Quantifying nitrate sources in a large reservoir for drinking water by using stable isotopes and a Bayesian isotope mixing model.


College of Environment, Zhejiang University of Technology, Hangzhou, 310032, China. [Email]


Drinking water reservoirs are threatened globally by anthropogenic nitrogen pollution. Hydrochemistry and isotopes were analyzed to identify spatial and temporal varieties of main nitrate sources in a large drinking water reservoir in East China. The results showed that NO3- was the main nitrogen form in both the dry and wet seasons, but dissolved organic nitrogen (DON) was increased in the wet season. The δ15N-NO3- values (+ 1.3‰ to + 11.8‰) and δ18O-NO3- values (+ 2.5‰ to + 13.5‰), combined with principal component analysis (PCA), indicated that chemical fertilizer was the main nitrate source during the dry season, while chemical fertilizer, soil N, and sewage/manure were the main nitrate sources during the wet season in the Qiandao Lake area. And, the nitrate isotopes showed the significant nitrification and assimilation in the Qiandao Lake area. A Bayesian isotopic mixing model (Stable Isotope Analysis in R) was applied to the spatial and seasonal trends in the proportional contribution of four NO3- sources (chemical fertilizer (CF), soil nitrogen (SN), sewage and manure (SM), and atmospheric deposition (AD)) in the Qiandao Lake area. It was revealed that CF was the most important nitrate source in the dry season, accounting for 53.4% with 19.2% of SM and 18.9% of SN, while the contribution of SN increased in the wet season, accounting for 31.6%, followed by CF (30.8%) and then SM (24.2%). The main nitrate sources in the urban area, rural area, and central lake area were CF and SN, accounting for 66.1% in the urban area, 71.7% in the rural area, and 68.2% in the central lake area. Measures should be made to improve chemical fertilizer use efficiency and to reduce nitrogen loss in the Qiandao Lake area. Graphical abstract .


Bayesian isotope mixing model,Nitrate,PCA,Reservoir,Stable isotopes,