Estimation of Iron Content Using Reflectance Spectroscopy in a Complex Soil System After a Loss-on-Ignition Pre-treatment

Nicolas Francos*, Nuphar Gedulter, Eyal Ben-Dor

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

Reflectance spectroscopy across the 400–2500 nm region is an important source of data for the approximation of soil attributes. Modeling is based on direct/indirect relationships between the spectral response and soil characteristics. Total iron content is an important property, controlling, among others, soil color, and aggregation stability. However, soil properties can have overlapping spectral features. Iron and organic matter (OM) are both active in the visible–near infrared region. Thus, OM content interferes with the assessment of iron content. Here, we suggest heating the sample to 405°C to remove OM content through the loss-on-ignition method before spectral assessment of the iron content. The suggested method improved iron prediction from [RPD = 1.57; R 2 = 0.61; RMSE = 0.81] to [RPD = 2.31; R 2 = 0.83; RMSE = 0.55].

Original languageEnglish
Pages (from-to)6866-6873
Number of pages8
JournalJournal of Soil Science and Plant Nutrition
Volume23
Issue number4
DOIs
StatePublished - Dec 2023

Funding

FundersFunder number
European Space Agency

    Keywords

    • Data analysis
    • Iron
    • Machine learning
    • Mediterranean soils
    • Organic matter
    • Soil spectroscopy

    Fingerprint

    Dive into the research topics of 'Estimation of Iron Content Using Reflectance Spectroscopy in a Complex Soil System After a Loss-on-Ignition Pre-treatment'. Together they form a unique fingerprint.

    Cite this