From Rigid Splits to Soft Frontiers: A Fuzzy Random Forest Framework for Eco-Efficiency Benchmarking of Wastewater Treatment Plants

Achieving carbon neutrality in the wastewater sector requires reconciling pollutant removal and degradation with greenhouse gas (GHG) emissions, positioning eco-efficiency assessment as a central task in sustainable wastewater treatment engineering. Existing methods based on data envelopment analysis (DEA) and efficiency evaluation trees are limited by the curse of dimensionality and step distortions caused by hard boundaries, making it difficult to simultaneously address the multisource heterogeneity of large-scale wastewater treatment plants (WWTPs) and the continuous nonlinear characteristics of biochemical reactions. This study proposes a hybrid analytical framework that integrates consensus matrix clustering with the Fuzzy-RFEAT (Fuzzy Random Forest Eco-Efficiency Assessment Tree), embedding a sigmoid-type fuzzy membership function into the random forest frontier estimation to replace the rigid binary splits of traditional tree models with a soft-boundary mechanism. An empirical evaluation based on 4999 municipal wastewater treatment plants in mainland China indicates that consensus clustering decomposes the samples into eight functional phenotypes with significant process and scale differentiation; Fuzzy-RFEAT consistently achieved the optimal fitting accuracy across all 8 phenotypes, with the lowest coefficient of variation (0.50) for the cross-cluster RMSE among the four comparison models, validating the global consistency of the soft-boundary strategy across different scenarios. Furthermore, through an importance analysis of efficiency scores, this study identified three types of eco-efficiency bottleneck structures─physical pretreatment-constrained, single-variable-dominated, and multivariable-synergistic-constrained─and proposed typology-based targeted intervention strategies accordingly. This framework provides a rigorous, data-driven benchmark to guide refined efficiency benchmarking and prioritize differentiated low-carbon retrofits in engineered wastewater treatment systems.