A method for fusion of S, C and X-band radar precipitation networks

Abstract

The present invention discloses a precipitation networking fusion method for S-band, C-band and X-band radars, belonging to the technical field of weather radar observation. The method comprises the following steps: using a weather radar for real-time observation, calculating and removing the influence of terrain and building shielding on radar inversion precipitation, statistically obtaining dual-polarization precipitation inversion operators of S-band, C-band and X-band radars applicable to the area where the weather radar is located, generating a radar dual-polarization mixed precipitation inversion method, performing coordinate conversion from a radar polar coordinate system to a geodetic coordinate system, calculating and obtaining the geodetic coordinates of radar observation pixels, mosaicking precipitation information of multiple radars, reducing discontinuity of fused precipitation products, generating a smooth precipitation rate field, selecting multiple rainfall processes within the mosaicked area, and observing rainfall using a rain gauge. The present invention can realize the complementary advantages of different radars, provide a large-scale and locally refined precipitation monitoring method, and meet the needs of large-scale precipitation monitoring and refined precipitation monitoring.

Description

Technical Field

[0001] The present invention belongs to the technical field of weather radar observation, and in particular relates to a precipitation networking fusion method for S-, C- and X-band radars. Background Art

[0002] Over the past few decades, weather radar has been widely used in applications such as disastrous weather monitoring. Weather radar can obtain precipitation information with high temporal and spatial resolution, and has unique advantages in precipitation monitoring and early warning. Today, many countries use long-wave radars such as S and C bands as important observation instruments and equipment for weather monitoring and weather forecasting, and have carried out radar networking to form a national radar network monitoring. In the past 30 years, with the introduction of new technologies such as dual polarization, the monitoring capabilities of the radar network have been greatly improved.

[0003] However, due to the Earth's curvature and terrain obstruction, a major limitation of S- or C-band radar systems is their inability to capture low-level weather processes. For example, China's new generation of Doppler weather radar (CINRAD) radars, comprising S- and C-band radars, form the CINRAD radar monitoring network. In central and eastern China, the average spacing between radar stations is 200 km, while in western China, the spacing is even greater. Within a 200 km range, the electromagnetic beam at the lowest elevation angle (0.50°) is approximately 5.1 km above the ground. The lack of low-level observations of weather systems, the low spatial resolution at long radial distances, and the slow scanning speed of traditional mechanical radars severely hamper the ability to accurately detect and identify weather features. To address the shortcomings of S- and C-band operational weather radar monitoring and improve urban flooding forecasting and warning, many cities in my country have established X-band dual-polarization Doppler weather radar monitoring networks. The construction of an X-band dual-polarization Doppler weather radar monitoring network can provide more detailed precipitation observations than S-band and C-band operational weather radars, giving X-band dual-polarization radars great potential for monitoring and early warning of severe weather events such as rainstorms. The National Science Foundation's Collaborative Adaptive Sensing of the Atmosphere (CASA) Engineering Research Center has introduced an innovative, collaborative, and dynamic sensing model, called Distributed Collaborative Adaptive Observation (DCAS), to establish an X-band radar observation network to overcome the resolution and detection range limitations of traditional weather radar networks (McLaughlin et al., 2009). During the five years of CASA's X-band radar operation, high-temporal and spatial resolution observations, case studies, and multidisciplinary basic research have demonstrated the feasibility of the DCAS observation model. Since the spring of 2012, CASA has established the first urban demonstration X-band radar observation network in the United States in Dallas-Fort Worth (DFW).

[0004] However, the following problems have arisen in the current practical application of X-band radar: X-band radar has a small detection range and insufficient lead time for monitoring, forecasting and warning; S / C-band radar has a large detection range but low spatial resolution, which is insufficient to meet the needs of refined precipitation monitoring in cities. Summary of the Invention

[0005] Problem to be solved

[0006] To address the problems that existing X-band radars have a small detection range and insufficient advance time for monitoring, forecasting and early warning, while S / C-band radars have a large detection range but low spatial resolution, which is insufficient to meet the needs of refined precipitation monitoring in cities, the present invention provides a precipitation networking fusion method for S-, C- and X-band radars.

[0007] Technical Solution

[0008] To solve the above problems, the present invention adopts the following technical solutions.

[0009] A method for integrating S-, C-, and X-band radar precipitation networks, comprising the following steps:

[0010] Step 1: Install multiple weather radars and use them for real-time observations to perform quality control on precipitation echoes and reduce interference from non-precipitation echoes.

[0011] Step 2: Obtain the geographic location, terrain, and building information of each weather radar installation site, calculate the radar obstruction at different azimuth and elevation angles, calculate and remove the impact of terrain and building obstruction on radar precipitation retrieval, and supplement the signal lost due to obstruction;

[0012] Step 3: Using the raindrop spectrum observation data, obtain the dual-polarization precipitation inversion operators for S-band, C-band, and X-band radars applicable to the area where the weather radar is located;

[0013] Step 4: Design a mixed precipitation inversion method. According to the intensity of the dual polarization quantity, switch to use the precipitation inversion operator to generate the radar dual polarization mixed precipitation inversion method to obtain the precipitation value inverted by a single radar.

[0014] Step 5: Perform coordinate conversion from the radar polar coordinate system to the geodetic coordinate system, and calculate the geodetic coordinates of the radar observation pixels;

[0015] Step 6: Calculate the weight of each grid point radar mosaic based on the height and width of each radar detection beam, and mosaic the precipitation information of multiple radars;

[0016] Step 7: Reduce the discontinuity of the fused precipitation product and smooth the edges of the X-band radar coverage to generate a smoothed precipitation rate field;

[0017] Step 8: Select multiple rainfall processes within the puzzle area, use rain gauge rainfall observations, analyze and evaluate the data, and obtain the precipitation assessment results.

[0018] Preferably, in step 1, the quality control of the precipitation echo and the reduction of interference from the non-precipitation echo are performed by using radar dual-polarization observation, and according to the difference in precipitation echo correlation coefficient and horizontal and vertical continuity between the precipitation echo and the non-precipitation echo in radar observation, the precipitation echo and the non-precipitation echo are identified, and the non-precipitation echo is removed.

[0019] Furthermore, the specific process of the quality control is as follows: the correlation coefficient is judged, and when the correlation coefficient is greater than the preset value of the correlation coefficient, the texture value of the correlation coefficient is judged. If the texture value is greater than the preset value of the texture, further calculation and judgment are performed. When the correlation coefficient is less than or equal to the preset value of the correlation coefficient, it is determined whether the radar to which the correlation coefficient belongs is in a hail area, a non-uniform filling area, and a melting layer. If so, further calculation and judgment are performed.

[0020] Furthermore, the further calculation and judgment is to perform electromagnetic interference echo removal, and the formula is as follows:

[0021] N1>NUM*70%∩N2>NUM*10%

[0022] When the valid observation value of the lowest layer exceeds 70% of the entire radial area and the valid observation value of the second layer is less than 10% of the entire radial area, it is judged as an electromagnetic interference echo and is removed.

[0023] Furthermore, the difference between horizontal and vertical continuity is calculated by observing the missing signal in the M*N window. When the signal is missing for more than half, the observation is considered to be a non-precipitation echo and is removed. M is the number of radar radial range bins, and N is the number of radar beams. The vertical reflectivity check is performed by calculating the vertical gradient of the reflectivity. The gradient calculation formula is as follows:

[0024] oeLh k =(Z k -Z k+1 ) / (h k+1 -h k )

[0025] In the formula, h k is the beam center height at the kth elevation angle of a given range library. When the vertical gradient of reflectivity exceeds the preset value, the observation is considered to be a non-precipitation echo and is removed.

[0026] When the remaining echo area is smaller than the preset area value, it is considered that there is no precipitation at that moment and is removed. The holes in the precipitation echo whose area is smaller than the preset precipitation area are filled using the surrounding observations.

[0027] Preferably, in step 2, the calculation of the situation where the radar is blocked at different azimuth and elevation angles is performed by sending, receiving and calculating electromagnetic waves, and the formula is as follows:

[0028]

[0029] h c represents the height of the electromagnetic wave beam center relative to the radar antenna, r represents the electromagnetic wave propagation distance, R0 is the actual radius of the earth, R e represents the equivalent earth radius, t represents the radar observation elevation angle;

[0030] Assuming the electromagnetic wave beam width is θ, then replacing t in the formula with t+θ / 2 and t-θ / 2 can respectively obtain the beam top height ht and beam bottom height hb, and the vertical cross section d of the beam at any radial distance can be expressed as:

[0031]

[0032] According to the Gaussian distribution of radar electromagnetic wave energy on the vertical section, combined with the high-precision digital elevation model, the terrain shielding rate of electromagnetic wave energy can be calculated. The formula is as follows:

[0033]

[0034] Where P is the radar electromagnetic wave energy obstruction rate, x is the horizontal distance from the beam center on the vertical cross section of the beam, y(x) is the surface elevation relative to the bottom of the electromagnetic wave beam, and f(x,y) is the Gaussian distribution function with a standard deviation σ of d / 6.

[0035] The lowest elevation angle with an obstruction rate less than the preset angle is selected as the elevation angle for precipitation inversion at this azimuth, forming a radar composite plane scanning elevation angle. The influence of obstructions such as terrain and buildings on the radar inversion precipitation is removed, and the signal caused by obstruction in this area is supplemented.

[0036] Preferably, the calculation formulas for the dual-polarization precipitation inversion operators of the S-band, C-band, and X-band radars in step 3 are as follows:

[0037]

[0038] where a, b, and c are the dual-polarization precipitation inversion operators for S-band, C-band, and X-band radars, respectively.

[0039] Preferably, the geodetic coordinates of the radar observation pixel calculated in step 5 are calculated based on the latitude and longitude coordinates of the radar station and the observation parameters, and the formula is as follows:

[0040] Lat2=arcsin(sin(Lat1)cos(θ)+cos(Lat1)sin(θ)cos(α))

[0041]

[0042] θ=s / a

[0043]

[0044] Among them, Lat1, Lon1, h r are the latitude, longitude and antenna height of the radar station, respectively; Lat2 and Lon2 are the latitude and longitude of the radar observation pixel, respectively; s is the propagation distance of the radar electromagnetic wave propagation path projected onto the ground; a is the radius of the earth; R e is the equivalent earth radius, r is the radar electromagnetic radial propagation distance, and t is the radar observation elevation angle.

[0045] Preferably, in step 6, the weight of each radar puzzle is calculated for each grid point, and the precipitation information of multiple radars is puzzled by calculating the weight of each radar puzzle at each puzzle grid point according to the height and width of each radar detection beam. For puzzles of precipitation information of multiple radars, for a given grid, the weight w is determined by the center height h of the radar electromagnetic wave. c and the beam vertical cross-sectional diameter d b OK, the formula is as follows:

[0046]

[0047] Preferably, in step 7, the edge of the X-band radar coverage is smoothed to generate a smooth precipitation rate field by setting a transition zone of a certain width from the edge to the radar center, and adjusting the X-band radar weight for the grids in the transition zone using a coefficient q, as shown in the following formula:

[0048] w′=q·w

[0049]

[0050] Where w is the adjusted weight, R is the maximum effective detection range of the radar, r is the distance between the current grid center and the radar station, S buff is the width of the transition zone.

[0051] A method for integrating precipitation networks of S-, C-, and X-band radars is proposed. The method comprises the following steps: installing multiple weather radars, using the weather radars for real-time observation, performing quality control on precipitation echoes, reducing interference from non-precipitation echoes, obtaining the geographical location, terrain, and building information of each weather radar installation site, calculating the obstruction of the radar at different azimuth and elevation angles, calculating and removing the impact of terrain and building obstruction on radar inversion precipitation, supplementing the signal lost due to obstruction, using raindrop spectrum observation data, statistically obtaining dual-polarization precipitation inversion operators for S-, C-, and X-band radars applicable to the area where the weather radar is located, designing a mixed precipitation inversion method, switching between precipitation inversion operators based on the strength of the dual polarization quantities, generating a radar dual-polarization mixed precipitation inversion method, and obtaining a radar dual-polarization mixed precipitation inversion method. The precipitation value inverted by a single radar is obtained, and the coordinate conversion from the radar polar coordinate system to the geodetic coordinate system is performed. The geodetic coordinates of the radar observation pixels are calculated. The weight of each grid radar puzzle is calculated based on the height and width of each radar detection beam. The precipitation information of multiple radars is puzzled to reduce the discontinuity of the fused precipitation product. The edges of the X-band radar coverage area are smoothed to generate a smooth precipitation rate field. Multiple rainfall processes within the puzzle area are selected, and rainfall observations are made using rain gauges. The data are analyzed and evaluated. The precipitation assessment results are obtained by fusing S- and C-band radars with X-band radars to achieve complementary advantages of different radars, providing a large-scale and locally refined precipitation monitoring method to meet the needs of large-scale precipitation monitoring and refined precipitation monitoring.

[0052] Beneficial effects

[0053] Compared with the prior art, the present invention has the following beneficial effects:

[0054] (1) The present invention uses a radar composite plane scanning elevation angle formation method. Based on the radar's geographical location information, terrain, and building information, combined with the principle of electromagnetic wave propagation, the shielding situation is calculated to eliminate the problem of missing observation signals caused by radar shielding, form a radar composite plane scanning elevation angle, and remove the impact of radar shielding caused by terrain, buildings, etc. on radar inversion precipitation;

[0055] (2) The present invention adopts a radar quality control method based on dual polarization quantities to determine the correlation coefficient, judge whether it is a precipitation echo, perform horizontal and vertical consistency checks, and remove electromagnetic interference echoes. Non-precipitation echoes near the radar center are effectively removed, while precipitation echoes are retained, thereby improving data reliability.

[0056] (3) The radar dual-polarization hybrid precipitation inversion method of the present invention combines the advantages of various inversion methods to improve the precipitation inversion accuracy and stability, and automatically switches to different inversion methods according to the signal strength of the dual-polarization observation quantity, thereby improving the precipitation inversion accuracy and stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solutions in the embodiments or exemplary embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments or exemplary descriptions. Obviously, the drawings described below are only some embodiments of the present application and should not be regarded as limiting the scope. For ordinary technicians in this field, other drawings can be obtained according to the drawings without paying creative work.

[0058] Figure 1 It is a schematic diagram of the steps of the present invention;

[0059] Figure 2 It is a schematic diagram of the technical solution flow of the present invention;

[0060] Figure 3 Schematic diagram of the quality control method of the present invention;

[0061] Figure 4 This is a graph showing the change in radar reflectivity before and after the quality control of the present invention;

[0062] Figure 5 Radar reflectivity graphs before and after adjustment observed by the X-band radar of the present invention;

[0063] Figure 6 This is a flow chart of the single-station radar mixed precipitation inversion method of the present invention;

[0064] Figure 7 The spatial distribution diagram of the fused radar precipitation rate before and after edge smoothing of the present invention;

[0065] Figure 8 This is a statistical indicator diagram of different precipitation intensities of existing products and the product of the present invention. DETAILED DESCRIPTION

[0066] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Generally, the components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations.

[0067] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application for protection, but merely represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0068] Example 1

[0069] like Figure 1 As shown in FIG, a method for integrating radar precipitation networks of S, C, and X bands is implemented by the following steps:

[0070] Install multiple weather radars and use them for real-time observations to perform quality control on precipitation echoes and reduce interference from non-precipitation echoes. This quality control and reduction of interference from non-precipitation echoes are achieved by using radar dual-polarization observations. Based on the differences in precipitation echo correlation coefficients, horizontal and vertical continuity in radar observations, precipitation echoes and non-precipitation echoes are identified and removed.

[0071] The correlation coefficient is judged. When the correlation coefficient is greater than a preset value of the correlation coefficient, the texture value of the correlation coefficient is judged. If the texture value is greater than the preset value of the texture, further calculation and judgment are performed. When the correlation coefficient is less than or equal to the preset value of the correlation coefficient, it is judged whether the radar to which the correlation coefficient belongs is in a hail area, a non-uniform filling area, or a melt layer. If so, further calculation and judgment are performed.

[0072] Further calculation and judgment is to remove the electromagnetic interference echo. The formula is as follows:

[0073] N1>NUM*70%∩N2>NUM*10%

[0074] When the valid observation value of the lowest layer exceeds 70% of the entire radial area and the valid observation value of the second layer is less than 10% of the entire radial area, it is judged as an electromagnetic interference echo and is removed.

[0075] The difference between horizontal and vertical continuity is calculated by observing the missing signal in the M*N window. When the signal is missing for more than half, the observation is considered to be a non-precipitation echo and is removed. M is the number of radar radial range bins, and N is the number of radar beams. The vertical check of reflectivity is performed by calculating the vertical gradient of reflectivity. The gradient calculation formula is as follows:

[0076] oeLh k =(Z k -Z k+1 ) / (h k+1 -h k )

[0077] In the formula, h k is the beam center height at the kth elevation angle of a given range library. When the vertical gradient of reflectivity exceeds the preset value, the observation is considered to be a non-precipitation echo and is removed.

[0078] When the remaining echo area is smaller than the preset area value, it is considered that there is no precipitation at that moment and is removed. The holes in the precipitation echo whose area is smaller than the preset precipitation area are filled using the surrounding observations.

[0079] Obtain the geographical location, terrain, and building information of each weather radar installation site, calculate the radar's obstruction at different azimuths and elevations, calculate and remove the impact of terrain and building obstructions on radar precipitation inversion, and supplement the lost signal. Calculate the radar's obstruction at different azimuths and elevations using electromagnetic waves for transmission, reception, and calculation. The formula is as follows:

[0080]

[0081] h c represents the height of the electromagnetic wave beam center relative to the radar antenna, r represents the electromagnetic wave propagation distance, R0 is the actual radius of the earth, R e represents the equivalent earth radius, t represents the radar observation elevation angle;

[0082] Assuming the electromagnetic wave beam width is θ, then replacing t in the formula with t+θ / 2 and t-θ / 2 can respectively obtain the beam top height ht and beam bottom height hb, and the vertical cross section d of the beam at any radial distance can be expressed as:

[0083]

[0084] According to the Gaussian distribution of radar electromagnetic wave energy on the vertical section, combined with the high-precision digital elevation model, the terrain shielding rate of electromagnetic wave energy can be calculated. The formula is as follows:

[0085]

[0086] Where P is the radar electromagnetic wave energy obstruction rate, x is the horizontal distance from the beam center on the vertical cross section of the beam, y(x) is the surface elevation relative to the bottom of the electromagnetic wave beam, f(x,y) is the Gaussian distribution function, and the standard deviation σ is d / 6;

[0087] The lowest elevation angle with an obstruction rate less than the preset angle is selected as the elevation angle for precipitation inversion at this azimuth, forming a radar composite plane scanning elevation angle. The influence of obstructions such as terrain and buildings on the radar inversion precipitation is removed, and the signal caused by obstruction in this area is supplemented.

[0088] Using raindrop spectrum observation data, we statistically obtained the dual-polarization precipitation inversion operators for S-band, C-band, and X-band radars that are applicable to the area where the weather radar is located. The calculation formulas for the dual-polarization precipitation inversion operators for S-band, C-band, and X-band radars are as follows:

[0089]

[0090] where a, b, and c are the dual-polarization precipitation inversion operators for S-band, C-band, and X-band radars, respectively.

[0091] A mixed precipitation inversion method is designed. According to the intensity of the dual polarization quantity, the precipitation inversion operator is switched to generate a radar dual polarization mixed precipitation inversion method to obtain the precipitation value inverted by a single radar. The coordinate conversion from the radar polar coordinate system to the geodetic coordinate system is performed to calculate the geodetic coordinates of the radar observation pixels. The geodetic coordinates of the radar observation pixels are calculated based on the latitude and longitude coordinates of the radar station and the observation parameters. The formula is as follows:

[0092] Lat2=arcsin(sin(Lat1)cos(θ)+cos(Lat1)sin(θ)cos(α))

[0093]

[0094] θ=s / a

[0095]

[0096] Among them, Lat1, Lon1, h r are the latitude, longitude and antenna height of the radar station, respectively; Lat2 and Lon2 are the latitude and longitude of the radar observation pixel, respectively; s is the propagation distance of the radar electromagnetic wave propagation path projected onto the ground; a is the radius of the earth; R e is the equivalent earth radius, r is the radar electromagnetic radial propagation distance, and t is the radar observation elevation angle.

[0097] According to the height and width of each radar detection beam, the weight of each grid point radar puzzle is calculated, and the precipitation information of multiple radars is puzzled. The weight of each grid point radar puzzle is calculated according to the height and width of each radar detection beam. The precipitation information of multiple radars is puzzled. For a given grid, the weight w is determined by the center height h of the radar electromagnetic wave. c and the beam vertical cross-sectional diameter d b OK, the formula is as follows:

[0098]

[0099] To reduce the discontinuity of the fused precipitation product, the edges of the X-band radar coverage are smoothed to generate a smooth precipitation rate field. To smooth the edges of the X-band radar coverage to generate a smooth precipitation rate field, a transition zone of a certain width is set from the edge to the radar center. The X-band radar weight is adjusted for the grids in the transition zone using the coefficient q. The formula is as follows:

[0100] w′=q·w

[0101]

[0102] Where w is the adjusted weight, R is the maximum effective detection range of the radar, r is the distance between the current grid center and the radar station, S buff is the width of the transition zone.

[0103] Multiple rainfall processes within the puzzle area were selected, and rainfall observations were made using rain gauges. The data were analyzed and evaluated to obtain the precipitation assessment results.

[0104] From the above description, it can be seen that in this example, by installing multiple weather radars, using weather radars for real-time observation, quality control of precipitation echoes, reducing the interference of non-precipitation echoes, obtaining the geographical location, terrain and building information of each weather radar installation site, calculating the obstruction of radars at different azimuths and elevation angles, calculating and removing the impact of terrain and building obstructions on radar inversion precipitation, supplementing the signal lost by obstruction, using raindrop spectrum observation data, statistically obtaining the dual-polarization precipitation inversion operators of S, C and X band radars applicable to the area where the weather radar is located, designing a hybrid precipitation inversion method, and The intensity of the precipitation is determined by switching to the precipitation inversion operator, generating a radar dual-polarization mixed precipitation inversion method, obtaining precipitation values ​​inverted by a single radar, performing coordinate conversion from the radar polar coordinate system to the geodetic coordinate system, calculating the geodetic coordinates of the radar observation pixels, calculating the weight of each grid radar puzzle based on the height and width of each radar detection beam, piecing together the precipitation information of multiple radars, reducing the discontinuity of the fused precipitation product, smoothing the edges of the X-band radar coverage area, generating a smoothed precipitation rate field, selecting multiple rainfall processes within the puzzle area, and using rain gauge rainfall observations to analyze and evaluate the data to obtain the precipitation assessment results.

[0105] Example 2

[0106] like Figure 2-8 As shown, multiple weather radars are installed and used for real-time observation to perform quality control on precipitation echoes, reduce the interference of non-precipitation echoes, and perform quality control on precipitation echoes. The method flow is as follows Figure 3 To reduce the interference of non-precipitation echoes, radar dual-polarization observations are used. According to the differences in precipitation echo correlation coefficients, horizontal and vertical continuity between precipitation echoes and non-precipitation echoes in radar observations, precipitation echoes and non-precipitation echoes are identified and non-precipitation echoes are removed;

[0107] The correlation coefficient is determined. When the correlation coefficient is greater than 0.85, the texture value of the correlation coefficient is determined. If the texture value is greater than the texture preset value, further calculation and judgment are performed. When the correlation coefficient is less than or equal to 0.85, it is determined whether the radar to which the correlation coefficient belongs is in the hail area, non-uniform filling area and melt layer. If so, further calculation and judgment are performed.

[0108] Further calculation and judgment is to remove the electromagnetic interference echo. The formula is as follows:

[0109] N1>NUM*70%∩N2>NUM*10%

[0110] When the valid observation value of the lowest layer exceeds 70% of the entire radial area and the valid observation value of the second layer is less than 10% of the entire radial area, it is judged as an electromagnetic interference echo and is removed.

[0111] The difference between horizontal and vertical continuity is calculated by observing the missing signal in the M*N window. When the signal missing exceeds half, the observation is considered to be a non-precipitation echo and is removed. M is the number of radar radial range bins, which is set to 3, and N is the number of radar beams, which is set to 3. The vertical check of reflectivity is performed by calculating the vertical gradient of reflectivity. The gradient calculation formula is as follows:

[0112] oeLh k =(Z k -Z k+1 ) / (h k+1 -h j )

[0113] In the formula, h k is the beam center height at the kth elevation angle of the given range library. When the vertical gradient of reflectivity exceeds 50 dBZ / km, the observation is considered to be a non-precipitation echo and is removed;

[0114] When the remaining echo area is smaller than the preset area value, it is considered that there is no precipitation at that moment and is removed. The holes in the precipitation echo whose area is smaller than the preset precipitation area are filled using the surrounding observations.

[0115] After the above steps, non-precipitation echoes are removed and precipitation echoes are retained to complete quality control. Figure 4 Comparing the reflectivity maps before and after quality control, it can be seen that the non-precipitation echo near the radar center has been effectively removed, while the precipitation echo has been retained.

[0116] Obtain the geographical location, terrain, and building information of each weather radar installation site, calculate the radar's obstruction at different azimuths and elevations, calculate and remove the impact of terrain and building obstructions on radar precipitation inversion, and supplement the lost signal. Calculate the radar's obstruction at different azimuths and elevations using electromagnetic waves for transmission, reception, and calculation. The formula is as follows:

[0117]

[0118] h c represents the height of the electromagnetic wave beam center relative to the radar antenna, r represents the electromagnetic wave propagation distance, R0 is the actual radius of the earth, R erepresents the equivalent earth radius, t represents the radar observation elevation angle;

[0119] Assuming the electromagnetic wave beam width is θ, then replacing t in the formula with t+θ / 2 and t-θ / 2 can respectively obtain the beam top height ht and beam bottom height hb, and the vertical cross section d of the beam at any radial distance can be expressed as:

[0120]

[0121] According to the Gaussian distribution of radar electromagnetic wave energy on the vertical section, combined with the high-precision digital elevation model, the terrain shielding rate of electromagnetic wave energy can be calculated. The formula is as follows:

[0122]

[0123] Where P is the radar electromagnetic wave energy obstruction rate, x is the horizontal distance from the beam center on the vertical cross section of the beam, y(x) is the surface elevation relative to the bottom of the electromagnetic wave beam, f(x,y) is the Gaussian distribution function, and the standard deviation σ is d / 6;

[0124] The lowest elevation angle with an obstruction rate of less than 50% is selected as the elevation angle for precipitation inversion in this azimuth, forming a radar composite plane scanning elevation angle. The influence of obstructions such as terrain and buildings on the radar inversion precipitation is removed, and the signal caused by obstruction in this area is supplemented.

[0125] Figure 5 This is the 0.9° elevation angle and composite plane scanning elevation reflectivity display of the Shenzhen Xichong radar. It can be seen that due to terrain obstruction, the observation signal to the north of the Xichong radar is seriously missing. Using the composite plane scanning elevation angle, the signal caused by the obstruction in this area is supplemented.

[0126] Using raindrop spectrum observation data, we statistically obtained the dual-polarization precipitation inversion operators for S-band, C-band, and X-band radars that are applicable to the area where the weather radar is located. The calculation formulas for the dual-polarization precipitation inversion operators for S-band, C-band, and X-band radars are as follows:

[0127]

[0128] where a, b, and c are the dual-polarization precipitation inversion operators for S-band, C-band, and X-band radars, respectively.

[0129] Design a hybrid precipitation retrieval method, the process is as follows Figure 6According to the intensity of the dual polarization, the precipitation inversion operator is switched to generate the radar dual polarization mixed precipitation inversion method to obtain the precipitation value inverted by a single radar; the coordinate conversion from the radar polar coordinate system to the geodetic coordinate system is performed to calculate the geodetic coordinates of the radar observation pixel. The geodetic coordinates of the radar observation pixel are calculated based on the latitude and longitude coordinates of the radar station and the observation parameters. The formula is as follows:

[0130] Lat2=arcsin(sin(Lat1)cos(θ)+cos(Lat1)sin(θ)cos(α))

[0131]

[0132] θ=s / a

[0133]

[0134] Among them, Lat1, Lon1, h r are the latitude, longitude and antenna height of the radar station, respectively; Lat2 and Lon2 are the latitude and longitude of the radar observation pixel, respectively; s is the propagation distance of the radar electromagnetic wave propagation path projected onto the ground; a is the radius of the earth; R e is the equivalent earth radius, r is the radar electromagnetic radial propagation distance, and t is the radar observation elevation angle.

[0135] According to the height and width of each radar detection beam, the weight of each grid point radar puzzle is calculated, and the precipitation information of multiple radars is puzzled. The weight of each grid point radar puzzle is calculated according to the height and width of each radar detection beam. The precipitation information of multiple radars is puzzled. For a given grid, the weight w is determined by the center height h of the radar electromagnetic wave. c and the beam vertical cross-sectional diameter d b OK, the formula is as follows:

[0136]

[0137] The fusion product has a "jump" from the X-band radar precipitation rate to the S- and C-band radar precipitation rates at the edge of the X-band radar coverage. To reduce the discontinuity of the fused precipitation product, the edge of the X-band radar coverage is smoothed to generate a smooth precipitation rate field. To smooth the edge of the X-band radar coverage and generate a smooth precipitation rate field, a transition zone of a certain width is set from the edge to the radar center. The X-band radar weight is adjusted for the grids in the transition zone using the coefficient q. The formula is as follows:

[0138] w′=q·w

[0139]

[0140] Where w is the adjusted weight, R is the maximum effective detection range of the radar, r is the distance between the current grid center and the radar station, S buff is the width of the transition zone.

[0141] Figure 6 The precipitation rate field before and after smoothing. It can be seen that after smoothing, the discontinuity of the precipitation rate field at the junction of different radars has been significantly suppressed, forming a relatively smooth precipitation rate field.

[0142] Multiple rainfall processes within the puzzle area were selected, and rainfall observations were made using rain gauges. The data were analyzed and evaluated to obtain the precipitation assessment results.

[0143] Example 3

[0144] Figure 8 The statistical parameter indicators of 11 precipitation processes of the business product and the product of the present invention at various precipitation intensities were counted. It can be seen that for the business product, as the precipitation intensity increases, the consistency between the product and the automatic station (i.e. CC) gradually decreases, the RMSE gradually increases, and the degree of dispersion gradually increases. The systematic deviation of the product (i.e. RMB) turns from positive to negative, indicating that the product overestimates weak precipitation and underestimates strong precipitation. The relative mean absolute error (RMAE) is large in the >=0.1mm level, exceeding 1.0, and the performance is poor. This is related to the serious overestimation of the product in weak precipitation. As the precipitation intensity increases, the RMAE gradually decreases. For the product of the present invention, in terms of overall indicators, except for the RMB in the >=5 and >=10mm levels, which is inferior to the business product, all other indicators are better than the business product. As the precipitation intensity increases, the consistency between the product of the present invention and the automatic station also gradually decreases, but the indicator value is relatively high. For short-term strong precipitation >20mm, the CC still reaches 0.74, which is better than the business product. RMSE increases with precipitation intensity, from 3.2 mm / h to 11.54 mm / h, but remains lower than the commercial product at the same level. RMB and RMAE vary slightly with precipitation intensity, with RMAE stabilizing around 0.3 and RMB around -0.2, indicating relatively stable performance across different precipitation intensities.

[0145] In summary, the product formed by the present invention can provide accurate precipitation products with high temporal and spatial resolution, and its accuracy and stability are better than those of commercial products.

[0146] The above-described embodiments merely represent preferred embodiments of the present invention. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art would be able to make numerous modifications, improvements, and substitutions without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention.

Claims

1. A method for integrating S-, C- and X-band radar precipitation networks, characterized in that: Use the following steps: Step 1: Install multiple weather radars and use them for real-time observations to perform quality control on precipitation echoes and reduce interference from non-precipitation echoes. Step 2: Obtain the geographic location, terrain, and building information of each weather radar installation site, calculate the radar obstruction at different azimuth and elevation angles, calculate and remove the impact of terrain and building obstruction on radar precipitation retrieval, and supplement the signal lost due to obstruction; Step 3: Using the raindrop spectrum observation data, obtain the dual-polarization precipitation inversion operators for S-band, C-band, and X-band radars applicable to the area where the weather radar is located; Step 4: Design a mixed precipitation inversion method. According to the intensity of the dual polarization quantity, switch to use the precipitation inversion operator to generate the radar dual polarization mixed precipitation inversion method to obtain the precipitation value inverted by a single radar. Step 5: Perform coordinate conversion from the radar polar coordinate system to the geodetic coordinate system, and calculate the geodetic coordinates of the radar observation pixels; Step 6: Calculate the weight of each grid point radar mosaic based on the height and width of each radar detection beam, and mosaic the precipitation information of multiple radars; Step 7: Reduce the discontinuity of the fused precipitation product and smooth the edges of the X-band radar coverage to generate a smoothed precipitation rate field; Step 8: Select multiple rainfall processes within the puzzle area, use rain gauge rainfall observations, analyze and evaluate the data, and obtain the precipitation assessment results.

2. The S-, C- and X-band radar precipitation network fusion method according to claim 1, characterized in that: In step 1, the quality control of precipitation echoes and the reduction of interference from non-precipitation echoes are performed by using radar dual-polarization observations, and according to the differences in precipitation echo correlation coefficients and horizontal and vertical continuity between precipitation echoes and non-precipitation echoes in radar observations, precipitation echoes and non-precipitation echoes are identified, and non-precipitation echoes are removed.

3. The S-, C- and X-band radar precipitation network fusion method according to claim 2, characterized in that: The specific process of the quality control is as follows: the correlation coefficient is judged. When the correlation coefficient is greater than the preset value of the correlation coefficient, the texture value of the correlation coefficient is judged. If the texture value is greater than the preset value of the texture, further calculation and judgment are performed. When the correlation coefficient is less than or equal to the preset value of the correlation coefficient, it is judged whether the radar to which the correlation coefficient belongs is in a hail area, a non-uniform filling area and a melt layer. If so, further calculation and judgment are performed.

4. The S-, C- and X-band radar precipitation network fusion method according to claim 3, characterized in that: The further calculation and judgment is to remove the electromagnetic interference echo, and the formula is as follows: N1>NUM*70%∩N2>NUM*10% When the valid observation value of the lowest layer exceeds 70% of the entire radial area and the valid observation value of the second layer is less than 10% of the entire radial area, it is judged as an electromagnetic interference echo and is removed.

5. The S-, C- and X-band radar precipitation network fusion method according to claim 4, characterized in that: The difference between horizontal and vertical continuity is calculated by observing the missing signal in the M*N window. When the signal is missing for more than half, the observation is considered to be a non-precipitation echo and is removed. M is the number of radar radial range bins, and N is the number of radar beams. The vertical check of reflectivity is performed by calculating the vertical gradient of reflectivity. The gradient calculation formula is as follows: vgdBZ k =(Z k -Z k+1 ) / (h k+1 -h k ) In the formula, h k is the beam center height at the kth elevation angle of a given range library. When the vertical gradient of reflectivity exceeds the preset value, the observation is considered to be a non-precipitation echo and is removed. When the remaining echo area is smaller than the preset area value, it is considered that there is no precipitation at that moment and is removed. The holes in the precipitation echo whose area is smaller than the preset precipitation area are filled using the surrounding observations.

6. The S-, C- and X-band radar precipitation network fusion method according to claim 1, characterized in that: In step 2, the calculation of the situation where the radar is blocked at different azimuth and elevation angles is performed by using electromagnetic waves for sending, receiving, and calculation. The formula is as follows: h c represents the height of the electromagnetic wave beam center relative to the radar antenna, r represents the electromagnetic wave propagation distance, R0 is the actual radius of the earth, R e represents the equivalent earth radius, t represents the radar observation elevation angle; Assuming the electromagnetic wave beam width is θ, then replacing t in the formula with t+θ / 2 and t-θ / 2 can respectively obtain the beam top height ht and beam bottom height hb, and the vertical cross section d of the beam at any radial distance can be expressed as: According to the Gaussian distribution of radar electromagnetic wave energy on the vertical section, combined with the high-precision digital elevation model, the terrain shielding rate of electromagnetic wave energy can be calculated. The formula is as follows: Where P is the radar electromagnetic wave energy obstruction rate, x is the horizontal distance from the beam center on the vertical cross section of the beam, y(x) is the surface elevation relative to the bottom of the electromagnetic wave beam, f(x,y) is the Gaussian distribution function, and the standard deviation σ is d / 6; The lowest elevation angle with an obstruction rate less than the preset angle is selected as the elevation angle for precipitation inversion at this azimuth, forming a radar composite plane scanning elevation angle. The influence of terrain and building obstruction on radar inversion precipitation is removed, and the signal caused by obstruction in this area is supplemented.

7. The S-, C- and X-band radar precipitation network fusion method according to claim 1, characterized in that: The calculation formulas for the dual-polarization precipitation inversion operators of the S-band, C-band, and X-band radars in step 3 are as follows: where a, b, and c are the dual-polarization precipitation inversion operators for S-band, C-band, and X-band radars, respectively.

8. The S-, C- and X-band radar precipitation network fusion method according to claim 1, characterized in that: The geodetic coordinates of the radar observation pixel calculated in step 5 are calculated based on the latitude and longitude coordinates of the radar station and the observation parameters. The formula is as follows: Lat2=arcsin(sin(Lat1)cos(θ)+cos(Lat1)sin(θ)cos(α)) θ=s / a Among them, Lat1, Lon1, h r are the latitude, longitude and antenna height of the radar station, respectively; Lat2 and Lon2 are the latitude and longitude of the radar observation pixel, respectively; s is the propagation distance of the radar electromagnetic wave propagation path projected onto the ground; a is the radius of the earth; R e is the equivalent earth radius, r is the radar electromagnetic radial propagation distance, and t is the radar observation elevation angle.

9. The S-, C- and X-band radar precipitation network fusion method according to claim 1, characterized in that: In step 6, the weight of each radar puzzle is calculated for each grid point. The precipitation information of multiple radars is puzzled based on the height and width of each radar detection beam. The weight of each radar puzzle at each puzzle grid point is calculated. The precipitation information of multiple radars is puzzled. For a given grid, the weight w is determined by the center height h of the radar electromagnetic wave. c and the beam vertical cross-sectional diameter d b OK, the formula is as follows:

10. The S-, C- and X-band radar precipitation network fusion method according to claim 1, characterized in that: In step 7, the edge of the X-band radar coverage is smoothed to generate a smooth precipitation rate field by setting a transition zone of a certain width from the edge to the radar center, and adjusting the X-band radar weight for the grid in the transition zone using the coefficient q. The formula is as follows: w'=q·w Where w is the adjusted weight, R is the maximum effective detection range of the radar, r is the distance between the current grid center and the radar station, S buff is the width of the transition zone.

Images