Amrit Chandan Patra¹, Annie Daliya N.² and Bibhas Sen³

1Geological Survey of India, Southern Region, Hyderabad-500068, Telangana, India 2Geological Survey of India, Training Institute, Hyderabad-500068, Telangana, India 3Geological Survey of India, State Unit: Maharashtra, Pune-411006, Maharashtra, India (*Corresponding Author, E-mail: bibhas.sen@gsi.gov.in)

Abstract

The present study provides an effective geostatistical methodology for extracting high-confidence information from geochemical stream sediment data collected on a grid-based sampling pattern over large areas. Unlike traditional deterministic interpolation methods, our approach employs exploratory data analysis of metal oxides to determine critical thresholds by examining the variability and anisotropy factors through variogram analysis of spatial data. To minimize the ‘nugget effect,’ which often causes predicted values to fall below critical thresholds, kriging with probability output surfaces was applied to generate geochemical probability maps indicating the likelihood of each oxide exceeding its statistically derived threshold. The methodology was tested using stream-sediment geochemical dispersion data from the Sittampundi Layered Anorthosite Complex (SAC) and the Ramagiri–Penakacherla Schist Belt (RPS). In the SAC, probable zones exceeding critical thresholds were delineated for Fe2O3, CaO, MgO, Ni, Co, and Cr based on 200 samples. Similarly, in the RPS, the method effectively identified such zones for Fe2O3, TiO2, CaO, MgO, and MnO using 770 samples. These case studies demonstrate the utility of geostatistically derived probability maps in delineating geochemical dispersion patterns from systematically collected stream-sediment data over extensive areas.

Keywords: Geostatistics, Kriging, Probability Mapping, Geochemical Dispersion