Analysing Land Cover Change with RS and GIS Methods in

advertisement
Analysing Land Cover Change with RS and GIS Methods in Osmancik Basin and its
Surroundings
Karatepe, A.
akaratepe@sakarya.edu.tr
İkiel,C.
cikiel@sakarya.edu.tr
Abstract:
This research involves land cover change of Osmancik Basin. The Basin, is located in the northern part
of Turkey, between 40⁰ 48’- 41⁰ 10’ north latitudes and 34⁰ 41’- 35⁰ 04’ east longitudes. Osmancik
Basin, which the tectonic process is effective, formed by accumulating of sediments bringing by
Kızılırmak river. Natural vegetation is involved in the Euxine province of Euro - Siberian
phytogeographical region. However natural vegetation, which is the large part of land cover in the
research area, is rapidly changed according to rapid urbanization and industrialization activities.
In this area, that stated as tree see and partly boggy in 1700, firstly agricultural areas opened. Also in
twenty century seen that these agricultural areas changed to urban and industrial areas.
The aim of this study is to research the permanent effects of human activities on natural land cover
change. For this purpose, satellite images which enclose 1987 - 2010 periods are used as data. These
images are analysed by using data images (Landsat TM) processing techniques Erdas Imagine 10.0 and
ArcGIS 10.0 software. Land cover change nomenclature is made due to Corine (Coordination of
Information on the Environment) level 2 (1- Urban fabric, 2- Forest, 3- Rice fields, 4- Shrub and/or
herbaceous vegetation association 5- Heterogeneous areas 5- İnlands wetlands). Furthermore,
observations made on this land are confirmed by the results of analysis of the images.
Keywords: Land cover change, Landsat TM, Osmancik Basin
1. Introduction:
The observed biophysical cover of the earth’s surface, termed land-cover, is composed of patterns that
occur due to a variety of natural and human-derived processes (Karnieli and Rozenstein , 2011). In
other words land cover is defined as the bio-physical features one can observe on the surface of the
earth (Di Gregorio and Jansen, 2000). On the other hand, Land-use refers to human activity on the land
that influenced by economic, cultural, political, historical, and land-tenure factors. Remotely-sensed
data (i.e., satellite or aerial imagery) can often be used to define land-use through observations of the
land-cover (Karnieli and Rozenstein, 2011).

Sakarya University, Arts and Sciences Faculty, Department of Geography, Sakarya, Turkey
Over the past few decades, land cover change has taken place around the study area. During this
process, many rural lands, such as forests and wetlands, have transformed to human settlements.
Initially, forest areas have been afforested than converted to the agriculture areas than finally
agricultural areas opened for urban settlements.
In this study, land cover changing of Osmancik Basin and its vicinity has examined by using remotely
sensed data from 1987 to 2010. During the study, Landsat Thematic Mapper (TM) images were used.
Because previous studies show that Landsat Thematic Mapper (TM) images with the spatial resolution
of 30 m are sufficient to accurately classify a large variety of landscapes from the homogeneous
tropical landscapes to the heterogeneous Mediterranean landscapes ((Karnieli and Rozenstein , 2011).
2. Study Area:
Study area (Figure. 1) is located in the western part of the Black Sea Region of Turkey, Osmancık
district of Çorum Province. Osmancık plain with an area of 300 square kilometers is located at the
bottom of the basin. Study area has a complex evolutionary history. Tectonic movements, volcanic
activities and erosional and sediment transportation processes have played important role in that
evolution.
Figure 1: The Location map of study area.
Osmancık Basin takes place on the North Anatolian fault zones which are the major tectonic structures
of Turkey. Due to this, seism and volcanism played an important role in the formation of Osmancık
Basin. On the other hand, fluvial processes have also important role for shaping of the basin. Thus
Kızılırmak river, which is the longest river of Turkey, enters the study area from the south western, and
it goes upward to the north a couple of kilometers and then turns left, approximately after 10 km, the
river leaves the basin from the north western side.
This area lies on the transition zone between, Black Sea climate zones from the north and continental
climate zone from the south. Average annual precipitation is 385 mm, temperature is 13, 5 C. The
study area is characterized by semi-arid climate according to Thornthwaite (1948) climate
classification.
Major economic activities in study are farming, livestock and industry, respectively. Rice farming is
very important on the alluviums near the Kızılırmak River. On the other hand, the city has also some
factories, such as brick-tile, rice and textile. In terms of the urban settlement, the city is getting bigger
because of the migration to the city from the neighbouring villages. Due to this, agricultural areas have
opened for the urban settlement.
3. Data:
A Landsat TM image acquired on July 27, 1987 and other Landsat TM image acquired on August 27,
2010 both of them with 30 meter spatial resolution were used to assess the land cover changes in the
study area. All these images were provided by a commercial data provider. The images were chosen to
be as closely as same date. The images were also required to have less than 20 % cloud cover. With
these criteria all images were cloud cover of 0 %. The characteristics of the image data are presented in
Table 1.
The other data used in this study for reference and analyses mainly include: (1) topographic maps at a
scale of 1/100.000; (2) detailed vegetation map obtained from TR Forestry and Water Affairs Minister
belongs to 2002; (3) ground reference data obtained from land survey with hand held GPS.
Type of imagery
Date
Spatial resolution (m)
Landsat TM
27.07.1987
30
Landsat TM
27.08.2010
30
Table1: Characteristics of the satellite data used for land cover change mapping in the study area.
4. Methodology:
The image processing was performed in four stages: 1) image pre-processing, 2) determination of land
cover types, 3) image classification 4) accuracy assessment. These applications were performed using
ERDAS Imagine 10.0 software. The flowchart of the image processing techniques was indicated in
Table 3.
4.1 Image pre-processing:
All images were clipped out according to the study area by using “Subset” function. The radiometric
corrections, systematic errors and Landsat TM images corrected using linear polynomial method based
on 1/100.000 scaled topographic map (produced by the General Directorate of Mapping) with 8 control
points and an RMS error of less than 0.5 pixels. All images were geo-referenced into the Universal
Transverse Mercator-UTM, WGS-84.
4.2 Determination of land covers types
The CORINE land cover nomenclature/ classification system was chosen and referred for the
classification system for this study. This methodology developed by the Joint Research Center, a center
affiliated with the European Union, utilizes LANDSAT TM … as its main database (European
Environment Agency, 2005). It was established by the European Commission to create a harmonized
Geographical Information System on the state of the environment in the European Community. The
legend of the CORINE land cover nomenclature/ classification is standard for the whole of Europe,
which as a result is quite extensive with 44 classes describing land cover (and partly land use)
according to a nomenclature of 44 classes organized hierarchically in three levels. The first level (five
classes) corresponds to the main categories of the land cover/land use (artificial surfaces, agricultural
areas, forests and semi-natural areas, wetlands, water bodies). The second level (15 classes) involves
the subclasses of the first level. The third and last level is composed of 44 classes and formed by the
sub-classes of second level classes (Heymann et al. 1994). In this study land cover legends include five
classes at the second level (Urban fabric, Heterogeneous agricultural areas, Forests, Heterogeneous
agricultural areas and Inland wetlands). The field work supported the image interpretation of land cover
types defined in the classification. Classification scheme and detailed descriptions were given in Table
2.
Class 1
Class 2
Urban fabric
Industrial, commercial and transport units
Artificial surfaces
Mine, dump and constructions sites
Artificial, nonagricultural vegetated areas
Arable land
Permanent crops
Agricultural areas
Pastures
Heterogeneous agricultural areas
Forests
Scrub and/or herbaceous vegetation associations
Forest and semi-natural areas
Open spaces with little or no vegetation
Inland wetlands
Wetlands
Maritime wetlands
Inland waters
Water bodies
Marine waters
Class 3
Continuous urban fabric
Discontinuous urban fabric
Industrial or commercial units
Road and rail networks and associated land
Port areas
Airports
Mineral extraction sites
Dump sites
Construction sites
Green urban areas
Sport and leisure facilities
Nonirrigated arable land
Permanently irrigated land
Rice fields
Vineyards
Fruit trees and berry plantations
Olive groves
Pastures
Annual crops associated with permanent crops
Complex cultivation patterns
Land principally occupied by agriculture, with
significant areas of natural vegetation
Agro-forestry areas
Broad-leaved forests
Coniferous forests
Mixed forests
Natural grasslands
Moors and heathland
Sclerophyllous vegetation
Transitional woodland-scrub
Beaches, dunes, sands
Bare rocks
Sparsely vegetated areas
Burnt areas
Glaciers and perpetual snow
Inland marshes
Peat bogs
Salt marshes
Salines
Intertidal flats
Water courses
Water bodies
Coastal lagoons
Estuaries
Sea and ocean
Table 2 : Corine Land Cover Nomenclature (European Environmental Agency, 2005)
4.3 Image classification
A good knowledge of the study area was achieved by a suitable image enhancement and related
literature. Furthermore Richards and Jia (1999) suggest fieldwork that develops knowledge of the area
with interviews, photography of characteristic surfaces, ground truth data in order to validate a
classification.
In this study, all images were independently classified using the supervised classification method of
maximum likelihood algorithm. Although many different methods have been devised to implement
supervised classification, the maximum likelihood is still one of the most widely used supervised
classification algorithms (Jensen, 1996). In supervised classification, spectral signatures are collected
from specified locations in the image by digitizing various polygons overlaying different land cover
types. The spectral signatures are then used to classify all pixels in the scene. Over 150 “user defined
polygon” were selected from the whole study area by drawing area of interest (aoi). In supervised
classification process, .aoi function reduces the chance of underestimating class variance since it
involved a high degree of user control. After the classification process, all signature sample points were
grouped as a class by “recode” function according to the determined land cover classification types in
this study (Table 2).
5. Results:
The outputs produced are land-cover data for Osmancik basin and its surrounding area for the years of
1987 and 2010. According to table 3 and figure 2-3, over 23-year period, it has seen that forests has
faced rapid decreased. Deforestation within this period was calculated 62 % between 1987 and 2010.
On the contrary, Inland wetlands have increased because of the built up new big dam named Obruk.
Table 3: Results of the land cover classification for 1987 and 2010 images showing area of each class,
class percentage and area changed.
Heterogeneous agricultural areas seem to be increased, but this condition could be depending on
converting forests to the opened agricultural areas in the rural settlements. In fact, in the city, arable
lands have been disappearing day by day because of the settlements. Especially, rice farming areas near
the Kızılırmak River have been opening for the new construction areas. So, in the same period, urban
fabric area has increased more than two times.
Figure 2: Land cover classification maps for each time step (1987 – 2010) according to the CORINE
Land Cover Nomenclature.
Figure 3: Land cover change ratio (%) throughout the study period (1987 – 2010).
6. Conclusions
This study evaluates land cover changes and urban expansion in the Osmancık Basin, Turkey, between
1987 and 2010 using satellite images. In this context, several image processing were applied to two
Landsat images collected over time (1987, and 2010) that provided recent and historical land cover
conditions for the study area.
The study area has undergone land cover change as a result of increasing population and
industrialization. A considerable increase in urban settlements has taken place as well as huge increase
in agricultural lands. The area of natural vegetation and agricultural have decreased considerably.
In this study, to investigate and to put forward these conditions that mentioned above, land cover
change from 1987 to 2010 was compared by using Remote Sensing and GIS. In the last few decades,
integrating GIS and Remote Sensing provided valuable information on the nature of land cover
changes.
Acknowledgement
This work was financially supported by the Scientific Research Project of Sakarya University.
References
Di Gregorio, A. and Jansen, L.J.M., (2001). Land-cover Classification System (LCCS): classification
concepts and user manual. FAO, Rome.
EEA, CORINE Land Cover, (2005). European Environment Agency. Acquired from
Heymann, Y., Steenmans, C., Croisille, G. and Bossard, M., (1994) CORINE Land Cover Technical
Guide (Luxembourg: Office for Official Publications of the European Communities).
Jensen J.R., (1996). Introductory Digital Image Processing: A Remote Sensing Perspective, 2nd
Edition, Prentice-Hall, Upper Saddle River, NJ.
Karnieli, A., Rozenstein, O. (2011). Comparison of methods for land-use classification incorporating
remote sensing and GIS inputs. Applied Geography 31 (2011) 533-544. Elsevier Publishing.
Richards, J.A.. and Jia, X., (1999). Remote Sensing Digital Image Analysis: An Introduction. 3rd
Edition, Springer, Berlin/Heidelberg/New York.
Download