Application of global sensitivity analysis in identification of herbicides cocktail effects at environment-related concentrations

Abstract

Listen

Translate article

<p indent="0mm">The use of herbicides in crop production continues to increase globally. By 2015, China had become one of the most important producers and consumers of herbicides in the world. Many herbicides are persistent organic pollutants and are used continuously for long periods of time, causing the high residual concentrations in the water surrounding farmland and negative effects on aquatic life. The concentrations of herbicides in the environmental water are variable and the types of herbicides are diverse. Herbicide mixtures in surface water may threaten aquatic life. However, in most areas of China, the hazard of herbicide mixtures associated with environmental exposure levels has not been assessed. This paper establishes a hazard identification assessment method for mixtures with environment-related concentrations. Firstly, the hazard of individual pollutant was assessed by hazard quotient (HQ), and the pollutants with high hazards (HQ≥1) to target organisms were further studied. Secondly, the global sensitivity analysis (GSA) based on Morris method was used to study combined effect, and identify the important or hidden drivers in chemical mixtures. In brief, the GSA experimental design consists of four steps: Input factors (target pollutants), Morris sampling, high-throughput screening, and the sensitivity analysis. Thirty-nine herbicides detected in the surface water of Tai Lake and Liao River were screened by HQ with <italic>Scenedesmus obliquus</italic> (<italic>S</italic>.<italic> obliquus</italic>) individually. The herbicides with high hazard were the input factors of GSA. Then, the Morris method was used to generate 150 samples of mixtures. The <italic>S</italic>.<italic> obliquus </italic>was exposed to 150 samples for <sc>24 h</sc> and subjected to high-throughput screening. Finally, the sensitivity analysis was used to identify the important or hidden drivers of the herbicide mixtures. The results showed that 14 herbicides had high hazards (HQ≥1) among 39 herbicides detected in Tai Lake and Liao River basin. Ametryn and diflufenican had higher hazards to algae. At the individual herbicide exposure of <sc>500 μg/L,</sc> most herbicides, except C5 and C10, significantly inhibited the growth of <italic>S</italic>.<italic> obliquus </italic>(<italic>P</italic>&lt;0.05). And the strongest growth inhibition effect caused by individual herbicide was about 40%. The results of 150 samples of mixtures with environment-related concentration showed that most of the mixtures had growth inhibition effects on <italic>S</italic>.<italic> obliquus</italic>. The total concentrations of the mixtures that consist of 14 herbicides are not greater than <sc>79 μg/L.</sc> And the strongest growth inhibition effect caused by mixture samples was over 50%. The toxicity of herbicide mixtures cannot be ignored. The growth inhibition effect of the mixtures that consist of 14 herbicides on <italic>S</italic>.<italic> obliquus</italic> was nonlinear, and there are strong interactions between herbicides. It is difficult to predict the toxic effects of the mixture system from the exposure concentration of individual herbicides. According to the <italic>μ*</italic> value, the most important factors in the combined pollution were terbutryne, bensulfuron methyl, atrazine, and ametryn. Bentazone had the lowest <italic>μ*</italic> and contributed little to the overall mixture system. And bensulfuron methyl was the hidden driver of the mixture system. The important drivers in the herbicide mixtures need to be focused, and their concentrations should be monitored regularly to control emissions and minimize the negative effect on aquatic life. Global sensitivity analysis based on Morris sampling can identify the important factors in cocktail effects of mixtures in limited samples, providing technical support for incorporating the actual toxic effects of mixtures into the risk assessment of pollutants.