Monitoring soil organic carbon variations in organic farming by proximal sensing

It has been widely It has been widely recognized the crucial role of soil monitoring in assessing the effectiveness of the soil measures and agro-environment schemes passed by Regional and local administration in complying the EU’s Common Agricultural Policy to soil protection. The costs of soil monitoring however are substantial and the reliability of the results often questionable. A reason is that the monitoring activity commonly consists of a network of sampling points which are then spatialized through different inference models. This methodology does not often enable a reliable estimation at the fi eld scale, that is the scale where the measures are foreseen and fi nanced. The organic carbon content of the ploughed layer is an important soil feature, which regulates many soil functions. It is therefore considered in Organic Farming (OF) and contemplated in many agro-environment schemes. The adoption of OF is expected to improve soil organic carbon content (SOC) of the fi elds as a whole. Nevertheless, the improvement might show local variations, because of the interaction between crop management and other factors, for instance, soil erosion. Aim of this work was to check the variations in SOC and stock between and within fi elds where CA was or not applied, using combined soil proximal sensors, namely γ-ray and Vis-NIR spectroscopy. The research work was carried out in four areas in western Sicily (Italy). Two adjacent arable fi elds were surveyed for each area, comparing OF (green and organic manure, 20-30 cm rotary tilling, chiselling, or ploughing, crop rotation) lasting from 3-10 years, with continuous traditional farming (TF, no rotation, mineral fertilization, 40-50 cm ploughing). The lithology was either clayey and calcareous-clayey fl ysch, or clays, silty-clays and marls, but the soils were similar and characterized by toposequences of Calcisols, Regosols, Cambisols, and Vertisols. The γ-radiometric survey was performed by “The Mole” sensor, made by “Medusa Systems”, while the Vis-NIR spectroscopy used Fieldspec 3Hi-res with the ASD Contact probe for Vis-NIR region (0.350-2.5 µm). The soils had similar texture, common vertic properties, presence of secondary carbonates, and low organic matter, but gravel content was very variable. The carbon stock (upper 30 cm) of the whole fi elds under OF resulted signifi cantly higher than in TF (on average of 36.0 versus 33.2 Mg. ha-1). In addition, TF fi elds showed very heterogeneous SOC, with very high and very low values, as a consequence of soil erosion, while in OF SOC was much more homogeneous, thus refl ecting both higher organic matter accumulation and lower soil loss from water erosion in similar morphological positions