Herein, we present the first systematic investigation to clarify the effect of hydrological rhythms on the concentrations and distributions of polycyclic aromatic hydrocarbons (PAHs) and phthalate esters (PAEs) in the Danjiangkou Reservoir. The results revealed that hydrological rhythms remarkably affected the PAH and PAE concentrations and distributions in the water body, wherein the PAH concentration peaked in the flood season while the PAE concentration remarkably increased in the dry season. This study represents methodological innovation, revealing significant heterogeneity of PAHs and PAEs across different water layers. The former compounds tended to accumulate in the water body’s bottom layer while the latter compounds had the highest concentration at the surface layer, which can be attributed to the different physicochemical properties and environmental transport behaviors of the two compound types. The overall concentrations of PAHs and PAEs fall within the international and domestic safety standards. The primary sources of these contaminants—coal and biomass combustion for PAHs and widespread use of plastic products for PAEs—are critical areas of regulatory focus. A machine learning model is proposed for the first time for predicting PAE concentrations in the Danjiangkou Reservoir, primarily based on the stacking model and supplemented by the random forest or XGBoost models.

中文翻译：

---

水文节律驱动大型油气藏内分泌干扰物特征、来源及浓度预测——以丹江口水库为例


在此，我们提出了第一个系统的研究，以阐明水文节律对丹江口油气中多环芳烃 （PAH） 和邻苯二甲酸酯 （PAE） 浓度和分布的影响。结果表明，水文节律显著影响了水体中 PAH 和 PAE 的浓度和分布，其中 PAH 浓度在汛期达到峰值，而 PAE 浓度在旱季显著增加。这项研究代表了方法学创新，揭示了不同水层中 PAH 和 PAE 的显著异质性。前者倾向于在水体底层积累，而后者在表层的浓度最高，这可以归因于两种化合物类型不同的物理化学性质和环境传输行为。PAH 和 PAE 的总体浓度符合国际和国内安全标准。这些污染物的主要来源——多环芳烃的煤和生物质燃烧以及 PAE 塑料产品的广泛使用——是监管关注的关键领域。首次提出了一种机器学习模型，用于预测丹江口油藏的PAE浓度，该模型主要基于堆积模型，并辅以随机森林或XGBoost模型。