen 1994).
groundwater potential (Table 1). The values of the The Karaburun Formation consists of a conglomerate
weightages are based on Musa et al. (2000). of polygenic pebbles (Oligocene). Conglomerates may
contain water within karstic cavities and in fractures.
However, this unit does not form an aquifer due to its
Geology-Hydrogeology limited outcrop in the investigated area.
A geological map of the study area was prepared using The Burdur Formation is composed of a marl, clay-
the Landsat ETM+ satellite image of the year 2000 stone, mudstone and weakly cemented sandstone with
composed of different analysis (contrast enhancement, tuff, tuffite and travertine and is Pliocene in age. The
spatial filters and band ratio) on the ETM+ 7,4,1 band sandstone and tuff layers within the Burdur Formation
together with field checking. The Kayikoy Formation, contain some groundwater depending on its thickness. In
Karaburun Formation, Burdur Formation and alluvium some dug wells within the Cine plain, travertine yields
are autochthonous units and the Gokcebağ Complex and some water.
Akdag Limestone, allochthonous units, all were observed In the research area, a Quaternary age unit was de-
in the research area (Figs. 3 and 4). scribed as alluvium. The unit consists of gravel, sand, clay
The Kayikoy Formation consists of sedimentary rock and silt, which are loosely cemented. The Burdur resi-
sequences, components of which are sandstone, con- dential area is located on the alluvium, which is described
glomerate, shale, clayey limestone and cherty limestone. as a good aquifer due to the presence of water in sand and

Hydrogeology Journal (2005) 13:826–834 DOI 10.1007/s10040-004-0378-5

828
Fig. 2 Methodology flowchart
for groundwater exploration

gravel beds. Yields from wells range between 8 to 55 l\sec were identified in the area. Minimum lineament length is
depending on the thickness of the alluvium and more 26.74 m and the longest lineament observed was 5,619 m.
permeable gravel layers. According to their hydrogeo- A lineament density map was prepared by IDW interpo-
logical properties, the lithological units in the investiga- lation method used with one grid cell per square km
tion area were evaluated and a table of suitable weigh- (Fig. 5). Suitable weightages for lineament density are
tages for geology is shown in Table 1. shown in Table 1. The values of the weightages were
proposed by Krishnamurty et al. (1996, 1997). Ground-
water potential in the area increases with higher lineament
Lineament Analysis density. The springs located in the research area were
Lineaments give a clue to movement and storage of digitized and were placed on the lineament map using the
groundwater (Subba et al. 2001) and therefore are im- Arc View program. In general, all springs were developed
portant guides for groundwater exploration. Recently, on lineaments (Fig. 5).
many groundwater exploration projects made in many
different countries have obtained higher success rates
when sites for drilling were guided by lineament mapping Drainage Density
(Teeuw 1995). Drainage pattern is one of the most important indicators
In the research area, the lineaments have been identi- of hydrogeological features, because drainage pattern,
fied with the aid of lineament filters and visual comments texture and density are controlled in a fundamental way
on the Landsat ETM+ 7 band. Most of the lineaments by the underlying lithology. In addition, the creek pattern
have a trend of N20-50W. A total of 171 lineaments is a reflection of the rate that precipitation infiltrates

Hydrogeology Journal (2005) 13:826–834 DOI 10.1007/s10040-004-0378-5

 829
Table 1 Weight values accord- Parameters Weight
ing to different parameters
Lithology Alluvium 70
Akdag Limestone 50
Karaburun Formation 40
Burdur Formation 30
Gkcebag Complex 15
Kayıkoy Formation 10
Lineament density (km/km2) 0–0.600 30
0.600–1.200 40
1.200–1.800 50
1.800–2.400 60
2.400–3.100 70
Drainage density (km/km2) 0–0.500 10
0.500–1.000 20
1.000–1.500 30
1.500–2.000 40
2.000–2.500 50
Topography elevation (m) 1,700–1,900 10
1,500–1,700 20
1,300–1,500 30
1,100–1,300 40
900–1,100 50
Slope gradient 64-90 10
25-63 20
10-24 30
4-9 40
0-3 50
Landuse Urban 10
Forest 20
Scrub 30
Agriculture 40
Water body 60
Annual rainfall (mm) 420–440 30
440–460 40
460–480 50
480–500 60
500–520 70

Fig. 3 Geological map

Hydrogeology Journal (2005) 13:826–834 DOI 10.1007/s10040-004-0378-5

830
Fig. 4 Generalized stratigraph-
ical column

compared with surface runoff. The infiltration-runoff re- Landuse

lationship is controlled largely by permeability, which is The landuse of the research area is characterized by a
in turn, a function of the rock type and fracturing of the mixture of forest cover, agricultural activities, residential
underlying rock or bedrock surface (Edet et al. 1998). area and water bodies. These were interpretable from
Creeks from the topographical map (1/100,000 scale) satellite image and landuse maps. A table of suitable
were digitized using the Arc Info program. A drainage weightages for landuse is shown in Table 1.
density map was prepared from drainage length per
square km (Fig. 6). In addition, a table of suitable
weightages for drainage density was also prepared and Annual Rainfall
presented in Table 1. An annual rainfall map of the research area was prepared
by the IDW method using the annual rainfall measured at
the meteorology stations and surrounding area. Grid cells
Topography Elevation and Slope (11 square km) were created on this map. A table of
Topographic data is a vital element in determining the suitable weightages for annual rainfall was prepared ac-
water table elevations. The topographic map was digitized cording to the map (Table 1). The values of the weightage
on a 1/100,000 scale base map (Fig. 7). Topographic are from Krishnamurty et al. (1996–1997).
contours were integrated into the GIS and used to derive a
digital elevation model (DEM). Values of slope angle
were taken from the DEM. Each grid cell represents the Spatial Analyses
value of a slope gradient or slope angle in degrees. The
table of suitable weightages for topography and slope are The final stage involves combining all thematic layers
shown in Table 1. using the method that is modified from Musa et al. 2000.

Hydrogeology Journal (2005) 13:826–834 DOI 10.1007/s10040-004-0378-5

 831
Fig. 5 Lineament density map

Fig. 6 Drainage density map

Hydrogeology Journal (2005) 13:826–834 DOI 10.1007/s10040-004-0378-5

832
Fig. 7 Topographical elevation
map

Fig. 8 Groundwater potential

prediction zone map

Hydrogeology Journal (2005) 13:826–834 DOI 10.1007/s10040-004-0378-5

 833

Kayıkoy Formation Gokcebag

The formula of the groundwater potential model (GP) as

1,100–1,300 / 1,700–1,900
shown below:

Agriculture-scrub-forest
GP ¼ Rf þ Lt þ Ld þ Lu þ Te þ S þ Dd ð1Þ

480–440 / 420–440
0.60–1.20 / 0–0.60
0.50–1.00/ 0–0.50
where;

Rf annual rainfall,

0–3, 4–9
Very low

Complex
Lt lithology (geology),
Ld lineament density,
Lu land use,
Te topography elevation,
S slope,
Karaburun Formation and Gokcebag

Dd drainage density.

1,300–1,500 / 1,500–1,700 A groundwater potential zone map of the research area

was prepared (Fig. 8). In order to produce the map, a GIS
model was used to integrate the thematic maps; geology,
lineament density, rainfall, drainage density, topography
Kayıkoy Formation

480–440 / 420–440
Agriculture—scrub
0.60–1.20 / 0–0.60
0.50–1.00/ 0–0.50

elevation, slope and land use. Each thematic layer con-

sisted of a grid cell. The grid cell in each of the thematic
0–3, 4–9

layers was categorized, depending on the contribution to

Complex

groundwater potential. Finally, all the thematic layers

Low

were integrated using the groundwater potential model to

produce the final derived layers. In addition, a general
evaluation table was built using data given in Table 1.
The outputs were then re-classed into four groups, high,
1,100–1,300 / 1,300–1,500

moderate, low and very low groundwater potential (Ta-

ble 2).
1.20–1.80 / 0.60–1.20
1.00–1.50 / 0.50–1.00

According to the Table 2 and Fig. 8, the areas having

Agriculture—scrub
Burdur Formation

high groundwater potential included the Akdag limestone

and alluvium unit that have a very high-high drainage
0–3, 4–9

density and high rainfall ratio. The Burdur Formation

Moderate

460–480

forms the areas with moderate groundwater potential. The

lineament density and drainage density are moderate-low
in the Burdur Formation. The areas having low ground-
water potential include the Karaburun Formation, Kayı-
koy Formation and Gkcebağ Complex. The lineament
1.80–2.40/2.40–3.10 (Alluv. NE)

density, drainage density and rainfall data of Karaburun

Formation, Kayıky Formation and Gkebağ Complex
Akdag Limestone, Alluvium

are low-very low. According to this general evaluation,

900–1,100 / 1,100–1,300

Table 2 and Fig. 8 correspond to each other.

1.50–2.00 /2.00 2.50

0–3, 4–9, 10–24

Agriculture—scrub

Result and Discussion

480–500

The indicators of groundwater occurrence are related to

High

rainfall distribution, land use, geology, topographic ele-

vation, slope and drainage features of the area. Satellite
data has proven to be very useful for surface study, es-
Table 2 The general evaluation

pecially in detecting surface features and characteristics

Groundwater potential zone

such as lineaments and geology. In order to predict the

groundwater potential zones, different thematic maps
Topography elevation

were prepared. These include annual rainfall distribution,

Geology / lithology

Lineament density

land use, geology, lineament density, topography eleva-

Drainage density

tion, slope and drainage density. Integrated assessment of

Rainfall (mm)

thematic maps using a model developed based on GIS

Land use

techniques was a suitable method for predicting ground-

water potential. In the research area, the surrounding
Slope

villages of Askeriye, Bugduz, Gelincik, Taskapı and

Hydrogeology Journal (2005) 13:826–834 DOI 10.1007/s10040-004-0378-5

834

Kayaaltı are important from the point of view of through remote sensing and a geographical information system.
groundwater potential according to this investigation. Int J Remote Sensing 7:1867–1884
Krishnamurty J, Arul Mani M, Jayaraman V, Manivel M (1997)
Those areas were investigated in detail for new water Selection of sites for artificial recharge towards groundwater
sources for Burdur city. development of water resources in India. Proc 18th Asian Conf
on Remote Sensing, Kuala Lumpur, 20–24 October
Musa KA, Juhari Mat A, Abdullah İ (2000) Groundwater prediction
potential zone in Langat Basin using the integration of remote
References sensing and GIS. The 21st Asian Conf on Remote Sensing,
Taipei (Taiwan)
Edet AE, Okereke CS, Teme SC, Esu EO (1998) Application of Subba Rao N, Chakradhar GKJ, Srinivas V (2001) Identification of
remote sensing data groundwater exploration: A case study of groundwater potential zones using remote sensing techniques in
the Cross River State, Southeastern Nigeria. Hydrogeology J and around Guntur Town. Andhra Pradesh, India 29 (1&2):69–
6:394–404 78
Ertunc A, Karaguzel R, Yagmurlu F, Turker A, Keskin N (2001) Tasdelen S (1994) Hydrogeological investigation of Cine Plain
Report of investigation of Burdur municipality and nearby (Burdur) and surroundings. Suleyman Demirel University, Phd
surroundings in reference to earthquake and residential com- Thesis, Isparta, Turkey
munities. Suleyman Demirel University, Isparta (in Turkish) Teeuw RM (1995) Groundwater exploration using remote sensing
Gormus M, Ozkul M (1995) Stratigraphy of Gonen-Atabey (Is- and a low-cost geographical information system. Hydrogeology
parta) and Aglasun (Burdur). Suleyman Demirel University, Journal 3(3):21–30
Isparta, Turkey, Publication 1, pp 43–64
Krishnamurty J, Venkatesa Kumar N, Jayaraman V, Manivel M
(1996) An approach to demarcate groundwater potential zones

Hydrogeology Journal (2005) 13:826–834 DOI 10.1007/s10040-004-0378-5