In the quest for sustainable agriculture, soil analysis stands as a cornerstone, guiding farmers and agronomists in making informed decisions. Traditional laboratory methods, while reliable, often come with significant drawbacks: they are time-consuming, expensive, and can be environmentally hazardous. Enter spectroscopy, a technique that has been gaining traction for its efficiency and eco-friendliness. Among the various spectroscopy methods, Near-Infrared (NIR) spectroscopy has emerged as a promising tool, requiring minimal sample preparation and no chemicals, while predicting multiple soil properties with a single scan.
However, the choice of equipment has been a contentious issue. Previous studies have reported inconsistent performance between conventional NIR spectroscopy and advanced Fourier-Transform Near-Infrared (FT-NIR) spectroscopy. To shed light on this, a recent study published in ‘Applied Sciences’ aimed to compare the predictive performance of conventional NIR and advanced FT-NIR spectroscopy for sixteen soil properties.
The study, led by Govind Dnyandev Vyavahare from the National Institute of Agricultural Sciences, Rural Development Administration in South Korea, collected soil samples from different land-use types across the country. The samples were analyzed using both spectroscopy techniques and traditional laboratory methods. Five models, including partial least squares regression (PLSR), Cubist, support vector machine (SVM), random forest (RF), and memory-based learning (MBL), were evaluated using 15-fold cross-validation to assess prediction accuracy.
The results were revealing. Conventional NIR spectroscopy yielded consistently higher accuracy for all soil properties than FT-NIR. “Despite their lower spectral resolution, NIR spectra provide robust predictive capability across a wide range of soil properties,” Vyavahare noted. Strong predictive accuracy was achieved for electrical conductivity (EC), organic matter (OM), available phosphorus (avl. P), total nitrogen (TN), and cation exchange capacity (CEC). In contrast, FT-NIR provided good prediction accuracy only for exchangeable potassium (Ex. K) and TN.
The average performance of NIR (R² = 0.67) outperformed FT-NIR spectroscopy (R² = 0.63) across all soil properties. These findings have significant implications for the agriculture sector. The study suggests that conventional NIR spectroscopy, which is generally more affordable and accessible, can provide reliable predictions for a wide range of soil properties. This could substantially reduce the investment cost for routine soil analysis, making advanced soil monitoring more accessible to farmers and agronomists worldwide.
The research also opens up new avenues for future developments. As Vyavahare explains, “The consistent performance of NIR spectroscopy across various soil properties indicates its potential for integration into precision agriculture practices.” This could lead to the development of portable, user-friendly NIR devices that farmers can use in the field, enabling real-time soil analysis and decision-making.
Moreover, the study highlights the importance of choosing the right spectroscopy technique for specific soil properties. As the field of agritech continues to evolve, such insights will be crucial in guiding the development of more efficient, accurate, and cost-effective soil analysis tools. The findings of this study not only advance our understanding of spectroscopy techniques but also pave the way for more sustainable and data-driven agriculture practices.