TURKISH JOURNAL OF AGRICULTURE AND FORESTRY, vol.38, no.6, pp.926-934, 2014 (SCI-Expanded)
The main objective of this study was to implement a spatial model with a multicriteria assessment approach for rice (Oryza sativa L.) land suitability by using geographic information system (GIS) techniques. The land mapping units resulting from the overlay of the selected theme layers had unique land quality information on which land suitability was based. The selected theme layers of rice include topographic factor (slope), soil physical factors (soil depth, soil texture, drainage, stoniness, hydraulic conductivity), and soil chemical factors (pH, electrical conductivity, CaCO3, soil fertility). These theme layers were collected from existing information. Spatial information of soil physical and soil chemical factors were formulated using a soil map database. Each land characteristic was also considered as a thematic layer in the GIS. In addition, each land quality layer associated with attributed data was digitally encoded in a GIS database. After combining these layers, a resultant map was produced. A land suitability rating model applied to the resultant polygonal layer provided the suitability classes for rice cultivation. Results showed that 79% of the study area was highly or moderately suitable for rice cultivation, whereas 21% of the area was unsuitable for rice cultivation due to soil and topographic conditions. The model was also validated with field studies. According to statistical analysis, there was a significant positive relationship between land suitability classes and grain yields. Grain yield was significantly affected by land suitability class at the level of P<0.001. As for LSD, 05 test results, the highest yield values of the 12 rice species studied were 7899 ha kg(-1) (cv. Hall Bey), 7605 ha kg(-1) (cv. Osmancik-97), and 7510 ha kg(-1) (cv. Duragan) in the 'highly suitable' class, whereas the lowest yield values were found to be 4764 ha kg(-1) (cv. Aromatik), 4403 ha kg(-1) (cv. Koral), and 4320 ha kg(-1) (cv. Gonen) in the 'marginally suitable' class. This study also confirms the ability of GIS to integrate spatial and attributed data and offer a quick and reliable method of land suitability assessment with high accuracy.