Ocean vortex region one-dimensional wave spectrum inversion method based on synthetic aperture radar

Abstract

The invention provides an ocean vortex region one-dimensional wave spectrum inversion method based on a synthetic aperture radar, and the method comprises the steps: recognizing an ocean vortex through HY-2 fusion data, collecting a Sentinel-1 dual-polarized SAR image covering a vortex region, and obtaining WAVEWATCH-III wave mode data; a hydrodynamic modulation transfer function is improved, and a vortex shear flow influence factor is introduced; an XGBoost machine learning model is adopted, and an ocean vortex region SAR one-dimensional sea wave spectrum inversion algorithm is established; and verifying the accuracy of the SAR one-dimensional sea wave spectrum inversion algorithm. According to the method, the precision limitation of traditional wave theory inversion in a vortex region is broken through, and the one-dimensional wave spectrum inversion precision is improved.

Description

Technical Field

[0001] This invention relates to the field of marine remote sensing inversion technology, specifically to a method for inverting one-dimensional wave spectra in marine eddy regions based on synthetic aperture radar. Background Technology

[0002] Mesoscale eddies in the ocean significantly modulate surface wave parameters (wave height, period, and direction) through shear flow fields and convergence-divergence effects, significantly impacting ocean energy transfer and extreme sea state prediction. Current observation techniques suffer from the following bottlenecks: limited data acquisition, with ship-based / buoy observations struggling to simultaneously cover the dynamic evolution of eddies; insufficient spatial resolution (>25km) of traditional remote sensing (altimeters / scatterometers), failing to resolve wave modulation details at eddy boundaries (typically 10-100km); limited swath coverage and long revisit periods for SWIM spectrometers; and a lack of physical mechanisms for spectral inversion in synthetic aperture radar (SAR): the classical hydrodynamic modulation transfer function (MTF) does not consider velocity gradient effects (divergence, vorticity, strain field) induced by eddy shear flow, leading to distorted spectral modeling. Furthermore, algorithms lack adaptability; empirical models (such as CWAVE) have poor generalization capabilities in strong shear zones, and theoretical algorithms (MPI / PFSM) rely on initial spectral guesses, resulting in significant error propagation in eddy-wave coupled fields. The verification methods are lacking, and there is a lack of high-resolution field observation data to support the verification of the wave spectrum in the vortex region.

[0003] Current SAR spectral inversion technology faces three major bottlenecks in mesoscale vortex environments: incomplete physical mechanisms, lack of flow field coupling, and insufficient verification data. These bottlenecks restrict the quantitative analysis of wave-vortex interaction mechanisms. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the present invention aims to provide a one-dimensional wave spectrum inversion method for ocean eddy regions based on synthetic aperture radar. This invention overcomes the accuracy limitations of traditional wave theory inversion in eddy regions and improves the accuracy of one-dimensional wave spectrum inversion.

[0005] To solve the above problems, the technical solution of the present invention is as follows:

[0006] A method for inverting one-dimensional wave spectra in ocean eddy regions based on synthetic aperture radar includes the following steps:

[0007] Ocean eddies were identified using HY-2 fusion data, and Sentinel-1 dual-polarization SAR images covering the eddy region were acquired to obtain WAVEWATCH-Ⅲ wave model data.

[0008] Improve the fluid dynamics modulation transfer function and introduce the vortex shear flow influence factor;

[0009] An algorithm for inverting the SAR wave spectrum in the ocean eddy region was established using the XGBoost machine learning model.

[0010] Verify the accuracy of the SAR one-dimensional wave spectrum inversion algorithm.

[0011] Preferably, the step of identifying ocean eddies using HY-2 fused data and acquiring Sentinel-1 dual-polarization SAR images covering the eddy region to obtain WAVEWATCH-III wave model data specifically includes: collecting HY-2 fused SLA data from 2022 to 2024 processed using vector geometry algorithms, acquiring more than 2,500 Sentinel-1 dual-polarization images covering the eddy region, and obtaining data from the WAVEWATCH-III model at the corresponding time and location.

[0012] Preferably, the step of improving the fluid dynamics modulation transfer function and introducing the vortex shear flow influence factor specifically includes: improving the fluid dynamics modulation transfer function to make it more suitable for the vortex region, wherein the formula for the improved fluid dynamics modulation transfer function is:

[0013] T hy =p0+p1ω+p2k+p3sinφ+(q0+q1ω+q2k+q3sinφ)i

[0014] Where k represents the wave number, k l The wave number represents the direction of the radar, ω is the angular frequency of the microscale waves on the sea surface, θ is the radar incident angle, θ is the wave direction relative to the azimuth direction, R is the radar's line-of-sight range, V is the flight direction of the satellite platform, and p and q are constants.

[0015] Preferably, the step of establishing a one-dimensional wave spectrum inversion algorithm for SAR in the ocean vortex region using the XGBoost machine learning model specifically includes: the one-dimensional wave spectrum inverted by this algorithm is determined by multiple variables: the VV polarized SAR intensity spectrum and three modulation transfer function parameters. For XGBoost-based spectral inversion, these parameters are used as input features, and the WW3 simulated spectrum is used as the target output, thus establishing a one-dimensional wave spectrum inversion algorithm for SAR in the ocean vortex region.

[0016] Preferably, the step of verifying the accuracy of the SAR one-dimensional wave spectrum inversion algorithm specifically includes: based on the established XGBoost algorithm, using SWIM data and HY-2 altimeter data to verify the SAR inversion spectrum results, thereby verifying the accuracy of the inversion algorithm.

[0017] Compared with existing technologies, this invention provides a one-dimensional wave spectrum inversion method for ocean eddy regions based on synthetic aperture radar. It uses WW3 wave model data as training to construct a one-dimensional wave spectrum inversion algorithm for ocean eddy regions, and verifies the accuracy of the results using SWIM data and HY-2 altimeter data. This invention breaks through the accuracy limitations of traditional wave theory inversion in eddy regions and improves the accuracy of one-dimensional wave spectrum inversion. Attached Figure Description

[0018] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0019] Figure 1 This is a flowchart of the one-dimensional wave spectrum inversion method for ocean eddy regions based on synthetic aperture radar according to the present invention.

[0020] Figure 2 This is a daily sea level anomaly (SLA) map from the HY-2 satellite altimeter on July 21, 2024.

[0021] Figure 3 Distribution map of more than 2,500 Sentinel-1 dual polarization images of the vortex region from 2022 to 2024;

[0022] Figure 4 This is a salient wave height (SWH) map measured by the Ocean-2 (HY-2) satellite altimeter on September 1, 2023.

[0023] Figure 5a Two-dimensional spectra were measured at SWIM on July 10, 2023, at [134.31°E, 13.81°N].

[0024] Figure 5b A comparison chart of the simulated spectrum of WW3 with the measurement results of HY-2 and SWIM;

[0025] Figure 6 Performance graph of XGBoost for training SAR spectral inversion;

[0026] Figure 7 The wave spectrum is obtained by SAR inversion at 17:33 UTC on March 6, 2023, and the wave spectrum is measured by SWIM at 18:37 UTC.

[0027] Figure 8 A comparison chart of SAR inversion results and SWIM data;

[0028] Figure 9 This is a comparison chart of the SAR inversion results and the HY-2 altimeter data. Detailed Implementation

[0029] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the scope of protection of the present invention.

[0030] Specifically, this invention provides a method for inverting one-dimensional wave spectra in ocean eddy regions based on synthetic aperture radar, such as... Figure 1 As shown, the method includes the following steps:

[0031] S1: Use HY-2 fusion data to identify ocean eddies and acquire Sentinel-1 dual-polarization SAR images covering the eddy region to obtain WAVEWATCH-Ⅲ wave model data;

[0032] Specifically, HY-2 fused SLA data from 2022 to 2024 were collected using vector geometry algorithms, and more than 2,500 Sentinel-1 dual-polarization images covering the vortex region were acquired to obtain data from the WAVEWATCH-III mode at the corresponding time and location. Figure 2 This is a daily sea level anomaly (SLA) map from the HY-2 satellite altimeter on July 21, 2024. The red and black circles represent cyclonic vortices and anticyclonic vortices, respectively. Figure 3 This is a distribution map of more than 2,500 Sentinel-1 dual-polarization images of the eddy region from 2022 to 2024. Spectral data from the Surface Wave Survey and Monitoring (SWIM) mission aboard the China-France Oceanographic Satellite (CFOSAT) and altimeter data from the Haiyang-2 (HY-2) satellite were used to validate the SAR inversion results. Figure 4 The effective wave height (SWH) was measured by the Ocean-2 (HY-2) satellite altimeter on September 1, 2023. Figure 5a Two-dimensional wave spectra were measured for SWIM located at [134.31°E, 13.81°N] on July 10, 2023. Figure 5b The results of the WW3 simulated spectrum comparison with the HY-2 and SWIM measurements show that the root mean square error (RMSE) is 0.32 meters, the correlation (r) is 0.95, and the scatter index (SI) is 0.18.

[0033] S2: Improved fluid dynamics modulation transfer function, introducing vortex shear flow influence factor;

[0034] Specifically, since the incident angle of space SAR radar is between 20° and 60°, Bragg scattering of microwaves at the microscale of the sea surface dominates the SAR imaging mechanism. The main modulation methods affecting wave imaging include tilting, hydrodynamic modulation, and velocity focusing. Currently, the divergence of long waves on the sea surface is adjusted through wave-current interaction to change the SAR sea surface roughness. In the context of strong shear currents caused by ocean eddies, the effect of hydrodynamic modulation is more significant, and the theoretical derivation of the hydrodynamic modulation transfer function (MTF) deviates from the actual situation.

[0035] Existing tilt modulation MTFs:

[0036] Speed-gathering MTFs:

[0037]

[0038] Hydrodynamic MTFs:

[0039]

[0040] The theoretical derivation of the hydrodynamic MTF deviates from the actual situation. Therefore, the formula for the hydrodynamic MTF has been modified as follows:

[0041] T hy =p0+p1ω+p2k+p3sinφ+(q0+q1ω+q2k+q3sinφ)i

[0042] Where k represents the wave number, k l The wave number represents the radar direction, ω is the angular frequency of the sea surface microscale wave, θ is the radar incident angle, θ is the wave direction relative to the azimuth direction, R is the radar's line-of-sight range, and V is the flight direction of the satellite platform. p and q are constants, as shown in Table 1 below.

[0043]

[0044] Table 1

[0045] S3: An algorithm for inverting the SAR wave spectrum in the ocean eddy region was established using the XGBoost machine learning model.

[0046] Specifically, the one-dimensional wave spectrum retrieved by this algorithm is determined by several variables: the VV-polarized SAR intensity spectrum and three modulation transfer function (MTF) parameters. For XGBoost-based spectral inversion, these parameters serve as input features, and the WW3-simulated spectrum serves as the target output, thus establishing an algorithm for SAR one-dimensional wave spectrum inversion in ocean eddy regions. Figure 6 The training performance was demonstrated, achieving an RMSE of approximately 2.1 after 200 iterations.

[0047] S4: Verify the accuracy of the SAR one-dimensional wave spectrum inversion algorithm.

[0048] Specifically, based on the XGBoost algorithm established above, SWIM data and HY-2 altimeter data were used to verify the SAR inversion spectral results, thus validating the accuracy of the inversion algorithm. Figure 7 The wave spectrum is obtained from SAR inversion at 17:33 UTC on March 6, 2023, and the wave spectrum is measured by SWIM at 18:37 UTC. Figure 8 A comparison of the SAR inversion results with SWIM data shows that the RMSE is 0.43 meters and r is 0.79. Figure 9 A comparison of SWH data retrieved from SAR and data from the HY-2 altimeter shows an RMSE of 0.39 meters and an r of 0.90.

[0049] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A method for inverting one-dimensional wave spectra in ocean eddy regions based on synthetic aperture radar, characterized in that, The method includes the following steps: Ocean eddies were identified using HY-2 fusion data, and Sentinel-1 dual-polarization SAR images covering the eddy region were acquired to obtain WAVEWATCH-Ⅲ wave model data. Improve the fluid dynamics modulation transfer function and introduce the vortex shear flow influence factor; An algorithm for inverting the SAR wave spectrum in the ocean eddy region was established using the XGBoost machine learning model. Verify the accuracy of the SAR one-dimensional wave spectrum inversion algorithm.

2. The method for one-dimensional wave spectrum inversion in ocean eddy regions based on synthetic aperture radar according to claim 1, characterized in that, The steps of identifying ocean eddies using HY-2 fusion data and acquiring Sentinel-1 dual-polarization SAR images covering the eddy region to obtain WAVEWATCH-III wave model data specifically include: collecting HY-2 fusion SLA data from 2022 to 2024 processed using vector geometry algorithms, acquiring more than 2,500 Sentinel-1 dual-polarization images covering the eddy region, and obtaining data from the WAVEWATCH-III model at the corresponding time and location.

3. The method for one-dimensional wave spectrum inversion in ocean eddy regions based on synthetic aperture radar according to claim 1, characterized in that, The step of improving the fluid dynamics modulation transfer function and introducing the vortex shear flow influence factor specifically includes: improving the fluid dynamics modulation transfer function to make it more suitable for the vortex region, wherein the formula for the improved fluid dynamics modulation transfer function is: T hy =p0+p1ω+p2k+p3sinφ+(q0+q1ω+q2k+q3sinφ)i Where k represents the wave number, k l The wave number represents the direction of the radar, ω is the angular frequency of the microscale waves on the sea surface, θ is the radar incident angle, θ is the wave direction relative to the azimuth direction, R is the radar's line-of-sight range, V is the flight direction of the satellite platform, and p and q are constants.

4. The method for inverting one-dimensional wave spectra in ocean eddy regions based on synthetic aperture radar according to claim 1, characterized in that, The steps for establishing a one-dimensional wave spectrum inversion algorithm for SAR in the ocean vortex region using the XGBoost machine learning model specifically include: the one-dimensional wave spectrum inverted by this algorithm is determined by multiple variables: the VV polarized SAR intensity spectrum and three modulation transfer function parameters. For XGBoost-based spectral inversion, these parameters are used as input features, and the WW3 simulated spectrum is used as the target output, thus establishing a one-dimensional wave spectrum inversion algorithm for SAR in the ocean vortex region.

5. The method for one-dimensional wave spectrum inversion in ocean eddy regions based on synthetic aperture radar according to claim 1, characterized in that, The steps for verifying the accuracy of the SAR one-dimensional wave spectrum inversion algorithm specifically include: based on the established XGBoost algorithm, using SWIM data and HY-2 altimeter data to verify the SAR inversion spectrum results, thereby verifying the accuracy of the inversion algorithm.

Images