Extending predictions of soil parameters for a national dataset

  • A. H.J. Robertson
  • E. Perez-Fernandez
  • N. J. Baggaley
  • B. M. McKenzie
  • I. J. Owen
  • A. Lilly
BÜCHI Labortechnik AG, Flawil, Switzerland
[email protected]
 Search for papers by this author
The James Hutton Institute, Craigiebuckler, Aberdeen, AB15 8QH, UK
[email protected]
 Search for papers by this author
The James Hutton Institute, Craigiebuckler, Aberdeen, AB15 8QH, UK
[email protected]
 Search for papers by this author
The James Hutton Institute, Craigiebuckler, Aberdeen, AB15 8QH, UK
[email protected]
 Search for papers by this author
The James Hutton Institute, Craigiebuckler, Aberdeen, AB15 8QH, UK
[email protected]
 Search for papers by this author
 Corresponding Author
The James Hutton Institute, Craigiebuckler, Aberdeen, AB15 8QH, UK
[email protected]
 Search for papers by this author

The National Soil Inventory of Scotland (NSIS) is an objective dataset and represents the most common soil types across Scotland. The soils in this unique dataset have been extensively characterised, with a wide range of chemical and physical parameters measured, including the near infrared reflectance (NIR) spectra. Significantly, the parameters measured have often been carried using more than one analytical method. In this work, we are looking to build on previously developed NIR calibrations for prediction of soil parameters and extend the useful information that can be gained from the NIR soil spectral data. We have examined in more detail the differences in the NIR correlations between elemental concentrations for some of the soil nutrients (Al, Fe, K, Mg, Mn, Na and P) measured in different ways. In addition, we are increasing the range of parameters used for calibration development to ones more directly linked to soil function. Here results are reported for correlations to soil aggregate stability data, which we are looking to apply to improving prediction of erosion risk.


Metrics

Downloads:

212

Abstract Views:

355