In soil spectral analysis, the spectral signatures of heavy metal ions do not manifest directly. Instead, they indirectly alter soil reflectance spectrum features by influencing the content and form of primary spectrally active components, such as organic matter, iron/manganese oxides, clay minerals, and moisture. Common heavy metals—including cadmium (Cd), lead (Pb), mercury (Hg), chromium (Cr), copper (Cu), zinc (Zn), and arsenic (As)—do not produce distinct characteristic absorption peaks in the visible to near-infrared (350–2500 nm) spectrum. However, they exhibit significant correlations with soil physicochemical properties (cation exchange capacity, organic matter content, pH, and iron oxide content), allowing for indirect prediction through spectral inversion. By utilizing field spectroradiometers to capture soil sample spectral datasets and applying spectral preprocessing and feature extraction methods (comprising Multiplicative Scatter Correction [MSC], Standard Normal Variate [SNV], first/second derivative transformations, and continuum removal), the representation of heavy metal informatics within the spectrum can be enhanced. Building upon these features, predictive regression models for heavy metal concentrations are constructed using algorithms such as Partial Least Squares Regression (PLSR), Support Vector Regression (SVR), or Random Forest (RF). Empirical studies demonstrate that the accuracy of optical remote sensing inversion for soil heavy metals satisfies the technical requirements for rapid field screening, making it highly applicable for preliminary classification and risk zone identification in large-scale soil pollution surveys. The HG-ispectra2500 field spectroradiometer, engineered by Hagorun Technology Limited, extends its spectral coverage into the short-wave infrared region, delivering a stable data acquisition infrastructure for these advanced spectral analyses. The binding mechanisms between distinct heavy metals and spectrally active components vary fundamentally. For instance, the complexation of heavy metals with organic matter triggers shifts in organic spectral features (such as absorption bands near 1400 nm, 1900 nm, and 2200 nm), whereas the substitution of iron ions within iron oxide crystal lattices by heavy metals directly impacts elemental iron spectral absorption features in the 400–550 nm and 850–1000 nm intervals. Deciphering these microscopic spectral response mechanisms serves as the theoretical foundation for constructing interpretable inversion models and enhancing prediction accuracy.

Unlike heavy metals, specific organic contaminants—such as certain petroleum hydrocarbons, polycyclic aromatic hydrocarbons (PAHs), and pesticide residues—contain explicit C-H and C-C functional groups that generate direct spectral absorption signatures in the near-infrared band. In petroleum hydrocarbon pollutants, C-H bonds exhibit characteristic absorption peaks near 1720 nm, 1760 nm, 2310 nm, and 2340 nm, providing direct spectral markers for the rapid identification of petroleum-contaminated soils via hyperspectral techniques. Field spectroradiometers execute in-situ scanning of suspected contamination zones under field conditions, enabling researchers to map contamination boundaries and relative pollution levels by contrasting the curves against clean reference soil spectra. In applying field spectroscopy to organic pollution detection, spectral resolution and signal-to-noise ratio (SNR) operate as critical performance benchmarks. Optimized spectral coverage paired with narrow-band sampling intervals captures the fine spectral features of organic contaminants, resolving key target informatics. The HG-ispectra2500 field spectroradiometer by Hagorun Technology Limited, backed by its inherent spectral resolution and SNR performance, adapts seamlessly to the identification and capture demands of organic pollution spectral peaks. In laboratory settings, multivariate calibration models built from spectral data of soils with known organic contaminant concentrations enable semi-quantitative predictions of total petroleum hydrocarbons and total PAHs in unknown samples. The overall investment cost of this analytical method is significantly lower than traditional Gas Chromatography-Mass Spectrometry (GC-MS) workflows, offering superior scenario adaptability and operational advantages. It should be noted that the spectral detection of soil organic pollution is heavily influenced by soil type, moisture variations, and background organic matter. By combining field sampling with laboratory spectral measurements and introducing transfer algorithms like External Parameter Orthogonalization (EPO) or Piecewise Direct Standardization (PDS), the impacts of environmental variables on spectral models can be mitigated. In the future, online discrimination systems for soil organic pollution that integrate spectral databases with cloud computing frameworks are poised to enable real-time monitoring and proactive risk forecasting for contaminated sites.

Soil salinization represents a major type of land degradation impacting agricultural yields and ecological environments. Saline ions (Na+, Ca2+, Mg2+, Cl-, SO42-, etc.) and their hydration processes generate specific spectral absorption features in the near-infrared band, particularly around the 1400 nm and 1900 nm water absorption bands, where increasing salt concentrations alter water molecule absorption intensities and peak shapes. Applying continuum removal and derivative transformations to soil spectra enables the extraction of spectral indices highly correlated with salt content (such as the Salinity Index [SI] and Normalized Difference Salinity Index [NDSI]), facilitating rapid estimation of soil salinity. Beyond salt concentrations, field spectroradiometers enable the rapid prediction and spatial mapping of key soil parameters, including organic matter, total nitrogen, cation exchange capacity, and moisture content. Organic matter displays a distinct spectral response in the 600–800 nm region, where its concentration correlates negatively with visible-band reflectance. Constructing organic matter prediction models by fusing spectral data with partial least squares regression enables the rapid acquisition of organic matter metrics across large batches of soil samples without relying on traditional potassium dichromate oxidation assays, significantly minimizing laboratory workloads and operational costs. During field-scale surveys, portable field spectroradiometers pair with synchronized GPS configurations to capture co-registered spectral and positional datasets. Integrated with Geographic Information System (GIS) tools, this workflow generates spatial distribution maps of critical soil parameters to identify contamination hotspots, evaluate migration trends, and guide strategic grid sampling. This spectroscopy-driven rapid assessment method is exceptionally suited for wide-area soil environmental quality audits, preliminary screening of contaminated brownfields, and tracking post-remediation efficacy.

Interested in Advancing Field Spectroscopy Applications?

Our engineering group delivers advanced technical consulting and integrated solutions

Hagorun Technology Limited  |  Focus · Dedication · Exploration · Vision