Annals of Valahia University of Targoviste. Geographical Series (2017), 17(1): 90-97
DOI: 10.1515/avutgs-2017-0009
ISSN (Print): 2393-1485, ISSN (Online): 2393-1493
© Copyright by Department of Geography. Valahia University of Targoviste
LANDFORM ANALYSIS USING TERRAIN ATTRIBUTES. A GIS
APPLICATION ON THE ISLAND OF IKARIA (AEGEAN SEA, GREECE)
Athanasios Skentos1, Anagnostopoulou Ourania2
Mott Macdonald Ltd, Sydenham Road, CR0 2EE, Croydon (UK)
E-mail: [email protected]
2
Faculty of Geology & Geoenvironment - University of Athens, Panepistimiopolis, 15771 Ilissia,
Greece E-mail: [email protected]
1
Abstract
The main objective of this study is to classify the landforms of Ikaria Island by conducting morphometric
analysis. The whole classification process is based on the calculation of the Topographic Position Index (TPI).
The delivered TPI landform classes are spatially correlated with the geology, slope, valley depth and the
topographic ruggedness of the island. The results of this study indicate the presence of two distinctive
landform units, affected mainly by the local geological setting.
Keywords: Landform classification, Geomorphometry, Topographic Positioning Index, Aegean Sea
1. INTRODUCTION
Over the past few years, the implementation of automated techniques for landform
classification has become very common among the scientific society (De Reu et al., 2013;
Grohmann and Riccomini, 2009; Ilia et al., 2013; Rigol-Sanchez, et al. 2015; Tagil and Jenness,
2008). What is more, the application of such techniques is not limited to geosciences. In particular,
landform classification has been proved a valuable processing tool for studies related to
archaeology, ecology, agriculture, forestry, rural planning, hazards, etc. (Ho and Umitsu, 2011;
Hoersch et al., 2002; Macmillan et al., 2003; Martin-Duque et al., 2003; Mcnab, 1993; Verhagen
and Dragut, 2012).
Concerning geomorphological research, studying the morphometric parameters of the terrain
is of utmost significance for the landform interpretation. In this study, the landforms of Ikaria are
classified by calculating Topographic Position Index (TPI). TPI is derived by conducting
geometrical and topological analysis using local statistics (user-defined calculus area). Since both
the geology and the ensuing topography determine significantly the landscape, the delivered TPI
landform classes are related to the geological setting and terrain attributes (slope, valley depth,
ruggedness) for further interpretation.
2. STUDY AREA AND GEOLOGICAL SETTING
Ikaria Island is located at the eastern part of the Aegean Sea, covering an area of 250km2
(Fig. 1). From a geomorphological point of view, the study area can be divided as follows; the
eastern part dominated by Atheras Mount, the plain with low hills of Messaria at the center, and the
western part featuring Pezi upland (Fig. 1). In general, the gradient of the relief is much lower close
to the south coast and the hydrographic network much more developed along the north flank of the
island (Stiros et al., 2011). Pezi, Chalaris, Charakas and Voutsides are the main rivers of Ikaria,
draining the area to the northern shore (Fig. 1).
90
Unauthenticated
Download Date | 6/17/17 1:59 AM
The presented complex geology of the island is due to its geotectonic position (Bozkurt and
Oberhansli, 2001; Ktenas and Marinos, 1969). Stiros et al., (2011) suggest that Ikaria’s geological
formations correspond to an anticline with a NE-trending axis. According to the geological map of
Photiades (2005), the study area may be divided in three distinctive geological patterns.
Figure 1. Location map and topography of Ikaria Island featuring main rivers
In particular, the western part of the island is dominated by granites, interrupted only by in
situ Holocene deposits (Fig. 2). Messaria region (at the center of the island) and the easternmost
part of the island consist of carbonate metamorphic rocks and schists. Moreover, Neogene
formations and Holocene deposits appear mainly at the end of the valleys, just before the beach
front (Fig. 2). Lastly, the eastern part of the island is dominated by gneisses, followed by
granodiorites and small extents of Holocene deposits (allocated at the mouth of the rivers). Beyond
these three distinctive patterns, a significant extent of molassic ophiolithic formation appears at the
easternmost part of the island (Fig. 2).
Figure 2. Simplified geological map of Ikaria Island based on PHOTIADES (2005)
91
Unauthenticated
Download Date | 6/17/17 1:59 AM
3. METHODOLOGY
3.1 Calculating terrain attribute
A high resolution DEM (25m2 pixel size) produced by large scale topographic data, is used
in order to compute the following topographic parameters; slope, valley depth, terrain ruggedness
and TPI (Fig. 3). QGIS and SAGA GIS open-software are used for applying the appropriate
algorithms. The pre-processing procedure includes a 25-meter radius Gaussian filter to smoothen
the grid data, and the application of the Planchon/Darboux algorithm to fill DEM’s depressions
(Planchon and Darboux, 2001).
Slope expresses the inclination of the surface of the terrain from either the horizontal or a
local base level (Senthilvelan, 2015).In this paper, slope is categorized in 5 classes (Fig. 3A). Valley
depth refers to the vertical distance to a channel network base level (Fig. 3B). The calculation
algorithm for the specific index consists of two major steps; Interpolation of a channel network base
level elevation, and subtraction of the channel network base level from the original elevations
(Conrad et al., 2015). The Terrain Ruggedness Index (TRI) provides a rapid, objective measure of
terrain heterogeneity (Riley et al., 1999). The algorithm calculates the sum change in elevation
between a grid cell and its eight neighbor grid cells. Therefore, an increased TRI shows increased
local relief heterogeneity (Fig. 3C). Finally, Topographic Position Index (TPI) is defined as the
difference between the elevation at a cell and the average elevation in a cell that surrounds it within
a predetermined radius (Weiss, 2001). TPI values above zero show locations that are higher than the
average of the local window e.g. ridges. In contrast, negative TPI values represent locations that are
lower e.g. valleys (Fig. 3D). Lastly, TPI values near zero are either flat areas (where the slope is
near zero) or areas of constant slope (where the slope of the point is significantly greater than zero).
Figure 3. DEM derived gradients of Ikaria Island (slope, valley depth, TRI, TPI)
92
Unauthenticated
Download Date | 6/17/17 1:59 AM
3.2 TPI Landform Classification
Therefore, TPI values calculated from two different neighbourhood sizes (small and large
ones) provide more detailed information about the general shape of the landscape rather than the
TPI values from just a single neighbourhood. Additionally, complex landforms may be identified
(Tagil and Jenness, 2008). For the purpose of this study, 10m and 200m radius are used for the
determination of the two neighbourhoods TPI classification (Fig. 4). Furthermore, the high and low
TPI values are distinguished by setting a threshold of ±1 STDEV.
The neighbourhood size selected for a TPI implementation is of utmost significance since
different sizes reveal different landforms (De Reu et al., 2013; Ilia et al., 2013; Verhagen and
Dragut, 2012; Weiss, 2001).
Figure 4. Landform Classification based on combined TPI (R1=10m & R2=200m)
4. RESULTS
The percentage distribution of each landform class is shown on figure 5. In particular, open
slopes are the dominant class of the study area (60.83%). Upper slopes cover an area of 10.40%,
followed by U-shaped valleys (8.36%). All other landforms are calculated between 2.5-5% each,
except local ridges and upland drainages who cover less than 1% of the total area.
At this point it should be mentioned that the application of TPI on coastal environments (and
in general next to the DEM boundaries) returns errors on the classified landforms. For instance,
coasts with steep slopes appear as canyons. However, the affected covering area along the coast is
quite small in contrast to the whole study area, resulting to limited influence on the delivering
classes. Moreover, the scale of the study indicates that such anomalies should not be taken into
consideration.
According to table 1, open slopes hold the greatest geodiversity. Granites that cover almost
the whole western part of the island are related with half of the open slopes of the study area
(30.98% in total of 61.28%). Moreover, more than half of the Holocene deposits are characterized
as open slopes (Table 1).
93
Unauthenticated
Download Date | 6/17/17 1:59 AM
Figure 5. Percentage distribution of TPI delivered landform classes
Marbles
Gneisses
Marine Neogene
deposits
Total
1,86%
0,02%
0,43%
0,00%
0,30%
1,93%
0,11%
0,02%
4,89%
Midslope Drainages,
Shallow Valleys
0,23%
2,09%
0,04%
0,36%
0,00%
0,24%
1,28%
0,11%
0,03%
4,38%
Upland Drainages,
Headwaters
0,00%
0,07%
0,00%
0,02%
0,00%
0,03%
0,08%
0,01%
0,00%
0,22%
U-shaped Valleys
0,41%
2,86%
0,02%
0,85%
0,00%
0,53%
3,32%
0,15%
0,03%
8,17%
Plains
0,60%
1,39%
0,07%
0,17%
0,00%
0,05%
0,10%
0,00%
0,05%
2,43%
Open Slopes
3,81%
30,98%
0,41%
5,46%
0,05%
3,58%
16,16%
0,69%
0,13%
61,28%
Upper Slopes, Mesas
0,11%
4,14%
0,04%
1,11%
0,05%
0,84%
3,95%
0,16%
0,00%
10,39%
Local Ridges, Hills in
Valleys
0,02%
0,27%
0,00%
0,06%
0,00%
0,03%
0,31%
0,02%
0,00%
0,71%
Midslope Ridges,
Small Hills in Plains
0,25%
2,03%
0,04%
0,40%
0,00%
0,25%
1,38%
0,12%
0,04%
4,50%
Mountain Tops, High
Ridges
0,03%
1,21%
0,01%
0,33%
0,02%
0,28%
1,09%
0,08%
0,00%
3,04%
Total
5,67%
46,90%
0,65%
9,18%
0,13%
6,14%
29,59%
1,43%
0,31%
100,00%
Granodiorites
Molassic ophiolitic
formation
0,22%
Recrystallized
limestones and
dolomites
Granites
Canyos , Deeply
Incised Streams
Schist and marble
alternations
Holocene
Table 1. Landform classes related to the geological setting
These are mainly in situ deposits formed by the weathering of the granites.
The
metamorphic rocks, covering almost the whole of the eastern part of the island, are related
approximately 50% to open slopes (Table 1). However, a significant percentage of metamorphic
rocks are characterized as upper slopes, U-shaped valleys and deeply incised streams (Table 1).
Taking everything into consideration, it appears to be that the study area could be divided
into two landform units (Fig. 6). Landform Unit 1 (LU1) covers the western part of the island where
granites are the dominant geological formation. Subsequently, Landform Unit 2 (LU2) covers the
94
Unauthenticated
Download Date | 6/17/17 1:59 AM
eastern part of Ikaria, characterized by the presence of Atheras Mount. On the western part, the
terrain seems to be smoother than the eastern one, since 30% of the slopes are classified from level
to moderate (6o-35o) (Fig. 6). In general, the central part of LU1 (upland) is characterized by near
flat areas directed SW-NE that transform into steep slopes just before reaching the beachfront. In
contrast, LU2, presents a more complex landscape associated mainly with the hydrographic
network. Therefore, slopes ranging from level to moderate are just over 15%. Steep slopes are the
dominant class reaching up to 68.04% (Fig. 6). Another index that shows the difference between
LU1 and LU2 is the valley depth. Specifically, over 45% of the LU1 presents valleys that are no
more than 20 meters deep (Fig. 6). In contrast, almost 70% of LU2’s valleys are deeper than 20
meters (Fig. 6). Concerning the ruggedness of the terrain, LU2 appears to be far more rugged than
LU1 (Fig. 6). In particular, almost 55% of the LU2 area ranges from highly rugged to extremely
rugged. Similar LU1 terrain characteristics are limited 37% (Fig. 6).
Figure 6. Spatial distribution of Landform Units (LUs) and local gradient statistics (slope, valley
depth, TRI)
5. DISCUSSION
All in all, it seems that both the geological setting and the topography of the island are
contributory factors to the geomorphological evolution of the study area. The landscape of LU1 is
mainly affected by the granites that cover almost the whole area. Thus, there is no highly developed
hydrographic network to boost the geomorphological processes of erosion and deposition.
Therefore, the Holocene deposits are mainly associated with the weathering of granites (in situ
deposition on upland or side debris along the cliffy shore). As a result, the landscape of LU1 is
characterized by an extensive near flat upland directed smoothly to the NE seafront, and cliffy
coasts on all other directions. In contrary, the geological setting of the LU2 consisting mainly of
gneisses and carbonate metamorphic rocks, allowed the development of an extensive hydrographic
network on both sides of Atheras Mount that shaped the local topography. In general, by the power
of fluvial processes, the relief of LU2 is getting smoother when reaching the beachfront.
95
Unauthenticated
Download Date | 6/17/17 1:59 AM
6. CONCLUSIONS
It is an undeniable fact that automated landform classification is a powerful geoprocessing
technique for geomorphometry. However, it should be stated that landform analysis can be
succeeded only by relating the TPI calculated landform classes with the geological setting of the
study area. Moreover, the DEM derived gradients (slope, valley depth, terrain ruggedness)
contributed significantly to the interpretation of the landscape of Ikaria Island that was further
divided in two landform units. Concerning future GIS applications, more DEM derived gradients
such as local relief, curvature, topographic openness, etc. could be analysed for morphometric
purposes. Lastly, the implementation of an algorithm based on pattern recognition (texture) rather
than differential geometry (geomorphometry) may deliver more accurate results along the coast.
REFERENCES
Bozkurt, E. & Oberhansli, R. (2001), Menderes Massif (Western Turkey): structural, meta-morphic
and magmatic evolution – a synthesis. International Journal of Earth Sciences, 89:4, 679 –
708, https://doi.org/10.1007/s005310000173
Conrad, O., Bechtel, B., Bock, M., Dietrich, H., Fiscer, E., Gerlitz, L., Wehberg, J., Wichmann, V.,
Bohner, J. (2015), System for Automated Geoscientific Analyses (SAGA) v. 2.1.4. Geosci.
Model Dev., 8, 1991-2007, https://doi.org/10.5194/gmd-8-1991-2015
De Reu, J., Bourgeois, J., Bats, M., Zwervaegher, A., Gerolini, V., De Smedt, P. Chu, W., Antrop,
M., De Maeer, P., Finke, P.; Van Meirvenne, M., Verniers, J., Crombe, P. (2013), Application
of the topographic position index to heterogeneous landscapes. Geomorphology, 186, 39-49,
http://doi.org/10.1016/j.geomorph.2012.12.015
Martin-Duque, J. F., Godfrey, A. E., Pedraza, J., Diez, A., Sanz, M. A., Carrasco, R. M., Bodoque,
J. M. (2003), Landform Classification for Land Use Planning in Developed Areas: An Example
in Segovia Province (Central Spain). Environmental Management, 32:4, 488-498,
http://doi:10.1007/s00267-003-2848-2
Grohmann, C.H. & Riccomini, C. (2009), Comparison of roving-window and search-window
techniques for characterizing landscape morphometry. Computers and Geosciences, 35:10,
2164-2169, http://doi.org/10.1016/j.cageo.2008.12.014
Ho, L. T. K. & Umitsu, M. (2011), Micro-landform classification and flood hazard assessment of
the Thu Bon alluvial plain, central Vietnam via an integrated method utilizing remotely sensed
data. Applied Geography, 31:3, 1082–1093, http://doi.org/10.1016/j.apgeog.2011.01.005
Hoersch, B., Braun, G., Schmidt, U. (2002), Relation between landform and vegetation in alpine
regions of Wallis, Switzerland. A multiscale remote sensing and GIS approach. Computers,
Environment and Urban Systems, 26:2-3, 113-139, http://doi.org/10.1016/S01989715(01)00039-4
Ilia, I., Rozos, D., Koumantakis, I. (2013), Landform classification using GIS techniques. The case
of Kimi municipality area, Euboea Island, Greece. Bulletin of the Geological Society of Greece,
Vol 47, 264-274, http://dx.doi.org/10.12681/bgsg.10940
Κtenas, Κ. & Μarinos, P. G. (1969), Ikaria island geology. ΙGME, vol. ΧΙΙΙ, VOL Geological and
Geophysical research, Νο 2, Αthens.
Macmillan, R. A., Martin, T. C., Earle, T. J., Mcnabb, T. H. (2003), Automated analysis and
classification of landforms using high-resolution digital elevation data: applications and issues.
Canadian Journal of Remote Sensing, 29:5, 592-606, http://doi.org/10.5589/m03-031
Mcnab, W. H. (1993), A topographic index to quantify the effect of mesoscale landform on site
productivity.
Canadian
Journal
of
Forest
Research,
23:6,
1100-1107,
http://doi.org/10.1139/x93-140
96
Unauthenticated
Download Date | 6/17/17 1:59 AM
Planchon, O. & Darboux, F. (2001), A fast, simple and versatile algorithm to fill the depressions of
digital elevation models. Catena, 46:2-3, 159-176, http://doi.org/10.1016/S03418162(01)00164-3
Photiades, A. (2005), Ikaria Island Sheet. Program for complying the geological map of Greece in
scale 1:50.000. IGME editions.
Riley, S.J., De Gloria, S.D., Elliot, R. (1999), A Terrain Ruggedness that Quantifies Topographic
Heterogeneity. Intermountain Journal of Science, 5:1-4, 23-27.
Rigol-Sanchez, J.P., Stuart, N., Pulido-Bosch, A. (2015), ArcGeomorphometry: A toolbox for
geomorphometric characterization of DEMs in the ArcGIS environment. Computers &
Geosciences, 85:A, 155-163, http://doi.org/10.1016/j.cageo.2015.09.020
Senthilvelan, A. (2015), A GIS Based Study on Slope Characteristics of Porandalar Watershed,
Amaravathi Sub-Basin, Tamil Nadu. Indian Journal of Applied Research, 5:12, 1-3.
Stiros, S. C., Laborel, J., Laborel-Deguen, F., Morhange, C. (2011), Quaternary and Holocene
coastal uplift in Ikaria Island, Aegean Sea. Geodinamica Acta, 24:3-4, 123-131.
Tagil, S. & Jenness, J. (2008), GIS-based automated landform classification and topographic,
landcover and geologic attributes of landforms around the Yazoren Polje, Turkey. Journal of
Applied Sciences, 8:6, 910-921, http://doi.org/10.3923/jas.2008.910.921
Verhagen, P. & Dragut, L. (2012), Object-based landform delineation and classification from
DEMs for archaeological predictive mapping. Journal of Archaeological Science, 39:3, 698703, http://doi.org/10.1016/j.jas.2011.11.001
Weiss, A.D. (2001), Topographic position and landforms analysis. ESRI Users Conference, San
Diego, CA, USA.
97
Unauthenticated
Download Date | 6/17/17 1:59 AM