Application of D-SEM to a catchment in Central Amazonia: calibration and validation of the carbon and nitrogen cycles

We applied the D-SEM model for predicting carbon (C) and nitrogen (N) dynamics in a natural forest ecosystem in a Central Amazonian microcatchment. A continuous collection of hydrological and water quality data (2002–2010) was used for model calibration and validation. Comparisons between predicted and measured hydrologic variables showed that the model captured seasonal and interannual discharges and DOC concentrations in stream water. The model also reproduced the monthly dynamics of DOC in the soil pool and captured seasonal changes in nitrogen transformations following seasonal variations in rainfall, including increased mineralization, nitrification, lixiviation, and decreased denitrification. Model estimates of monthly DON and NH4 in stream water concentration were comparable to measured values (0.37 ± 0.22; 0.065 ± 0.06, respectively). Overall, this study demonstrated that the D-SEM model can predict water, C and N dynamics in drained forests under natural conditions in the Amazon. However, discrepancies between the magnitudes of the estimates suggest that the model requires refinement for further application in tropical forested catchments.