摘要
Accurate assessment of greenhouse gas emissions from wastewater treatment plants and water supply systems is essential for achieving climate goals and environmental sustainability. As regulatory requirements intensify and treatment technologies evolve, selection of appropriate emissions quantification methodologies is a strategic concern. Four methodological approaches—emission factor method, mechanistic modeling, machine learning, life cycle assessment—are critically assessed. Each is evaluated according to its theoretical foundations, operational complexities, accuracy, data requirements, and scalability. The emission factor method, simple and cost-effective, lacks sensitivity to local operational dynamics. Mechanistic models provide detailed causal insights from physicochemical and biological processes, however, demand precise input data and extensive calibration. Machine learning techniques detect nonlinear patterns, facilitating real-time predictive analytics, but require large, high-quality datasets and face interpretability challenges. LCA delivers a comprehensive, cradle-to-grave evaluation, incorporating direct and indirect emissions, although it requires significant resource investment and specialized expertise. Findings underscore the need for a hybrid framework—integrating dynamic emission factors, real-time analytics, and life cycle perspectives—can enhance accuracy, adaptability, and strategic relevance. The study proposes future research directions, including the development of context-specific emission factor databases, incorporation of socio-economic parameters into ML models, and seamless integration of mechanistic insights into LCA workflows.