The southern regions of China are rich in ion-adsorbed rare earth mineral resources, primarily distributed in ecologically fragile red soil hilly areas. Recent decades of mining activities have caused severe environmental damage, exacerbating ecological security (ES) risks due to the inherent fragility of the red soil hilly terrain. However, the mechanisms through which multiple interacting factors influence the ES of rare earth mining areas (REMA) remain unclear, and an effective methodological framework to evaluate these interactions dynamically is still lacking. To address these challenges, this study develops an innovative dynamic ES evaluation and earlywarning simulation framework, integrating Variable Weight (VW) theory and the Bayesian Network (BN) model. This framework enhances cross-stage comparability and adapts to evolving ecological conditions while leveraging the BN model's diagnostic inference capabilities for precise ES predictions. A case study was conducted in the Lingbei REMA. The main findings of the study are as follows: (1) From 2000 to 2020, the overall ES of the mining area exhibited a dynamic trend of deterioration, followed by improvement, and ultimately stabilization. (2) Scenario S27 (high vegetation health status and high per capita green space coverage) significantly reduces the probability of the ES reaching the extreme warning level. (3) The evaluation and simulation framework developed in this study provides a more accurate representation of the ES level distribution and its variations, with probabilistic predictions of ES demonstrating high accuracy. This study is of great significance for improving regional ES, supporting the optimization of ecological restoration strategies under multi-objective scenarios, and promoting the coordinated development of nature and resource utilization.

Climate change impacts water supply dynamics in the Upper Rio Grande (URG) watersheds of the US Southwest, where declining snowpack and altered snowmelt patterns have been observed. While temperature and precipitation effects on streamflow often receive the primary focus, other hydroclimate variables may provide more specific insight into runoff processes, especially at regional scales and in mountainous terrain where snowpack is a dominant water storage. The study addresses the gap by examining the mechanisms of generating streamflow through multi-modal inferences, coupling the Bayesian Information Criterion (BIC) and Bayesian Model Averaging (BMA) techniques. We identified significant streamflow predictors, exploring their relative influences over time and space across the URG watersheds. Additionally, the study compared the BIC-BMA-based regression model with Random Forest Regression (RFR), an ensemble Machine Learning (RFML) model, and validated them against unseen data. The study analyzed seasonal and long-term changes in streamflow generation mechanisms and identified emergent variables that influence streamflow. Moreover, monthly time series simulations assessed the overall prediction accuracy of the models. We evaluated the significance of the predictor variables in the proposed model and used the Gini feature importance within RFML to understand better the factors driving the influences. Results revealed that the hydroclimate drivers of streamflow exhibited temporal and spatial variability with significant lag effects. The findings also highlighted the diminishing influence of snow parameters (i. e., snow cover, snow depth, snow albedo) on streamflow while increasing soil moisture influence, particularly in downstream areas moving towards upstream or elevated watersheds. The evolving dynamics of snowmelt-runoff hydrology in this mountainous environment suggest a potential shift in streamflow generation pathways. The study contributes to the broader effort to elucidate the complex interplay between hydroclimate variables and streamflow dynamics, aiding in informed water resource management decisions.

Climate change will create significant challenges to agriculture. The effects on livestock productivity and crop production are highly dependent on weather conditions with consequences for food security. If agriculture is to remain a viable industry and to maintain future food security, the adaptations and the ideal timeframes for their implementation to mitigate against climate change impacts will be essential knowledge. This study aims to show how farms will be affected and will need to adapt to climate change, based on a holistic examination of the entire farming process. A modified Bayesian belief network (BBN) was used to investigate climate change impacts on livestock, crops, soil, water use, disease, and pesticide use through the use of 48 indicators (comprising climate, agricultural, and environmental). The seasonal impact of climate change on all aspects of farming was investigated for three different climate forcing scenarios (RCPs 2.6, 4.5, and 8.5) for four timeframes (2030, 2050, 2080, and 2099). The results suggest that heat stress and disease in both livestock and crops will require adaptations (e.g., shelter infrastructure being built, new crops, or cultivators grown). Pest intensity is expected to rise, leading to increased pesticide use and greater damage to crops and livestock. Higher temperatures will likely cause increased drought and irrigation needs, while increasing rain intensity might lead to winter flooding. Soil quality maintenance will rely increasingly on fertilisers, with significant decreases in quality if unsustainable. Crop yield will be dependent on new crops or cultivators that can cope with a changing climate being successful and market access; failure to do so could lead to substantial decrease, in food security. Impacts are more significant from 2080 onwards, with the severity of impacts dependent on season.

Multi-source precipitation products (MSPs) are critical for hydrologic modeling, but their spatial and temporal heterogeneity and uncertainty present challenges to simulation accuracy that need to be addressed urgently. This study assessed the impact of different precipitation data sources on hydrologic modeling in an arid basin. There were seven precipitation products and meteorological station interpolated data that were used to drive the hydrological model, and we evaluated their performance by fusing the six precipitation products through the dynamic bayesian averaging algorithm. Ultimately, the runoff simulation uncertainty was quantified based on the DREAM algorithm, and the information transfer entropy was used to quantify the differences in hydrologic simulation processes driven by different precipitation data. The results showed that CMFD and ERA5 weights were higher, and the DBMA fused precipitation annual mean value was about 309.83 mm with good simulation accuracy (RMSE of 1.46 and R-2 of 0.75). The simulation was satisfactory (NSE >0.80) after parameter calibration and data assimilation for all driving data, with CHIRPS and TRMM performed better in the common mode, and HRLT and CMFD performed excellently in the glacier mode. The DREAM algorithm indicated less uncertainty for DBMA, CHIRPS and HRLT data. The entropy of information transfer revealed that precipitation occupied a significant position in information transfer, especially affecting evapotranspiration and surface soil moisture. CMFD and TPS CMADS were highest in snow water equivalent information entropy, and CHIRPS and TPS CMADS were highest in evapotranspiration information entropy. CDR, CHIRPS, ERA5-Land and IDW STATION had the highest snow water equivalent information entropy, DBMA and CMORPH had the highest runoff information entropy, CHIRPS and TRMM had the highest soil moisture information entropy, whereas ERA5, HRLT, and TPS CMADS had the highest evapotranspiration information entropy in glacial mode. This study reveals significant differences between different precipitation data sources in hydrological modeling of arid basin, which is an important reference for future water resources management and climate change adaptation strategies.

The constitutive model is essential for predicting the deformation and stability of rock-soil mass. The estimation of constitutive model parameters is a necessary and important task for the reliable characterization of mechanical behaviors. However, constitutive model parameters cannot be evaluated accurately with a limited amount of test data, resulting in uncertainty in the prediction of stress-strain curves. This paper proposes a Bayesian analysis framework to address this issue. It combines the Bayesian updating with the structural reliability and adaptive conditional sampling methods to assess the equation parameter of constitutive models. Based on the triaxial and ring shear tests on shear zone soils from the Huangtupo landslide, a statistical damage constitutive model and a critical state hypoplastic constitutive model were used to demonstrate the effectiveness of the proposed framework. Moreover, the parameter uncertainty effects of the damage constitutive model on landslide stability were investigated. Results show that reasonable assessments of the constitutive model parameter can be well realized. The variability of stress-strain curves is strongly related to the model prediction performance. The estimation uncertainty of constitutive model parameters should not be ignored for the landslide stability calculation. Our study provides a reference for uncertainty analysis and parameter assessment of the constitutive model.

Due to their distinct geotechnical and structural features, soft rock tunnels pose serious issues because of their seismic sensitivity. These tunnels, often constructed in formations with lower shear strength and higher deformability, are particularly susceptible to damage during earthquakes. Fragility curves, which graphically represent the probability that a structure may sustain damage up to or beyond a particular threshold as a function of seismic intensity, are essential tools for evaluating the seismic resilience of these infrastructures. This research looks closely at the use of fragility curves to assess the seismic vulnerability of soft rock tunnels. Exploring the fundamental concepts and methodologies involved in constructing fragility curves, including seismic hazard analysis, structural modeling, damage state definition, data collection and statistical analysis is looked at first. The review highlighted the integration of soft rock characteristics such as strength and deformation properties into the fragility assessment process. Key developments in the topic are covered such as how machine learning and Bayesian inference might improve the precision and usefulness of fragility curves. The paper identified key findings such as the high sensitivity of fragility curves to geotechnical properties and seismic intensity levels and emphasized the importance of accurate data collection and model calibration. Important gaps in seismic risk evaluations are filled by integrating cutting-edge methodologies, such as Bayesian inference and real-time machine learning models that clarify the seismic behaviour of soft rock tunnels in the real world. For the purpose of strengthening earthquake-resistant infrastructure in earthquake-prone areas, engineers, scholars and policymakers are given practical insights.

Soft clays exhibit significant challenges in geotechnical engineering due to their low permeability, high compressibility, and susceptibility to settlement under applied loads. These geological factors pose unique difficulties in predicting long-term settlement accurately and efficiently, particularly through Class C prediction methods that involve iterative processes with complex numerical models. To address these challenges, this study presents an efficient approach for Class C prediction of long-term settlement in soft clays. This approach integrates Bayesian updating with structural reliability methods (BUS) and the general simplified Hypothesis B method which is a semi-analytical method based on one-dimensional elastic visco-plastic (1D EVP) model. Unlike previous research that used Response Surface Model (RSM) with polynomial function for consolidation evaluation, the proposed approach enhances both accuracy and performance consistency under varying conditions. Additionally, by leveraging analytical solutions instead of iterative small-time steps required by Finite Element Method (FEM) or Finite Difference Method (FDM), the computational efficiency is also enhanced. The effectiveness of the proposed approach is demonstrated through its application to an embankment improved with prefabricated vertical drain (PVD) in Ballina, New South Wales, Australia. Comparative analyses demonstrate that the predicted settlements from this study, using only the monitoring settlement data collected prior to the 76th day of the project, align closely with the results from established RSM and FEM-based Bayesian back analysis approaches. The obtained results also indicate that the predicted settlements, based on 76 days of monitoring data, closely match field measurements at various depths, whether relying solely on settlement data or integrating additional pore water pressure data. For the Ballina embankment, over 40,000 consolidation analyses required for a single BUS simulation can be completed within 10 h using the general simplified Hypothesis B method, compared to months it might take with FEM or FDM approaches. This makes the proposed approach a practical tool for geotechnical engineers, enabling reliable settlement predictions early in the project timeline while maintaining low computational costs.

In uncoupled consolidation analysis, settlement and pore water pressure are solved independently, whereas in coupled analysis, they are solved simultaneously to ensure continuity (i.e., the volume change in soil due to compression must equal the water volume change caused by dissipation). This study investigates the coupling effects of soil deformation and pore water pressure dissipation in the back analysis of soft soil settlements. It further evaluates the suitability of both coupled and uncoupled constitutive models with different types of monitoring data, providing practical guidance for selecting consolidation models and achieving reliable long-term predictions. The one-dimensional governing equations for soft soil consolidation, incorporating prefabricated vertical drains and creep deformation, are first reviewed. A case study of a trial embankment in Ballina, New South Wales, Australia, is then used to demonstrate the impact of coupling effects and monitoring data on settlement predictions. The results show that considering coupling effects not only improves long-term settlement predictions but also reduces uncertainties in the updated soil parameters, especially when both settlement and pore water pressure data are used.

Accurate estimation of soil properties is crucial for reliability-based design in engineering practices. Conventional empirical equations and prevalent data-driven models rarely consider uncertainty quantification in both measurement and modelling processes. This study tailors three uncertainty quantification methods including Bayesian learning, Markov chain Monte Carlo and ensemble learning into data-driven modelling, in which support vector regression is selected as the baseline algorithm. The compression index of clay is adopted as an example for model training and testing. In this context, Bayesian learning and Markov chain quantify uncertainty by considering the distribution of function and hyper-parameters, respectively, while different sampled data are employed to explore model uncertainty. These models are evaluated in terms of accuracy, reliability and cost-effectiveness and also compared with Gaussian process regression, etc. The results reveal that based on built-in structural risk minimization, sparse solution and uncertainty quantification, developed models can capture more accurate and reliable correlations from actual measured data over other methods. Their practicability and generalization ability are also verified on a new creep index database. The proposed probabilistic methods are also compiled into a user-friendly platform, showing a significant potential to enrich the data-driven modelling framework and be applied in other geotechnical properties.

Tunneling operations in modern construction demand meticulous evaluation of their impact on nearby structures. A primary concern is the potential for soil subsidence, which could damage adjacent buildings. Complicating matters is the challenge of accurately modeling such settlement and the consequent damage, a critical process for informed decision-making during construction projects. By employing Bayesian updating, we refine our models by acquiring posterior distributions for key parameters. We put forth an analytical method for profiling ground settlement and follow this by calculating the strain on an equivalent beam, which serves as a proxy for building damage. This results in a distribution of strain values that allows for an assessment of how varying certain length parameters affects the probability of maintaining a safe distance between the tunneling activities and the surrounding buildings. With this probabilistic approach, one can propose a recommended safety distance as a guideline for construction practices.