This study concerns the applied use of the natural radioactivity in soils. The relevance of airborne radiometric (gamma ray) survey data to peat mapping is now well established and such data have been used in a stand-alone sense and as covariates in machine learning algorithms. Here we present a method to use these data to accurately map the boundaries of peat (raised bogs). This has the potential to assist with the estimation of carbon stocks using a property-based assessment of soil. The significance of such regionally-uniform survey data lies in the subsurface information carried by the measurement which contrasts with the surficial nature of many other covariates. Soils attenuate radiometric flux by virtue of their bulk density (and associated carbon content) and water saturation level. The high attenuation levels in low density, wet peat materials give rise to a distinctive soil response. Here an entirely physics-based assessment of flux attenuation is carried out both theoretically and empirically. Radiometric data from the ongoing Tellus airborne survey of Ireland are used. The study area is characterised by an extensive assemblage of discrete raised peat bogs in a framework of largely mineral soils. Peat is detected by a property contrast with adjacent soils and so we consider all soils within the study area. The relatively low lateral resolution of the airborne data is demonstrated by modelling and we examine the behaviour of a combined spatial derivative of the data. The procedure allows the identification of the edges of the 128 peat polygons considered and indicates other additional potential areas of subsurface peat. The data appear to resolve the differences that exist across three available soil/peat databases that are used for the validation of the results obtained.
Read full abstract