and 
ado 
of 
ter, 
| de 
la y 
  
Musaoglu, Nebiye 
  
MULTITEMPORAL FOREST CHANGE ANALYSIS 
A CASE STUDY : SARIYER REGION 
Nebiye MUSAOGLU 
nmusaoglu@srv.ins.itu.edu.tr 
Istanbul Technical University Civil Engineering Faculty 
Geodesy and Photogrammetry Department, Remote Sensing Division 
80626 Maslak/ISTANBUL 
Working Group VII/2 
KEY WORDS: Sariyer, Landsat TM, GCP, RMS, ISODATA 
ABSTRACT 
In this study, differentiability of vegetation at Sariyer, an area with rich forest due to its geographical situation and 
general characteristics, has been analysed through the satellite images used jointly with ground data, and temporal 
changes which occurred in forest lands have been determined. In order to evaluate multitemporal data sets (June 1984 
and June 1997 Landsat TM) Sariyer region, located in the north of Istanbul was selected. In differentiating types of 
trees through satellite images, classification algorithms have been used by taking as basis the ground measurements 
carried out at Istanbul University, The Forest of Education and Research of the Faculty of Forestry which was selected 
as the test area. Firstly uncontrolled classification and then controlled classification algorithms have been applied to the 
test areas with different dates and to the entire area of study. 
Accuracy assessment were made to the classification results and the zonal values obtained were temporally compared. 
1 INTRODUCTION 
A major role that remote sensing and GIS techniques serve is to inventory and monitor biophysical and man-made 
features on the surface of the Earth. While some of the data may be static and does not change over time, much of the 
biophysical and man-made features are dynamic and constantly changing. It is important that such changes be 
inventoried accurately so that the physical and human processes at work can be more fully understood (Estes, 1992; 
Jensen and Narumalani, 1992; Ramsey, 2000). In order to maintain his life and in the meantime leave the best for the 
generations to come, mankind has to utilize the natural resources he has in the most economic manner and protect them. 
Parallel to the developing technology and increasing scientific researches today, advancements in the science of remote 
sensing are major helpers of mankind in this field. Changes made on the surface of the earth today are more extensive 
and occur more rapidly than ever before. Planners and resource managers, need a reliable mechanism to assess these 
consequences by detecting, monitoring and analysing land use changes quickly and efficiently. To meet the demand for 
current and accurate data multitemporal remotely sensed images are increasingly used as one of the data sources of land 
use change analysis. 
It is necessary to monitor and know at certain time intervals the amounts and the changes in time ofthe forest resources 
in terms of parameters such as tree wealth, growing sites and tree species (Coskun et al., 1998). As the forest inventory 
to be made by means of ground studies is a time-consuming, costly and difficult task, the studies which have been made 
show that satellite images provide, for this purpose, sufficient and reliable information in a short period of time. 
Satellite data are important sources of information to monitor at a regional scale the use of forests and land 
(Varjo, 1995; Botkin, 1984; Elijah, 1996). Studies which have been carried out through the satellite data at various 
dates are being widely used in analysing the use of land as well as the temporal changes in the shorelines and in the 
forest ecosystem(Peters et al., 1993; Jacobberger - Jellison, 1994; Chavez and MacKinnon, 1994; Johnson, 1994; 
Dobson et al., 1995; Michener and Houjoulis, 1997). In forestry works and differentiating the types of forest stands, 
generally 3., 4. and 5. channels of Landsat TM sensor are used. Satellite images, which are widely used in the topics of 
  
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam 2000. 939