Abstract:

Under climate change, the increasing frequency of high-intensity rainfall has complicated non-point source pollution (NPS) dynamics, necessitating further research to elucidate emission patterns and support refined pollution control in water source areas. Based on field monitoring data from moderate, heavy, and torrential rain events, this study analyzed the dynamic emission patterns and first-flush effects of NPS pollution from typical underlying surfaces in water source areas. Runoff coefficients ranged from 0.55 to 0.99 for roads and residential areas, but only 0-0.71 for paddies, vegetable fields, and forests, with peak runoff delayed by 4-12 h in the latter. The results showed that under the three rainfall events, the runoff coefficients of traffic roads and residential areas ranged from 0.55 to 0.99, whereas those of paddy fields, vegetable fields, and forestlands ranged from 0 to 0.71, with the peak runoff of the latter lagging 4-12 h behind that of former. The peak fluxes of COD, TN, and TP from traffic roads were 0.10-0.53, 0.01-0.04, and 0.0047-0.0006 g·h⁻¹·m⁻² higher than those from residential areas, respectively. Under torrential rainfall, paddy and vegetable fields exhibited even higher peak fluxes, reaching 1.3-1.6 times (COD), 4.0-5.4 times (TN), and 2.9-4.6 times (TP) the fluxes of traffic roads. Under moderate rainfall, the Event Mean Concentrations (EMCs) of COD, TN, and TP from both traffic roads and rural residential areas were 1.4-1.7 times, 6.5-7.8 times, and 5.9-6.6 times higher than those under torrential rainfall, respectively. Notably, moderate rainfall events following longer antecedent dry periods significantly elevated pollutant EMCs. Under torrential rainfall, the pollution load coefficients of COD, TN, and TP were relatively highest in vegetable fields, being 1.2-1.5 times higher than those in paddy fields, 1.4-6.8 times higher than those on traffic roads, 3.2-8.1 times higher than those in forestlands, and 2.2-10.6 times higher than those in rural residential areas. Regarding the initial flush ratio (R_MFF30) of the three pollutants, under moderate and heavy rain, the order was COD> TN>TP, whereas under torrential rainfall, TP and TN were generally greater than COD. For COD and TN, the R_MFF30 was generally higher for traffic roads and rural residential areas than for vegetable fields, paddy fields, and forestlands. Intercepting the initial 30% of runoff can reduce 30%‒72% of the COD load and 33%‒39% of the TN load. Conversely, for TP, the R_MFF30 was higher in vegetable fields, paddy fields and forestlands. Capturing 30% of the runoff reduced the TP load by 58%‒62%, compared with a reduction of only 20%‒52% from traffic roads and rural residential areas. Therefore, during moderate-to-heavy rainfall events, priority attention should be given to non-point source pollution from impervious underlying surfaces, particularly during rainfall events following extended dry periods. During high-intensity torrential rainfall events, additional focus must be placed on agricultural non-point source pollution, as its flux peaks and initial flush loads may adversely affect the water quality in source water areas.

Key words: water source area, non-point source (NPS) pollution, event mean concentration, load coefficient, first flush ratio

摘要： 气候变化下高强度降雨频发，非点源污染流失更为复杂，其排放规律有待于进一步研究，以期为水源地水环境污染精细化防控提供依据。根据中雨、大雨和大暴雨下野外监测数据，分析水源地典型下垫面非点源污染动态排放规律和初期冲刷效应。结果表明：3场降雨下交通道路和村镇住宅径流系数为0.55－0.99，水田、菜地和林地径流系数为0－0.71，径流峰值后者比前者滞后4－12 h；交通道路COD、TN、TP通量峰值分别高于村镇住宅0.10－0.53、0.01－0.04、0.0047－0.0006 g·h⁻¹·m⁻²，大暴雨下水田和菜地更高，分别是交通道路的1.3－1.6、4.0－5.4和2.9－4.6倍；交通道路和村镇住宅COD、TN、TP的降雨事件平均浓度（EMC）中雨是大暴雨的1.4－1.7、6.5－7.8和5.9－6.6倍，晴天日数长的中雨事件污染物EMC更高；大暴雨下3种污染物负荷系数菜地相对最高，分别是水田、交通道路、林地和村镇住宅的1.2－1.5、1.4－6.8、3.2－8.1和2.2－10.6倍；3种污染物初期冲刷比率R_MFF30中雨和大雨下COD>TN>TP，大暴雨下TP和TN总体大于COD；COD和TN的R_MFF30交通道路和村镇住宅总体高于菜地、水田和林地，截留30%径流分别拦截30%－72%和33%－39%负荷；TP的R_MFF30菜地、水田和林地高于交通道路和村镇住宅，截留30%径流分别拦截58%－62%和20%－52%负荷。因此，在中大雨时应重点关注不透水下垫面非点源污染，特别是前期晴天日数较长的降雨事件；在高峰值大暴雨时还应关注农田非点源污染，其通量峰值和初期冲刷负荷可能会对水源地水质造成不利影响。

关键词: 水源地, 非点源污染, 事件平均浓度, 负荷系数, 初期冲刷比率