中文摘要：

【目的】选择描述黄土高原石灰性土壤氮素矿化过程的合适模型，明确可溶性有机氮（soluble organic nitrogen，SON）对矿化模型拟合效果的影响。【方法】采用长期间隙淋洗淹水培养方法，研究了10种黄土高原主要农业土壤在包括和不包括浸提液中有机氮的情况下氮素矿化过程，在此基础上选择①有效积温式；②一级反应式（One-pool模型）；③两部分一级反应式（Two-pool模型）；④带常数项一级反应式（Special模型）对这两种情况下的氮素矿化曲线进行拟合。【结果】发现4种模型在拟合包括SON的氮素矿化曲线时效果更好。从模型均方根误差、估计标准误差、参数误差以及与作物吸氮量的相关分析等比较发现，One-pool模型拟合效果最差，Two-pool模型和Special模型优于有效积温模型，而Special模型参数精度及与作物吸氮量的相关性更高。【结论】综合比较认为Special模型能较好反映石灰性土壤在淹水条件下的氮素矿化过程。

英文摘要：

[Objective] In order to choose a proper model for depicting soil nitrogen mineralization process on the Loess Plateau, and find out the effect of soluble organic nitrogen （SON） on the study of soil nitrogen mineralization dynamics and model fitting. [Method] Nitrogen mineralization process of 10 farmland soils on the Loess Plateau was investigated. Two aspects of the long-term alternate leaching water-logged incubation were examined： cumulative mineralized nitrogen without soluble organic nitrogen （SON） and with SON. Net cumulative mineralized nitrogen over 217 days was analyzed in terms of 4 models of N mineralization curves with and without SON： （1）Effective accumulated temperature model （EATM）; （2）One-pool first order kinetic model （One-pool model）; （3）Two-pool first order kinetic model （Two-pool model）; （4）Combined first-order, zero-order kinetic model （Special model）. [Result] It showed that 4 models gave better results to nitrogen mineralization curves with SON. The comparison of model standard error estimate, relative error of parameter estimations and correlation analysis of model parameter with N-uptake by crop indicated that the One-pool model provided the worst fit, both Two-pool model and Special model fit the data closer than effective cumulated temperature model, but the precision of parameter estimations and correlation with N-uptake by crop using special model was relatively high. [Conclusion] Results from this study showed that Special model could simulate preferably N mineralization process of calcareous soil on the Loess Plateau with long-term alternate leaching water-logged incubation method.