Modeling nitrogen loadings from agricultural soils in southwest China with modified DNDC

Degradation of water quality has been widely observed in China, and loadings of nitrogen (N) and other nutrients from agricultural systems play a key role in the water contamination. Process-based biogeochemical models have been applied to quantify nutrient loading from nonpoint sources at the watershed scale. However, this effort is often hindered by the fact that few existing biogeochemical models of nutrient cycling are able to simulate the two-dimensional soil hydrology. To overcome this challenge, we launched a new attempt to incorporate two fundamental hydrologic features, the Soil Conservation Service curve and the Modified Universal Soil Loss Equation functions, into a biogeochemistry model, Denitrification-Decomposition (DNDC). These two features have been widely utilized to quantify surface runoff and soil erosion in a suite of hydrologic models. We incorporated these features in the DNDC model to allow the biogeochemical and hydrologic processes to exchange data at a daily time step. By including the new features, DNDC gained the additional ability to simulate both horizontal and vertical movements of water and nutrients. The revised DNDC was tested against data sets observed in a small watershed dominated by farmlands in a mountainous area of southwest China. The modeled surface runoff flow, subsurface drainage flow, sediment yield, and N loading were in agreement with observations. To further observe the behaviors of the new model, we conducted a sensitivity test with varied climate, soil, and management conditions. The results indicated that precipitation was the most sensitive factor determining the rate of N loading from the tested site. A Monte Carlo test was conducted to quantify the potential uncertainty derived by variations in four selected input parameters. This study demonstrates that it is feasible and effective to use enhanced biogeochemical models such as DNDC for quantifying N loadings by incorporating basic hydrological features into the model framework.