Pedosphere 23(6): 767–778, 2013 ISSN 1002-0160/CN 32-1315/P c  2013 Soil Science Society of China Published by Elsevier B.V. and Science Press Qualitative Land Suitability Evaluation for Main Irrigated Crops in the Shahrekord Plain, Iran: A Geostatistical Approach Compared with Conventional Method *1 Y. SAFARI 1 , I. ESFANDIARPOUR-BOROUJENI 2,∗2 , A. KAMALI 2 , M. H. SALEHI 3 and M. BAGHERI-BODAGHABADI 3 1 College of Agriculture, Vali-e-Asr University of Rafsanjan, Rafsanjan 77139-36417 (Iran) 2 Department of Soil Science, College of Agriculture, Vali-e-Asr University of Rafsanjan, Rafsanjan 77139-36417 (Iran) 3 Department of Soil Science, College of Agriculture, Shahrekord University, Shahrekord 88186-34141 (Iran) (Received January 10, 2013; revised September 16, 2013) ABSTRACT The purity of soil map units and their quality for various uses like land suitability evaluation are always questioned. The main objective of this study was to compare the precision of qualitative land suitability classification based on geostatistical and conventional soil mapping methods for main irrigated crops in the Shahrekord Plain, Central Iran. A regular grid sampling method consisting 104 sample points was designed and soil samples were collected. Ordinary kriged maps were achieved for all studied soil attributes after physico-chemical analyses. Afterward, to combine kriged maps and ecological requirements of the studied crops, a script was designed in ILWIS 3.4 software and consequently, kriged qualitative land suitability maps were generated. Conventional qualitative land suitability was also mapped based on the representative pedon analysis in each soil map unit. Finally, comparison of the conventional and kriged maps was carried out using the statistical method, error matrix. The results showed that the overall accuracies of wheat, sugar beet, potato and alfalfa maps were 39.8%, 24.3%, 18.7% and 18.6% at subclass category, respectively, whereas these values increased to 80.9%, 82.3%, 23.7% and 82.3% at class level, respectively. Hence, it can be stated that thanks to the relative facility of conventional soil survey compared with geostatistical methods, this method can be expressed as a preferable way for handling a usual land suitability evaluation design; but using soil map units as land suitability delineations may lead to unsatisfactory results in estimation of quantity and type of existing limitations. Key Words: land use planning, ordinary kriging, representative pedon, soil variability Citation: Safari, Y., Esfandiarpour-Boroujeni, I., Kamali, A., Salehi, M. H. and Bagheri-Bodaghabadi, M. 2013. Qualitative land suitability evaluation for main irrigated crops in the Shahrekord Plain, Iran: A geostatistical approach compared with conventional method. Pedosphere. 23(6): 767–778. INTRODUCTION Land use planning (LUP) is one of the most im- portant objectives of land suitability evaluation (LSE) (FAO, 2007; Niekerk, 2010). A good understanding of soil properties’ spatial variability pattern is essential requirement for accurate LUP and precise land ma- nagement in the soil maps (Miao et al., 2006; Santra et al., 2008). However, land suitability units are sepa- rated regardless these variations. In other words, ab- breviation of the soil variability thorough a soil map unit to a typical value, i.e., representative pedon, may cause the precision of the land suitability maps to be declined, especially when these maps are used in site- specific soil and crop management (Salehi et al., 2003; Ziadat, 2007). Additionally, conventional soil survey according to individual observations ignores the conti- nuous nature of soil and landscape variation, resul- ting in the misclassification of sites to discrete and sharply bounded areas (Keshavarzi et al., 2010). Hence, some researchers in recent years have tried to characte- rize the soil properties’ spatial variability patterns and their effects on land suitability in different scales (Mo- hamed, 2000; Braimoh et al., 2004; Emadi et al., 2010; Keshavarzi et al., 2010). To achieve such an important purpose, the spatial estimation methods using statisti- cal and geostatistical estimators have been extensively used (Li and Heap, 2008). Now that the differences of soil attributes are studied regardless spatial correlation between soil samples in the classical statistics (Webster and Oliver, 2001; Junior et al., 2006), soil scientists have focused on the geostatistical methods, i.e., dif- ∗1 Supported by the Vali-e-Asr University of Rafsanjan, Iran (No.518). ∗2 Corresponding author. E-mail: iesfandiarpour@yahoo.com.