1Department of Soil Science and Agricultural Chemistry, College of Agriculture, Jodhpur, Agriculture University, Jodhpur, 342304, Rajasthan, India
2Soil Technology and Carbon Sequestration Laboratory, Department of Soil Science and Agricultural Chemistry, Institute of Agricultural Sciences, Banaras Hindu University, Varanasi, 221005, Uttar Pradesh, India
3ICAR-IARI, Gauria Karma, Hazaribag, 825405, Jharkhand, India
4Department of Farm Engineering, Institute of Agricultural Sciences, Banaras Hindu University, Varanasi, 221005, Uttar Pradesh, India
5Department of Soil Science and Agricultural Chemistry, Siksha O Anusandhan University, Bhubaneswar, 751003, Odisha, India
6Department of Plant Physiology, College of Agriculture, Nagaur, Agriculture University, Jodhpur, 342304, Rajasthan, India
*Corresponding author (Email: seemachahar94@gmail.com)
The point measurement of common soil properties and their use in fertilizer predictions over an area in India has often backfired due to large existent spatial variability. Mapping of soil properties with acknowledge of its spatial variability has widespread applications in environmental modelling, precision farming and sustainable agriculture. Hence a study was carried out in the middle Indo Gangetic plain zone of India to map the spatial variability of some important soil properties that are used in fertilizer predictions. A total of 280 surface soil samples from the plough layer (0–15 cm depth) were collected and analyzed for pH, electrical conductivity, organic carbon, available P, available K, sand, silt and clay. The values of these properties ranged from 5.12 to 9.20, 0.01 to 1.71 dS m-1, 1.2 to 8.70 g kg-1, 8.27 to 24 kg ha-1, 114.9 to 277.8 kg ha-1, 19.0 to 58.0%, 13.7 to 56.0 % and 12.5 to 52.5%, respectively following non-normal distribution for electrical conductivity and normal distribution for rest of the parameters. The coefficient of variation was highest for organic carbon (43.7%) and lowest for soil pH (0.15%), suggesting that use of arithmetic mean and standard deviation would serve as a poor estimator of these parameters for unknown locations. Geostatistics revealed that Gaussian model best fitted the experimental semi-variogram for pH and sand content; exponential model for organic carbon, available P, available K, silt and clay content and spherical model for electrical conductivity. The nugget/sill ratio indicated that all the studied soil properties were strongly spatially dependent except pH and sand content, which were moderately spatially dependent. Cross validation statistics showed that the surface soil maps prepared by ordinary kriging were robust and would sever much better when predicating values of the soil parameters in unsampled locations, often need in modelling at the landscape scale, and could serve as a primary guide for identifying management zones for precision farming and region-specific nutrient management and designing future soil sampling strategies in the intensively cultivated soils of the middle Indo Gangetic plains of eastern India.
Kriging, Semivariogram, Nugget, Sill, Spatial dependence, Indo-Gangetic plains, Inceptisols
