Source facilities can manage stormwater runoff pollutants while also limiting runoff volume in China’s sponge city construction. However, there is no apparent link between source volume control and runoff pollution reduction. This research uses SWMM to develop a model based on the basic data of a city’s catchment zone in northern China in order to statistically examine the response relationship between the initial runoff volume capture of sponge city construction and the reduction effect of stormwater runoff pollution. Based on the dimensionless constant of the runoff pollution control curve, we suggested a novel approach for evaluating runoff pollution control effectiveness. This method was used to examine the response relationship between runoff volume control and runoff pollution control in three typical initial runoff volume capture facilities in the study area, namely bioretention, permeable pavement, rain barrels, and terminal storage tanks, under various design rainfall conditions. The dimensionless constant (Dc) has good practical application value and can quantitatively quantify the control effect of various facilities on stormwater runoff pollution. The Dc value of the source facilities is higher than that of the terminal control facilities, indicating that the source facilities have a higher ability to reduce rainwater runoff pollution than the terminal facilities, particularly the infiltration facilities. The research ideas and evaluation methods proposed in this paper provide an alternative approach for selecting and optimizing facilities during the planning and designing stages of sponge city construction, which can also be used in the monitoring and evaluation stage after completion to carry out a more effective evaluation of sponge facilities’ effects.

Generative artificial intelligence (GAI) enables students to offload cognitive tasks to an external system, yet the consequences of such cognitive offloading for the development of critical thinking—a core dimension of human intelligence—remain underexplored. Drawing upon cognitive offloading theory and distributed cognition theory, this study investigates the heterogeneity of critical thinking outcomes among Chinese university students who use GAI, focusing on how different patterns of human–AI collaboration relate to cognitive autonomy relinquishment. A questionnaire survey was administered to 353 university students across multiple provinces in China. Cluster analysis and regression analysis were employed to identify distinct user profiles and to examine predictors of critical thinking gains and cognitive autonomy. Four distinct user profiles emerged, ranging from “simple Q&A users” (25.2%) to “critical co-thinkers” (15.6%). Learning motivation was the strongest predictor of both critical thinking gains (β = 0.42) and lower cognitive autonomy relinquishment (β = −0.35). Notably, offloading depth positively predicted cognitive autonomy relinquishment (β = 0.25), revealing a paradoxical pattern: sophisticated GAI use was associated with greater dependence. A “high depth–high dependence” subgroup (25.8%) was identified, disproportionately composed of female students and Information and Communication Technology (ICT) majors. The findings challenge the assumption that deeper GAI engagement automatically yields cognitive benefits. Because all constructs were measured through self-report, the findings are interpreted as reflecting students’ perceptions of their cognitive behaviors and abilities; the methodological implications of this design are discussed in detail. Educational interventions should prioritize metacognitive training over technical skill development to ensure that cognitive offloading enhances rather than undermines critical thinking.

Solving the problem of waterlogging is of great significance to the protection of historic districts and urban renewal. To solve the problem of waterlogging, it is necessary to comprehensively consider technical aspects such as the “Major-Minor-Micro” drainage system and the connection of urban/watershed flood control systems. At the same time, attention should be paid to the role of management in the entire system. However, there are a series of problems in the current waterlogging prevention and control system in China’s historic districts. The effectiveness improvement plan of the waterlogging prevention and control system is not sufficiently targeted. The various subsystems in the waterlogging prevention and control system play a role in poor coordination. The waterlogging prevention and control system does not work closely with the flood control system. Different management departments have weak coordination in the prevention and control of waterlogging. Aiming at the problems in the waterlogging prevention and control system in historic districts, this paper proposed a series of countermeasures. The waterlogging prevention and control system in historic districts should make full use of the current advanced concepts and practical experience of stormwater management. In order to enhance the role of the waterlogging prevention and control system in historic districts, specific issues should be analyzed in detail. The waterlogging prevention and control system in historic districts needs to be divided into different levels. The waterlogging prevention and control system should be aligned with historic attributes and be highly pertinent. The waterlogging prevention and control system should be closely coordinated with the larger-scale flood control system. At the same time, the intelligent management platform should be used to increase the supervision of the whole process of waterlogging in historic districts.

Urban flooding poses escalating threats to high-density cities, yet the nonlinear mechanisms linking rainfall characteristics and urban morphology to waterlogging severity remain poorly understood. This study proposes a rainfall-stratified explainable machine learning framework to distinguish deep from shallow inundation at the block scale, taking Shanghai as a case study. Four models (XGBoost, random forest, SVM, and logistic regression) were compared via nested cross-validation and Bayesian optimization, with XGBoost identified as the optimal model. Three physically distinct rainfall dimensions and multi-dimensional urban morphological indicators were incorporated as predictive features. SHAP-based attribution and PDP were employed to unveil the driving mechanisms behind inundation severity, characterizing scenario-dependent shifts in driver dominance and nonlinear threshold effects. Urban morphology primarily governs spatial risk under non-extreme rainfall, with building shape coefficient (BSC) remaining the primary driver overall. Meteorologically, waterlogging severity surges beyond critical thresholds for maximum hourly rainfall (>18.40 mm/h) and total volume (>139 mm), while duration exhibits an inverted U-shaped response. Morphologically, a high BSC (>0.39 m⁻¹) is consistently associated with elevated deep inundation probability, whereas higher SDBV (>54,155 m³) and greater DR (>582 m) are associated with a severity-attenuating effect. These findings provide threshold-driven insights for integrating morphological resilience into urban renewal and sustainable flood adaptation strategies in high-density metropolises.