极端降雨事件对非点源氮素淋失过程的影响机制研究

Diffuse nitrogen leaching is closely related to hydrological processes, the intensity and frequency of extreme rainfall events have been increasing in China since 50 years ago, runoff generation and nitrogen leaching processes along with hydrological processes under extreme rainfall events are different from that under normal rainfall conditions. In this study, a small typical headwater sub-catchment located at downstream of LeAn River will be selected as study area. By combining in-situ hill slope rainfall-runoff experiments, high frequency hydrological water quality monitoring, and catchment distributed nitrogen leaching modelling, three objectives are covered, including (i) investigating the runoff separation and nitrogen leaching regimes along with different flow pathways under extreme rainfall events; (ii) modifying the calculation modes of key parametrers of hydrological and nitrogen leaching processes, optimising modelling time step; (iii) revealing the impacts of extreme rainfall events on streamwater nitrogen concentration dynamics and nitrogen fluxes by comparing modelling results with that derived from simulations under normal rainfall events. Using multi-source data fusion technique, the information contents of field experiments and high frequency hydrological water quality monitoring data are exploited in model parameter calibration, it enables to improve parameter identification and decrease the uncertainties of parameter calibration and model predictions. The outcomes of this study will enrich the understanding of nitrogen leaching along with runoff pathways under extreme rainfall events, improve modelling accuracy of streamwater nitrogen concentration and nitrogen loads, as well as provide guidelines for development of optical water resource management program.

非点源氮素淋失与水文过程密切相关，中国近50年来极端降雨事件的强度和频次增加，极端降雨下径流过程和氮素随水文过程的迁移与非极端降雨不同，目前这些方面的研究尚显不足。本研究选择乐安河集水域下游典型源区小流域作为研究区，综合运用野外坡面小区径流实验、水文水质高频监测、流域同位素水化学流量过程线分割、流域分布式氮素迁移模拟等方法，研究极端降雨事件下径流分割和氮素随不同径流路径的迁移机制；改进现有模型水文和氮素迁移过程关键参数的计算模式，优化模型的计算步长；模拟比较极端、非极端降雨条件下河流氮素浓度变化和流域氮素输出通量。该研究基于多源数据融合技术，将野外实验、水文水质高频监测等数据信息用于模型参数率定，以达到改善参数辨识，减小参数率定和模型预测不确定性的目标。研究成果有助于加深对极端降雨事件下氮素径流淋失机制的认识，提高氮素浓度和输出通量的模拟精度，为制定流域水资源综合管理措施提供科学依据。

项目立足于近50年来我国极端降雨事件增加，河流水质恶化，而传统低频（月/季节尺度）水质监测数据无法满足精准率定流域营养盐迁移模型的问题，综合运用高频水质观测、流域分布式营养盐输移模拟（HYPE）等方法，研究氮、磷随迁移机制，评价不同监测频率的水质数据对模拟精度和预测不确定性的影响。采用基于贝叶斯理论和随机采样的参数率定与不确定性分析算法DREAM，比较了每两周一次和每日一次硝态氮监测数据对氮素输移模拟的影响。结果表明，两种频率监测数据率定获得的氮素输移过程参数最大似然值相似，日步长数据率定获得的参数后验分布不确定性范围更窄；日步长数据率定的模型能够更好地再现日步长和每两周步长的硝态氮浓度数据；日步长数据大大减小了参数率定不确定性引起的预测不确定性，而对总预测不确定性的减小程度有限；模型结构偏差和监测误差是预测不确定性的主要来源，这些因素限制了高频水质数据的效应，适合的采样频率取决于模型结构和监测误差。. 采用HYPE模型模拟了典型亚热带季风流域（乐安河流域）的径流与总磷输移。采用PEST进行参数敏感性分析发现径流模拟对代表性土地利用（林地）的潜在蒸散发速率、代表性土壤类型（红壤）的最上层径流衰减系数最为敏感，磷素输出模拟对地表径流土壤侵蚀方程的指数、吸附/解吸附方程的指数和系数最为敏感。HYPE模型能够很好地模拟不同气象水文条件下的径流及其时空变化（NSE≥0.8，相对偏差绝对值≤14.5%），模型能较好地捕捉总磷浓度与负荷的变化，对溶解态磷和颗粒态磷的模拟偏差较大。总磷负荷与径流呈密切相关性，暴雨径流事件对总磷流失的贡献最大，表明流域磷素流失主要受暴雨径流控制。. 项目研究成果对优化制定取决于流域水文水环境研究和管理目标的流域水质监测方案，揭示流域氮、磷输移及时空变化机制，制定流域面源污染削减方案，改善和保护河湖水环境具有重要的科学价值和应用前景。