Probability mapping of Deficiency or Excess of Soil Available Phosphorus and Organic Carbon by Indicator Kriging

Probability mapping of soil proprieties Including soil organic carbon and nutrients based on threshold values play important role in optimum land management and sustainable agriculture development. In order to map based on critical limits need to use non parametric geostatistics methods such as kriging indicator. In this study, using indicator kriging, probability map of deficiency or excess of soil available phosphorusand organic carbon was prepared in part of Eslamabad-e Gharbagriculture research station in Kermanshah province. The soil legacy data including 217 surface soil samples on an almost regular grid, about 50 m apart is used to perform research. After preliminary statistical processing, indicator variogram was drawn based on threshold values 1.2% and 15ppm for soil organic carbon and available phosphorus, respectively. Exponential model as the best model fitted to indicator variogram is selected based on minimum residual sum of squares. The necessary data for validation was achieved through leave-one-out cross-validation (LOOCV).For assessing the performance of indicator kriging is used the area under the curve (AUC) of receiver operating characteristic (ROC). Area under the ROC curve was calculated for phosphorus and soil organic carbon was 0.796 and 0.800, respectively. These results indicated that the indicator kriging to predict these variables have relatively good performance. In general, it can be said that using fertilizer based on probability maps of soil fertility properties lead to reduce cost and increase sustainable food production.