Soybean cultivation has been steadily expanding in the Cerrados since its introduction in the 1970’s and nowadays 50% of the national production is produced in the region. Brazil is the world’s first exporter and the second largest producer of soybean with a current production of over 50 million Tn. The increasing global market of soybeans, soy meal and soy oil (Smaling et al., 2008) and the great agricultural potential of the Cerrados (Neto, 2004), further expansion is expected. This has led to a growing concern of the sustainability of the larger expansion of soybean and the management of the already established soybean cultivations. In addition to losses of biodiversity, soil quality has been proposed as critical to agricultural and environmental sustainability. Soil Organic Matter (SOM) is the main indicator to evaluate soil quality in agroecosystems. SOM plays a significant role in physical, chemical and biological soil properties. It is a nutrient source for plant growth and It is the major terrestrial C pool. Soil Organic Carbon (SOC) stocks are expected to decrease when native lands are converted to arable lands. However, contradictory results have been reported in the Brazilian Cerrados. Losses or gains of SOC depend on several factors such as climate, soil type (%clay), quantity and quality of litter or crop residues, changes in the microbial community, changes in soil nitrogen cycling and cropland management practices.
To promote the maintenance or storage of SOC depends on understanding the factors that control SOC dynamics. This study attempted to analyse SOC dynamics with a systems approach by analysing experimental data and using a simple summary model. Models, which are defined as a simplified representation of a complex reality, may give a better understanding of the main processes that are involved in SOC dynamics within the system. Long-term experiments with different land uses and soil management practices were selected in Embrapa Cerrados (Planaltina, DF) and Embrapa Beef Cattle (Campo Grande, MS). In both sites, treatments with different tillage practices and crop sequences in soybean systems were set up in 1991/92 and 1993/94, respectively. Simulations were compared to field data collected from PhD thesis that reported SOC stocks after 11 years of the experiment establishment. Model simulations were considered satisfactory. They showed a depletion of SOC stocks in all treatments after eleven years of soybean cultivation. However, greater reduction was found in conventional tillage (CT) systems in both experimental sites. Soil C sequestration rates of No tillage (NT) systems, with CT systems as a reference, were 0.43 and 0.35 Mg C ha-1 yr-