Metro lines buried in soft soil in coastal regions will inevitably suffer long-term uneven deformation during operation. In this study, the long-term deformation characteristics of metro lines in Shanghai soft deposits are analyzed based on extensive field data of four metro lines, i.e., Shanghai metro lines (SMLs) 1–4. Settlement processes of complete lines and typical sections including both the tunnel and the station are illustrated. The cumulative longitudinal settlements of the four lines in 2008 and 2015 are presented and comparatively analyzed. The bending deformation process is analyzed and severe bending deformation is examined using the curvature radius and the relative deflection. All the extensive field data indicate that (1) SML 1 had the largest cumulative settlement and SML 4 showed the smallest cumulative settlement. The cumulative settlement is closely related to the time when the metro line was built. SMLs 1–4 settled before 2008 and uplifted entirely and slightly since then. (2) The deformation of the metro tunnel was gentle, whereas the deformation of the elevated metro line showed periodic fluctuations. (3) The bending deformation was more severe in the elevated metro line than those in the metro tunnel, although SML 3 was built later than SML 1. (4) Severe bending deformation appeared at the connections among the station and tunnel, tunnel entrance or exit, and side passageway. (5) The curvature radius was more suitable for examining the severe bending deformation of the elevated metro and the relative deflection was more suitable for examining that of the metro tunnel.

Constructed transport facilities in soft deposits deform significantly due to long-term land subsidence. A great deal of research has been carried out to study the land subsidence–induced deformations of metro tunnels, with a few focusing on the performance of highways in soft deposits. Based on extensive field data of the Shanghai Middle Ring Highway (SMRH), this paper investigates the long-term settlement behavior of the SMRH, which consists of ground road and elevated road. Annual longitudinal settlement profiles of the SMRH are presented to study the responses of the ground road and elevated road to land subsidence. The accumulated longitudinal settlements are compared to the ground surface settlements in association with geological conditions. Moreover, a comparison of the deformation processes between the elevated road and subsoils is presented. Field data indicate that the whole highway, including ground roads and elevated roads, settled in the earlier stage, but part of the elevated roads heaved since 2009. In the ground road part, ground road settled with the ground surface. However, in the elevated road part, the ground surface settled more significantly than the elevated road, with differential settlements ranging from 15 to 40 mm. The settlements of the elevated roads were invariably close to the total compression of soils below the pile tips and the elevated road heaved due to rebound of these soils.

Three-dimensional (3D) geological models are essential for geological analysis and mineral resource estimation. Although conventional on-site survey methods, such as boreholes, provide local engineering geological information for 3D geological modeling, accurately predicting strata in areas with sparse borehole data remains a challenge. This study proposes a 3D geological modeling method using the Transformer model under the conditions of sparse borehole data. First, a K-dimensional tree was used to identify boreholes adjacent to the target point, and a borehole context sequence was constructed using stratigraphic information from neighboring boreholes. Subsequently, the relationship between the target point and its adjacent borehole sequence was calculated using the multi-head attention mechanism of the Transformer model. Finally, trained Transformer encoders were used to predict the stratigraphic category of the target point, and the normalized information entropy was used to quantify uncertainty during the modeling process. Experimental results showed that the accuracy of the method was 0.86, outperforming the accuracy and uncertainty of a recurrent neural network. The root mean square error is smaller than the inverse distance weight and Kriging. Compared to other methods, the proposed method can more accurately describe the geometric shape and distribution of geological bodies and reveal the sedimentary laws of the study area.

Pumping artesian water is mandatory in deep excavation above a shallow confined aquifer with a high piezometric head. In this study, a comprehensive monitoring program is performed to investigate the responses of far-field ground and groundwater to the pumping of artesian water in deep excavations. The artesian water level was monitored by observation wells, and the pore pressures along the depth were monitored by a cluster of vibrating wire piezometers. The ground surface settlement and subsurface settlement were surveyed by level instruments and multiple position extensometers, respectively. These measured data are analyzed comprehensively and thoroughly. Moreover, a simplified prediction method of far-field groundwater drawdown and ground surface settlements induced by dewatering is proposed. Results indicate that the response of artesian water levels is closely related to the discharge and recharge flow. Because of the vertical drainage, the pore pressures in the lower portion of the aquitard above the confined aquifer decrease slightly during the dewatering process. However, the pore pressures in poorly conductive soils barely increase during the short dated recharge. The developments of the ground surface settlements are consistent with the artesian water level. The ground surface beyond 4H_e settled significantly after dewatering, and the influence zone of dewatering on ground surface settlements extended to 10H_e, where H_e represents the final excavation depth. The measured data reveal that dewatering in the confined aquifer resulted in the compression of all soil layers. The comparison between predicted results and the monitoring data confirms the practicability of the proposed method.

Land subsidence, a slow-onset geohazard, poses a severe threat to cities worldwide. However, the lack of quantification in terms of intensity, probability, and hazard zoning complicates the assessment and understanding of the land subsidence risk. In this study, we employ a weighted Bayesian model to explicitly present the spatial distribution of land subsidence probability and map hazard zoning in Shanghai. Two scenarios based on distinct aquifers are analyzed. Our findings reveal the following: (1) The cumulative land subsidence probability density functions in Shanghai follow a skewed distribution, primarily ranging between 0 and 50 mm, with a peak probability at 25 mm for the period 2017–2021. The proportions of cumulative subsidence above 100 mm and between 50 and 100 mm are significantly lower for 2017–2021 compared to those for 2012–2016, indicating a continuous slowdown in land subsidence in Shanghai. (2) Using the cumulative subsidence from 2017–2021 as a measure of posterior probability, the probability distribution of land subsidence under the first scenario ranges from 0.02 to 0.97. The very high probability areas are mainly located in the eastern peripheral regions of Shanghai and the peripheral areas of Chongming District. Under the second scenario, the probability ranges from 0.04 to 0.98, with high probability areas concentrated in the eastern coastal area of Pudong District and regions with intensive construction activity. (3) The Fit statistics for Scenario I and Scenario II are 67% and 70%, respectively, indicating a better fit for Scenario II. (4) High-, medium-, low-, and very low-hazard zones in Shanghai account for 14.2%, 48.7%, 23.6%, and 13.5% of the city, respectively. This work develops a method based on the weighted Bayesian model for assessing and zoning land subsidence hazards, providing a basis for land subsidence risk assessment in Shanghai.