Soil dielectric response to chemical fertilizers in Northern India's key agricultural areas

Abstract

This study investigates the impact of various artificial nutrient components, specifically Nitrogen (N), Phosphorus (P), Potassium (K), and their combination (NPK), on the dielectric properties of soil (ε′ and ε''). The research focuses on examining the physical, chemical, and dielectric properties of soil influenced by different concentrations of these fertilizers, both individually and in combination. Dielectric analysis was performed within the radio and microwave spectrum (200 MHz-14 GHz) using a modified probe arrangement with an Agilent 85070E open-ended coaxial probe and a vector network analyzer.

The results revealed a complex interplay between N, P, K, and NPK concentrations and soil characteristics. The varying chemical compositions significantly altered the soil's physical and chemical properties, as detailed in the tabulated results. Soils treated with K exhibited the highest dielectric parameter (ε′ and ε'') values, followed by P and N and NPK combined. Advanced modeling techniques, including Response Surface Methodology (RSM) and Machine Learning (ML), were employed to predict soil dielectric properties (ε′ and ε'') as functions of nutrient concentrations, temperature, and frequency. The RSM models demonstrated high precision, with R² values of 0.9982, 0.9958, 0.9913, and 0.9962 for ε′ of N, P, K, and NPK, respectively. However, the accuracy of these models decreased for ε''. To address this limitation, various ML regression models were analyzed for ε′ and ε'', yielding high accuracy and enhanced prediction values, with MAE, MSE, RMSE, and R² scores of 0.378, 0.196, 0.615, and 0.9945 for ε′ and 0.045, 0.0034, 0.212, and 0.95 for ε'', respectively. This research highlights the significant effects of N, P, K, and NPK on soil dielectric behavior, providing valuable insights into nutrient-soil interactions. The findings have practical implications for agricultural practices, offering a non-destructive method to assess soil nutrient levels and optimize fertilization strategies for enhanced crop productivity.