冬小麦生育期农田尺度下土壤硝态氮淋失动态的数值模拟

Numerical simulation on nitrate leaching at field scale during the growth of winter wheat

在北京通州区永乐店田间试验的基础上 ,假设土壤由一系列不发生相互作用的一维土柱组成 ,根据实测的土壤有机质含量 ,假定土壤有机氮的矿化作用速率常数 (零级动力学 )和有机质含量成正比 ,运用 HYDRUS- 1D软件 ,分别就考虑和不考虑土壤有机氮的矿化速率的空间变异性这两种方案 ,对 2 0 0 0～ 2 0 0 1年冬小麦生长条件下农田尺度土壤氮素转化和硝态氮淋失规律进行了数值分析。两种方案的模拟结果表明 :考虑和不考虑土壤有机氮矿化速率的空间变异性对剖面 2 5 0 cm埋深处硝态氮淋失量的影响很小 ,其差异主要在于前者对土壤氮素的矿化量、固持及反硝化量、作物吸氮量的影响更大 ,其空间变异性高于不考虑矿化速率时的结果。剖面 2 5 0 cm埋深处平均的土壤水渗透量和累积硝态氮淋失量分别为 2 .2 5 mm、0 .0 0 984 m g/cm2 ,变异系数大于 1.4 6 ,属于强变异性。对模拟结果进行地统计学分析 ,结果表明 :剖面 2 5 0 cm埋深处的土壤水渗透量和硝态氮淋失量的半方差函数为纯块金形式 ,在空间上表现为相互独立。考虑有机氮矿化速率空间变异性时的土壤氮素净转化量、吸氮量均可用球状模型描述 ,其变程与土壤有机质含量的变程接近 ,约为 4 .7m;

It is critical to understand nitrogen transformation and nitrate leaching at field scale for establishing agricultural ecosystem. However, nitrate transport in soils is a complicated process, affected by physical, chemical, and biological properties of the porous media, because of the spatial variability of soils. Mathematical model, which can provide essential information to us, is a very effective and convenient tool to solve the problem on water quality degradation resulted from nitrate leaching. In order to obtain more accurate simulation and prediction on nitrate leaching at field scale, we incorporated spatial variability of soil hydraulic and nitrogen mineralization parameters in the model. Field experiment was conducted in a 27×27m^2 plot (3m×3m grid) in Yongledian, Tongzhou District, Beijing. Physical and chemical properties, e.g., texture, OM, pH, EC, were measured at all 100 grid nodes. Based on the assumption that the field was composed of a number of non-interacting one-dimensional columns, nitrate-nitrogen leaching was simulated during the growing season of winter wheat. According to measured soil organic matter content, it was assumed that zero order mineralization rate constant was appropriate to soil organic matter content and then its spatial distribution was obtained. Two scenarios, with and without considering spatial variability of mineralization rate, were incorporated in this study to assess the impact of mineralization on N leaching at the depth of 250cm. By HYDRUS-1D software, N transformation, soil water percolation, and nitrate leaching at the depth of 250cm were simulated at each sampling point. The results showed that there was little difference between nitrate leaching with and without considering spatial variability of mineralization, but some transformation terms and N uptake were much different. The average soil water percolation and nitrate leaching at field scale were 2.25mm, 0.00984 mg/cm^2 (CV>1.46) respectively. Results of Geostatistics showed that semi-variance of water percolation and nitrate leaching amount can be described by pure nugget model. Spatial variability of mineralization mainly affected mineralization, immobilization & denitrification, net transformation, and N uptake. Semi-variance of net transformation and N uptake with considering spatial variability of mineralization can be fitted by spherical model and their ranges were similar to that of organic matter (about 4.7m). In the other case, semi-variance of net transformation without considering spatial variability of mineralization can be fitted by linear without sill model and there was no finite semi-variance. Furthermore, the relationships between nitrate leaching and bulk density, K_S,θ_S, and α were obviously significant, and R^2 were about 0.65, 0.31, 0.22, 0.15 respectively. This study provides a sight that nitrate leaching shows spatial variability at field scale because of spatial variation of soil properties. Mineralization mainly affects nitrogen transformation, and shows less effect on nitrate leaching at the depth of 2.5m. Geostatistics analysis showes that semi-variance of nitrate leaching amount can be described by pure nugget model. The coefficient of determination between bulk density was highest (0.65). Therefore, further research is still needed to test the results being obtained in this study under different conditions.