GEOLOGICA CARPATHICA, 49, 5, BRATISLAVA, OCTOBER 1998
MORPHOSTRUCTURE PATTERNS IN SATELLITE
MULTISPECTRAL LANDSAT TM IMAGES
Earthquake Planning and Protection Organization, Xanthou 32, 154 51 Athens, Greece; firstname.lastname@example.org
(Manuscript received February 2, 1998; accepted in revised form September 1, 1998)
Abstract: The possibility of tectonic observations using multispectral remote sensing data is based mainly on
morphostructures, which could lead directly to the structural interpretation of the area. This work aims to contribute to
the creation of a morphostructure-code with a corresponding key-legend, in order to facilitate the acquisition of
tectonic information by users, who do not specialize in remote sensing. The studied area covers a great part of NW
Greece (Epirus Prefecture). During the summer and autumn of 1996, this region was affected by strong and disastrous
earthquakes (up to 5.6 R). Data from Landsat TM have been selected and processed creating a false colour composite
image and then analyzed and interpreted in order to detect the tectonic features.
Key words: NW Greece, Epirus, tectonism, morphostructure, Remote Sensing, Landsat.
The direct recognition of tectonic features on satellite multi-
spectral data, is based on the concept of morphostructures
which could lead, through their detection and interpretation, to
the tectonic analysis of an area. The recognition of the surfa-
cial traces of faults and folds is one of the advantages of re-
mote sensing. For many years the term lineament (OLeary
et al. 1976) was used to describe linear features many of which
correspond to known geological structures. The possibility of
structural observation from satellite data optical or radar, is
based mainly on morphotectonics. The term morphotectonics
is simply a contradiction for tectonic geomorphology, that is,
the study of processes and forms related to any form of tecton-
ic activity (Embleton 1987).
Recently, the term morphoneotectonic has been intro-
duced. It is based on the concept that tectonic movements
have brought about changes in the Earths surface. In general
these changes appear more marked and evident, the more re-
cent and bigger are the movements (Panizza 1991). The man-
ifestation of tectonic features depends mainly on the charac-
ter of the rock and on the operating climatic conditions.
Chorowicz (1984) mentioned the importance of pattern rec-
ognition for geological remote sensing applications, conclud-
ing that fundamental geological objects have their distinctive
shape and geomorphological expressions and that on the dig-
ital images, their automatic recognition by computer is possi-
ble. Faults, joints and lineaments most likely have a rectilin-
ear exposure on images and their determination depends on
morphological features or on particular patterns. The recog-
nition of folds can be done through geomorphology or the
pattern of the elementary geological features. The simple
recognition of a fault does not seem to be the most important
factor, but the type and the dynamic that characterized it. Of
course low-angle faults (nappes) are difficult to interprete
since the images provide planar views from above. Such
faults have a strongly curved or irregular surfacial trace and
can be inferred on the basis of discordance of the foliation
(Greiling 1983; Otsuki 1985; Gupta 1991). In the past, lists
of classifications of faults and their photo-interpretation cri-
teria have been created (Reading 1980; Slemmons 1982;
Scanvic 1983). This paper presents an attempt to directly rec-
ognize the types of structure and their movement on the basis
of morphotectonic features.
Fig. 1. Sketch map with the study area.
Location and tectonic setting of the area
The Prefecture of Epirus is located in North-western Greece
and Konitsa town is in the northern part of it (Fig. 1). The area
consists of a great thickness of sedimentary sequence which
geotectonically belongs to the Ionian Zone. Locally in the
north-eastern part of the region, there are sedimentary forma-
tions of the Pindos Zone which are thrust above the Ionian
Zone. The most recent sediments in the area are fluvial terrac-
es of the Aoos, Sarantaporos and Voidomatis rivers.
It is obvious that intense tectonic activity affected the area
predominantly during the Alpine orogenetic compressive
stress. The result of this is the presence of anticlinal and syn-
clinal forms with a NNW-SSE axial direction. These forms
are interrupted by a NE-SW transverse fracture zone, named
as the Konitsa fracture zone, which is a product of a later ex-
tensional stress. A characteristic element of this zone is the
attenuation with smaller fractures in the two edges. Another
fault system with a NW-SE direction has also affected the
area. The rivers Aoos and Voidomatis are related to this fault
system (Papanikolaou & Parcharidis 1996).
In the 26 of July 1996 a strong earthquake was occurred in
the area with magnitude of 5.2 R followed by less intense
earthquakes, while a new greater one affected the area
(M=5.6 R). Dangerous phenomena such as landslides and
rock falls have been observed along the active faults.
Remote sensing data
A digital image of Landsat 5 TM was selected covering the
area (subscene of the 185-032 full scene), with 7 spectral
bands, dating 24-6-93 with 0 % cloud cover, 2500 columns
and 2500 lines for each band and pixel size 30
An important parameter for structural applications is the
azimuth and elevation of the sun at the moment of acquisi-
tion of the scene. In this case the direction of illumination
(from ESE) is more or less perpendicular to the main struc-
tures. The suns elevation is high (beginning of summer) but
this does not create any problem for the information recorded
because the area is characterized by high relief.
First the raw data was radiometrically and geometrically
corrected so that the image can be represented on a planar
surface and have the integrity of a map. The next step was to
combine the adequate spectral bands and display them in the
RGB (Red, Green, Blue) colour system. Different combina-
Fig. 2. False Colour Composite image created from 7, 4, 1 spectral bands of Landsat TM satellite as RGB. The white squares represent
the epicenters of the main seismic events.
MORPHOSTRUCTURE PATTERNS IN SATELLITE MULTISPECTRAL LANDSAT TM IMAGES 309
tions of the TM bands can be displayed to create different
composite effects. Different colour schemes can be used to
bring out or enhance the feature under study. There are 120
possible colour combinations of the data for a large number
of applications. Theory and experience, however, show that a
small number of colour combinations are suitable for most
applications. The optimum band combination is determined
by the terrain, climate and nature of the interpretation project
(Sabins 1997). In the present study the selected bands are 1,
4, 7 forming the false colour composite image 7, 4, 1 as RGB
(Fig. 2). This combination provides the maximum range of
colour signatures for the rock outcrops and is optimum for
geological interpretation in semiarid areas. The image shows
a region, mainly, of mountains and small plains composed of
bare rocks or soil, and cover by scrub and extensive forested
areas. The bare rock and soil are represented by magenta and
pink tones, the scrub by darker greens and browns and the
forested slopes by shades of green. Cultivation of cereals,
fruit and vegetables are represented by the bright green
patches on the flatter ground especially in the Konitsa Valley
and along the rivers. In the center of the image the Tymfi Mt.
and the Konitsa Valley are recognized and the Aoos, Saran-
taporos and Voidomatis rivers too (in blue colour in the im-
age), forming a rectangular type of network. The anticlinal
forms of Mitsikeli and Dusko, with a NNW-SEE direction
can be recognized, but are interrupted, in the middle of the
image by the transverse fracture zone of Konitsa.
Two visits to the area, organized by E.P.P.O. (Earthquake
Planning and Protection Organization), have been made ac-
cording to a program of field studies, mapping and evalua-
tion of the geological structures followed by remote sensing
observation. In situ observations and measurements have
been done mainly in the geological structures which were al-
ready recognized on the image.
The seismological data from the last seismic events have
also been taken in account (magnitude, coordinates of the
epicenters and depths). Earthquake epicentral distribution
may delineate active faults. In Figure 2 the epicenters of the
main seismic events are plotted (the coordinates of the epi-
centers were provided by the Earthquake Planning and Pro-
tection Organization seismological data base). These are dis-
tributed along the Sarantaporos River and the northwestern
slope of Tymfi Mountain.
Recognition, description and analysis
of the structural patterns
The manifestation of tectonic features depends on many pa-
rameters, especially on the rock type and operating climatic
conditions. The freshness of appearance and type of geomor-
phic expression of faults is related to the age of faulting (Mat-
suda 1975). Geomorphological investigations into faulting can
yield considerable information (Doornkamp 1986). A consid-
erable number of geomorphological direct and indirect indica-
tors of tectonic or recent activity exist such as: distortion of
river terraces (Popp 1971), prominent high angle scarps (Cot-
ton 1948), fresh sygmoid-shaped of ridges (Migiros et al.
1995), shutter ridges that is topographic ridges that have been
offset laterally to shut off drainage channels (Cotton 1948),
segmentation and deformation of alluvial fans (Hook 1972;
Bull 1977), displacement of synthetic structures (Rogers &
Nason 1971), faceted ridges created when scarps cut a topo-
graphic ridge (Thornbury 1954), formation of sag ponds (Cot-
ton 1948), offset drainage channels which are especially sig-
nificant because they also indicate the sense and amount of
lateral displacement along a fault (Adams 1975; Parcharidis
1996), river capture (Biancotti 1979). Drainage is also a sensi-
tive indicator of neotectonic events. Streams and rivers can ei-
ther be displaced by such an event or have their gradients
changed. In either case the response may be quite rapid. How-
ever, complications exist, because causes other than tectonics
can produce similar changes (Cooke & Doornkamp 1990) and
last but non least, arcuate scarps or sets of concentric scarps
(Slemmons 1982). The above geomorphological characteris-
tics could be detected on remotely sensed imagery at appropri-
ate products and scales, interpreted and correlated allowing us
to assess composite fault systems and the overall sense of
movement of blocks of the crust which the faults bound. The
heterogeneities of the crusts nature and the stress, involved dur-
ing a period of strong deformation, mean that all the types of
fault may occur in an area. Overview of the geomorphology al-
lows delineation of key locations for morphotectonic investiga-
tion. The representation of morphostructures, related to faults,
through spectral reflectance of the terrain characteristics have
been recognized and described for the studied area as follows:
aBayonnette type structures (Figs. 3 and 4)
More than one are located in the transversal fracture zone
of Konitsa with a NE-SW direction. They are presented like
bayonnettes, perfectly linear and about 10 km in length.
The main topographic feature is the scarp easily recognized
from the shadow effect with moderate slope (the sun azimuth
is perpendicular to the structure) and the presence of the veg-
etation along the topographic depression of the scarp. The
structure corresponds to a fault with a horizontal movement
and an extension regime in the central part of it (better devel-
opment of vertical displacement). A very characteristic fea-
ture is the attenuation of the stress patterns in the two edges
of the structure recorded in the image by a minor scarp relief
with a polyline edge.
bStrike slip faults
The Sarantaporos structure (Fig. 5) coincides with the
straight part of the homonymous river running through the
area with a NE-SW direction and length of 30 km. The total
length of the structure is greater but it is not clearly recognized
in the image. The drainage network of the area, including the
Sarantaporos River, could be characterized as rectangular,
which means that is strongly controlled by the tectonic regime.
The bluish colour in the river basin corresponds to conglomer-
ates and deposits transported by the river. The basin shows the
geomorphic characteristics of a strike-slip fault. The above
conclusion is based on facts such as the development of a
strike-slip type basin which is elongated, parallel to the strike-
slip system, and in this case is relatively deep in relation to its
width. Although the displacement along the strike-slip faults is
dominantly horizontal, the most obvious motion at any one
place may be oblique-slip. This vertical movement is not so
clear in the Sarantaporos fault. In the area just before the Sa-
rantaporos meets Aoos the fault forms a larger basin due to the
curvature formed by the fault (extension). In addition, mor-
phological criteria, confirming the type of the fault, include a
characteristic assemblage of landforms such as offset or de-
flected streams, small scarps, shutter ridges, combined with in
situ information (presence of thermal springs along the basin,
commonly associated with strike-slip zones). It is very inter-
esting that in the middle, the river basin is crossed by a fault
with a NW-SE direction which provokes a small scale dis-
placement (dextral movement) of the rivers route.
The structure along the artificial lake of Aoos sources (Fig.
6) is also a strike-slip fault. It is presented as a linear feature
with a NNW-SSE direction. The typical geomorphic features
are, the linear canyon, eroded and non scarps, a pond activat-
ed actually as a technical lake for energy production, drag-
ging of crests easy detected in the southeastern block (in the
center of the image).
cThe Konitsa fault (Fig. 7)
The Konitsa fault is a long well marked fracture zone with
a NE-SW direction interrupting the Alpine fold system with
a NW-SE axis direction. It is classified as a normal fault but
also with a component of horizontal movement according to
the geomorphic features recognized on the figure. In this
case it is interesting to study the slope and the basic ele-
ments of it. The crest of the slope produced by faulting seems
to be sharp only in the area of the Konitsa Basin, this conti-
nuity is interrupt by small sharp breaks with channels that
cross the scarp. The free face, presented as a straight seg-
ment, is under shadow because the azimuth of the suns illu-
mination is perpendicular to the direction of the structure.
Locally the free face is modified by the accumulation of de-
bris and gulling. The debris slope is clearly recognized in the
image covered by vegetation (light green color in the figure).
The Konitsa Basin is the result of the extensive stress of the
fault and the Aoos River which crosses the basin is of braid-
ed type, with a flow direction from NE to SW, due to the lack
of a thick amount of sediments in the basin. In the southwest-
ern part, the fault interrupts the anticline form of Mitsikeli
provoking the displacement of the folds axis, enhancing the
horizontal component of the structure which seems to be of
The super-synoptic view provided by satellite images is
ideal for detecting or re-evaluating the tectonic patterns over
Fig. 4. Bayonnette type structure, extension zone (1), attenuation
of the stress (2).
Fig. 3. Bayonnette type structure, extension zone (1), attenuation
of the stress (2).
MORPHOSTRUCTURE PATTERNS IN SATELLITE MULTISPECTRAL LANDSAT TM IMAGES 311
Fig. 6. Structure with NW-SE direction along the artificial lake (1), dragging crests (2).
Fig. 5. Sarantaporos River structure with NE-SW direction, material transported by the river and deposited in small basins (1), a small
displacement in the river route, due to the activity of a fault, with dextral movement (2).
Fig. 7. Konitsa fault with NE-SW direction where the basic slope elements are well recognized, crest of the slope (1), free face (2) and
debris slope (3).
large areas. In addition the synoptic view of the satellite imag-
es enables widely separated pieces of evidence to be linked in
their continuation. The best results come from areas of high re-
lief and recent activity, where the faults are well expressed and
the movement directions is clear. Difficulties in the interpreta-
tion could arise when their surface expression is deep eroded.
Photographic interpretation or computer aided interpretation
of faults is often more reliable than their detection in the field. If
multispectral images are available morphostructures may be
more distinct on particular spectral bands. For example interpre-
tations using Landsat images may be enhanced by using infrared
bands, as shadows are sharp and vegetation patterns are distinct.
MORPHOSTRUCTURE PATTERNS IN SATELLITE MULTISPECTRAL LANDSAT TM IMAGES 313
A plethora of more sophisticated processing techniques could
enhance the information contained in the image even more.
In conclusion the remote sensing techniques and data can
be effective in detecting, delineating and describing the
character of active faults, and the near future will be very
promising when data of very high resolution (12 m pixel
size) is available.
Adams D.P., 1975: Geomorphic evidence for late Holocene tilting
in southern San Mateo County, California. J. Res. US Geol.
Surv., 8, 7276.
Biancotti A., 1979: Relations between morphology and tectonic in
Cuneese basin. Geografia Fisicia e Dinamica Quaternaria, 2,
516 (in Italian).
Bull W. B., 1977: The alluvial fan environment. Prog. Phys. Geog.
Chorowicz J., 1984: Importance of pattern recognition for geolog-
ical remote sensing applications and new look at geological
maps. Remote Sensing for geological mapping. In: Teleki P.
& Weber C. (Eds.): Documents BRGM n. 82, Publication
IUGS n. 18, 2940.
Cooke R.U. & Doornkamp J.C., 1990: Geomorphology in environ-
mental management: An introduction. Clarendon Press, Oxford.
Cotton C.A., 1948: Landscape. CUP, Cambridge.
Doornkamp J.C., 1986: Geomorphological approaches to the study
of neotectonics. J. Geol. Soc. (London), 143, 335342.
Embleton C., 1987: Neotectonic and morphotectonic research. Z.
Geomorphol. N. F., Suppl. Bd. 63, 17.
Greiling R., 1983: Fracture patterns, foliations and low-angle
thrusts, interpreted from Landsat images and aerial photo-
graphs from two basement culminations in the north-central
Scandinavian Caledonides. International Basement Tectonics
Assoc., Publ. 4, 321330.
Gupta R.P., 1991: Remote Sensing Geology. Springer-Verlag, Berlin.
Hook R.B., 1972: Geomorphic evidence for late-Wisconsin and
Holocene tectonic deformation, Death Valley, California.
Geol. Soc. Amer. Bull., 83, 20732098.
Matsuda T., 1975: Magnitude and recurrence interval of earth-
quakes from a fault. Earthquake, Ser. 2, 28, 269283 (in Japa-
nese, abstract in English).
Migiros G., Pavlopoulos A. & Parcharidis I., 1995: Recognition of
fracture zones by using spatial models and remote sensing
data: an application in western Attica (Greece). Proc. of the
XV congress of the Carpatho-Balcan Geol. Ass., Sept. 1995,
Athens (Gr), Geol. Soc. Greece, Sp. Pub. No. 4, 10411049.
OLeary D.W., Friedman J.D. & Pohn H.A., 1976: Lineaments,
linear, lineations some standards for old terms. Geol. Soc.
Am. Bull., 87, 14631469.
Otsuki K., 1985: Plate tectonics of Eastern Eurasia in the Light of
fault Systems. Science Reports of the Tohuku University, Sen-
dai, Second Series (Geology), 55, 2, 141251.
Panizza M., 1991: Geomorphology and seismic risk. Earth-Sc.
Rev., 31, 1120.
Papanikolaou D. & Parcharidis I., 1996: Landsat MSS and TM
data for local and regional tectonic observations in the Epi-
rusKonitsa area. Symposium on Remote sensing applica-
tions, Athens 28-29/11/96. Abstracts.
Parcharidis I., 1996: Integration of ERS-1 satellite data and
DEMderived spatial models for a geo-structural scenario in
the Kozani basin (Greece). International Meeting on results
of the May 13, 1995 earthaquake of West Macedonia: one
year after. Abstract.
Popp N., 1971: Hydrogeographical and geomorphological aspects
regarding the problem of the recent vertical movements of the
crust in Romania. Z. Geomorphol., 15, 445459 (in German).
Reading H.G., 1980: Characteristics and recognition of strike-slip
fault systems. Spec. Publ. Int. Assoc. Sediment., 4, 726.
Rogers T.H. & Nason R.D., 1971: Active displacement on the Ca-
laveras fault zone at Hollister, California. Bull. Seismological
Soc. Am., 61, 399416.
Sabins F.F., 1997: Remote Sensing: Principles and interpretation.
W.H. Freeman and Co., New York.
Scanvic J.C., 1983: The use of Remote Sensing in geosciences.
Bureau de recherches geol. Et minieres, Manuels et meth-
odes, n. 7 ( in French).
Slemmons D.B., 1982: A procedure for analyzing fault-controlled
lineaments and the activity of faults. Proceed. 3
fer. On basement Tectonics, In: OLeary D.W & Earl L.J.
(Eds.): International basement tectonics Association, 33.
Thornbury W.D., 1954: Principles of geomorphology. Wiley, New York.