A flood simulation method and system that couples underlying surface and river channel dynamics

Abstract

This invention discloses a flood simulation method and system that couples the dynamic characteristics of the underlying surface and the river channel, relating to the fields of watershed hydrological simulation and disaster prevention and mitigation. The method includes: multi-source data acquisition and preprocessing; prediction of underlying surface evolution and hydrophysical parameterization; inversion of river channel dynamic roughness based on water-sediment coupling simulation and inverse Manning's formula; coupling the future runoff generation parameter field and dynamic roughness sequence to a distributed hydrological model for simulation; and quantitatively decoupling the contributions of underlying surface and river channel changes to flood characteristics through multi-scenario comparative experiments. This invention overcomes the shortcomings of existing technologies, such as the disconnect between underlying surface and river channel processes and the static nature of roughness parameters, achieving full-chain dynamic coupling simulation. It significantly improves the physical realism and prediction accuracy of flood simulation under complex environments, and is particularly suitable for data-scarce high-altitude watersheds.

Description

Technical Field

[0001] This invention relates to the field of watershed hydrological simulation and disaster prevention and mitigation technology, and in particular to a flood simulation method and system that couples the underlying surface and the dynamic characteristics of the river channel. Background Technology

[0002] With the intensification of global climate change and human activities (such as high-altitude ecological restoration projects and land-use changes), the flood evolution characteristics of watersheds have undergone profound changes. Existing flood simulation and risk assessment methods have the following main shortcomings when dealing with complex high-altitude watersheds: The underlying surface is disconnected from river channel evolution: Traditional flood simulations typically focus on changes in the land surface (such as the impact of vegetation restoration on infiltration) or changes in river hydraulic properties (such as the impact of riverbed evolution on flow velocity). However, in actual physical processes, changes in runoff generation on the underlying surface and the confluence evolution of the river channel are intertwined, and existing models struggle to quantitatively describe the combined impact of their synergistic effects on flood processes within a unified framework.

[0003] Staticization of channel resistance parameters: Traditional confluence simulations often use static Manning roughness, which ignores its dynamic characteristics as it changes with flow, water depth, sediment erosion and deposition, and vegetation cover during floods. Especially in high-altitude cold basins or complex river network areas where data is scarce, the lack of detailed measured topographic data can lead to significant deviations in the simulation of flood evolution velocity and peak value, making it impossible to accurately reproduce the flood evolution process. These are crucial for flood forecasting of downstream reservoirs and river channels.

[0004] The physical mechanisms of the simulation process are weakened: the output of existing prediction models (such as simple land use prediction models) is often difficult to directly and effectively couple to the physical confluence process, resulting in flood response research under environmental change scenarios remaining at the statistical level and lacking a refined characterization based on physical mechanisms. Summary of the Invention

[0005] To address the problems of existing flood simulation and risk assessment methods in dealing with complex mountainous watersheds, such as the disconnect between underlying surface and channel evolution, static channel resistance parameters, and weakened physical mechanisms in the simulation process, this invention proposes a flood simulation method and system that couples the dynamic characteristics of the underlying surface and the channel. By constructing a unified framework that integrates underlying surface evolution prediction, channel dynamic roughness inversion, and distributed hydrophysical simulation, it achieves a realistic simulation of the entire dynamic coupling process of "land surface runoff generation-channel confluence" and quantitatively analyzes the contributions of different driving factors, thereby significantly improving the physical realism and prediction accuracy of flood simulation under complex environmental disturbances.

[0006] This application discloses a flood simulation method that couples the underlying surface with the dynamic characteristics of the river channel, including the following steps: S1. Obtain the raw data of the target watershed and preprocess it to generate basic data with a unified spatial reference frame. The raw data includes topographic data, land use / cover data, meteorological and hydrological data, and spatial socio-economic and spatial location data. S2. Based on the processed historical land use and land cover data, conduct land use expansion analysis, predict and generate future land use and land cover scenario data, and based on the preset land use type and hydrophysical parameter mapping rules, transform the future land use and land cover scenario data into a spatially distributed future runoff parameter field. S3. A hydrodynamic model is used to perform forward simulation of water-sediment coupling. Based on the Manning formula, the dynamic roughness of the river channel is inverted through real-time hydraulic variables. According to the preset characteristic duration triggering mechanism, asynchronous inversion is performed in combination with the physical feedback of the current hydrodynamic state, and the generated characteristic roughness values ​​are fed back to the confluence constraint parameters for real-time updates. S4. Integrate the future runoff parameter field generated in S2 and the dynamic channel roughness sequence obtained by inversion in S3 into the physical simulation engine of the distributed hydrological model, and couple them to obtain a full-scale distributed hydrological model. S5. Establish a controlled simulation experiment with multiple scenarios, run the full-scale distributed hydrological model of S4, extract flood characteristic indicators for each scenario, and quantitatively calculate the contribution of underlying surface changes and channel roughness changes to flood characteristics based on the control variable method.

[0007] Preferably, the preprocessing in S1 includes: The digital elevation model is used to calculate depression filling, flow direction and runoff accumulation, extract key geomorphic parameters, and discretize them into a computational grid with a specified spatial resolution. Coordinate system 1 and resampling were performed on multi-period historical land use and land cover raster data; Spatial interpolation is performed on meteorological station data to generate a spatiotemporal continuous field; Spatial processing of socioeconomic and locational data is used to construct driving factors for predicting underlying surface evolution.

[0008] Preferably, the mapping rule between land use type and hydrophysical parameters in S2 is as follows: based on the land use type of the grid unit, and coupled with the topographic index and soil type attribute of the grid unit, runoff generation physical parameters, including saturated hydraulic conductivity and vegetation interception capacity, are dynamically assigned to each grid unit through a preset cross-attribute lookup table or empirical function.

[0009] Preferably, S3 includes: S31. Water-sediment coupling forward simulation: Using a hydrodynamic model based on the HLL format, under set boundary conditions, perform coupled calculations of hydrodynamics, sediment transport, and riverbed deformation, and synchronously update and output water level, flow velocity, and riverbed topography. S32, Feature Duration Trigger: Set a feature time period. When the cumulative duration of the coupled simulation reaches the feature time period, roughness inversion is triggered. S33. Roughness inversion during characteristic time period: At the triggering time, extract the hydraulic radius, energy slope and flow velocity calculated from the latest riverbed topography, substitute them into the inverse operation form of the Manning formula, and calculate the equivalent comprehensive roughness value of each river grid in the current characteristic time period. S34. Roughness physical constraint and feedback update mechanism: Apply physical constraints to the equivalent comprehensive roughness value obtained by inversion, update the characteristic roughness value obtained by inversion, and if the characteristic roughness value exceeds the physical constraint range, initiate a weighted correction based on physical consistency, and feed the corrected characteristic roughness value back to the hydrodynamic model as the initial resistance boundary condition for the next stage of calculation.

[0010] Preferably, S4 includes: By constructing a mapping function between land use / cover type and hydrophysical parameters, the future runoff generation parameter field is input into the distributed hydrological model in file format; The dynamic roughness sequence is input in real time into the physical confluence operator of the distributed hydrological model, replacing the static Manning coefficient.

[0011] Preferably, the multiple control simulation experiments described in S5 include: Baseline scenario: using current land use and static channel roughness; Scenario of underlying surface change: using future land use and static channel roughness; River channel roughness variation scenario: using current land use and future dynamic roughness; Comprehensive change scenario: adopting future land use and future dynamic roughness.

[0012] Preferably, the quantitative calculation of contribution in S5 includes: The contribution ratio of the underlying surface to the peak flow:

[0013] in, For the peak flow, This is a simulated peak flood flow under a scenario of underlying surface change. The simulated peak flow rate is based on the baseline scenario. To simulate peak flow under a comprehensive coupled scenario, It is a simulated peak flow under the scenario of varying river channel roughness; The contribution of channel roughness to the advancement of peak time:

[0014] in, This represents the peak time shift caused by the dynamic evolution of river channel roughness; negative values ​​indicate an earlier peak and positive values ​​indicate a later peak. The simulated peak time under the scenario of varying river channel roughness. The simulated peak time is the baseline scenario.

[0015] It can also calculate its proportion relative to the overall change.

[0016]

[0017] in, The contribution weight of the change in river channel roughness relative to the total shift in peak time under the comprehensive scenario. It is the simulated peak time under a comprehensive coupled scenario.

[0018] This application also discloses a flood simulation system that couples the dynamic characteristics of the underlying surface and the river channel, used to implement the above-mentioned flood simulation method that couples the dynamic characteristics of the underlying surface and the river channel, including: The data preprocessing module is used to acquire and process multi-source data to build a unified spatial reference frame and physical property matrix; The underlying surface prediction and parameterization module is used to learn transformation rules based on historical LULC data, generate future LULC scenarios, and transform them into a distributed runoff generation parameter field. The dynamic roughness inversion module is used to run the water-sediment coupled hydrodynamic model and invert and update the dynamic channel roughness based on the Manning formula and the characteristic duration triggering mechanism. The coupled simulation engine module is used to integrate the distributed runoff generation parameter field and the dynamic channel roughness sequence to perform distributed hydrophysical simulation. The scenario analysis and decoupling module is used to design and run multiple sets of controlled simulation experiments, extract flood characteristic indicators, and quantitatively analyze the contribution of each driving factor.

[0019] Preferably, the dynamic roughness inversion module is specifically used to perform hydrodynamic calculations based on the HLL format, adopt a heterogeneous time step strategy to synchronously simulate the flow and riverbed deformation process, and at preset characteristic time trigger points, use real-time updated river channel geometric parameters and kinematic parameters to perform roughness inversion and feedback updates.

[0020] Preferably, the coupled simulation engine module uses the BTOP model as the core physics simulation engine.

[0021] The beneficial effects of this invention are: (1) A complete physical mechanism of land-river bidirectional coupling simulation has been realized: For the first time, land use evolution prediction, river dynamic roughness inversion and distributed hydrophysical model are deeply integrated, overcoming the defect of the two being disconnected in the traditional method and improving the physical reality of the simulation process.

[0022] (2) It breaks through the limitations of static river resistance parameters: the HLL-DRM method realizes the dynamic updating of roughness with water and sediment processes, which significantly improves the simulation accuracy of flood wave propagation speed and flood peak morphology, and provides a critical forecast period for flood prevention and early warning.

[0023] (3) Provides a quantitative impact decoupling analysis tool: Through multi-scenario control experiments and the control variable method, it is possible to clearly and quantitatively separate the respective contributions of underlying surface changes and river roughness changes to flood characteristics, providing a precise basis for engineering benefit assessment and risk management.

[0024] (4) Enhanced applicability to areas with scarce data: By using readily available observation data to dynamically invert and compensate for complex terrain information, the method’s dependence on high-precision mapping data is reduced, and the application boundaries of high-precision flood simulation technology in remote areas are expanded. Attached Figure Description

[0025] Figure 1 This is a flowchart of a flood simulation method that couples the dynamic characteristics of the underlying surface and the river channel according to an embodiment of the present invention. Figure 2 This is a schematic diagram showing the location of the study area in an embodiment of the present invention; Figure 3 This illustrates the daily flow variations under different model scenarios in embodiments of the present invention. Figure 4 A comparison of flood peak time and amplitude in typical events according to embodiments of the present invention; Figure 5 This represents the distribution of the maximum peak flow difference under different scenarios in this embodiment of the invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided with reference to the accompanying drawings and embodiments.

[0027] One embodiment of this application discloses a flood simulation method that couples the underlying surface and the dynamic characteristics of the river channel, the process of which is as follows: Figure 1 As shown, it specifically includes the following S1-S5.

[0028] S1, Multi-source data acquisition and preprocessing, aims to build a unified spatial reference framework and physical property matrix for subsequent calculations.

[0029] Topographic analysis: Obtain the watershed digital elevation model (DEM), and extract key geomorphic parameters such as river network topology, sub-watershed boundaries, slope, aspect and topographic index through filling depressions, flow direction and runoff accumulation calculation, and discretize them into a computing grid with a specified spatial resolution (such as 1km grid).

[0030] Land use / cover (LULC) data processing: acquire multi-period historical LULC raster data of the watershed, perform coordinate system I, resample to the simulation cell resolution, and establish a land use classification system (e.g., forest, grassland, cultivated land, construction land, water area, bare land, etc.).

[0031] Meteorological and hydrological data: Collect time-series data on precipitation, temperature, evaporation, etc., from meteorological stations (or grid points) in and around the basin. Use spatial interpolation methods (such as Thiessen polygons and Kriging interpolation) to generate a spatiotemporal continuous field covering the entire basin from the station data. Collect measured water level and flow process data from major hydrological control stations, perform quality control, remove outliers, and appropriately interpolate missing data.

[0032] Socioeconomic development and spatial location data: Collect and spatialize socioeconomic data (population density, GDP, nighttime light index, etc.) and location elements (transportation network, urban distance, nature reserve boundaries, etc.), calculate the characteristic distance of each grid unit, and construct the driving force factors for predicting the evolution of the underlying surface.

[0033] S2, Prediction of underlying surface evolution scenarios and hydrophysical parameterization, transforms land use change trends into a quantitative, physically meaningful distributed runoff generation parameter field.

[0034] Land use expansion analysis: Based on processed historical multi-period LULC data, an evolutionary model based on cellular automata (CA) logic is used to learn historical LULC transformation rules and establish the expansion potential of each land use type.

[0035] Future scenario generation: Combining the quantified driving factors in S1, based on the expansion potential of various land uses, the transition probability between types is calculated using the Markov chain model to generate land use demand for specific future years (such as 2040 and 2050). Then, the CA model is run to perform spatial iterative simulation to generate a future LULC spatial distribution map.

[0036] Physical mapping of runoff generation parameters: Addressing the shortcomings of traditional methods that simply assign fixed parameter values, resulting in weak physical meaning, this embodiment constructs and applies a multi-level physical parameter lookup rule. This rule not only considers land type but also couples the grid's topographic index and soil type attributes. Through a pre-defined cross-attribute lookup table or empirical function, it dynamically assigns more physically realistic parameters such as saturated hydraulic conductivity and vegetation interception capacity to each grid, thereby generating a future runoff generation parameter field with high spatial heterogeneity and clear physical meaning. This improvement achieves a high-fidelity conversion from qualitative land type to quantitative hydrophysical parameters.

[0037] The method proposed in this embodiment breaks through the traditional simple category assignment method. By associating the multi-attribute features of pixels, it transforms the evolved underlying surface information into a distributed hydrological parameter field with clear physical meaning, which can cover key parameters such as nonlinear saturated hydraulic conductivity and vegetation interception capacity field, thereby providing a high-precision physical constraint basis for land surface runoff simulation.

[0038] S3. Dynamic Roughness Inversion of the Channel (HLL-DRM). A spatiotemporal dynamic roughness parameter inversion method is established that varies with flow conditions and channel state. This method substantially optimizes the application of existing hydrodynamic models and Manning's formula. The specific process is as follows: Physical inversion principle: Based on Manning's formula in hydraulics, a dynamic roughness inversion model is constructed. This model utilizes hydraulic variables (flow velocity) acquired or simulated in real time. Hydraulic radius Energy slope The roughness coefficient is calculated in real time using the Manning formula:

[0039] S31. Hydrodynamic and Sediment Coupled Forward Simulation. Using a hydrodynamic model based on the HLL numerical scheme, under given hydraulic boundaries, heterogeneous time-step calculations of "one-step hydrodynamics, multi-step sediment transport, and riverbed deformation" are performed, with water level output synchronously. ) and flow rate ( The updated riverbed topography and vegetation cover distribution are used to construct a four-dimensional dynamic physical field. This strategy significantly improves the efficiency and numerical stability of long-term coupled computations by differentiating between fast-changing (water flow) and slow-changing (riverbed) processes, while ensuring full simulation of key physical processes.

[0040] S32. Characteristic Duration Triggering Mechanism. To avoid numerical noise and computational redundancy caused by time-step inversion, an asynchronous triggering mechanism based on simulation duration is constructed. That is, after completing the calculation of a preset "characteristic time period" (such as containing several water-sediment coupling steps), a roughness inversion is initiated uniformly to capture the comprehensive resistance state within that time period. When the process reaches a preset node, the resistance state of the characteristic time period is captured (to avoid numerical oscillations), and the current instantaneous cross-sectional geometric parameters (corrected by riverbed scouring and deposition) and kinematic parameters are automatically retrieved.

[0041] S33. Roughness Inversion During Feature Time Periods. At the triggering moment, the algorithm extracts the hydraulic radius calculated in real-time from the latest riverbed topography. With Nengpo Combined with flow rate Substituting the values ​​into the Manning inverse operation form (Formula (1)), the equivalent comprehensive roughness value of each channel grid in the current characteristic period is calculated. This value inherently includes the influence of riverbed scouring and deposition deformation on resistance. The key is that the geometric parameters used in the inversion are updated in real time with scouring and deposition, so that the inverted roughness value inherently includes the resistance effect of dynamic changes in channel morphology, breaking through the limitations of the fixed cross-section assumption.

[0042] S34. Roughness physical constraints and feedback update mechanism: Physical boundary constraints: The equivalent comprehensive roughness obtained by inversion is subject to physical constraints based on the material properties of the watershed channel (e.g., for high-altitude cold source areas, it is limited to between 0.042 and 0.065), resulting in characteristic roughness values. .

[0043] Outlier weighted correction: If the inverted value deviates from the above physical constraint range, a weighted correction procedure based on physical consistency is initiated: by calculating the average resistance level of the current flow condition under the historical flood sequence of the same magnitude, a weight allocation algorithm (e.g., assigning 70% weight to the historical experience value and 30% weight to the current inverted value) is used for smoothing to eliminate numerical pseudo-solutions caused by instantaneous hydraulic fluctuations and ensure the numerical stability of long-term simulations.

[0044] Feedback-driven logic: The corrected characteristic roughness value is used as the output at the current moment and injected into the underlying resistance matrix of the hydrodynamic solver in real time as the initial resistance boundary condition for the next calculation period.

[0045] Key control parameters: The system strictly adheres to physical stability conditions during execution, with the CFL coefficient (Courant-Friedrichs-Lewy) set to 0.9, the maximum roughness limit set to 0.055, and the minimum water depth calculation threshold set to 0.001m.

[0046] The aforementioned method, named HLL-DRM, refers to a method for identifying drag based on instantaneous flow field data output in HLL format, performing discontinuous feedback iterations. This process forms a closed-loop adaptive correction flow of "simulation-inversion-update-resimulation". This mechanism transforms roughness from a static input parameter into an endogenous variable dynamically determined by the system state. The characteristic roughness values ​​are then used as confluence constraint parameters to update the flood simulation framework, simulating the flood evolution state of the watershed at specific stages of its evolution.

[0047] S4, full-scale distributed hydrological model coupling, aims to integrate medium- and long-term underlying surface evolution parameters with short-term channel dynamic roughness parameters into a distributed physical simulation framework based on physical mechanisms, so as to achieve full-scale, high-fidelity simulation from watershed runoff generation to channel confluence.

[0048] Physical simulation engine selection: A distributed hydrological model with complete physical mechanisms (such as the BTOP model) was chosen as the core simulation engine. The core advantage of this engine lies in the high physical realism of its runoff calculations, effectively overcoming the limitations of traditional models in terms of time-scale adaptability. Its computational architecture supports both long-term hydrological process simulations with a daily timescale and refined flood evolution simulations with an hourly timescale, maintaining excellent computational accuracy and numerical stability.

[0049] Embedding of runoff generation parameter field: The spatially distributed parameter field (such as the saturated hydraulic conductivity of grid cells and vegetation interception capacity) generated by S2, which reflects the future underlying surface state, is input into the runoff generation calculation of the simulation engine in file format through the mapping function between the land use / cover type (LULC) and hydrophysical parameters, to characterize the long-term evolution of watershed infiltration capacity and runoff generation mechanism caused by land use change.

[0050] Dynamic Driving of Confluence Parameter Field: The dynamic channel roughness sequence, obtained by inverting S3 using the HLL-DRM method and varying over time, is coupled in real-time to the engine's physical confluence operator, replacing the static Manning coefficient. This allows the confluence calculation to respond to instantaneous changes in channel resistance, accurately simulating the dynamic adjustment of flood wave velocity and peak morphology. This modification is not a simple parameter replacement, but rather a change in the resistance expression paradigm of the confluence process. It enables the model to respond to instantaneous changes in channel resistance, thereby accurately simulating the adjustment process of flood wave velocity and peak morphology caused by dynamic changes in resistance, achieving a substantial functional enhancement to the traditional static parameter hydrological model.

[0051] The full-scale distributed hydrological model of this embodiment is obtained by selecting the physical simulation engine, embedding the runoff generation parameter field, and dynamically driving the coupling of the confluence parameter field.

[0052] S5. Multi-scenario flood simulation and contribution decoupling analysis: Through controlled variable experiments, the sensitivity and response characteristics of floods to environmental changes are quantitatively analyzed.

[0053] This embodiment establishes a controlled simulation experiment including at least the following four scenarios: Baseline Scenario (BL): Using current land use and static channel roughness.

[0054] Underlying surface change scenario (LUC): using future land use and static channel roughness.

[0055] River channel roughness variation scenario (RRC): using current land use and future dynamic roughness.

[0056] Integrated Coupled Scenario (COMB): Employing future land use and future dynamic roughness.

[0057] Simulation and Feature Extraction: Under the same meteorological data, the full-scale distributed hydrological model coupled with S4 was run to extract flood hydrographs for each scenario. Changes in key characteristic indicators were quantified, including peak flow (…). ), time of arrival of flood peak ( ), total flood volume ( )wait.

[0058] Quantitative decoupling of contribution: Based on the principle of the control variable method, the contribution of each key characteristic indicator is quantitatively calculated.

[0059] The contribution of the underlying surface to the peak flow can be estimated using the following formula:

[0060] in, For the peak flow, This is a simulated peak flood flow under a scenario of underlying surface change. The simulated peak flow rate is based on the baseline scenario. To simulate peak flow under a comprehensive coupled scenario, The simulated peak flow rate is given under the scenario of varying river roughness.

[0061] The contribution of channel roughness to the earlier peak time can be directly reflected by the difference in peak time between the RRC and BL scenarios:

[0062] in, This represents the peak time shift caused by the dynamic evolution of river channel roughness; negative values ​​indicate an earlier peak and positive values ​​indicate a later peak. The simulated peak time under the scenario of varying river channel roughness. The simulated peak time is the baseline scenario.

[0063] It can also calculate its proportion relative to the overall change.

[0064]

[0065] in, The contribution weight of the change in river channel roughness relative to the total shift in peak time under the comprehensive scenario. It is the simulated peak time under a comprehensive coupled scenario.

[0066] Another embodiment of this application discloses a flood simulation system that couples the underlying surface with the dynamic characteristics of the river channel, for implementing the above simulation method, including: The data preprocessing module is used to acquire and process multi-source data to build a unified spatial reference frame and physical property matrix; The underlying surface prediction and parameterization module is used to learn transformation rules based on historical LULC data, generate future LULC scenarios, and transform them into a distributed runoff generation parameter field.

[0067] The dynamic roughness inversion module is used to run a water-sediment coupled hydrodynamic model and invert and update the dynamic channel roughness based on the Manning formula and a characteristic time-triggered mechanism. Specifically, the dynamic roughness inversion module performs hydrodynamic calculations based on the HLL format, employs a heterogeneous time-step strategy to synchronously simulate the flow and riverbed deformation processes, and performs roughness inversion and feedback updates at preset characteristic time-triggered points using real-time updated channel geometric parameters and kinematic parameters.

[0068] The coupled simulation engine module integrates the distributed runoff generation parameter field and the dynamic channel roughness sequence to perform distributed hydrophysical simulation. In this embodiment, the coupled simulation engine module uses the BTOP model as the core physical simulation engine.

[0069] The scenario analysis and decoupling module is used to design and run multiple sets of controlled simulation experiments, extract flood characteristic indicators, and quantitatively analyze the contribution of each driving factor.

[0070] In a specific embodiment, a flood response evolution simulation was conducted in a high-altitude cold-climate watershed. Taking a source area (the watershed above Gangtuo station) as an example, the 2040 LULC scenario was generated using the land use cover dataset (CLCD) and the CA land use prediction method. A dynamic roughness model (HLL-DRM) was constructed based on the HLL format hydrodynamic solver and Manning's formula inversion, with the BTOP model selected as the coupling engine. Results after implementation according to the above S1-S5 process showed that the Nash efficiency coefficient (NSE) of the time-scale flood process simulation reached 0.86-0.89 at key hydrological stations, verifying the reliability of the method at a fine scale. The LUC scenario reduced peak flow by 5-15%, exhibiting a "peak-shaving" effect; the RRC scenario advanced the peak flow by 2-3 hours; the COMB integrated scenario simulation revealed the complex phenomenon of "overall runoff suppression" and "local peak amplification," demonstrating the unique value of the coupling framework in identifying systemic risks. Specifically, as follows... Figure 2 As shown, the study area covers approximately 16.1 × 10⁴ km², with an elevation ranging from 2600 m to 6500 m. Technical inputs include integrated multi-source data on meteorology, hydrology, topography (30 m DEM), land use (CLCD dataset), soil, and socioeconomic data. Figure 2 In the middle (a), the geographical location of the u study area is shown. Figure 2 (b) shows the annual average temperature distribution. Figure 2 (c) represents the annual average precipitation distribution.

[0071] By establishing a simulated scenario (control experimental group), this embodiment quantitatively characterizes the contribution of different physical factors to the flood response through four types of comparative experiments: Baseline Scenario (BL): The current land use and river roughness are used as the historical verification and comparison baseline.

[0072] Underlying Surface Evolution Scenario (LUC): Input future land use data predicted by the PLUS model, keep river parameters constant, and analyze the impact of land surface changes in isolation.

[0073] River Characteristics Evolution Scenario (RRC): Input the future dynamic roughness parameters derived from HLL-DRM inversion, keep the underlying surface unchanged, and analyze the influence of confluence velocity.

[0074] Integrated Coupled Scenario (COMB): Simultaneously inputting the predicted underlying surface and dynamic roughness parameters, we explore the nonlinear coupling effect of the two factors.

[0075] The simulation results are as follows: Improved model simulation accuracy, such as Figure 3As shown, the daily-scale NSE is 0.68 (calibrated) for ZMD station and 0.70 (calibrated) for GT station; the hourly-scale NSE is 0.86 for ZMD station and 0.89 for GT station, indicating that the model proposed in this application can accurately capture the flood peak process on the hourly scale.

[0076] In the impacts of land use change (LUC scenario), peak flow reduction is 5%–15%. Due to vegetation restoration and enhanced soil infiltration, total runoff initially decreases, exhibiting a "peak shaving and valley filling" effect. However, in subsequent stages, as vegetation communities stabilize and soil structure improves, the proportion of groundwater runoff increases, eventually restoring the total flow to a new equilibrium state.

[0077] The impact of changes in channel roughness (RRC scenario) Figure 4 As shown, the reduction in channel roughness significantly accelerates the propagation speed of flood waves. Reduced roughness implies decreased channel resistance, which fundamentally alters the kinematic characteristics of flood waves. Due to minimized energy loss caused by friction, flood waves maintain a higher energy concentration, resulting in steeper hydrological hydrographs and earlier peak arrival times. Under the RRC effect, peak flows at key hydrological stations occurred approximately 2-3 hours earlier. This conclusion demonstrates the advantages of this application in refining (hour-scale) flood risk prediction.

[0078] Coupled nonlinear effects (COMB scenario) such as Figure 5 As shown, Figure 5 In the middle (a), the intensity of land use change (LUC-BL) is represented. Figure 5 In the middle (b), the roughness variation intensity (RRC-BL) is represented. Figure 5 In the middle (c), the nonlinear interaction intensity is represented. When both factors act together, the flood volume and propagation velocity become "decoupled". Although the overall runoff is suppressed by the underlying surface, the accelerated confluence of the river channel causes the originally staggered tributary floods to overlap synchronously in specific river sections, resulting in a nonlinear phenomenon of local flood peak amplification.

[0079] In summary, this application is the first to couple the land use prediction model (PLUS), the dynamic roughness model (HLL-DRM), and the distributed hydrological model (BTOP) to achieve dynamic joint simulation of runoff generation and confluence processes, improve physical realism, provide a flood attribution analysis method for dual land-river regulation, and support refined flood control decision-making.

[0080] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A flood simulation method coupling underlying surface and river channel dynamic characteristics, characterized in that, Includes the following steps: S1. Obtain the raw data of the target watershed and preprocess it to generate basic data with a unified spatial reference frame. The raw data includes topographic data, land use / cover data, meteorological and hydrological data, and spatial socio-economic and spatial location data. S2. Based on the processed historical land use and land cover data, conduct land use expansion analysis, predict and generate future land use and land cover scenario data, and based on the preset land use type and hydrophysical parameter mapping rules, transform the future land use and land cover scenario data into a spatially distributed future runoff parameter field. S3. A hydrodynamic model is used to perform forward simulation of water-sediment coupling. Based on the Manning formula, the dynamic roughness of the river channel is inverted through real-time hydraulic variables. According to the preset characteristic duration triggering mechanism, asynchronous inversion is performed in combination with the physical feedback of the current hydrodynamic state, and the generated characteristic roughness values ​​are fed back to the confluence constraint parameters for real-time updates. S4. Integrate the future runoff parameter field generated in S2 and the dynamic channel roughness sequence obtained by inversion in S3 into the physical simulation engine of the distributed hydrological model, and couple them to obtain a full-scale distributed hydrological model. S5. Establish a controlled simulation experiment with multiple scenarios, run the full-scale distributed hydrological model of S4, extract flood characteristic indicators for each scenario, and quantitatively calculate the contribution of underlying surface changes and channel roughness changes to flood characteristics based on the control variable method.

2. The flood simulation method for coupling underlying surface and river channel dynamic characteristics according to claim 1, characterized in that, The preprocessing described in S1 includes: The digital elevation model is used to calculate depression filling, flow direction and runoff accumulation, extract key geomorphic parameters, and discretize them into a computational grid with a specified spatial resolution. Coordinate system 1 and resampling were performed on multi-period historical land use and land cover raster data; Spatial interpolation is performed on meteorological station data to generate a spatiotemporal continuous field; Spatial processing of socioeconomic and locational data is used to construct driving factors for predicting underlying surface evolution.

3. The flood simulation method for coupling underlying surface and river channel dynamic characteristics according to claim 2, characterized in that, The mapping rule between land use type and hydrophysical parameters described in S2 is as follows: based on the land use type of the grid cell, and coupled with the topographic index and soil type attribute of the grid cell, runoff generation physical parameters, including saturated hydraulic conductivity and vegetation interception capacity, are dynamically assigned to each grid cell through a preset cross-attribute lookup table or empirical function.

4. The flood simulation method for coupling underlying surface and river channel dynamic characteristics according to claim 3, characterized in that, S3 includes: S31. Water-sediment coupling forward simulation: Using a hydrodynamic model based on the HLL format, under set boundary conditions, perform coupled calculations of hydrodynamics, sediment transport, and riverbed deformation, and synchronously update and output water level, flow velocity, and riverbed topography. S32, Feature Duration Trigger: Set a feature time period. When the cumulative duration of the coupled simulation reaches the feature time period, roughness inversion is triggered. S33. Roughness inversion during characteristic time period: At the triggering time, extract the hydraulic radius, energy slope and flow velocity calculated from the latest riverbed topography, substitute them into the inverse operation form of the Manning formula, and calculate the equivalent comprehensive roughness value of each river grid in the current characteristic time period. S34. Roughness physical constraint and feedback update mechanism: Apply physical constraints to the equivalent comprehensive roughness value obtained by inversion, update the characteristic roughness value obtained by inversion, and if the characteristic roughness value exceeds the physical constraint range, initiate a weighted correction based on physical consistency, and feed the corrected characteristic roughness value back to the hydrodynamic model as the initial resistance boundary condition for the next stage of calculation.

5. The flood simulation method for coupling underlying surface and river channel dynamic characteristics according to claim 4, characterized in that, S4 includes: By constructing a mapping function between land use / cover type and hydrophysical parameters, the future runoff generation parameter field is input into the distributed hydrological model in file format; The dynamic roughness sequence is input in real time into the physical confluence operator of the distributed hydrological model, replacing the static Manning coefficient.

6. The flood simulation method for coupling underlying surface and river channel dynamic characteristics according to claim 5, characterized in that, The multiple controlled simulation experiments described in S5 include: Baseline scenario: using current land use and static channel roughness; Scenario of underlying surface change: using future land use and static channel roughness; River channel roughness variation scenario: using current land use and future dynamic roughness; Comprehensive change scenario: adopting future land use and future dynamic roughness.

7. The flood simulation method for coupling underlying surface and river channel dynamic characteristics according to claim 6, characterized in that, The quantitative calculation of contributions in S5 includes: The contribution ratio of the underlying surface to the peak flow: in, For the peak flow, This is a simulated peak flood flow under a scenario of underlying surface change. The simulated peak flow rate is based on the baseline scenario. For simulating peak flood flow under a comprehensive coupled scenario; The contribution of channel roughness to the advancement of peak time: in, This represents the peak time shift caused by the dynamic evolution of river channel roughness; negative values ​​indicate an earlier peak and positive values ​​indicate a later peak. The simulated peak time under the scenario of varying river channel roughness. The simulated peak time is the baseline scenario; The contribution of channel roughness variation to the total peak time offset under the combined scenario: in, The contribution weight of the change in river channel roughness relative to the total shift in peak time under the comprehensive scenario. It is the simulated peak time under a comprehensive coupled scenario.

8. A flood simulation system coupling underlying surface and river channel dynamic characteristics, characterized in that, A flood simulation method for implementing the coupled underlying surface and river channel dynamic characteristics as described in any one of claims 1-7, comprising: The data preprocessing module is used to acquire and process multi-source data to build a unified spatial reference frame and physical property matrix; The underlying surface prediction and parameterization module is used to learn transformation rules based on historical LULC data, generate future LULC scenarios, and transform them into a distributed runoff generation parameter field. The dynamic roughness inversion module is used to run the water-sediment coupled hydrodynamic model and invert and update the dynamic channel roughness based on the Manning formula and the characteristic duration triggering mechanism. The coupled simulation engine module is used to integrate the distributed runoff generation parameter field and the dynamic channel roughness sequence to perform distributed hydrophysical simulation. The scenario analysis and decoupling module is used to design and run multiple sets of controlled simulation experiments, extract flood characteristic indicators, and quantitatively analyze the contribution of each driving factor.

9. The flood simulation system for coupling underlying surface and river channel dynamics according to claim 8, characterized in that, The dynamic roughness inversion module is specifically used to perform hydrodynamic calculations based on the HLL format. It adopts a heterogeneous time step strategy to synchronously simulate the flow and riverbed deformation process, and at preset characteristic time trigger points, it uses real-time updated river channel geometric parameters and kinematic parameters to perform roughness inversion and feedback updates.

10. The flood simulation system for coupling underlying surface and river channel dynamics according to claim 9, characterized in that, The coupled simulation engine module uses the BTOP model as the core physics simulation engine.

Images