The experiment of loess crack development under dry–wet cycle conditions is of great significance for the study of groundwater preferential flow channels and the prevention and control of infrastructure engineering disasters in loess areas. The loess samples in Chencang District of Baoji City, Shaanxi Province, were taken as the samples in the test. The multiple humidification and dehumidification tests were used to simulate multiple rainfall evaporation, and the moisture content changes in the loess samples during the dry–wet cycle were calculated. With the help of digital image technology, the fracture parameters of the loess samples were extracted, and the variation law of crack parameters was analyzed by combining fractal dimension, Bayesian factor, and Pearson correlation coefficient. The findings indicate that variations in soil moisture content and the number of dry and wet cycles contribute to fluctuations in soil evaporation rates, resulting in varying degrees of soil cracking development. The increase in the number of dry and wet cycles leads to evident soil shrinkage, an accelerated water evaporation process, pronounced surface deterioration, and a higher degree of crack development. The rate of crack propagation varies at different locations, with a higher rate observed in the horizontal plane compared to the vertical plane. The influence of temperature and humidity varies due to the different dimensions of cracks (horizontal and vertical). Horizontal crack development is primarily influenced by temperature, while vertical crack development is primarily influenced by humidity. Temperature and humidity inhibit each other. When one factor is dominant, the other indirectly affects crack development by influencing the dominant factor. The research findings can serve as a valuable reference for effectively mitigating and minimizing the impact of crack development-induced disasters.

Groundwater is one of the main water supplies for Zhengzhou City. Changes in groundwater depth negatively affect the water supply and geological stability of the city. Based on groundwater monitoring data from 1999 to 2024, this paper investigates the spatiotemporal changes in shallow groundwater depth in Zhengzhou and their driving factors. Trend analysis, ordinary Kriging interpolation, and Pearson correlation analysis were first employed to characterize the spatiotemporal evolution patterns of groundwater depth. A principal component regression (PCR) framework was then applied to reduce multicollinearity among hydro-climatic and water-use variables and to quantify their integrated relationships with groundwater depth at individual monitoring wells. The main findings are as follows: (1) Over the past 25 years, groundwater depth in Zhengzhou has gradually increased, with an average decline rate of 0.68 m per year. The most rapid increase occurred from 1999 to 2012. (2) Spatially, groundwater depth is greatest in the central urban area (average 35.2 m) and decreases with distance from the city center, forming a depression cone in the downtown area. (3) Anthropogenic factors dominate groundwater depth variation. Groundwater extraction shows a strong positive correlation with groundwater depth, and standardized regression coefficients indicate that domestic water consumption and groundwater supply are the key drivers of groundwater deepening and recovery, respectively. Rapid urbanization has led to water table decline, while changes in precipitation have had relatively minor effects. (4) Socioeconomic factors, such as expansion of construction land, population growth, and industrialization, have also contributed to the increase in groundwater depth by raising water demand and reducing infiltration recharge. The PCR-based analysis provides a parsimonious and spatially explicit basis for understanding groundwater-depth dynamics in a rapidly urbanizing plain city and offers practical guidance for targeted groundwater management, extraction control, and ecological water replenishment in Zhengzhou and similar regions.

The manual identification of non-metallic pipelines in ground-penetrating radar (GPR) images is inefficient and heavily experience-dependent. Existing deep-learning methods suffer from performance degradation caused by image noise, signal attenuation, and false anomalies. To address these challenges, this paper proposes YOLOv8-WT, which introduces a novel WTConv-ATT module combining wavelet transform and multi-dimensional dynamic attention. This module performs multi-level wavelet decomposition to extract frequency-domain features and enhance global and low-frequency information perception; meanwhile spatial-channel-pixel attention adaptively generates feature fusion weights to suppress noise interference. Several existing well-established modules (Wise-IoU, C2f-FSDA, CBAM, EMA) are also integrated to further boost detection performance. Experimental results on merged public GPR datasets show that compared with the YOLOv8s baseline, the proposed model achieves increases of 1.27, 4.59, and 2.55 percentage points in Precision, Recall, and mAP50, reaching 92.98%, 91.83%, and 96.76%, respectively. YOLOv8-WT obtains promising detection performance for non-metallic pipelines under mixed-dataset conditions. Further validation is still required for unknown real-world GPR scenarios.