Data fusion of Vis-NIR and pXRF with machine learning for predicting andic properties in volcanic soils
Volcanic soils exhibit distinctive andic properties, including organic carbon (OC), bulk density (Bd), phosphate retention (Pret), and the sum of ammonium oxalate-extractable aluminum and half iron (Alo + 0.5Feo), and play an important role in agricultural production and ecological functions. However, conventional determination of andic properties relies on destructive, labor-intensive, and time-consuming wet chemistry analyses. Soil spectroscopic techniques such as visible and near-infrared (Vis-NIR) spectroscopy and portable X-ray fluorescence (pXRF) provide rapid and non-destructive alternatives. This study aimed to elucidate the relationships between andic properties and signals from Vis-NIR and pXRF, and to evaluate the accuracy of sensor-based models for predicting andic properties and soil classification. A total of 93 soil samples were collected from 24 pedons of volcanic soils (0–60 cm depth) in northern Taiwan, including Andisols and Inceptisols. Predictive models were developed using individual sensors and a data fusion approach calibrated with partial least squares regression (PLSR) and the Cubist algorithm. Both Vis-NIR and pXRF signals demonstrated associations with andic properties, while data fusion of these two sensors markedly improved model performance. In particular, the Vis-NIR + pXRF-based model calibrated with Cubist yielded good predictive performance, achieving R2 and LCCC values exceeding 0.90 for OC, Pret, and Alo + 0.5Feo, and R2 = 0.83 and LCCC = 0.89 for Bd. Moreover, 23 out of the 24 studied pedons were correctly classified. Integrating Vis-NIR and pXRF provided an efficient approach for assessing andic properties, improving management and resource-use efficiency in volcanic soils, and supporting sustainable smart agriculture.
Authors: Po-Hui Wu, Budiman Minasny, Yin-Chung Huang, José A.M. Demattê, Zeng-Yei Hseu
Keywords: Andisols; Modeling; Soil classification; Soil health; Soil spectroscopy
Spectral Evolution is a leading manufacturer of field-portable UV-Vis-NIR remote sensing spectroradiometers and spectrometers for applications including geological remote sensing, ground-truthing, spectral remote sensing, environmental and climate research, crop and soil research, vegetative studies, forestry and canopy studies, radiometric calibration transfer, upwelling and downwelling measurement and more. Our UV-Vis-NIR spectrometers and spectroradiometers provide the best combination of high resolution and high sensitivity resulting in the most precise field portable instruments available on the market.
Spectral Evolution UV-Vis-NIR spectroradiometers provide:
- Fast, full-spectrum UV-Vis-NIR measurements – 350-2500 nm with just one scan
- Superior reliability in the field with 100% diode array optics with no moving parts
- Wireless Bluetooth® Connection
- Lithium-ion rechargeable batteries for field use
- Lightweight and rugged
- Rugged handheld tablet for field use running DARWin™ LT on Windows 10 with digital camera, and GPS
- Field switchable optics for varying target sizes and measurement modes
All Spectral Evolution UV-Vis-NIR spectroradiometers include our exclusive DARWin™ SP Data Acquisition and Analysis software for one-touch, full range spectroscopic analysis of samples. DARWin™ SP automatically saves data as ASCII files for use with 3rd party software without post-processing.