Using Stage Sampling Technique to Determine Saturated Hydraulic Conductivity Measurement Points on the Farm

Saturated hydraulic conductivity is important physical parameters of water movement in the soil, especially in determining the distance between the drain lines is the drainage projects. High cost and Time-consuming the field measurement and high variability of this parameter, a correct understanding of the changes in the amounts of saturated hydraulic conductivity is difficult in projects. So selecting the location and number of measuring is the first step to determining the spatial variability of Hydraulic conductivity which is corresponds to reality. Therefore, it is necessary to examine the impact the new sampling methods on the accuracy of geostatistics estimates. Optimal sampling design, is the stage sampling. In this method, sampling of a parameter will done in two or more stages and data from each step, will form the basis of the next step designation. This study is done in 40 thousand hectares, within Siahrud to Talar River of Mazandaran province to compare the stage sampling method based on the standard error of geostatistical methods, with conventional sampling methods (single-stage) on improving the estimation of spatial variability of saturated hydraulic conductivity. In this study, four different proposed scenarios were examined. One scenario based on common sampling method and three scenarios based on the stage sampling. The results indicate that the use of all scenarios, including stage sampling, resulting in fewer errors, better estimation and stronger Spatial structure in determination the saturated hydraulic conductivity values. Among the stage methods, the spatially balanced design approach was less accurate relative to the standard error of prediction index. The RMSE, MAE and ASE values in the scenario designed by kriging standard error of prediction than conventional sampling methods reduced respectively from 5.56, 4.7 and 17.3 to 3.79, 2.84 and 3.86 m.d-1 and so improving 39.6, 31.8 and 77.7 percent these three indicators error.