Termite bioturbation constitutes a dominant yet under quantified driver of soil matrix restructuring in tropical agroecosystems, profoundly influencing water retention, nutrient cycling, and microstructural architecture processes intimately linked to soil dielectric permittivity. This investigation deploys high-resolution frequency-domain dielectric spectroscopy (10 MHz–14 GHz, 501 discrete frequencies; 0–60 °C) to systematically characterize termite-induced perturbations in real (ε′) and imaginary (ε″) permittivity components across composite soil samples from various termitarium-inhabited sites in northern India’s fertile alluvial plains. Termite-affected soils exhibited statistically significant elevations in ε′ versus undisturbed controls and dramatic ε″ amplification, correlating strongly with bioturbation-mediated enhancements: + 18% volumetric moisture retention, 2.1 × organic matter enrichment, pH acidification and bulk density reduction across five textural classes (sand–clay loam). Inverse-quartic Response Surface Methodology resolved ε′ nonlinear dispersion physics, achieving R2 = 0.9818, Adj.R2 = 0.9808, Pred.R2 = 0.9804, adequate precision = 124.06, and CV = 3.40% superior performance reflecting Debye relaxation linearization in reciprocal space. ANOVA paradox traces to dominant frequency polynomials. SVR outperformed: ε′ (R2 = 0.99, RMSE = 0.60, MAE = 0.38); ε″ (R2 = 0.82, RMSE = 285.7), validated via fivefold stratified CV and 20% holdout testing confirming zero overfitting. ε″ heteroscedasticity evidences multi-Lorentzian superposition resistant to polynomial bases. Custom brass sample holders preserved bulk density during open-ended coaxial probe (85070E) + VNA (E5071C) measurements (ε′ error = 1.5%, ε″ = 2.3%). USDA-compliant grid sampling, sieve analyses (< 600 μm: 60%), and real-time field validation (r = 0.92, n = 3 sites) across termite activity stages (initial → post-infestation) established > 97% detection specificity. Applications encompass precision agriculture (termite-aware irrigation), microwave drying (ε′-optimized energy dosing), remote sensing (bioturbation mapping), and soil engineering (non-destructive termite damage assessment). Laboratory findings necessitate field-scale multi-species validation, long-term environmental cycling durability, Cole–Cole hybridized neural architectures for ε″ mastery, and techno-economic analyses scaling dielectric proxies to global soil management. This work pioneers dielectric spectroscopy as a bio-intelligent paradigm for subterranean pest detection, bridging soil physics, ecological engineering, and precision agriculture toward resilient agroecosystems. Graphical Abstract

Investigating the Impact of Termite Bioturbation on Agricultural Soil Dielectric Properties: Implications for Detection, Monitoring, and Precision Agriculture

期刊论文
发表日期：

2026年2月5日

查看全文

Prachi Palta,
Aastha Palta,
Ankur Kumar

摘要

Termite bioturbation constitutes a dominant yet under quantified driver of soil matrix restructuring in tropical agroecosystems, profoundly influencing water retention, nutrient cycling, and microstructural architecture processes intimately linked to soil dielectric permittivity. This investigation deploys high-resolution frequency-domain dielectric spectroscopy (10 MHz–14 GHz, 501 discrete frequencies; 0–60 °C) to systematically characterize termite-induced perturbations in real (ε′) and imaginary (ε″) permittivity components across composite soil samples from various termitarium-inhabited sites in northern India’s fertile alluvial plains. Termite-affected soils exhibited statistically significant elevations in ε′ versus undisturbed controls and dramatic ε″ amplification, correlating strongly with bioturbation-mediated enhancements: + 18% volumetric moisture retention, 2.1 × organic matter enrichment, pH acidification and bulk density reduction across five textural classes (sand–clay loam). Inverse-quartic Response Surface Methodology resolved ε′ nonlinear dispersion physics, achieving R² = 0.9818, Adj.R² = 0.9808, Pred.R² = 0.9804, adequate precision = 124.06, and CV = 3.40% superior performance reflecting Debye relaxation linearization in reciprocal space. ANOVA paradox traces to dominant frequency polynomials. SVR outperformed: ε′ (R² = 0.99, RMSE = 0.60, MAE = 0.38); ε″ (R² = 0.82, RMSE = 285.7), validated via fivefold stratified CV and 20% holdout testing confirming zero overfitting. ε″ heteroscedasticity evidences multi-Lorentzian superposition resistant to polynomial bases. Custom brass sample holders preserved bulk density during open-ended coaxial probe (85070E) + VNA (E5071C) measurements (ε′ error = 1.5%, ε″ = 2.3%). USDA-compliant grid sampling, sieve analyses (< 600 μm: 60%), and real-time field validation (r = 0.92, n = 3 sites) across termite activity stages (initial → post-infestation) established > 97% detection specificity. Applications encompass precision agriculture (termite-aware irrigation), microwave drying (ε′-optimized energy dosing), remote sensing (bioturbation mapping), and soil engineering (non-destructive termite damage assessment). Laboratory findings necessitate field-scale multi-species validation, long-term environmental cycling durability, Cole–Cole hybridized neural architectures for ε″ mastery, and techno-economic analyses scaling dielectric proxies to global soil management. This work pioneers dielectric spectroscopy as a bio-intelligent paradigm for subterranean pest detection, bridging soil physics, ecological engineering, and precision agriculture toward resilient agroecosystems.

Graphical Abstract