Istanbul 2004 
t Information 
naging, CIPA 
ctives 10 save 
n. Psprin ger- 
Journal of 
gn with the 
don. 
ge by means 
formation on 
urkey. 
y. 2., 2001. 
| University, 
\nalysis and 
abase. In 
pplications, 
il, Mapping 
DETERMINING CURRENT LAND COVER AND POTENTIAL HAZELNUT 
PLANTATION AREAS USING GIS & RS 
IN NE BLACK-SEA REGION OF TURKEY, TRABZON 
S.Reis ? > 
7 . a 
, I. Yomralioglu 
ar = . . = . ; : 
KTU, Faculty of Architecture and Engineering, Dept. of Geodesy and Photogrammetry Engineering, GISLab, 
61080, Trabzon, Turkey — (sreis, tahsin) @ktu.edu.tr 
Commission PS WG 11/6 
Keywords: GIS, RS, Land Cover, Agriculture, Analysis, Landsat ETM+, Hazelnut, Trabzon 
ABSTRACT 
In Turkey, hazelnut grows mostly in the Eastern Black Sea Region of the country, and has also a vital impact on regional economical 
development. Most important task is to determine what areas exactly hazelnut grows and how it can be grown-up productively in the 
region. The principle aim of this study is, firstly, to create a methodology for determining existing and potential hazelnut plantation 
areas, using GIS/Remote Sensing techniques in order to help hazelnut policies developers and economists. This study was done in the 
province of Trabzon which is one of the most hazelnut produced area of Turkey. In this study, Landsat ETM+ image was used to 
generate a current land cover classification. Using the supervised classification method, overall accuracy was determined as 84.7%. 
As for hazelnut growing areas, an 87.9% overall accuracy percentage was reached. It was determined that there is an area of 71.401 
hectares hazelnut plantation area in the Trabzon province. Because hazelnut production exceeds the consumption in Turkey, some 
new regulations have been released to find new land use policies since 1983. By putting into practice these regulations that restrict 
hazelnut plantation areas, a more efficient and productive hazelnut harvest policy should be reached. As a result, potential hazelnut 
areas were determined and examined with respect to raster-based spatial analysis functions of GIS. 
I. INTRODUCTION 
Turkey has the highest hazelnut production rate in the world. 
The rate is 77% of the whole world production (Koksal, 
2000). Hazelnut is harvested mostly in the Black Sea Region 
of Turkey. There are shortages in such data as hazelnut 
growing up areas, characteristics of the hazelnut gardens 
(topography, soil type, age and densities of hazelnut groves, 
number of the types in the hazelnut gardens, etc.), consumed 
hazelnut amounts in domestic market and remaining hazelnut 
amount in private sector at the end of the each harvest season 
in Turkey. Especially in the studies specifying hazelnut 
locations, valuable data is much more intensive. However, 
the most important and primary requirement of having an 
effective hazelnut policy is the existence of the reliable data 
on these issues (State Planning Organization, 1999). 
In order to supervise hazelnut harvesting process to increase 
production rates in every year, spatially based information 
produced by different institutions should be accurately 
collected and analyzed in such an effective way. Today, 
Geographic Information System (GIS) is a powerful tool for 
collecting, processing and analyzing of spatial information 
(Longley, et al , 2001; Yomralioglu, 2002). Remote Sensing 
(RS) has also an important role providing data in quick mode 
for GIS, especially for determining of land cover analyses 
(Maxwell, et al, 2003; Oetter, et al, 2000). RS is accepted as 
an effective way to acquire the data for agricultural products 
with accurate and time related manner in wide areas. 
Especially, Landsat satellite images have been used 
successfully for classifying different crop types since 1970 
(Lobell, et al, 2003; Cohen et al., 2002). 
Corresponding author 
In the early 1980’s, because of the hazelnut production was 
more than the domestic consumption, exportation, and raised 
stocks, restriction of the hazelnut areas was raised. A Law 
numbered 2844 and named as “The Act about Planning of the 
Hazelnut Production and Determining of the Planting Areas” 
was put into practice in 1983 to provide the preduction is 
realized in most appropriate areas and respect to demand 
(Turkey Official Gazette, 2002). With this law, it has been 
targeted that the hazelnut areas would be reduced by 100.000 
hectares (European Commission, 2000). With the new 
arrangements with this law, the provinces where hazelnut 
harvested were restricted and some criteria were introduced 
for the production process for some provinces in 2002. Thus, 
it has been aimed that the areas suitable for hazelnut 
harvesting and the rest of it could be specified. However, 
applying of this law has not been realized yet. In this study, a 
procedure has been developed for specifying the most 
appropriate hazelnut growing areas in order to provide a base 
data for agricultural decision-makers. 
2. METHODS 
2.1 The Study Area 
The Province of Trabzon is situated between 39° 15' - 40° 15' 
WE Longitudes and 41^ 8' - 40° 30' NS Latitudes in the 
middle of East Black Sea Region of Turkey (Figure 1). In the 
province, increasing elevation beginning from the sea-level 
exceeds 3000 m in some areas. Generally, mountains, hills 
and high plateaus take part in inner land of the region. The 
province is the most important trade centre in the region with 
the features of having a rich historical and cultural wealth.