Destructive earthquakes result in significant damage to a wide variety of buildings. The resulting damage data is crucial for evaluating the seismic resilience of buildings in the region and investigating urban resilience. Field damage data from 38 destructive earthquakes in Sichuan Province were collected, classified, and statistically analysed according to the criteria of the latest Chinese seismic intensity scale for evaluating building damage levels. Meanwhile, the construction features and seismic damage characteristics of these buildings were also examined. These results facilitated the development of a damage probability matrix (DPM) for various building typologies, such as raw-soil structures (RSSs), stone-wood structures (SWSs), brick-wood structures (BWSs), masonry structures (MSs), and reinforced concrete frame structures (RCFSs). The damage ratio was employed as the parameter for vulnerability assessment, and a comprehensive analysis was performed on the differences in damage levels among all buildings in various intensity zones and time frames. Furthermore, the DPMs were further refined by simulating additional data from high-intensity zones to more accurately represent the seismic resistance of existing buildings in multiple-intensity zones. Vulnerability prediction models were developed using the biphasic Hill model, which elucidates varying damage trends across different construction typologies. Finally, empirical fragility curves were established based on horizontal peak ground acceleration (PGA) as the damage indicator. This study is based on multiple seismic damage samples from various regions, accounting for the influence of earthquake age. The DPMs, representative of the regional characteristics of Sichuan Province, were developed for different building types. Furthermore, multidimensional vulnerability regression models and empirical fragility curves are established based on these DPMs. These models and curves provide a theoretical foundation for seismic disaster scenario simulations and the seismic capacity analysis of buildings within Sichuan Province.

Landslides pose significant risks to human life and infrastructure, particularly in mountainous regions like Inje, South Korea. This study aims to develop detailed landslide susceptibility maps (LSMs) using statistical (i.e., Frequency Ratio (FR), Logistic Regression (LR)) models and a hybrid integrated approach. These models incorporated various factors influencing landslides, including aspect, elevation, rainfall, slope, soil depth, slope length, and landform, derived from comprehensive geospatial datasets. The FR method assesses the likelihood of landslides based on historical occurrences relative to specific factor classes, while the LR method predicts landslide susceptibility through the statistical modeling of multiple predictor variables. The results from the FR, LR, and hybrid methods showed that the cumulative area covered by high and very high landslide susceptibility zones was 13.8%, 13.0%, and 14.28%, respectively. The results were validated using Receiver Operating Characteristic (ROC) curves and the Area Under the Curve (AUC), revealing AUC values of 0.83 for FR, 0.86 for LR, and 0.864 for the hybrid method, indicating high predictive accuracy. Subsequently, we used K-mean clustering algorithms on the hybrid LSI to identify the higher LSI cluster of the region. Furthermore, sensitivity analysis based on landslide density confirmed that all methods accurately identified high-risk areas. The resulting LSMs provide critical insights for land-use planning, infrastructure development, and disaster risk management, enhancing predictive accuracy and aiding in the prevention of future landslide damage.

Understanding slope stability is crucial for effective risk management and prevention of slides. Some deterministic approaches based on limit-equilibrium and numerical methods have been proposed for the assessment of the safety factor (SF) for a given soil slope. However, for risk analyses of slides of earth dams, a range of SFs is required due to uncertainties associated with soil strength properties as well as slope geometry. Recently, several studies have demonstrated the efficiency of artificial neural network (ANN) models in predicting the SF of natural and artificial slopes. Nevertheless, such techniques operate as black-box models, prioritizing predictive accuracy without suitable interpretability. Alternatively, multivariate polynomial regression (MVR) models offer a pragmatic interpretability strategy by combining the analysis of variance with a response surface methodology. This approach overcomes the difficulties associated with the interpretability of the black-box models, but results in limited accuracy when the relationship between independent and dependent variables is highly nonlinear. In this study, two models for a quick assessment of slope SF in earth dams are proposed considering the MVR and the ANN models. Initially, a synthetic dataset was generated considering different soil properties and slope geometries. Then, both models were evaluated and compared using unseen data. The results are also discussed from a geotechnical point of view, showing the impact of each input parameter on the assessment of the SF. Finally, the accuracy of both models was measured and compared using a real-case database. The obtained accuracy was 78% for the ANN model and 72% for the MVR one, demonstrating a great performance for both proposed models. The efficacy of the ANN model was also observed through its capacity to reduce false negatives (a stable prediction when it is not), resulting in a model more favorable to safety assessment.

Saltwater intrusion (SWI) exposed the significant risk to rice production in the tropical lowland delta, especially under the contact of climate change. This study have developed the economic loss functions for both direct and indirect losses caused by SWI after investigating several regression models (such as: Ordinary Least Squares (OLS), Fixed Effects Model (FEM), Random Effects Model (REM), and Feasible Generalized Least Squares (FGLS), based on the 85 questionaires colleted in the tropical rice fields located in Ho Chi Minh City (HCMC). Direct damages were estimated based on cultivated area, rice yield, and salinity levels; while indirect damages were included the costs of water pumping, soil improvement, and irrigation infrastructure construction. The results showed that rice yield decreases sharply when salinity exceeds the threshold level of 1.5 parts per thousand, and indirect costs account for 9% of total damages. The new finding of this study is integrating indirect factors (water pumping, soil improvement, and irrigation infrastructure construction) into the economic loss function, enabling the estimation of both direct and indirect damages cause by SWI; which is a critical tool for water related disasters prevention and management, or land use planning, or developing socio-economic strategies to ensure food security for the deltas strongly affected by SWI.

In complex physical systems, conventional differential equations fall short in capturing non-local and memory effects. Fractional differential equations (FDEs) effectively model long-range interactions with fewer parameters. However, deriving FDEs from physical principles remains a significant challenge. This study introduces a stepwise data-driven framework to discover explicit expressions of FDEs directly from data. The proposed framework combines deep neural networks for data reconstruction and automatic differentiation with Gauss-Jacobi quadrature for fractional derivative approximation, effectively handling singularities while achieving fast, high-precision computations across large temporal/spatial scales. To optimize both linear coefficients and the nonlinear fractional orders, we employ an alternating optimization approach that combines sparse regression with global optimization techniques. We validate the framework on various datasets, including synthetic anomalous diffusion data, experimental data on the creep behavior of frozen soils, and single-particle trajectories modeled by L & eacute;vy motion. Results demonstrate the framework's robustness in identifying FDE structures across diverse noise levels and its ability to capture integer-order dynamics, offering a flexible approach for modeling memory effects in complex systems.

Structural damage and foundation leakage are major concerns for earthen dams. To minimize seepage, cutoff walls are typically installed beneath the dam core to act as impermeable barriers. While concrete cutoff walls are widely used, their limited ductility and strength incompatibility with foundation soil present design challenges. Plastic concrete, a modified form of conventional concrete incorporating bentonite and pond ash, offers improved ductility and reduced brittleness, making it a suitable alternative. This study investigates the use of pond ash-based flowable fill as a replacement for normal concrete in plastic concrete cutoff walls. The unconfined compressive strength (UCS) of plastic concrete mixes was analyzed using four advanced regression machine learning algorithms: multivariate adaptive regression splines, extreme neural network (ENN), extreme gradient boosting (XGBoost), and gradient boosting machine (GBM). Several performance indices were used to evaluate model accuracy. The MARS model achieved the highest accuracy, with R2 = 0.990 for training and R2 = 0.963 for testing, followed by XGBoost, GBM, and ENN. SHAP analysis revealed that curing period has the most significant positive effect on UCS, followed by water and cement contents, while bentonite showed the least impact. Key properties were evaluated to determine an optimal mix design. This research enhances the understanding of CLSM-based plastic concrete and supports its application in cutoff walls by developing accurate UCS prediction models, contributing to the improved suitability and sustainability of dam foundation systems.

Landslides are recognized as major natural geological hazards in the mountainous region, and they are accountable for enormous human causalities, damage to properties, and environmental issues in the Teesta River basin, Sikkim, India. GIS approaches are widely used in landslide susceptibility mapping (LSM) that can help relevant authorities to mitigate landslide risk. The binary logistic regression is applied to estimate the landslide susceptibility zonation (LSZ) in the upper Teesta River basin areas. The landslide inventory data are subdivided into training data sets (70%) for applying algorithms in models and testing data sets (30%) for testing model accuracy. The LSZ mapping is designed after analyzing multicollinearity test of 14 landslide CFs and the result shows that the VIF value is less than 10, and TOL is greater than 0.1, respectively. There is no multicollinearity for the 14 conditioning landslides factors. The upper Teesta River basin is categorized into five groups: very low-to-very high landslide susceptibility zones. The results highlighted that most of the middle and southern parts of the study region are highly prone to landslides compared to the other parts. The susceptibility of landslide in the upper Teesta River basin areas validated by performing the Receiver Operating Characteristics (ROC) curve, which showed an 83% confidence level. The present research demonstrated landslide vulnerability circumstances for the Teesta River basin, Sikkim, an area prone to landslides, emphasizing the need for an effective mitigation and management roadmap.

Hazardous waste from metal processing industries increases heavy metal contamination in ecosystems, threatening environmental health and regional sustainability. This study suggests a resilient and human-centered environmental monitoring approach that incorporates machine learning and decision analytics to address these challenges in line with Industry 5.0's goals. By utilising a PRINCIPAL COMPONENT REGRESSION (PCR)-based predictive model, the approach addresses variability in environmental data, predicting levels of heavy metals like lead, zinc, nickel, arsenic, and cadmium, frequently beyond regulatory thresholds. The suggested PCR-based model outperforms conventional models by lowering mean absolute error (MAE) to 2.9339, mean absolute percentage error (MAPE) to 0.0358, and nearly the same mean square error (MSE). This study introduces a more interpretable and computationally efficient alternative to existing predictive models by introducing a novel integration of PCR with machine learning for environmental monitoring. By predicting and optimising environmental outcomes, validation against test datasets confirmed its ability to optimise impurity control. After process adjustments, the average concentrations of lead, nickel, and cadmium were reduced from 13.23 to 11.26 mg/L, 2.83 to 2.70 mg/L, and 2.15 to 1.88 mg/L, respectively. This research supports sustainability, resilience, and decisionmaking aligned with Industry 5.0, offering scalable solutions and insights for global industries.HighlightsChemical plants' environmental risk is evaluated using a machine learning algorithmFor better monitoring, the PCR method forecasts process variables and interactionsIt identifies the key factors that affect the environmental risks in soil and waterAs a result, the local ecosystem's levels of toxic metals have notably decreasedInsights for managing environmental risks aligned with Industry 5.0 principles

In the processes of seismic design of underground structures, selecting a reasonable input ground motion is very important, which can cause severe damage to underground structures. To quantitatively evaluate the seismic damage potential of ground motions on multi-storey underground structures and solve the problem that single intensity measures are inadequate in accurately indicating the seismic damage potential of ground motions, this paper taking the input ground motions in the seismic design of underground structures as the research object, and constructing some composite intensity measures that can effectively characterize the damage potential of input ground motion. Firstly, considering that the underground structures in different characteristic period-type sites will exhibit different seismic responses under the same excitation, the soil-structure system is divided into four-period bands. Then, four representative periods are selected from four-period bands respectively, and the corresponding four soil-structure system numerical models are established. Subsequently, 40 ground motions are selected for elastoplastic numerical analysis, and the results of the numerical analysis were used as sample data to construct composite intensity measures corresponding to soil-structure systems in each period band using the partial least squares regression method. Finally, 100 additional ground motions were used to verify the correlation between the composite intensity measures and the seismic damage of underground structures. The results show that the correlation coefficient between the composite intensity measures and the seismic damage of multi-storey underground structures is better than those of commonly used single intensity measures.

The number of studies concerning the shear strength of resedimented alluvial soils is extremely limited compared to the studies conducted on fine-grained marine sediments, since alluvial soils are generally tested in remolded or reconstituted state especially in the studies investigating their liquefaction potential. In this study, estimation models were developed to predict cohesion (c) and internal friction angle (phi) parameters of a fine-grained alluvial soil using resedimented samples. A total of 60 undisturbed soil samples were obtained from Bafra district of Samsun province (Turkiye) by core drilling. A cone penetration test with pore water pressure measurement (CPTu) was also carried out alongside each borehole to determine the over-consolidation ratios of the samples. Physical-index property determinations and triaxial tests were conducted on the undisturbed samples. 20 sample sets were created with known physical, index, and strength characteristics. The samples are classified as CH, CL, MH, and ML according to the Unified Soil Classification System, with liquid and plastic limits ranging from 31.6-75% and 19.3 to 33.6% respectively. The c and phi values of the samples varied from 4.1 to 46.1 kPa and 26 to 35 degrees respectively. The samples were then resedimented in the laboratory under conditions reflecting their original in-situ properties, and triaxial tests were repeated. The c and phi values of the resedimented samples ranged from 5.3 to 24.5 kPa and 28 to 32 degrees respectively. The results indicate that the c values of the resedimented samples are generally lower than those of the undisturbed samples, whereas upper and lower bounds for phi values are similar. Multivariate regression analyses (MVR) were utilized to develop estimation models for predicting c and phi using strength and physical properties of 20 soil samples as independent variables. Three estimation models with R-2 values varying between 0.723 and 0.797 were proposed for c and phi which are statistically significant for p <= 0.05. Using artificial neural networks (ANN), the estimation models developed by MVR were replicated to validate the models. ANN yielded very similar results to the MVR, where the R-2 values for the correlations between c and phi values predicted by both methods varied from 0.852 to 0.955. The results indicate that c and phi values of undisturbed samples can be estimated with acceptable accuracy by determining basic physical and index properties of the disturbed samples and shear strength parameters of the resedimented samples. This approach, which enables the reuse of disturbed soil samples, can be used when undisturbed soil samples cannot be obtained from the field due to economic, logistical, or other reasons. Further research on the shear strength parameters of resedimented alluvial soils is needed to validate the estimation models developed in this study and enhance their applicability to a wider range of alluvial soils.