Open AccessAssessment of Geostatistical Models for the Major Soil Nutrients for Tumkur District of Karnataka, India
Author(s) -
Leena H.U.*,
Premasudha B.G.,
P.K. Basavarja,
H. Mohamed Saqeebulla,
G. V. Gangamrutha
Publication year - 2019
Publication title -
international journal of recent technology and engineering
Language(s) - English
Resource type - Journals
ISSN - 2277-3878
DOI - 10.35940/ijrte.d9606.118419
Subject(s) - kriging , soil fertility , environmental science , mathematics , multivariate interpolation , soil test , nutrient , soil map , soil science , fertilizer , agricultural engineering , agriculture , interpolation (computer graphics) , statistics , soil water , agronomy , computer science , geography , engineering , ecology , biology , animation , computer graphics (images) , bilinear interpolation , archaeology
Digitization of agriculture has tremendously increased in the adoption of various advanced techniques in the Indian agricultural sector. One of the core agriculture objectives is preserving soil fertility. To achieve this efficient soil fertility management alongside an effective spatial distribution of soil nutrient properties is required. The main objective of this study is to evaluate and propose the best interpolation technique on estimating the soil nutrients status to provide site-specific fertilizer recommendations through the Soil Test Crop Response target yield approach. In this study, we have focused on three major soil nutrients viz., nitrogen (N), phosphorus (P2O5) and potassium (K2O) for evaluation. The benchmarking study has considered four most successive interpolation techniques like Ordinary Kriging (OK), Radial Basis Function (RBF), Inverse Distance Weighted (IDW), and Global Polynomial Function (GPI). The evaluation and analytical results proved Ordinary Kriging is better by securing the highest accuracy against other interpolation techniques concerning RMSE and ME for interpreting the soil nutrients N, P2O5, and K2O. The interpreted values are also cross-validated with actual soil test samples with an accuracy of more than 85% for each nutrient. Nevertheless, these results are dependent on the number of actual soil test samples and the accuracy of the designed network with overall accuracy between the interpreted and the actual data.