EGU24-19741, updated on 11 Mar 2024
https://doi.org/10.5194/egusphere-egu24-19741
EGU General Assembly 2024
© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.

Digital soil analysis and mapping using in-situ Vis-NIR spectroscopy – Challenges and future perspectives

Akos-Etele Csibi, Hans Sanden, Pavel Baykalov, Ruth Pereira, Anabela Cachada, Boris Rewald, and David Perry
Akos-Etele Csibi et al.
  • BOKU, Forest Ecology and Soil Sciences, Forestry, Vienna, Austria (etele4sure@gmail.com)

The use of Vis-NIR spectroscopy in digital soil mapping is emerging as a fast, viable option to provide 
spatial and temporal information on specific soil parameters that serve as good indicators for soil 
health. While MIR spectroscopy tends to be a much more reliable (high-precision) tool for different 
soil properties estimations, currently only NIR can be adapted for rapid in-situ soil surveys.
The Subterra Green device, developed by “S4 Mobile Laboratories”, equipped with a Visible and an 
FTIR spectrometer can optimally capture spectra until 90 cm underground down to a 1 cm resolution. 
With a carefully selected sampling pattern, a survey of several hectares can be conducted in a matter 
of few days as a single insertion takes about 2-6 minutes.
One scope of the PHENET project is to carry out soil surveys in different locations with varying soil 
types, from the humid continental zones of Austria to the temperate oceanic climate of Portugal. 
This will be done by creating models which are verified with laboratory biochemical analysis of soil 
samples. Previous scientific resource concluded that some soil properties like the soil water content 
or texture can have a major effect on the recorded spectra, so when building up a database for 
machine learning models from different site surveys (with unique spatial and temporal conditions) a 
lot of external factors should be taken into consideration and pre-processing techniques selected, 
like external parameter orthogonalization or calibration spiking for creating an accurately predicting 
model for soil parameters prediction. The aim is to provide estimations of soil organic carbon and 
nitrogen stocks as well as interpolated maps in different soil depths. Being able to do fast and highresolution soil maps using in-situ Vis-NIR soil spectroscopy makes it possible to improve precision 
agriculture and monitor soil properties over space and time.

How to cite: Csibi, A.-E., Sanden, H., Baykalov, P., Pereira, R., Cachada, A., Rewald, B., and Perry, D.: Digital soil analysis and mapping using in-situ Vis-NIR spectroscopy – Challenges and future perspectives, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19741, https://doi.org/10.5194/egusphere-egu24-19741, 2024.