ds at the object 
n using Landsat 
Turkey, the best 
nd cover types is 
nagery for these 
. land cover/use 
the Atatürk Dam 
th water reserve 
am Lake with a 
nded land is 817 
Atatürk dam is 
n area of 8,724 
jir is 526 m. The 
rigated from the 
entral part of the 
rigation District. 
n of rolling hills 
ria (Fig 2). Two 
in the study are 
9?Q'Q"E 
(August 2011). 
ey with the study 
le. 
es of the Eastern 
ence. The annual 
    
  
  
  
  
  
average temperature is 18 °C and annual rainfall is around 350 
mm. There is significant seasonal variation in precipitation, with 
most precipitation occurring between November and April. The 
area receives almost no rain during the summer, at which time 
irrigation becomes crucial. 
The Sanliurfa-Harran Plain Irrigation Project is the first realized 
scheme within the Lower Euphrates Project. The water is 
brought to Sanliurfa- Harran Plain by the Sanliurfa tunnel 
system consisting of two parallel tunnels each 26.4-km long and 
with an 7.62-m inside diameter and a carrying capacity of 328 
m? /sec. These tunnels were completed successively in 1995 and 
in 1998. The irrigation capacity of Sanliurfa and Harran canals 
are almost 500 km? and 1000 km”. 
Several different crops including cotton, cereals, maize, and 
vegetables are cultivated in the Harran Plain. While there is a 
variety, cotton and cereals dominate the agricultural scene in 
any given year (Ozdogan et al., 2006). 
3. METHODOLOGY 
Remote sensing has been an effective tool for monitoring 
imigated fields under a variety of climatic conditions and 
locations. The thematic analysis of multi-temporal data series 
requires differences between images to result exclusively from 
changes in surface properties, necessitating a precise geometric 
and radiometric correction of incorporated images (Song et al., 
2001). 
Monitoring the changes in the Atatürk Dam Lake and summer 
irrigated fields in the Harran Plain require multiple sources of 
data. In the first part of this study, land cover/use changes on 
agricultural fields under the Atatürk Dam Lake and its vicinity 
have been identified between the periods of 1984 to 1992. 
Reserve changes and inundated agricultural fields have been 
identified by change detection using multi temporal Landsat 
imagery within these periods. 
After the 10-year period of completion and the filling up of the 
reservoir in 1992, Landsat and meteorological time-series 
analyses are examined to assess the impact of the Atatürk Dam 
Lake on irrigated agricultural areas in the Harran Plain. For the 
last 9-year period from 2002 to 2011, the relationships between 
seasonal water reserve changes and irrigated plains under 
changing climatic factors primarily driving vegetation activity 
on the watershed have been analyzed consecutively using the 
appropriate tools. A total number of 99 Landsat images have 
been used in order to constitute time series analysis and to 
determine the changes on these fields in conjunction with 
climatic datasets. For all images, geometric corrections 
including digital elevation information and Tasseled Cap 
transformations were carried out to attain changes in surface 
Res and denoting disturbance of Landsat reflectance 
ata. 
Tasseled Cap transformation was originally developed for early 
Landsat sensors (multispectral scanner and thematic mapper) 
(Crist & Cicone, 1984; Kauth & Thomas, 1976), its linear 
coefficients have more recently been modified for applicability 
to Enhanced Thematic Mapper Plus (ETM+) imagery. In order 
to use the tasseled cap coefficients (Huang et al., 2002) for the 
Landsat 5 TM sensor, conversion of the Landsat 5 TM DN data 
into data that is equivalent to data recorded by the Landsat 7 
ETM+ sensor is needed due to the calibration differences 
between the two sensors. This process is described by 
Vogelmann et al. (2001) in reverse; that is, they converted from 
  
  
   
   
   
   
   
  
   
  
   
   
  
     
   
    
  
   
    
  
   
    
     
   
   
   
   
  
   
   
  
   
   
  
  
  
  
  
  
  
   
  
   
   
   
   
   
   
  
  
  
  
   
  
  
   
  
  
Landsat 7 ETM+ data to Landsat 5 TM equivalent. To convert 
from Landsat 5 TM DN data to Landsat 7 ETM+ DN data, the 
following expression is used (Eq. 1): 
DN! —(slope, * DN5) -intercept ; (1) 
where DN7 is the Landsat 7 ETM+ equivalent DN data, DNS is 
the Landsat 5 TM DN data, and the slope and intercept are 
band-specific numbers. 
Before converting to reflectance data, all images with DN 
values were converted to radiance. While radiance is the 
quantity actually measured by the Landsat sensors, a conversion 
to reflectance facilitates better comparison among different 
scenes. It was obtained by removing differences caused by the 
position of the sun and the differing amounts of energy output 
by the sun in each band. The reflectance can be thought of as a 
“planetary albedo,” or fraction of the sun’s energy that is 
reflected by the surface. During the conversion from DN data to 
reflectance, it is possible to create small negative reflectance 
values which are set to zero. 
Each image was taken during the mid-to-late summer months 
from 1992 through 2011 in order to investigate the change 
detection on Harran Plain and environmental impacts due to 
water reserve changes in Atatürk Dam Lake. All images were 
radiometrically corrected and then converted into reflectance 
values and then tasseled cap procedure was used to create a 
vegetation index that measures three vegetation dimensions— 
brightness, greenness and wetness (Crist and Krauth, 1986). 
Table 1 gives an overview of the Landsat based Tasseled Cap 
coefficients (Huang et al., 2002). 
Table 1. Tassled Cap coefficients for Landsat (Huang et al., 
2002). 
Band 1 Band2 Band 3 Band 4 Band5 Band 7 
B 0.3561 0.3972 0.3904 0.6966 0.2286 0.1596 
G  -0344  -03544  -0.4556 0.6966 -0.0242  -0.2630 
W 0.2626 0.2141 0.0926 0.0656 -0.7629 -0.5388 
  
  
Remote sensing change detection is performed using the 
Disturbance Index described in Healey et al. (2005), an index 
specifically designed to detect changes in vegetated land cover 
types. The Disturbance Index is a transformation of the Tasseled 
Cap data space and is calculated using the three normalized 
Tasseled Cap indices (brightness, greenness and wetness) from 
Landsat TM/ETM--data (Healey et al., 2005; Kauth & Thomas, 
1976; Masek et al., 2008). The index is computed as a linear 
combination of the three normalized Tasseled Cap values (Eq. 
2): 
G-G WW 
A eran) 
Bg Go W, 
  
  
where B,, G, , W, are the normalized (rescaled) brightness, 
greenness, and wetness, indices respectively, and B PE Gu» 
Wu and By, Gg, W,, are mean and standard deviation of 
these three Tasseled Cap spaces. The re-scaling process 
normalizes pixel values across Tasseled Cap bands with respect 
to overall changes in reflectance, such as seasonal changes or 
changes induced by directional reflectance effects, thereby