Flood-landslide coupling simulation method considering mutual feedback mechanism and parameter uncertainty

Abstract

This invention discloses a flood-landslide coupled simulation method considering feedback mechanisms and parameter uncertainties, relating to the field of natural disaster simulation technology. The method includes: performing iterative calculations of a hydrological model and a landslide model within a preset time step; generating a landslide state index indicating the landslide state of a target area based on the hydrological parameters output by the hydrological model within the time step; dynamically updating one or more runoff generation parameters in the hydrological model for calculation in the next time step in response to the landslide state index indicating a landslide in the target area; and dynamically updating one or more confluence parameters in the hydrological model for calculation in the next time step in response to the landslide state index indicating a landslide in the target area. By constructing a complete bidirectional feedback closed loop, the simulation accuracy and physical realism can be significantly improved, and the reliability of the prediction results can be enhanced.

Description

Technical Field

[0001] This invention relates to the field of natural disaster simulation technology, and in particular to a flood-landslide coupled simulation method that considers feedback mechanisms and parameter uncertainties. Background Technology

[0002] Floods and landslides are two interconnected and highly destructive natural disasters. Floods are often caused by heavy rainfall, snowmelt, reservoir failures, or ice blockage, leading to a rapid increase in the volume of rivers and lakes, a surge in water levels exceeding the carrying capacity of river channels or dams, and ultimately, flooding. This process not only directly threatens people's lives and property, inundates farmland and buildings, and damages infrastructure such as transportation and communication, but also, due to the strong scouring and prolonged soaking, damages mountain structures, inducing landslides. Therefore, in the field of disaster simulation, coupled models are commonly used to assess the risk of the flood-landslide disaster chain. Most of these models are based on a one-way coupling mechanism, that is, using the calculation results of hydrological models, such as rainfall, runoff, or soil moisture content, as input to the landslide model to calculate and predict slope stability, thereby determining whether rainfall will induce landslides.

[0003] However, the above schemes neglect the two-way interaction between the two. In actual physical processes, landslides significantly alter surface characteristics. For example, landslides may bury the original surface, or they may block rivers, hindering normal flow processes. Current one-way coupling models fail to capture this two-way physical interaction, resulting in incomplete physical mechanisms and biased simulation results that do not accurately reflect the dynamic evolution of the entire disaster chain, thus affecting the accuracy and reliability of predictions. Summary of the Invention

[0004] The main objective of this invention is to provide a flood-landslide coupled simulation method that considers mutual feedback mechanisms and parameter uncertainties, in order to solve the problem of low simulation accuracy caused by the lack of a two-way feedback mechanism between landslides and flood runoff generation and confluence processes.

[0005] To achieve the above objectives, this application provides a flood-landslide coupled simulation method considering mutual feedback mechanisms and parameter uncertainties, the method comprising: Within a preset time step, iterative calculations are performed on the hydrological model and the landslide model; Within the time step, based on the hydrological parameters output by the hydrological model, a landslide state index indicating the landslide state of the target area is generated by the landslide model. In response to the landslide state index indicating that a landslide has occurred in the target area, one or more runoff generation parameters in the hydrological model used for the calculation of the next time step are dynamically updated. In response to the landslide state index indicating that a landslide has occurred in the target area, one or more confluence parameters in the hydrological model used for the calculation of the next time step are dynamically updated.

[0006] This application also provides a flood-landslide coupled simulation device, comprising: The iterative calculation module is used to perform iterative calculations of the hydrological model and the landslide model within a preset time step; The status assessment module is used to generate landslide status indicators indicating the landslide status of the target area based on the hydrological parameters output by the hydrological model within the time step. The runoff feedback module is used to dynamically update one or more runoff parameters in the hydrological model for the calculation of the next time step in response to the landslide state index indicating that a landslide has occurred in the target area. The confluence feedback module is used to dynamically update one or more confluence parameters in the hydrological model for the calculation of the next time step in response to the landslide state index indicating that a landslide has occurred in the target area.

[0007] In another aspect, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, causes the processor to perform the steps of the method described in the first aspect.

[0008] This application provides a flood-landslide coupled simulation method considering feedback mechanisms and parameter uncertainties. It involves iterative calculations of a hydrological model and a landslide model within a preset time step. Within this time step, based on the hydrological parameters output by the hydrological model, a landslide state index indicating the landslide state of a target area is generated by the landslide model. In response to the landslide state index indicating a landslide in the target area, one or more runoff generation parameters in the hydrological model used for calculation in the next time step are dynamically updated. Finally, in response to the landslide state index indicating a landslide in the target area, one or more confluence parameters in the hydrological model used for calculation in the next time step are also dynamically updated.

[0009] This application has the following beneficial effects: 1. Improved simulation accuracy and physical realism: By introducing a two-way feedback mechanism between landslides and runoff generation and confluence, a more complete closed loop of the physical process is constructed. This enables dynamic and collaborative simulation of the entire flood-landslide disaster chain evolution, resulting in a higher degree of agreement between simulation results and measured data. 2. Enhanced reliability of prediction results: By preferentially adopting probability-based calculation methods, the uncertainty of model input parameters can be effectively addressed. This upgrades traditional deterministic judgments to quantitative assessments of landslide occurrence probabilities, providing more scientific quantitative information for disaster risk management and emergency decision-making. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] in: Figure 1 A flowchart illustrating a flood-landslide coupled simulation method considering feedback mechanisms and parameter uncertainties, provided for an embodiment of this application; Figure 2 A flowchart of an AFORM algorithm based on HLRF iteration provided for an embodiment of this application; Figure 3 A scatter plot of predicted and measured flow rates provided in an embodiment of this application; Figure 4 A schematic diagram illustrating the change of the objective function during the optimization process, provided as an embodiment of this application; Figure 5 This application provides a schematic diagram of landslide probability assessment under different parameter combinations as an embodiment of the present application. Figure 6 This is a schematic diagram of a flood-landslide coupling simulation device provided in an embodiment of this application. Detailed Implementation

[0012] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0013] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0014] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0015] The embodiments of this application are described below with reference to the accompanying drawings.

[0016] Please see Figure 1 This is a flowchart illustrating a flood-landslide coupled simulation method considering feedback mechanisms and parameter uncertainties, provided in an embodiment of this application. Figure 1 As shown, the method includes: 101. Perform iterative calculations of the hydrological model and the landslide model within a preset time step.

[0017] In this embodiment of the application, the subject executing the method can be a flood-landslide coupling simulation device, which can be an electronic device in practical applications.

[0018] The hydrological model in this embodiment can simulate the runoff generation and confluence processes of rainfall within a watershed. Specifically, based on the rainfall data and underlying surface conditions in the input data, the hydrological state of each spatial unit (e.g., a grid) at different time points can be calculated. Key hydrological parameters output, such as slope soil moisture content and pore water pressure, will serve as inputs to drive the landslide model.

[0019] The landslide model in this embodiment can assess slope stability. Specifically, it can receive hydrological parameters from a hydrological model and, combined with topographic and geological information from the input data, calculate landslide state indicators for the target area to determine whether a landslide has occurred.

[0020] In this embodiment, the time step can be set as needed. This time step is the basic time unit for iterative calculation of the model, for example, it can be set to 1 hour. This allows simulation of entering an iterative calculation loop with a time step period. Within each time step t, subsequent steps are executed.

[0021] Hydrological models can acquire rainfall data at the current time step and the watershed state (such as soil moisture content) stored at the end of the previous time step, and perform hydrological calculations. As an optional implementation, a distributed hydrological model can be used, which divides the study area into multiple grid cells. For each grid cell, the model calculates the change in soil moisture content. After calculation, the hydrological model can output a set of hydrological parameters, where the key parameter is the average soil moisture content of each grid cell.

[0022] Specifically, the hydrological model can simulate the spatiotemporal variations of water volume and energy flux in each grid within a watershed. Its core components include runoff generation and runoff concentration. Precipitation is first intercepted by the vegetation canopy, and the remaining precipitation is assessed based on infiltration capacity to determine whether there is excess runoff. Besides the excess runoff, precipitation infiltrating into the soil is divided into runoff segments according to a three-layer soil generalization model, taking into account soil evapotranspiration. For each grid, precipitation is segmented according to infiltration capacity, and the infiltration capacity curve is shown below: (1) (2) (3) in, i For grid infiltration capacity; The maximum infiltration capacity of the grid is a function of the maximum water storage capacity of the soil; A For the first i The proportion of the flow-generating area of ​​each grid; b i These are the shape parameters of the curve; This refers to the amount of water that seeps in. Maximum water storage capacity; This represents the average moisture content of the three soil layers. The clean rain that falls onto the soil surface; R This is net rainfall.

[0023] 102. Within the above time step, based on the hydrological parameters output by the above hydrological model, a landslide state index indicating the landslide state of the target area is generated through the above landslide model.

[0024] In this embodiment, hydrological parameters (especially soil moisture content) output by the hydrological model can be used as the main input to the landslide model. In this embodiment, the landslide model can be a deterministic model based on limit equilibrium theory, such as an infinite slope model. This model can calculate the stability safety factor of the slope. F s To assess its stability.

[0025] The hydrological model outputs soil moisture content, which is used as input to the landslide model. The landslide model in this embodiment can be controlled using a safety factor. F s The formula for indicating whether slope slippage has occurred is as follows:

[0026] In the formula, For effective soil viscosity; Effective stress; Angle of internal friction; This is the coefficient relating soil stress and matrix suction. The weight of the soil slope per unit width; Slope; For soil air pressure; - Matrix suction. Safety factor. F s The value is a classification value: a value greater than 1 indicates that the slope is stable; a value equal to 1 indicates that it is in a critical state; and a value less than 1 indicates that the slope is unstable.

[0027] Calculated safety factor F s This is the landslide state index in this embodiment.

[0028] Furthermore, a landslide assessment can be conducted. The assessment criteria can be: if... F s If the value is less than 1, then the slope of the grid cell is considered unstable and a landslide has occurred; if... F s If the value is ≥1, the slope is considered to be in a stable or critically stable state, and no landslide has occurred.

[0029] If the judgment result is negative (i.e., all grids), F s If ≥ 1), it means that no new landslides have occurred in the current time step. The process will jump directly to the next time step t+1 and return to the step of performing hydrological model calculations. The original hydrological model parameters will be used to continue the calculation.

[0030] Accordingly, if the judgment result is yes (i.e., at least one grid is...), F s If <1), it means a landslide event has occurred. At this time, the update of the execution feedback parameters can be triggered, including the update of the production and confluence parameters, that is, steps 103 and 104 can be executed.

[0031] 103. In response to the landslide state index indicating that a landslide has occurred in the target area, dynamically update one or more runoff generation parameters in the hydrological model used for the calculation of the next time step.

[0032] Specifically, this step updates runoff parameters to simulate the impact of landslides on surface runoff conditions. Landslides form landslide bodies and may cause the original vegetation cover to be replaced by exposed rock and soil, creating areas with reduced runoff area, which significantly alters the runoff characteristics of that area.

[0033] Optionally, one or more of the following updates may be performed: Based on the area affected by the landslide, the effective runoff area of ​​the above hydrological model is reduced; Introducing an infiltration capacity attenuation factor to reduce the maximum soil infiltration capacity in the landslide area; A shape correction factor is introduced to reduce the shape parameter of the flow process curve.

[0034] Specifically, after a landslide occurs, the slope characteristics change significantly. Firstly, the runoff-generating area changes; landslide burial or deposition reduces the effective runoff-generating area of ​​the slope, updating to: (5) In the formula, A 0 represents the original runoff area; The area of ​​runoff generated after the landslide; The area affected by the landslide is approximated by the area of ​​a circle, with the radius of the circle being 50% of the grid length.

[0035] Secondly, after a landslide, the destruction of the surface structure and the accumulation of debris generally lead to a decrease in infiltration capacity, which can be expressed by the following formula: (6) In the formula: and These represent the infiltration capacity before and after a landslide, respectively. This is the infiltration capacity attenuation factor, representing the ratio of the landslide value to the original value. It can be set according to different situations, such as: 0.5 when the surface is covered by sediment, 0.8 when the slope is exposed by mudflow, and 0.2 when the accumulation zone is blocked by mud and rocks. A conservative value can be taken as 0.5.

[0036] also, b i Reflecting the steepness of the runoff curve, after a landslide, the slope roughness and loose structure lead to a faster, more concentrated, and steeper runoff process. Therefore, generally... b i It will become smaller, calculated as follows: (7) In the formula: and The steepness of the runoff curves before and after the landslide is shown, respectively. This is a correction factor; the smaller the value, the steeper the runoff and the faster the response. It can be further refined based on slope and landslide type. For steep slopes and debris flow type landslides... The value is 0.5; for gentle slopes and shallow landslides, The value is 0.8.

[0037] 104. In response to the landslide state index indicating that a landslide has occurred in the target area, dynamically update one or more confluence parameters in the hydrological model used for the calculation of the next time step.

[0038] Specifically, this step can update the confluence parameters to simulate the blocking effect caused by a landslide entering the river channel. In this embodiment, this effect can be mitigated by introducing a landslide blocking coefficient. λ ( t ) to quantify. Optionally, one or more of the following updates can be performed: A landslide blocking coefficient is introduced, and the runoff flow rate at the next time step is corrected using the landslide blocking coefficient. Landslides or collapses into river channels can block flow paths, forming barrier lakes. Introducing a blocking coefficient can address this issue. λ ( t To correct the flow rate: (8) In the formula: Q 0( t () represents the flow rate under conditions without landslides; The flow rate under landslide conditions; The landslide blocking coefficient is calculated as follows: (9) Where: when the landslide volume V slide ( t ) Exceeding the cross-sectional volume of the river channel ( A river ( t) and H river ( t The product of ), where A river ( t) The cross-sectional area of ​​the river channel. H river ( t (This refers to the depth of the river channel). A value close to 0 indicates complete blockage. The landslide volume is represented by the product of the landslide's affected area and the corresponding soil layer thickness.

[0039] Based on the above process, the hydrological model and the landslide model are synchronously iterated with a unified time step, and the slope water content, runoff, landslide criteria and feedback parameters are transmitted in real time to form a closed loop of two-way coupled dynamic calculation of flood-landslide.

[0040] After completing steps 103 and 104, all updated runoff generation and confluence parameters are integrated into feedback parameters, which are sent back to the hydrological model at time step t+1 when the landslide occurs. Upon entering the next time step t+1, the hydrological model will use the updated model parameters to calculate flood runoff generation and confluence.

[0041] Through the aforementioned iterative cycle, this embodiment constructs a dynamic closed-loop system. Changes in hydrological processes (such as increased soil moisture content) may trigger landslides, and the occurrence of landslides will, in turn, instantly and quantitatively alter the boundary conditions and core parameters of the hydrological model, thereby affecting the subsequent flood evolution process. This two-way coupling mechanism enables the simulation results to reflect real physical phenomena such as increased runoff after a landslide, earlier flood peaks, and higher peak values, thus significantly improving the accuracy and reliability of the simulation.

[0042] Steps 103 and 104 can be performed in any order.

[0043] As an optional implementation, this embodiment provides a coupled simulation method based on probability assessment, building upon the aforementioned embodiments. The main difference in this embodiment lies in the assessment method of landslide state, which aims to address the inherent uncertainties in geophysical and hydrological parameters within the model, thereby providing more scientifically sound and valuable prediction results.

[0044] The main difference in this embodiment lies in the internal implementation of the landslide model, which can integrate a probability calculation unit to handle parameter uncertainties and calculate the instability probability.

[0045] The overall process of the method can be partially referred to Figure 1 The steps of the illustrated embodiment differ, but there are differences in the specific execution of landslide model calculation and landslide judgment.

[0046] Figure 2 A flowchart of an AFORM algorithm based on HLRF iteration is provided for an embodiment of this application.

[0047] Specifically, when the landslide model receives hydrological parameters, it no longer simply calculates a deterministic safety factor, but instead initiates a probabilistic calculation process, which may include: First, the aforementioned slope stability model based on a deterministic safety factor is transformed into a probabilistic model aimed at calculating the probability of slope instability. This transformation can be achieved by defining a limit state function. g ( X This is implemented using ), and the function is typically defined as: (10) (11) in, X This represents a set of basic random variables that affect slope stability. X = ( X 1, X 2, …, X n ) TIn this embodiment, these random variables may include the effective cohesion of the soil, the effective internal friction angle, etc. Since these parameters are spatially variable and difficult to measure precisely, it is more reasonable to treat them as random variables. They can usually be assumed to follow a specific probability distribution, such as a normal distribution or a log-normal distribution, whose mean and standard deviation can be determined based on regional geological survey data or empirical data. g ( X ) is the limit state function; F S ( X ) represents the functional form of the landslide model; f ( X ) is a random variable X The joint probability distribution function. When g ( X When ) < 0, it indicates that the slope is in a failure state (i.e. F s <1); when g ( X When )>0, it indicates that the slope is in a safe state; g ( X If ) = 0, it represents the limit state surface. P f The probability of slope instability is often difficult to calculate. FORM provides the following approximate solution method: (12) In the formula: Ф(•) is the standard normal cumulative distribution function; As a reliability index, representing the shortest distance from the origin to the limit state surface in standard normal space, it can be calculated through the following process: make Z = g ( X ) = 0, x * = ( x 1*, x 2*, …, x n *)T is a point on the limit state surface, i.e.: g ( x *)=0, at point x Expanding the function using Taylor series and taking down to the first-order term, we have: (13) Utilizing the properties of linear combinations of mutually independent normally distributed random variables, Z The mean and standard deviation are as follows: (14) (15) Then, the reliability index is obtained: (16) set up X i The standardized random variable is Y i ,Right now: After substituting into equation (13), we divide by equation (15) and then substitute into equation (16), and after simplification, we get: (17) Define variables from the above formula X i The sensitivity coefficient is: (18) Equation (17) can be written as: (19) In the original X In space x *Corresponding to standard normal random variable Y Points in space y * indicates the verification point. Equation (19) represents the result of... y The limit state surface of *. Based on geometric relationships, the verification point is located at... Y The coordinates in space are: (20) Reverse calculation X The coordinates in space are: (twenty one) The above calculations repeatedly involve converting to standard normal space during iteration, which reduces computational efficiency. To address this, a fast recursive algorithm (Hasofer-Lind-Rackwitz-Fiessler, HLRF) is introduced. This method can directly convert the original normal space... X Iterative solution of the most likely failure point in space x i * It can efficiently handle cases involving relevant non-normal random variables, and its recursive formula is as follows: (twenty two) (twenty three) (twenty four) In the formula: It is a diagonal matrix; for x k First i The equivalent normal standard deviation of a random variable;R The correlation matrix of random variables; T k It is an orthogonal transformation matrix; It is the equivalent normal mean of the random variable; In order to be in x k The gradient vector of the limit state function at the given point.

[0048] When the iteration converges to the most likely failure point x i *At that time, the new reliability index β f It can be calculated using the following formula: (25) Through the above calculations, new calculations are continuously iterated. x *, until the two times before and after. x If the difference is less than the allowable error, output the current value. β f The value is calculated using equation (12) to determine the instability probability (e.g., ...). Figure 2 (As shown). After all grid points in the study area have completed the calculations according to the above steps, the final landslide probability spatial distribution map is output.

[0049] In one alternative implementation, another embodiment provides an alternative method for updating confluence parameters by directly modifying topographic data to simulate the impact of landslides on the river confluence process.

[0050] The system architecture and overall process of this embodiment are largely similar to those of the aforementioned embodiments, but it employs a different technical solution in implementing the step of updating the bus parameters. Specifically, it may include: 1. Calculate the geometric parameters of the landslide body: When the landslide model determines that a landslide has occurred, it estimates the volume of the landslide body, V_slide. Based on the extent of the landslide's influence, it can then identify the set of grid cells where the landslide body accumulates within the river channel and calculate the total accumulation area, A_dep. 2. Calculate the average accumulation thickness: Based on the volume and area, calculate the average accumulation thickness ΔH = V_slide / A_dep of the landslide body in the river channel.

[0051] 3. Update DEM data: For each affected grid, read its original elevation value H_old stored in the input data, and then update it with a new elevation value H_new = H_old + ΔH. This update operation can be performed directly on the DEM data in memory.

[0052] 4. Implicit Update of Confluence Path: After modifying the DEM data, the update of the confluence parameters is essentially complete. In this scheme, the confluence parameters are not specific coefficients, but rather the terrain data itself, because terrain is the most fundamental parameter determining the confluence path and flow velocity.

[0053] Subsequently, the process proceeds to the next time step, t+1, for calculation.

[0054] In the new time step t+1, when the hydrological model is invoked to perform confluence calculations, it will operate based on the latest, modified DEM data. The hydrological model's confluence algorithm (such as the D8 algorithm) will recalculate the flow direction for each grid based on the new elevation data. Since the grid elevation in the landslide accumulation area has been raised, forming a de facto "dam," the confluence algorithm will naturally calculate the following effects: the upstream flow path will be blocked; the flow will converge in front of the "dam," leading to a rise in upstream water levels and an expansion of the inundation area; if the risen water level exceeds a saddle in the surrounding terrain, the flow may even change course, forming a new confluence path; the downstream flow will be significantly reduced, or even completely stopped, due to the upstream interception.

[0055] This embodiment, by directly manipulating underlying topographic data, can simulate the landslide-dammed river effect in a more mechanistic and visual way. It not only reflects changes in flow but also simulates the formation, evolution, and inundation impact of landslide-dammed lakes on upstream areas, providing a solid data and physical foundation for further assessment of secondary disaster chains such as the risk of lake outburst. While this approach may be computationally more complex, the physical realism and information richness of its simulation results far surpass those of coefficient-based methods, making it particularly suitable for scenarios requiring high-precision simulation of landslide-dammed rivers and their subsequent evolution.

[0056] To more clearly illustrate the method of the embodiments of this application, the results are verified and analyzed below in conjunction with experimental data.

[0057] Figure 3 This application provides a scatter plot of predicted and measured flow rates.

[0058] To quantify the simulation accuracy, particularly considering the feedback from landslides to hydrological processes, after taking into account the two-way coupling mechanism, and to clarify the correlation between simulation and measured results, a scatter plot was drawn between all measured and simulated flood disaster points in the study area. For example... Figure 3As shown, it is clear that the correlation coefficient of the model before considering the two-way coupling mechanism is 0.70, while the correlation coefficient after considering it rises to 0.78, and the data points are closer to the red dashed line, indicating that the difference between the predicted and measured values ​​has decreased. This means that the constructed two-way coupling model possesses a stronger physical mechanism and simulation accuracy. Future research could further optimize the model parameters or introduce more influencing factors to further improve the model's prediction accuracy.

[0059] Figure 4 This is a schematic diagram illustrating the changes of an objective function during the optimization process, as provided in an embodiment of this application.

[0060] like Figure 4 As shown, this illustrates the comparison between simulated and measured flows in a hydrological-landslide coupled model, considering the feedback effect of landslides on hydrological processes. The uncertainty band is constructed using multiple simulations with different representative parameter combinations during the landslide-feedback flood process, aiming to reflect the model's simulated response range under different landslide feedback intensities. Figure 4 As can be seen, compared to the original model without considering landslide feedback, the uncertainty bands obtained after adding the feedback mechanism not only clearly show the variation range of the model simulation results, but also significantly improve the model's ability to capture flood processes. Especially under certain parameter combinations, the model's simulation of the flow process and the measured flow are significantly better matched, and the fitting effect between the peak flood time and the peak flow is closer to the measured values. This phenomenon indicates that landslides, as an important geomorphic process, have an undeniable impact on rainfall-runoff processes during extreme rainfall events. By introducing the feedback effect of landslides on changes in underlying surface runoff generation and runoff paths, the model can more realistically reflect the dynamic characteristics of flood processes in complex mountainous watersheds, thus effectively capturing phenomena such as sudden increases in flow and earlier or later peak floods caused by topographic changes. Furthermore, the uncertainty bands also reveal the model's sensitivity to parameter selection and the potential error range, providing a scientific basis for subsequent model optimization, parameter calibration, and risk decision-making.

[0061] in addition, Figure 5 This is a schematic diagram illustrating a landslide probability assessment under different parameter combinations, provided as an embodiment of this application.

[0062] from Figure 5The ROC curves show that the probability of landslides varies depending on the combination of parameters in the coupled model. Among the 50 sets of parameters tested, the accuracy of the parameters from the 18th to the 28th set continuously improved, from 0.59 to 0.75, and finally stabilized. This indicates that as the coupled model samples in the parameter space, it can update the predicted state in real time, thereby continuously improving its discriminative ability. Moreover, the accuracy of the landslide probability predicted by the model is obviously higher under certain parameter combinations, further illustrating the importance of considering parameter uncertainty.

[0063] In summary, this application creatively constructs a hydrological-landslide coupled model considering a two-way feedback mechanism. Based on the original model, it considers the feedback correction of runoff generation conditions by landslides and the blocking effect of landslides on the confluence process, transmitting slope water content, runoff, landslide criteria, and feedback parameters in real time, forming a two-way coupled calculation framework for flood-landslide. The results show that in long-sequence tests, the AUC value of flood forecast accuracy exceeds 0.8, and the landslide forecast accuracy exceeds 0.7, proving that the model has high accuracy in the coupled forecast results of flood and landslide disasters. From a spatiotemporal distribution perspective, the flood and landslide risk assessment results are largely consistent with the location and time of actual disasters, further verifying the reliability of the simulation results. The correlation coefficient between the measured and simulated results of the model before considering the two-way coupling mechanism was 0.70, while the correlation coefficient after considering it increased to 0.78. By selecting different representative parameter combinations to form uncertainty bands, the variation range of the model simulation results is clearly shown, and the model's ability to capture flood processes is significantly improved, making it more consistent with the measured flood hydrograph. Moreover, the accuracy of the simulated landslide probability assessment under different parameter combinations is further improved, which once again proves that the constructed model has high accuracy.

[0064] Based on the description of the foregoing method embodiments, this application also provides a flood-landslide coupled simulation device.

[0065] Figure 6 This is a schematic diagram of a flood-landslide coupled simulation device provided in an embodiment of this application. Figure 6 As shown, the flood-landslide coupled simulation device 600 includes: The iterative calculation module 610 is used to perform iterative calculations of the hydrological model and the landslide model within a preset time step; The state assessment module 620 is used to generate landslide state indicators indicating the landslide state of the target area based on the hydrological parameters output by the hydrological model within the time step. The runoff feedback module 630 is used to dynamically update one or more runoff parameters in the hydrological model for the calculation of the next time step in response to the landslide state index indicating that a landslide has occurred in the target area. The confluence feedback module 640 is used to dynamically update one or more confluence parameters in the hydrological model for the calculation of the next time step in response to the landslide state index indicating that a landslide has occurred in the target area.

[0066] It is understood that the relevant content concerning each module in the above-mentioned device has been described in detail in the foregoing method embodiments, and specific details can be found in the method embodiments; that is, the flood-landslide coupling simulation device provided in this application can perform the following... Figure 1 Any steps in the illustrated embodiments will not be described in detail here.

[0067] Based on the description of the foregoing method and apparatus embodiments, this application also provides typical watershed implementation examples and parameter setting examples, which may specifically include: 1. Watershed selection and data preparation A pilot watershed of a first-order tributary of the Xiaojiang River in a certain county was selected as a demonstration area. The watershed covers an area of ​​3.8 km², with an elevation of 1850–2450 m, a main channel length of 3.2 km, and an average slope of 18°. Historically, a rainstorm on August 22, 2016 triggered three shallow landslides, forming barrier lakes. The measured data are complete and meet the verification requirements.

[0068] DEM: 1 m × 1 m LiDAR data, resampled to 10 m × 10 m grids, totaling 380 × 140 grids (denoted as N=53200), with an elevation error ≤0.2 m; Land use: Secondary classification vector map, Manning n value is obtained according to the following conversion relationship, where forest land n=0.065, cultivated land n=0.035, and settlement n=0.028; Table 1 is a land use-Manning n correspondence table provided in the embodiments of this application.

[0069]

[0070] Table 1 Soil type: 39 boreholes were drilled on site and classified into two categories according to the original formula (4): "gravelly clay" and "silty loam". The results showed that: Gravelly clay: c′=18 kPa, φ′=24°, θs=0.45, θr=0.08, Ks=2.3×10 -6 m / s; Silty loam: c′=9 kPa, φ′=31°, θs=0.41, θr=0.05, Ks=4.1×10 -6 m / s; The spatial distribution of the two types of soil is assigned according to the grid to form c′ and φ′ raster layers, which are used as inputs for the original formula (4); Rainfall: An automatic rain gauge station is set up at the outlet of the basin with a time resolution of 10 min. In this embodiment, the 40-hour sequence from 14:00 on August 21, 2016 to 06:00 on August 23, 2016 is extracted. The cumulative rainfall is 132 mm and the maximum rainfall intensity in 1 hour is 38 mm. This is directly used as the Psoil input in the original formula (3). Initial conditions: The TDR soil moisture observation at 14:00 on August 21, 2016 was used as the initial value. After gridding, W (t=0) was obtained, which satisfies the W variable in the original formula (3).

[0071] 2. Model running steps (refer to) Figure 1 (The process should be adjusted accordingly) The time step Δt = 1 h, with a total of 40 steps, is as follows: Figure 1 The loop executes steps 101–104 in the specified range. a) Hydrological model runoff generation and confluence calculation Using the aforementioned “three-layer soil generalization + Green-Ampt infiltration” scheme, each grid independently runs equations (1)–(3) to output W(t), Psoil(t), and R(t) at that moment; the confluence part uses the D8 algorithm, Manning n is assigned according to the above land use, and the slope is calculated in real time from the 10 m DEM to obtain Q0(t) (flow rate without landslide). b) The probability assessment of the landslide model uses W(t) as the pore water pressure uw input, combined with the c′ and φ′ grids, and calculates the deterministic safety factor Fs according to the original equation (4); c′ and φ′ are regarded as normal random variables, the mean is the above grid value, and the standard deviation is taken as σc′=2 kPa and σφ′=3° according to the literature. Figure 2The HLRF algorithm iteration shown (convergence threshold |Δβ|<0.01, maximum 30 times) yields Pf(t); in the demonstration, the probability threshold Pf≥0.7 is set to determine a landslide, consistent with the original text "Pf as a landslide state index". c) Update the runoff parameters for the grid with Pf≥0.7, according to equation (5)–(7): based on the landslide radius r=0.5×10 m=5 m, the circle area Aslide=πr²=78.5 m², substitute into equation (5) to get A′=A0–Aslide; the landslide type in this watershed is shallow scour type, according to the original text "when the slope surface is exposed by mudflow scour αi=0.8", directly substitute into equation (6) to get i′m=0.8 im; the slope is >15°, according to the original text "αb=0.5", substitute into equation (7) to get b′i=0.5 bi. d) For the confluence parameter update, for the cells that are simultaneously located in the river channel grid and have Pf≥0.7, the landslide thickness Hslide=1.2 m (average value measured on site) and volume Vslide=Aslide×Hslide=94.2 m³ are used; the product of the river channel cross-sectional area (measured 8.4 m²) and the river channel depth (measured 1.1 m) is 9.24 m³, which is substituted into equation (9) to get λ(t)=0.90; the flow rate is corrected according to equation (8): Q(t)=λ(t) Q0(t)=0.90 Q0(t); at the same time, the DEM elevation is updated according to the original text: Hnew=Hold+ΔH, where ΔH=Vslide / Adep, Adep is the actual accumulation area of ​​the landslide falling into the river channel (measured on site 60 m²), so ΔH=1.57 m, directly replacing the corresponding grid elevation for the next time step D8 to recalculate the flow direction.

[0072] 3. Result Verification After running with the above parameters, the simulated flood peak occurred at 18:00 on August 22, 2016, with a peak flow of 52.7 m³ / s; the actual measured flood peak occurred at 19:00, with a peak flow of 55.3 m³ / s. Landslide Location: The simulated high-probability area (Pf>0.7) spatially overlaps with two of the three measured landslide locations. The third landslide is located outside the watershed boundary (not included in the calculation), satisfying the condition... Figure 5 The requirement is "AUC > 0.7"; The simulated length of the backwater of the landslide dammed lake was 420 m, while the actual measured length was 380 m. The simulation results are consistent with the measured flow process line trend, with a correlation coefficient of 0.78, which meets the requirements. Figure 3 The accuracy improvement pattern is shown; the spatial overlap between high-probability landslide areas and measured disaster points is better than that of... Figure 5 The level is AUC>0.7.

[0073] Using the specific watershed, specific observation values, and relevant formulas mentioned in this application, technicians can fully reproduce the "two-way feedback" process in this application.

[0074] In one embodiment, a computer-readable storage medium is also provided, which stores a computer program that, when executed by a processor, causes the processor to perform any of the steps in the above method embodiments.

[0075] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0076] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0077] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A flood-landslide coupled simulation method considering feedback mechanisms and parameter uncertainties, characterized in that, The method includes: Within a preset time step, iterative calculations are performed on the hydrological model and the landslide model; Within the time step, based on the hydrological parameters output by the hydrological model, a landslide state index indicating the landslide state of the target area is generated by the landslide model. In response to the landslide state index indicating that a landslide has occurred in the target area, one or more runoff generation parameters in the hydrological model used for the calculation of the next time step are dynamically updated. In response to the landslide state index indicating that a landslide has occurred in the target area, one or more confluence parameters in the hydrological model used for the calculation of the next time step are dynamically updated.

2. The flood-landslide coupled simulation method considering mutual feedback mechanisms and parameter uncertainties according to claim 1, characterized in that, The dynamic updating of one or more runoff generation parameters in the hydrological model used for the calculation of the next time step includes one or more of the following: Based on the area affected by the landslide, the effective runoff area of ​​the hydrological model is reduced; Introducing an infiltration capacity attenuation factor to reduce the maximum soil infiltration capacity in the landslide area; A shape correction factor is introduced to reduce the shape parameter of the flow process curve.

3. The flood-landslide coupled simulation method considering mutual feedback mechanisms and parameter uncertainties according to claim 2, characterized in that, The value of the infiltration capacity attenuation factor and the value of the shape correction factor are dynamically determined according to the type of landslide.

4. The flood-landslide coupled simulation method considering mutual feedback mechanisms and parameter uncertainties according to claim 1, characterized in that, The dynamic updating of one or more confluence parameters in the hydrological model used for the calculation of the next time step includes: A landslide blocking coefficient is introduced, and the runoff flow rate at the next time step is corrected using the landslide blocking coefficient.

5. The flood-landslide coupled simulation method considering mutual feedback mechanisms and parameter uncertainties according to claim 4, characterized in that, The landslide blocking coefficient is calculated based on the ratio of the volume of the landslide body to the cross-sectional volume of the river channel at the location of the landslide body.

6. The flood-landslide coupled simulation method considering mutual feedback mechanisms and parameter uncertainties according to claim 1, characterized in that, The dynamic updating of one or more confluence parameters in the hydrological model used for the calculation of the next time step includes: Calculate the average thickness of the landslide based on its volume and area of ​​accumulation. Based on the average deposition thickness, the digital elevation model data of the river channel grid affected by the landslide is updated so that the hydrological model can recalculate the confluence path at the next time step.

7. The flood-landslide coupled simulation method considering mutual feedback mechanisms and parameter uncertainties according to claim 1, characterized in that, The process of generating landslide state indicators that indicate the landslide state of the target area using the landslide model includes: The slope stability model based on the deterministic safety factor is transformed into a probabilistic model with the objective of calculating the probability of slope instability. Based on the aforementioned probability model, an iterative algorithm based on the first-order reliability method is used to calculate the instability probability of the target area, and the instability probability is used as the landslide state index.

8. The flood-landslide coupled simulation method considering mutual feedback mechanisms and parameter uncertainties according to claim 7, characterized in that, The iterative algorithm is the Hasofer-Lind-Rackwitz-Fiessler fast recursive algorithm.

9. A flood-landslide coupled simulation device, characterized in that, include: The iterative calculation module is used to perform iterative calculations of the hydrological model and the landslide model within a preset time step; The status assessment module is used to generate landslide status indicators indicating the landslide status of the target area based on the hydrological parameters output by the hydrological model within the time step. The runoff feedback module is used to dynamically update one or more runoff parameters in the hydrological model for the calculation of the next time step in response to the landslide state index indicating that a landslide has occurred in the target area. The confluence feedback module is used to dynamically update one or more confluence parameters in the hydrological model for the calculation of the next time step in response to the landslide state index indicating that a landslide has occurred in the target area.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, the processor performs the steps of the method as described in any one of claims 1-8.

Images