Spatial variability and geo-statistics application for mapping of soil properties and nutrients in semi arid district Kohat of Khyber Pakhtunkhwa (Pak

摘要

Spatial variability and its importance was kept in view and this project was designed to model spatial variability of soil properties and their mapping in semi arid district Kohat of Khyber Pakhtunkhwa (KP) province of Pakistan. Soil sampling was done on a grid system using Global Positioning System (GPS) from two depths i.e. 0-15 and 15-45 cm during 2004 and were analyzed for soil physical properties (soil texture and bulk density), soil chemical properties (pH, ECe, SAR, lime and organic matter) and soil fertility status (Mineral N, AB-DTPA extractable P, K, Zn, Cu, Fe and Mn, and HCl extractable boron). Geostatistical techniques of semivariogram analysis and kriging were used to model the spatial variability and interpolation of data values at un sampled locations and mapping in the district. Semivariogram analysis showed that the soil separates viz. sand, silt and clay content in Kohat district showed spatial patterns in both surface as well as subsoil. In the surface soil, the data were described by linear models for all the three soil separates. However, in the subsoil, silt content was described by a spherical model with a range of 30.38 km. Semivariogram analysis of the data on soil pH was described by a spherical model in both the depths with a range of 12.55 km in the surface soil and 8.26 km in the subsoil. Lime content in the surface soil was described by a linear model while in the subsoil, it was described by a spherical model with a range of 5.50 km. Organic matter content in the surface soil was described by a linear model. Potash content of the surface and subsoil was described by linear models showing strong spatial patterns in surface and very poor structure in subsoil. Manganese content was described by a spherical model in the subsoil with a range of 20.19 km. Iron content was described by linear models with a poor structure in surface and strong spatial structure in subsoil. Boron content in both the depths was described by spherical models with a range of 15.70 km in surface soil and 4.32 km in the subsoil. The data on various measured soil properties and the semivariogram models developed were used to estimate the soil test values at unsampled locations using geostatistical technique of kriging. Maps were developed using Surfer 6.04 programme and the areas were delineated into low, medium and high levels of plant nutrients.S.