Key Laboratory of Pollution Processes and Environmental Criteria (Ministry of Education), Tianjin Key Laboratory of Environmental Remediation and Pollution Control, College of Environmental Science and Engineering, Nankai University, Tianjin 300350, China.
Key Laboratory of Pollution Processes and Environmental Criteria (Ministry of Education), Tianjin Key Laboratory of Environmental Remediation and Pollution Control, College of Environmental Science and Engineering, Nankai University, Tianjin 300350, China.
Environ Int. 2022 Apr;162:107172. doi: 10.1016/j.envint.2022.107172. Epub 2022 Mar 12.
Microplastic (MP) contamination has become an increasingly serious environmental problem. However, the risks of MP contamination in complex global climatic and geographic scenarios remain unclear. We established a multifeature superposition analysis boosting (MFAB) machine learning (ML) approach to address the above knowledge gap. MFAB-ML identified and predicted the importance, interaction networks and superposition effects of multiple features, including 34 characteristic variables (e.g., MP contamination and climatic and geographic variables), from 1354 samples distributed globally. MFAB-ML analysis achieved realistic and significant results, in some cases even opposite to those obtained using a single or a few features, revealing the importance of considering complicated scenarios. We found that the microbial diversity in East Asian seas will continually decrease due to the superposition effects of MPs with ocean warming; for example, the Chao1 index will decrease by 10.32% by 2065. The present work provides a powerful approach to identify and predict the multifeature superposition effects of pollutants on realistic environments in complicated climatic and geographic scenarios, overcoming the bias from general studies.
微塑料（MP）污染已成为一个日益严重的环境问题。然而，在复杂的全球气候和地理情景下，MP 污染的风险仍不清楚。我们建立了一种多特征叠加分析增强（MFAB）机器学习（ML）方法来解决上述知识空白。MFAB-ML 从全球分布的 1354 个样本中识别和预测了多个特征（如 MP 污染以及气候和地理变量）的重要性、相互作用网络和叠加效应。MFAB-ML 分析实现了现实和显著的结果，在某些情况下，甚至与使用单个或少数特征获得的结果相反，揭示了考虑复杂情景的重要性。我们发现，由于 MPs 与海洋变暖的叠加效应，东亚海域的微生物多样性将持续减少；例如，到 2065 年，Chao1 指数将减少 10.32%。本研究提供了一种强大的方法，可用于识别和预测污染物在复杂气候和地理情景下对现实环境的多特征叠加效应，克服了一般研究的偏差。
