A multi-scale mapping method between dual-polarization radar grid and hydrological simulation units

Abstract

The present invention discloses a multi-scale mapping method for dual-polarization radar grids and hydrological simulation units, comprising the following steps: inputting dual-polarization radar raw data, performing data quality control using median filtering and mathematical methods; screening rainfall echoes based on polarization parameters to determine their three-dimensional spatial coordinates; comprehensively considering the initial velocity of rainfall particles, echo height, and the microphysical processes of meteorological factors, calculating the two-dimensional coordinates of the rainfall particles landing on the ground surface through projection transformation, and reflecting hydrological units of different spatial scales based on the two-dimensional grid pixel size; constructing a rainfall spatiotemporal cube, whose changes along the time axis reflect the temporal evolution of rainfall spatial distribution, and obtaining the rainfall distribution grid and rainfall intensity mapped from dual-polarization radar rainfall echoes to multi-scale hydrological simulation units within a specified time period. The present invention effectively reduces the spatial mapping deviation of dual-polarization radar rainfall inversion, improves the spatiotemporal resolution of rainfall measurement data, and enhances the accuracy of rainfall-based hydrological simulation and prediction.

Description

Technical Field

[0001] The present invention belongs to the fields of geographic information system technology, application of Doppler radar technology and hydrological and meteorological coupling, and particularly relates to a multi-scale mapping method of a dual-polarization radar grid and a hydrological simulation unit. Background Art

[0002] Radar-based quantitative rainfall estimation is one of the primary applications of weather radar. Weather radar can quickly and accurately acquire echo information within its detection range, enabling relatively accurate calculations of rainfall intensity and its spatial distribution. Since the mid-20th century, weather radar equipment has evolved from conventional radar to a new generation of Doppler radar. Dual-polarization Doppler weather radars are now mainstream worldwide, and my country has also begun gradually implementing dual-polarization upgrades to this new generation of Doppler weather radars.

[0003] Currently, efficient coupling and spatial mapping of dual-polarization radar grids and hydrological simulation units is key to improving hydrological simulation accuracy and the effective forecast period of disasters. The finer the hydrological simulation unit, the more computational resources are consumed. However, real-time simulation and disaster forecasting require hydrological models to respond as quickly as possible. This requires dynamic and timely adjustment of the hydrological simulation unit refinement based on rainfall factors. For example, when rainfall intensity is low, the surface can be generalized and larger hydrological simulation units can be used. When rainfall varies significantly in time and space, finer hydrological simulation units can be used to more accurately locate inundated areas. However, current dual-polarization radar rainfall inversion suffers from widespread spatial mapping errors. Radar observations are three-dimensional rainfall near the ground, while hydrological simulations require two-dimensional rainfall input at the ground surface. The complex microphysical processes of raindrops near the ground result in significant errors in traditional simple vertical mapping methods. Furthermore, the comprehensive management and integrated modeling of radar rainfall data require further exploration.

[0004] Therefore, it is necessary to develop an efficient and accurate multi-scale mapping method between dual-polarization radar grids and hydrological simulation units. Summary of the Invention

[0005] To solve the above problems, the present invention discloses a multi-scale mapping method between dual-polarization radar grids and hydrological simulation units. The method fully takes into account the spatiotemporal changes, rainfall intensity, rainfall amount and other characteristics of dual-polarization radar grid rainfall data when dividing hydrological simulation units, and uses dual-polarization radar rainfall data as the dynamic attribute of the hydrological simulation unit, aiming to solve the spatial mapping deviation problem that is common in the current dual-polarization radar inversion of rainfall. The method performs multi-scale grid mapping of hydrological units on rainfall particles and constructs a rainfall space-time cube, which effectively improves the spatial resolution and accuracy of the radar inversion results, and can more accurately predict and simulate the impact of rainfall on hydrological units.

[0006] To achieve the above object, the technical solution of the present invention is as follows:

[0007] A multi-scale mapping method between a dual-polarization radar grid and a hydrological simulation unit comprises the following steps:

[0008] Step S1: Data quality control, rainfall echo screening, rainfall intensity estimation, projection conversion and construction of rainfall space-time cube are performed on the dual-polarization radar raw data in the rainfall area to achieve multi-scale mapping between the dual-polarization radar grid and the hydrological simulation unit;

[0009] Step S2: Perform data quality control on the dual polarization radar raw data, including the differential propagation phase shift parameter Φ DP Denoising, differential propagation phase constant K DP Estimation, dual polarization radar detection parameter Z H With Z DR Attenuation correction;

[0010] Step S3: using the interval boundary values ​​of the polarization parameter membership functions of the dual-polarization radar as the threshold interval for screening rainfall echoes, thereby screening rainfall echoes in heavy rain weather;

[0011] Step S4: According to Z H The rainfall conditions reflected by the parameters take into account the influence of raindrop deformation on quantitative rainfall estimation under different rainfall intensities, and are combined with Z H and K DP Calculate rainfall intensity;

[0012] Step S5: Calculate the three-dimensional spatial coordinates of the rainfall echo based on the polar coordinate storage format of the dual-polarization radar raw data, calculate the coordinates of the rainfall particles landing on the ground based on the wind field and temperature conditions in the rainfall inversion area, and construct a two-dimensional grid based on the pixel size of the surface hydrological unit;

[0013] Step S6: establishing a dual-polarization radar rainfall measurement database, and storing the radar data processing results that have completed rainfall estimation and projection conversion;

[0014] Step S7: Divide the time segments according to the length of the simulated rainfall time axis and the time scale, calculate the rainfall distribution within the segments in combination with the dual-polarization radar rainfall data stored in the database, and merge multiple segments into a rainfall space-time cube.

[0015] Furthermore, due to the influence of the detection environment and the equipment itself, the dual-polarization radar data is prone to signal noise and measurement errors. Therefore, before using the dual-polarization radar data for qualitative and quantitative applications, it is necessary to perform quality control on each physical parameter. The specific process includes: DP Denoising, K DP Estimate, Z H With Z DR Attenuation correction.

[0016] Therefore, the step S2 includes the following steps:

[0017] Step S21: Select the radial library as the smoothing window and use median filtering to smooth the differential propagation phase shift parameter Φ DP Perform denoising;

[0018] Step S22: Based on the median filter Φ DP , use the least squares method to calculate the differential propagation phase constant K DP First, select 2n+1 radial libraries of Φ in a radial direction. DP , use the least squares method to fit Φ DP The slope of K is used as the K of the n+1th radial library position DP Value, calculation formula:

[0019]

[0020] In the formula, r i represents the distance from the i-th radial bin to the radar antenna, To estimate the distance from the center of the window to the radar antenna, is the average of the 2n+1 radial library differential propagation phase shifts.

[0021] Step S23: Based on the median filter Φ DP and the initial differential phase pair Z H and Z DR The electromagnetic waves emitted by the radar are absorbed and scattered by particles such as gas and rain during the propagation process, resulting in energy attenuation, which leads to Z H and Z DR The parameter is small, while the differential propagation phase shift parameter Φ DP Not affected by electromagnetic wave attenuation. Calculate the differential offset and adjust the Z H and Z DR The parameters are supplemented and corrected, and the calculation formula is:

[0022]

[0023]

[0024] In the formula, Z H and Z DR are the horizontal reflectivity factor and differential reflectivity factor obtained by dual polarization radar detection, Z H 'With Z DR' is the horizontal reflectivity factor and differential reflectivity factor after attenuation correction; a1 and a2 are linear correction parameters. For C-band dual-polarization radar, a1 is 0.08dBZ / ° and a2 is 0.02dB / °; Φ DP is the differential propagation phase shift parameter of the echo position after median filtering, Φ DP0 is the initial differential phase, and the value of the initial phase in each radar scanning radial direction is different.

[0025] Furthermore, scattered echoes received by polarimetric radar are generated by a variety of targets, such as raindrops, hailstones, snowflakes, organisms, and ground objects. Therefore, when using radar echo data to calculate rainfall intensity, it is first necessary to filter out rainfall echoes. Because different objects or particles vary in shape, size, density, and spatial orientation, their scattering characteristics for radar electromagnetic waves differ. Consequently, the polarization parameters of these various targets vary in value. By comprehensively utilizing these polarization parameters, rainfall echoes can be filtered out during heavy rain.

[0026] Therefore, the step S3 includes the following steps:

[0027] Step S31: determining the interval boundary values ​​of each polarization parameter membership function of the dual-polarization radar target object by fuzzy logic method;

[0028] Step S32: Using the determined boundary value as a threshold interval for screening rainfall echoes, the rainfall echoes in heavy rain weather are screened.

[0029] Furthermore, there are many mathematical relationships when using dual-polarization radar data to calculate rainfall intensity. Among them, using the horizontal reflectivity factor to calculate rainfall intensity is the most commonly used method for radar quantitative rainfall estimation. Dual-polarization radar can not only provide a reflectivity factor related to precipitation, but also provide multiple polarization parameters containing droplet spectrum information. The differential propagation phase constant K DP It is not only not affected by the radar system calibration, but also related to the shape and size of raindrops. Therefore, it can also be used as one of the dual-polarization quantitative rainfall estimation operators. DP The parameter calculation of rainfall intensity in heavy rain weather is more reliable.

[0030] Therefore, the step S4 includes the following steps:

[0031] Step S41: According to Z H The magnitude of the parameter value determines the intensity of rainfall;

[0032] Step S42: Use Z according to the rainfall intensity H and K DP Calculate the rainfall intensity using the following formula:

[0033]

[0034] When Z H When the parameter value is less than 35dBZ, it means that the rainfall is weak, K DP The parameter value is easily affected by noise and is not suitable for rainfall intensity calculation in this case. (ZH) The calculation will be more accurate. H When it is greater than or equal to 35dBZ, it means that the rainfall is heavy and the raindrops are prone to deformation. (ZH) There will be a large error in the calculation of rainfall intensity. In this case, it is better to use R (KDP) Calculate the relationship.

[0035] Furthermore, the raw data of the dual-polarization radar is stored in polar coordinates, which is expressed by the elevation angle. The azimuth angle θ and slant range R represent the spatial position of the echo. Due to the elevation angle, the radar detects rain particles in the air. To map rain echoes to a surface grid at different scales, radar rainfall data must be expressed in a plane rectangular coordinate system or a geodetic coordinate system. Therefore, determining the aerial coordinates of rain echoes and projecting them onto the surface are two prerequisites for achieving multi-scale spatiotemporal representation.

[0036] Therefore, the step S5 includes the following steps:

[0037] Step S51: Calculate the spatial three-dimensional coordinates of the rainfall echo using the polar coordinates stored in the dual-polarization radar raw data. The calculation formula is:

[0038]

[0039]

[0040]

[0041] In the formula, X, Y, and Z are the three-dimensional coordinates of the rainfall echo, X0, Y0, and h are the plane coordinates and elevation of the radar antenna position, respectively, and R is the slant range. is the elevation angle of the radar antenna, θ is the azimuth angle of the scanning radial direction, R 2 / (2·R m ′) is a correction term that takes into account the curvature of the earth and the refraction of electromagnetic waves.

[0042] Step S52: Combined with the initial velocity of the rainfall particles, the height of the echo position, the wind field, and the temperature, the two-dimensional coordinates of the rainfall particles falling on the surface hydrological unit are calculated through geographic coordinate projection transformation. The relationship between the rainfall falling on the surface and the rainfall detection results in the air is expressed as:

[0043]

[0044] p=g(X,Y,Z,temp,k dp ,Z H , Z DR )

[0045] In the formula, geom and p are the coordinates of the rainfall falling to the ground and the rainfall intensity, the horizontal coordinate X, the vertical coordinate Y, the vertical coordinate Z, and the differential propagation phase constant K dp , horizontal reflectivity factor Z H , differential reflectivity factor Z DR ,speed is a parameter that can represent the position and state of rainfall particles in the air. and temp represent the wind field and temperature conditions in the rainfall retrieval area, respectively.

[0046] Step S53: Reflecting hydrological units of different spatial scales according to different pixel sizes to obtain a multi-scale rainfall grid.

[0047] Furthermore, in the process of using dual-polarization radar data for quantitative rainfall estimation, radar meteorology research mostly ignores the time information of the scanning radial and the three-dimensional coordinate information of the rainfall echo, which makes it inconvenient to further embed the radar rainfall data products into the data framework. Therefore, there is an urgent need to model and store the dual-polarization radar data and its rainfall measurement results to improve the flexibility and versatility of dual-polarization radar rainfall measurement data.

[0048] Therefore, the step S6 includes the following steps:

[0049] Step S61: constructing a dual-polarization radar rainfall measurement data logic model and creating a dual-polarization radar rainfall measurement database;

[0050] Step S62: Provide rainfall echo screening, rainfall intensity calculation, coordinate conversion and geographic projection, and multi-scale division interface, and store the radar data processing results that have completed rainfall estimation and projection conversion into the data warehouse.

[0051] Furthermore, in urban rainstorm and waterlogging simulation and other hydrological and meteorological application fields, spatial distribution information of rainfall with different time accuracy is often required as input. The rainfall information detected by dual-polarization radar equipment can accurately describe the rainfall field in the simulation area from both time and space dimensions.

[0052] The rainfall space-time cube model uses the temporal evolution of rainfall spatial distribution in a region by measuring the change in rainfall intensity or amount over time on a two-dimensional grid. The temporal scale is represented by the change in the temporal dimension, and the spatial scale is represented by the pixel size of the two-dimensional grid. The advantage of using the space-time cube model for multi-scale classification of radar rainfall inversion results lies in its simple and clear description of rainfall temporal and spatial variations, and its ease of integration into various specialized models.

[0053] Therefore, the step S7 includes the following steps:

[0054] Step S71: Divide the time segments according to the inverted rainfall time axis length and time scale, and divide t0 to t n Divide into n layers evenly, each layer is the spatial rainfall distribution at a certain moment;

[0055] Step S72: The time interval is further divided into k segments. Surface rainfall point data are filtered through spatial queries with additional time conditions. These point data are then used for spatial interpolation. Interpolation algorithm parameters are selected and executed according to the spatial scale and interpolation range requirements to obtain rainfall raster data within the minimum time segment. Finally, the rainfall distribution for time dimension index i is calculated through segmented summation.

[0056] Step S73: Repeat step S72 to calculate the rainfall distribution of n layers, and finally superimpose them on the time axis to form a rainfall space-time cube of the rainfall inversion area.

[0057] The beneficial effects of the present invention are:

[0058] (1) The present invention combines median filtering, least squares method, linear correction and other means to achieve quality control of dual-polarization radar data and improve the accuracy of dual-polarization radar grid; establishes a dual-polarization radar rainfall data logic model and creates a dual-polarization radar rainfall database, and provides multiple functional interfaces to facilitate comprehensive management and integrated modeling of radar rainfall data.

[0059] (2) The present invention comprehensively considers the microphysical processes of meteorological factors on rainfall particle groups during the projection conversion process, effectively solving the spatial mapping deviation problem that is common in dual-polarization radar rainfall inversion; by constructing a rainfall space-time cube, a multi-scale hydrological simulation cell network can be generated as needed, accurately describing the rainfall field in the simulation area from both time and space dimensions. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 is a flow chart of the present invention;

[0061] Figure 2 The rainfall time series data result of the observation station described in the present invention;

[0062] Figure 3 This is a specific application block diagram of the present invention;

[0063] Figure 4 The quality control result of the dual-polarization radar data according to the present invention;

[0064] Figure 5This is the result of a case study of quantitative rainfall estimation using dual-polarization radar rainfall measurement data according to the present invention;

[0065] Figure 6 The dual-polarization radar rainfall measurement data logic model of the present invention;

[0066] Figure 7 This is a flowchart of the dual-polarization radar rainfall data storage process and various operation interfaces described in the present invention;

[0067] Figure 8 This is a flowchart of constructing a rainfall space-time cube using dual-polarization radar rainfall measurement data according to the present invention;

[0068] Figure 9 Schematic diagram of constructing a rainfall space-time cube for the dual-polarization radar rainfall measurement data of the present invention;

[0069] Figure 10 This is the result diagram of constructing a rainfall space-time cube using dual-polarization radar rainfall measurement data as described in the present invention. DETAILED DESCRIPTION

[0070] The present invention will be further described below with reference to the accompanying drawings and specific embodiments. It should be understood that the following specific embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention.

[0071] according to Figure 1 As shown, the multi-scale mapping method of the dual-polarization radar grid and the hydrological simulation unit of the present invention specifically includes the following steps:

[0072] Step S1: Data quality control, rainfall echo screening, rainfall intensity estimation, projection conversion and construction of rainfall space-time cube are performed on the dual-polarization radar raw data in the rainfall area to achieve multi-scale mapping between the dual-polarization radar grid and the hydrological simulation unit;

[0073] Step S2: Perform data quality control on the dual polarization radar raw data, including the differential propagation phase shift parameter Φ DP Denoising, differential propagation phase constant K DP Estimation, dual polarization radar detection parameter Z H With Z DR Attenuation correction;

[0074] Step S3: using the interval boundary values ​​of the polarization parameter membership functions of the dual-polarization radar as the threshold interval for screening rainfall echoes, thereby screening rainfall echoes in heavy rain weather;

[0075] Step S4: According to Z H The rainfall conditions reflected by the parameters take into account the influence of raindrop deformation on quantitative rainfall estimation under different rainfall intensities, and are combined with Z H and K DPCalculate rainfall intensity;

[0076] Step S5: Calculate the three-dimensional spatial coordinates of the rainfall echo based on the polar coordinate storage format of the dual-polarization radar raw data, calculate the coordinates of the rainfall particles landing on the ground based on the wind field and temperature conditions in the rainfall inversion area, and construct a two-dimensional grid based on the pixel size of the surface hydrological unit;

[0077] Step S6: establishing a dual-polarization radar rainfall measurement database, and storing the radar data processing results that have completed rainfall estimation and projection conversion;

[0078] Step S7: Divide the time segments according to the length of the simulated rainfall time axis and the time scale, calculate the rainfall distribution within the segments in combination with the dual-polarization radar rainfall data stored in the database, and merge multiple segments into a rainfall space-time cube.

[0079] In this example, the NUIST-C dual-polarization radar data of a certain area on June 10, 2017 were used for data processing and spatiotemporal multi-scale rainfall inversion. The inversion results were compared with the rainfall observation data of 39 rain gauges in the area on that day. The rainfall time series data results of the rainfall stations are as follows: Figure 2 As shown by, Figure 2 Analysis shows that the peak rainfall period on June 10, 2017 was from 9:00 to 10:00.

[0080] Implementation steps such as Figure 3 As shown:

[0081] Step 1: Use Python language and pyart toolkit to perform data quality control on regional NUIST-C radar data, including differential propagation phase shift parameter Φ DP Denoising, differential propagation phase constant K DP Estimation, dual polarization radar detection parameter Z H With Z DR Attenuation correction. The detailed steps are as follows:

[0082] Step (1): Use median filtering to filter Φ DP For denoising, choosing a larger median filter window will result in Φ DP is over-smoothed, and a smaller window cannot filter out the continuous Φ DP After comparative analysis of abnormally large values, 11 radial bins were selected as smoothing windows. DP After the radial data is median filtered, most of the outliers within 20 km from the radar antenna are filtered out, and the Φ values ​​beyond 20 km from the radar antenna are filtered out. DP The value gradually increases with the distance, such as Figure 4 shown.

[0083] Step (2): Based on the median filter Φ DP , use the least squares method to calculate the differential propagation phase constant K DP First, select 2n+1 radial libraries of Φ in a radial direction. DP , use the least squares method to fit Φ DP The slope of K is used as the K of the n+1th radial library position DP Value, calculation formula:

[0084]

[0085] In the formula, r i represents the distance from the i-th radial bin to the radar antenna, To estimate the distance from the center of the window to the radar antenna, is the average of the 2n+1 radial library differential propagation phase shifts.

[0086] In this embodiment, the value of n is selected as 15 according to the size of the simulation area, that is, the fitting window is 31 radial bins, and the differential propagation phase constant K is calculated by the least square method. DP Make an estimate and get K DP Estimation results, such as Figure 5 As shown. In the range of 20km to 80km from the radar antenna, K DP The values ​​are all greater than 0 and less than 3° / km, indicating that the area is mainly a rainfall area and K DP The estimation effect is good, and no obvious fluctuations appear until outside the bright band of the zero-degree layer.

[0087] Step (3): Based on the median filter Φ DP and the initial differential phase pair Z H and Z DR The parameters are linearly corrected. The electromagnetic waves emitted by the radar are absorbed and scattered by particles such as gas and rain during the propagation process, resulting in energy attenuation. H and Z DR The parameter is small, while the differential propagation phase shift parameter Φ DP Not affected by electromagnetic wave attenuation. Calculate the differential offset and adjust the Z H and Z DR The parameters are supplemented and corrected, and the calculation formula is:

[0088]

[0089]

[0090] In the formula, Z H and Z DR are the horizontal reflectivity factor and differential reflectivity factor obtained by dual-polarization radar detection, ZH 'With Z DR ' is the horizontal reflectivity factor and differential reflectivity factor after attenuation correction. a1 and a2 are linear correction parameters. For C-band dual-polarization radar, a1 is 0.08dBZ / ° and a2 is 0.02dB / °. Φ DP is the differential propagation phase shift parameter of the echo position after median filtering, Φ DP0 is the initial differential phase, and the value of the initial phase in each radar scanning radial direction is different.

[0091] In this embodiment, by judging the continuous larger CC parameter value and the smaller Φ DP Standard deviation determines Φ DP0 (That is, search 9 radial libraries from 2 km away from the radar. If all CC parameter values ​​in the interval are greater than 0.9, and Φ DP If the standard deviation is less than 10°, the average value of the interval is used as the initial differential phase in the radial direction of the interval. Z is calculated using the linear correction formula H and Z DR Parametric attenuation results.

[0092] Step 2: Use the threshold intervals of polarization parameters of different targets to screen rainfall echoes in heavy rain weather. In this embodiment, Z H The threshold range of the parameter is 25~60dBZ, which can effectively filter out clear sky echoes and hail echoes; Z DR The threshold range of the parameter is selected as 0.5~4.0dB, K DP The threshold range for the parameters was selected to be 0.2 to 10.0° / km. These two conditions can filter out ice and snow echoes to a certain extent. The threshold range for the CC parameter was selected to be 0.98 to 1. This condition can reduce the interference of biological and ground clutter. Finally, the filtered rainfall echoes were obtained.

[0093] Step 3: According to Z H The parameter value is used to judge the intensity of rainfall, and Z H and K DP Calculate the rainfall intensity using the following formula:

[0094]

[0095] When Z H When the parameter value is less than 35dBZ, it means that the rainfall is weak, K DP The parameter value is easily affected by noise and is not suitable for rainfall intensity calculation in this case. (ZH) The calculation will be more accurate. H When it is greater than or equal to 35dBZ, it means that the rainfall is heavy and the raindrops are prone to deformation. (ZH)There will be a large error in the calculation of rainfall intensity. In this case, it is better to use R (KDP) In this example, Z is used in conjunction with the rainfall conditions. H and K DP The rainfall intensity R of the filtered rainfall echo is calculated by the parameter to obtain the quantitative rainfall estimation result. Figure 5 shown.

[0096] Step 4: In this embodiment, the polar coordinates stored in the dual-polarization radar raw data are used to calculate the spatial three-dimensional coordinates of the rainfall echo. The calculation formula is:

[0097]

[0098]

[0099]

[0100] In the formula, X, Y, and Z are the three-dimensional coordinates of the rainfall echo, X0, Y0, and h are the plane coordinates and elevation of the radar antenna position, respectively, and R is the slant range. is the elevation angle of the radar antenna, θ is the azimuth angle of the scanning radial direction, R 2 / (2·R m ′) is a correction term that takes into account the curvature of the earth and the refraction of electromagnetic waves.

[0101] The microphysical process of the rainfall particle group is simulated by comprehensively considering the initial velocity of the rainfall particles, the height of the echo position, the wind field, and the temperature. The two-dimensional coordinates of the rainfall particles falling on the surface hydrological unit are calculated through geographic coordinate projection transformation. The relationship between the rainfall falling on the surface and the rainfall detection results in the air is expressed as:

[0102]

[0103] p=g(X,Y,Z,temp,K dp , Z H , Z DR )

[0104] In the formula, geom and p are the coordinates of the rainfall falling to the ground and the rainfall intensity, the horizontal coordinate X, the vertical coordinate Y, the vertical coordinate Z, and the differential propagation phase constant K dp , horizontal reflectivity factor Z H ,speed is a parameter that can represent the position and state of rainfall particles in the air. and temp represent the wind field and temperature in the rainfall inversion area, respectively. In this embodiment, different pixel sizes of 100 meters, 200 meters, 500 meters, 1000 meters, and 2000 meters are set to reflect hydrological units of different spatial scales, thereby obtaining a multi-scale rainfall grid.

[0105] Step 5: Construct a data table structure logical model for dual-polarization radar rainfall data based on the object-relational database PostgreSQL and PostGIS extensions. Use Python and tools such as psycopg2, gdal, pyart, and pyproj to complete radar data processing and storage. The detailed steps are as follows:

[0106] Step (1): Construct a logical model of radar rainfall data such as Figure 6 The following diagrams show the Radar class table and the RadarEcho class table for radar stations. Each class diagram contains three components from top to bottom: feature name, feature spatiotemporal data and attribute variables, and feature operation interface. Create a radar rainfall database in PostgreSQL.

[0107] Step (2): Import the radar rainfall data processed from step 1 to step 4 into the radar rainfall database, and use Python functions to operate PostgreSQL to implement various interfaces including rainfall echo screening, rainfall intensity calculation, coordinate conversion and geographic projection, multi-scale division, etc. The dual polarization radar data storage process and various operation interfaces are as follows: Figure 7 shown.

[0108] Step 6: If Figure 8 As shown in the figure, the dual-polarization radar rainfall data stored in the PostgreSQL database is processed into a rainfall space-time cube for a specified area. The detailed steps are as follows:

[0109] Step (1): Split the space-time cube. Split the time segments according to the time axis length and time scale of the space-time cube to obtain the time t0~t n The space-time cube of a region between t0 and t1 is divided into two parts according to the time scale Δt (which must be greater than or equal to the period T of radar data acquisition). n The data is evenly divided into n layers, each layer representing the spatial rainfall distribution at a certain moment. In this embodiment, T is the radar data acquisition period of 7.5 minutes.

[0110] Step (2): Calculate the rainfall distribution within the segment. i ~t (i+1)In the layer, for kT≤Δt≤(k+1)T, to fully utilize radar rainfall data, the time interval can be evenly divided into k segments. Surface rainfall point data can be filtered through spatial queries with additional time conditions. These point data are then used for spatial interpolation. Interpolation algorithm parameters are selected and executed according to the spatial scale and interpolation range requirements to obtain rainfall raster data within the minimum time segment. Finally, the rainfall distribution for time dimension index i is calculated through segmented summation.

[0111] Step (3): Merge multiple segments into a space-time cube. Repeat step (2) to calculate the rainfall distribution of n layers, and finally superimpose the rainfall space-time cube of the rainfall inversion area on the time axis. The schematic diagram of the rainfall space-time cube is as follows: Figure 9 This embodiment finally obtains 25 sets of radar rainfall spatiotemporal cube data with different spatiotemporal resolutions in the simulation area, all of which are stored as NetCDF files. Figure 10 These are the dual-polarization radar rainfall inversion results at different spatial scales during this period.

[0112] It should be noted that the above content merely illustrates the technical idea of ​​the present invention and cannot be used to limit the scope of protection of the present invention. For ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications all fall within the scope of protection of the claims of the present invention.

Claims

1. A multi-scale mapping method of dual-polarization radar grid and hydrological simulation unit, characterized in that: The steps include: Step S1: Data quality control, rainfall echo screening, rainfall intensity estimation, projection conversion and construction of rainfall space-time cube are performed on the dual-polarization radar raw data in the rainfall area to achieve multi-scale mapping between the dual-polarization radar grid and the hydrological simulation unit; Step S2: Perform data quality control on the dual polarization radar raw data, including the dual polarization radar differential propagation phase shift parameter Φ DP Denoising, differential propagation phase constant K DP Estimation and detection parameter Z H With Z DR Attenuation correction; Step S3: using the interval boundary values ​​of the polarization parameter membership functions of the dual-polarization radar as the threshold interval for screening rainfall echoes, thereby screening rainfall echoes in heavy rain weather; Step S4: According to Z H The rainfall conditions reflected by the parameters take into account the influence of raindrop deformation on quantitative rainfall estimation under different rainfall intensities, and are combined with Z H and K DP Calculate rainfall intensity; Step S5: Calculate the three-dimensional spatial coordinates of the rainfall echo based on the polar coordinate storage format of the dual-polarization radar raw data, calculate the coordinates of the rainfall particles landing on the ground based on the wind field and temperature conditions in the rainfall inversion area, and construct a two-dimensional grid based on the pixel size of the surface hydrological unit; Step S6: establishing a dual-polarization radar rainfall measurement database, and storing the radar data processing results that have completed rainfall estimation and projection conversion; Step S7: Divide the time segments according to the length of the simulated rainfall time axis and the time scale, calculate the rainfall distribution within the segments in combination with the dual-polarization radar rainfall data stored in the database, and merge multiple segments into a rainfall space-time cube.

2. The multi-scale mapping method of dual-polarization radar grid and hydrological simulation unit according to claim 1, characterized in that: The step S2 comprises the following steps: Step S21: Select the radial library as the smoothing window and use median filtering to smooth the differential propagation phase shift parameter Φ DP Perform denoising; Step S22: Based on the median filter Φ DP , use the least squares method to calculate the differential propagation phase constant K DP Make an estimate and calculate the formula: In the formula, r i represents the distance from the i-th radial bin to the radar antenna, To estimate the distance from the center of the window to the radar antenna, is the average value of the differential propagation phase shift of 2n+1 radial bins; Step S23: Based on the median filter Φ DP and the initial differential phase pair Z H and Z DR The parameters are linearly corrected, and the calculation formula is: In the formula, Z H and Z DR are the horizontal reflectivity factor and differential reflectivity factor obtained by dual polarization radar detection, Z H 'With Z DR ' is the horizontal reflectivity factor and differential reflectivity factor after attenuation correction; a1 and a2 are linear correction parameters. For C-band dual-polarization radar, a1 is 0.08dBZ / ° and a2 is 0.02dB / °; Φ DP is the differential propagation phase shift parameter of the echo position after median filtering, is the initial differential phase, and the value of the initial phase in each radar scanning radial direction is different.

3. The multi-scale mapping method of dual-polarization radar grid and hydrological simulation unit according to claim 1, characterized in that: The step S3 comprises the following steps: Step S31: determining the interval boundary values ​​of each polarization parameter membership function of the dual-polarization radar target object by fuzzy logic method; Step S32: Using the determined boundary value as a threshold interval for screening rainfall echoes, the rainfall echoes in heavy rain weather are screened.

4. The multi-scale mapping method of dual-polarization radar grid and hydrological simulation unit according to claim 1, characterized in that: The step S4 comprises the following steps: Step S41: According to Z H The magnitude of the parameter value determines the intensity of rainfall; Step S42: Use Z according to the rainfall intensity H and K DP Calculate the rainfall intensity using the following formula:

5. The multi-scale mapping method of dual-polarization radar grid and hydrological simulation unit according to claim 1, characterized in that: The step S5 comprises the following steps: Step S51: Calculate the spatial three-dimensional coordinates of the rainfall echo using the polar coordinates stored in the dual-polarization radar raw data. The calculation formula is: In the formula, X, Y, and Z represent the three-dimensional coordinates of the rainfall echo, X0, Y0, and h are the plane coordinates and elevation of the radar antenna position, and R is the slant range. is the elevation angle of the radar antenna, θ is the azimuth angle of the scanning radial direction, R 2 / (2·R m ′) is a correction term taking into account the curvature of the earth and the refraction of electromagnetic waves; Step S52: Combining the initial velocity of the rainfall particles, the echo position height, the wind field, and the temperature information, the two-dimensional coordinates of the rainfall particles falling on the surface hydrological unit are calculated through geographic coordinate projection conversion. The relationship between the rainfall falling on the surface and the rainfall detection results in the air is expressed as: p=g(X,Y,Z,temp,K dp ,Z H ,Z DR ) In the formula, geom and p are the coordinates of the rainfall falling to the ground and the rainfall intensity, the horizontal coordinate X, the vertical coordinate Y, the vertical coordinate Z, and the differential propagation phase constant K dp , horizontal reflectivity factor Z H , differential reflectivity factor Z DR ,speed is a parameter that can represent the position and state of rainfall particles in the air. and temp represent the wind field and temperature conditions in the rainfall retrieval area, respectively; Step S53: Reflecting hydrological units of different spatial scales according to different pixel sizes to obtain a multi-scale rainfall grid.

6. The multi-scale mapping method of dual-polarization radar grid and hydrological simulation unit according to claim 1, characterized in that: The step S6 comprises the following steps: Step S61: constructing a dual-polarization radar rainfall measurement data logic model and creating a dual-polarization radar rainfall measurement database; Step S62: Provide rainfall echo screening, rainfall intensity calculation, coordinate conversion and geographic projection, and multi-scale division interface, and store the dual-polarization radar data processing results that have completed rainfall estimation and projection conversion into data.

7. The multi-scale mapping method of dual-polarization radar grid and hydrological simulation unit according to claim 1, characterized in that: The step S7 includes the following steps: Step S71: Divide the time segments according to the inverted rainfall time axis length and time scale, and divide t0 to t n Divide into n layers evenly, each layer is the spatial rainfall distribution at a certain moment; Step S72: The time interval is further divided into k segments. Surface rainfall point data are filtered out through spatial query with additional time conditions. These point data are used for spatial interpolation. Interpolation algorithm parameters are selected and executed according to the spatial scale and interpolation range requirements to obtain rainfall raster data within the minimum time segment. Finally, the rainfall distribution with time dimension index i is calculated through segmented summation. Step S73: Repeat step S72 to calculate the rainfall distribution of n layers, and finally superimpose them on the time axis to form a rainfall space-time cube of the rainfall inversion area.

Images