Description:FIELD OF THE INVENTION
[0001] The present invention relates to the field of natural disaster prediction and early warning systems, and more particularly to a hybrid landslide risk assessment system and computational method thereof for analyzing and predicting landslide triggers through integrated processing of temporal rainfall data and seismic accelerogram data.

DESCRIPTION OF THE RELATED ART
[0002] The following description of the related art is intended to provide background information pertaining to the field of disclosure. This section may include certain aspects of the art that may be related to various features of the present disclosure. However, it should be appreciated that this section is used only to enhance the understanding of the reader with respect to the present disclosure, and not as admissions of the prior art.
[0003] Landslides represent one of the most devastating natural disasters worldwide, causing significant loss of life, property damage, and economic disruption. These geological hazards are typically triggered by various environmental factors, with rainfall and earthquakes being the two most prevalent triggers. The complexity of landslide mechanisms increases substantially when multiple triggering factors occur simultaneously or in close temporal succession, such as when seismic events are accompanied by heavy rainfall.
[0004] Traditional landslide prediction methods have generally focused on single-trigger analysis, examining either rainfall-induced or earthquake-induced landslides in isolation. Rainfall-triggered landslide assessment methods typically employ threshold-based approaches that correlate precipitation intensity and duration with landslide occurrence. However, these methods often fail to account for the cumulative effects of antecedent rainfall that can progressively weaken slope stability over extended periods.
[0005] Similarly, earthquake-induced landslide assessment methods, such as Newmark's displacement analysis, focus primarily on seismic ground motion parameters while neglecting the influence of soil moisture conditions and recent precipitation patterns. This single-factor approach leads to incomplete risk assessments, particularly in regions where both seismic activity and heavy rainfall are common.
[0006] Conventional landslide prediction methods also suffer from methodological limitations in their processing workflows. These methods typically involve sequential analysis where rainfall and seismic data are processed independently and then combined post-hoc, leading to loss of important interaction effects. The lack of parallel processing methods that can simultaneously analyze multiple data streams while maintaining their temporal relationships significantly reduces prediction accuracy.
[0007] Furthermore, existing computational methods for landslide risk assessment typically require extensive manual calibration and lack automated workflows for integrating physical models with statistical approaches. The absence of hybrid processing methods that can synthesize deterministic physics-based models with probabilistic statistical models limits their effectiveness in providing comprehensive risk assessments.
[0008] Current methods also fail to provide efficient computational frameworks for generating spatial landslide trigger probability maps in real-time. The computational burden of processing high-resolution temporal and spatial data using conventional methods makes it impractical for operational early warning systems that require rapid response times.
[0009] Therefore, there exists a need for an improved landslide risk assessment system and computational method thereof that can effectively integrate multiple trigger factors through hybrid processing approaches, account for temporal exposure patterns using distributed lag models, and provide real-time probability estimates that distinguish between different triggering mechanisms. Such a system and method would significantly enhance disaster preparedness and risk mitigation efforts in landslide-prone regions.

OBJECTS OF THE PRESENT DISCLOSURE
[0010] Some of the objects of the present disclosure, which at least one embodiment herein satisfies are as listed herein below.
[0011] An object of the present disclosure is to provide a hybrid landslide risk assessment system and computational method thereof for automated landslide trigger prediction which employs a distributed lag nonlinear model processor and a Newmark displacement processor stored in memory to classify landslide triggers into rainfall-induced, earthquake-induced, or combined-effect categories based on temporal rainfall data and seismic accelerogram analysis.
[0012] An object of the present disclosure is to provide a hybrid landslide risk assessment system including a hybrid integration module that synthesizes outputs from both a distributed lag nonlinear model processor calculating cumulative rainfall exposure effects and a Newmark processor computing earthquake-induced displacement values to generate adaptive landslide trigger probability estimates based on temporal lag relationships and spectrum-compatible artificial accelerograms.
[0013] An object of the present disclosure is to provide a computational method for landslide risk assessment that implements parallel processing where the temporal rainfall measurements are routed to a distributed lag nonlinear model processor for calculating cross-basis functions while seismic accelerogram data are simultaneously routed to a Newmark processor for displacement calculations, with both outputs integrated through logistic regression with distributed lag terms.
[0014] An object of the present disclosure is to provide a hybrid landslide risk assessment system that generates real-time spatial landslide trigger probability maps distinguishing between rainfall-triggered, earthquake-triggered, and combined-effect landslides by processing multiple environmental parameters including temporal rainfall measurements, seismic accelerogram data, and lithological parameters through an integrated dual-processor architecture.
[0015] An object of the present disclosure is to provide a hybrid landslide risk assessment system and method thereof that overcomes limitations of single-factor analysis by capturing non-linear interactions between rainfall exposure patterns and seismic ground motion through cross-basis function generation, enabling accurate prediction of complex multi-trigger landslide events for early warning applications.

SUMMARY
[0016] This section is provided to introduce certain objects and aspects of the present disclosure in a simplified form that are further described below in the detailed description. This summary is not intended to identify the key features or the scope of the claimed subject matter.
[0017] The present disclosure generally relates to natural disaster prediction and geological hazard assessment systems. More particularly, the present disclosure relates to a hybrid landslide risk assessment system that employs distributed lag nonlinear models and Newmark displacement analysis for accurate landslide trigger prediction through integration of temporal rainfall data and seismic accelerogram analysis, providing enhanced hazard assessment capabilities through probability-based classification and spatial visualization mechanisms for improved disaster preparedness and early warning applications.
[0018] An aspect of the present disclosure relates to a hybrid landslide trigger prediction system for automated risk assessment and classification. The system includes rainfall data receivers, earthquake data receivers, and geological data receivers operatively connected to a control unit for acquiring environmental parameters. The system includes a distributed lag nonlinear model processor and a Newmark processor connected to the control unit for parallel processing of rainfall and seismic data respectively. The system includes a hybrid integration module electrically connected to both processors that synthesizes statistical and physical model outputs using logistic regression with distributed lag terms. The system includes a memory unit storing instructions that enable the control unit to process received data, calculate cumulative exposure effects and displacement values, and generate landslide trigger probability estimates. The system includes processing components that identify whether landslides are triggered by rainfall, earthquake, or combined effects based on temporal lag relationships and spectrum-compatible artificial accelerograms. The system includes display outputs for transmitting integrated probability estimates as spatial landslide trigger maps for decision support.
[0019] In another aspect, the present disclosure relates to a method for landslide trigger prediction and risk assessment. The method includes receiving temporal rainfall measurements, seismic accelerogram data, and lithological parameters at respective data receivers. The method includes routing rainfall measurements to a distributed lag nonlinear model processor for calculating cross-basis functions representing cumulative exposure effects. The method includes simultaneously routing seismic data to a Newmark processor for generating artificial accelerograms and computing displacement values. The method includes transferring outputs from both processors to a hybrid integration module for combined analysis. The method includes executing logistic regression with distributed lag terms to generate adaptive landslide trigger probability estimates that distinguish between rainfall-triggered, earthquake-triggered, and combined-effect landslides. The method includes displaying spatial probability maps on output devices for early warning and risk management applications.
[0020] Various objects, features, aspects and advantages of the inventive subject matter will become more apparent from the following detailed description of preferred embodiments, along with the accompanying drawing figures in which like numerals represent like components.

BRIEF DESCRIPTION OF DRAWINGS
[0021] The accompanying drawings are included to provide a further understanding of the present disclosure and are incorporated in and constitute a part of this specification. The drawings illustrate exemplary embodiments of the present disclosure and, together with the description, serve to explain the principles of the present disclosure. The diagrams are for illustration only, which thus is not a limitation of the present disclosure.
[0022] FIG. 1 illustrates an exemplary block diagram of the hybrid landslide trigger prediction system, in accordance with an embodiment of the present disclosure.
[0023] FIG. 2 illustrates an exemplary receiver operating characteristic (ROC) curve comparing the performance of three models: basic GAM, GAM with pre-seismic rainfall, and the distributed lag nonlinear model for landslide trigger prediction, in accordance with an embodiment of the present disclosure.
[0024] FIG. 3 illustrates an exemplary spatial distribution map depicting (a) cumulative rainfall sum overlayed with probability of rainfall trigger and (b) Newmark's displacement values overlayed on geological formations for the study area, in accordance with an embodiment of the present disclosure.
[0025] FIG. 4 illustrates an exemplary three-dimensional surface plot depicting the relationship between rainfall exposure, temporal lag periods, and landslide trigger probability estimates generated by the distributed lag nonlinear model, in accordance with an embodiment of the present disclosure.
[0026] FIG. 5 illustrates an exemplary flow diagram depicting the method for predicting landslide triggers, in accordance with an embodiment of the present disclosure.

DETAILED DESCRIPTION
[0027] The ensuing description provides exemplary embodiments only, and is not intended to limit the scope, applicability, or configuration of the disclosure. Rather, the ensuing description of the exemplary embodiments will provide those skilled in the art with an enabling description for implementing an exemplary embodiment. It should be understood that various changes may be made in the function and arrangement of elements without departing from the spirit and scope of the invention as set forth.
Definitions:
Trigger Probability: A numerical value ranging from 0 to 1 calculated by the hybrid integration module representing the likelihood that a landslide will be triggered by rainfall, earthquake, or combined effects, compared against field-verified data for validation.
Predetermined Categories: Classification groups for landslide triggers including rainfall-induced, earthquake-induced, and combined-effect landslides, each associated with specific probability ranges and spatial distribution patterns for hazard mapping.
Cross-basis Functions: Mathematical representations combining main basis functions and lag basis functions that capture the non-linear relationship between rainfall exposure and landslide occurrence over temporal periods.
Spectrum-compatible Artificial Accelerograms: Synthetic ground motion records generated to match recorded seismic parameters including peak ground acceleration and Arias intensity for site-specific displacement calculations.
Distributed Lag Terms: Statistical parameters in logistic regression that account for the delayed and cumulative effects of rainfall exposure on slope stability over predetermined time periods.
[0028] An aspect of the present disclosure relates to a hybrid landslide risk assessment system for automated trigger prediction including rainfall data receivers positioned to acquire temporal rainfall measurements over predetermined periods. The system includes earthquake data receivers processing seismic accelerogram data and peak ground acceleration values. The system includes geological data receivers inputting lithological parameters including rock type and slope characteristics. The system includes processors executing distributed lag nonlinear models and Newmark analysis stored in memory for trigger classification. The system includes a hybrid integration module synthesizing outputs based on logistic regression. The system includes display outputs presenting spatial probability maps. The system includes memory storing models and geological parameters. The system includes processing modules that analyze multi-source data and generate probability estimates. The system includes a network module interfacing with external databases for enhanced prediction accuracy.
[0029] Various embodiments of the present disclosure are described using FIGs. 1 to 5.
[0030] FIG. 1 illustrates an exemplary block diagram of the hybrid landslide trigger prediction system, in accordance with an embodiment of the present disclosure.
[0031] In an embodiment, referring to FIG. 1, the system (100) for hybrid landslide trigger prediction can include an input/output unit including rainfall data receiver (114), earthquake data receiver (116), geological data receiver (118), and display output (120). The system can include a control unit (112) serving as the main controller coordinating data flow between components. The processing architecture can include a distributed lag nonlinear model (DLNM) processor (102) and a Newmark processor (104) operating in parallel. The system can include a memory unit (106) storing models and parameters, a network module (108) for external database access, and a hybrid integration module (110) synthesizing outputs. The interconnections can enable simultaneous processing of multiple data streams maintaining temporal relationships.
[0032] In an embodiment, the rainfall data receiver (114) can be implemented in the system (100) to acquire temporal rainfall measurements at predetermined intervals. The receiver (114) can interface with rain gauge networks capturing precipitation data at hourly or daily resolution. The receiver can validate data integrity checking for missing values or anomalous readings. The temporal measurements can span predetermined periods from days to weeks capturing antecedent conditions. The receiver can format data into time series compatible with DLNM processor requirements. The continuous data acquisition can enable real-time monitoring of evolving rainfall patterns. The receiver (114) can include, but is not limited to, automated weather station interfaces, satellite precipitation receivers, and radar-based rainfall sensors, to provide comprehensive precipitation monitoring.
[0033] In an embodiment, the earthquake data receiver (116) can be implemented in the system (100) to process seismic accelerogram data from strong motion networks. The receiver (116) can extract peak ground acceleration values and complete time histories from three-component records. The receiver can identify the horizontal component with maximum PGA for displacement calculations. The seismic data can include hypocentral parameters enabling distance-based attenuation calculations. The receiver can synchronize seismic events with rainfall time series for combined analysis. The real-time processing can enable rapid post-earthquake landslide assessment. The receiver (116) can interface with, but not limited to, national seismic networks, local accelerometer arrays, and global earthquake catalogs.
[0034] In an embodiment, the geological data receiver (118) can be implemented in the system (100) to input lithological parameters essential for stability analysis. The receiver (118) can process rock type classifications from geological databases including formation names and material properties. The slope characteristics can be extracted from digital elevation models at specified resolutions. The receiver can integrate shear strength parameters including cohesion and friction angle values. The lithological data can enable factor of safety calculations for critical acceleration determination. The spatial resolution can match landslide inventory points ensuring accurate parameter assignment. The receiver (118) can access geological survey databases, geotechnical investigation reports, and remote sensing-derived geological maps.
[0035] In an embodiment, the control unit (112) can be implemented in the system (100) to orchestrate data flow and processing sequences across all components. The control unit (112) can receive environmental parameters from all data receivers simultaneously. Upon data reception, the control unit can validate completeness and initiate parallel processing pathways. The unit can route temporal rainfall measurements and lithological parameters to the DLNM processor (102). Concurrently, the unit can direct seismic accelerogram data and geological parameters to the Newmark processor (104). The control unit can monitor processing status preventing bottlenecks in the computational pipeline. The synchronized operation can ensure coherent multi-modal analysis maintaining temporal relationships.
[0036] In an embodiment, the distributed lag nonlinear model processor (102) can be implemented in the system (100) to calculate cumulative rainfall exposure effects. The processor (102) can generate cross-basis functions combining exposure-response relationships with temporal lag structures. The main basis functions can capture non-linear rainfall-trigger relationships including threshold effects. The lag basis functions can model the temporal distribution of rainfall influence over predetermined periods. The processor can estimate regression coefficients through maximum likelihood optimization. The cross-basis expansion can enable flexible modeling of complex exposure patterns. The processor (102) can output probability surfaces representing trigger likelihood across rainfall-lag combinations.
[0037] In an embodiment, the Newmark processor (104) can be implemented in the system (100) to compute earthquake-induced displacement values. The processor (104) can calculate critical acceleration using factor of safety derived from geological parameters. The processor can generate spectrum-compatible artificial accelerograms matching recorded ground motion characteristics. The double integration of acceleration exceeding critical values can yield cumulative displacement. The processor can account for site-specific amplification effects based on geological conditions. The displacement thresholds can classify earthquake contribution to landslide triggering. The processor (104) can output spatial displacement maps for integration with rainfall analysis.
[0038] In an embodiment, the memory unit (106) can be implemented in the system (100) to store computational models, parameters, and intermediate results. The memory (106) can maintain the trained DLNM model including basis function specifications and regression coefficients. The unit can store geological databases with shear strength parameters for various rock types. The memory can buffer temporal datasets enabling efficient access during iterative calculations. The predetermined thresholds for trigger classification can be stored for rapid comparison. The memory architecture can support parallel read operations from multiple processors. The efficient data structures can minimize computational latency during real-time processing.
[0039] In an embodiment, the network module (108) can be implemented in the system (100) to interface with external databases enhancing local measurements. The module (108) can access precipitation databases through application programming interfaces retrieving regional rainfall patterns. The module can connect to seismic data repositories obtaining comprehensive earthquake catalogs. The geological surveys can be accessed for updated lithological classifications and material properties. The network protocols can ensure secure data transmission maintaining data integrity. The caching mechanisms can store frequently accessed external data reducing network latency. The module (108) can enable system scalability through cloud-based data integration.
[0040] In an embodiment, the hybrid integration module (110) can be implemented in the system (100) to synthesize outputs from both processors generating unified probability estimates. The module (110) can receive cross-basis functions from the DLNM processor representing rainfall exposure effects. Simultaneously, the module can obtain displacement values from the Newmark processor indicating seismic contributions. The logistic regression with distributed lag terms can combine both inputs accounting for interaction effects. The module can calculate probability distributions across trigger categories: rainfall-induced, earthquake-induced, and combined effects. The adaptive probability estimates can reflect site-specific conditions through geological modulation. The integration can preserve the temporal structure of rainfall effects while incorporating instantaneous seismic impacts.
[0041] In an embodiment, the display output (120) can be implemented in the system (100) to visualize spatial landslide trigger probability maps. The output (120) can render color-coded maps distinguishing trigger categories through distinct symbology. The probability gradients can be displayed using continuous color scales enhancing interpretation. The display can overlay multiple data layers including rainfall, displacement, and geology. The interactive features can enable zooming and querying specific locations for detailed information. The temporal animation can show evolving probability patterns as rainfall accumulates. The output (120) can generate reports summarizing high-risk areas for decision support.
[0042] FIG. 2 illustrates an exemplary receiver operating characteristic (ROC) curve comparing the performance of three models: basic GAM, GAM with pre-seismic rainfall, and the distributed lag nonlinear model for landslide trigger prediction, in accordance with an embodiment of the present disclosure.
[0043] In an embodiment, referring to FIG. 2, illustrating the ROC curve comparison (200) implemented for model performance evaluation within the system (100). The curve (200) depicts sensitivity on the y-axis ranging from 0 to 1.0 versus specificity on the x-axis from 1.0 to 0. Upon analyzing the curves, the distributed lag model (green line) can demonstrate superior performance with an AUC of 83%. The GAM with 3-weeks average rainfall (red line) can achieve an AUC of 80% showing intermediate performance. The basic Lag-0 model (blue line) can exhibit the lowest performance with an AUC of 77%. The diagonal reference line can represent random classification with AUC of 50%. The area between each curve and the diagonal can quantify the improvement over random prediction.
[0044] In an embodiment, the ROC curves (200) can validate the enhanced prediction capability of the distributed lag nonlinear model over conventional approaches. The curves can be generated through leave-one-out spatial cross-validation ensuring robust performance estimates. The distributed lag model's superior AUC can result from its ability to capture temporal exposure patterns. The improvement from 77% to 83% AUC can represent a significant enhancement in operational landslide prediction. The curves can guide threshold selection for operational deployment balancing sensitivity and specificity requirements. The performance metrics can enable objective comparison between different modeling approaches.
[0045] FIG. 3 illustrates an exemplary spatial distribution map depicting (a) cumulative rainfall sum overlayed with probability of rainfall trigger and (b) Newmark's displacement values overlayed on geological formations for the study area, in accordance with an embodiment of the present disclosure.
[0046] In an embodiment, referring to FIG. 3 illustrating the spatial distribution maps (300) for the system's study area demonstrating geographic patterns of landslide triggers. The map (300a) can display cumulative rainfall from August 22nd to September 18th with values ranging from 53.2 mm to 519.3 mm represented by blue gradient shading. The overlayed circles can indicate probability of rainfall trigger with size corresponding to probability values from 0-0.3 (small), 0.31-0.6 (medium), to 0.61-0.82 (large). The map can reveal high rainfall accumulation in western districts (W) exceeding 400 mm. The northern district (N) can show moderate rainfall with concentrated landslide occurrences near Lachen and Lachung. The spatial correlation between rainfall patterns and trigger probabilities can validate the model's predictive capability.
[0047] In an embodiment, the map (300b) can illustrate Newmark's displacement values ranging from 0-10 cm (yellow circles) to 41-200 cm (red circles) overlayed on geological formations. The Kanchenjunga Gneiss formation (brown) can show predominance of low displacement values indicating rainfall-dominated triggers. The Gorubathan formation (light blue) can exhibit higher displacement values suggesting earthquake influence. The Chungthang formation (dark blue) can demonstrate mixed trigger patterns. The geological control on trigger mechanisms can be evident from the spatial clustering of displacement values. The maps can enable targeted hazard assessment based on combined rainfall-seismic-geological factors.
[0048] FIG. 4 illustrates an exemplary three-dimensional surface plot depicting the relationship between rainfall exposure, temporal lag periods, and landslide trigger probability estimates generated by the distributed lag nonlinear model, in accordance with an embodiment of the present disclosure.
[0049] In an embodiment, referring to FIG. 4, the three-dimensional surface plot (400) for landslide trigger probability visualization can demonstrate the complex non-linear relationships captured by the distributed lag nonlinear model processor (102). The surface plot (400) can include the x-axis representing rainfall amount in millimeters ranging from 0 to 80 mm, the y-axis depicting lag periods in days from 0 to 25 days, and the z-axis showing probability estimates (Effect) ranging from 0 to 8 on a logarithmic scale. The plot (400) can reveal peak probability values occurring at high rainfall amounts (60-80 mm) with minimal lag and at moderate rainfall (40-60 mm) with 20-25 day lag periods. The surface curvature can indicate the distributed lag effect where antecedent rainfall from weeks prior continues influencing current landslide susceptibility. The visualization can enable identification of critical rainfall-lag combinations requiring enhanced monitoring.
[0050] In an embodiment, the surface plot (400) can be generated by the distributed lag nonlinear model processor (102) through cross-basis function calculations integrating temporal and exposure dimensions. The processor can compute probability surfaces by evaluating the interaction between rainfall intensity and temporal distribution patterns. The three-dimensional representation can capture the cumulative exposure effects that linear models fail to represent. The surface gradients can indicate the rate of probability change with varying rainfall and lag combinations. The plot (400) can guide early warning thresholds by identifying high-risk rainfall-lag scenarios. The visualization can be updated in real-time as new rainfall data becomes available enabling dynamic risk assessment.
[0051] FIG. 5 illustrates an exemplary flow diagram depicting the method for predicting landslide triggers, in accordance with an embodiment of the present disclosure.
[0052] In an embodiment, referring to FIG. 5 illustrating a step-by-step operational process (500) for the system (100) outlining systematic trigger prediction methodology. At step (502) receiving temporal rainfall measurements, seismic accelerogram data, and lithological parameters establishes the multi-source data foundation. The process transmits all received data to the control unit for centralized processing at step (504). The system accesses external databases through the network module supplementing local measurements. At step (506) the system routes data to respective processors based on data type. At step (508) the DLNM processor calculates cross-basis functions for rainfall analysis. Simultaneously at step (510) the Newmark processor computes displacement values from seismic data. At step (512) outputs transfer to the hybrid integration module. The system executes logistic regression at step (514) combining both processor outputs. At step (516) the system generates adaptive probability estimates identifying trigger categories. The probability estimates transmit to display output at step (518). Finally at step (520) spatial trigger maps display for decision support.
[0053] In an embodiment, the method for landslide trigger prediction can be implemented in step (502), receiving temporal rainfall measurements can involve activating the rainfall data receiver (114) to acquire precipitation data from gauge networks. The measurements can span predetermined periods capturing both antecedent and concurrent rainfall. The temporal resolution can range from hourly to daily depending on data availability and system requirements. The receiver can implement quality control procedures identifying and flagging anomalous values. The measurements can include rainfall intensity, duration, and cumulative amounts. The data reception can occur continuously enabling real-time hazard assessment. The buffering mechanisms can prevent data loss during communication interruptions.
[0054] In an embodiment, the method can be implemented where receiving seismic accelerogram data involves the earthquake data receiver (116) processing strong motion records. The accelerograms can include three-component time histories from multiple recording stations. The receiver can extract ground motion parameters including PGA, PGV, and Arias intensity. The hypocenter information can enable distance calculations for each landslide location. The data format can comply with standard seismological formats ensuring compatibility. The automatic triggering can initiate data reception upon earthquake detection. The multi-station data can enable spatial interpolation of ground motion fields.
[0055] In an embodiment, the method can be implemented where receiving lithological parameters involves the geological data receiver (118) accessing formation maps and material databases. The parameters can include rock type classifications at various taxonomic levels. The shear strength properties including cohesion and friction angle can be assigned based on lithology. The slope angles can be extracted from digital elevation models at landslide locations. The geological structure information can indicate potential failure planes. The spatial join operations can link geological attributes to landslide inventory points. The parameter validation can ensure realistic value ranges preventing computational errors.
[0056] In an embodiment, the method for trigger prediction can be implemented in step (504), transmitting data to the control unit can involve establishing synchronized data streams from all receivers. The control unit (112) can implement buffer management preventing data overflow during peak events. The data packets can include timestamps enabling temporal alignment across different sources. The transmission protocols can ensure data integrity through checksums and acknowledgments. The priority queuing can handle time-critical seismic data with minimal latency. The control unit can log all received data maintaining audit trails for system validation. The error handling routines can manage communication failures gracefully.
[0057] In an embodiment, the method can be implemented in step (506), routing data from the control unit can involve intelligent distribution based on data types and processing requirements. The rainfall measurements can be packaged with corresponding temporal indices for DLNM processing. The seismic accelerograms can be paired with site-specific geological parameters for Newmark analysis. The routing logic can prevent data misalignment maintaining coherent processing streams. The parallel pathways can enable simultaneous execution maximizing computational efficiency. The metadata preservation can ensure traceability throughout the processing pipeline. The load balancing can distribute computational tasks preventing processor bottlenecks.
[0058] In an embodiment, the method can be implemented in step (508), processing rainfall data in the DLNM processor can involve constructing cross-basis matrices from temporal measurements. The processor (102) can apply basis expansion using natural splines for smooth function representation. The main basis can model the exposure-response relationship capturing non-linear effects. The lag basis can represent the temporal distribution of rainfall influence. The maximum likelihood estimation can optimize model parameters given observed trigger data. The cross-validation can prevent overfitting ensuring generalization capability. The processor can generate probability surfaces across the rainfall-lag space as shown in the three-dimensional plot (400).
[0059] In an embodiment, the method can be implemented in step (510), processing seismic data in the Newmark processor can involve calculating site-specific critical accelerations. The processor (104) can compute factor of safety using geological parameters and slope characteristics. The artificial accelerogram generation can match target spectra from recorded ground motions. The numerical integration can accumulate displacement when acceleration exceeds critical values. The processor can apply empirical corrections for topographic amplification effects. The displacement calculations can consider both horizontal components selecting maximum values. The spatial interpolation can estimate displacement at unrecorded locations as visualized in map (300b).
[0060] In an embodiment, the method can be implemented in step (512), transferring outputs to the hybrid integration module can involve formatting processor results for combined analysis. The DLNM processor output can include probability matrices indexed by rainfall amount and lag period. The Newmark processor output can include displacement values with associated confidence intervals. The data transfer can maintain full numerical precision preventing rounding errors. The synchronization can ensure outputs correspond to identical spatial locations and time periods. The metadata can document processing parameters enabling reproducibility. The module (110) can validate output completeness before integration.
[0061] In an embodiment, the method can be implemented in step (514), executing logistic regression with distributed lag terms can involve formulating the combined statistical model. The regression can incorporate cross-basis functions as predictors representing rainfall effects. The displacement values can enter as additional covariates capturing seismic contributions. The interaction terms can model synergistic effects between rainfall and earthquakes. The distributed lag terms can preserve the temporal structure of rainfall influence. The maximum likelihood optimization can estimate regression coefficients. The model diagnostics can assess goodness-of-fit and residual patterns validated through ROC curves (200).
[0062] In an embodiment, the method can be implemented in step (516), generating adaptive landslide trigger probability estimates can involve evaluating the fitted logistic model. The probability calculations can generate values for each trigger category at every spatial location. The adaptive nature can reflect local conditions through site-specific parameter values. The confidence intervals can quantify prediction uncertainty based on data availability. The probability thresholds can classify triggers into dominant categories. The spatial smoothing can reduce noise while preserving significant gradients. The estimates can update dynamically as new data becomes available.
[0063] In an embodiment, the method can be implemented in step (518), transmitting probability estimates to display output can involve preparing visualization-ready datasets. The transmission can include probability values, confidence intervals, and trigger classifications. The spatial indexing can enable efficient rendering of large datasets. The data compression can reduce transmission bandwidth without losing critical information. The metadata can specify color scales, symbology, and map projections. The incremental updates can refresh only changed regions improving display performance. The output module (120) can acknowledge successful data reception.
[0064] In an embodiment, the method can be implemented in step (520), displaying spatial landslide trigger probability maps can involve rendering geographic visualizations. The maps can use color gradients representing probability magnitudes from low (green) to high (red) as demonstrated in spatial distribution maps (300). The different symbols can distinguish trigger categories: circles for rainfall, triangles for earthquake, squares for combined. The base maps can provide geographic context including terrain, roads, and settlements. The interactive features can enable panning, zooming, and location queries. The legend can explain symbology and provide interpretation guidance. The export functions can generate high-resolution images for reports and presentations.
[0065] The method can further include performing spatial cross-validation where the system evaluates prediction accuracy using field-verified landslide data. The leave-one-out approach can assess model performance across different geographic regions. The validation metrics can include accuracy, precision, recall, and F1-scores for each trigger category achieving the 83% AUC shown in ROC curves (200). The spatial autocorrelation analysis can ensure independence of validation samples. The error analysis can identify systematic biases requiring model refinement. The validation results can guide threshold optimization for operational deployment. The continuous validation can maintain model reliability as new events occur.
[0066] The method can include updating model parameters where the system incorporates new landslide events improving prediction capability. The incremental learning can adjust regression coefficients based on recent observations. The temporal windows can be updated capturing evolving rainfall patterns reflected in the surface plot (400). The geological databases can integrate new geotechnical investigations. The seismic catalogs can include recent earthquakes expanding the magnitude-distance range. The automated workflows can ensure timely model updates. The version control can track model evolution enabling rollback if needed.
[0067] In an embodiment, the system (100) can handle complex multi-trigger scenarios by detecting rainfall accumulation patterns through temporal analysis, automatically processing concurrent seismic events, analyzing combined effects using hybrid integration, determining dominant trigger mechanisms through probability comparison, generating early warnings based on threshold exceedances, ensuring accurate hazard assessment through multi-modal validation, and enabling proactive risk management through real-time monitoring.
[0068] The described disclosure presents an advanced hybrid landslide risk assessment system (100) integrated with parallel processing architecture that offers several novel features distinguishing it from conventional single-trigger analysis methods. The system (100) can automatically classify triggers when environmental thresholds are exceeded, enabling real-time hazard assessment without human intervention. The rainfall data receiver (114) can capture temporal patterns while the earthquake data receiver (116) processes seismic events that coordinate with processors executing physics-based and statistical models. During hazard analysis, the combination of DLNM and Newmark processors ensures comprehensive trigger identification across diverse geological settings. The automated system (100) ensures consistent risk assessment without requiring extensive calibration. The network-enabled operation effectively enables integration with regional monitoring systems.
[0069] In an exemplary embodiment, the system (100) can operate within validated performance parameters ensuring reliable trigger prediction across varying environmental conditions. The system (100) can demonstrate prediction accuracy exceeding 83% for trigger classification through hybrid model optimization as validated in ROC curves (200). The processing architecture can achieve probability calculations within 500 milliseconds from data reception to map generation. The distributed lag modeling can capture rainfall effects over 25-day periods with daily resolution as shown in the surface plot (400). The system (100) can process multiple landslide events simultaneously supporting regional-scale assessment visible in spatial maps (300). These validated performance metrics make the system (100) practical, efficient, and scalable for operational landslide early warning applications.
[0070] While considerable emphasis has been placed herein on the preferred embodiments, it will be appreciated that many embodiments can be made and that many changes can be made in the preferred embodiments without departing from the principles of the disclosure. These and other changes in the preferred embodiments of the disclosure will be apparent to those skilled in the art from the disclosure herein, whereby it is to be distinctly understood that the foregoing descriptive matter to be implemented merely as illustrative of the disclosure and not as limitation.
[0071] If the specification states a component or feature “may”, “can”, “could”, or “might” be included or have a characteristic, that particular component or feature is not required to be included or have the characteristic.
[0072] As used in the description herein and throughout the claims that follow, the meaning of “a,” “an,” and “the” includes plural reference unless the context clearly dictates otherwise. Also, as used in the description herein, the meaning of “in” includes “in” and “on” unless the context clearly dictates otherwise.
[0073] Moreover, in interpreting the specification, all terms should be interpreted in the broadest possible manner consistent with the context. In particular, the terms “comprises” and “comprising” should be interpreted as referring to elements, components, or steps in a non-exclusive manner, indicating that the referenced elements, components, or steps may be present, or utilized, or combined with other elements, components, or steps that are not expressly referenced. Where the specification claims refer to at least one of something selected from the group consisting of A, B, C ….and N, the text should be interpreted as requiring only one element from the group, not A plus N, or B plus N, etc.
[0074] While the foregoing describes various embodiments of the proposed disclosure, other and further embodiments of the proposed disclosure may be devised without departing from the basic scope thereof. The scope of the proposed disclosure is determined by the claims that follow. The proposed disclosure is not limited to the described embodiments, versions or examples, which are included to enable a person having ordinary skill in the art to make and use the invention when combined with information and knowledge available to the person having ordinary skill in the art.

ADVANTAGES OF THE PRESENT DISCLOSURE
[0075] The present disclosure provides a hybrid landslide risk assessment system that enables automated prediction of diverse landslide triggers through integrated DLNM and Newmark processing achieving classification accuracy exceeding 83% with sub-second probability calculations, eliminating single-factor analysis limitations and reducing uncertainty in hazard assessment for disaster-prone regions.
[0076] The present disclosure provides a hybrid landslide risk assessment system that implements parallel processing architecture with distributed lag nonlinear model and Newmark processors to provide real-time trigger classification capturing temporal rainfall effects over 25-day periods while simultaneously computing seismic displacements, enabling comprehensive multi-hazard assessment without extensive calibration requirements.
[0077] The present disclosure provides a hybrid landslide risk assessment system that utilizes cross-basis function generation integrating main basis and lag basis functions to enhance prediction reliability through temporal modeling, capturing cumulative antecedent rainfall effects and non-linear exposure-response relationships to improve early warning capabilities for rainfall-induced landslides.
[0078] The present disclosure provides a hybrid landslide risk assessment system that employs spectrum-compatible artificial accelerogram generation matched to recorded ground motion parameters enabling site-specific displacement calculations, accounting for geological variations and topographic amplification effects to accurately assess earthquake contributions to slope failures.
, Claims:1. A hybrid landslide trigger prediction system (100) comprising:
at least one input unit comprising at least one rainfall data receiver (114), at least one earthquake data receiver (116), and at least one geological data receiver (118), each configured to receive corresponding environmental parameters;
at least one rainfall data receiver (114) to acquire temporal rainfall measurements over predetermined time periods;
at least one earthquake data receiver (116) to process seismic accelerogram data and peak ground acceleration values;
at least one geological data receiver (118) to input lithological parameters comprising rock type and slope characteristics;
a processing unit comprising at least one distributed lag nonlinear model processor (102) and at least one Newmark processor (104) to analyze rainfall and seismic data respectively;
at least one control unit (112) electrically connected to the rainfall data receiver (114), the earthquake data receiver (116), the geological data receiver (118), the distributed lag nonlinear model processor (102), and the Newmark processor (104);
at least one hybrid integration module (110) electrically connected to both the distributed lag nonlinear model processor (102) and the Newmark processor (104) to synthesize physical and statistical model outputs;
at least one network module (108) electrically connected to the control unit (112) to interface with external databases;
at least one memory unit (106) operatively coupled to the control unit (112), the distributed lag nonlinear model processor (102), the Newmark processor (104), and the hybrid integration module (110), wherein the memory unit (106) stores instructions which, when executed by the control unit (112), cause the system to:
receive temporal rainfall measurements from the rainfall data receiver (114), seismic accelerogram data from the earthquake data receiver (116), and lithological parameters from the geological data receiver (118);
process the received data by routing rainfall measurements to the distributed lag nonlinear model processor (102) for calculating cumulative exposure effects and routing seismic data to the Newmark processor (104) for displacement calculations;
determine trigger probability values based on the processed outputs from both processors using logistic regression with distributed lag terms;
generate landslide trigger probability estimates by combining cross-basis functions from the distributed lag nonlinear model processor (102) with displacement values from the Newmark processor (104) in the hybrid integration module (110), wherein the probability estimates are adaptively calculated based on temporal lag relationships and spectrum-compatible artificial accelerograms, thereby identifying whether landslides are triggered by rainfall, earthquake, or combined effects; and
transmit the integrated probability estimates to at least one display output (120) for spatial visualization of landslide trigger maps.
2. The system as claimed in claim 1, wherein the distributed lag nonlinear model processor (102) is configured to process rainfall data with temporal lags.
3. The system as claimed in claim 1, wherein the Newmark processor (104) generates artificial accelerograms that are spectrally matched to recorded ground motion parameters.
4. The system as claimed in claim 1, wherein the hybrid integration module (110) employs logistic regression with distributed lag terms for combining model outputs.
5. The system as claimed in claim 1, wherein the memory unit (106) comprises random access memory for storing temporal datasets and geological parameters.
6. The system as claimed in claim 1, wherein the network module (108) interfaces with external precipitation databases and seismic data repositories through application programming interfaces.
7. A method (500) for predicting landslide triggers, the method comprising:
receiving (502) temporal rainfall measurements at the rainfall data receiver (114), seismic accelerogram data at the earthquake data receiver (116), and lithological parameters at the geological data receiver (118);
transmitting (504) the received temporal rainfall measurements, seismic accelerogram data and lithological parameters from the respective receivers (114, 116, 118) to the control unit (112) for centralized processing;
accessing external databases through the network module (108) to supplement received data;
routing (506) the temporal rainfall measurements and lithological parameters from the control unit (112) to the distributed lag nonlinear model processor (102) and simultaneously routing the seismic accelerogram data and lithological parameters from the control unit (112) to the Newmark processor (104);
processing (508) the rainfall measurements in the distributed lag nonlinear model processor (102) to calculate cross-basis functions representing cumulative rainfall exposure effects with temporal lag relationships;
simultaneously processing (510) the seismic accelerogram data in the Newmark processor (104) to generate spectrum-compatible artificial accelerograms and compute displacement values;
transferring (512) the calculated cross-basis functions from the distributed lag nonlinear model processor (102) and the computed displacement values from the Newmark processor (104) to the hybrid integration module (110);
executing (514) logistic regression with distributed lag terms in the hybrid integration module (110) by combining the cross-basis functions with the displacement values;
generating (516) adaptive landslide trigger probability estimates from the integrated model in the hybrid integration module (110), wherein the estimates identify rainfall-triggered, earthquake-triggered, or combined-effect landslides;
transmitting (518) the probability estimates from the hybrid integration module (110) to the display output (120) for spatial visualization; and
displaying (520) spatial landslide trigger probability maps on the display output (120) for decision support.
8. The method as claimed in claim 7, wherein the step of calculating cross-basis functions comprises generating main basis functions and lag basis functions.
9. The method as claimed in claim 7, wherein the step of computing displacement values comprises calculating critical acceleration based on factor of safety.
10. The method as claimed in claim 7, further comprising performing spatial cross-validation using field-verified landslide data.