Decoupling effects, efficiency evolution, and driving mechanisms of carbon emissions in water-rich cities of Guangxi

Based on data from prefectural-level cities in Guangxi from 2006 to 2023, this study examines the carbon emissions decoupling states in water-rich cities using the Tapio decoupling model. Carbon emission efficiency is measured through an SBM-DEA model with undesirable outputs and the Malmquist method, followed by spatiotemporal evolution analysis and Markov chain simulations. The driving mechanisms are investigated using the Lasso regression in machine learning and the Tobit model. The findings reveal that, the dominant decoupling type in Guangxi water-rich water-rich cities is growth-negative decoupling, while carbon emission efficiency exhibits a fluctuating upward trend, primarily driven by technological catch-up effects. Temporally, the evolution of carbon emission efficiency can be divided into three phases: low-efficiency fluctuation, unstable preparatory transition, and gradual high-efficiency transformation. Cities within the Beibu Gulf Urban Agglomeration, supported by regional planning policies, demonstrated leading performance. Spatially, the center of gravity of carbon emission efficiency in Guangxi water-rich water-rich cities shifted southward, forming a synergistic pattern where larger cities drive improvements in surrounding smaller cities. Industrial structure, innovation capacity, and household consumption significantly drive carbon emission efficiency, with industrial structure and residential consumption exerting significantly stronger effects in cities influenced by the Beibu Gulf Urban Agglomeration planning.