Comparison of geostatistical interpolation models (kriging) to estimate soil salinity and wheat yield (a case study: army field of Aq qala)

In the past decade, data acquired from the geographical information system (GIS), Global Positioning System (GPS) and geostatistics have an important role in the study of spatial distribution of soil properties and the results often show that changes in soil properties could occur from very small distances (a few mm) to long distances (several kilometers). This study compared different methods of interpolation (kriging) to determine the best model to map soil salinity and yield variables in the field of military, AQ-Qala. In order to investigate the effects of soil salinity and its changes during wheat growing season, 101 ground control points were taken in the field of military, AQ-Qala, based on systematic sampling selection, and EC and pH levels were measured in two stages along with corresponded yield in harvesting stage. In order to interpolate salinity levels, ordinary, universal and disjunctive kriging in combination with five models of semivarograms were tested. For this, the field was divided to four separate sections and the models were tested separately. Final results showed that among three kriging methods and five applied models, ordinary kriging with an exponential model and the universal kriging with exponential models were the best models to estimate soil salinity and wheat yield, respectively . In this study, a significant relation was found between salinity differences for two samplings and wheat yield, as 4.5 gr yield reduction was demonstrated per salinity unit increase. Also, results of soil testing revealed that EC value for each parcel is different than others which is related to topography and parcel area, as the parcel A with the least difference between ECs had the highest yield. Results showed that EC difference had a pronounced variation which could be used to interpret yield differences among four parcels. According to estimates by interpolation methods used for predicting wheat yield in 4 studied units (A, B, C and D), unit A had the most, while unit B the least yield range. Unit A had the lowest salinity and on the other hand, only one cultivar (Koohdasht) was grown in unit A, while in the other units more than one cultivar (Line 17, Morvarid, Koohdasht and N8019) was planted. Overall results revealed that GIS along with available information could be used as a powerful tool for detecting the effects of abiotic factors effects (including salinity) on the agroecosystems performance. Also, these results emphasizes on this reality that the fields are faced by widespread spatial variations of different factors which need different management options.