Disaster risk situation generation method and system based on dynamic grading, and medium

Abstract

The embodiment of the invention provides a disaster risk situation generation method and system based on dynamic grading, and a medium, and relates to the technical field of multi-disaster chain type disaster dynamic early warning. Comprising the following steps: performing rainstorm flood peak flow deduction by taking an infiltration loss parameter array as a constraint condition and combining an instantaneous rainstorm peak time-space field and a persistent rainfall characteristic time-space field to generate a disaster-inducing potential curved surface array; performing grid-level disaster susceptibility evaluation on the area static factors of the disaster monitoring area and outputting a reference susceptibility array; and constructing a dynamic grading-probability coupling array by coupling the disaster-causing potential curved surface array and the reference susceptibility array, executing disaster risk situation prediction based on real-time rainfall data, and outputting disaster chain risk probability distribution. The problems that in the prior art, a static single disaster type early warning system is adopted for split type static early warning, rainstorm-mountain torrent-geology chain type dynamic conduction risks cannot be captured, the dynamic conduction process of chain disaster is difficult to quantify, disaster chain response lags behind, and precision is insufficient are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of multi-disaster chain dynamic early warning, and particularly relates to a disaster risk situation generation method and system based on dynamic grading and a medium. BACKGROUND

[0002] The prior art adopts a static single-disaster early warning system for fragmented static early warning, which independently evaluates rainstorm, mountain flood or geological disaster risks depending on fixed experience parameters, cannot quantify the dynamic transmission process of rainstorm triggering mountain flood and mountain flood intensifying geological secondary disasters, and ignores the real-time correction mechanism of soil moisture dynamic change on rainfall loss parameters, resulting in systematic lag in response to disaster chains.

[0003] At the same time, the spatial accuracy of the traditional model is limited to the basin scale and cannot realize grid-level risk positioning, and lacks the spatio-temporal coupling mechanism of dynamic disaster-causing factors and static disaster-brewing environment, making it difficult to capture the chain risk evolution law of rainstorm triggering mountain flood and mountain flood inducing landslide debris flow, resulting in serious lack of timeliness and accuracy of early warning, and ultimately restricting the accurate prevention and control of disaster chain risk zoning and grading.

[0004] In summary, the prior art adopts a static single-disaster early warning system for fragmented static early warning, cannot capture the rainstorm-mountain flood-geological chain dynamic transmission risk, and cannot quantify the dynamic transmission process of chain disaster, resulting in lag in response to disaster chains and insufficient accuracy.

[0005] It should be noted that the information disclosed in this BACKGROUND section is only intended to increase an understanding of the general context of the present application and should not be considered as admitting or in any form suggesting that this information forms the prior art that is already known to those skilled in the art. SUMMARY

[0006] In view of the above defects or improvement needs of the prior art, the present application provides a disaster risk situation generation method and system based on dynamic grading, which solves the problem of the prior art that the static single-disaster early warning system is used for fragmented static early warning, cannot capture the rainstorm-mountain flood-geological chain dynamic transmission risk, and cannot quantify the dynamic transmission process of chain disaster, resulting in lag in response to disaster chains and insufficient accuracy. The technical effect of achieving minute-level dynamic early warning response, spatial accurate positioning and zoning and grading prevention and control decision support for rainstorm-mountain flood-geological disaster chains is achieved, and the specific technical scheme is as follows:

[0007] According to a first aspect of the present application, a disaster risk situation generation method based on dynamic grading is provided, which comprises:

[0008] The original time series rainfall signal returned by the rain gauge station deployed in the disaster monitoring area is subjected to multi-scale decomposition to capture the instantaneous rainstorm peak spatiotemporal field and the persistent rainfall characteristic spatiotemporal field. The disaster monitoring area is gridded based on a preset grid scale to obtain a disaster monitoring grid array. The infiltration loss parameter array is constructed by performing dynamic inversion of the infiltration loss parameter in the disaster monitoring grid array. The grid-level rainstorm flood peak flow is calculated based on the infiltration loss parameter array as the constraint condition and in combination with the instantaneous rainstorm peak spatiotemporal field and the persistent rainfall characteristic spatiotemporal field to generate a disaster-causing potential surface array. The disaster monitoring area is taken as the spatial calculation carrier to perform grid-level disaster-prone area assessment on the regional static factors of the disaster monitoring area to output a reference-prone-area array. The disaster-causing potential surface array and the reference-prone-area array are coupled through Bayesian spatiotemporal covariance to construct a dynamic grading-probability coupled array. Real-time rainfall data is loaded into the dynamic grading-probability coupled array to perform grid-level disaster risk situation prediction to output a disaster chain risk probability distribution.

[0009] In an embodiment, the original time series rainfall signal returned by the rain gauge station is subjected to multi-scale decomposition to capture the instantaneous rainstorm peak spatiotemporal field and the persistent rainfall characteristic spatiotemporal field, and the following processing is further performed:

[0010] The decomposition level parameter is set according to the regional rainstorm characteristics of the disaster monitoring area. The wavelet basis function is selected according to the rainstorm disaster-causing mechanism of the disaster monitoring area. The discrete wavelet transform decomposition is performed on the original time series rainfall signal by using the wavelet basis function with the decomposition level parameter as the transform order constraint to obtain the high-frequency detail coefficient sequence and the low-frequency approximation coefficient. After the short-duration rainstorm component is reconstructed based on the high-frequency detail coefficient sequence, the local maximum value is extracted through spatial interpolation to form the instantaneous rainstorm peak spatiotemporal field. The long-duration rainfall component is reconstructed based on the low-frequency approximation coefficient, and the rainfall intensity field is generated through gridded sliding average processing as the persistent rainfall characteristic spatiotemporal field output.

[0011] In an embodiment, the grid-level rainstorm flood peak flow is calculated based on the infiltration loss parameter array as the constraint condition and in combination with the instantaneous rainstorm peak spatiotemporal field and the persistent rainfall characteristic spatiotemporal field to generate the disaster-causing potential surface array, and the following processing is further performed:

[0012] Based on the infiltration loss parameter array, the parameters of the standard rainstorm formula are dynamically corrected, and a corrected rainstorm formula array is output. The instantaneous rainstorm peak spatiotemporal field and the continuous rainfall characteristic spatiotemporal field are fused at the grid level to generate a multi-duration rainstorm intensity field array mapped to the disaster monitoring grid array. The infiltration loss of the multi-duration rainstorm intensity field array is dynamically subtracted from the corrected rainstorm formula array to generate a net rainfall intensity field array. Based on the net rainfall intensity field array, iterative reasoning of the confluence time is performed to generate a duration peak flow array. The duration peak flow array is traversed according to a preset disaster critical threshold to determine the disaster potential and construct a disaster potential array. Using the duration dimension of the duration peak flow array as the time coordinate axis, the multi-duration rainstorm intensity field array and the disaster potential array are organized into a three-dimensional structure of duration, rainstorm intensity, and disaster potential to generate the disaster potential surface array.

[0013] In one implementation, based on the net rainfall intensity field array, a convergence time iterative inference is performed to generate a time-based peak flow array, and the following processing is also performed:

[0014] The slope runoff time and gully runoff time of the generated net rainfall intensity field array are iteratively calculated to obtain a two-dimensional runoff time array; based on the two-dimensional runoff time array and the net rainfall intensity field array, the peak flow duration inference is performed to generate the duration peak flow array.

[0015] In one implementation, a dynamic tiering-probability coupled array is constructed by coupling the catastrophic potential surface array and the baseline susceptibility array using Bayesian spatiotemporal covariance, and the following processing is also performed:

[0016] The disaster potential surface array is dynamically categorized based on the disaster potential value to obtain a disaster potential categorized array; the baseline susceptibility array is statically categorized based on the susceptibility level value to obtain a susceptibility categorized array; a conditional probability judgment matrix is ​​constructed based on historical disaster data; after aligning the disaster potential categorized array and the susceptibility categorized array, the conditional probability judgment matrix is ​​loaded and the risk level probability distribution array is calculated and output; the risk level probability distribution array is probabilistically serialized to locate the dominant risk level array; the disaster potential categorized array, the susceptibility categorized array, the risk level probability distribution array, and the dominant risk level array are superimposed to generate the dynamic categorization-probability coupling array.

[0017] In one implementation, using the disaster monitoring grid array as a spatial computing carrier, a grid-level disaster susceptibility assessment is performed on the regional static factors of the disaster monitoring area, outputting a baseline susceptibility array, and the following processing is also performed:

[0018] The topographic slope distribution, lithological distribution, fault structure, and historical disaster point distribution that constitute the static factors of the region are rasterized according to the spatial resolution of the disaster monitoring grid array to generate a gridded factor array; multi-factor information weighted fusion is performed on the gridded factor array to output a comprehensive information array; the comprehensive information array is traversed using a preset susceptibility level classification rule to perform grid-level disaster susceptibility classification and output the baseline susceptibility array.

[0019] In one implementation, the infiltration loss parameters are dynamically inverted in the disaster monitoring grid array to construct an infiltration loss parameter array, and the following processing is also performed:

[0020] The interactive remote sensing satellite system obtains multispectral remote sensing data of the disaster monitoring area; based on the disaster monitoring grid array, it performs gridded inversion of soil moisture content on the multispectral remote sensing data and outputs a soil moisture content array; it maps the soil moisture content array to a soil dryness and wetness classification array; using the soil dryness and wetness classification array as a query condition, it iterates through the soil loss parameter comparison table to obtain the infiltration loss parameter array.

[0021] According to a second aspect of the present invention, a disaster risk situation generation system based on dynamic grading and classification is provided, the system comprising:

[0022] A multi-scale rainfall decomposition module is used to perform multi-scale decomposition on the raw time-series rainfall signals transmitted back from rain gauges, capturing the spatiotemporal field of instantaneous rainstorm peaks and the spatiotemporal field of continuous rainfall characteristics, wherein the rain gauges are deployed in the disaster monitoring area; a spatial gridding module is used to grid the disaster monitoring area based on a preset grid scale to obtain a disaster monitoring grid array; a dynamic inversion module for infiltration loss is used to dynamically invert infiltration loss parameters in the disaster monitoring grid array to construct an infiltration loss parameter array; a disaster-causing potential surface generation module is used to combine the spatiotemporal field of instantaneous rainstorm peaks and the characteristic spatiotemporal field of continuous rainfall with the infiltration loss parameter array as a constraint. The system employs a characteristic spatiotemporal field to estimate peak rainfall and flood flows at the grid level, generating a disaster-causing potential surface array. A static susceptibility assessment module uses the disaster monitoring grid array as a spatial computational carrier to perform grid-level disaster susceptibility assessment on regional static factors of the disaster monitoring area, outputting a baseline susceptibility array. A Bayesian disaster chain coupling module couples the disaster-causing potential surface array and the baseline susceptibility array using Bayesian spatiotemporal covariance, constructing a dynamic tiered-probability coupled array. A real-time risk prediction module loads real-time rainfall data into the dynamic tiered-probability coupled array to perform grid-level disaster risk situation prediction, outputting a disaster chain risk probability distribution.

[0023] A third aspect of the present invention discloses a computer-readable storage medium storing a computer program for performing any step of the first aspect of the present invention.

[0024] Beneficial effects of the embodiments of the present invention:

[0025] In the solution provided by this invention, the original time-series rainfall signals transmitted back from rain gauge stations are decomposed at multiple scales to capture the spatiotemporal field of instantaneous rainstorm peak and the spatiotemporal field of continuous rainfall characteristics. The rain gauge stations are deployed in the disaster monitoring area. The disaster monitoring area is gridded based on a preset grid scale to obtain a disaster monitoring grid array. Infiltration loss parameters are dynamically inverted within the disaster monitoring grid array to construct an infiltration loss parameter array. Using the infiltration loss parameter array as a constraint, and combining the instantaneous rainstorm peak spatiotemporal field and the spatiotemporal field of continuous rainfall characteristics, grid-level rainstorm peak flow is estimated to generate a disaster-causing potential surface array. Using the disaster monitoring grid array as a spatial computing carrier, grid-level disaster susceptibility assessment is performed on the regional static factors of the disaster monitoring area, outputting a baseline susceptibility array. The disaster-causing potential surface array and the baseline susceptibility array are coupled using Bayesian spatiotemporal covariance to construct a dynamic grading-probability coupling array. Real-time rainfall data is loaded into the dynamic grading-probability coupling array to perform grid-level disaster risk situation prediction, outputting a disaster chain risk probability distribution. This invention achieves the technical effect of providing minute-level dynamic early warning response, precise spatial positioning, and zoned and graded prevention and control decision support for the rainstorm-flash flood-geological disaster chain. Of course, implementing any product or method of this invention does not necessarily require achieving all of the advantages described above simultaneously. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 The diagram shows a flowchart of the disaster risk situation generation method based on dynamic classification and grading provided by the present invention.

[0028] Figure 2 A schematic diagram of the structure of the disaster risk situation generation system based on dynamic classification and grading provided by the present invention is shown.

[0029] Figure labeling: Rainfall multi-scale decomposition module 01, spatial gridding module 02, infiltration loss dynamic inversion module 03, disaster potential surface generation module 04, static susceptibility assessment module 05, Bayesian disaster chain coupling module 06, real-time risk prediction module 07. Detailed Implementation

[0030] To facilitate understanding of the present invention, a more complete description of the invention will be given below with reference to the accompanying drawings, which illustrate preferred embodiments of the invention. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein; rather, these embodiments are provided to enable a more thorough and complete understanding of the disclosure of the invention.

[0031] Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0032] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0033] Unless otherwise expressly stated, throughout the specification and claims, the term "comprising" or its variations such as "including" or "comprises" shall be understood to include the stated elements or components without excluding other elements or other components.

[0034] The present invention provides a method, system and medium for generating disaster risk situation based on dynamic classification and grading, which is used to solve the problem that the existing technology uses a static single disaster early warning system for fragmented static early warning, which cannot capture the dynamic transmission risk of rainstorm-flash flood-geological chain, and is difficult to quantify the dynamic transmission process of chain disasters, resulting in delayed response and insufficient accuracy of disaster chain.

[0035] Example 1: See Figure 1 The flowchart of the disaster risk situation generation method based on dynamic grading and classification provided in this embodiment of the invention includes:

[0036] A100: Perform multi-scale decomposition on the raw time-series rainfall signals transmitted back from the rain gauge stations to capture the spatiotemporal field of the instantaneous rainstorm peak and the spatiotemporal field of the continuous rainfall characteristics, wherein the rain gauge stations are deployed in the disaster monitoring area.

[0037] In one implementation, the original time-series rainfall signal transmitted back from the rain gauge is decomposed into multiple scales to capture the spatiotemporal field of the instantaneous rainstorm peak and the spatiotemporal field of the continuous rainfall characteristics. Step A100 may further include:

[0038] A110: Set the decomposition level parameters according to the regional rainstorm characteristics of the disaster monitoring area.

[0039] A120: Select wavelet basis functions based on the rainstorm disaster-causing mechanism in the disaster monitoring area.

[0040] A130: Using the decomposition level parameters as the transform order constraint, the original time-series rainfall signal is subjected to discrete wavelet transform decomposition using the wavelet basis functions to obtain a high-frequency detail coefficient sequence and low-frequency approximation coefficients.

[0041] A140: After reconstructing the short-duration rainstorm component based on the high-frequency detail coefficient sequence, local maxima are extracted by spatial interpolation to form the spatiotemporal field of the instantaneous rainstorm peak.

[0042] A150: Based on the low-frequency approximation coefficients, the long-duration rainfall components are reconstructed, and a rainfall intensity field is generated through gridded moving average processing, which is used as the spatiotemporal field output of the continuous rainfall characteristics.

[0043] Specifically, in this embodiment, the disaster monitoring area is a small watershed and a high-incidence area of ​​geological disasters for calculating rainstorm peaks. An interactive remote sensing satellite system and a rainfall monitoring network obtain regional rainstorm characteristics characterizing the disaster monitoring area. These regional rainstorm characteristics include the frequency of short-duration heavy rainfall and the duration of continuous rainfall. Based on these regional rainstorm characteristics, the depth of wavelet decomposition is determined to set the decomposition level parameters.

[0044] The specific decomposition level parameters need to match the typical duration distribution of rainstorms in the disaster monitoring area. For example, areas with frequent thunderstorms need a higher level to capture minute-level rainstorm peaks, while monsoon-prone areas focus on durations of more than an hour.

[0045] For different disaster-causing modes, such as landslide-sensitive areas needing to focus on short-duration peak values ​​and debris flow-prone areas focusing on cumulative rainfall, appropriate wavelet basis functions should be selected.

[0046] For example, the Db4 wavelet is suitable for capturing sudden heavy rainfall, while the Symlet wavelet is more sensitive to the characteristics of continuous rainfall. The optimal basis function needs to be determined in combination with the geological risk type.

[0047] Using the preset decomposition level parameters as the order constraint of the wavelet transform, the original rainfall time series signal is decomposed by discrete wavelet transform using the selected wavelet basis functions.

[0048] This operation separates the original signal in the time-frequency domain into a high-frequency detail coefficient sequence that reflects short-period rainstorm fluctuations, such as instantaneous intensity changes on a 10–30 minute scale, and a low-frequency approximation coefficient that characterizes long-period rainfall trends, such as cumulative processes on a 1–6 hour scale.

[0049] For example, after the raw signal of a rain gauge station for 1 hour is decomposed into 3 layers, the high-frequency coefficients D1-D3 capture 3 minute-level rainstorm pulses with peak values ​​of 35 mm / h, 42 mm / h and 38 mm / h respectively, while the low-frequency coefficient A3 outputs a gentle base rainfall trend with an average intensity of 12 mm / h.

[0050] Based on the reconstruction of short-duration rainstorm components using high-frequency detail coefficient sequences, the Kriging spatial interpolation method is preferred to transform discrete station data into a continuous spatial field. Then, the local rainfall intensity maxima of each grid cell are extracted. These maxima represent the instantaneous rainstorm peak intensity and are used for critical judgment of flash flood disasters.

[0051] For example, the reconstructed 10-minute rainstorm component shows that a certain grid cell has a rainfall intensity peak of 45mm at a specific moment. After spatial interpolation, it forms a spatiotemporal field of instantaneous rainstorm peak covering the entire disaster monitoring area. Its spatial resolution is aligned with the disaster monitoring grid and can be directly input into the flood peak flow calculation model. The flood peak flow calculation model is used to deduce rainstorm and flood peak flow to support flash flood disaster early warning.

[0052] With a time resolution set at the hour level, long-duration rainfall components are reconstructed using low-frequency approximation coefficients. Transient fluctuations are smoothed through gridded moving average processing to generate a spatiotemporal field of continuous rainfall characteristics that reflects the intensity of continuous rainfall. This spatiotemporal field of continuous rainfall characteristics quantifies the gradual impact of cumulative rainfall on soil moisture content and is the core dynamic input for geological disaster early warning.

[0053] For example, after the reconstructed 1-hour resolution rainfall field is averaged by moving average, a certain grid cell maintains an average intensity of 28 mm / h for 6 consecutive hours, and the output is a gridded continuous rainfall field. Its time smoothing characteristics are consistent with the disaster mechanism of landslides induced by long-term rainfall.

[0054] This embodiment performs multi-scale decomposition on the original time-series rainfall signals transmitted back from rain gauge stations, capturing the spatiotemporal field of instantaneous rainstorm peaks and the spatiotemporal field of continuous rainfall characteristics, achieving the technical effect of providing accurate dynamic input for subsequent calculation of flash flood disaster potential and geological disaster early warning.

[0055] A200: The disaster monitoring area is gridded based on a preset grid scale to obtain a disaster monitoring grid array.

[0056] In this embodiment, the preset grid scale is determined based on the terrain complexity and the spatial distribution density of disaster hazard points. It is used to unify spatial computing units and to grid the disaster monitoring area based on the preset grid scale to obtain a disaster monitoring grid array. This enables spatial alignment and fusion calculation of multi-source data, providing a standardized spatial benchmark for subsequent disaster potential surface generation and static susceptibility assessment.

[0057] A300: Dynamically invert the infiltration loss parameters in the disaster monitoring grid array to construct an infiltration loss parameter array.

[0058] In one implementation, the infiltration loss parameters are dynamically inverted in the disaster monitoring grid array to construct an infiltration loss parameter array. Step A300 may further include:

[0059] A310: Interactive remote sensing satellite systems acquire multispectral remote sensing data of the disaster monitoring area.

[0060] A320: Based on the disaster monitoring grid array, perform soil moisture content gridding inversion on the multispectral remote sensing data and output a soil moisture content array.

[0061] A330: Map the soil moisture content array to a soil dryness / wetness classification array.

[0062] A340: Using the soil dryness and wetness classification array as the query condition, traverse and query the soil loss parameter comparison table to obtain the infiltration loss parameter array.

[0063] Specifically, multi-band spectral data of the disaster monitoring area is acquired through a satellite remote sensing system. The band composition of the multi-band spectral data includes, but is not limited to, visible light and near-infrared bands. The multi-band spectral data carries surface reflectance information, providing a physical basis for soil moisture content inversion.

[0064] For example, the NDVI index can be calculated using the Band 8 (near-infrared) and Band 4 (red) bands of the Sentinel-2 satellite as an indicator of vegetation cover and soil moisture.

[0065] Satellite data is resampled according to the spatial resolution of the disaster monitoring grid array. The soil volumetric water content of each grid cell is calculated using an inversion algorithm for the water-sensitive band. Finally, the soil water content array corresponding to the disaster monitoring grid array is output.

[0066] For example, for a 10m×10m grid, based on the negative correlation between Band11 reflectance and soil moisture, the water content of a certain grid is found to be 22.3%. In this scenario, the thermal inertia model is preferred for the inversion algorithm.

[0067] The continuous values ​​in the soil moisture content array are mapped to discrete classification levels according to preset dryness and wetness thresholds. The preferred mapping rules are as follows: when the soil moisture content is higher than 30%, it is classified as a wet level, for example, the inversion value of 35% is classified into this level; when it is between 20% and 30%, it is classified as a moderately wet level, for example, the value of 25%; when it is lower than 20%, it is classified as a dry level, for example, the measured value of 15%. Through hierarchical mapping, the soil dryness and wetness classification array is output.

[0068] Using the soil dryness and wetness classification array as input, and combining it with the static lithological data of the grid cells, a predefined soil loss parameter comparison table is traversed and queried to obtain dynamic infiltration parameters.

[0069] For example, when the wetness level of a certain grid is medium, by cross-matching soil type and wetness state, the rainstorm loss coefficient f=1.02 and rainstorm loss index m=0.69 are directly obtained. Finally, the (f,m) parameter pairs corresponding to each grid unit are output to form a spatially full-coverage infiltration loss parameter array.

[0070] The lookup mechanism in this embodiment achieves dynamic parameter inversion through a combination mapping of soil type and moisture status, ensuring that the parameter values ​​are logically consistent with historical verification data.

[0071] This embodiment constructs an array of infiltration loss parameters, providing key soil moisture constraint parameters for the subsequent dynamic correction of the rainstorm formula, thereby indirectly achieving the technical effect of improving the accuracy of flood peak flow calculation.

[0072] A400: Using the infiltration loss parameter array as a constraint, and combining the instantaneous rainstorm peak spatiotemporal field and the continuous rainfall characteristic spatiotemporal field, the grid-level rainstorm peak flow is estimated to generate a disaster-causing potential surface array.

[0073] In one implementation, using the infiltration loss parameter array as a constraint, and combining the instantaneous rainfall peak spatiotemporal field and the continuous rainfall characteristic spatiotemporal field, a grid-level rainfall peak flow is calculated to generate a disaster-causing potential surface array. Step A400 may further include:

[0074] A410: Based on the infiltration loss parameter array, perform dynamic parameter correction of the standard rainstorm formula and output the corrected rainstorm formula array.

[0075] A420: The instantaneous rainfall peak spatiotemporal field and the continuous rainfall characteristic spatiotemporal field are fused at the grid level to generate a multi-duration rainfall intensity field array mapped to the disaster monitoring grid array.

[0076] A430: The net rainfall intensity field array is generated by dynamically subtracting the infiltration loss of the multi-duration rainfall intensity field array using the modified rainfall formula array.

[0077] A440: Based on the net rainfall intensity field array, perform convergence time iterative reasoning to generate a time-based peak flow array.

[0078] A450: Based on the preset disaster critical threshold, traverse the time-based peak flow array to determine the disaster potential and construct a disaster potential array.

[0079] A460: Using the duration dimension of the duration peak flow array as the time coordinate axis, the multi-duration rainstorm intensity field array and the disaster potential array are organized into the disaster potential surface array according to the three-dimensional structure of duration, rainstorm intensity, and disaster potential.

[0080] In one implementation, a peak flow duration array is generated by performing convergence time iterative inference based on the net rainfall intensity field array, and step A440 may further include:

[0081] A441: The slope runoff time and gully runoff time of the generated net rainfall intensity field array are iteratively calculated to obtain a two-dimensional runoff time array.

[0082] A442: Perform peak flow duration inference based on the two-dimensional confluence time array and net rainfall intensity field array to generate the duration peak flow array.

[0083] A500: Using the disaster monitoring grid array as a spatial computing carrier, perform grid-level disaster susceptibility assessment on the regional static factors of the disaster monitoring area and output a baseline susceptibility array.

[0084] First, it should be understood that the standard rainstorm formula is: In the formula, S is the rainstorm parameter, reflecting the regional rainstorm characteristics; t is the rainstorm duration; and m is the rainstorm loss index in the soil loss parameter, characterizing the attenuation characteristics of rainstorm intensity as duration increases. The formula for calculating effective rainfall intensity in this embodiment is: , where f is the rainstorm loss coefficient in the soil loss parameters.

[0085] This embodiment dynamically corrects the parameter m in the standard rainstorm formula based on the infiltration loss parameter array, so that the rainstorm formula is adapted to the current soil moisture state, and outputs a corrected rainstorm formula array.

[0086] The spatiotemporal field of instantaneous rainfall peak and the spatiotemporal field of continuous rainfall characteristics are fused according to grid cells to generate a rainfall intensity field array covering multiple durations.

[0087] For example, a certain grid provides an instantaneous peak rainfall intensity of 45 mm / h over 10 minutes, and a continuous rainfall field provides an average intensity of 40 mm / h over 3 hours. After fusion, the multi-duration rainstorm intensity sequence of the grid is output, with i=65 mm / h at t=0.5h, i=50 mm / h at t=1h, and i=40 mm / h at t=3h.

[0088] This embodiment uses the modified rainstorm formula array to dynamically subtract the infiltration loss of the multi-duration rainstorm intensity field array, generating a net rainstorm intensity field array. For example, if a certain grid has a rainstorm intensity i = 50 mm / h and f = 1.02 at t = 1 h, then the net rainstorm intensity i 净 =48.98mm / h.

[0089] Based on the net rainfall intensity field array, the slope runoff time and gully runoff time are simultaneously calculated for each grid cell. The dynamic coupling solution of the two runoff times is achieved through iterative calculation. Specifically, the slope runoff time is first calculated based on the topographic parameters of the grid and the initially assumed peak flow value, according to the slope velocity coefficient table. Simultaneously, the gully runoff time is calculated based on gully characteristics, such as tortuosity and blockage, combined with the gully velocity coefficient table (e.g., 0.13 for sparsely vegetated riverbeds).

[0090] Since both depend on the unknown quantity of peak flow, they need to be iterated until the results converge. For example, after three iterations, the slope time stabilizes at 0.2 hours and the channel time stabilizes at 0.8 hours. The final output is a two-dimensional confluence time array containing the time dimensions of the slope and the channel.

[0091] The peak flow rate is calculated based on a two-dimensional confluence time array and a net rainfall intensity field array. The core of this method is to determine the design storm duration by using the total confluence time. Specifically, in this embodiment, the slope and gully confluence times are added together for each grid to obtain the total duration, such as 0.2 + 0.8 = 1 hour. This total duration is used as the design storm period, and the corresponding net rainfall intensity value is extracted from the net rainfall intensity field. For example, a duration of 1 hour corresponds to a net rainfall intensity of 48.98 mm / h.

[0092] The peak flow rate for that duration is then calculated by substituting the formula into the inference formula. For example, using a comprehensive coefficient of 0.278, a runoff coefficient of 0.85, and a drainage area of ​​0.5 km², the peak flow rate is calculated to be 32.6 m³ / s. Finally, a peak flow rate array covering all grids and the design duration is generated.

[0093] This embodiment follows the hydrological principle of determining the duration of a rainstorm based on the confluence time, ensuring consistency between the flood peak calculation and the actual hydrological response process.

[0094] The time-based peak flow array in this embodiment is essentially a spatially gridded multi-time-based peak flow dataset. Each grid cell stores a set of peak flows under different design rainstorm durations, which are subsequently used to construct the disaster potential surface and drive Bayesian disaster chain coupling analysis.

[0095] Based on the preset disaster threshold, such as a debris flow triggering flow of 30 m³ / s, the value of each grid cell in the time-based peak flow array is iterated to perform a binary judgment of disaster potential. If the peak flow of the grid exceeds the threshold, it is marked as high disaster potential, such as 1; otherwise, it is marked as low disaster potential, such as 0.

[0096] For example, if the peak flow of a certain grid is 32.6 m³ / s after 1 hour, exceeding the critical threshold of 30 m³ / s, then the potential value of that cell is output as Z=1. Finally, a disaster potential array covering the entire grid and all durations is constructed. This array quantifies the triggering risk intensity of rainstorm peaks on flash floods, providing a data foundation for the risk dimension of the three-dimensional disaster potential surface.

[0097] Using the time dimension of the historical peak flow array as the time coordinate axis, such as t=0.5h, 1h, 3h, the multi-history rainstorm intensity field array and the disaster potential array are organized in a three-dimensional spatial structure to generate a three-dimensional disaster potential surface array composed of duration, rainstorm intensity, and disaster potential.

[0098] Each grid cell outputs a sequence of point sets. For example, at t=1h, a certain grid cell has a rainfall intensity of i=50mm / h and a disaster potential of Z=1, forming a surface point (1,50,1); at t=3h, i=40mm / h and Z=0, forming a point (3,40,0). By aggregating all historical point sets, a surface model reflecting the dynamic risk evolution is formed, generating the disaster potential surface array, which directly supports Bayesian disaster chain coupling analysis.

[0099] A500: Using the disaster monitoring grid array as a spatial computing carrier, perform grid-level disaster susceptibility assessment on the regional static factors of the disaster monitoring area and output a baseline susceptibility array.

[0100] In one implementation, using the disaster monitoring grid array as a spatial computing carrier, a grid-level disaster susceptibility assessment is performed on the regional static factors of the disaster monitoring area, and a baseline susceptibility array is output. Step A500 may further include:

[0101] A510: The topographic slope distribution, lithological distribution, fault structure and historical disaster point distribution that constitute the static factors of the region are rasterized according to the spatial resolution of the disaster monitoring grid array to generate a gridded factor array.

[0102] A520: Perform multi-factor information weighted fusion on the gridded factor array to output a comprehensive information array.

[0103] A530: The comprehensive information array is traversed using a preset susceptibility level classification rule to perform grid-level disaster susceptibility classification and output the baseline susceptibility array.

[0104] Specifically, this embodiment uses a disaster monitoring grid array as a spatial computing carrier, and standardizes and rasterizes regional static factors such as topographic slope distribution, lithological distribution, fault structure distribution and historical disaster point distribution according to grid resolution to generate a factor data layer that is spatially aligned with the grid array.

[0105] For example, slope data is calculated and converted into slope values ​​for each grid using a digital elevation model, lithology data is classified and coded into lithology types for grid units based on geological maps, and historical disaster points are converted into gridded disaster frequency values ​​through spatial density analysis, ultimately forming a gridded array covering all static factors.

[0106] The specific construction process of the factor data layer is as follows:

[0107] For four types of static factor data—topographic slope, lithology, fault structure, and historical disaster points—rasterization transformation is performed based on the preset spatial resolution of the disaster monitoring grid array. Specifically, continuous slope data is generated into a slope raster array using bilinear interpolation, discrete lithology data is converted from vector surface to raster to generate a lithology coding array, fault structure data is generated into an influence intensity raster based on buffer distance, and historical disaster points are generated into a disaster probability raster through kernel density analysis. The output is a gridded factor array with a unified structure.

[0108] Multi-factor weighted fusion analysis was performed on the rasterized static factor array. First, the contribution of each factor to the occurrence of disasters was quantified independently. For example, in the topographic slope factor, steep slope areas were assigned a higher contribution value, such as a steep slope area with a slope greater than 25 degrees being assigned a higher contribution value of 0.35. In the lithology factor, the contribution value of fragile rock layers was set to 0.42. Fault structures were assigned a contribution value according to distance attenuation, such as 0.28 within 1 kilometer of a fault. Historical disaster points were assigned a contribution value according to spatial density, such as 0.51 for high-density areas.

[0109] Subsequently, the contribution values ​​of each factor within the grid cell are weighted and superimposed according to a preset weight ratio. For example, the weighted value of a certain grid is 0.89. Finally, the output is a comprehensive information array reflecting the probability of disaster occurrence across the entire region. The higher the value, the greater the probability of disaster occurrence.

[0110] According to the preset threshold rules for susceptibility level classification, the disaster susceptibility classification of each grid cell value in the comprehensive information array is performed. Specifically, when the grid information value is below 0.5, it is classified as low susceptibility level, such as a grid value of 0.35 being marked as level 1; when it is in the range of 0.5 to 1.2, it is classified as medium susceptibility level, such as a value of 0.89 being marked as level 2; when it is above 1.2, it is classified as high susceptibility level, such as a value of 1.35 being marked as level 3.

[0111] The grading process in this embodiment follows the principle of information content model, that is, the higher the value, the greater the probability of disaster occurrence. Finally, the baseline susceptibility array corresponding to the susceptibility level code of each grid cell is output. This array is directly input into the Bayesian disaster chain coupling module as a static risk background.

[0112] A600: A dynamic grading-probability coupled array is constructed by coupling the catastrophic potential surface array and the baseline susceptibility array through Bayesian spatiotemporal covariance.

[0113] In one implementation, a dynamic tiered-probability coupled array is constructed by coupling the catastrophic potential surface array and the baseline susceptibility array through Bayesian spatiotemporal covariance. Step A600 may further include:

[0114] A610: Dynamically classify the disaster potential surface array according to the disaster potential value to obtain a disaster potential classification array.

[0115] A620: Staticly divide the baseline susceptibility array according to the susceptibility level value to obtain a susceptibility-divided array.

[0116] A630: Construct a conditional probability judgment matrix based on historical disaster data.

[0117] A640: After aligning the catastrophic potential grading array and the susceptibility grading array, load and execute the conditional probability judgment matrix to calculate and output the risk level probability distribution array.

[0118] A650: Perform probability serialization on the risk level probability distribution array to locate the dominant risk level array.

[0119] A660: The dynamic grading-probability coupling array is generated by superimposing the disaster potential grading array, the susceptibility grading array, the risk level probability distribution array, and the dominant risk level array.

[0120] This embodiment integrates a dynamically evolving disaster-causing potential surface array and a static baseline susceptibility array through a Bayesian spatiotemporal covariance coupling mechanism, and deeply analyzes the interaction effects of rainstorm peaks and geological environment as dynamic disaster-causing factors and static disaster-inducing backgrounds in the spatiotemporal dimension.

[0121] By employing grid-level spatiotemporal alignment technology, the disaster-causing potential classification results for each unit are ensured to accurately match the susceptibility classification results. Based on a conditional probability matrix constructed from historical disaster statistics, the probability of disaster occurrence under different combinations of risk levels is quantified, ultimately generating a dynamic classification-probability coupling array that integrates multi-level risk probability distributions and dominant early warning labels. This dynamic classification-probability coupling array supports dynamic hierarchical early warning of disaster chain risks, enabling full-process quantitative assessment of chain disasters from rainstorms to flash floods to landslides / debris flows.

[0122] Specifically, for each grid cell in the disaster-causing potential surface array, dynamic grading is performed based on its disaster-causing potential value. The specific rules for dynamic grading are as follows: three threshold ranges are set: low, medium, and high. When the potential value is below the low threshold, it is classified as low disaster-causing potential; when it is in the medium range, it is classified as medium disaster-causing potential; and when it reaches the high threshold, it is classified as high disaster-causing potential. For example, if a grid cell's potential value exceeds the debris flow triggering threshold due to a rainstorm peak flow, it is marked as 1, representing the high level.

[0123] The final output is a disaster potential classification array that covers the entire space. Its classification criteria strictly correspond to the dynamic parameter α level in the disaster chain model, serving as the dynamic input for Bayesian coupling.

[0124] Similarly, for each grid cell in the baseline susceptibility array, static grading is performed according to a preset susceptibility level range. The specific rules for static grading are as follows: low susceptibility ranges are mapped to low susceptibility levels, medium susceptibility ranges are mapped to medium susceptibility levels, and high susceptibility ranges are mapped to high susceptibility levels. For example, if a grid cell is assessed as having a high susceptibility level of 3 due to a combination of steep slope and fragile lithology, it will be classified as a high-susceptibility cell.

[0125] In this embodiment, the static classification results correspond to the static parameter β level in the disaster chain model, ensuring that the inherent risks of the geological background and the dynamic rainstorm peaks are accurately matched on the spatial grid.

[0126] It should be noted that this embodiment does not specifically limit the numerical range of static and dynamic grading, and can be specifically set according to the actual application scenario.

[0127] Based on long-term historical disaster data, this study statistically analyzes the actual frequency of disasters under different combinations of risk levels, constructing a conditional probability judgment matrix. Specifically, it collects samples of combinations of α and β from historical events, such as combinations of high catastrophic potential and high susceptibility, and then calculates the proportion of disaster triggers for each combination relative to the total sample size. For example, statistics show that the probability of disaster occurrence for the α-high + β-high combination is 92%, while the probability for the α-medium + β-low combination is 35%. Finally, a probability lookup table covering all risk level combinations is generated, serving as the core basis for Bayesian inference.

[0128] In this embodiment, the disaster potential classification array and the susceptibility classification array are strictly aligned according to grid coordinates and time dimension, and the conditional probability judgment matrix is ​​loaded to perform risk calculation.

[0129] Specifically, for each grid cell, the corresponding multi-level risk probability distribution is obtained by querying the matrix based on its α and β levels. For example, if a grid cell has both high α and high β levels, the probability of risk level R33***** is 92%, and the probability of R32**** is 8%. The multi-level probability set for each grid cell is output, forming a probability distribution array that characterizes the uncertainty of disaster chain risk.

[0130] Where R represents the risk level, the first digit after R is the level of the static parameter β, 1=low, 2=medium, 3=high, the second digit after R is the level of the Wie dynamic parameter α, 1=low, 2=medium, 3=high, and the suffix asterisk indicates the severity of the risk, and the more asterisks, the higher the risk. The specific number of asterisks does not limit the degree of risk quantification in this embodiment.

[0131] Extract the probability value sequence of each risk level within the grid, and determine the level corresponding to the maximum probability as the dominant risk. If there are multiple similar high probability values, such as R33 probability of 45% and R32**** probability of 43%, then select the highest level according to the preset high risk priority rule. For example, if the probability of R33***** in a certain grid reaches 92%, it is directly marked as the dominant risk.

[0132] This embodiment selects and outputs a global dominant risk level array based on the dominant risk, providing a clear direction for early warning actions.

[0133] By superimposing and fusing the α-level labels of the catastrophic potential grading array, the β-level labels of the susceptibility grading array, the multi-probability values ​​of the risk level probability distribution array, and the decision labels of the dominant risk level array, a dynamic grading-probability coupling array integrating grading-probability-decision multi-dimensional elements is generated.

[0134] This embodiment uses the Bayesian spatiotemporal covariance coupling method to jointly analyze the dynamically changing disaster-causing potential surface array and the static baseline susceptibility array. It fully considers the interaction effects of disaster-causing factors and disaster-inducing environment in the time and space dimensions, and constructs a graded-probability coupling array that can reflect the dynamic risk evolution of the rainstorm-flash flood-geological disaster chain.

[0135] Its core technology lies in establishing a spatiotemporal alignment mechanism to ensure that the dynamic classification results of the disaster potential of each grid unit are accurately matched with the static classification results of the susceptibility. By quantifying the probability of risk occurrence under different combinations of risk levels through a conditional probability matrix, a multi-dimensional output array that integrates the probability distribution of risk levels and the dominant decision label is finally generated, providing a quantitative decision basis for disaster chain early warning.

[0136] A700: Loads real-time rainfall data into the dynamic grading-probability coupling array to perform grid-level disaster risk situation prediction and outputs the disaster chain risk probability distribution.

[0137] Specifically, this embodiment is used to dynamically update and predict the disaster chain risk situation based on real-time rainfall data, and output the disaster chain risk probability distribution that represents the dominant risk level at the grid level across the entire region. The specific data processing procedure is as follows:

[0138] First, by performing multi-scale decomposition step A100 on the rainfall signal of real-time rainfall data, an updated instantaneous rainstorm peak field and a continuous rainfall characteristic field are generated. Then, the rainstorm flood peak flow estimation process is driven to recalculate the disaster-causing surface array.

[0139] Subsequently, the hazard potential grading array is dynamically adjusted based on the updated hazard potential values ​​and spatiotemporally aligned with the static baseline hazard grading array. Next, a conditional probability judgment matrix is ​​loaded to perform Bayesian coupling calculations, outputting the updated multi-level risk probability distribution for each grid cell. Finally, the dominant risk level is extracted through probability serialization to generate a real-time hazard chain risk probability distribution array covering the entire domain.

[0140] This process enables minute-level response to dynamic rainstorm data and static geological background. For example, if the real-time rainstorm peak flow rate of a certain grid jumps to the supercritical threshold of 35 m³ / s, its α level will be raised to high. After high coupling with the inherent β level, it will trigger the R33***** risk level and directly output a red warning command to support real-time and precise prevention and control of disaster chain risks.

[0141] This embodiment is based on the dynamic decomposition of real-time rainfall signals to drive grid-level flood peak flow extrapolation, quantifies the disaster potential caused by rainstorms, combines static geological susceptibility zoning to classify background risk levels, and uses the Bayesian spatiotemporal covariance mechanism to fuse dynamic disaster potential and static susceptibility according to matrix-based conditional probability to generate dominant risk levels. Finally, the risk situation is updated by loading real-time rainfall data, achieving the technical effects of minute-level response, precise spatial positioning, and zoned and graded prevention and control decision support for the entire process of disaster chain from rainstorm triggering to secondary geological disasters.

[0142] Example 2: Based on the same inventive concept as the disaster risk situation generation method based on dynamic grading and classification in the foregoing examples, this invention provides a disaster risk situation generation system based on dynamic grading and classification. See [link to example]. Figure 2 As shown, the system includes:

[0143] The rainfall multi-scale decomposition module 01 is used to perform multi-scale decomposition on the original time-series rainfall signal transmitted back by the rain gauge station, and to capture the spatiotemporal field of the instantaneous rainstorm peak and the spatiotemporal field of the continuous rainfall characteristics. The rain gauge station is deployed in the disaster monitoring area.

[0144] Spatial gridding module 02 is used to grid the disaster monitoring area based on a preset grid scale to obtain a disaster monitoring grid array.

[0145] The infiltration loss dynamic inversion module 03 is used to perform dynamic inversion of infiltration loss parameters in the disaster monitoring grid array and construct an infiltration loss parameter array.

[0146] The disaster potential surface generation module 04 is used to generate a disaster potential surface array by using the infiltration loss parameter array as a constraint condition and combining the instantaneous rainstorm peak spatiotemporal field and the continuous rainfall characteristic spatiotemporal field to perform grid-level rainstorm peak flow estimation.

[0147] The static susceptibility assessment module 05 is used to perform grid-level disaster susceptibility assessment on the regional static factors of the disaster monitoring area using the disaster monitoring grid array as a spatial computing carrier, and output a baseline susceptibility array.

[0148] The Bayesian disaster chain coupling module 06 is used to construct a dynamic tiered-probability coupling array by coupling the disaster potential surface array and the baseline susceptibility array through Bayesian spatiotemporal covariance.

[0149] The real-time risk prediction module 07 is used to load real-time rainfall data into the dynamic grading-probability coupling array to perform grid-level disaster risk situation prediction and output the disaster chain risk probability distribution.

[0150] In one implementation, the rainfall multi-scale decomposition module 01 is used for:

[0151] Decomposition level parameters are set according to the regional rainstorm characteristics of the disaster monitoring area; wavelet basis functions are selected according to the rainstorm disaster-causing mechanism of the disaster monitoring area; using the decomposition level parameters as the transform order constraint, the wavelet basis functions are used to perform discrete wavelet transform decomposition on the original time-series rainfall signal to obtain high-frequency detail coefficient sequences and low-frequency approximation coefficients; after reconstructing the short-duration rainstorm components based on the high-frequency detail coefficient sequences, local maxima are extracted through spatial interpolation to form the spatiotemporal field of the instantaneous rainstorm peak; the long-duration rainfall components are reconstructed based on the low-frequency approximation coefficients, and the rainfall intensity field is generated through gridded moving average processing, which is used as the spatiotemporal field output of the continuous rainfall characteristics.

[0152] In one implementation, the disaster-causing potential surface generation module 04 is used for:

[0153] Based on the infiltration loss parameter array, the parameters of the standard rainstorm formula are dynamically corrected, and a corrected rainstorm formula array is output. The instantaneous rainstorm peak spatiotemporal field and the continuous rainfall characteristic spatiotemporal field are fused at the grid level to generate a multi-duration rainstorm intensity field array mapped to the disaster monitoring grid array. The infiltration loss of the multi-duration rainstorm intensity field array is dynamically subtracted from the corrected rainstorm formula array to generate a net rainfall intensity field array. Based on the net rainfall intensity field array, iterative reasoning of the confluence time is performed to generate a duration peak flow array. The duration peak flow array is traversed according to a preset disaster critical threshold to determine the disaster potential and construct a disaster potential array. Using the duration dimension of the duration peak flow array as the time coordinate axis, the multi-duration rainstorm intensity field array and the disaster potential array are organized into a three-dimensional structure of duration, rainstorm intensity, and disaster potential to generate the disaster potential surface array.

[0154] In one implementation, the disaster-causing potential surface generation module 04 is used for:

[0155] The slope runoff time and gully runoff time of the generated net rainfall intensity field array are iteratively calculated to obtain a two-dimensional runoff time array; based on the two-dimensional runoff time array and the net rainfall intensity field array, the peak flow duration inference is performed to generate the duration peak flow array.

[0156] In one implementation, the Bayesian disaster chain coupling module 06 is used for:

[0157] The disaster potential surface array is dynamically categorized based on the disaster potential value to obtain a disaster potential categorized array; the baseline susceptibility array is statically categorized based on the susceptibility level value to obtain a susceptibility categorized array; a conditional probability judgment matrix is ​​constructed based on historical disaster data; after aligning the disaster potential categorized array and the susceptibility categorized array, the conditional probability judgment matrix is ​​loaded and the risk level probability distribution array is calculated and output; the risk level probability distribution array is probabilistically serialized to locate the dominant risk level array; the disaster potential categorized array, the susceptibility categorized array, the risk level probability distribution array, and the dominant risk level array are superimposed to generate the dynamic categorization-probability coupling array.

[0158] In one implementation, the static susceptibility assessment module 05 is used for:

[0159] The topographic slope distribution, lithological distribution, fault structure, and historical disaster point distribution that constitute the static factors of the region are rasterized according to the spatial resolution of the disaster monitoring grid array to generate a gridded factor array; multi-factor information weighted fusion is performed on the gridded factor array to output a comprehensive information array; the comprehensive information array is traversed using a preset susceptibility level classification rule to perform grid-level disaster susceptibility classification and output the baseline susceptibility array.

[0160] In one implementation, the infiltration loss dynamic inversion module 03 is used for:

[0161] The interactive remote sensing satellite system obtains multispectral remote sensing data of the disaster monitoring area; based on the disaster monitoring grid array, it performs gridded inversion of soil moisture content on the multispectral remote sensing data and outputs a soil moisture content array; it maps the soil moisture content array to a soil dryness and wetness classification array; using the soil dryness and wetness classification array as a query condition, it iterates through the soil loss parameter comparison table to obtain the infiltration loss parameter array.

[0162] Example 3: A computer-readable storage medium that can be used to store software programs, computer executable programs, and modules to perform various functional applications and data processing of computer devices, that is, to realize the above-mentioned disaster risk situation generation method based on dynamic classification and grading.

[0163] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

[0164] The foregoing description of specific exemplary embodiments of the invention is for illustrative and explanatory purposes. These descriptions are not intended to limit the invention to the precise forms disclosed, and it will be apparent that many changes and variations can be made in accordance with the foregoing teachings. The exemplary embodiments were chosen and described in order to explain the specific principles of the invention and its practical application, thereby enabling those skilled in the art to implement and utilize various different exemplary embodiments of the invention, as well as various different choices and variations. The scope of the invention is intended to be defined by the claims and their equivalents.

Claims

1. A method for generating disaster risk status based on dynamic grading and classification, characterized in that, include: The raw time-series rainfall signals transmitted back from the rain gauges are decomposed at multiple scales to capture the spatiotemporal field of the instantaneous rainstorm peak and the spatiotemporal field of the continuous rainfall characteristics. The rain gauges are deployed in the disaster monitoring area. The disaster monitoring area is gridded according to a preset grid scale to obtain a disaster monitoring grid array; Dynamic inversion of infiltration loss parameters is performed on the disaster monitoring grid array to construct an infiltration loss parameter array; Using the infiltration loss parameter array as a constraint, and combining the instantaneous rainstorm peak spatiotemporal field and the continuous rainfall characteristic spatiotemporal field, grid-level rainstorm peak flow is estimated to generate a disaster-causing potential surface array. Using the disaster monitoring grid array as a spatial computing carrier, a grid-level disaster susceptibility assessment is performed on the regional static factors of the disaster monitoring area, and a baseline susceptibility array is output. A dynamic grading-probability coupled array is constructed by coupling the catastrophic potential surface array and the baseline susceptibility array using Bayesian spatiotemporal covariance. Real-time rainfall data is loaded into the dynamic grading-probability coupling array to perform grid-level disaster risk situation prediction and output the disaster chain risk probability distribution.

2. The disaster risk situation generation method based on dynamic grading and classification as described in claim 1, characterized in that, The raw time-series rainfall signals transmitted from rain gauges are decomposed at multiple scales to capture the spatiotemporal fields of instantaneous rainstorm peaks and persistent rainfall characteristics, including: Decomposition level parameters are set based on the regional rainstorm characteristics of the disaster monitoring area; Wavelet basis functions were selected based on the rainstorm-induced disaster mechanism in the disaster monitoring area. Using the decomposition level parameters as transform order constraints, the discrete wavelet transform decomposition of the original time-series rainfall signal is performed using the wavelet basis functions to obtain the high-frequency detail coefficient sequence and low-frequency approximation coefficients; After reconstructing the short-duration rainstorm component based on the high-frequency detail coefficient sequence, local maxima are extracted by spatial interpolation to form the spatiotemporal field of the instantaneous rainstorm peak. Based on the low-frequency approximation coefficients, the long-duration rainfall components are reconstructed, and a rainfall intensity field is generated through gridded moving average processing, which serves as the spatiotemporal field output of the continuous rainfall characteristics.

3. The disaster risk situation generation method based on dynamic grading and classification as described in claim 1, characterized in that, Using the infiltration loss parameter array as a constraint, and combining the instantaneous rainfall peak spatiotemporal field and the continuous rainfall characteristic spatiotemporal field, a grid-level rainfall peak flow estimation is performed to generate a disaster-causing potential surface array, including: Based on the infiltration loss parameter array, the parameters of the standard rainstorm formula are dynamically corrected, and the corrected rainstorm formula array is output. The instantaneous rainfall peak spatiotemporal field and the continuous rainfall characteristic spatiotemporal field are fused at the grid level to generate a multi-duration rainfall intensity field array mapped to the disaster monitoring grid array; The net rainfall intensity field array is generated by dynamically subtracting the infiltration loss of the multi-duration rainfall intensity field array from the modified rainfall formula array. Based on the net rainfall intensity field array, iterative reasoning of the confluence time is performed to generate a time-based peak flow array; Based on the preset disaster critical threshold, the time-based peak flow array is traversed to determine the disaster potential and construct a disaster potential array; Using the duration dimension of the duration peak flow array as the time coordinate axis, the multi-duration rainstorm intensity field array and the disaster potential array are organized into the disaster potential surface array according to the three-dimensional structure of duration, rainstorm intensity, and disaster potential.

4. The disaster risk situation generation method based on dynamic grading and classification as described in claim 3, characterized in that, Based on the net rainfall intensity field array, iterative inference of the confluence time is performed to generate a time-bound peak flow array, including: The slope runoff time and gully runoff time of the generated net rainfall intensity field array are iteratively calculated to obtain a two-dimensional runoff time array; Based on the two-dimensional confluence time array and net rainfall intensity field array, the peak flow duration inference is performed to generate the duration peak flow array.

5. The disaster risk situation generation method based on dynamic grading and classification as described in claim 1, characterized in that, A dynamic tiered-probability coupled array is constructed by coupling the catastrophic potential surface array and the baseline susceptibility array using Bayesian spatiotemporal covariance, including: The disaster potential surface array is dynamically divided according to the disaster potential value to obtain a disaster potential grading array. The baseline susceptibility array is statically divided according to the susceptibility level value to obtain a susceptibility-divided array; Construct a conditional probability judgment matrix based on historical disaster data; After aligning the catastrophic potential grading array and the susceptibility grading array, load the conditional probability judgment matrix to calculate and output the risk level probability distribution array. The probability distribution array of the risk levels is serialized to locate the dominant risk level array; The dynamic grading-probability coupling array is generated by superimposing the disaster-causing potential grading array, the susceptibility grading array, the risk level probability distribution array, and the dominant risk level array.

6. The disaster risk situation generation method based on dynamic grading and classification as described in claim 1, characterized in that, Using the disaster monitoring grid array as a spatial computing platform, a grid-level disaster susceptibility assessment is performed on the regional static factors of the disaster monitoring area, outputting a baseline susceptibility array, including: The topographic slope distribution, lithological distribution, fault structure, and historical disaster point distribution that constitute the static factors of the region are rasterized according to the spatial resolution of the disaster monitoring grid array to generate a gridded factor array. Perform multi-factor information weighted fusion on the gridded factor array to output a comprehensive information array; The comprehensive information array is traversed using a preset susceptibility level classification rule to perform grid-level disaster susceptibility classification and output the baseline susceptibility array.

7. The disaster risk situation generation method based on dynamic grading and classification as described in claim 1, characterized in that, Dynamic inversion of infiltration loss parameters is performed on the disaster monitoring grid array to construct an infiltration loss parameter array, including: Interactive remote sensing satellite systems acquire multispectral remote sensing data of the disaster monitoring area; Based on the disaster monitoring grid array, soil moisture content gridding inversion is performed on the multispectral remote sensing data to output a soil moisture content array; The soil moisture content array is mapped to a soil dryness and wetness classification array; Using the soil dryness and wetness classification array as the query condition, the soil loss parameter comparison table is traversed to obtain the infiltration loss parameter array.

8. A disaster risk situation generation system based on dynamic grading and classification, characterized in that, For implementing the method steps of any one of claims 1 to 7, including: The rainfall multi-scale decomposition module is used to perform multi-scale decomposition on the raw time-series rainfall signals transmitted back from the rain gauge station, and to capture the spatiotemporal field of the instantaneous rainstorm peak and the spatiotemporal field of the continuous rainfall characteristics, wherein the rain gauge station is deployed in the disaster monitoring area; The spatial gridding module is used to grid the disaster monitoring area based on a preset grid scale to obtain a disaster monitoring grid array. The infiltration loss dynamic inversion module is used to dynamically invert infiltration loss parameters in the disaster monitoring grid array and construct an infiltration loss parameter array. The disaster-causing potential surface generation module is used to generate a disaster-causing potential surface array by using the infiltration loss parameter array as a constraint condition and combining the instantaneous rainstorm peak spatiotemporal field and the continuous rainfall characteristic spatiotemporal field to perform grid-level rainstorm peak flow estimation. The static susceptibility assessment module is used to perform grid-level disaster susceptibility assessment on the regional static factors of the disaster monitoring area using the disaster monitoring grid array as a spatial computing carrier, and output a baseline susceptibility array; The Bayesian disaster chain coupling module is used to couple the disaster potential surface array and the baseline susceptibility array through Bayesian spatiotemporal covariance to construct a dynamic tiered-probability coupling array. The real-time risk prediction module is used to load real-time rainfall data into the dynamic grading-probability coupling array to perform grid-level disaster risk situation prediction and output the disaster chain risk probability distribution.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the disaster risk situation generation method based on dynamic grading and classification as described in any one of claims 1-7.

Images