Quantifying spatial uncertainty of soil organic matter content using conditional sequential simulations: A case study in Emilia Romagna Plain (Northern Italy)

The development and the application of soil organic matter (SOM) indices have recently received much attention by soil scientists. Inventories of soil carbon for modeling, monitoring and mapping programs are currently being realized at different scales, ranging from global to national, to sub-regional, in order to provide operative tools for land use planners and decision makers. This paper presents the results of a research project on the estimation of SOM spatial distribution in the Emilia Romagna Plain, focusing on the effect of spatial uncertainty on the development and mapping of SOM content. In this work, the analysis has been carried out adopting and comparing two distinct geostatistical approaches to assess the uncertainty regarding the content of SOM jointly over several locations on a regular grid: a parametric approach based on the assumption of a multi-gaussian data distribution and a non-parametric approach based on the transformation of data in indicator variables through selection of operatively significant sill values. Two different SOM content class maps are presented with associated uncertainty. The two maps, although almost coincident in terms of SOM content class distribution in the study area, show a significantly different overall degree of spatial uncertainty, with the non-parametric approach based map showing the highest overall accuracy. Key words: Soil organic matter inventory, spatial variability, stochastic simulations, indicators uncertainty