Airborne radiometric data for digital soil mapping of peat at broad and local scales

Abstract

Peat soils are high in soil organic matter (SOM) and are recognised stores of carbon. Knowledge of the spatial distribution of peat soils is becoming the focus of many studies and is related closely to peatland mapping. Accurate maps of peat soils have many applications of international importance e.g., gaseous emission inventory reporting or soil organic carbon stock accounting. Traditional mapping methods include in-situ soil auger sampling or peat probing (for depth) while modern methods also incorporate satellite data (optical and radar). However, both methods have limitations. Traditional sampling often omits boundaries and transition zones between peat and mineral soils, while satellite data only measure the surface and may not be able to penetrate landcover, potentially omitting areas of peat under, for example, grassland or forestry. Radiometrics is a measurement of naturally occurring gamma radiation. Peat soils attenuate this radiation through high soil moisture content. For the present study in Ireland, the supervised classification of gridded airborne radiometric data, acquired over multiple years, is performed using neural network pattern recognition to identify areas of peat and non-peat soils. Classification confidence values are used to identify the transition zone between these soil types, providing a simplified visualisation of this transition. Validation is performed using Loss on Ignition (LOI %) point data and several different (blanket bog, raised bog, transition zone) sites in Ireland, showing classified data can detect the presence of peat soils from broad to local scales. Airborne geophysical methods, in particular airborne radiometrics, can bridge the gap between the accuracy of point measurement and the spatial coverage of satellite data to identify peat soils by providing uniform data and objective analysis. The resulting map is a step towards understanding the true spatial distribution of peat soils in Ireland, including transition zones.

Keywords

• Peat
• Airborne geophysics
• Carbon
• Neural networks
• Gamma rays
• Radiometrics