As urbanization accelerates, the contamination of urban soil and the consequent health implications stemming from urban expansion are increasingly salient. In recent years, a plethora of cities and regions nationwide have embarked on rigorous soil geological surveys with a focus on environmental quality, yielding invaluable foundational data. This research aims to develop scientifically robust and rational land-use planning strategies while assessing the levels of heavy metal pollution and associated risks. The urban agglomeration encompassing Zhengzhou, Luoyang, and Kaifeng (referred to as Zheng–Bian–Luo Urban Agglomeration) in Henan Province was designated as the study area. Leveraging the Nemerow comprehensive index method alongside the Hakanson potential ecological risk assessment method, this study delved into the pollution levels and potential ecological ramifications of nine heavy metals, namely Cr, Mn, Ni, Cu, Zn, As, Cd, Pb, and Co. Research indicates that the hierarchy of individual potential ecological risks ranges from most to least significant as follows: Cd > Pb > Cr > Ni > Cu > Zn > As > Mn > Co. The concentrations of Cd in both Zhengzhou and Kaifeng surpassed the established background levels. Furthermore, the mean single-factor pollution index values for the heavy metals Cd and Zn exceeded 1, signifying a state of minor pollution. The Nemerow comprehensive index P of Cd and Zn is between 1 < P_comp ≤ 2, which is considered mild pollution. The comprehensive P values of the other seven metal elements are all less than 0.7, reaching a clean (alert) level. Predominantly, the primary potential risk factor in the superficial soil of the Zheng–Bian–Luo urban agglomeration is Cd, while the ecological risk implications associated with other heavy metal elements are comparatively minimal. The soil environmental quality within the designated study area remains secure, although certain localized areas pose potential risks of heavy metal pollution. A comprehensive assessment of the current state of soil heavy metal pollution is essential to establish a theoretical foundation and provide technical support for soil environmental protection, pollution mitigation, and sustainable utilization.

We propose Mamba - Physics-Informed Neural Network(Mamba-PINN), a novel data–physics integrated resilience assessment model, to evaluate the anti-typhoon disturbance ability of coastal sewage treatment systems in Zhuhai. The increasing frequency of extreme weather events poses significant challenges to urban infrastructure; yet, existing methods often fail to capture the complex spatio-temporal dynamics of typhoon impacts. Our approach combines a Mamba neural network for high-frequency monitoring data processing with Physics-Informed Neural Networks (PINN) to quantify process recovery dynamics under typhoon conditions. The Mamba network extracts critical storm impact features, while the PINN embeds fluid inertia and microbial activity inhibition mechanisms to model system responses. Furthermore, we introduce a disturbance–recovery resilience index to provide a quantitative measure of system robustness, enabling targeted adaptive transformations for coastal sewage plants. The proposed method addresses the limitations of purely data-driven or physics-based models by integrating both paradigms, offering a more comprehensive understanding of resilience mechanisms. Experimental results demonstrate the model’s effectiveness in capturing nonlinear interactions between typhoon disturbances and treatment process recovery. This work contributes to the sustainable development of coastal cities by providing a scientifically grounded framework for infrastructure adaptation under climate change.