Method for inverting seawater particulate matter backscattering coefficient profile through active and passive fusion

Abstract

The invention discloses a method for inverting a seawater particulate matter backscattering coefficient profile through active and passive fusion. The method comprises the steps of S1, obtaining a seawater particulate matter laser radar ratio based on a passive optical satellite; s2, eliminating land noise photons and atmospheric noise photons from the photon elevation and latitude and longitude data acquired by the active detection laser radar satellite; s3, after the land noise photons and the atmospheric noise photons are removed, aligning the sea surface, and accumulating the photons according to the elevation after alignment to obtain a detected seawater signal profile; s4, performing deconvolution and refractive index correction on the detected seawater signal profile to obtain a real ocean laser radar signal profile; and S5, according to the laser radar ratio and the real ocean laser radar signal profile, using a Fernald method to iteratively invert the seawater particulate matter backscattering coefficient profile. According to the method, high-precision ocean bbp profile inversion can be realized, and the method is of great significance to global ocean optical property vertical distribution.

Description

Technical Field

[0001] The present invention belongs to the technical field of marine lidar remote sensing detection, and particularly relates to a method for actively and passively fusing and inverting the backscattering coefficient profile of seawater particles. Background Technique

[0002] The ocean is the largest active carbon reservoir on Earth, and the photosynthesis of phytoplankton in the upper ocean contributes approximately 50% of the global net primary productivity. The upper ocean is not only the main habitat of plankton but also the key interface for air-sea energy exchange. Optical properties are the external manifestations of seawater components and are of great significance for marine ecosystem and biochemistry research. Detecting the vertical distribution of upper-layer optical properties can drive multiple marine scientific research, such as phytoplankton layer detection, vertical migration of marine species, and harmful algal bloom prediction.

[0003] In-situ, shipborne, and airborne instruments can provide accurate optical property detection, but they are limited by space, time, and energy consumption constraints and cannot achieve global detection. The emergence of satellite technology has provided an effective way for global detection of marine optical properties. Passive ocean color remote sensing satellites have been developed for more than 40 years. Satellites such as MODIS (Moderate Resolution Imaging Spectroradiometer) and OLCI (Ocean and Land Color Imager) have achieved high-precision inversion of optical parameters, providing valuable data support for global marine ecological research. However, they cannot achieve vertical optical property detection.

[0004] Spaceborne lidar can effectively make up for the limitations of passive remote sensing. The ICESat-2 (Ice, Cloud, and Land Elevation Satellite 2) mission was launched in 2018 and carried the photon-counting lidar ATLAS (Advanced Topographic Laser Altimeter System) at 532 nm, with sub-meter vertical resolution, which is the only spaceborne instrument that can achieve continuous water body profile detection. Currently, many scholars have explored the vertical K d inversion. Chinese patents with publication numbers CN119126073A, CN119355685A, etc. have disclosed various K d inversion methods and achieved relatively high precision. However, regarding b bp , the current method is to use the wind speed as auxiliary data to calibrate the lidar constant disclosed in the Chinese patent with publication number CN114089366A, and then substitute it into the lidar equation to invert b bp . The accuracy of this method is severely affected by the accuracy of the wind speed, and the backscattering conversion factor varies greatly in different regions.

[0005] Therefore, there is an urgent need for a method that is not affected by the detection accuracy of wind speed and can effectively solve the problems caused by the value problem and the wind speed accuracy problem for ICESat-2 b bpMethod for reducing the accuracy of profile inversion. Summary of the Invention

[0006] The present invention provides a method for actively and passively fused inversion of the backscattering coefficient profile of seawater particulate matter, which can achieve high-precision ocean b bp profile inversion, which is of great significance for the vertical distribution of global ocean optical properties.

[0007] A method for actively and passively fused inversion of the backscattering coefficient profile of seawater particulate matter, comprising:

[0008] S1: Obtaining the lidar ratio of seawater particulate matter based on passive optical satellites;

[0009] S2: In the photon elevation and longitude and latitude data obtained by the active detection lidar satellite, removing land noise photons and atmospheric noise photons;

[0010] S3: After removing land noise photons and atmospheric noise photons, aligning the sea surface, and accumulating the photons according to the aligned elevation to obtain the detected seawater signal profile;

[0011] S4: Performing deconvolution and refractive index correction on the detected seawater signal profile to obtain the true ocean lidar signal profile;

[0012] S5: According to the lidar ratio and the true ocean lidar signal profile, using the Fernald method to iteratively invert the backscattering coefficient profile of seawater particulate matter.

[0013] The specific process of step S1 is as follows:

[0014] According to the passive optical satellite, the global monthly average 531nm absorption coefficient a(531) and the total backscattering coefficient b b (531) with an observation resolution of 4km are obtained, and the global total diffuse attenuation coefficient K d (531) is further calculated;

[0015] By subtracting the diffuse attenuation coefficient K of water molecules d (531) from the global total diffuse attenuation coefficient K dw (531), the total particulate diffuse attenuation coefficient K of the ocean is obtained dp (531);

[0016] By subtracting the backscattering coefficient b of seawater b (531) from the total backscattering coefficient b bw (531), the total particulate backscattering coefficient b of the ocean is obtained bp (531);

[0017] Using the formula β p π= b bp (531) / 2πχ p π Calculate the backscattering function β of particulate matter for passive optical satellites pp π , and further use the formula R p = K dp (531) / β pp π Calculate the lidar ratio R of particulate matter p , where χ p π represents the backscattering conversion factor.

[0018] Furthermore, the formula for calculating the global total diffuse attenuation coefficient K d (531) is:

[0019] K d (531) = a(531) + 4.18(1 - 0.52e -10.8a(531) ) × b b (531).

[0020] Furthermore, the diffuse attenuation coefficient K of water molecules dw (531) = 0.0452m -1 , and the seawater backscattering coefficient b bw (531) = 8.5 × 10 -4 m -1 .

[0021] In step S2, the specific process of removing land noise photons is as follows:

[0022] Adopt a segmented sampling and judgment method, segment by 10 consecutive 3 photons. If the surface types of the first photon and the last photon in each segment are the same, it is considered that the surface types of all photons in this segment are the same; find the photon segments with inconsistent surface types at the beginning and end and further refine the segmentation until all critical photons with inconsistent surface types before and after are found. Finally, find all photon segments with land attributes according to the critical photons and remove them.

[0023] In step S2, the specific process of removing atmospheric noise photons is as follows:

[0024] Divide all photons according to the time accuracy Δt, and detect whether there are atmospheric noise photons within Δt. If there are, delete all atmospheric noise photons within this Δt;

[0025] Among them, the allowable height range of ocean photons is plus or minus 50m from the mean sea level. Photons outside this range are atmospheric noise photons.

[0026] The specific process of step S3 is:

[0027] The photons after removing the land noise photons and atmospheric noise photons are divided into several segments according to the along-track distance. For each segment of photons, the photon height distribution is statistically analyzed at a vertical interval of 0.15 m to generate a histogram.

[0028] Use a Gaussian function to fit the histogram, calculate the average μ and standard deviation σ of the distribution Gaussian parameters of the sea surface photons, and determine the sea surface photon range as μ ± 3σ.

[0029] For the sea surface photons within the determined range, the elevation average value is calculated every 7 m as the sea surface elevation. The photons less than μ - 3σ are water body photons. Subtract the corresponding sea surface elevation from the water body photons to achieve sea surface alignment; accumulate the photons according to the aligned elevation to obtain the detected sea water signal profile.

[0030] The specific process of step S4 is as follows:

[0031] The detected sea water signal profile is the convolution of the true ocean lidar signal profile and the system impulse response curve, and its matrix form is expressed as:

[0032]

[0033] In the formula, z i (i = 1, 2,..., n) represents different water depth positions in the ocean lidar signal profile, S(z) is the detected sea water signal profile, S c (z) is the true ocean lidar signal profile, and F(z) is the system impulse response curve;

[0034] First, according to S c (z) = S desert (z) -1 S(z) performs deconvolution to remove the influence of the post-pulse, and then performs water depth refraction correction to obtain the true ocean lidar signal profile;

[0035] Among them, S desert (z) represents the detected desert surface echo signal profile; during water depth refraction correction, due to the influence of the sea-air interface and the near-vertical incidence of the active detection lidar satellite, the actual depth of the water body photons relative to the sea surface is 0.75 times the water depth recorded by the lidar.

[0036] The specific process of step S5 is as follows:

[0037] Write the equation of the true ocean lidar signal profile as

[0038]

[0039] Among them, C is the lidar constant; in step S1, the lidar ratio R of global sea water particles has been obtainedp , the lidar of water molecules is better than Therefore, rewrite the detected seawater signal profile equation as:

[0040]

[0041] In the formula, z c is the maximum detection depth, and the initial value is obtained from the true ocean lidar signal profile by the slope method ; K d The profile is continuously iterated by the Fernald method, and the depth is obtained by expanding from bottom to top; then according to the particulate matter backscattering function profile β p π (z) is obtained; the seawater particulate matter backscattering coefficient b bp (z) is calculated by 2πβ p π (z)χ p π .

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] 1. The b bp profile inverted by the present invention is not limited by the wind speed, and there is an obvious improvement in accuracy in the low wind speed region compared with the existing method.

[0044] 2. The b bp profile implemented by the present invention is independent of the value of χ p π , and is not subject to regional limitations. Because β p π is divided by χ p π when obtained in step S1, and the parameter is multiplied when calculating b bp in step S5, because its value has no influence on the result. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 is a flowchart of a method for actively and passively fused inversion of seawater particulate matter backscattering coefficient profile according to the present invention.

[0046] Figure 2 is a comparison of the b bp profile (left) inverted by the present invention and the existing method (right) in the Mediterranean and Black Sea regions. DETAILED DESCRIPTION OF THE INVENTION

[0047] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be noted that the following embodiments are intended to facilitate the understanding of the present invention, but do not impose any limitations on it.

[0048] In the embodiment of the present invention, the ICESat-2 ATL03 data in the Mediterranean Sea and Red Sea regions from January 2020 to December 2023 is taken as an example (30°N - 47°N, 5°W - 42°E).

[0049] As Figure 1 shown, a method for fusing active and passive inversion of the backscattering coefficient profile of seawater particulate matter includes the following steps:

[0050] Step S1: Obtain the lidar ratio of seawater particulate matter based on the passive optical satellite MODIS.

[0051] According to the global monthly average 531nm absorption coefficient a(531) and total backscattering coefficient b b (531) provided by MODIS with an observation resolution of 4km, the global total diffuse attenuation coefficient is calculated, and the formula is K d (531) = a(531) + 4.18(1 - 0.52e -10.8a(531) ) × b b (531).

[0052] The total diffuse attenuation coefficient K dp (531) of marine total particulate matter can be obtained by subtracting the diffuse attenuation coefficient K d (531) of water molecules. K dw (531) = 0.0452m -1 , the total backscattering coefficient b bp (531) of marine total particulate matter can be obtained by subtracting the backscattering coefficient b b (531) of seawater. b bw (531) = 8.5×10 -4 m -1 is obtained, and the particulate backscattering function of the passive optical satellite can be calculated by β pp π = b bp (531) / 2πχ p π . After that, the lidar ratio R p can be calculated as K dp (531) / β pp π .

[0053] Step S2: Remove land and atmosphere photons from the lidar satellite ICESat-2.

[0054] In the ICESat-2 ATL03 file, the elevation, longitude, and latitude information of the returned photons is recorded, but the surface type tags carried by it are not accurate. Therefore, it is necessary to first remove the photons scattered from the land according to the longitude, latitude information of the photons and the land boundary longitude and latitude dataset. Since the number of photons is very large (the number of photons in a strong light beam reaches 10 7 magnitude), the method of traversing photons to determine whether they are land photons is time-consuming. Considering the large scales of land and ocean, there must be a continuous and relatively long section of photons with the same surface type. The method of segmented sampling and judgment can be adopted. If the surface types of the first photon and the last photon in each segment of photons (10 consecutive photons) are the same, it is considered that the surface types of all photons in this segment are the same. Find the photon segments where the surface types of the first and last photons are inconsistent and further refine the segmentation (10 3 →10 3 →10 2 →1), until all critical photons with inconsistent surface types before and after are found. Finally, find all photon segments with land attributes according to the critical photons and remove them.

[0055] The method of removing atmospheric noise photons is to divide all photons according to the time precision Δt, detect whether there are atmospheric noise photons within Δt, and if so, delete all photons within this Δt. Since the land noise photons have been removed previously, for the sources of most of the remaining photons, the allowable range of the height of ocean photons is set to be plus or minus 50 m from the mean sea level. Photons outside this range are atmospheric noise photons.

[0056] Step S3: After removing land noise photons and atmospheric noise photons, align the sea surface, accumulate the photons according to the aligned elevation, and obtain the detected sea water signal profile.

[0057] After removing land and atmospheric photons, the sea surface needs to be aligned. Divide the photons into several segments according to the along-track distance. For each segment of photons, generate a histogram by statistically analyzing the photon height distribution at a vertical interval of 0.15 m. Fit the histogram with a Gaussian function, calculate the Gaussian parameters of the distribution of sea surface photons, the mean μ and the standard deviation σ, and determine the range of sea surface photons as μ±3σ. Calculate the elevation average value of sea surface photons within the determined range every 7 m as the sea surface elevation. Photons less than μ - 3σ are water body photons. Subtract the corresponding sea surface elevation from the water body photons to achieve sea surface alignment. Accumulate the photons according to the aligned elevation to obtain the sea water signal profile.

[0058] Step S4: Deconvolution of the ICESa-2 signal profile and refractive index correction.

[0059] The detected sea water signal profile is the convolution of the true ocean lidar signal profile and the system impulse response curve, and its matrix form is expressed as:

[0060]

[0061] In the formula, z i (i = 1, 2, ..., n) represents different water depth positions in the marine lidar signal profile. S(z) is the detected seawater signal profile, and S c (z) is the true marine lidar signal profile, and F(z) is the system impulse response curve;

[0062] First, according to S c (z) = S desert (z) -1 S(z) performs deconvolution to remove the influence of the post-pulse, and then performs water depth refraction correction to obtain the true marine lidar signal profile;

[0063] Among them, S desert (z) represents the detected desert surface echo signal profile; during water depth refraction correction, due to the influence of the sea-air interface and the near-vertical incidence of the active detection lidar satellite, the actual depth of water photons relative to the sea surface is 0.75 times the water depth recorded by the lidar.

[0064] Step S5: Perform deconvolution and refractive index correction on the detected seawater signal profile to obtain the true marine lidar signal profile.

[0065] The equation of the true marine lidar signal profile can be written as

[0066]

[0067] In the formula, C is the lidar constant. In step S1, the lidar ratio R of global seawater particulate matter has been obtained p , and the lidar ratio of water molecules Therefore, this equation can be written as:

[0068]

[0069] In the formula, z c is the maximum detection depth, and the initial value can be obtained from the true marine lidar signal profile by the slope method , and the profile can be continuously iterated by the Fernald method, and the depth is extended from bottom to top. Then, according to d the particulate backscattering function profile β can be obtained. The seawater particulate backscattering coefficient b p π (z) can be obtained from 2πβ bp (z)χ p π (z)χ p πCalculated. Since β is obtained in step S1 p π and divided by χ when p π this parameter is multiplied in this step, because its value has no influence on the result.

[0070] For b obtained from step S5 bp The contour line and the b of BGC-Argo bp Are compared and verified in the Mediterranean Sea and Black Sea regions. The matching spatial interval is set to 0.3°, and the time interval is set to 5 days. The results are as Figure 2 Shown, proving that this method has a greater improvement in accuracy in the nearshore area compared with the traditional wind speed calibration method.

[0071] The above-described embodiments have detailed the technical solutions and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the present invention. Any modification, supplement, and equivalent replacement made within the scope of the principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for actively and passively fusing and inverting the seawater particulate backscattering coefficient profile, characterized in that, Including: S1: Obtain the lidar ratio of seawater particles based on passive optical satellites; S2: In the photon elevation and longitude-latitude data obtained by the active detection lidar satellite, eliminate the land noise photons and atmospheric noise photons; S3: After eliminating the land noise photons and atmospheric noise photons, align the sea surface, accumulate the photons according to the aligned elevation, and obtain the detected seawater signal profile; S4: Perform deconvolution refractive index correction on the detected seawater signal profile to obtain the true ocean lidar signal profile; S5: According to the lidar ratio and the true ocean lidar signal profile, use the Fernald method to iteratively invert the seawater particle backscattering coefficient profile.

2. The method for actively and passively fusing and inverting the seawater particulate matter backscattering coefficient profile according to claim 1, wherein The specific process of step S1 is as follows: Based on passive optical satellites, the global monthly average 531 nm absorption coefficient a(531) and total backscattering coefficient b b (531) with an observation resolution of 4 km are obtained, and the global total diffuse attenuation coefficient K d (531) is further calculated; Through the global total diffuse attenuation coefficient K d (531) Subtract the diffuse attenuation coefficient K dw (531) of water molecules to obtain the total particulate matter diffuse attenuation coefficient K dp (531); Through the total backscattering coefficient b b (531) subtract the seawater backscattering coefficient b bw (531), to obtain the total marine particulate backscattering coefficient b bp (531); Using the formula β pp π = b bp (531) / 2πχ p π calculate the backscattering function β of particulate matter for passive optical satellites pp π , and further use the formula R p = K dp (531) / β pp π to calculate the lidar ratio R of particulate matter p , where χ p π represents the backscattering conversion factor.

3. The method for inversely calculating the seawater particulate matter backscattering coefficient profile by active-passive fusion according to claim 2, characterized in that Calculate the global total diffuse attenuation coefficient K d (531) The formula is as follows: K d (531) = a(531) + 4.18(1 - 0.52e -10.8a(531) ) × b b (531).

4. The method for actively and passively integrated inversion of the seawater particulate backscattering coefficient profile according to claim 2, wherein Diffuse attenuation coefficient K of water molecules dw (531) = 0.0452 m -1 , Backscattering coefficient b of seawater bw (531) = 8.5×10 - 4 m -1 .

5. The method for actively and passively integrated inversion of the seawater particulate matter backscattering coefficient profile according to claim 1, wherein In step S2, the specific process of eliminating land noise photons is as follows: Adopt the method of segmented sampling and judgment, with 10 consecutive 3 photons for segmentation. If the surface types of the first photon and the last photon in each segment of photons are the same, it is considered that the surface types of all photons in this segment are the same; find the photon segments with inconsistent surface types at the beginning and end and further refine the segmentation until all critical photons with inconsistent surface types before and after are found. Finally, find all photon segments with land attributes based on the critical photons and eliminate them.

6. The method for actively and passively integrated inversion of the backscattering coefficient profile of seawater particulate matter according to claim 1, wherein In step S2, the specific process of eliminating atmospheric noise photons is as follows: Divide all photons according to the time precision Δt, detect whether there are atmospheric noise photons within Δt, and if so, delete all atmospheric noise photons within that Δt; Among them, the allowable height range of ocean photons is plus or minus 50m from the mean sea level, and photons outside this range are atmospheric noise photons.

7. The method for actively and passively integrated inversion of the seawater particulate matter backscattering coefficient profile according to claim 1, characterized in that The specific process of step S3 is as follows: Divide the photons after eliminating land noise photons and atmospheric noise photons into several segments according to the along-track distance. For each segment of photons, generate a histogram by statistically counting the photon height distribution at a vertical interval of 0.15 meters; Fit the histogram with a Gaussian function, calculate the average μ and standard σ of the distribution Gaussian parameters of the sea surface photons, and determine the sea surface photon range as μ ± 3σ; For the sea surface photons within the determined range, calculate the elevation average every 7m as the sea surface elevation. Photons less than μ - 3σ are water body photons. Subtract the corresponding sea surface elevation from the water body photons to achieve sea surface alignment; Accumulate the photons according to the aligned elevation to obtain the detected seawater signal profile.

8. The method for actively and passively integrated inversion of the backscattering coefficient profile of seawater particulate matter according to claim 1, wherein The specific process of step S4 is as follows: The detected seawater signal profile is the convolution of the true ocean lidar signal profile and the system impulse response curve, and its matrix form is expressed as: where z i (i = 1, 2, ..., n) represents different water depth positions in the ocean lidar signal profile, S(z) is the detected seawater signal profile, S c (z) is the true ocean lidar signal profile, and F(z) is the system impulse response curve; First, according to S c (z) = S desert (z) -1 Perform deconvolution to remove the influence of the post-pulse, and then perform water depth refraction correction to obtain the true ocean lidar signal profile; Among them, S desert (z) represents the detected echo signal profile of the desert surface; during water depth refraction correction, due to the influence of the sea-air interface and the near-vertical incidence of the active detection lidar satellite, the actual depth of water photons relative to the sea surface is 0.75 times the water depth recorded by the lidar.

9. The method for actively and passively fusing and inverting the backscattering coefficient profile of seawater particulate matter according to claim 1, wherein The specific process of step S5 is as follows: Write the true ocean lidar signal profile as where C is the lidar constant; in step S1, the lidar ratio R of global seawater particles has been obtained p , and the lidar ratio of water molecules Therefore, the detected seawater signal profile equation is rewritten as: In the formula, z c is the maximum detection depth, and its initial value is obtained from the true ocean lidar signal profile by the slope method; K d The profile is continuously iterated by the Fernald method, and the depth is obtained by expanding from bottom to top; then according to the particulate matter backscattering function profile β p π (z); the seawater particulate backscattering coefficient b bp (z) is calculated by 2πβ p π (z)χ p π is obtained.

Images