Radar precipitation measurement method and device based on 45-degree slant de-polarization ratio

Abstract

The application provides a radar precipitation measurement method and device based on a 45-degree slant depolarization ratio, which uses a radar with or modified to realize 45-degree slant polarization to collect single-shot double-receiving two-channel complex voltage data, calculates a slant baseline linear depolarization ratio and a reflectivity factor Z; identifies a melting layer according to a vertical pointing profile observation collected by the radar in a vertical pointing mode, and applies upper and lower boundaries of the melting layer to a near-horizontal scanning mode to realize liquid rainfall identification; performs rain attenuation correction on and Z, and according to the rain attenuation corrected and Z, looks up a table or a neural network to inverse precipitation parameters such as rainfall intensity, mass-weighted mean diameter, normalized intercept, number concentration and water content. The application improves the precision and robustness of precipitation measurement.

Description

Technical Field

[0001] This invention relates to the field of meteorological forecasting technology, and in particular to a radar precipitation measurement method and device based on a 45-degree depolarization ratio. Background Technology

[0002] Quantitative precipitation estimation (QPE) using radar is a key technology in weather forecasting, urban flood control, and watershed hydrological early warning. Traditional single-polarization weather radars can only obtain reflectivity factors. The difficulty in accurately obtaining microphysical information such as precipitation magnitude and morphology makes precise precipitation observation challenging. To improve the ability to observe cloud precipitation microphysics and estimate quantitative precipitation, dual-polarization radar technology has been widely developed. For example, weather radar measures echoes and phases using a dual-transmitter, dual-receiver system on two orthogonally polarized channels (horizontal and vertical), enabling the acquisition of reflectivity factors. It can also obtain differential reflectance that is sensitive to particle phase, particle size, size / shape, and water content. Differential phase (Comparison of differential propagation phase) ), correlation coefficient A series of polarization variables are used to improve the accuracy of quantitative inversion of precipitation parameters.

[0003] Furthermore, dual-polarization technology is also used in cloud radar. Existing cloud radars typically point the antenna beam nearly vertically, employing a single-transmitter, dual-receiver system (such as the MIRA-35C). Two mutually orthogonal linear polarization channels are set in a plane perpendicular to the propagation direction (conventionally still denoted as "horizontal" H and "vertical" V, but in this case, they are both on the horizontal plane). The receiver can simultaneously output the local polarization power, cross-polarization power, and phase information of both channels, thereby constructing the Doppler spectrum and reflectivity factor on the vertical profile. and linear depolarization ratio When the direction is vertical, since the average major axis of the raindrop is horizontal and the propagation direction is vertical, the scattering characteristics of the two orthogonal linear polarizations on the horizontal plane are approximately equivalent in liquid raindrops, H / V– In areas of pure rainfall, the intensity is typically very low, only rising significantly in the melting layer (bright band) and solid precipitation areas. Therefore, existing cloud radars mostly utilize vertical pointing... Identifying the height and thickness of the melting layer for cloud physics research or to assist in bright band identification in weather radar is difficult to provide stable constraints for precipitation parameters (especially particle size). It is worth noting that, also because the scattering power and phase of the two orthogonal linear polarizations of hydrocondensate particles are approximately equal when pointing vertically downwards, cloud radar typically does not employ a dual-transmitter, dual-receiver system. The true value is close to 0 dB. The true value is close to 0 deg / km.

[0004] In recent years, to further expand the application areas of cloud radar, the industry has begun to explore using single-transmitter dual-receiver cloud radar for horizontal PPI scanning or RHI scanning. Initially, this type of radar primarily borrowed from conventional weather radar, employing horizontal and vertical polarization channels. However, unlike conventional weather radar, the polarization parameters of single-transmitter dual-receiver radar are... While it can be used to indicate melting particles or tilted ice crystal particles, for rainfall, It can only represent the oscillations of larger (flatter) raindrops (with a flatter shape), but its ability to quantitatively describe rainfall information is insufficient.

[0005] To improve sensitivity to particle shape, some studies have proposed measuring the linear depolarization ratio (SLDR) of the slant base under a 45-degree linear polarization substrate. For example, a study on the 35 GHz MIRA-35 cloud radar proposed that by operating the radar in SLDR mode and utilizing the sensitivity of the SLDR measured by scanning to particle geometry, the vertical distribution of particle shapes in mixed-phase clouds can be inverted. Researchers have also attempted to use observations from scanning polarized W-band radar in SLDR mode to identify different ice crystal behaviors (such as dendritic, columnar, graupel, and hail), demonstrating the significant application potential of the SLDR model in cloud / ice microphysical inversion. However, these works primarily focus on cloud and ice particle shape and microphysical processes in mixed-phase clouds, without aiming at quantitative precipitation estimation, and do not provide systematic inversion methods for microphysical parameters (especially particle size and concentration) and intensity of precipitation.

[0006] In light of the above background, existing technologies for quantitative precipitation measurement based on cloud radar have the following shortcomings:

[0007] 1. Vertical pointing cloud radar missions are primarily focused on "cloud measurement," lacking a dedicated measurement system for precipitation. Most existing cloud radars operate in vertical pointing mode. Primarily used for cloud and melting layer identification, it lacks a quantitative precipitation estimation method designed for near-horizontal precipitation fields. The reflectivity factor Z suffers significant rain attenuation in the short-wavelength band, and without differential phase or other self-consistent correction methods, long-distance estimation exhibits large biases.

[0008] 2. H / V mode In liquid precipitation, its measurability is poor, making it difficult to use as a particle shape constraint. When scattering perpendicularly, the differences in scattering characteristics in orthogonal directions are very small, leading to... Extremely low in pure rain. In near-horizontal scan horizontal + vertical planning mode. It is also mainly used to distinguish between solid and liquid phases, rather than as a core variable for precipitation parameter inversion.

[0009] 3. Already available Cloud radar research has not yet established a complete inversion technology path for precipitation. Existing SLDR model cloud radar research mainly focuses on particle shape or ice cloud properties, using... As a shape-sensitive quantity, it does not construct a quantitative inversion method for precipitation parameters, does not consider rain attenuation correction strategies, and does not organically combine the detection of the melting layer during vertical scanning with near-horizontal SLDR rainfall measurement. Summary of the Invention

[0010] This invention provides a radar precipitation measurement method and apparatus based on a 45-degree depolarization ratio, which can solve the following technical problems:

[0011] 1. How to construct a 45-degree oblique polarization operating mode adapted to the hardware structure of the modified cloud radar, so that information related to raindrop shape can be transmitted through... Entering the measurable dynamic range;

[0012] 2. How to utilize the two working modes, with or without phase, and By using lookup tables or multilayer perceptron (MLP) neural networks, the rainfall intensity R and mass-weighted average diameter D can be stably retrieved. m Normalized intercept N w Number concentration N t Precipitation parameters such as water content (LWC);

[0013] 3. How to use the profile in the vertical pointing mode of cloud radar to identify the height and thickness of the melting layer, and provide bright band correction and quality control for SLDR rain measurement in near-horizontal mode;

[0014] 4. In short-wavelength environments with strong attenuation, how can phase be utilized? Rain attenuation correction is performed when phase is not available. – Radial consistency enables phase-free self-consistent correction.

[0015] This invention provides a radar precipitation measurement method based on a 45-degree depolarization ratio, comprising:

[0016] Based on the two-channel complex voltage data acquired by the radar, the linear depolarization ratio of the slant baseline is calculated. And reflectivity factor Z, the radar needs to adopt a 45-degree slant linear polarization single-transmission dual-receive mode;

[0017] Based on the vertical pointing profile acquired by the radar in vertical pointing mode, the upper and lower boundaries of the melting layer are identified. After the lower boundary of the melting layer is mapped to the radar horizontal scanning mode for observation, liquid precipitation is identified and mixed phase or ice phase precipitation is shielded.

[0018] Regarding the linear depolarization ratio of the sloping baseline Rain attenuation correction was performed using the reflectivity factor Z;

[0019] Based on the linear depolarization ratio of the slope after rain attenuation correction The precipitation parameters are obtained by inversion with the reflectivity factor Z, including precipitation intensity, mass-weighted average diameter, normalized intercept, number concentration, and water content.

[0020] According to the present invention, a radar precipitation measurement method based on the 45-degree depolarization ratio is provided, which calculates the linear depolarization ratio of the slant line based on the two-channel complex voltage data acquired by the radar. Before the reflectivity factor Z, it also includes:

[0021] When the radar is unable to scan horizontally, an elevation and azimuth mechanism is added to it, or the scanning function is enabled to achieve near-horizontal scanning;

[0022] When the radar polarization mode is horizontal / vertical polarization, its transmission linear polarization direction is rotated by 45 degrees in a plane perpendicular to the beam propagation direction to achieve oblique 45-degree linear polarization transmission.

[0023] The radar maintains a dual-channel coherent receiving structure and reserves a vertical pointing mode for melting layer identification.

[0024] According to the present invention, a radar precipitation measurement method based on the 45-degree depolarization ratio is provided, which calculates the linear depolarization ratio of the slant line based on the two-channel complex voltage data acquired by the radar. The reflectivity factor Z includes:

[0025] Decouple the two-channel complex voltage data acquired by the radar;

[0026] Calculate the common polarization power and cross polarization power between the two channels of complex voltage data after decoupling;

[0027] The linear depolarization ratio of the sloping baseline is calculated based on the ratio of the common polarization power to the cross polarization power after noise depolarization in the power domain. ;

[0028] After noise reduction of the common polarization power, the reflectivity factor Z is obtained by converting it using radar constant and observation distance;

[0029] Noise removal and filtering are performed on the linear depolarization ratio (SLDR) of the sloping baseline in regions where the signal-to-noise ratio (SNR) is less than a preset threshold.

[0030] According to the present invention, a radar precipitation measurement method based on a 45-degree depolarization ratio is provided, which identifies the melting layer based on the vertical pointing profile acquired by the radar in vertical pointing mode, comprising:

[0031] In each vertically pointing profile, the reflectivity factor Z profile and the linear depolarization ratio (SLDR) profile with 45-degree oblique angle are calculated as height.

[0032] The center of the zero-degree layer bright band is estimated based on the local peak value in the reflectivity factor Z profile.

[0033] The upper and lower boundaries of the melting layer are determined based on the height range corresponding to the rise of the SLDR profile at a 45-degree angle.

[0034] Based on the height and upper and lower boundaries of the melt layer, the data is mapped to radar horizontal scanning mode observations to identify liquid precipitation and shield mixed-phase or ice-phase precipitation. Subsequently, quantitative inversion is only carried out on liquid precipitation areas.

[0035] According to the present invention, a radar precipitation measurement method based on a 45-degree depolarization ratio is provided, wherein the linear depolarization ratio of the slant line is... Rain attenuation correction is performed using the reflectivity factor Z, including:

[0036] For devices equipped with phase measurement capabilities, the differential propagation phase shift is calculated after recombination from a 45-degree oblique substrate back to an H / V substrate via substrate transformation. and differential propagation phase shift rate ;

[0037] Using the differential propagation phase shift rate Based on the empirical relationship or self-consistent algorithm of the ratio attenuation, the linear depolarization ratio of the sloping baseline is... Rain attenuation correction was performed using the reflectivity factor Z;

[0038] For devices lacking phase measurement capabilities, a method based on the reflectivity factor Z and the linear depolarization ratio of the slant baseline is used. The self-consistent rain attenuation correction method with optimal radial consistency is applied to the linear depolarization ratio of the sloping baseline. Rain attenuation correction is performed using the reflectivity factor Z.

[0039] According to the present invention, a radar precipitation measurement method based on a 45-degree depolarization ratio is provided, which is based on the linear depolarization ratio of the slant line after rain attenuation correction. Inversion with the reflectivity factor Z yields precipitation parameters, including:

[0040] Based on LUT or MLP neural network, the linear depolarization ratio of the sloping baseline after rain attenuation correction is calculated. The precipitation parameters are obtained by inverting the reflectance factor Z.

[0041] The present invention provides a radar precipitation measurement method based on the 45-degree depolarization ratio, which uses the linear depolarization ratio of the slant line corrected for rain attenuation by a LUT. Inversion with the reflectivity factor Z yields precipitation parameters, including:

[0042] Based on the T-matrix scattering model, a slant-based linear depolarization ratio was constructed. A lookup table showing the correlation between reflectance factor Z and precipitation parameters;

[0043] The observed linear depolarization ratio of the sloping baseline After multidimensional interpolation of the reflectivity factor Z, the corresponding precipitation parameters are obtained through the lookup table.

[0044] The present invention provides a radar precipitation measurement method based on the 45-degree depolarization ratio, which uses an MLP neural network to correct the linear depolarization ratio of the slant line after rain attenuation. Inversion with the reflectivity factor Z yields precipitation parameters, including:

[0045] Rain attenuation corrected slope baseline linear depolarization ratio The reflectivity factor Z, radar elevation angle, ambient temperature, signal-to-noise ratio (SNR), and radar height above the ground are used as inputs to the MLP neural network to obtain the precipitation parameters output by the MLP neural network.

[0046] The present invention also provides a radar precipitation measurement device based on a 45-degree depolarization ratio, comprising:

[0047] The calculation module is used to calculate the linear depolarization ratio of the slant baseline based on the two-channel complex voltage data acquired by the radar. The reflectivity factor Z, the radar transmits at a 45-degree oblique linear polarization.

[0048] The identification module is used to identify the upper and lower boundaries of the melting layer based on the vertical pointing profile acquired by the radar in vertical pointing mode. After mapping the lower boundary of the melting layer to the radar horizontal scanning mode for observation, it identifies liquid precipitation and shields mixed phase or ice phase precipitation.

[0049] The correction module is used to adjust the linear depolarization ratio of the slant line. Rain attenuation correction was performed using the reflectivity factor Z;

[0050] The inversion module is used to determine the linear depolarization ratio of the sloping baseline after rain attenuation correction. The precipitation parameters are obtained by inversion with the reflectivity factor Z, including precipitation intensity, mass-weighted average diameter, normalized intercept, number concentration, and water content.

[0051] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the radar precipitation measurement method based on the 45-degree depolarization ratio described above.

[0052] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the radar precipitation measurement method based on the 45-degree depolarization ratio as described above.

[0053] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the radar precipitation measurement method based on the 45-degree depolarization ratio described above.

[0054] The radar precipitation measurement method and device based on a 45-degree depolarization ratio provided by this invention have the following advantages:

[0055] 1. This invention transforms depolarization information, which is difficult to quantify in the H / V model, into a core variable that can be used for precipitation particle size inversion. Using a 45-degree slant polarization regime, this invention utilizes the linear depolarization ratio of the slant baseline. High sensitivity to scattering anisotropy brings information related to raindrop shape and size into the measurable dynamic range; and Combined use to achieve precipitation parameters (precipitation intensity) Mass-weighted average diameter D m Normalized intercept N w Number concentration N t Quantitative inversion of moisture content (LWC, etc.).

[0056] 2. A complete precipitation inversion and rainfall attenuation correction framework specifically designed for cloud radar slant-based depolarization ratio was constructed. Compared with existing... Cloud radar is mainly used for cloud / ice particle shape inversion. This invention designs two quantitative inversion paths, LUT and MLP, for precipitation particle size, concentration and rainfall intensity, and combines two working scenarios with / without phase to form a quantitative inversion technology chain for precipitation parameters that takes into account both accuracy and feasibility.

[0057] 3. Fully utilize the sensitivity of the linear depolarization ratio to the melting layer in the vertical scanning mode to achieve an integrated design for vertical cloud measurement and near-horizontal rainfall measurement. This invention not only uses the high sensitivity of the linear depolarization ratio to the melting layer in the vertical scanning mode to identify the melting layer, but also feeds back its information for near-horizontal liquid rainfall identification and SLDR rainfall measurement quality control, enabling the same cloud radar to work collaboratively in both vertical and near-horizontal modes, thus improving the overall system's adaptability to complex phase environments.

[0058] 4. It exhibits better long-distance quantitative precipitation monitoring capabilities under short-wavelength heavy rainfall attenuation conditions. This is due to... or – With consistency correction, this invention can still recover the mid-to-long-range reflectivity factor and [other properties] relatively well in the strongly attenuated frequency bands such as X / Ka / W. To achieve more accurate precipitation parameters (precipitation intensity) Mass-weighted average diameter D m Normalized intercept N w Number concentration N t Estimated using water content (LWC, etc.). Compared to relying solely on uncorrected estimates. This significantly reduces the underestimation of long-distance rainfall compared to traditional methods.

[0059] 5. Low modification cost and easy to promote in existing cloud radar networks. This invention only requires adding near-horizontal scanning capability to the cloud radar and achieving 45-degree oblique polarization through mechanical rolling, while keeping the receiving link basically unchanged. The hardware modification is small, making it easy to deploy quickly in existing cloud radar networks and expand its application range from vertical cloud measurement to near-horizontal rainfall measurement. Attached Figure Description

[0060] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0061] Figure 1 This is a schematic flowchart of the radar precipitation measurement method based on the 45-degree depolarization ratio provided by the present invention.

[0062] Figure 2 This is an example of the radar precipitation measurement method based on the 45-degree depolarization ratio provided by the present invention. Based on observational raindrop spectrum simulations in East China, rainfall intensity is retrieved using an MLP neural network. and mass-weighted average diameter Normalized intercept Number concentration Moisture content Precision;

[0063] Figure 3 This is a schematic diagram of the radar precipitation measurement device based on a 45-degree depolarization ratio provided by the present invention. Detailed Implementation

[0064] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0065] The following is combined with Figure 1 This invention describes a radar precipitation measurement method based on a 45-degree depolarization ratio, comprising:

[0066] Step 101: Calculate the linear depolarization ratio of the slant baseline based on the two-channel complex voltage data acquired by the radar. The reflectivity factor Z, the radar transmits at a 45-degree oblique linear polarization.

[0067] Step 102: Identify the upper and lower boundaries of the melting layer based on the vertical pointing profile acquired by the radar in vertical pointing mode. After mapping the lower boundary of the melting layer to the radar horizontal scanning mode for observation, identify liquid precipitation and shield mixed phase or ice phase precipitation.

[0068] Step 103, adjust the linear depolarization ratio of the inclined line. Rain attenuation correction was performed using the reflectivity factor Z;

[0069] Step 104, based on the linear depolarization ratio of the sloping baseline after rain attenuation correction. The precipitation parameters are obtained by inversion with the reflectivity factor Z, including precipitation intensity, mass-weighted average diameter, normalized intercept, number concentration, and water content.

[0070] The overall concept of this embodiment is as follows:

[0071] 1. At the hardware level, by simply modifying existing cloud radar to enable near-horizontal scanning capability and a 45-degree oblique linear polarization transmission system, while retaining the vertical pointing mode, dual-mode operation of vertical cloud measurement and near-horizontal rainfall measurement can be achieved on the same platform. The added horizontal scanning not only allows for the measurement of reflectivity factor... radial velocity Spectral width It can also measure the depolarization ratio of the slant base. . use Sensitivity to rainfall shape, and Joint inversion of precipitation particle size and rainfall intensity.

[0072] 2. In vertical pointing mode, combined with sounding temperature data, utilizing Z and Vertical profiles identify the height and thickness of the melt layer for precipitation identification and quality control under near-horizontal scanning.

[0073] 3. Based on observed or ideal raindrop spectra, use scattering simulations to construct precipitation parameters (rainfall intensity). Mass-weighted average diameter D m Normalized intercept N w Number concentration N t Water content (LWC, etc.) and radar parameters ( and The mapping relationship between radar observation variables and precipitation parameters is established online using lookup tables or multilayer perceptrons.

[0074] 4. For equipment with phase measurement capabilities, the phase is calculated by combining the 45-degree oblique substrate back to the H / V substrate through substrate transformation. and Combined with existing Rain attenuation correction method, for and Perform along-the-path correction; for equipment without phase measurement capabilities, use a method based on... – A self-consistent rain attenuation correction method with optimal radial consistency.

[0075] This embodiment uses "radar modification to 45-degree oblique polarization" and " and With "precipitation inversion" as the core, and supplemented by "vertical pointing linear depolarization ratio melting layer identification" and "rain attenuation correction", a complete precipitation measurement technology chain of single-launch dual-receive cloud radar has been constructed.

[0076] Based on the above embodiments, this embodiment calculates the slant baseline linear depolarization ratio based on the two-channel complex voltage data acquired by the radar. Before the reflectivity factor Z, it also includes:

[0077] When the radar is unable to scan horizontally, an elevation and azimuth mechanism is added to it, or the scanning function is enabled to achieve near-horizontal scanning;

[0078] When the radar polarization mode is horizontal / vertical polarization, its transmission linear polarization direction is rotated by 45 degrees in a plane perpendicular to the beam propagation direction to achieve oblique 45-degree linear polarization transmission.

[0079] The radar maintains a dual-channel coherent receiving structure and reserves a vertical pointing mode for melting layer identification.

[0080] Based on an X-band cloud radar with an operating frequency of approximately 10 GHz, the radar was originally designed as a vertical pointing, single-transmitter dual-receiver system with two orthogonal linearly polarized receiving channels. Its main parameters include: antenna aperture: approximately 1 m, half-power beamwidth approximately 2 degrees; transmission mode: single-pulse transmission, linear polarization; reception mode: coherent reception through two channels (local polarity / cross polarity); original operating mode: fixed vertical pointing (elevation angle 90 degrees), output reflectivity Z and LDR vertical profile.

[0081] To implement the 45-degree depolarization rain measurement method, the radar underwent the following mechanical modifications:

[0082] Scanning mechanism modification: Add elevation and azimuth motors to the original antenna support to enable the radar to perform RHI scanning within the elevation angle range of 0-10 degrees and 360-degree PPI (horizontal) scanning at a fixed small elevation angle, while retaining the original vertical pointing mode.

[0083] Polarization roll structure modification: A controllable roll mechanism around the beam axis is added at the connection between the feed and the antenna, rotating the transmission polarization direction by 45 degrees relative to the original H / V reference axis through mechanical locking. After the modification, the radar transmits with a 45-degree oblique linear polarization during near-horizontal scanning, and can maintain the original polarization or use the same 45-degree oblique polarization as needed in vertical pointing mode.

[0084] Coherent reception and calibration: Maintain the original two-channel coherent reception structure, and calibrate the gain and phase of the two channels through noise injection and periodic solar observation (within the allowable elevation angle range) to ensure amplitude consistency better than 0.2 dB, phase consistency better than 2 degrees, and main lobe cross-polarization isolation better than 35 dB.

[0085] The radar can operate according to the following observation scheme: perform a PPI scan every 5 minutes, with the elevation angle fixed at 2 degrees; insert a vertical pointing profile every 10 minutes for melting layer identification; set the pulse accumulation number to 64 and the radial gate length to 100m.

[0086] Based on the above embodiments, this embodiment calculates the linear depolarization ratio of the slant baseline based on the two-channel complex voltage data acquired by the radar. The reflectivity factor Z includes:

[0087] The two-channel complex voltage data acquired by the radar are decoupled, and amplitude-phase compensation and cross-coupling elimination are performed using the calibrated 2×2 system error matrix.

[0088] Calculate the common polarization power and cross polarization power between the two channels of complex voltage data after decoupling;

[0089] After subtracting the noise, the common polarization power for noise depolarization is obtained. and cross-polarization power Then, the linear depolarization ratio of the sloping baseline is calculated on each pixel. ;

[0090] After noise reduction of the common polarization power, the reflectivity factor Z is obtained by converting it using radar constant and observation distance;

[0091] Noise removal and filtering are performed on the linear depolarization ratio (SLDR) of the sloping baseline in regions where the signal-to-noise ratio (SNR) is less than a preset threshold.

[0092] By employing methods such as internal noise injection, solar observation, or uniform light rain observation, the gain, phase, and cross-coupling of the receiving channel are estimated, and a 2×2 system error matrix is ​​constructed. During data processing, the observed voltage is decoupled and zero-point corrected, while the noise power of both channels is estimated.

[0093] Notice, The calculations also need to consider the radar constants of the two polarization channels. After obtaining the observation data, for low signal-to-noise ratio regions... Noise removal and filtering are performed to ensure statistical stability.

[0094] To suppress positive bias in weak echo regions, the SLDR calculation employs a noise estimation-based debiasing algorithm in regions with SNR < 10 dB, and performs 3×3 pixel median filtering spatially. Based on the above embodiments, this embodiment identifies the melting layer according to the vertical pointing profile acquired by the radar in vertical pointing mode, including:

[0095] In each vertically pointing profile, calculate the reflectivity factor Z profile as a function of height and the linear depolarization ratio (SLDR) profile at a 45-degree slant.

[0096] The center of the zero-degree layer bright band is estimated based on the local peak value in the reflectivity factor Z profile.

[0097] The upper and lower boundaries of the melt layer are determined based on the height range corresponding to the rise of the SLDR profile at a 45-degree depolarization ratio.

[0098] Based on the height and upper and lower boundaries of the melt layer, the data is mapped to radar horizontal scanning mode observations to identify liquid precipitation and shield mixed-phase or ice-phase precipitation. Subsequently, quantitative inversion is only carried out on liquid precipitation areas.

[0099] Based on the above embodiments, this embodiment focuses on the linear depolarization ratio of the sloping baseline. Rain attenuation correction is performed using the reflectivity factor Z, including:

[0100] For devices with phase measurement capabilities, the system outputs a dual-channel complex phase, and calculates the differential propagation phase shift after recombination from a 45-degree oblique substrate back to an H / V substrate via substrate transformation. and differential propagation phase shift rate ;

[0101] Using the differential propagation phase shift rate Based on the empirical relationship or self-consistent algorithm of the ratio attenuation, the linear depolarization ratio of the sloping baseline is... The reflectivity factor Z is corrected for rain attenuation along the path, and the corrected Z and SLDR are used as input variables to be fed into the LUT or MLP inversion process.

[0102] For devices that lack phase measurement capabilities, there is a lack of... and The method employs a method based on reflectivity factor Z and linear depolarization ratio of the sloping baseline. The self-consistent rain attenuation correction method with optimal radial consistency is applied to the linear depolarization ratio of the sloping baseline. Rain attenuation correction is performed using the reflectivity factor Z.

[0103] The method based on radial consistency of Z and SLDR includes: estimating path attenuation segment by segment along the radial direction so that the corrected Z is as consistent as possible with the Z and SLDR obtained by back-calculation from the inverted drop spectrum through the forward scattering model in an overall sense (such as minimizing the difference between the two in the mean square sense). Rain attenuation correction under phase-free conditions is achieved through this self-consistent recursive method.

[0104] In a typical case of moderate to heavy rain, this embodiment compares the regional cumulative rainfall obtained by inversion with the observation of ground rain gauges. The results show that in the range of 0 to 40 km, the bias is close to 0 and the correlation coefficient is greater than 0.9, which is better than the estimation effect of traditional single Z-R relationship and the case of uncorrected Z.

[0105] Based on the above embodiments, this embodiment uses the linear depolarization ratio of the slant line after rain attenuation correction. Inversion with the reflectivity factor Z yields the rainfall intensity. Mass-weighted average diameter Normalized intercept Number concentration Moisture content Precipitation parameters include:

[0106] Based on lookup table (LUT) or multilayer perceptron (MLP) neural network, the linear depolarization ratio of the slant line after rain attenuation correction is calculated. By inverting the reflectivity factor Z, precipitation parameters such as rainfall intensity, mass-weighted average diameter, normalized intercept, number concentration, and water content are obtained.

[0107] Based on the above embodiments, this embodiment uses LUT to correct the linear depolarization ratio of the slant line after rain attenuation. Inversion with the reflectivity factor Z yields the rainfall intensity. Mass-weighted average diameter Normalized intercept Number concentration Moisture content Precipitation parameters include:

[0108] Based on the T-matrix scattering model, a slant-based linear depolarization ratio was constructed. A lookup table for reflectivity factor Z and precipitation parameters such as rainfall intensity, mass-weighted average diameter, normalized intercept, number concentration, and water content;

[0109] Online observations After performing multidimensional interpolation, the corresponding result is obtained through the lookup table. , , , as well as Precipitation parameters.

[0110] In the offline phase, based on the electromagnetic scattering characteristics of the X-band, D is selected. m In 0.5 to 3 mm, N w In 10 2 Up to 10 5 mm -1 m -3 , In the parameter space of [–1,4], the corresponding Z and SLDR are calculated using a T-matrix model to construct a LUT, taking into account different temperatures and elevation angles. In the actual implementation, linear interpolation is used to balance accuracy and storage requirements.

[0111] During the online phase, for each near-horizontal observation point, its Z and SLDR information are read, and interpolation is performed in the LUT to obtain the corresponding... , , , as well as .

[0112] Based on the above embodiments, this embodiment uses an MLP neural network to correct the linear depolarization ratio of the slant baseline after rain attenuation. Inversion with the reflectivity factor Z yields the rainfall intensity. Mass-weighted average diameter Normalized intercept Number concentration Moisture content Precipitation parameters include:

[0113] Rain attenuation corrected slope baseline linear depolarization ratio The reflectivity factor Z is used as the input to the MLP neural network to obtain the output of the MLP neural network. , , , as well as .

[0114] The MLP neural network is pre-trained on simulated data and then fine-tuned on joint observations from measured radar and ground rain gauges / drop spectrometers to achieve an inversion that combines physical constraints with data-driven approaches.

[0115] The two inversion paths can be implemented separately or run in parallel within the same system to verify each other and obtain more stable and reliable inversion results.

[0116] The output includes , , , as well as The system generates a rasterized product containing information such as the height of the melting layer, and adds a quality flag to each pixel. Optionally, the radar inversion results are fused with ground rain gauge and droplet spectrometer data for bias correction.

[0117] In this embodiment, the following variables are selected as network input features: Observations: Z and SLDR are used as rainfall input information. The network output is: R (mm / h), D m (mm) (m -3 mm -1 ), (m -3 )and (g / m 3 ).

[0118] Network structure example: The input layer is followed by two fully connected hidden layers with 128 and 64 nodes respectively. The hidden layer activation function is ReLU (which can be supplemented with Batch Normalization and Dropout to improve training stability and suppress overfitting). To handle the differences in the dimensions and dynamic range of different inversion quantities, a variable-level transformation and scaling strategy is used on the output: for larger dynamic ranges... , In the logarithmic field (e.g.) Return, for , use Transformation, and Linear regression is maintained; subsequently, all transformed outputs are standardized to ensure comparable contributions of each dimension to the loss. Weighted mean square error is used for regression error; simultaneously, a physical consistency constraint based on forward scattering / empirical radar relationships is introduced to penalize inconsistent inversion results generated by the network output under the physical model, thereby improving interpretability and generalization ability. Model training employs a training / validation / test set partitioning combined with an early stopping strategy, selecting the model with the minimum validation set loss as the optimal inversion model.

[0119] Training and Fine-tuning: First, the MLP is pre-trained across the entire parameter space using a large amount of simulated data generated based on the T-matrix model, enabling the network to learn. and The basic relationships are established. Based on this, several precipitation cases with good ground rain gauges and drop spectrometers are selected, and the actual observation data are cleaned and paired to fine-tune the network so that its weights can be adapted to the system bias and noise characteristics of the X-band cloud radar.

[0120] During online operation, for each near-horizontal observation point, the observed Z, SLDR, and auxiliary features are input into the MLP to obtain R and D. m , , , The effects of random noise can be further reduced by using a simple 3×3 spatial window smoothing and a rolling average over time.

[0121] Figure 2 This is an example of the radar precipitation measurement method based on the 45-degree depolarization ratio provided by the present invention. The inversion uses an MLP neural network, and the evaluation data is the raindrop spectrum observation in East China. Figure 2 (a) is the retrieved rainfall intensity ; Figure 2 (b) is the mass-weighted average diameter ; Figure 2 (c) is the normalized intercept. ; Figure 2 (d) is the number concentration ; Figure 2 (e) is the water content. .

[0122] like Figure 2 As shown, based on radar observations ( and The established neural network inversion framework can retrieve precipitation parameters relatively well, with scatter points distributed on both sides of the unit line. The overall correlation coefficients on the test set are generally high. , and The correlation coefficients reached 0.978, 0.953, and 0.901, respectively. exist The domain correlation coefficient exceeds 0.848; The inversion correlation coefficient was slightly worse, but still reached 0.773. The main reason for the performance difference between different variables is that the sensitivity and identifiability of each output parameter to radar observation information are different. This uncertainty also exists in the inversion of conventional dual-transmitter dual-receiver dual-polarization radar.

[0123] The radar precipitation measurement device based on a 45-degree depolarization ratio provided by the present invention will be described below. The radar precipitation measurement device based on a 45-degree depolarization ratio described below can be referred to in correspondence with the radar precipitation measurement method based on a 45-degree depolarization ratio described above.

[0124] like Figure 3 As shown, the device includes a calculation module 201, an identification module 202, a correction module 203, and an inversion module 204, wherein:

[0125] The calculation module 201 is used to calculate the linear depolarization ratio of the slant baseline based on the two-channel complex voltage data acquired by the radar. And reflectivity factor Z, the radar transmits at a 45-degree oblique linear polarization;

[0126] The identification module 202 is used to identify the upper and lower boundaries of the melting layer based on the vertical pointing profile acquired by the radar in the vertical pointing mode, and after mapping the lower boundary of the melting layer to the radar horizontal scanning mode for observation, identify liquid precipitation and shield mixed phase or ice phase precipitation.

[0127] Correction module 203 is used to adjust the linear depolarization ratio of the slant line. Rain attenuation correction was performed using the reflectivity factor Z;

[0128] Inversion module 204 is used to determine the linear depolarization ratio of the sloping baseline after rain attenuation correction. By inverting the reflectivity factor Z, precipitation parameters such as rainfall intensity, mass-weighted average diameter, normalized intercept, number concentration, and water content are obtained.

[0129] This embodiment uses "radar modification to 45-degree oblique polarization" and " and With "precipitation inversion" as the core, and assisted by "vertical pointing mode" "Melting layer identification" and "rain attenuation correction" have constructed a complete rainfall measurement technology chain for single-launch dual-receive cloud radar.

[0130] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A radar precipitation measurement method based on a 45-degree depolarization ratio, characterized in that, include: Using two-channel complex voltage data acquired by radar, the linear depolarization ratio of the slant baseline is calculated. The radar employs a 45-degree slant linear polarization single-transmission dual-receive mode, along with a reflectivity factor Z. Based on the vertical pointing profile acquired by the radar in vertical pointing mode, the upper and lower boundaries of the melting layer are identified. After the lower boundary of the melting layer is mapped to the radar horizontal scanning mode for observation, liquid precipitation is identified and mixed phase or ice phase precipitation is shielded. Regarding the linear depolarization ratio of the sloping baseline Rain attenuation correction was performed using the reflectivity factor Z; Based on the linear depolarization ratio of the slope after rain attenuation correction The precipitation parameters are obtained by inversion with the reflectivity factor Z, including precipitation intensity, mass-weighted average diameter, normalized intercept, number concentration, and water content. Identifying the melting layer based on the vertical pointing profile acquired by the radar in vertical pointing mode includes: In each vertically pointing profile, the reflectivity factor Z profile and the linear depolarization ratio (SLDR) profile with 45-degree oblique angle are calculated as height. The center of the zero-degree layer bright band is estimated based on the local peak value in the Z-profile of the reflectivity factor. The upper and lower boundaries of the melting layer are determined based on the height range corresponding to the rise of the SLDR profile at a 45-degree angle. Based on the height and upper and lower boundaries of the melt layer, the data is mapped to radar horizontal scanning mode observations to identify liquid precipitation and shield mixed-phase or ice-phase precipitation. Quantitative inversion is then carried out only on liquid precipitation areas. Regarding the linear depolarization ratio of the sloping baseline Rain attenuation correction is performed using the reflectivity factor Z, including: For devices equipped with phase measurement capabilities, the differential propagation phase shift is calculated after recombination from a 45-degree oblique substrate back to an H / V substrate via substrate transformation. and differential propagation phase shift rate ; Using the differential propagation phase shift rate Based on the empirical relationship or self-consistent algorithm of the ratio attenuation, the linear depolarization ratio of the sloping baseline is... Rain attenuation correction was performed using the reflectivity factor Z; For devices lacking phase measurement capabilities, a method based on the reflectivity factor Z and the linear depolarization ratio of the slant baseline is used. The self-consistent rain attenuation correction method with optimal radial consistency is applied to the linear depolarization ratio of the sloping baseline. Rain attenuation correction is performed using the reflectivity factor Z.

2. The radar precipitation measurement method based on 45-degree depolarization ratio according to claim 1, characterized in that, Based on the two-channel complex voltage data acquired by the radar, the linear depolarization ratio of the slant baseline is calculated. And reflectivity factor Z, also includes: When the radar is unable to scan horizontally, an elevation and azimuth mechanism is added to it, or the scanning function is enabled to achieve near-horizontal scanning; When the radar polarization mode is horizontal / vertical polarization, its transmission linear polarization direction is rotated by 45 degrees in a plane perpendicular to the beam propagation direction to achieve oblique 45-degree linear polarization transmission. The radar maintains a dual-channel coherent receiving structure and reserves a vertical pointing mode for melting layer identification.

3. The radar precipitation measurement method based on a 45-degree depolarization ratio according to claim 1, characterized in that, Using two-channel complex voltage data acquired by radar, the linear depolarization ratio of the slant baseline is calculated. and reflectivity factor Z, including: Decouple the two-channel complex voltage data acquired by the radar; Calculate the common polarization power and cross polarization power between the two channels of complex voltage data after decoupling; The linear depolarization ratio of the sloping baseline is calculated based on the ratio of the common polarization power to the cross polarization power after noise depolarization in the power domain. ; After noise reduction of the common polarization power, the reflectivity factor Z is obtained by converting it using radar constant and observation distance; Noise removal and filtering are performed on the linear depolarization ratio (SLDR) of the sloping baseline in regions where the signal-to-noise ratio (SNR) is less than a preset threshold.

4. The radar precipitation measurement method based on a 45-degree depolarization ratio according to claim 1, characterized in that, Based on the linear depolarization ratio of the slope after rain attenuation correction And the reflectivity factor Z, to obtain precipitation parameters, including: Based on lookup table (LUT) or multilayer perceptron (MLP) neural network, the linear depolarization ratio of the sloping baseline after rain attenuation correction is analyzed. The precipitation parameters are obtained by inverting the reflectance factor Z.

5. The radar precipitation measurement method based on a 45-degree depolarization ratio according to claim 4, characterized in that, Based on LUT, the linear depolarization ratio of the sloping baseline after rain attenuation correction is calculated. Inversion with the reflectivity factor Z yields precipitation parameters, including: Based on the T-matrix scattering model, a slant-based linear depolarization ratio was constructed. A lookup table showing the correlation between reflectance factor Z and precipitation parameters such as rainfall intensity, mass-weighted average diameter, normalized intercept, number concentration, and water content; The observed linear depolarization ratio of the sloping baseline After multidimensional interpolation with the reflectivity factor Z, the corresponding precipitation parameters are obtained through the lookup table.

6. The radar precipitation measurement method based on a 45-degree depolarization ratio according to claim 4, characterized in that, Based on MLP neural network, the linear depolarization ratio of the sloping baseline after rain attenuation correction is calculated. Inversion with the reflectivity factor Z yields precipitation parameters, including: Rain attenuation corrected slope baseline linear depolarization ratio The precipitation parameters are obtained by using the reflectance factor Z as input to the MLP neural network.

7. A radar precipitation measurement device based on a 45-degree depolarization ratio, characterized in that, The radar precipitation measurement method based on a 45-degree depolarization ratio, applied to any one of claims 1-6, includes: The calculation module is used to calculate the linear depolarization ratio of the slant baseline based on the two-channel complex voltage data acquired by the radar. The radar employs a 45-degree slant linear polarization single-transmission dual-receive mode, along with a reflectivity factor Z. The identification module is used to identify the upper and lower boundaries of the melting layer based on the vertical pointing profile acquired by the radar in vertical pointing mode. After mapping the lower boundary of the melting layer to the radar horizontal scanning mode for observation, it identifies liquid precipitation and shields mixed phase or ice phase precipitation. The correction module is used to adjust the linear depolarization ratio of the slant line. Rain attenuation correction was performed using the reflectivity factor Z; The inversion module is used to determine the linear depolarization ratio of the sloping baseline after rain attenuation correction. The precipitation parameters are obtained by inversion with the reflectivity factor Z, including precipitation intensity, mass-weighted average diameter, normalized intercept, number concentration, and water content.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the radar precipitation measurement method based on the 45-degree depolarization ratio as described in any one of claims 1 to 6.

Images