Stochastic Environmental Research and Risk Assessment最新文献

The escalating climate instability and extreme weather events significantly jeopardize food security. The study assessed the impact of long-term climatic variables and extreme weather events on soybean and wheat yields in rainfed central India. To address inherent spatial variability, the study area was divided into homogeneous zones based on rainfall and soil parameters. Crop yields were correlated with a comprehensive set of driving variables at seasonal and monthly scales within each zone. Machine learning algorithms, including Random Forest Regression (RFR) and Neural Networks (NN), were employed to analyze crop yield anomalies caused by climate and weather extremes. The Sobol’ index was utilized for global sensitivity analysis to identify key parameters. Results showed significant negative correlations between thermo-meteorological parameters and yields of both monsoon soybean and winter wheat across multiple districts. Soybean yield exhibited a notable positive correlation with hydro-meteorological parameters, while wheat yield displayed a significant positive correlation with cold temperature extremes. RFR and NN demonstrated similar performance, with Root Mean Square Error (RMSE) values ranging from 0.27 to 0.39 t/ha for soybean and 0.4 to 0.6 t/ha for wheat. The Sobol’ index highlighted the high sensitivity of soybean yield to rainfall and rainy days during July and August, corresponding to the crop development and flowering stages. In contrast, wheat yield was primarily influenced by temperature extremes, particularly cold nights and hot days during the reproductive-maturity stage. These crop- and growth-stage-specific analyses of meteorological parameters are essential for devising effective strategies to adapt and mitigate climate emergencies.

Dam failure may lead to significant casualties among downstream residents. Therefore, it is crucial to study a reliable method to quantitatively assess the loss of life (LOL) caused by dam failure for emergency response to dam failure incidents. Based on a statistical analysis of typical dam failure accidents in China and the research on the formation mechanism of LOL, the study quantified the factors influencing LOL using fuzzy theory and constructed a quantitative database for the LOL. Then, it proposed an innovative algorithm integrating the grey wolf optimization (GWO) algorithm and the random forest (RF) model. Finally, a data-driven assessment model for the LOL caused by dam failure was developed by combining the gray correlation analysis of the factors. The performance of the GWO-RF model was validated using a dataset of the LOL caused. The proposed model was used to assess the LOL in typical dam failure events. The results indicate that the model has higher accuracy, with an average absolute error of approximately 945 persons, significantly lower than 2529 persons in the Graham method. Thus, it can effectively estimate the LOL caused by dam failure. This study developed a novel method for quantitatively assessing the LOL caused by dam failure, which could also serve as a reference for modeling disaster consequences in other fields.

Accurate and efficient identification of pollution sources is a key process that assists in the treatment of water pollution incidents. The ensemble Kalman filter (EnKF) has been proven to be an effective approach for identifying pollution source parameters (e.g., source location, release time, and mass released). In this paper, a method involving multiple observations of reconstruction (MOR) is proposed for reconstructing multidimensional state vectors for assimilation based on pollutant concentration monitoring techniques. The newly reconstructed state variables have dimensionless characteristics that decouple the source mass from the parameter group to be identified before assimilation is performed. This approach can mitigate the interference of assimilation caused by nonmain source parameters. As a result, the pollution sources and material dispersion coefficients can be simultaneously identified at limited observation sites. Then, a set of synthetic numerical examples with 7 scenarios is assembled to investigate and compare the unique characteristics of the derived state variables during assimilation. A laboratory experiment for unknown parameter identification based on monitoring the chemical oxygen demand (COD) concentration is carried out in an annular flume to verify the applicability of the method in real events. The results show that the EnKF combined with the MOR method based on the decoupling pattern performs well in identifying pollution sources and dispersion coefficients simultaneously. The method can still perform excellently in identifying parameters in practice when some data in the observation sequences are lost, with relative errors of pollution source parameters being controlled within 4%. The relative errors of the identified transverse and longitudinal dispersion coefficients are 39% and 12%, respectively. Overall, by evaluating the original data, reconstructing the dataset, and combining it with the EnKF method, it is proven that the MOR–EnKF method is an effective measure for identifying high-dimensional unknown parameter groups.

The research investigates the significance of identifying structure discontinuities, such as cracks, in concrete dams to ensure dam safety and stability. A novel automatic image classification method is developed, employing Deep Dense Transfer Learning (DDTL) with pre-trained models, including EfficientNetB1, ResNet50, and a hybrid model to identify the detection of cracks in sub-surfaces at pillow dams in Sichuan province, China. The developed model was trained, validated, and tested, with the Hybrid model demonstrating superior performance. The results showed that the DDTL models had high classification accuracies, surpassing Convolutional identification techniques for sub-surface cracks. Consequently, this study suggests that automatic image classification techniques can effectively identify and localize structural defects in concrete dams. This is an innovative approach to predicting normal borehole images and crack recognition using CCTV borehole images.

Non-stationary sediment load analysis is critical for river engineering design and water resource management. Traditional sediment load frequency analysis methods usually assume stationarity, which can lead to inconsistent results in a changing environment because they cannot account for factors such as time variations. Here, we use generalized additive models for location, scale and shape (GAMLSS) to establish non-stationary models with time, precipitation and streamflow as covariates (named Model 1 and Model 2, respectively), and compare their fitting effects with stationary models (parameters unchanged: Model 0). In this study, the sediment load of the Jinsha River Basin in southwest China was analyzed. Outcomes indicate that: (1) the research area's sediment load decreased significantly, with a significant change point in 2002 (p < 0.1); (2) the goodness of fit indices (global fitting deviation: GD, AIC criterion and SBC criterion) based on Model 2 are smaller than the values of the other two models. The other two models' sediment load quantile design values are within Model 2's range. (3) Compared with Model1, precipitation and streamflow as covariates in Model 2 are more able to capture the non-stationary features of sediment load frequency. Furthermore, Model 2 can more accurately forecast future changes in sediment load when external physical factors are considered. The findings of this research can serve as a scientific foundation for decision makers to carry out water conservancy planning and design and river management and development.

Soil depth plays a pivotal role in determining hillslope stability, understanding hydrogeology, promoting optimal vegetation growth, and comprehensively elucidating soil erosion dynamics. In this study, two robust artificial intelligence methodologies, quantile regression forest (QRF) and deep neural network (DNN), were employed to predict spatial variations in soil depth across a digital terrain. Particularly during periods of intense rainfall, shallow landslides pose recurrent threats to human safety and property integrity. Thus, the identification of potential landslide-prone regions becomes imperative for mitigating associated risks. During slope stability analyses, soil depth assumes significance; nonetheless, data regarding soil depth from areas prone to landslides are rarely obtained. The main objective of this study is to explore the impact of incorporating soil depth spatial distributions on the predictive capabilities of shallow landslide model within a given terrain. By leveraging two distinct spatial soil depth distributions, a comprehensive analysis of slope stability analysis was conducted. The significance of soil depth spatial distribution, particularly when employing DNN-generated data, is underscored in refining predictions and preventing overestimations of landslide-prone or stable regions. Notably, integration of DNN-derived soil depth data into the infinite slope model yielded a marked enhancement in the accuracy of factor of safety (FS) distributions, achieving an impressive 86.9% accuracy rate while QRF-derived FS has shown 74.7% accuracy. This analytical approach, while straightforward, offers a powerful tool for evaluating slope instability and forecasting shallow landslides, thereby facilitating proactive mitigation measures.

Gaussian processes are popular in spatial statistics, data mining and machine learning because of their versatility in quantifying spatial variability and in propagating uncertainty. Although there has been a prolific research activity about Gaussian processes over Euclidean domains, only recently this research has extended to non-Euclidean manifolds. This paper digs into vector-valued Gaussian processes defined over the product of a hypersphere and a Euclidean space of arbitrary dimension, which are of interest in various disciplines of the natural sciences and engineering. Under mild regularity conditions, we establish a surprising one-to-one correspondence between matrix-valued kernels associated with vector Gaussian processes over the product space, and what we term partial ultraspherical and Fourier transforms that are taken over either the sphere or the Euclidean subspace. The properties of our approach are illustrated in terms of new parametric classes of matrix-valued kernels for product spaces of a hypersphere crossed with a Euclidean space. We also provide two algorithms that allow for fast simulation of approximately Gaussian (in the sense of the central limit theorem) processes in such product spaces.

Slope failures possess destructive power that can cause significant damage to both life and infrastructure. Monitoring slopes prone to instabilities is therefore critical in mitigating the risk posed by their failure. The purpose of slope monitoring is to detect precursory signs of stability issues, such as changes in the rate of displacement with which a slope is deforming. This information can then be used to predict the timing or probability of an imminent failure in order to provide an early warning. Most approaches to predicting slope failures, such as the inverse velocity method, focus on predicting the timing of a potential failure. However, such approaches are deterministic and require some subjective analysis of displacement monitoring data to generate reliable timing predictions. In this study, a more objective, probabilistic-learning algorithm is proposed to detect and characterise the risk of a slope failure, based on spectral analysis of serially correlated displacement time-series data. The algorithm is applied to satellite-based interferometric synthetic radar (InSAR) displacement time-series data to retrospectively analyse the risk of the 2019 Brumadinho tailings dam collapse in Brazil. Two potential risk milestones are identified and signs of a definitive but emergent risk (27 February 2018-26 August 2018) and imminent risk of collapse of the tailings dam (27 June 2018-24 December 2018) are detected by the algorithm as the empirical points of inflection and maximum on a risk trajectory, respectively. Importantly, this precursory indication of risk of failure is detected as early as at least five months prior to the dam collapse on 25 January 2019. The results of this study demonstrate that the combination of spectral methods and second order statistical properties of InSAR displacement time-series data can reveal signs of a transition into an unstable deformation regime, and that this algorithm can provide sufficient early-warning that could help mitigate catastrophic slope failures.

The identification of groundwater contamination sources (IGCSs) is an important requirement for the remediation and treatment of groundwater contamination. The data assimilation methods such as ensemble Kalman filter (EnKF) and ensemble smoother (ES) have been applied to IGCSs in recent years and obtained good identification results. The unscented kalman filter (UKF) is also a data assimilation method with the potential to simultaneously identify hydraulic conductivity and GCSs. However, when UKF is applied to identify hydraulic conductivity and GCSs, it is necessary to use the observed data at different times separately, which increases the complexity of the update process and this may result in low identification accuracy. ES is a variant of EnKF that updates the system parameters with all observed data in all time periods, which makes ES faster and easier to implement than EnKF. Therefore, inspired by the ES, an unscented kalman smoother (UKS) based on UKF was proposed for simultaneously identifying the hydraulic conductivity and GCSs in this study. The UKS can use the data observed in all time periods simultaneously, while it is also simpler to operate and the calculation speed is faster. Present studies have shown that ES can solve IGCS problems. Thus, ES was also applied to identify the hydraulic conductivity and GCSs in this study, and its identification performance was compared with UKS. In contrast to previous applications of ES to IGCSs, both UKS and ES were set up with stop iteration conditions instead of only performing one update process, and thus both methods applied multiple update processes. The results showed that compared with ES, the identification results obtained by UKS were characterized by greater stability, higher accuracy, and the iterative process required less iteration process and computational time.

This study examines how the variability of the Manning coefficient (n) affects the position of hydraulic jumps downstream of hydraulic structures. Using a robust finite volume method and random field theory, the study investigates the impact of spatial variations in n on hydraulic jump characteristics. Two scenarios are considered: one with a stilling basin and one without. Both one-dimensional and two-dimensional spatial distributions of n are analyzed. The results show that without a stilling basin, there are significant variations in the location of hydraulic jumps in the riverbed. The uncertainty in the location of the hydraulic jump is much higher than the uncertainty in the values of conjugate depths. Additionally, one-dimensional spatial distribution of n leads to higher standard deviations in the estimated location compared to two-dimensional distribution. In scenarios with a stilling basin, increasing riprap length causes the hydraulic jump to move upstream, while standard deviation remains constant.