The acceleration of urbanization has significantly exacerbated climate change due to excessive anthropogenic carbon emissions and air pollutants. Based on data from 281 prefecture-level cities in China between 2015 and 2021. The spatiotemporal co-evolution of urban carbon emissions and air pollutants was analyzed through map visualization and kernel density estimation, revealing non-equilibrium and heterogeneity. Extreme gradient boosting (XGBoost) multiscale geographically weighted regression models(MGWR) and SHAP theory from game theory were employed to deeply investigate the disparities in relevance, spatial heterogeneity, and multiscale fluctuations of carbon emissions and air pollution. The main results are summarized as follows: (1) Between 2015 and 2018, CO₂ emissions exhibited significant fluctuations, while SO₂ and PM_2.5 concentrations decreased markedly. (2) The XGBoost-SHAP model identified seven key driving factors, demonstrating high precision, the SHAP model is used to explain the model and reveal the influence of driving factors on carbon emissions. (3) The concentrations of CO₂, SO₂, and PM_2.5 were positively correlated, the influence of each factor exhibited significant spatiotemporal differences, with varying directions of fluctuation across different regions. Thus, the symbiotic relationship between carbon emissions and air pollutants can inform decision-making for regional planning and sustainable urban development.

As a high-risk production unit, natural gas supply enterprises are increasingly recognizing the need to enhance production safety management. Traditional process warning methods, which rely on fixed alarm values, often fail to adequately account for dynamic changes in the production process. To address this issue, this study utilizes deep learning techniques to enhance the accuracy and reliability of natural gas load forecasting. By considering the benefits and feasibility of integrating multiple models, a VMD-CNN-LSTM-Self-Attention interval prediction method was innovatively proposed and developed. Empirical research was conducted using data from natural gas field station outgoing loads. The primary model constructed is a deep learning model for interval prediction of natural gas loads, which implements a graded alarm mechanism based on 85%, 90%, and 95% confidence intervals of real-time observations. This approach represents a novel strategy for enhancing enterprise safety production management. Experimental results demonstrate that the proposed method outperforms traditional warning models, reducing MAE, MAPE, MESE, and REMS by 1.13096 m³/h, 1.3504%, 7.6363 m³/h, 1.6743 m³/h, respectively, while improving R² by 0.04698. These findings are expected to offer valuable insights for enhancing safe production management in the natural gas industry and provide new perspectives for the industry’s digital and intelligent transformation.