CN-121980906-A - Method for reconstructing temperature field of ocean subsurface by combining space-based microwaves and lasers

Abstract

A method for reconstructing a temperature field of a marine subsurface layer by combining space-based microwaves and lasers solves the technical problems of lack of direct participation of subsurface parameters and low reconstruction accuracy in the acquisition of a current marine subsurface layer temperature profile, and the method for reconstructing the temperature field of the marine subsurface layer comprises the following steps of 1 determining an error of a laser radar inversion marine subsurface layer temperature; step 2, generating input and output data sets, step 3, constructing a subsurface temperature profile reconstruction model, training, and step 4, evaluating the efficiency of the trained model. The invention obviously improves the reconstruction precision of the temperature field and breaks through the technical limitation of the traditional parameter model depending on the pure sea surface.

Inventors

• TIAN XIAOQING
• LI RUIFENG
• YUAN ZHI

Assignees

• 中国空间技术研究院

Dates

Publication Date

20260505

Application Date

20251212

Claims (9)

1. 1. A method for reconstructing a temperature field of a marine subsurface fused with space-based microwaves and laser is characterized by comprising the following steps: step 1, determining an error of the temperature of a laser radar inversion ocean subsurface; step 2, generating input and output data sets; Step3, constructing a subsurface temperature profile reconstruction model and training; and 4, performing efficiency evaluation on the trained model.
2. 2. The method for reconstructing a temperature field of a marine subsurface layer according to claim 1, wherein in step 1, a laser detection method based on brillouin scattering is used for inverting the water temperature of the marine subsurface layer, and an inversion error of the subsurface layer temperature is shifted by brillouin frequency Brillouin linewidth Is determined by the measurement accuracy of (a).
3. 3. The method for reconstructing a temperature field of a marine subsurface according to claim 2, wherein in step 1, the theoretical error of the laser radar inversion of the temperature of the marine subsurface is as follows: wherein T represents temperature and S represents salinity.
4. 4. The method for reconstructing the ocean subsurface temperature field according to claim 1, wherein in the step 2, input data is in two configurations, wherein the subsurface water temperature is added in the group 1, and comprises the sea surface temperature, the sea surface height and the 3m subsurface water temperature simulated by adding noise; outputting data, namely, a true value of the subsurface temperature profile of 0-200 m.
5. 5. The method for reconstructing a marine subsurface temperature field according to claim 1, wherein the constructing a subsurface temperature profile reconstruction model and training in step 3 specifically comprises: selecting a Random Forest (Random Forest) regression model to reconstruct a subsurface temperature profile; inputting variables, namely, space-based observation data and subsurface water temperature simulation data; outputting a target of 0-200m subsurface temperature profile; and (3) a true value reference, namely ARGO observation sections are used as output references for model training and verification.
6. 6. The method of claim 5, wherein the performance assessment includes precision index calculation, contribution analysis, noise margin verification.
7. 7. The method of reconstructing a temperature field of a marine subsurface according to claim 6, wherein the accuracy index calculation is to calculate a root mean square error of the reconstruction result based on ARGO measured profiles.
8. 8. The method of reconstructing a marine subsurface temperature field as recited in claim 6 wherein the contribution analysis is comparing RMSE changes from different experimental groups to quantify the contribution of subsurface water temperature to accuracy improvement.
9. 9. The method for reconstructing a temperature field of a marine subsurface according to claim 6, wherein the noise margin verification is to analyze the influence of different noise levels on the accuracy and evaluate the practical effectiveness of the laser radar shallow water temperature.

Description

Method for reconstructing temperature field of ocean subsurface by combining space-based microwaves and lasers Technical Field The invention relates to a method for reconstructing a temperature field of a marine subsurface layer by combining space-based microwaves and lasers, belongs to the technical field of marine subsurface layer temperature monitoring, and can be applied to marine environment monitoring, disaster early warning, climate research and the like. Background At present, the reconstruction of the ocean subsurface temperature profile (namely the underwater temperature profile) mainly depends on a space-based remote sensing technology to acquire sea surface parameters and a model method based on the sea surface parameters. The core steps of the prior art scheme are as follows: And acquiring sea surface data by using space-based microwave remote sensing. The surface of the ocean is observed from space using sensors such as a microwave radiometer and a radar altimeter mounted on satellites. And obtaining parameters such as Sea Surface Temperature (SST), sea Surface Height (SSH) and the like by inversion through measuring microwave signals emitted by the sea surface. The space-based observation has the advantages of wide coverage range and strong periodicity, but can only directly detect sea surface parameters, and cannot acquire subsurface (underwater) temperature information. And constructing a model based on sea surface parameters. An empirical or physical model is established using statistical correlations between sea surface measurements (e.g., sea surface temperature and altitude) and subsurface temperatures. Common methods include using regression analysis, machine learning algorithms (such as neural networks) or data assimilation techniques, taking sea surface parameters as inputs, and outputting the model as a subsurface temperature profile or three-dimensional temperature field, but inputting data is limited to sea surface observations. Although the above prior art techniques are widely used in marine monitoring, they suffer from the following key drawbacks: The existing method completely depends on sea surface parameters (such as sea surface temperature and sea height) as input, and cannot directly acquire or utilize temperature data of the subsurface layer (shallow underwater). This results in a reconstruction model that ignores the varying nature of the subsurface itself, based only on the indirect correlation of the subsurface to the subsurface, adding a source of error. The reconstruction accuracy is limited, because the model input is limited to sea surface observation, the fine structure (such as thermocline or small scale change) of the subsurface layer cannot be captured, and the reconstruction accuracy of the temperature profile and the underwater three-dimensional temperature field is low. Actual measurement verification shows that the error of the method is obviously increased under a complex marine environment (such as vortex or frontal area), and the high-precision application requirement cannot be met. The real-time performance of the reconstruction data is insufficient, and the model based on the sea surface parameters is highly dependent on statistical correlation assumption and is easy to fail under the condition of sea dynamic change (such as seasonal fluctuation or extreme events). Meanwhile, the model parameter calibration depends on limited measured data, so that generalization capability is poor, and the model parameter calibration is difficult to adapt to complex conditions of different sea areas worldwide. Disclosure of Invention The invention aims to solve the technical problems that the defects of the prior art are overcome, the underwater shallow surface water temperature directly measured by a laser radar is added into a reconstruction algorithm, and the technical problems of lack of direct participation of subsurface parameters in the acquisition of the current ocean subsurface temperature profile and low reconstruction accuracy are solved. The invention aims at realizing the following technical scheme: a method for reconstructing a temperature field of a marine subsurface fused by space-based microwaves and laser comprises the following steps: step 1, determining an error of the temperature of the inversion ocean subsurface of the laser radar The invention adopts a laser detection method (wavelength 532 nm) based on Brillouin scattering, and is used for inverting the water temperature of the ocean subsurface. Compared with raman scattering, brillouin scattering has two major core advantages: The anti-interference performance is strong, the scattering wavelength is close to the incident light, and the light is not easily influenced by sunlight and fluorescence; the detection efficiency is high, and the reflection section interacted with the water body is 1-2 orders of magnitude larger than the Raman scattering. By measuring the brillouin scattering spectrum,