In engineering geology and geotechnical engineering , it is well recognized that subsurface soils/rocks are natural materials and exhibit variability in stratigraphy due to the complex geological formation processes they have undergone. Knowledge of subsurface soil stratigraphy is of great importance to geotechnical engineers . However, accurate and reliable interpretation of subsurface soil stratigraphy is challenging due to the limited number of site investigation boreholes available at the site and the highly heterogeneous properties of soil stratigraphy (e.g., interbedded or non-ordered layers). This paper proposes an improved data-driven machine learning framework boosted with the neighborhood aggregation technique for modelling three-dimensional (3D) subsurface soil stratigraphy in a more general and robust manner. Neighborhood aggregation, a technique often adopted in graph network learning, is integrated into this framework to regulate and improve the prediction results of classical machine learning models. The proposed framework is then cross-validated using 165 real site investigation boreholes for four selected machine learning models respectively. Cross-validation results suggest that the eXtreme Gradient Boosting (XGBoost) and Random Forest (RF) models are more suitable for the task of soil stratigraphy prediction than Artificial Neural Network (ANN) and Support Vector Machine (SVM). Particularly, the XGBoost and RF are also amenable to neighborhood aggregation and can yield around 5% improvement in terms of average borehole prediction accuracy after introducing neighborhood aggregation. The improved machine learning framework allows for explicit 1D to 3D geological modelling , uncertainty quantification , and convenient visualization. The proposed framework facilitates digital transformation of geological and geotechnical site investigation . • Neighborhood aggregation is introduced into machine learning for robust 3D soil stratigraphy modelling. • Performances of different machine learning algorithms are compared. • The proposed framework is validated using a borehole database in Singapore. • The new framework enables explicit 1D to 3D stratigraphy modelling, uncertainty quantification, and visualization.

Determining the location and boundary of underground obstructions and/or anomalies is a common problem and often a great challenge for tunneling and underground construction. In this study, geotechnical investigations (penetration tests and borehole drilling/sampling) and geophysical investigations (surface wave method and cross-hole seismic method) were conducted with the aim of identifying the location and boundary of rock obstructions in Changi East, Singapore. The surface wave method is frequently used in the sites with lateral homogeneity in previous studies, but its application in the sites with rock obstructions is rare. The experimental results of this study indicate that the surface wave method is also able to determine the upper surface of rock obstructions, but difficult to identify the lateral and bottom boundaries of rock obstructions. To improve the precision of detection, the full waveform inversion (FWI) method was used to process the data from the cross-hole seismic survey. The results indicate that the inversion precision of P-wave is higher than that of S-wave. The horizontal and vertical ranges of rock obstructions in the P-wave inversion results are 14–26 m and 7.5–11.0 m respectively, roughly consistent with the results of penetration tests (about 15–25 m) and borehole log (8.85–10.80 m). This result proves that the sequential application of first-arrival time analysis and FWI can effectively delineate the boundary of rock obstructions. Finally, the results of various detection methods were analyzed and compared in this study. Considering the advantages of various methods, we propose a cost-effective and high-precision workflow containing both geotechnical and geophysical investigations.

This study evaluates the seismic impact of the Bukit Timah granite massif in Singapore, emphasizing the role of granite weathering in local site effects. A 2.5D geomodel was constructed using geophysical data, boreholes, and geological maps, from which five 2D cross-sections were extracted for dynamic numerical modeling. Simulations were performed to assess resonance frequencies, amplification patterns, peak ground acceleration (PGA), and shear wave velocity (Vs). The dominant frequency ranged between 1.8 and 3.8 Hz, with amplification factors up to 4.5, consistent with HVSR data. The close agreement between HVSR and modeling results highlights the influence of weathered granite layers on site response. A 2D Vs model was derived from the inversion of HVSR and synthetic curves. The findings provide detailed insights into lithological and topographic effects, especially lateral heterogeneities in weathering. This study demonstrates the reliability of numerical modeling for microzonation in granitic terrains and supports its use in enhancing seismic risk assessments and urban planning in weathered rock regions.

Cement-treated ground often possesses significant spatial variation in strength and stiffness. Studies on the detection and stiffness measurement of weak zones in cement-treated ground using seismic geophysical methods remain limited to date. The work examines the feasibility of using seismic travel time tomography to detect and measure the stiffness of weak zones in cement-treated ground, through 1-g modelling of a cross-hole setup with weak zones of prescribed sizes and stiffnesses. Using bender elements as transmitters and receivers, shear wave velocity profiles across the weak zones are mapped. Sizes, locations, and stiffnesses of the weak inclusions are also inferred from shear wave travel times using GeoTomCG. The results indicate that although the sizes and locations of weak zones can be reliably detected using first-arrival time-picking, the stiffnesses is significantly over-estimated. The latter is due to the arrival of the diffracted waves around the weak zone masking the arrival of the transmitted waves. A modified time-picking method, using the wavelet transform, is developed and is shown to give more reliable stiffness measurements of the weak zone, but not the shape. Combining direct and wavelet time-picking allows the sizes, shapes, locations, and stiffnesses of the weak zones to be more reliably resolved.

The field measurement of shear wave velocity (Vs) is essential for geotechnical design practices, as it directly provides the initial tangent shear modulus at very small strain levels (γs < 10-6) in geo-materials. The small strain shear stiffness is a fundamental soil property for assessing dynamic loading responses, ground vibrations, and static deformation problems related to shallow and deep foundations. In addition, the Vs is one of the critical elements in evaluating seismic ground hazards such as site amplification and liquefaction potential. Various field and laboratory geotechnical site investigation programs in Kazakhstan have been conducted to understand basic soil behavior. However, in-situ geophysical seismic surveys such as surface reflection and refraction tests and down-hole and cross-hole tests were generally not included in the site investigation program in Kazakhstan, and a few limited seismic surveys have been carried out for specific projects. In most prior construction projects, the small strain shear stiffness was assessed by limited data using general empirical correlations from other in-situ measurements or selective laboratory testing programs that may result in significant uncertainties. In this study, in-situ dynamic soil characteristics of loam soils using active MASW (multi-channel analysis of surface waves) testing are evaluated to obtain comprehensive insights for geotechnical boundary value problems. The resulting Vs profiles are in good agreement with a-priory known geotechnical information (e.g., borehole logs) of sites. Thus, to minimize potential uncertainties of dynamic soil properties estimation via in-situ tests, MASW methods are suggested for construction works in Kazakhstan.

The South and Southeast Kazakhstan regions exhibit notable seismicity due to intricate tectonic interactions, albeit experiencing infrequent catastrophic earthquakes. Proximate to the convergence of the Eurasian and Indian plates, this region witnesses frequent seismic activity, particularly in cities like Almaty near mountainous terrain. Given the significant seismic activity, comprehensive site characterization is imperative. Traditionally, evaluating dynamic soil properties relies on conventional borehole logging techniques. However, the emergence of the multi-channel analysis of surface waves (MASW) offers advantages in cost, time efficiency, and non-invasiveness. Despite its benefits, MASW remains underutilized in Kazakhstan and is absent from local building codes, unlike neighboring CIS countries. This study aims to demonstrate the applicability of the MASW method for site characterization in Kazakhstan's seismic regions. Through extensive work, shear wave velocity (ð ð ) values were estimated and compared with reference data obtained from seismic refraction and dilatometer testing. The results showed significant agreement, highlighting the suitability and effectiveness of the MASW method in Kazakhstan.

Compacted cement-stabilized soils are important geomaterials for construction projects involving backfilling, embankment construction, and ground improvement. In general, unconfined compressive strength (UCS) serves as an important design parameter for such types of materials, but it remains difficult to characterize the UCS owing to complex interactions of various influencing factors. This study aims to obtain practical insights into the effects of water and fines content on the UCS of cement-stabilized clayey sand, which is commonly encountered in reclamation projects. The experimental results show that the water content and fines content affect the strength in a combined manner. The strength decreases monotonically with increasing water content for clean sand, while it first increases and then decreases with the increase in water content for cemented clayey sands. In addition, the effects of fines content on the UCS are also affected by water content. The threshold water content and threshold fines content were identified and characterized. The maximum strength corresponding to these threshold parameters was also characterized. Finally, a new sequential framework to predict the UCS of cement-stabilized clayey sands is proposed.

In engineering geology and geotechnical engineering , it is well recognized that subsurface soils/rocks are natural materials and exhibit variability in stratigraphy due to the complex geological formation processes they have undergone. Knowledge of subsurface soil stratigraphy is of great importance to geotechnical engineers . However, accurate and reliable interpretation of subsurface soil stratigraphy is challenging due to the limited number of site investigation boreholes available at the site and the highly heterogeneous properties of soil stratigraphy (e.g., interbedded or non-ordered layers). This paper proposes an improved data-driven machine learning framework boosted with the neighborhood aggregation technique for modelling three-dimensional (3D) subsurface soil stratigraphy in a more general and robust manner. Neighborhood aggregation, a technique often adopted in graph network learning, is integrated into this framework to regulate and improve the prediction results of classical machine learning models. The proposed framework is then cross-validated using 165 real site investigation boreholes for four selected machine learning models respectively. Cross-validation results suggest that the eXtreme Gradient Boosting (XGBoost) and Random Forest (RF) models are more suitable for the task of soil stratigraphy prediction than Artificial Neural Network (ANN) and Support Vector Machine (SVM). Particularly, the XGBoost and RF are also amenable to neighborhood aggregation and can yield around 5% improvement in terms of average borehole prediction accuracy after introducing neighborhood aggregation. The improved machine learning framework allows for explicit 1D to 3D geological modelling , uncertainty quantification , and convenient visualization. The proposed framework facilitates digital transformation of geological and geotechnical site investigation . • Neighborhood aggregation is introduced into machine learning for robust 3D soil stratigraphy modelling. • Performances of different machine learning algorithms are compared. • The proposed framework is validated using a borehole database in Singapore. • The new framework enables explicit 1D to 3D stratigraphy modelling, uncertainty quantification, and visualization.