ASSESSMENT OF MULTI-TEMPORAL LAND USE / COVER
CHANGES USING REMOTELY SENSED IMAGERY,
A CASE STUDY : TUZLA REGION IN ISTANBUL, TURKEY
Filiz SUNAR, Cankut ORMECI, Sinasi KAYA, Nebiye MUSAOGLU
|. T.U. Civil Engineering Faculty, Remote sensing Dept., Maslak 80626 Istanbul, TURKEY
Commission VII, Working Group 9
KEY WORDS : Remote Sensing, Land use, Landsat, Classification, Change Detection.
ABSTRACT
Natural environments are changed due to natural processes, disasters, human and other interventions. One of the
major human interventions is the unplanned urbanization. The rapid growth of population and illegal settlements in
urban centers and industrial areas have put enormous strains on urban services and amenities, which causes the
overgrowing pressure on land resources. To meet the demand for current and accurate data, multi-temporal remotely
sensed images are increasingly used as one of the data sources for land use change analysis. The objective of this
study was to detect land-cover changes by using 3 Landsat TM imagery in the Tuzla region, Istanbul. with different
change detection (i.e. image overlay, image differencing, principal component analysis) methods. Each change
detection methods used was assessed with their ability to detect specific changes.
1. Introduction
The urbanization is now being accepted as the
essential infrastructure for economic development of a
region and in index of material progress and prosperity
of society. The rapid growth of population and illegal
settlements (i.e., slums) in urban centers and new
industrial structures have put enormous strains on
urban services and amenities, which causes the
overgrowing pressure on land resources. One of the
major tasks is the acquisition and analysis of latest
information upon for the sufficient administration and
planning. The conventional methods of urban data
collection and analysis can not deliver necessary
information in a timely and cost effective fashion
(Michalak, 1993]. To meet the demand for current
and accurate data, multi-temporal remotely sensed
images are increasingly used as one of the data
sources of land use change analysis. Data from the
Landsat TM, with its synoptic and regular (16 day and
year-to-year) coverage offer potential for detection and
inventory of disturbances and other changes that occur
in land use, cover type and cover condition in areas of
interest [Ormeci et al, 1994].
In this study, land-cover changes and intense
development pressure were analyzed by using the
multi-temporal (12 June 1984, 1 September 1990 and
6 September 1992) Landsat TM imagery in the Tuzla
region, Istanbul with different change detection
methods with their ability to detect specific changes.
2. Study Area
During the last two decades, the urban expansion in
Istanbul, biggest city of Turkey, is mostly unplanned
and arbitrary which has an encroached upon the
agricultural land. In this study, the impacts of the
urban growth and land use changes in Tuzla region,
678
located in the east coast line of Istanbul, was
evaluated. At the beginning this region has developed
as an small village having flat, productive agricultural
lands and had a summer residential area characteristic
due to its natural beauties, than it showed rapid
developments because of the new transportation axes
and regional investments. This caused rapid
population growth due to migration from the rural
areas, and creation of new job facilities and services.
The other reason for the rapid growth of the
urbanization and land developments in the region was
the establishment of new dockyard in 1985. Today,
this area and the Lake Kamil Abdus, which is accepted
as an natural and archeological site and protection
area, are threaten by this unplanned expansion and
developments (Figure 1).
2.1. Materials and technique of analysis
To evaluate the population growth and unplanned
urbanization temporally, 12 June 1984, 1 September
1990, and 6 September 1992 dated 3 LANDSAT TM
satellite data (4 bands of each data set) of the study
area were used in addition to an earlier mosaic of
aerial photographs and field check. The key
characteristics of the data used are given in Table 1.
Table 1 . The key characteristics
of the data used.
Parameter : LANDSAT TM -
Date 12.08.1984
1.09.1990
6.09.1992
No. Bands 7
Wavelength Region 450-12500 nm
Pixel Size 30m
Thermal Band:120 m
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996