Backscattering differential phase estimation method for polarized X-band weather radar
By eliminating redundant phases and outliers in polarized X-band weather radar, eliminating cumulative deviations, and filling the blank area with image interpolation method, the problem of low backscattering differential phase estimation accuracy in polarized X-band weather radar is solved. Improved the accuracy of rainfall inversion and melt layer classification.
Patent Information
- Application Number
- CN202411176934.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-26
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2044-08-26
AI Technical Summary
In polarized X-band weather radar, the estimation of backscattering differential phase has a problem of distance integral accumulation deviation, resulting in a decrease in accuracy, especially at a distance that is greater, affecting the accuracy of rainfall inversion and melt layer classification.
By obtaining the backscattered differential phase profile in each radial direction of the radar, removing redundant phases and outliers, splicing the paths, eliminating the accumulated deviations, and filling the blank areas with image interpolation method, the final backscattered differential phase profile is obtained.
The estimation accuracy of the backscattering differential phase is significantly improved, the influence of distance integral accumulation deviation is eliminated, and the reliability of rainfall inversion and melt layer classification is improved.
Smart Images

Figure CN119044974A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of meteorological radars, and in particular relates to a backscattering differential phase estimation method for a polarization X-band weather radar. Background Art
[0002] In recent years, with the urgent need for local severe convective rainfall monitoring and the continuous expansion of corresponding rainfall products, more and more polarimetric X-band weather radars have played an important role in rainfall observation. Compared with S-band and C-band radars, X-band radars have significant advantages in monitoring fast-developing storms at close range, due to their small size and low cost, higher resolution and greater mobility.
[0003] Backscatter differential phase δ hv It is an important polarization parameter of polarization X-band weather radar. hv It contains a lot of precipitation information, but is insensitive to the number concentration of the raindrop spectrum (the total number of raindrops per unit volume), but is highly correlated with the shape of raindrops and is a good parameter for inverting characteristic raindrop size. hv It helps to mitigate the uncertainties associated with the variation of raindrop spectrum and unknown raindrop shape, which can reduce the overestimation of rainfall rate when large raindrops or melted hail occur. In addition, δ hv It also helps to identify and classify hydrometeors contained in the melt layer. However, the delta hv is so small that it can be ignored, so it is usually used as a “redundant” addition to the measured differential phase Ψ DP The researchers are mainly interested in how to hv Accurately estimate the propagation differential phase shift rate K in the presence of DP , rather than measuring and using δ hv Therefore, the lack of precise δ hv The estimation method makes δ hv Systematic research and related applications have not been well explored.
[0004] However, in the X-band, the scattering of electromagnetic waves by moderate rain is already in a non-Rayleigh state, at which time δ hv is significant and can easily reach the same level as K DP The same order of magnitude. Therefore, for the X-band hv Make an accurate estimate and then based on δ hv Applications such as rainfall inversion and melt layer classification are promising and feasible. hv In order to ensure δ hv To obtain the resolution, we need to first estimate K DP , and estimate K DP Integrate along the path to obtain the propagation differential phase ΦDP , rather than directly filtering to obtain Φ DP However, the integration process accumulates K DP The deviation of the library is greater when the distance from the radar is farther away. This cumulative deviation will undoubtedly reduce δ hv Accuracy of estimates. Summary of the invention
[0005] To solve the above problems, the present invention provides a backscatter differential phase estimation method for polarization X-band weather radar, which can eliminate the influence of the deviation of distance integral accumulation on the backscatter differential phase estimation accuracy.
[0006] A backscatter differential phase estimation method for polarization X-band weather radar comprises the following steps:
[0007] S1: Obtain the backscatter differential phase profile δ of the radar in each radial direction hv ;
[0008] S2: Eliminate the backscatter differential phase profile δ in each radial direction of the radar hv The redundant phase on the backscattered image is removed to obtain the redundant backscattered differential phase profile.
[0009] S3: De-redundant backscatter differential phase profiles from each radial direction of the radar Filter out the backscattered differential phase segment as the road section sample;
[0010] S4: splicing the road section samples screened in each radial direction according to the set requirements to obtain the splicing path in each radial direction;
[0011] S5: Eliminate the accumulated deviations of the splicing paths in each radial direction respectively to obtain a non-deviation splicing path in each radial direction;
[0012] S6: Eliminate the backscattering differential phase outliers of the unbiased splicing path in each radial direction respectively to obtain a smooth splicing path in each radial direction;
[0013] S7: The backscattering differential phase corresponding to the blank distance library from which the outliers have been removed in the smooth stitching path in each radial direction is filled by the image interpolation method to obtain the final backscattering differential phase profile in each radial direction.
[0014] Furthermore, in step S1, the backscatter differential phase profile δ hv The method to obtain is:
[0015] S11: In the current radial direction, the estimated propagation differential phase shift rate is integrated along the transmission path of the radar signal to obtain the propagation differential phase on each range library:
[0016]
[0017] The range library is the detection interval evenly divided by the detection radius of the radar in the current radial direction, r k represents the position of the kth distance library, Φ DP (r k ) represents the propagation differential phase at the kth range bin, s represents the integral variable; K DP (r k ) represents the propagation differential phase shift rate at the kth range bin;
[0018] S12: Subtract the initial radar phase Φ from the differential phase measured by the radar at each range library DP0 , propagation differential phase Φ DP (r k ), and obtain the backscattering differential phase δ at the kth range bin hv (r k ):
[0019] δ hv (r k )=Ψ DP (r k )-Φ DP0 -Φ DP (r k )
[0020] Among them, DP (r k ) represents the measured differential phase of the radar at the kth range bin;
[0021] S13: The backscattered differential phase δ at all range bins hv (r k ) is combined to obtain the backscattering differential phase profile δ hv .
[0022] Furthermore, in step S2, the de-redundant backscatter differential phase profile of the radar in any radial direction is The method to obtain is:
[0023] S21: Attenuation correction is performed based on the linear mapping relationship between the differential attenuation rate and the propagation differential phase shift rate:
[0024] A DP (r k )=γ DP K DP (r k )
[0025] Among them, A DP (r k ) is the differential attenuation rate at the kth distance bin, γDP is a constant coefficient, K DP (r k ) is the propagation differential phase shift rate at the kth range bin;
[0026] S22: Integrate both sides of the linear mapping relationship along the path to obtain the relationship between the cumulative attenuation and the propagation differential phase as follows:
[0027] PIA DP (r k )=γ DP Φ DP (r k )
[0028] Among them, PIA DP (r k ) is the cumulative attenuation at the kth distance bin, Φ DP (r k ) is the propagation differential phase at the kth range bin;
[0029] S23: Based on the cumulative attenuation PIA DP (r k )Get the intrinsic differential reflectivity at each distance library:
[0030] Z DRint (r k )=Z DRatt (r k )+PIA DP (r k )
[0031] Among them, Z DRint (r k ) is the intrinsic differential reflectivity before attenuation at the kth distance bin, Z DRatt (r k ) is the differential reflectivity measured by the radar at the kth range bin;
[0032] S24: The intrinsic differential reflectivity Z at all distance bins DRint (r k ) to obtain the intrinsic differential reflectivity profile Z DRint ;
[0033] S25: According to the intrinsic differential reflectivity profile Z in the current radial direction DRint Get the standard deviation σ of the intrinsic differential reflectivity of each distance library in the current radial direction ZDRint (r k ), where any distance library r k The standard deviation of the intrinsic differential reflectivity σ ZDRint (r k ) is obtained as follows:
[0034] Select the intrinsic differential reflectivity profile Z of sliding windows of different sizes in the current radial direction DRint Intercept the data and obtain the intercepted data to the current distance library r k The intrinsic differential reflectivity Z on DRint (r k ) contains the standard deviation of all intrinsic differential reflectances in each sliding window; the average value of the standard deviations corresponding to all sliding windows is taken as the current distance library r k The corresponding intrinsic differential reflectivity standard deviation σ ZDRint (r k );
[0035] S26: For any distance library r on the current radial direction k , construct the intercept distance library r k And the minimum length is L min , the maximum is L max A series of scale intervals, and then the intrinsic differential reflectivity profile Z in the current radial direction DRint The current distance library r is in each scale interval k Select the distance library combination Δr(r m ; r n ):
[0036] |Z DRint (r m )-Z DRint (r n )|<σ ZDRint (r k )
[0037] Among them, r m ≤r k ≤r n , and r m ≠r n , L min To set the upper limit of the distance library, L max To set the upper limit of the distance library;
[0038] S27: For each distance library combination Δr(r m ; r n ), the measured differential phase profile Ψ of the radar DP The measured differential phase between the mth distance library and the nth distance library is taken as the differential phase sample ΔΨ DP (r m ; r n );
[0039] S28: Backscattering differential phase profile δ hv The upper filter out all differential phase samples ΔΨ DPThe corresponding backscattered differential phase sample Δδ hv (r m ; r n );
[0040] S29: Backscatter differential phase profile δ in the current radial direction hv Remove all backscattered differential phase samples Δδ hv (r m ; r n ), and obtain the de-redundant backscattered differential phase profile
[0041] Furthermore, in step S3, the method for obtaining road section samples in any radial direction is:
[0042] S31: De-redundant backscatter differential phase profile based on the current radial direction Get the backscatter differential phase standard deviation of each range library on the current radial direction Among them, any distance library r k The standard deviation of the backscattered differential phase The method to obtain is:
[0043] Select sliding windows of different sizes to obtain backscatter differential phase profiles in the current radial direction Intercept the data and obtain the intercepted data to the current distance library r k Backscatter differential phase The standard deviation of all backscattered differential phases contained in each sliding window of the current range library r is taken as the average value of the standard deviations corresponding to all sliding windows. k The corresponding backscatter differential phase standard deviation
[0044] S32: For any distance library r on the current radial direction k , construct the intercept distance library r k And the minimum length is L min , the maximum is L max A series of scale intervals, and then the backscattering differential phase profile in the current radial direction The current distance library r is in each scale interval k Select the distance library combination Δr(r m ; r n ):
[0045]
[0046] Among them, r m ≤r k ≤r n , and r m ≠rn , L min To set the upper limit of the distance library, L max To set the upper limit of the distance library;
[0047] S33: For each distance library combination Δr(r m ; r n ), in the backscattered differential phase profile The backscattered differential phase segment between the mth distance library and the nth distance library is cut out as the road section sample
[0048] Furthermore, in step S4, the setting requirements satisfied by the splicing path in any radial direction are:
[0049] There are no completely overlapping road segment samples in the spliced path;
[0050] The length of the spliced path covers all distance libraries in the current radial direction;
[0051] The length of the segment samples of the spliced path is greater than the set value.
[0052] Furthermore, in step S5, the method for obtaining the unbiased splicing path in any radial direction is:
[0053] S51: Get the sum of deviations of each distance library of the unbiased stitching path in the current radial direction:
[0054]
[0055] Among them, CumBias(r k ) is the kth distance library r of the unbiased splicing path in the current radial direction k The sum of deviations on the radial direction, M is the distance library r of the unbiased splicing path in the current radial direction k The total number of previous road segment samples, is the distance library r of the unbiased splicing path in the current radial direction k The deviation corresponding to the previous i-th road section sample, r mi is the starting point distance library in the i-th road section sample, r ni is the terminal distance library in the i-th road section sample;
[0056] S52: The backscattered differential phase δ of each distance library of the unbiased stitching path on the current radial direction is calculated. hv (r k ) minus the corresponding deviation sum CumBias(r k ) to obtain the unbiased splicing path in the current radial direction.
[0057] Furthermore, in step S6, the method for obtaining a smooth splicing path in any radial direction is:
[0058] S61: Eliminate the backscattering differential phase greater than 12° in the unbiased stitching path in the current radial direction to obtain a preprocessing path in the current radial direction;
[0059] S62: Use the numerical sliding window [K, K+△κ] to calculate the propagation differential phase shift rate profile K in the current radial direction DP The propagation differential phase shift rate data segment is intercepted, where K∈[min(K DP ),max(K DP )-△κ], △κ is the setting constant, min(K DP ) is the propagation differential phase shift rate profile K in the current radial direction DP The minimum value of the propagation differential phase shift rate on the DP ) is the propagation differential phase shift rate profile K in the current radial direction DP The maximum value of the propagation differential phase shift rate on ;
[0060] S63: Compare the preprocessing path in the current radial direction with the propagation differential phase shift rate profile K in the current radial direction DP According to the distance library, the backscattering differential phase data segment corresponding to the propagation differential phase shift rate data segment intercepted in step S62 is then found from the preprocessing path in the current radial direction;
[0061] S64: Eliminate the preprocessing paths that do not belong to the current radial direction. The backscattered differential phase in the interval is obtained to obtain the final smooth splicing path, where is the average value of the backscattering differential phase of the backscattering differential phase data segment obtained in step S63, It is the average value of the standard deviation of the backscattering differential phase of each distance library contained in the backscattering differential phase data segment obtained in step S63.
[0062] Beneficial effects:
[0063] The present invention provides a backscattering differential phase estimation method for polarization X-band weather radar, which eliminates δ hv On the basis of the outliers, the deviation of distance integral accumulation is eliminated to improve δ hv The accuracy of estimation is improved by applying the present invention to the data of typical storm events observed by the Netherlands polarization X-band weather radar to perform backscatter differential phase estimation. The experimental results show that the estimation method of the present invention improves the reliability of the backscatter differential phase and significantly eliminates the influence of the deviation of the distance integral accumulation on the accuracy of the backscatter differential phase estimation. BRIEF DESCRIPTION OF THE DRAWINGS
[0064] Figure 1 A schematic flow chart of a backscatter differential phase estimation method for polarization X-band weather radar provided by the present invention;
[0065] Figure 2 The measured differential reflectivity graph provided by the present invention;
[0066] Figure 3 The measured differential phase diagram provided by the present invention;
[0067] Figure 4 The estimated propagation differential phase shift rate map provided by the present invention;
[0068] Figure 5 The intrinsic differential reflectivity diagram after attenuation correction provided by the present invention;
[0069] Figure 6 The backscattering differential phase map after eliminating the accumulated deviation and outliers provided by the present invention;
[0070] Figure 7 The relationship diagram between the backscattering differential phase and the propagation differential phase shift rate provided by the present invention. DETAILED DESCRIPTION
[0071] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.
[0072] like Figure 1 As shown, a backscatter differential phase estimation method for polarization X-band weather radar includes the following steps:
[0073] S1: Obtain the backscatter differential phase profile δ of the radar in each radial direction hv ;
[0074] Among them, the backscatter differential phase profile δ of the radar in any radial direction is hv The method to obtain is:
[0075] S11: In the current radial direction, the estimated propagation differential phase shift rate is integrated along the transmission path of the radar signal to obtain the propagation differential phase on each range library:
[0076]
[0077] The range library is the detection interval evenly divided by the detection radius of the radar in the current radial direction, r k represents the position of the kth distance library, Φ DP (r k ) represents the propagation differential phase at the kth range bin, s represents the integral variable; K DP(r k ) represents the propagation differential phase shift rate at the kth range bin;
[0078] S12: Subtract the initial radar phase Φ from the differential phase measured by the radar at each range library DP0 , propagation differential phase Φ DP (r k ), and obtain the backscattering differential phase δ at the kth range bin hv (r k ):
[0079] δ hv (r k )=Ψ DP (r k )-Φ DP0 -Φ DP (r k )
[0080] Among them, DP (r k ) represents the measured differential phase of the radar at the kth range bin;
[0081] S13: The backscattered differential phase δ at all range bins hv (r k ) is combined to obtain the backscattering differential phase profile δ hv .
[0082] S2: Eliminate the backscatter differential phase profile δ in each radial direction of the radar hv The redundant phase on the backscattered image is removed to obtain the redundant backscattered differential phase profile.
[0083] Among them, the de-redundant backscatter differential phase profile of the radar in any radial direction is The method to obtain is:
[0084] S21: Attenuation correction is performed based on the linear mapping relationship between the differential attenuation rate and the propagation differential phase shift rate:
[0085] A DP (r k )=γ DP K DP (r k )
[0086] Among them, A DP (r k ) is the differential attenuation rate at the kth distance bin, γ DP is a constant coefficient, K DP (r k ) is the propagation differential phase shift rate at the kth range bin;
[0087] S22: Integrate both sides of the linear mapping relationship along the path to obtain the relationship between the cumulative attenuation and the propagation differential phase as follows:
[0088] PIA DP (r k )=γ DP Φ DP (r k )
[0089] Among them, PIA DP (r k ) is the cumulative attenuation at the kth distance bin, Φ DP (r k ) is the propagation differential phase at the kth range bin;
[0090] S23: Based on the cumulative attenuation PIA DP (r k )Get the intrinsic differential reflectivity at each distance library:
[0091] Z DRint (r k )=Z DRatt (r k )+PIA DP (r k )
[0092] Among them, Z DRint (r k ) is the intrinsic differential reflectivity before attenuation at the kth distance bin, Z DRatt (r k ) is the differential reflectivity measured by the radar at the kth range bin;
[0093] S24: The intrinsic differential reflectivity Z at all distance bins DRint (r k ) to obtain the intrinsic differential reflectivity profile Z DRint ;
[0094] S25: According to the intrinsic differential reflectivity profile Z in the current radial direction DRint Get the standard deviation σ of the intrinsic differential reflectivity of each distance library in the current radial direction ZDRint (r k ), where any distance library r k The standard deviation of the intrinsic differential reflectivity σ ZDRint (r k ) is obtained as follows:
[0095] Select the intrinsic differential reflectivity profile Z of sliding windows of different sizes in the current radial direction DRint Intercept the data and obtain the intercepted data to the current distance library r kThe intrinsic differential reflectivity Z on DRint (r k ) contains the standard deviation of all intrinsic differential reflectances in each sliding window; the average value of the standard deviations corresponding to all sliding windows is taken as the current distance library r k The corresponding intrinsic differential reflectivity standard deviation σ ZDRint (r k );
[0096] S26: For any distance library r on the current radial direction k , construct the intercept distance library r k And the minimum length is L min , the maximum is L max A series of scale intervals, and then the intrinsic differential reflectivity profile Z in the current radial direction DRint The current distance library r is in each scale interval k Select the distance library combination Δr(r m ; r n ):
[0097] |Z DRint (r m )-Z DRint (r n )|<σ ZDRint (r k )
[0098] Among them, r m ≤r k ≤r n , and r m ≠r n , L min To set the upper limit of the distance library, L max To set the upper limit of the distance library;
[0099] S27: For each distance library combination Δr(r m ; r n ), the measured differential phase profile Ψ of the radar DP The measured differential phase between the mth distance library and the nth distance library is taken as the differential phase sample ΔΨ DP (r m ; r n );
[0100] S28: Backscattering differential phase profile δ hv The upper filter out all differential phase samples ΔΨ DP The corresponding backscattered differential phase sample Δδ hv (r m ; r n );
[0101] S29: Backscatter differential phase profile δ in the current radial direction hv Remove all backscattered differential phase samples Δδ hv (r m ; r n ), and obtain the de-redundant backscattered differential phase profile
[0102] S3: De-redundant backscatter differential phase profiles from each radial direction of the radar Filter out the backscattered differential phase segment as the road section sample;
[0103] The method for obtaining road section samples on any radial direction is as follows:
[0104] S31: De-redundant backscatter differential phase profile based on the current radial direction Get the backscatter differential phase standard deviation of each range library on the current radial direction Among them, any distance library r k The standard deviation of the backscattered differential phase The method to obtain is:
[0105] Select sliding windows of different sizes to obtain backscatter differential phase profiles in the current radial direction Intercept the data and obtain the intercepted data to the current distance library r k Backscatter differential phase The standard deviation of all backscattered differential phases contained in each sliding window of the current range library r is taken as the average value of the standard deviations corresponding to all sliding windows. k The corresponding backscatter differential phase standard deviation
[0106] S32: For any distance library r on the current radial direction k , construct the intercept distance library r k And the minimum length is L min , the maximum is L max A series of scale intervals, and then the backscattering differential phase profile in the current radial direction The current distance library r is in each scale interval k Select the distance library combination Δr(r m ; r n ):
[0107]
[0108] Among them, r m ≤r k ≤r n , and r m ≠r n , Lmin To set the upper limit of the distance library, L max To set the upper limit of the distance library;
[0109] S33: For each distance library combination Δr(r m ; r n ), in the backscattered differential phase profile The backscattered differential phase segment between the mth distance library and the nth distance library is cut out as the road section sample
[0110] S4: splicing the road section samples screened in each radial direction according to the set requirements to obtain the splicing path in each radial direction;
[0111] Among them, the setting requirements satisfied by the splicing path in any radial direction are:
[0112] There are no completely overlapping road segment samples in the spliced path;
[0113] The length of the spliced path covers all distance libraries in the current radial direction;
[0114] The length of the segment samples of the spliced path is greater than the set value.
[0115] That is to say, the length of the spliced single road segment sample is as long as possible, the path samples are not overlapped as much as possible, and all distance libraries are traversed by the spliced path samples as much as possible.
[0116] S5: Eliminate the accumulated deviations of the splicing paths in each radial direction respectively to obtain a non-deviation splicing path in each radial direction;
[0117] The method for obtaining the unbiased splicing path in any radial direction is as follows:
[0118] S51: Get the sum of deviations of each distance library of the unbiased stitching path in the current radial direction:
[0119]
[0120] Among them, CumBias(r k ) is the kth distance library r of the unbiased splicing path in the current radial direction k The sum of deviations on the radial direction, M is the distance library r of the unbiased splicing path in the current radial direction k The total number of previous road segment samples, is the distance library r of the unbiased splicing path in the current radial direction k The deviation corresponding to the previous i-th road section sample, r mi is the starting point distance library in the i-th road section sample, r ni is the terminal distance library in the i-th road section sample;
[0121] S52: The backscattered differential phase δ of each distance library of the unbiased stitching path on the current radial direction is calculated. hv (r k ) minus the corresponding deviation sum CumBias(r k ) to obtain the unbiased splicing path in the current radial direction.
[0122] S6: Eliminate the backscattering differential phase outliers of the unbiased splicing path in each radial direction respectively to obtain a smooth splicing path in each radial direction;
[0123] In step S6, the method for obtaining the smooth splicing path in any radial direction is:
[0124] S61: Eliminate the backscattering differential phase greater than 12° in the unbiased stitching path in the current radial direction to obtain a preprocessing path in the current radial direction;
[0125] S62: Use the numerical sliding window [K, K+△κ] to calculate the propagation differential phase shift rate profile K in the current radial direction DP The propagation differential phase shift rate data segment is intercepted, where K∈[min(K DP ),max(K DP )-△κ], △κ is the setting constant, min(K DP ) is the propagation differential phase shift rate profile K in the current radial direction DP The minimum value of the propagation differential phase shift rate on the DP ) is the propagation differential phase shift rate profile K in the current radial direction DP The maximum value of the propagation differential phase shift rate on ;
[0126] S63: Compare the preprocessing path in the current radial direction with the propagation differential phase shift rate profile K in the current radial direction DP According to the distance library, the backscattering differential phase data segment corresponding to the propagation differential phase shift rate data segment intercepted in step S62 is then found from the preprocessing path in the current radial direction;
[0127] S64: Eliminate the preprocessing paths that do not belong to the current radial direction. The backscattered differential phase in the interval is obtained to obtain the final smooth splicing path, where is the average value of the backscattering differential phase of the backscattering differential phase data segment obtained in step S63, It is the average value of the standard deviation of the backscattering differential phase of each distance library contained in the backscattering differential phase data segment obtained in step S63.
[0128] S7: The backscattering differential phase corresponding to the blank distance library from which the outliers have been removed in the smooth stitching path in each radial direction is filled by the image interpolation method to obtain the final backscattering differential phase profile in each radial direction.
[0129] The following uses the polarization X-band (9.45 GHz) weather radar (IDRA) data of the International Center for Telecommunications and Radar Research (IRCTR) located at the Cabauw Atmospheric Research Experiment Station (CESAR) observation station in the Netherlands (NL) to verify the content of the present invention.
[0130] The radar is 213m above the ground, with a working distance and range resolution of 15.3km and 0.03km respectively. The scanning height is fixed at 0.50°, the antenna rotates 360° within 1min, and the beam width is 1.8°. A convective storm event occurred in the Netherlands on September 10, 2011. For the radar data at 1946UTC, Z is discarded. DRatt ,Ψ DP Medium linear depolarization ratio L DR (dB) greater than -18dB to remove particles other than rain and / or areas with low signal-to-noise ratio (SNR). In addition, to reduce the impact of noise, Z DRatt ,Ψ DP Speckle filtering was performed respectively. DRatt ,Ψ DP The PPIs are as follows Figure 2 and Figure 3 shown.
[0131] This embodiment provides a backscatter differential phase estimation method for a polarization X-band weather radar, which specifically includes the following steps:
[0132] Step 1 specifically includes:
[0133] In a certain radial direction, the estimated propagation differential phase shift rate K DP (° / km)(K DP Domain Figure 4 As shown in the figure, the propagation differential phase Φ is obtained by integrating along the path. DP (°) Profile (r k represents the position of the kth distance bin, where k represents any distance bin on the radial direction):
[0134]
[0135] The differential phase Ψ measured from the radar DP (°) Subtract the radar initial phase Φ from the profile DP0 (°), then subtract the integral obtained from Φ DP The profile is δ hv (°) Profile:
[0136] δ hv (r k )=Ψ DP (r k )-Φ DP0 -Φ DP (r k )
[0137] Repeat the above steps on all remaining radial directions.
[0138] Step 2 specifically includes:
[0139] Step 21, attenuation correction:
[0140] In a certain radial direction, through A DP (dB / km) and K DP The empirical linear relationship between the two is used to make attenuation correction:
[0141] A DP (r k )=γ DP K DP (r k )
[0142] According to the results of X-band radar theoretical simulation, γ DP =0.03. Integrate both sides of the above equation along the path to obtain the cumulative attenuation PIA DP (dB) and Ф DP The relationship is:
[0143] PIA DP (r k )=γ DP Φ DP (r k )
[0144] Then the attenuation-corrected estimated intrinsic differential reflectivity Z DRint (dB) is:
[0145] Z DRint (r k )=Z DRatt (r k )+PIA DP (r k )
[0146] Repeat the above operation on all remaining radial directions to obtain Z DRint Domain, such as Figure 5 shown.
[0147] Step 22, ΔΨ DP Sample search:
[0148] In a certain radial direction, for the distance library k, calculate the Z in the multiple sliding windows where it is located DRint The average of the standard deviations is used as the standard deviation σ of the distance library k ZDRint (r k ). According to the results of X-band radar theoretical simulation, Δr(r m ; r n ) is in the range [3km; 5km], and the unit distance library length is 0.03km as the step length. In this interval, select all Δr(r m ; r n )(r m ≤r k ≤r n , where r m ≠r n ):
[0149] |Z DRint (r m )-Z DRint (r n )|<σ ZDRint (r k )
[0150] For each Δr(r m ; r n ), the corresponding ΔΨ DP Δδ in the sample hv (r m ; r n ) can be ignored and the above operation is repeated in all remaining radial directions.
[0151] Step 3 specifically includes:
[0152] Step 31, path sample screening:
[0153] In a certain radial direction, hv Profile and Ψ DP The profile corresponds to hv Screen out all Δr(r m ; r n ), the corresponding Δδ hv The samples are called path samples:
[0154]
[0155] in is the distance from the library k to the multiple sliding windows δ hv The mean of the standard deviations, Δδ, of these path samples hv (r m ; r n) cannot be ignored, and the above operation is repeated in all remaining radial directions.
[0156] Step 4 specifically includes:
[0157] Step 41, path sample splicing:
[0158] In a certain radial direction, the path samples are spliced, and the splicing process must meet the following requirements:
[0159] 1) The length of the single-path sample should be as long as possible;
[0160] 2) Try not to overlap path samples;
[0161] 3) All distance libraries are traversed by spliced path samples as much as possible;
[0162] Repeat the above steps on all remaining radial directions.
[0163] Step 42, cumulative deviation elimination:
[0164] In a certain radial direction, for distance library k, the cumulative deviation of the distance library is eliminated according to the sum of the deviations between the first and last distance libraries of all previous path samples:
[0165]
[0166] Repeat the above operation on all remaining radial directions to eliminate the entire δ hv The accumulated deviation in the domain.
[0167] Step 5 specifically includes:
[0168] Elimination of δ hv δ in the domain hv For abnormal values >12°, δ hv Domain and K DP Corresponding to the domain, the numerical sliding window [K, K+△κ] is used to adjust the hv Eliminate δ on the domain hv In the interval The value outside, where K∈[min(K DP ),max(K DP )-△κ], △κ is defined according to the following formula:
[0169]
[0170] Step 6 specifically includes:
[0171] Fill in the gaps due to δ by using image interpolation algorithm hv δ due to outlier elimination hv The blank of the domain is obtained, that is, the δ after eliminating the cumulative deviation and outliers hv Domain, such as Figure 6 shown.
[0172] δ hv With K DP The relationship diagram is as follows Figure 7 The green one is the scatter plot between the two after eliminating the cumulative deviation, while the red one is the scatter plot between the two without eliminating the cumulative deviation. With K DP The blue one is to eliminate the cumulative deviation. With K DP The relationship diagram of gold is to eliminate the cumulative deviation With K DP The relationship diagram, where is δ hv Mean value, σ δhv is δ hv Standard deviation. Pink is δ hv With K DP Compared with the δ without eliminating the cumulative deviation hv Estimates, eliminating the cumulative bias of δ hv A large reduction in outliers in estimates and Follow K DP The strong consistency between the changing profile and the simulated fitting profile indirectly verifies the good effect of the method of the present invention on eliminating the cumulative deviation and the δ hv Improved accuracy of estimates.
[0173] In summary, in the X-band, the scattering of electromagnetic waves by moderate rain is already in a non-Rayleigh state. hv is significant and can easily reach the same level as K DP The same order of magnitude. Therefore, for the X-band hv Make an accurate estimate and then based on δ hv Applications such as rainfall inversion and melt layer classification are promising and feasible. hv In order to ensure δ hv To obtain the resolution, we need to first estimate K DP , and estimate K DP Integrate along the path to obtain the propagation differential phase Φ DP , rather than directly filtering to obtain Φ DP However, the integration process accumulates K DP The deviation of the library is greater when the distance from the radar is farther away. This cumulative deviation will undoubtedly reduce δ hv The present invention proposes a backscatter differential phase estimation method for polarized X-band weather radar, which eliminates δ hv On the basis of the outliers, the deviation of distance integral accumulation is eliminated to improve δhv The invention is applied to perform backscatter differential phase estimation on data of typical storm events observed by the Netherlands polarization X-band weather radar, thereby improving the reliability of the data.
[0174] Of course, the present invention may have many other embodiments. Without departing from the spirit and essence of the present invention, those skilled in the art may certainly make various corresponding changes and modifications based on the present invention, but these corresponding changes and modifications should all fall within the scope of protection of the claims attached to the present invention.
Claims
1. A backscatter differential phase estimation method for polarization X-band weather radar, characterized in that: The following steps are involved: S1: Obtain the backscatter differential phase profile δ of the radar in each radial direction hv ; S2: Eliminate the backscatter differential phase profile δ in each radial direction of the radar hv The redundant phase on the backscattered image is removed to obtain the redundant backscattered differential phase profile. S3: De-redundant backscatter differential phase profiles from each radial direction of the radar Filter out the backscattered differential phase segment as the road section sample; S4: splicing the road section samples screened in each radial direction according to the set requirements to obtain the splicing path in each radial direction; S5: Eliminate the accumulated deviations of the splicing paths in each radial direction respectively to obtain a non-deviation splicing path in each radial direction; S6: Eliminate the backscattering differential phase outliers of the unbiased splicing path in each radial direction respectively to obtain a smooth splicing path in each radial direction; S7: The backscattering differential phase corresponding to the blank distance library from which the outliers have been removed in the smooth stitching path in each radial direction is filled by the image interpolation method to obtain the final backscattering differential phase profile in each radial direction.
2. The backscatter differential phase estimation method for polarization X-band weather radar according to claim 1, characterized in that: In step S1, the backscatter differential phase profile δ hv The method to obtain is: S11: In the current radial direction, the estimated propagation differential phase shift rate is integrated along the transmission path of the radar signal to obtain the propagation differential phase on each range library: The range library is the detection interval evenly divided by the detection radius of the radar in the current radial direction, r k represents the position of the kth distance library, Φ DP (r k ) represents the propagation differential phase at the kth range bin, s represents the integral variable; K DP (r k ) represents the propagation differential phase shift rate at the kth range bin; S12: Subtract the initial radar phase Φ from the differential phase measured by the radar at each range library DP0 , propagation differential phase Φ DP (r k ), and obtain the backscattering differential phase δ at the kth range bin hv (r k ): d hv (r k )=Ψ DP (r k )-F DP0 -F DP (r k ) Among them, DP (r k ) represents the measured differential phase of the radar at the kth range bin; S13: The backscattered differential phase δ at all range bins hv (r k ) is combined to obtain the backscattering differential phase profile δ hv .
3. The backscatter differential phase estimation method for polarization X-band weather radar according to claim 1, characterized in that: In step S2, the de-redundant backscatter differential phase profile of the radar in any radial direction is The method to obtain is: S21: Attenuation correction is performed based on the linear mapping relationship between the differential attenuation rate and the propagation differential phase shift rate: A DP (r k )=γ DP K DP (r k ) Among them, A DP (r k ) is the differential attenuation rate at the kth distance bin, γ DP is a constant coefficient, K DP (r k ) is the propagation differential phase shift rate at the kth range bin; S22: Integrate both sides of the linear mapping relationship along the path to obtain the relationship between the cumulative attenuation and the propagation differential phase as follows: PIA DP (r k )=γ DP F DP (r k ) Among them, PIA DP (r k ) is the cumulative attenuation at the kth distance bin, Φ DP (r k ) is the propagation differential phase at the kth range bin; S23: Based on the cumulative attenuation PIA DP (r k )Get the intrinsic differential reflectivity at each distance library: WITH DRint (r k )=Z DRatt (r k )+PIA DP (r k ) Among them, Z DRint (r k ) is the intrinsic differential reflectivity before attenuation at the kth distance bin, Z DRatt (r k ) is the differential reflectivity measured by the radar at the kth range bin; S24: The intrinsic differential reflectivity Z at all distance bins DRint (r k ) to obtain the intrinsic differential reflectivity profile Z DRint ; S25: According to the intrinsic differential reflectivity profile Z in the current radial direction DRint Get the standard deviation σ of the intrinsic differential reflectivity of each distance library in the current radial direction ZDRint (r k ), where any distance library r k The standard deviation of the intrinsic differential reflectivity σ ZDRint (r k ) is obtained as follows: Select the intrinsic differential reflectivity profile Z of sliding windows of different sizes in the current radial direction DRint Intercept the data and obtain the intercepted data to the current distance library r k The intrinsic differential reflectivity Z on DRint (r k ) contains the standard deviation of all intrinsic differential reflectances in each sliding window; the average value of the standard deviations corresponding to all sliding windows is taken as the current distance library r k The corresponding intrinsic differential reflectivity standard deviation σ ZDRint (r k ); S26: For any distance library r on the current radial direction k , construct the intercept distance library r k And the minimum length is L min , the maximum is L max A series of scale intervals, and then the intrinsic differential reflectivity profile Z in the current radial direction DRint The current distance library r is in each scale interval k Select the distance library combination Δr(r m ; r n ): Among them, r m ≤r k ≤r n , and r m ≠r n , L min To set the upper limit of the distance library, L max To set the upper limit of the distance library; S27: For each distance library combination Δr(r m ; r n ), the measured differential phase profile Ψ of the radar DP The measured differential phase between the mth distance library and the nth distance library is cut out as the differential phase sample ΔΨ DP (r m ; r n ); S28: Backscattering differential phase profile δ hv The upper filter out all differential phase samples ΔΨ DP The corresponding backscattered differential phase sample Δδ hv (r m ; r n ); S29: Backscatter differential phase profile δ in the current radial direction hv Remove all backscattered differential phase samples Δδ hv (r m ; r n ), and obtain the de-redundant backscattered differential phase profile 4. The backscatter differential phase estimation method for polarization X-band weather radar according to claim 1, characterized in that: In step S3, the method for obtaining the road section samples in any radial direction is: S31: De-redundant backscatter differential phase profile based on the current radial direction Get the backscatter differential phase standard deviation of each range library on the current radial direction Among them, any distance library r k The standard deviation of the backscattered differential phase The method to obtain is: Select sliding windows of different sizes to obtain backscatter differential phase profiles in the current radial direction Intercept the data and obtain the intercepted data to the current distance library r k Backscatter differential phase The standard deviation of all backscattered differential phases contained in each sliding window of the current range library r is taken as the average value of the standard deviations corresponding to all sliding windows. k The corresponding backscatter differential phase standard deviation S32: For any distance library r on the current radial direction k , construct the intercept distance library r k And the minimum length is L min , the maximum is L max A series of scale intervals, and then the backscattering differential phase profile in the current radial direction The current distance library r is in each scale interval k Select the distance library combination Δr(r m ; r n ): Among them, r m ≤r k ≤r n , and r m ≠r n , L min To set the upper limit of the distance library, L max To set the upper limit of the distance library; S33: For each distance library combination Δr(r m ; r n ), in the backscattered differential phase profile The backscattered differential phase segment between the mth distance library and the nth distance library is cut out as the road section sample 5. The backscatter differential phase estimation method for polarization X-band weather radar according to claim 1, characterized in that: In step S4, the setting requirements satisfied by the splicing path in any radial direction are: There are no completely overlapping road segment samples in the spliced path; The length of the spliced path covers all distance libraries in the current radial direction; The length of the segment samples of the spliced path is greater than the set value.
6. The backscatter differential phase estimation method for polarization X-band weather radar according to claim 1, characterized in that: In step S5, the method for obtaining the unbiased splicing path in any radial direction is: S51: Get the sum of deviations of each distance library of the unbiased stitching path in the current radial direction: Among them, CumBias(r k ) is the kth distance library r of the unbiased splicing path in the current radial direction k The sum of deviations on the radial direction, M is the distance library r of the unbiased splicing path on the current radial direction k The total number of previous road segment samples, is the distance library r of the unbiased splicing path in the current radial direction k The deviation corresponding to the previous i-th road section sample, is the starting point distance library in the i-th road section sample, is the terminal distance library in the i-th road section sample; S52: The backscattered differential phase δ of each distance library of the unbiased stitching path on the current radial direction is calculated. hv (r k ) minus the corresponding deviation sum CumBias(r k ) to obtain the unbiased splicing path in the current radial direction.
7. The backscatter differential phase estimation method for polarization X-band weather radar according to claim 1, characterized in that: In step S6, the method for obtaining a smooth splicing path in any radial direction is: S61: Eliminate the backscattering differential phase greater than 12° in the unbiased stitching path in the current radial direction to obtain a preprocessing path in the current radial direction; S62: Use the numerical sliding window [K, K+△κ] to calculate the propagation differential phase shift rate profile K in the current radial direction DP The propagation differential phase shift rate data segment is intercepted, where K∈[min(K DP ),max(K DP )-△κ], △κ is the setting constant, min(K DP ) is the propagation differential phase shift rate profile K in the current radial direction DP The minimum value of the propagation differential phase shift rate on the DP ) is the propagation differential phase shift rate profile K in the current radial direction DP The maximum value of the propagation differential phase shift rate on ; S63: Compare the preprocessing path in the current radial direction with the propagation differential phase shift rate profile K in the current radial direction DP According to the distance library, the backscattering differential phase data segment corresponding to the propagation differential phase shift rate data segment intercepted in step S62 is then found from the preprocessing path in the current radial direction; S64: Eliminate the preprocessing paths that do not belong to the current radial direction. The backscattered differential phase in the interval is obtained to obtain the final smooth splicing path, where is the average value of the backscattering differential phase of the backscattering differential phase data segment obtained in step S63, It is the average value of the standard deviation of the backscattering differential phase of each distance library contained in the backscattering differential phase data segment obtained in step S63.
Citation Information
Patent Citations
X-band dual-polarization weather radar precipitation estimation method
CN108562904A
X-band dual-polarization weather radar quantitative precipitation measurement method and system
CN110488393A
Method and system for processing differential phase of dual-polarization radar
CN113466868A
Quality control method for differential phase of dual-polarization meteorological radar
CN113777567A
X wave band type dual polarization phase control differential reflectivity calibration method based on light rain method
CN115932764A
Cited By
X-band dual-polarization radar differential propagation phase shift quality control method
CN120446890A
