Challenges in stratigraphic modeling arise from underground uncertainty.While borehole exploration is reliable,it remains sparse due to economic and site constraints.Electrical resistivity tomography(ERT)as a cost-eff...Challenges in stratigraphic modeling arise from underground uncertainty.While borehole exploration is reliable,it remains sparse due to economic and site constraints.Electrical resistivity tomography(ERT)as a cost-effective geophysical technique can acquire high-density data;however,uncertainty and nonuniqueness inherent in ERT impede its usage for stratigraphy identification.This paper integrates ERT and onsite observations for the first time to propose a novel method for characterizing stratigraphic profiles.The method consists of two steps:(1)ERT for prior knowledge:ERT data are processed by soft clustering using the Gaussian mixture model,followed by probability smoothing to quantify its depthdependent uncertainty;and(2)Observations for calibration:a spatial sequential Bayesian updating(SSBU)algorithm is developed to update the prior knowledge based on likelihoods derived from onsite observations,namely topsoil and boreholes.The effectiveness of the proposed method is validated through its application to a real slope site in Foshan,China.Comparative analysis with advanced borehole-driven methods highlights the superiority of incorporating ERT data in stratigraphic modeling,in terms of prediction accuracy at borehole locations and sensitivity to borehole data.Informed by ERT,reduced sensitivity to boreholes provides a fundamental solution to the longstanding challenge of sparse measurements.The paper further discusses the impact of ERT uncertainty on the proposed model using time-lapse measurements,the impact of model resolution,and applicability in engineering projects.This study,as a breakthrough in stratigraphic modeling,bridges gaps in combining geophysical and geotechnical data to address measurement sparsity and paves the way for more economical geotechnical exploration.展开更多
Spatial interpolation has been frequently encountered in earth sciences and engineering.A reasonable appraisal of subsurface heterogeneity plays a significant role in planning,risk assessment and decision making for g...Spatial interpolation has been frequently encountered in earth sciences and engineering.A reasonable appraisal of subsurface heterogeneity plays a significant role in planning,risk assessment and decision making for geotechnical practice.Geostatistics is commonly used to interpolate spatially varying properties at un-sampled locations from scatter measurements.However,successful application of classic geostatistical models requires prior characterization of spatial auto-correlation structures,which poses a great challenge for unexperienced engineers,particularly when only limited measurements are available.Data-driven machine learning methods,such as radial basis function network(RBFN),require minimal human intervention and provide effective alternatives for spatial interpolation of non-stationary and non-Gaussian data,particularly when measurements are sparse.Conventional RBFN,however,is direction independent(i.e.isotropic)and cannot quantify prediction uncertainty in spatial interpolation.In this study,an ensemble RBFN method is proposed that not only allows geotechnical anisotropy to be properly incorporated,but also quantifies uncertainty in spatial interpolation.The proposed method is illustrated using numerical examples of cone penetration test(CPT)data,which involve interpolation of a 2D CPT cross-section from limited continuous 1D CPT soundings in the vertical direction.In addition,a comparative study is performed to benchmark the proposed ensemble RBFN with two other non-parametric data-driven approaches,namely,Multiple Point Statistics(MPS)and Bayesian Compressive Sensing(BCS).The results reveal that the proposed ensemble RBFN provides a better estimation of spatial patterns and associated prediction uncertainty at un-sampled locations when a reasonable amount of data is available as input.Moreover,the prediction accuracy of all the three methods improves as the number of measurements increases,and vice versa.It is also found that BCS prediction is less sensitive to the number of measurement data and outperforms RBFN and MPS when only limited point observations are available.展开更多
Water quality in surface bodies remains a pressing issue worldwide.While some regions have rich water quality data,less attention is given to areas that lack sufficient data.Therefore,it is crucial to explore novel wa...Water quality in surface bodies remains a pressing issue worldwide.While some regions have rich water quality data,less attention is given to areas that lack sufficient data.Therefore,it is crucial to explore novel ways of managing source-oriented surface water pollution in scenarios with infrequent data collection such as weekly or monthly.Here we showed sparse-dataset-based prediction of water pollution using machine learning.We investigated the efficacy of a traditional Recurrent Neural Network alongside three Long Short-Term Memory(LSTM)models,integrated with the Load Estimator(LOADEST).The research was conducted at a river-lake confluence,an area with intricate hydrological patterns.We found that the Self-Attentive LSTM(SA-LSTM)model outperformed the other three machine learning models in predicting water quality,achieving Nash-Sutcliffe Efficiency(NSE)scores of 0.71 for COD_(Mn)and 0.57 for NH_(3)N when utilizing LOADEST-augmented water quality data(referred to as the SA-LSTMLOADEST model).The SA-LSTM-LOADEST model improved upon the standalone SA-LSTM model by reducing the Root Mean Square Error(RMSE)by 24.6%for COD_(Mn)and 21.3%for NH_(3)N.Furthermore,the model maintained its predictive accuracy when data collection intervals were extended from weekly to monthly.Additionally,the SA-LSTM-LOADEST model demonstrated the capability to forecast pollution loads up to ten days in advance.This study shows promise for improving water quality modeling in regions with limited monitoring capabilities.展开更多
基金the financial support from the National Key R&D Program of China(Grant No.2021YFC3001003)Science and Technology Development Fund,Macao SAR(File No.0056/2023/RIB2)Guangdong Provincial Department of Science and Technology(Grant No.2022A0505030019).
文摘Challenges in stratigraphic modeling arise from underground uncertainty.While borehole exploration is reliable,it remains sparse due to economic and site constraints.Electrical resistivity tomography(ERT)as a cost-effective geophysical technique can acquire high-density data;however,uncertainty and nonuniqueness inherent in ERT impede its usage for stratigraphy identification.This paper integrates ERT and onsite observations for the first time to propose a novel method for characterizing stratigraphic profiles.The method consists of two steps:(1)ERT for prior knowledge:ERT data are processed by soft clustering using the Gaussian mixture model,followed by probability smoothing to quantify its depthdependent uncertainty;and(2)Observations for calibration:a spatial sequential Bayesian updating(SSBU)algorithm is developed to update the prior knowledge based on likelihoods derived from onsite observations,namely topsoil and boreholes.The effectiveness of the proposed method is validated through its application to a real slope site in Foshan,China.Comparative analysis with advanced borehole-driven methods highlights the superiority of incorporating ERT data in stratigraphic modeling,in terms of prediction accuracy at borehole locations and sensitivity to borehole data.Informed by ERT,reduced sensitivity to boreholes provides a fundamental solution to the longstanding challenge of sparse measurements.The paper further discusses the impact of ERT uncertainty on the proposed model using time-lapse measurements,the impact of model resolution,and applicability in engineering projects.This study,as a breakthrough in stratigraphic modeling,bridges gaps in combining geophysical and geotechnical data to address measurement sparsity and paves the way for more economical geotechnical exploration.
基金supported by grants from the Research Grants Council of Hong Kong Special Administrative Region,China(Project No.City U 11213119 and T22-603/15N)The financial support is gratefully acknowledgedfinancial support from the Hong Kong Ph.D.Fellowship Scheme funded by the Research Grants Council of Hong Kong,China。
文摘Spatial interpolation has been frequently encountered in earth sciences and engineering.A reasonable appraisal of subsurface heterogeneity plays a significant role in planning,risk assessment and decision making for geotechnical practice.Geostatistics is commonly used to interpolate spatially varying properties at un-sampled locations from scatter measurements.However,successful application of classic geostatistical models requires prior characterization of spatial auto-correlation structures,which poses a great challenge for unexperienced engineers,particularly when only limited measurements are available.Data-driven machine learning methods,such as radial basis function network(RBFN),require minimal human intervention and provide effective alternatives for spatial interpolation of non-stationary and non-Gaussian data,particularly when measurements are sparse.Conventional RBFN,however,is direction independent(i.e.isotropic)and cannot quantify prediction uncertainty in spatial interpolation.In this study,an ensemble RBFN method is proposed that not only allows geotechnical anisotropy to be properly incorporated,but also quantifies uncertainty in spatial interpolation.The proposed method is illustrated using numerical examples of cone penetration test(CPT)data,which involve interpolation of a 2D CPT cross-section from limited continuous 1D CPT soundings in the vertical direction.In addition,a comparative study is performed to benchmark the proposed ensemble RBFN with two other non-parametric data-driven approaches,namely,Multiple Point Statistics(MPS)and Bayesian Compressive Sensing(BCS).The results reveal that the proposed ensemble RBFN provides a better estimation of spatial patterns and associated prediction uncertainty at un-sampled locations when a reasonable amount of data is available as input.Moreover,the prediction accuracy of all the three methods improves as the number of measurements increases,and vice versa.It is also found that BCS prediction is less sensitive to the number of measurement data and outperforms RBFN and MPS when only limited point observations are available.
基金supported by the Strategic Priority Research Program of the Chinese Academy of Sciences(XDA23040502)National Natural Science Foundation of China(41890823)Key Laboratory of Water Cycle and Related Land Surface Processes,Institute of Geographic Sciences and Natural Resources Research,Chinese Academy of Sciences(No.WL2019003).
文摘Water quality in surface bodies remains a pressing issue worldwide.While some regions have rich water quality data,less attention is given to areas that lack sufficient data.Therefore,it is crucial to explore novel ways of managing source-oriented surface water pollution in scenarios with infrequent data collection such as weekly or monthly.Here we showed sparse-dataset-based prediction of water pollution using machine learning.We investigated the efficacy of a traditional Recurrent Neural Network alongside three Long Short-Term Memory(LSTM)models,integrated with the Load Estimator(LOADEST).The research was conducted at a river-lake confluence,an area with intricate hydrological patterns.We found that the Self-Attentive LSTM(SA-LSTM)model outperformed the other three machine learning models in predicting water quality,achieving Nash-Sutcliffe Efficiency(NSE)scores of 0.71 for COD_(Mn)and 0.57 for NH_(3)N when utilizing LOADEST-augmented water quality data(referred to as the SA-LSTMLOADEST model).The SA-LSTM-LOADEST model improved upon the standalone SA-LSTM model by reducing the Root Mean Square Error(RMSE)by 24.6%for COD_(Mn)and 21.3%for NH_(3)N.Furthermore,the model maintained its predictive accuracy when data collection intervals were extended from weekly to monthly.Additionally,the SA-LSTM-LOADEST model demonstrated the capability to forecast pollution loads up to ten days in advance.This study shows promise for improving water quality modeling in regions with limited monitoring capabilities.
