LanMei basin multi-disaster risk early warning method, system, equipment and medium

Abstract

The invention provides a LanMei basin multi-disaster risk early warning method, system and equipment and a medium, and relates to the technical field of disaster early warning. The method comprises the steps that a multi-disaster cooperative processing engine is planned to be constructed, and disaster coupling calculation and dynamic early warning with high temporal-spatial resolution are achieved. Forecasting and early warning of short-term and medium-term extreme events such as rainstorm, high temperature and typhoon and disasters such as derived flood, drought and geological landslide are achieved, and coupling modeling is achieved. And performing normalization and visualization processing on the multi-disaster forecasting result, outputting a risk early warning result, and displaying the multi-disaster risk in a visual map and chart form. According to the scheme provided by the invention, intelligent early warning support is provided for disaster prevention and reduction in the LanMei basin, and the accuracy and timeliness of disaster early warning are improved.

Description

Technical Field

[0001] The present invention belongs to the field of disaster early warning, and particularly relates to a method, system, device and medium for multi-disaster risk early warning in the Lancang-Mekong River Basin. Background Art

[0002] At present, the meteorological monitoring and forecasting capabilities in the Mekong River Basin within the Southeast Asian region are generally weak. There is an urgent need to carry out research and development on regional meteorological disaster monitoring and early warning technologies, improve regional disaster prevention and mitigation capabilities, and jointly enhance regional early warning capabilities. The existing meteorological disaster monitoring and early warning technologies in China mainly have the following problems and deficiencies: First, the problem of data heterogeneity. Meteorological multi-source data includes station observations, satellite retrievals, numerical forecasts, etc. The data formats are mixed (NetCDF / GRIB / HDF / TXT, etc.), and the spatio-temporal resolutions and physical element descriptions are not unified. This leads to the need for a variety of differential parsing tools, and the entire processing process is extremely complex, greatly increasing the difficulty of data integration and analysis, and reducing the data processing efficiency and accuracy.

[0003] Second, there is a bottleneck in service efficiency. In practical applications, the query and visualization requirements for meteorological data are very diverse. Users hope to quickly obtain the required information from large-scale data sets and display or even analyze and evaluate it through visualization tools. Traditional query and visualization technologies are difficult to meet the real-time response requirements of large-scale heterogeneous data, and the system latency is significant in high-concurrency scenarios, affecting the user experience.

[0004] The forecast and early warning content is single. Most of the existing early warning service content mainly focuses on meteorological element forecasts such as temperature and precipitation or circulation situation forecasts, lacking complex disaster type forecast and early warning content related to weather and climate such as floods, droughts, geological landslides, and ecological monitoring in overseas regions. Functions such as parallel computing of multi-disaster type forecast models, visualization display of risk early warning results, and interactive evaluation of multi-disaster type monitoring are also basically lacking. Summary of the Invention

[0005] To solve the above technical problems, the present invention proposes a technical solution for a method, system, device and medium for multi-disaster risk early warning in the Lancang-Mekong River Basin to solve the above technical problems.

[0006] The first aspect of the present invention discloses a method for multi-disaster risk early warning in the Lancang-Mekong River Basin, and the method includes: Step S1: Parse NetCDF data, GRIB data, HDF data and TXT data, and output a binary file with a unified format; perform spatio-temporal consistency processing on the binary file, and output a data file with a unified spatial resolution and time resolution; Step S2: For data with a resolution higher than the target spatial resolution, interpolation is applied for data downscaling; for data with a resolution lower than the target spatial resolution, interpolation is applied for data upscaling. Step S3: If there are missing or incorrect hourly data in the data file, linear interpolation is performed based on the data at the previous and subsequent time points; non-hourly data is resampled in time to be converted into hourly data, and the converted hourly data is obtained; the linearly interpolated hourly data and the converted hourly data are merged to obtain the meteorological data input to the model. Step S4: Initialize and configure the multi-hazard collaborative processing engine to obtain grid base terrain data, soil infiltration characteristic parameters, and vegetation interception and evapotranspiration parameters. Step S5: Apply the meteorological data, grid base terrain data, soil infiltration characteristic parameters, and vegetation interception and evapotranspiration parameters to construct a flood forecasting model, a drought forecasting model, and a geological landslide risk forecasting model. Step S6: Conduct multi-hazard coupling calculations using the flood forecasting model, the drought forecasting model, and the geological landslide risk forecasting model; based on the results of the multi-hazard coupling calculations, output and visualize the multi-hazard risk warning results.

[0007] According to the method of the first aspect of the present invention, in the step S1, Set the spatial resolution to 0.1°×0.1° and the time resolution to hourly, perform spatio-temporal consistency processing on the binary file, and output a data file with a unified spatial resolution and time resolution.

[0008] According to the method of the first aspect of the present invention, in the step S2, the application of interpolation for data downscaling of data with a resolution higher than the target spatial resolution includes: Read the target resolution parameters, generate the target grid coordinate system, perform downscaling on the data with a resolution higher than the target spatial resolution using the Kriging interpolation method, calculate the target grid values, and write them into the downscaled data file. The application of interpolation for data upscaling of data with a resolution lower than the target spatial resolution includes: Based on the target resolution grid, according to the values of the low-resolution data points within the preset range of each grid point and their distances from the grid point, calculate the value of the grid point through the inverse distance weighting formula.

[0009] According to the method of the first aspect of the present invention, in step S3, for daily precipitation data, the total daily precipitation is evenly distributed over 24 hours of the day to obtain an hourly precipitation data estimate; or in combination with the precipitation diurnal variation characteristic model of the region, using historical hourly precipitation data, the average proportion of hourly precipitation in the daily total is statistically calculated, and the daily precipitation data is distributed to the hourly scale.

[0010] According to the method of the first aspect of the present invention, in step S4, the initialization configuration of the multi-disaster collaborative processing engine includes: Generate grid-based terrain data with a resolution of 10 km within the basin range according to digital elevation model data; Determine the infiltration characteristic parameters of the soil based on soil type data; Calculate the vegetation interception and evapotranspiration parameters using vegetation cover data.

[0011] According to the method of the first aspect of the present invention, in step S5, the application of the meteorological data, grid-based terrain data, soil infiltration characteristic parameters, and vegetation interception and evapotranspiration parameters to construct flood forecasting models, drought forecasting models, and geological landslide risk forecasting models includes: The input meteorological data is used as the real-time driving data of the model. Combining with the initialized geographical information parameters, simulate the grid infiltration and runoff generation of complex terrain and soil characteristics in different grids under the influence of different precipitation spatio-temporal distributions within the basin; by calculating the generation amounts of surface runoff and subsurface runoff, and using the hydrodynamic grid confluence mechanism, calculate and predict the dynamic propagation process of floods within the basin and the changes in hydrological elements; the geographical information parameters include: grid-based terrain data, soil infiltration characteristic parameters, and vegetation interception and evapotranspiration parameters; hydrological elements include: flow rate and water level; Extract precipitation and temperature from the meteorological data, calculate the potential evapotranspiration by the Penman-Monteith method, and then calculate the SPEI values at different time scales based on the precipitation data and potential evapotranspiration; use the SPEI values as indicators for drought monitoring and prediction, and adopt an autoregressive moving average model to predict future SPEI values based on the historical SPEI value sequence, so as to judge the development trend of drought; Combine the precipitation data, terrain slope data, and soil stability data in the meteorological data to establish a geological landslide risk assessment index; use machine learning algorithms to analyze the relationship between the geological landslide risk assessment index and the probability of geological landslide occurrence, construct a geological landslide risk forecasting model, and conduct real-time assessment and prediction of the geological landslide risks in different regions within the basin.

[0012] According to the method of the first aspect of the present invention, in step S6, the output and visualization of multi-disaster risk early warning results according to the results of multi-disaster coupling calculation include: Perform multi-threaded parallel normalization processing on the forecast results of each disaster type calculated by the multi-disaster collaborative processing engine and the results of multi-disaster coupling calculation to obtain normalized data; Put the normalized data into the channels of a grayscale image to generate a grayscale image result file; Compress the grayscale image result file using a lossless compression algorithm; Utilize WebGL graphics rendering technology to directly load and render the compressed grayscale image result file, and display the multi-disaster risk warning results in the form of maps and charts.

[0013] The second aspect of the present invention discloses a multi-disaster risk warning system for the Lancang-Mekong River Basin, and the system includes: A first processing module, configured to parse NetCDF data, GRIB data, HDF data, and TXT data, and output a binary file with a unified format; perform spatio-temporal consistency processing on the binary file, and output a data file with a unified spatial resolution and time resolution; A second processing module, configured to, for data with a resolution higher than the target spatial resolution, apply an interpolation method for data downscaling processing; for data with a resolution lower than the target spatial resolution, apply an interpolation method for data upscaling processing; A third processing module, configured to, if there are missing or incorrect hourly data in the data file, perform linear interpolation based on the data at the previous and subsequent time points; perform time resampling on non-hourly data to convert it into hourly data, and obtain the converted hourly data; merge the linearly interpolated hourly data and the converted hourly data to obtain the meteorological data input to the model; A fourth processing module, configured to perform initialization configuration on the multi-disaster collaborative processing engine to obtain grid base terrain data, soil infiltration characteristic parameters, and vegetation interception and evapotranspiration parameters; A fifth processing module, configured to apply the meteorological data, grid base terrain data, soil infiltration characteristic parameters, and vegetation interception and evapotranspiration parameters to construct a flood forecast model, a drought forecast model, and a geological landslide risk forecast model; A sixth processing module, configured to perform multi-disaster coupling calculation using the flood forecast model, the drought forecast model, and the geological landslide risk forecast model; according to the results of the multi-disaster coupling calculation, perform output and visualization of multi-disaster risk warning results.

[0014] The third aspect of the present invention discloses an electronic device. The electronic device includes a memory and a processor. When the processor executes a computer program, the steps in any one of the methods for multi-disaster risk warning of the Lancang-Mekong River Basin in the first aspect of the present disclosure are implemented.

[0015] In the fourth aspect of the present invention, a computer-readable storage medium is disclosed. A computer program is stored on the computer-readable storage medium. When the computer program is executed by a processor, the steps in a multi-disaster risk early warning method for the Lancang-Mekong River Basin in any one of the first aspects of the present disclosure are implemented.

[0016] In summary, by combining a multi-source heterogeneous data adaptive parsing engine and a multi-disaster meteorological risk forecasting collaborative computing technology, the present invention solves the key problems in meteorological services in the Lancang-Mekong River Basin. In terms of data heterogeneity, traditional methods require manual writing of parsing codes for different formats, which is time-consuming and prone to human errors. The multi-source heterogeneous data adaptive parsing engine of the present invention can quickly convert different formats of data into a unified format, and the data preprocessing efficiency can be increased by ten times. In traditional technologies, inconsistent spatio-temporal benchmarks lead to difficulties in data integration. The spatio-temporal normalization model established by the present invention automatically matches data with different resolutions and time scales. This ensures the physical consistency of data fusion and provides a reliable basis for subsequent data analysis, with a significant improvement in data integration efficiency compared to the prior art, and the analysis results are also more reliable.

[0017] In terms of the service efficiency bottleneck, traditional query and visualization technologies have poor real-time response capabilities and serious delays in high-concurrency scenarios when facing large-scale heterogeneous data. The present invention uses a grayscale map result file for front-end rendering display, and normalizes the queried meteorological data to generate a grayscale map. The grayscale map file is compressed using a lossless compression algorithm, greatly reducing the file size and improving the front-end and back-end transmission efficiency, providing a smoother and more efficient experience for users.

[0018] In the early warning technology, the parallel computing and coupling of a multi-disaster meteorological risk forecasting model for overseas regions are filled. The present invention develops a multi-disaster meteorological risk forecasting collaborative computing technology, researches and develops disaster weather forecasting and early warning technologies and products such as heavy rain, high temperature, and typhoon to support meteorological services in the Lancang-Mekong River Basin, and adapts to the meteorological and hydrological underlying surface characteristics of the Lancang-Mekong River Basin to develop regionalized flood, drought, and geological landslide early warning technologies and products. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0020] Figure 1 It is a flowchart of a multi-disaster risk early warning method for the Lancang-Mekong River Basin according to an embodiment of the present invention; Figure 2Structural diagram of a multi-hazard risk early warning system for the Lancang-Mekong River Basin according to an embodiment of the present invention; Figure 3 Structural diagram of an electronic device according to an embodiment of the present invention. Detailed implementation manners

[0021] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are only a part rather than all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0022] The first aspect of the present invention discloses a multi-hazard risk early warning method for the Lancang-Mekong River Basin. Figure 1 As shown in the flowchart of a multi-hazard risk early warning method for the Lancang-Mekong River Basin according to an embodiment of the present invention, Figure 1 the method includes: Step S1: Parse NetCDF data, GRIB data, HDF data and TXT data, and output a binary file with a unified format; perform spatio-temporal consistency processing on the binary file, and output a data file with a unified spatial resolution and time resolution; Step S2: For data with a resolution higher than the target spatial resolution, apply the interpolation method for data downscaling; for data with a resolution lower than the target spatial resolution, apply the interpolation method for data upscaling; Step S3: If there are missing or incorrect hourly data in the data file, perform linear interpolation based on the data at the previous and subsequent time points; perform time resampling on the non-hourly data to convert it into hourly data, and obtain the converted hourly data; merge the linearly interpolated hourly data and the converted hourly data to obtain the meteorological data input to the model; Step S4: Initialize and configure the multi-hazard collaborative processing engine to obtain grid base terrain data, soil infiltration characteristic parameters, and vegetation interception and evapotranspiration parameters; Step S5: Apply the meteorological data, grid base terrain data, soil infiltration characteristic parameters, and vegetation interception and evapotranspiration parameters to construct a flood forecasting model, a drought forecasting model, and a geological landslide risk forecasting model; Step S6: Apply the flood forecasting model, the drought forecasting model, and the geological landslide risk forecasting model to carry out multi-hazard coupling calculations; according to the results of the multi-hazard coupling calculations, output and visualize the multi-hazard risk early warning results.

[0023] In step S1, parse NetCDF data, GRIB data, HDF data, and TXT data, and output a binary file with a unified format; perform spatio-temporal consistency processing on the binary file to output a data file with a unified spatial resolution and temporal resolution.

[0024] In some embodiments, in step S1, set the spatial resolution to 0.1°×0.1° and the temporal resolution to hourly, perform spatio-temporal consistency processing on the binary file, and output a data file with a unified spatial resolution and temporal resolution.

[0025] Specifically, the NetCDF format is often used to store scientific data and has a complex data structure. Call the NetCDF library to read the file header, extract variables (such as temperature, humidity), dimensions (longitude, latitude, altitude level), and timestamps. Map variable names to a unified physical element code according to rules (such as TMP → temperature, RH → humidity). Write the three-dimensional data (altitude level × longitude × latitude) into the binary file in descending order of altitude level and ascending order of longitude / latitude. The output is a unified binary file (element + time naming) (such as <TEMP_2024051001.bin>) to ensure the input consistency for subsequent spatio-temporal normalization processing.

[0026] The GRIB format is mainly used for storing meteorological numerical forecast data. The plugin uses the GRIB decoding library to parse the GRIB message. It first identifies the GRIB version number, and there are slight differences in data organization for different versions of the GRIB format. Then parse key parts such as the indicator section and identification section to obtain information such as the physical elements of the data (such as air pressure, wind speed), grid definition (spatial resolution), and time identification. The output is a unified binary file in the same format as S1 to ensure the input consistency for subsequent spatio-temporal normalization processing.

[0027] The HDF format can store various types of data and has a hierarchical data structure. Traverse the HDF group structure to locate the meteorological dataset (such as / VIIRS / Channel_IR). Extract the data matrix and geolocation information (such as scan angle, projection parameters), and convert them into longitude-latitude grids. For multi-band data (such as infrared, visible light), store them as separate files respectively. The output is a unified binary file in the same format as S1 to ensure the input consistency for subsequent spatio-temporal normalization processing.

[0028] For TXT format meteorological data, it is usually a simple text record, which may contain station observation data, etc. The plugin reads the data line by line according to the predefined text format rules. For example, for a TXT file where each line contains information such as station number, time, temperature, humidity, etc., the output is a unified binary file in the same format as S1 to ensure the input consistency for subsequent spatio-temporal normalization processing.

[0029] According to the application requirements, the spatial resolution is set to 0.1°×0.1° and the temporal resolution is set to hourly. The binary file is processed for spatiotemporal consistency and a data file with uniform spatial resolution and temporal resolution (element + time naming + resolution) is output (e.g.<TEMP_2024051001_0.1deg.bin> ).

[0030] In step S2, for data with a resolution higher than the target spatial resolution, an interpolation method is applied to downscale the data; for data with a resolution lower than the target spatial resolution, an interpolation method is applied to upscale the data.

[0031] In some embodiments, in step S2, applying an interpolation method to downscale data with a resolution higher than the target spatial resolution includes: Read the target resolution parameters, generate the target grid coordinate system, use the Kriging interpolation method to downscale the data with a resolution higher than the target spatial resolution, calculate the target grid value, and write the downscaled data file; For data with a resolution lower than the target spatial resolution, applying the interpolation method to perform data upscaling processing includes: Based on the target resolution grid, the value of the grid point is calculated using an inverse distance weighted formula according to the value of the low-resolution data point within a preset range of each grid point and its distance from the grid point.

[0032] Specifically, for data with a resolution higher than the target spatial resolution (the original resolution of satellite data is 0.01°×0.01°), data downscaling is performed, the target resolution parameter (0.1°×0.1°) is read, and the target grid coordinate system (starting longitude, ending longitude, 0.1) (starting latitude, ending latitude, 0.1) is generated. Kriging interpolation is used for downscaling, the target grid value is calculated, and written into the downscaled binary file. The output naming rules (such as<TEMP_2024051001_0.1deg.bin> ).

[0033] The Kriging interpolation method is as follows:

[0034] (λ i is the weight, determined by the variogram model).

[0035] For data with a spatial resolution lower than the target resolution (the numerical weather prediction model has a resolution of 0.25°×0.25°), data upscaling is performed using the inverse distance weighted interpolation method. Based on the target resolution grid, for each grid point, the value of the grid point is calculated according to the values of the surrounding low-resolution data points and their distances from the grid point through the inverse distance weighted formula. The inverse distance weighted formula is as follows:

[0036] (where Z ( x 0) is the value of the target grid point, Z ( x i ) is the value of the surrounding low-resolution data point, d ( x i , x 0) is the distance between the data point and the target grid point, p is the weight exponent taken as 1).

[0037] Write the calculated value of the target grid point into a new binary file to complete the upscaling process. The output naming rule is (e.g., <TEMP_2024051001_0.1deg.bin>).

[0038] In step S3, if there are missing or incorrect hourly data in the data file, linear interpolation is performed based on the data at the previous and next time points; non-hourly data is resampled in time to convert it into hourly data, and the converted hourly data is obtained; the linearly interpolated hourly data and the converted hourly data are merged to obtain the meteorological data input to the model.

[0039] In some embodiments, in step S3, for daily precipitation data, the total daily precipitation is evenly distributed over 24 hours of the day to obtain an estimated value of hourly precipitation data; or in combination with the precipitation diurnal variation characteristic model of the region, using historical hourly precipitation data, the average proportion of hourly precipitation in the daily total is statistically calculated, and the daily precipitation data is distributed to the hourly scale.

[0040] Specifically, for hourly data, first check the continuity and accuracy of the data timestamps. By traversing the timestamps in the data, determine whether there are missing or incorrect situations. When it is found that there are missing or incorrect timestamps in the temperature data, linear interpolation is performed based on the data at the previous and next time points. If the timestamp of the temperature at a certain moment is missing, and the temperatures at the previous and next adjacent moments are T1 and T2 respectively, and the time interval is Δt, and the time interval between the missing moment and the previous moment is Δt1, then the estimated value of the temperature at the missing moment is calculated as follows,

[0041] For precipitation data, when there are missing or incorrect measurements, the data at adjacent time points and the records of adjacent stations are combined for judgment and correction. If the precipitation at a certain station and time point is missing, and there is precipitation recorded at an adjacent station at that time, and it is judged that precipitation may exist based on the precipitation trends at adjacent time points, the precipitation data of the adjacent station and the historical precipitation ratio of this station are referred to for correction.

[0042] The naming rule for the output after correction (such as <TEMP_2024051001_0.1deg.bin>).

[0043] First, check the data timestamps. For non-hourly data such as every 3 hours, every 6 hours, and daily data, perform time resampling to convert it into hourly data. For daily precipitation data, evenly distribute the total daily precipitation over the 24 hours of the day to obtain an estimated hourly precipitation data value. It is also possible to combine the precipitation diurnal variation characteristic model of this region and use historical hourly precipitation data to statistically calculate the average proportion of hourly precipitation in the daily total (by season / weather type). For example, in a certain region, summer precipitation is mostly concentrated in the afternoon (60% from 12:00 to 18:00), then a higher weight is assigned, and the daily precipitation data is more reasonably distributed to the hourly scale. For the total daily precipitation , distribute according to the weight ratio:

[0044] where is the weight for the h-th hour, is the total daily weight.

[0045] Through this method, the daily precipitation data is more accurately distributed to the hourly scale, and the naming rule for the output is (such as <TEMP_2024051001_0.1deg.bin>).

[0046] In step S4, initialize and configure the multi-hazard collaborative processing engine to obtain grid-based terrain data, soil infiltration characteristic parameters, and vegetation interception and evapotranspiration parameters.

[0047] In some embodiments, in the said step S4, the initialization and configuration of the multi-hazard collaborative processing engine includes: Generate grid-based terrain data with a resolution of 10 km within the basin range based on digital elevation model data; Determine the soil infiltration characteristic parameters based on soil type data; Calculate the vegetation interception and evapotranspiration parameters using vegetation cover data.

[0048] Specifically, based on the output binary data file after spatio-temporal normalization (such as <TEMP_2024051001_0.1deg.bin>), it is used as the unified input data for the multi-hazard collaborative processing engine. These data contain various meteorological elements (temperature, precipitation, wind speed, etc.) within the Lancang-Mekong Basin, and have been standardized with a spatial resolution of 0.1°×0.1° and an hourly temporal resolution.

[0049] Meanwhile, in combination with geographical information data such as topographic and geomorphic data (Digital Elevation Model DEM), soil type data, and vegetation cover data within the basin, the multi-hazard collaborative processing engine is initialized and configured. Specifically, it includes: Based on the Digital Elevation Model (DEM) data, grid-based terrain data with a resolution of 10 km within the basin is generated for grid division and terrain simulation of the subsequent flood model; According to the soil type data, the infiltration characteristic parameters of the soil are determined to provide a basis for soil infiltration calculation in the flood model; Using the vegetation cover data to calculate parameters such as vegetation interception and evapotranspiration to improve the simulation of the hydrological process within the basin by the model.

[0050] In step S5, the meteorological data, grid-based terrain data, soil infiltration characteristic parameters, and vegetation interception and evapotranspiration parameters are applied to construct a flood forecasting model, a drought forecasting model, and a landslide risk forecasting model.

[0051] In some embodiments, in step S5, the application of the meteorological data, grid-based terrain data, soil infiltration characteristic parameters, and vegetation interception and evapotranspiration parameters to construct a flood forecasting model, a drought forecasting model, and a landslide risk forecasting model includes: Adopting a distributed hydrological-hydrodynamic model based on physical processes (VIC-CaMa-Flood model). The input meteorological data serves as the real-time driving data of the model. Combining the initialized geographical information parameters, it simulates the grid infiltration and runoff generation of complex terrain and soil characteristics in different grids under the influence of different precipitation spatio-temporal distributions within the basin; by calculating the generation amounts of surface runoff and subsurface runoff, and using the hydrodynamic grid confluence mechanism, it calculates and predicts the dynamic propagation process of floods within the basin and the changes in hydrological elements; the geographical information parameters include: grid-based terrain data, soil infiltration characteristic parameters, and vegetation interception and evapotranspiration parameters; the hydrological elements include: flow rate and water level; Construct a drought prediction model using indicators such as the Standardized Precipitation Evapotranspiration Index (SPEI). Extract precipitation and temperature from the meteorological data, calculate the potential evapotranspiration through the Penman-Monteith method, and then calculate the SPEI values at different time scales based on the precipitation data and potential evapotranspiration; use the SPEI values as indicators for drought monitoring and prediction, and adopt an autoregressive moving average model to predict future SPEI values based on the historical SPEI value sequence, so as to judge the development trend of drought. Combine the precipitation data, terrain slope data, and soil stability data in the meteorological data to establish geological landslide risk assessment indicators; use machine learning algorithms to analyze the relationship between geological landslide risk assessment indicators and the probability of geological landslides, construct a geological landslide risk prediction model, and conduct real-time assessment and prediction of the geological landslide risks in different regions of the basin.

[0052] In step S6, apply the flood prediction model, drought prediction model, and geological landslide risk prediction model to carry out multi-hazard coupling calculations; according to the results of the multi-hazard coupling calculations, output and visualize the multi-hazard risk warning results.

[0053] In some embodiments, in the step S6, the outputting and visualizing the multi-hazard risk warning results according to the results of the multi-hazard coupling calculations includes: Perform multi-threaded parallel normalization processing on the forecast results of each hazard and the results of the multi-hazard coupling calculations obtained by the multi-hazard collaborative processing engine to obtain normalized data; Put the normalized data into the channels of a grayscale image to generate a grayscale image result file; Compress the grayscale image result file using a lossless compression algorithm; Use WebGL (Web Graphics Library) graphics rendering technology to directly load and render the compressed grayscale image result file, and display the multi-hazard risk warning results in the form of maps and charts.

[0054] Specifically, on the basis of completing the calculations of each single-hazard forecasting model, multi-hazard coupling calculations are carried out. Taking the coupling of typhoon and flood as an example: To achieve the coupling calculation of typhoon path prediction and the 100-meter hourly interval flood model, regional meteorological numerical model forecasts are carried out in the area affected by the typhoon. The typhoon evolution results are integrated with the meteorological numerical model to generate a regional-scale meteorological element forecast field with a temporal and spatial resolution of 10 km every 3 hours. Subsequently, the meteorological element forecast field is processed through steps S005 to S009, and a binary data file after spatio-temporal normalization (such as <TEMP_2024051001_0.1deg.bin>) is output. This data is applied to drive the distributed hydrological-hydrodynamic model within the Lancang-Mekong River Basin to calculate the grid runoff generation and confluence processes of the hydrological-hydrodynamic model with a resolution of 10 km every 3 hours in the area affected by the typhoon. Subsequently, based on the historical base flood threshold data and the regional high-precision base river network underlying surface data, a 100-meter-level flood risk forecast, flood inundation depth and extent forecast within the calculation area are realized through an interpolation downscaling algorithm, thereby realizing a one-way loose coupling architecture between typhoon path prediction and the hydrological-hydrodynamic model, and carrying out flood risk forecasting under the influence of typhoons. For example, when a typhoon approaches a certain area in the Lancang-Mekong River Basin, according to the typhoon path prediction, the meteorological field data of the future hourly precipitation intensity distribution in this area is input into the hydrological-hydrodynamic model, and the flood inundation depth, inundation extent and flood risk are calculated in real-time and rolled, realizing the collaborative calculation of the two and improving the early warning accuracy.

[0055] The forecast results of each hazard (such as flood water level, drought level, geological landslide risk level, etc.) and the multi-hazard comprehensive risk assessment results calculated by the multi-hazard collaborative processing engine are subjected to multi-threaded parallel normalization processing. Taking the two-dimensional array data corresponding to one element (such as drought level) at one time step (such as 01:00 on May 10, 2024) as an example. First, calculate the minimum value and the maximum value of this two-dimensional array data, let the minimum value be min and the maximum value be max. Then, normalize each data point value in the array, and the formula is: normalized_value = (value - min) / (max - min) × 255. In this way, the data value is mapped to the range of 0 - 255, and the result file after normalization processing (such as <TEMP_2024051001_0.1deg_nv.bin>) is output for subsequent placement into the grayscale image channel.

[0056] Put the normalized data into the channels of the grayscale image to generate a grayscale image result file (such as <TEMP_2024051001_0.1deg_nv.png>). When generating the grayscale image, different mapping methods can be selected according to the characteristics of meteorological data and visualization requirements. For example, for temperature data, low temperatures can correspond to black (value 0) in the grayscale image, high temperatures can correspond to white (value 255), and intermediate temperature values can correspond to different shades of gray. For precipitation data, no precipitation can correspond to black, and heavy precipitation can correspond to white, visually showing the distribution of precipitation intensity through grayscale changes.

[0057] Perform data compression and transmission on the output file after normalization processing (such as <TEMP_2024051001_0.1deg_500hpa.png>). In terms of data compression, use a lossless compression algorithm (such as the PNG-LZ77 algorithm) to compress and output the grayscale image result file. The PNG-LZ77 algorithm reduces the file size by finding repetitive byte sequences in the data and representing them with shorter codes. After compression, when the grayscale image result file is transmitted between the front and back ends, the transmission efficiency can be greatly improved. For example, an uncompressed grayscale image file originally sized 1MB may be reduced to a few hundred KB after being compressed by the PNG-LZ77 algorithm, significantly shortening the transmission time.

[0058] Use visualization technology to display the multi-hazard risk warning results in an intuitive form such as maps and charts. When rendering on the front end, use front-end graphics rendering technologies such as WebGL to directly load the compressed grayscale image result file (such as <TEMP_2024051001_0.1deg_500hpa.png>) for rendering. WebGL can utilize the graphics processing capabilities of the browser to quickly render the grayscale image. At the same time, provide an interactive query function for multi-hazard warning results. Users can view detailed risk warning information by selecting different conditions such as time and region, providing support for disaster prevention and mitigation decision-making in the Lancang-Mekong River Basin.

[0059] In summary, by combining the multi-source heterogeneous data adaptive parsing engine and the multi-hazard meteorological risk forecasting collaborative computing technology, the present invention solves the key problems in meteorological services in the Lancang-Mekong River Basin. In terms of data heterogeneity, traditional methods require manual writing of parsing codes for different formats, which is time-consuming and prone to human errors. The multi-source heterogeneous data adaptive parsing engine of the present invention can quickly convert data in different formats into a unified format, and the data preprocessing efficiency can be increased by ten times. In traditional technologies, inconsistent spatio-temporal benchmarks lead to difficulties in data integration. The spatio-temporal normalization model established by the present invention automatically matches data with different resolutions and time scales. This ensures the physical consistency of data fusion and provides a reliable basis for subsequent data analysis, resulting in a significant improvement in data integration efficiency compared to existing technologies and more reliable analysis results.

[0060] Regarding the bottleneck of service efficiency, traditional query and visualization technologies have poor real-time response capabilities when facing large-scale heterogeneous data, and there are serious delays in high-concurrency scenarios. The present invention uses the grayscale map result file for front-end rendering display, and normalizes the queried meteorological data to generate a grayscale map. The grayscale map file is compressed using a lossless compression algorithm, significantly reducing the file size and improving the front-end and back-end transmission efficiency, providing a smoother and more efficient experience for users.

[0061] In terms of early warning technology, it fills the parallel computing and coupling of the multi-hazard meteorological risk forecasting model for overseas regions. The present invention develops the multi-hazard meteorological risk forecasting collaborative computing technology, researches and develops disaster weather forecasting and early warning technologies and products such as heavy rain, high temperature, typhoon, etc. to support meteorological services in the Lancang-Mekong River Basin, and adapts to the characteristics of the meteorological and hydrological underlying surface in the Lancang-Mekong River Basin to develop regionalized flood, drought, and geological landslide early warning technologies and products.

[0062] The second aspect of the present invention discloses a multi-hazard risk early warning system for the Lancang-Mekong River Basin. Figure 2 FIG. is a structural diagram of a multi-hazard risk early warning system for the Lancang-Mekong River Basin according to an embodiment of the present invention; as Figure 2 shown, the system 100 includes: A first processing module 101, configured to parse NetCDF data, GRIB data, HDF data, and TXT data, and output a binary file with a unified format; perform spatio-temporal consistency processing on the binary file, and output a data file with a unified spatial resolution and time resolution; A second processing module 102, configured to perform data downscaling processing on data with a resolution higher than the target spatial resolution by using the interpolation method; perform data upscaling processing on data with a resolution lower than the target spatial resolution by using the interpolation method; The third processing module 103 is configured to perform linear interpolation based on the data at the previous and subsequent time points if there are missing or incorrect hourly data in the data file; perform time resampling on the non-hourly data to convert it into hourly data, obtaining the converted hourly data; and merge the linearly interpolated hourly data and the converted hourly data to obtain the meteorological data input to the model. The fourth processing module 104 is configured to perform initialization configuration on the multi-hazard collaborative processing engine to obtain grid base terrain data, soil infiltration characteristic parameters, and vegetation interception and evapotranspiration parameters. The fifth processing module 105 is configured to apply the meteorological data, grid base terrain data, soil infiltration characteristic parameters, and vegetation interception and evapotranspiration parameters to construct a flood forecasting model, a drought forecasting model, and a geological landslide risk forecasting model. The sixth processing module 106 is configured to perform multi-hazard coupled calculations using the flood forecasting model, the drought forecasting model, and the geological landslide risk forecasting model; and output and visualize the multi-hazard risk warning results based on the results of the multi-hazard coupled calculations.

[0063] According to the system of the second aspect of the present invention, the first processing module 101 is specifically configured to set the spatial resolution to 0.1°×0.1° and the time resolution to hourly, perform spatio-temporal consistency processing on the binary file, and output a data file with a unified spatial resolution and time resolution.

[0064] Specifically, the NetCDF format is commonly used to store scientific data and has a complex data structure. Call the NetCDF library to read the file header, extract variables (such as temperature, humidity), dimensions (longitude, latitude, altitude layer), and timestamps. Map the variable names to a unified physical element code according to the rules (such as TMP→temperature, RH→humidity). Write the three-dimensional data (altitude layer×longitude×latitude) to the binary file in descending order of altitude and ascending order of longitude / latitude. The output is a unified binary file (element + time naming) (such as <TEMP_2024051001.bin>), ensuring the input consistency for subsequent spatio-temporal normalization processing.

[0065] The GRIB format is mainly used for storing meteorological numerical prediction data. The plugin uses the GRIB decoding library to parse the GRIB messages. It first identifies the GRIB version number, and there are slight differences in data organization for different versions of the GRIB format. Then it parses key parts such as the indication section and the identification section to obtain information such as the physical elements of the data (such as air pressure, wind speed), grid definition (spatial resolution), and time identification. The output is a unified binary file in the same S1 format, ensuring the input consistency for subsequent spatio-temporal normalization processing.

[0066] The HDF format can store various types of data and has a hierarchical data structure. Traverse the HDF group structure to locate the meteorological dataset (such as / VIIRS / Channel_IR). Extract the data matrix and geolocation information (such as scan angle, projection parameters), and convert them into a longitude-latitude grid. For multi-band data (such as infrared, visible light), store them in separate files respectively. Output as a unified binary file in the same S1 format to ensure the input consistency for subsequent spatio-temporal normalization processing.

[0067] For meteorological data in TXT format, it is usually a simple text record that may contain station observation data, etc. The plugin reads the data line by line according to the predefined text format rules. For example, for a TXT file where each line contains information such as station number, time, temperature, humidity, etc., output as a unified binary file in the same S1 format to ensure the input consistency for subsequent spatio-temporal normalization processing.

[0068] According to the application requirements, set the spatial resolution to 0.1°×0.1° and the time resolution to hourly, and perform spatio-temporal consistency processing on the binary file, and output a data file with unified spatial resolution and time resolution (element + time naming + resolution) (such as <TEMP_2024051001_0.1deg.bin>).

[0069] According to the system of the second aspect of the present invention, the second processing module 102 is specifically configured that for data with a resolution higher than the target spatial resolution, the data downscaling process using the interpolation method includes: Read the target resolution parameters, generate the target grid coordinate system, and use the Kriging interpolation method to perform downscaling processing on the data with a resolution higher than the target spatial resolution, calculate the target grid values, and write them into the downscaled data file; For data with a resolution lower than the target spatial resolution, the data upscaling process using the interpolation method includes: Based on the target resolution grid, according to the values of the low-resolution data points within the preset range of each grid point and their distances from the grid point, calculate the value of the grid point through the inverse distance weighting formula.

[0070] Specifically, for data with a resolution higher than the target spatial resolution (the original resolution of satellite data is 0.01°×0.01°), perform data downscaling processing, read the target resolution parameters (0.1°×0.1°), generate the target grid coordinate system (starting longitude, ending longitude, 0.1)(starting latitude, ending latitude, 0.1), use the Kriging interpolation method for downscaling processing, calculate the target grid values, and write them into the downscaled binary file, and the output naming rule is (such as <TEMP_2024051001_0.1deg.bin>).

[0071] The Kriging interpolation method is as follows:

[0072] (where λ i is the weight, determined by the variogram model).

[0073] For data with a spatial resolution lower than the target resolution (numerical weather prediction model 0.25°×0.25°), data upscaling is performed using the inverse distance weighted interpolation method. Based on the target resolution grid, for each grid point, the value of the grid point is calculated through the inverse distance weighted formula according to the values of the surrounding low-resolution data points and their distances from the grid point. The inverse distance weighted formula is as follows:

[0074] (where Z ( x 0) is the value of the target grid point, Z ( x i ) is the value of the surrounding low-resolution data point, d ( x i , x 0) is the distance between the data point and the target grid point, p and the weight exponent is taken as 1).

[0075] Write the calculated value of the target grid point into a new binary file to complete the upscaling process. The output naming rule is (such as <TEMP_2024051001_0.1deg.bin>).

[0076] According to the system of the second aspect of the present invention, the third processing module 103 is specifically configured to, for daily precipitation data, evenly distribute the total daily precipitation amount over 24 hours of the day to obtain an hourly precipitation data estimate; or combine the precipitation diurnal variation characteristic model of the region and use historical hourly precipitation data to statistically calculate the average proportion of hourly precipitation in the daily total amount, and distribute the daily precipitation data to the hourly scale.

[0077] Specifically, for hourly data, first check the continuity and accuracy of the data timestamps. By traversing the timestamps in the data, determine whether there are missing or incorrect situations. When it is found that there are missing or incorrect timestamps in the temperature data, linear interpolation is performed based on the data at the previous and subsequent time points. If the temperature timestamp is missing at a certain moment, and the temperatures at the adjacent previous and subsequent moments are T1 and T2 respectively, and the time interval is Δt, and the time interval between the missing moment and the previous moment is Δt1, then the estimated value of the temperature at the missing moment is calculated as follows,

[0078] For precipitation data, when there are missing or incorrect measurements, the data at adjacent time points and the records of adjacent stations are combined for judgment and correction. If the precipitation at a certain station and time is missing, and there is precipitation recorded at an adjacent station at that time, and it is judged that there may be precipitation based on the precipitation trend at adjacent time points, the precipitation data of the adjacent station and the historical precipitation ratio of this station are referred to for correction.

[0079] The naming rule for the corrected output is (e.g., <TEMP_2024051001_0.1deg.bin>).

[0080] First, check the data timestamps. For non-hourly data such as every 3 hours, every 6 hours, and daily data, perform time resampling to convert it into hourly data. For daily precipitation data, evenly distribute the total daily precipitation over 24 hours of the day to obtain an estimated hourly precipitation data value. It is also possible to combine the precipitation diurnal variation characteristic model of this region, use historical hourly precipitation data, and statistically calculate the average proportion of hourly precipitation in the daily total (by season / weather type). For example, in a certain region, summer precipitation is mostly concentrated in the afternoon (the proportion from 12:00 to 18:00 is 60%), then a higher weight is assigned to more reasonably distribute the daily precipitation data to the hourly scale. For the total daily precipitation , distribute according to the weight ratio:

[0081] where is the weight of the h-th hour, and is the total daily weight.

[0082] By this method, the daily precipitation data is more accurately distributed to the hourly scale, and the naming rule for the output is (e.g., <TEMP_2024051001_0.1deg.bin>).

[0083] According to the system of the second aspect of the present invention, the fourth processing module 104 is specifically configured that the initialization configuration of the multi-disaster collaborative processing engine includes: Generate grid base terrain data with a resolution of 10 km within the basin range based on digital elevation model data; Determine the soil infiltration characteristic parameters based on soil type data; Calculate the vegetation interception and evapotranspiration parameters using vegetation cover data.

[0084] Specifically, based on the output spatio-temporally normalized binary data file (such as <TEMP_2024051001_0.1deg.bin>), it is used as the unified input data of the multi-disaster collaborative processing engine. These data contain various meteorological elements (temperature, precipitation, wind speed, etc.) within the Lancang-Mekong River Basin, and have been standardized with a spatial resolution of 0.1°×0.1° and an hourly time resolution.

[0085] Meanwhile, in combination with geographical information data such as topographic and geomorphic data (Digital Elevation Model - DEM), soil type data, and vegetation cover data within the basin, the multi - disaster collaborative processing engine is initialized and configured. Specifically, it includes: Based on the Digital Elevation Model (DEM) data, grid - based terrain data with a resolution of 10 km within the basin is generated for grid division and terrain simulation of the subsequent flood model; According to the soil type data, the infiltration characteristic parameters of the soil are determined to provide a basis for the soil infiltration calculation in the flood model; Using the vegetation cover data, parameters such as vegetation interception and evapotranspiration are calculated to improve the simulation of the hydrological process within the basin by the model.

[0086] According to the system of the second aspect of the present invention, the fifth processing module 105 is specifically configured that the construction of the flood forecasting model, drought forecasting model, and geological landslide risk forecasting model by applying the meteorological data, grid - based terrain data, soil infiltration characteristic parameters, and vegetation interception and evapotranspiration parameters includes: Adopt a distributed hydrological - hydrodynamic model (VIC - CaMa - Flood model) based on physical processes. The input meteorological data is used as the real - time driving data of the model. Combining the initialized geographical information parameters, it simulates the grid infiltration and runoff generation of complex terrain and soil characteristics in different grids under the influence of different precipitation spatio - temporal distributions within the basin; by calculating the generation amounts of surface runoff and subsurface runoff, and using the hydrodynamic grid confluence mechanism, the dynamic propagation process of floods within the basin and the changes of hydrological elements are calculated and predicted; the geographical information parameters include: grid - based terrain data, soil infiltration characteristic parameters, and vegetation interception and evapotranspiration parameters; hydrological elements include: flow rate and water level; Use indicators such as the Standardized Precipitation Evapotranspiration Index (SPEI) to construct a drought forecasting model. Precipitation and temperature are extracted from the meteorological data, the potential evapotranspiration is calculated by the Penman - Monteith method, and then the SPEI values at different time scales are calculated based on the precipitation data and potential evapotranspiration; using the SPEI values as indicators for drought monitoring and prediction, an autoregressive moving average model is adopted to predict the future SPEI values according to the historical SPEI value sequence, thereby judging the development trend of drought; Combining the precipitation data, terrain slope data, and soil stability data in the meteorological data, a geological landslide risk assessment index is established; using machine learning algorithms, the relationship between the geological landslide risk assessment index and the probability of geological landslide occurrence is analyzed, and a geological landslide risk forecasting model is constructed to conduct real - time assessment and prediction of the geological landslide risks in different regions within the basin.

[0087] For the system according to the second aspect of the present invention, the sixth processing module 106 is specifically configured to perform multi-hazard risk warning result output and visualization based on the result of multi-hazard coupling calculation, including: Perform multi-threaded parallel normalization processing on the forecast results of each hazard type calculated by the multi-hazard collaborative processing engine and the result of multi-hazard coupling calculation to obtain normalized data; Put the normalized data into the channels of a grayscale image to generate a grayscale image result file; Compress the grayscale image result file using a lossless compression algorithm; Use WebGL graphics rendering technology to directly load and render the compressed grayscale image result file, and display the multi-hazard risk warning results in the form of maps and charts.

[0088] Specifically, on the basis of completing the calculation of each single-hazard forecast model, multi-hazard coupling calculation is carried out. Taking the coupling of typhoon and flood as an example: To realize the coupling calculation of typhoon path prediction and the 100-meter hourly interval model of flood, regional meteorological numerical model forecasting is carried out in the area affected by the typhoon, and the typhoon evolution result is fused with the meteorological numerical model to generate a regional-scale meteorological element forecast field with a spatial-temporal resolution of 10 km every 3 hours. Subsequently, the meteorological element forecast field is processed through steps S005 to S009, and a binary data file after spatio-temporal normalization (such as <TEMP_2024051001_0.1deg.bin>) is output. Applying this data and driving the distributed hydrological-hydrodynamic model within the Lancang-Mekong River Basin, the grid runoff and confluence processes of the hydrological-hydrodynamic model with a resolution of 10 km every 3 hours in the area affected by the typhoon are calculated. Subsequently, based on the historical base flood threshold data and the regional high-precision base river network underlying surface data, a 100-meter-level flood risk forecast, flood inundation depth and range forecast within the calculation area are realized through an interpolation downscaling algorithm, so as to realize a one-way loose coupling architecture between typhoon path prediction and the hydrological-hydrodynamic model, and carry out flood risk forecasting under the influence of typhoons. For example, when a typhoon approaches a certain area in the Lancang-Mekong River Basin, according to the typhoon path prediction, the meteorological field data of the future hourly precipitation intensity distribution in this area is input into the hydrological-hydrodynamic model, and the flood inundation depth, inundation range and flood risk are calculated in real-time and rolled, realizing the collaborative calculation of the two and improving the warning accuracy.

[0089] The forecast results of various disaster types calculated by the multi-disaster collaborative processing engine (such as flood water levels, drought levels, geological landslide risk levels, etc.) and the multi-disaster comprehensive risk assessment results are subjected to multi-threaded parallel normalization processing. Taking the two-dimensional array data corresponding to one element (such as the drought level) at one time step (such as 01:00 on May 10, 2024) as an example. First, calculate the minimum and maximum values of this two-dimensional array data, let the minimum value be min and the maximum value be max. Then, perform normalization processing on each data point value in the array, and the formula is: normalized_value = (value - min) / (max - min) × 255. In this way, the data values are mapped to the range of 0 - 255, and the output result file of the normalization process is such as <TEMP_2024051001_0.1deg_nv.bin>), so as to be put into the grayscale image channel later.

[0090] Put the normalized data into the channel of the grayscale image to generate a grayscale image result file (such as <TEMP_2024051001_0.1deg_nv.png>). When generating the grayscale image, different mapping methods can be selected according to the characteristics of meteorological data and visualization requirements. For example, for temperature data, low temperatures can correspond to black (value 0) in the grayscale image, high temperatures can correspond to white (value 255), and intermediate temperature values can correspond to different shades of gray. For precipitation data, no precipitation can correspond to black, and heavy precipitation can correspond to white, visually showing the distribution of precipitation intensity through gray-scale changes.

[0091] Perform data compression and transmission on the output file after normalization processing (such as <TEMP_2024051001_0.1deg_500hpa.png>). In terms of data compression, a lossless compression algorithm (such as the PNG-LZ77 algorithm) is used to compress the grayscale image result file and output. The PNG-LZ77 algorithm reduces the file size by finding repeated byte sequences in the data and representing them with shorter encodings. After compression, when the grayscale image result file is transmitted between the front end and the back end, the transmission efficiency can be greatly improved. For example, an uncompressed grayscale image file originally with a size of 1MB may be reduced to a few hundred KB after being compressed by the PNG-LZ77 algorithm, and the transmission time is significantly shortened.

[0092] Using visualization technology, the multi-hazard risk early warning results are presented in intuitive forms such as maps and charts. When rendering on the front end, front-end graphics rendering technologies such as WebGL are used to directly load the compressed grayscale image result file (such as <TEMP_2024051001_0.1deg_500hpa.png>) for rendering. WebGL can utilize the graphics processing capabilities of the browser to quickly render grayscale images. Meanwhile, an interactive query function for multi-hazard early warning results is provided. Users can view detailed risk early warning information by selecting different conditions such as time and region, providing support for disaster prevention and mitigation decision-making in the Lancang-Mekong River Basin.

[0093] The third aspect of the present invention discloses an electronic device. The electronic device includes a memory and a processor. The memory stores a computer program. When the processor executes the computer program, the steps in any one of the multi-hazard risk early warning methods in the first aspect disclosed by the present invention are implemented.

[0094] Figure 3 For a structural diagram of an electronic device according to an embodiment of the present invention, as Figure 3 shown, the electronic device includes a processor, a memory, a communication interface, a display screen, and an input device connected through a system bus. Among them, the processor of the electronic device is used to provide computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the electronic device is used to communicate with external terminals in a wired or wireless manner. The wireless manner can be achieved through WIFI, carrier networks, near-field communication (NFC), or other technologies. The display screen of the electronic device can be a liquid crystal display screen or an electronic ink display screen. The input device of the electronic device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads provided on the housing of the electronic device, or an external keyboard, touchpad, or mouse, etc.

[0095] Those skilled in the art can understand that Figure 3 the structure shown in

[0096] is only a structural diagram of a part related to the technical solution of the present disclosure, and does not constitute a limitation on the electronic device to which the application solution of the present application is applied. The specific electronic device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0097] Please note that the technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as within the scope described in this specification. The above embodiments only express several implementation manners of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A multi-hazard risk early warning method for the Lancang-Mekong River Basin, characterized in that, The method includes: Step S1: Parse NetCDF data, GRIB data, HDF data, and TXT data, and output a binary file with a unified format; perform spatio-temporal consistency processing on the binary file, and output a data file with a unified spatial resolution and temporal resolution; Step S2: For data with a resolution higher than the target spatial resolution, apply interpolation method for data downscaling; for data with a resolution lower than the target spatial resolution, apply interpolation method for data upscaling; Step S3: If there are missing or incorrect hourly data in the data file, perform linear interpolation based on the data at the previous and next time points; perform time resampling on non-hourly data to convert it into hourly data, and obtain the converted hourly data; merge the linearly interpolated hourly data and the converted hourly data to obtain the meteorological data for model input; Step S4: Initialize and configure the multi-hazard collaborative processing engine to obtain grid-based terrain data, soil infiltration characteristic parameters, and vegetation interception and evapotranspiration parameters; Step S5: Apply the meteorological data, the grid-based terrain data, the soil infiltration characteristic parameters, and the vegetation interception and evapotranspiration parameters to construct a flood forecasting model, a drought forecasting model, and a geological landslide risk forecasting model; Step S6: Apply the flood forecasting model, the drought forecasting model, and the geological landslide risk forecasting model to carry out multi-hazard coupling calculations; according to the results of the multi-hazard coupling calculations, output and visualize the multi-hazard risk warning results.

2. The multi-hazard risk early warning method for the Lancang-Mekong River Basin according to claim 1, wherein, In the step S1, Set the spatial resolution to 0.1°×0.1° and the temporal resolution to hourly, perform spatio-temporal consistency processing on the binary file, and output a data file with a unified spatial resolution and temporal resolution.

3. The multi-hazard risk early warning method for the Lancang-Mekong River Basin according to claim 1, wherein, In the step S2, for the data with a resolution higher than the target spatial resolution, the data downscaling process using the interpolation method includes: Read the target resolution parameters, generate a target grid coordinate system, use the Kriging interpolation method to perform downscaling on the data with a resolution higher than the target spatial resolution, calculate the target grid values, and write them into the downscaled data file; For the data with a resolution lower than the target spatial resolution, the data upscaling process using the interpolation method includes: Based on the target resolution grid, according to the values of the low-resolution data points within the preset range of each grid point and their distances from the grid point, calculate the value of the grid point through the inverse distance weighted formula.

4. A multi-hazard risk early warning method for the Lancang-Mekong River Basin according to claim 1, characterized in that In the step S3, for daily precipitation data, evenly distribute the total daily precipitation amount over 24 hours of the day to obtain an estimated hourly precipitation data value; or combine the precipitation diurnal variation characteristic model of the region, use historical hourly precipitation data, statistically calculate the average proportion of hourly precipitation in the daily total, and distribute the daily precipitation data to the hourly scale.

5. A multi-hazard risk early warning method for the Lancang-Mekong River Basin according to claim 1, characterized in that, In the step S4, the initialization and configuration of the multi-hazard collaborative processing engine include: Generate grid-based terrain data with a resolution of 10 km within the basin range according to the digital elevation model data; Determine the soil infiltration characteristic parameters based on the soil type data; Vegetation interception and evapotranspiration parameters were calculated using vegetation cover data.

6. A multi-hazard risk early warning method for the Lancang-Mekong River Basin according to claim 1, characterized in that In step S5, applying the meteorological data, the grid base terrain data, the soil permeability characteristic parameters and the vegetation interception and evapotranspiration parameters to construct a flood forecast model, a drought forecast model and a geological landslide risk forecast model includes: The input meteorological data is used as the real-time driving data of the model, combined with the geographical information parameters of the initial configuration, to simulate the grid infiltration and runoff generation of complex terrain and soil characteristics in different grids under the influence of different spatial and temporal distribution of precipitation in the basin; by calculating the amount of surface runoff and underground runoff, and using the hydrodynamic grid confluence mechanism, the dynamic propagation process of floods in the basin and the changes in hydrological elements are calculated and estimated; the geographical information parameters include: grid base terrain data, soil permeability characteristic parameters and vegetation interception and evapotranspiration parameters; hydrological elements include: flow and water level; Extracting precipitation and temperature from the meteorological data, calculating potential evapotranspiration by the Penman-Monteith method, and then calculating SPEI values at different time scales based on precipitation data and potential evapotranspiration; using the SPEI value as an indicator for drought monitoring and prediction, using an autoregressive moving average model, predicting future SPEI values based on a historical SPEI value sequence, thereby determining the development trend of drought; Combined with precipitation data, terrain slope data and soil stability data in meteorological data, a geological landslide risk assessment index is established. Using machine learning algorithms, the relationship between geological landslide risk assessment indicators and the probability of geological landslides is analyzed, and a geological landslide risk prediction model is constructed to conduct real-time assessment and prediction of geological landslide risks in different areas within the basin.

7. A multi-hazard risk early warning method for the Lancang-Mekong River Basin according to claim 1, characterized in that In step S6, the output and visualization of multi-hazard risk warning results according to the results of multi-hazard coupling calculations include: The forecast results of each disaster type calculated by the multi-disaster collaborative processing engine and the results of the multi-disaster coupling calculation are normalized by multi-threaded parallel processing to obtain normalized data; Put the normalized data into the channel of the grayscale image to generate a grayscale image result file; Use lossless compression algorithm to compress the grayscale image result file; Using WebGL graphics rendering technology, the compressed grayscale image result file is directly loaded for rendering, and the multi-hazard risk warning results are displayed in the form of maps and charts.

8. A multi-hazard risk early warning system for the Lancang-Mekong River Basin, characterized in that, The system comprises: The first processing module is configured to parse NetCDF data, GRIB data, HDF data and TXT data, and output a binary file with a unified format; perform spatiotemporal consistency processing on the binary file, and output a data file with a unified spatial resolution and temporal resolution; The second processing module is configured to apply the interpolation method to downscale the data for data with a resolution higher than the target spatial resolution; and apply the interpolation method to upscale the data for data with a resolution lower than the target spatial resolution; A third processing module, configured to perform linear interpolation based on data at adjacent time points if there are missing or incorrect hourly data in the data file; perform time resampling on non-hourly data to convert it into hourly data, obtaining the converted hourly data; and merge the linearly interpolated hourly data and the converted hourly data to obtain the meteorological data for model input. A fourth processing module, configured to perform initialization configuration on the multi-hazard collaborative processing engine to obtain grid base terrain data, soil infiltration characteristic parameters, and vegetation interception and evapotranspiration parameters. A fifth processing module, configured to construct a flood forecasting model, a drought forecasting model, and a geological landslide risk forecasting model by applying the meteorological data, the grid base terrain data, the soil infiltration characteristic parameters, and the vegetation interception and evapotranspiration parameters. A sixth processing module, configured to perform multi-hazard coupled calculations by applying the flood forecasting model, the drought forecasting model, and the geological landslide risk forecasting model; and output and visualize the multi-hazard risk warning results based on the results of the multi-hazard coupled calculations.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor. When the processor executes the computer program stored in the memory, the steps in a method for multi-hazard risk warning in the Lancang-Mekong River Basin according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium. When the computer program is executed by a processor, the steps in a method for multi-hazard risk warning in the Lancang-Mekong River Basin according to any one of claims 1 to 7 are implemented.

Images