PREDICTING SOIL ORGANIC CARBON IN NORTHWESTERN BRAZIL USING VISIBLE, NEAR-INFRARED, AND MID-INFRARED SPECTROSCOPY AND MACHINE LEARNING ALGORITHM

- 319763
Poster
Favorite this paper
How to cite this paper?
Abstract

Soil organic carbon (SOC) modeling is essential for assessing soil fertility, soil health, and carbon sequestration potential. Visible and near-infrared (Vis-NIR) and mid-infrared (MIR) spectroscopy have gained global popularity for SOC prediction as a cost-effective method in socioeconomic and environmental terms. This study explores the use of Vis-NIR and MIR spectroscopy to model organic matter (OM) content in soils from northeastern Brazil. The dataset included 629 samples from 7 of the 9 states in the region. We used the Cubist model to predict OM content, with R software aiding in statistical processing. Our results showed that the MIR technique performed better, with an R² of 0.78 for validation data and 0.92 for training data than Vis-NIR range. This suggests the potential of spectroscopy techniques to support or even replace conventional physical-chemical analysis methods, presenting a viable alternative, especially when comprehensive and optimized calibration databases are available.

Share your ideas or questions with the authors!

Did you know that the greatest stimulus in scientific and cultural development is curiosity? Leave your questions or suggestions to the author!

Sign in to interact

Have a question or suggestion? Share your feedback with the authors!

Institutions
  • 1 USP/ESALQ
  • 2 Universidade de São Paulo - Escola Superior de Agricultura “Luiz de Queiroz” (ESALQ/USP)
  • 3 ESALQ
Track
  • 28. Hyperspectral Remote Sensing
Keywords
Cross-validation
soil health
diffuse reflectance spectroscopy
organic matter