三峡库区典型流域硝态氮输出特征及归因分析

为研究流域硝态氮来源、输出特征及驱动因素，应用SWAT模型对三峡库区梅溪河和大宁河流域径流和硝态氮负荷进行模拟，进而解析流域水文过程及硝态氮来源，并基于随机森林模型量化了不同影响因素（气候、土地利用、土壤类型、地形）对径流和硝态氮负荷的影响程度。结果表明：（1）不同土地利用类型间硝态氮负荷差异显著，年负荷强度表现为园地（20.41 kg/hm²）>旱地（12.51 kg/hm²）>水田（10.31 kg/hm²）>建设用地（7.09 kg/hm²）>林地（0.62 kg/hm²）>草地（0.46 kg/hm²），旱地是梅溪河（80%）和大宁河流域（67%）硝态氮输出的主要来源；（2）梅溪河和大宁河流域基流系数分别为67%和62%，基流是硝态氮主要运移途径，分别贡献68%和60%的硝态氮输出；（3）径流分配和硝态氮输出具有明显季节变异性，旱季基流对2个流域径流和硝态氮的贡献均在70%以上，雨季地表径流输出的硝态氮分别占36%和42%；（4）降雨量是影响总径流的主要因素，土壤类型是影响地表径流和基流的主要因素；土地利用是影响不同径流途径硝态氮的主要因素，其次是土壤类型，二者相对重要性之和大于70%。综上，环境土地利用冲突是造成硝态氮流失的根本原因，源头控制仍是三峡库区面源污染防控的关键环节；在地表径流控制的基础上亟待纳入旱地和园地基流途径控制策略。

Export Characteristics of Nitrate-N and Its Dominant Factors in Typical Basins of the Three Gorges Reservoir Area

In order to study the sources, export characteristics and influencing factors of nitrate-N, the SWAT model was implemented to simulate the runoff and nitrate-N loads in the Meixi River and Daning River basins in the Three Gorges Reservoir Area (TGRA). Then, the hydrological processes and sources of nitrate-N in the basins were analyzed. Based on the random forest model, the effects of different influencing factors (i.e., climate, land use, soil type and topography) on runoff and nitrate-N loads were also quantified. The results showed that:(1) The nitrate-N load varied significantly among different land use types, and the annual load intensity followed the order of orchard (20.41 kg/hm²)>dryland (12.51 kg/hm²)>paddy field (10.31 kg/hm²)>construction land (7.09 kg/hm²)>forestland (0.62 kg/hm²)>grassland (0.46 kg/hm²). Dryland was found to be the main source of nitrate-N export in the Meixi River (80%) and Daning River basins (67%).(2) The base flow coefficient in the Meixi River and Daning River basins was 67% and 62%, respectively, and base flow was the main transport pathway of nitrate-N, contributing 68% and 60% of nitrate-N export in the two basins, respectively. (3) Runoff distribution and nitrate-N export had obvious seasonal variation, the base flow contributed above 70% to both runoff and nitrate-N in these two basins in the dry season, and surface runoff contributed 36% and 42% to nitrate-N export in the two basins in the rainy season, respectively. (4) Rainfall was the main factor affected total runoff, and soil type was the main factor affected surface runoff and base flow. Land use was the main factor affected nitrate-N in different runoff pathways,followed by soil type, and the sum of their relative importance was over 70%. In conclusion, environmental land use conflicts were the root cause of nitrate-N loss, and source control was still a key of non-point source pollution prevention and control in the TGRA. In addition to surface runoff control, it was urgent to incorporate dry land and orchard base flow pathway control strategies.