Assessing the effects of integrated crop-livestock systems and associated management practices on soil carbon and nitrogen cycling processes using the Daycent model

Abstract

The adoption of crop-livestock integration systems has been used in Brazil as a strategy to achieve a more sustainable and productive agriculture, with a perspective of its implementation in areas of monoculture and extensive grazing, many of which under degradation. The integrated systems present organic compounds inputs with quantities and composition different from the traditional systems, as well as water and heat fluxes, what leads to marked changes in the magnitude of carbon (C) and nitrogen (N) cycling processes. In this regard, there is a lack of information about the magnitude of C and N cycling in integrated systems on an agroecosystem scale, especially regarding the quantitative importance of greenhouse gas emissions and changes in the organic matter stocks over the time. Therefore, the present proposal has the objective of evaluating the effects of crop-livestock integration systems and management practices on the C and N cycling processes in the soil using the DAYCENT model, which has an extensive history of use in this capacity. Given that the model has never been used before for integrated agricultural systems in Brazil, the project will consist of parameterizing and evaluating the model for this purpose. To that end, the team of the Núcleo Interdisciplinar de Planejamento Energético of UNICAMP in partnership with the University of New Hampshire will compile a set of literature data and perform field measurements to adjust the coefficients of the DAYCENT model equations in conditions of crop-livestock integration system of the Brazil. Once parameterized, data on greenhouse gas emissions (CO2, N­2O, and NO), C and N total of the soil and N mineralization will be used in the validation, with 50 % of the data used for eventual tuning and the other 50 % reserved for independent evaluation. Regression equations between observed and predicted values will be adjusted, determining the performance of the model by the significance of the coefficients and by the determination coefficient, among other metrics. In addition, an evaluation of the magnitude of the error of the predictions will be carried out to identify possible trends (sub and overestimate conditions). A version control software GitHub (www.github.com) will be used to track all files associated with the analyzes performed, as well as support collaboration among project members and to support a more transparent and repeatable analysis. The parameterization and validation of DAYCENT for crop-livestock integration systems represent an important advance for science and Brazilian agriculture.