Spatial Variability and Geostatistical Mapping of DTPA-Extractable Cadmium in Surface Soils of Varanasi District, Uttar Pradesh, India
Himanshu Pandey
Department of Soil Science & Agricultural Chemistry, Institute of Agricultural Sciences, Banaras Hindu University, Uttar Pradesh, India.
Satish Kumar Singh
Department of Soil Science & Agricultural Chemistry, Institute of Agricultural Sciences, Banaras Hindu University, Uttar Pradesh, India.
Pramod Kumar Sharma
Department of Soil Science & Agricultural Chemistry, Institute of Agricultural Sciences, Banaras Hindu University, Uttar Pradesh, India.
Shubham Jaiswal *
Department of Soil Science & Agricultural Chemistry, Institute of Agricultural Sciences, Banaras Hindu University, Uttar Pradesh, India.
Shashank Kumar Singh
Department of Soil Science & Agricultural Chemistry, Institute of Agricultural Sciences, Banaras Hindu University, Uttar Pradesh, India.
Kajal Singh
Department of Soil Science & Agricultural Chemistry, Institute of Agricultural Sciences, Banaras Hindu University, Uttar Pradesh, India.
Ramesh Kumar Patel
Department of Soil Science & Agricultural Chemistry, Institute of Agricultural Sciences, Banaras Hindu University, Uttar Pradesh, India.
*Author to whom correspondence should be addressed.
Abstract
Background: Cadmium (Cd) availability in agricultural soils can show strong local variation, making district-scale averages insufficient for identifying areas of concern.
Aims: To describe the spatial variability of DTPA-extractable cadmium (DTPA-Cd) in surface soils of Varanasi district, map it by ordinary kriging and assess the uncertainty of the map.
Study Design: Geo-referenced field survey with geostatistical analysis.
Place and Duration of Study: Agricultural soils of Varanasi district, Uttar Pradesh, India, sampled between April and May 2026; laboratory analysis was conducted at the Department of Soil Science & Agricultural Chemistry, Institute of Agricultural Sciences, Banaras Hindu University, India.
Methodology: Ninety-eight surface soil samples (0 to 15 cm) were collected with GPS coordinates and extracted with DTPA. Data were ln-transformed, and spherical, exponential and Gaussian models were fitted to the experimental semivariogram. Ordinary kriging (OK) and inverse distance weighting (IDW) were compared by leave-one-out cross-validation, and the OK prediction and kriging variance were mapped on a 250 m grid.
Results: DTPA-Cd ranged from 0.020 to 6.154 mg kg−1 (mean 0.139, median 0.065 mg kg−1), with a coefficient of variation of 441.9%; 15.3% of samples exceeded 0.1 mg kg−1. The Gaussian model gave the best fit, with a nugget/sill ratio of 42.3% (moderate spatial dependence) and a range parameter of 18.0 km. OK performed slightly better than IDW (RMSE values of 0.616 and 0.619, respectively, on the ln scale), but both explained only 11 to 12% of the variation. Predicted DTPA-Cd ranged from 0.043 to 0.346 mg kg−1, with the highest quartile in the central and eastern parts of the district. Kriging variance was lowest in densely sampled areas and highest in the sparsely sampled centre and along the boundary.
Conclusion: Available Cd in the district is generally low but includes isolated high values. The map shows broad regional tendencies, and the extent of the high class depends strongly on a single extreme sample, which needs field verification and denser sampling.
Keywords: Cadmium, DTPA, semivariogram, ordinary kriging, Indo-Gangetic Plain, soil contamination