An antarctic ice sheet surface liquid water content inversion method based on genetic algorithm optimization of radiation transmission model

Abstract

The present application relates to a kind of Antarctic ice sheet surface liquid water content inversion method based on genetic algorithm optimization radiation transfer model, comprising: S1, the radiation brightness temperature data of L band is acquired, and the Antarctic ice sheet brightness temperature distribution map of different observation angles containing time information is constructed based on radiation brightness temperature data;S2, based on the Antarctic ice sheet brightness temperature distribution map, the observed brightness temperature is obtained, after the pre-processing of observed brightness temperature, the V polarization and H polarization brightness temperature data of the incident angle of 22.5 ° to 62.5 ° are screened and the daily average is calculated;S3, the multilevel radiation transfer model of Antarctic ice sheet is constructed, and the modeling of key parameters is carried out;S4, the brightness temperature of each layer structure is calculated and the Kirchhoff coefficient is obtained, and the radiation transfer brightness temperature of atmosphere bottom is calculated based on Kirchhoff coefficient and brightness temperature;S5, fitness function is constructed and the optimal snow density and liquid water content are solved using genetic algorithm.Compared with prior art, the present application overcomes the limitation that traditional empirical threshold method can only carry out binary detection of snow melting state, and realizes the quantification of liquid water content in snow.

Description

Technical Field

[0001] This invention relates to the field of remote sensing inversion technology, and in particular to a method for inverting the liquid water content on the surface of the Antarctic ice sheet based on a genetic algorithm-optimized radiative transfer model. Background Technology

[0002] Antarctica is one of the most important components of the global climate system, therefore, research on Antarctica plays a crucial role in understanding the global climate system. Current research on Antarctica largely focuses on the inversion of temperatures within the Antarctic ice sheet. For example, Chinese patent application CN118114555A provides a method for inverting Antarctic radiative brightness temperature and ice sheet internal temperature using a generalized simulated annealing optimized radiative transfer model. By employing a generalized simulated annealing optimized radiative transfer model and a GSA metaheuristic optimization algorithm, it solves the accuracy problem of monitoring Antarctic ice sheet internal temperature, improves the accuracy and reliability of ice sheet temperature inversion, and provides a more precise method for Antarctic ice sheet temperature analysis. However, research on the surface liquid water content of the Antarctic ice sheet is scarce. The surface liquid water content of the Antarctic ice sheet is related to its meltwater, which is mainly concentrated in the Antarctic summer (November to February of the following year), and typically occurs in relatively warm areas such as coastal ice shelves, glaciers, and the Antarctic Peninsula. As global temperatures rise, meltwater events in Antarctica are increasing, exacerbating the physical weakening of ice shelves and potentially leading to larger-scale glacial mass loss through accelerated ice flow and ice shelf collapse. Meltwater also forms melt pools, reducing the albedo of the ice sheet surface and increasing the amount of solar radiation absorbed, thus accelerating the melting process and creating a vicious cycle. Therefore, monitoring liquid water content is crucial for characterizing seasonal and interannual climate change in Antarctica and for assessing the future stability of the ice sheet.

[0003] Given the rarity of in-situ measurements related to meltwater or liquid water, the use of remote sensing technology to detect the melting process of polar ice and snow has become a core topic in polar research, with the advent of spaceborne microwave radiometry. Water has a higher absorption rate in the microwave range than dry snow, therefore wet snow containing liquid water causes drastic changes in microwave signals. Various algorithms have been developed for active microwave remote sensing sensors, such as the European Space Agency's Sentinel-1 and Canada's RADARSAT-2, as well as passive microwave remote sensing sensors, such as AMSRE and SSM / I. These algorithms have been widely used in detecting ice and snow melt in Greenland and Antarctica. Early microwave remote sensing technologies mainly relied on higher frequency bands, such as the Ku band (12-18 GHz), sensors, which, due to their shallow penetration depth, typically only identified surface melting conditions. To overcome this limitation, the use of lower frequency band sensors, such as the L-band and C-band, has been proposed to obtain deeper information on the physical properties of ice and snow. In particular, the L-band, due to its ability to penetrate ice and snow and provide physical information about the surface and subsurface layers, has been widely used for meltwater detection in Antarctica. Many detection algorithms generate binary indices about the surface melting state based on changes in microwave brightness temperature, used to reveal the spatiotemporal distribution of Antarctic ice and snow melt. Although most current passive microwave detection methods can effectively identify the surface melting state, these algorithms fail to quantify the liquid water content in snow cover. This limitation makes it difficult to accurately assess the surface energy balance in Antarctica and its potential impact on the global climate system.

[0004] Therefore, providing a method for quantifying the liquid water content on the surface of the Antarctic ice sheet is a technical problem that needs to be solved. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the existing technology and provide a method for inverting the liquid water content on the surface of the Antarctic ice sheet based on a genetic algorithm-optimized radiative transfer model.

[0006] The objective of this invention can be achieved through the following technical solutions: According to a first aspect of the present invention, a method for inverting the liquid water content on the surface of the Antarctic ice sheet based on a genetic algorithm-optimized radiative transfer model is provided, comprising: S1. Obtain L-band radiation brightness temperature data, and construct a brightness temperature distribution map of the Antarctic ice sheet from different observation angles, including time information, based on the radiation brightness temperature data. S2. Based on the aforementioned Antarctic ice sheet brightness temperature distribution map, obtain the observed brightness temperature. After preprocessing the observed brightness temperature, select V-polarized and H-polarized brightness temperature data with incident angles from 22.5° to 62.5° and calculate the daily average value. S3. Construct a radiative transfer model of the Antarctic ice sheet, including the dry snow layer, wet snow layer, atmosphere, and basement, and model the key parameters. S4. Calculate the brightness temperature of each layer and obtain the Kirchhoff coefficient of each layer. Calculate the brightness temperature of atmospheric bottom radiation transmission based on the Kirchhoff coefficient and brightness temperature. S5. Based on the daily average value and the brightness temperature of atmospheric bottom radiative transfer, construct a fitness function, and use a genetic algorithm to solve for the optimal snow density and liquid water content based on the fitness function.

[0007] As a preferred technical solution, the preprocessing includes: Based on satellite orbital characteristics, brightness temperature data of V-polarization and H-polarization with an incident angle range of 22.5° to 62.5° in the L-band were screened and the standard deviation was calculated. Brightness temperature data with a standard deviation greater than or equal to the preset value were screened out. Based on the 5° observation angle interval of SMOS satellite observations, the aforementioned incident angle range is divided into 8 segments, resulting in... For each observation angle, the time interval of brightness temperature data is less than the preset time interval, and the gap is filled by linear interpolation. The gap is discarded if it is greater than or equal to the preset time interval. Based on satellite orbit characteristics and polar observation coverage limitations, the latitude limit was selected as 83°S, and the pixel values ​​of the region south of the latitude limit in the Antarctic ice sheet brightness temperature distribution map were set to empty. The brightness temperature from the top of the atmosphere to the bottom of the atmosphere is corrected using an empirical atmospheric correction method.

[0008] As a preferred technical solution, the empirical atmospheric correction method is as follows: ， in, This indicates the corrected brightness temperature value; Indicates the observed brightness temperature; Indicates the polarization mode, when When indicates vertical polarization, Time indicates horizontal polarization; Indicates the i-th observation angle; Indicates the equivalent atmospheric temperature; This indicates atmospheric transmittance.

[0009] As a preferred technical solution, the key parameters include the complex dielectric constant of the dry snow layer and the wet snow layer, the liquid water content, and the snow density. The method for modeling the key parameters is as follows: The wet snow layer includes dry snow matrix and liquid water. A model of the relationship between liquid water content and complex permittivity of wet snow layer is constructed using the Maxwell-Garnett mixing model. A model of snow density and complex permittivity of dry snow layer was constructed using empirical formulas.

[0010] As a preferred technical solution, the method for constructing a model of liquid water content and complex permittivity of wet snow layer is as follows: , in, This represents the liquid water content in the wet snow layer, and is an unknown value. This represents the complex permittivity of the dry snow matrix. The complex dielectric constant of liquid water is represented by a known value. The factor is represented, and its calculation method is as follows: , Indicates a constant value, and .

[0011] As a preferred technical solution, the method for constructing a model of snow density and complex permittivity of dry snow layer is as follows: , in, Indicates the complex permittivity of the dry snow layer; This represents the snow density, which is an unknown value. This represents the density ratio of accumulated snow to theoretical dry snow, and its calculation method is as follows: .

[0012] As a preferred technical solution, the method for calculating the brightness temperature of each layer structure includes: For the substrate, obtain the vertical polarization brightness temperature within a preset range of the Brewster angle, and calculate the average value of the vertical polarization brightness temperature, which is the brightness temperature of the substrate. For the aforementioned atmosphere, the sky brightness temperature is calculated as follows: ， in, Indicates the brightness temperature of the sky; Represents the equivalent atmospheric temperature, and , This indicates the air temperature at a height of 2 meters above the ground. Represents cosmic microwave background radiation; Indicates atmospheric optical thickness, and has , This indicates the height of the ice sheet surface relative to sea level; Indicates the first i One observation angle; For the dry snow layer, its brightness temperature is set to null based on the properties of the dry snow layer; For the aforementioned wet snow layer, its brightness temperature is set to a constant value based on the properties of the wet snow layer, i.e. .

[0013] As a preferred technical solution, the method for calculating the brightness temperature of atmospheric bottom radiation transmission is as follows: , in, Indicates the brightness temperature of radiative transfer at the bottom of the atmosphere; Denotes the Kirchhoff coefficient of the j-th layer, and when Time represents the basis, when When indicates the dry snow layer, Time indicates the wet snow layer; Indicates the brightness temperature of the j-th layer; The Kirchhoff coefficient representing the sky; This indicates that the sky is at an observation angle of 0°. Brightness temperature; Indicates snow density; Indicates the liquid water content; Indicates the polarization mode, when When indicates vertical polarization, The time indicates horizontal polarization.

[0014] As a preferred technical solution, the fitness function is characterized by: , in, Indicates snow density; Indicates the liquid water content; Indicates the number of observed angles; Indicates the observation angle as of The daily average brightness temperature observed by the satellite; This indicates the brightness temperature of radiation transmission at the bottom of the atmosphere.

[0015] As a preferred technical solution, the method for solving the optimal snow density and liquid water content includes: S51. Initialize the genetic algorithm parameters. Within the range of values, encode the snow density and liquid water content into two-dimensional vectors using real number encoding, and use them as initial candidate solution individuals to construct a population. S52. Calculate the fitness function value of each initial candidate solution individual, and take the candidate solution individual corresponding to the minimum fitness function value as the global optimal solution; Iterative execution steps S53~S5: S53. Select several candidate solution individuals from the population using the roulette wheel betting method, calculate the fitness function value of each candidate solution individual, put the candidate solution individual corresponding to the minimum fitness function value into the mating pool, and repeat the process until the mating pool is full. S54. Randomly select two parent candidate solution individuals from the mating pool and perform gene crossover and gene mutation operations to generate new candidate solution individuals; S55. For the population in the current iteration, an elite retention strategy is adopted to copy a preset number of candidate solutions in the population to the next generation, and form a new population with the new candidate solution individuals. S56. Calculate the fitness function of each candidate solution individual in the new population, and select the candidate solution individual corresponding to the minimum fitness function value to update the global optimal solution; S57. Determine whether the iteration termination condition is met. If yes, end the iteration and output the global optimal solution for the current iteration. Decode the solution to obtain the optimal snow density and liquid water content. Otherwise, return to execute S53.

[0016] Compared with existing technologies, this invention proposes a method for inverting the liquid water content on the surface of the Antarctic ice sheet using low-frequency passive microwave remote sensing data (L-band). Based on the characteristics of satellite orbits and the limitations of polar observation coverage, the observed brightness temperature values ​​of the Antarctic ice sheet surface are subjected to error removal, missing value filling, and brightness temperature correction to make the observed values ​​more accurate. A layered radiative transfer model is then constructed, modeling key parameters based on snow density and liquid water content. A fitness function is constructed based on the daily average values ​​of the observed values ​​and the radiative brightness temperature simulated by the radiative transfer model. A genetic algorithm is then used to optimize the snow density and liquid water content in the radiative transfer model, effectively reducing the error between simulated and observed data, improving the accuracy of liquid water content inversion, and overcoming the limitation of traditional empirical threshold methods that can only perform binary detection of snow melting states. This successfully quantifies the liquid water content on the surface of the Antarctic ice sheet. Attached Figure Description

[0017] Figure 1 This is a flowchart of the method steps of the present invention; Figure 2 This is a flowchart of the method of the present invention; Figure 3 (3a) is a brightness temperature distribution map of the substrate in an embodiment of the present invention; (3b) is a brightness temperature distribution map of the substrate obtained by calculating the annual average value of the vertical polarization brightness temperature near the Brewster angle; (3c) is a brightness temperature standard deviation distribution map of the substrate brightness temperature. Figure 4 This is the convergence curve of the genetic algorithm of this invention; Figure 5 This is a graph showing the simulated brightness temperature and observed brightness temperature values ​​of the radiative transfer model under different observation angles of the present invention. Figure 6 This is a comparison chart of liquid water content and melting days derived by the present invention. Figure 7 This is a comparison chart of liquid water content, melting days, and surface temperature (SMB) of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0019] Example 1 A method for simultaneously inverting snow liquid water content and snow density using L-band data has been successfully applied to the ablation zone of the Greenland ice sheet. This method, by integrating the liquid water inversion results, has verified the potential of the L-band in polar ice and snow detection. However, its application in broader polar environments, especially under different ice and snow conditions in Antarctica, has not yet been validated. Therefore, this invention provides a method for inverting the surface liquid water content of the Antarctic ice sheet based on a genetic algorithm-optimized radiative transfer model. This method utilizes low-frequency passive microwave remote sensing data (L-band) to invert the surface liquid water content of the Antarctic ice sheet. The method flow is as follows: Figure 1 As shown, the framework is as follows Figure 2 As shown.

[0020] Specifically, this includes: S1. Obtain L-band radiation brightness temperature data, and construct a brightness temperature distribution map of the Antarctic ice sheet from different observation angles, including time information, based on the radiation brightness temperature data.

[0021] S11. Obtain L-band radiative brightness temperature data of the SMOS satellite in the Antarctic region from the CATDS website.

[0022] S12. Process the radiation brightness temperature data, including spatial projection processing of the data, selection of observation angle and polarization information, and determination of time range and data subset.

[0023] S13. Calculate the average brightness temperature and time standard deviation of each pixel in Antarctica, and construct a brightness temperature distribution map of the Antarctic ice sheet from different observation angles from the beginning to the end of the year.

[0024] S2. Based on the brightness temperature distribution map of the Antarctic ice sheet, obtain the observed brightness temperature. After preprocessing the observed brightness temperature, select the V-polarized and H-polarized brightness temperature data with an incident angle of 22.5° to 62.5° and calculate the daily average value.

[0025] S21. Based on satellite orbital characteristics, select V-polarized and H-polarized brightness temperature data with an L-band incident angle range of 22.5° to 62.5° and calculate the standard deviation, then filter out brightness temperature data with a standard deviation greater than or equal to a preset value.

[0026] Specifically, the L-band radiometer of the SMOS satellite can acquire observation data in the range of 0° to 65°. However, due to the influence of satellite orbit characteristics, the spatial coverage integrity of satellite observations is best in the angle range of 22.5° to 62.5°. If the angle range is too narrow or too wide, it will lead to insufficient spatial distribution of observations, affecting the interpolation effect in the next step, thus reducing the inversion accuracy.

[0027] Furthermore, satellite-observed brightness temperatures are affected by random factors such as observation noise, atmospheric interference, and short-term changes in snow cover. Daily or single-angle brightness temperature values ​​may have deviations. The standard deviation quantifies the dispersion of brightness temperature data in the time series. If the standard deviation is small, it indicates that the brightness temperature observation is stable, less affected by random interference, and the mean is highly reliable. Conversely, a large standard deviation indicates that the data fluctuates drastically and there may be unresolved interference (such as atmospheric noise or abnormal changes in snow cover). Therefore, it is necessary to empirically filter out brightness temperature data with a standard deviation greater than or equal to a preset value.

[0028] S22. Based on the 5° observation angle interval of the SMOS satellite observation values, the aforementioned incident angle range is divided into 8 segments, resulting in... For each observation angle, gaps in the brightness temperature data that are less than a preset time are filled using linear interpolation, while gaps that are greater than or equal to the preset time are discarded.

[0029] Specifically, the SMOS satellite data observation angle interval is 5°, and the range from 22.5° to 62.5° is divided into 9 satellite observation angles. For these 9 satellite observation angles, the gaps in the brightness temperature data with an interval of less than 3 days are filled with missing values ​​using linear interpolation, and the brightness temperature data with an interval of more than or equal to 3 days are removed. This ensures that the pixel grid has brightness temperature observation values ​​of 9 angles for retrieval every day, thereby improving data reliability.

[0030] S23. Based on satellite orbit characteristics and polar observation coverage limitations, the latitude limit is selected as 83°S, and the pixel values ​​of the area south of the latitude limit in the Antarctic ice sheet brightness temperature distribution map are set to empty.

[0031] Specifically, due to limitations in satellite orbit design and polar observation coverage, the area south of 83°S latitude is affected by factors such as polar night and satellite observation angle, resulting in very few effective observation days per year. To ensure the reliability of the inversion results, the pixel values ​​south of 83°S latitude are set to empty to eliminate insufficient observation constraints caused by the limited number of observation days per year.

[0032] S24. The brightness temperature from the top of the atmosphere to the bottom of the atmosphere is corrected using the empirical atmospheric correction method. The expression is as follows: ， in, This indicates the corrected brightness temperature value; Indicates the observed brightness temperature; Indicates the polarization mode, when When indicates vertical polarization, Time indicates horizontal polarization; Indicates the i-th observation angle; Indicates the equivalent atmospheric temperature; This indicates atmospheric transmittance.

[0033] S3. Construct a radiative transfer model of the Antarctic ice sheet, including the dry snow layer, wet snow layer, atmosphere, and basement, and model the key parameters.

[0034] Regarding the radiative transfer model, the dry snow layer is the lower layer of snow, covering a semi-infinite ice substrate; the wet snow layer may contain liquid water, covering the surface of the dry snow layer.

[0035] Key parameters include the complex permittivity of dry and wet snow layers, liquid water content, and snow density. The method for modeling these key parameters is as follows: S31. The wet snow layer consists of dry snow matrix and liquid water. The Maxwell-Garnett mixing model is used to construct a model of the liquid water content and the complex permittivity of the wet snow layer.

[0036] , in, This represents the liquid water content in the wet snow layer, and is an unknown value. The complex permittivity of the dry snow matrix; The complex dielectric constant of liquid water is represented by a known value. The factor is represented, and its calculation method is as follows: , Indicates a constant value, and .

[0037] S32. Construct a model of snow density and complex permittivity of dry snow layer using empirical formulas.

[0038] , in, Indicates the complex permittivity of the dry snow layer; This represents the snow density, which is an unknown value. This represents the density ratio of accumulated snow to theoretical dry snow, and its calculation method is as follows: .

[0039] S4. Calculate the brightness temperature of each layer and obtain the Kirchhoff coefficient of each layer. Calculate the brightness temperature of the bottom atmospheric radiation transmission based on the Kirchhoff coefficient and the brightness temperature.

[0040] S41. For the brightness temperature of the substrate, obtain the vertical polarization brightness temperature within the preset range of the Brewster angle, as follows: Figure 3 As shown in (3a), its standard deviation is calculated as shown in (3b). After removing outliers based on the standard deviation, the average value of the vertical polarization brightness temperature is calculated, which is the brightness temperature of the substrate.

[0041] The Kirchhoff coefficients for the basis can be obtained directly from existing models in this field.

[0042] S42. Regarding the brightness temperature of the atmosphere, the method for calculating the sky brightness temperature is as follows: ， in, Indicates the brightness temperature of the sky; Represents the equivalent atmospheric temperature, and , This parameter represents the air temperature at a height of 2 meters above the ground. It is commonly used to estimate atmospheric temperature and reflects the air temperature conditions near the Earth's surface. This represents the cosmic microwave background radiation, which is usually considered to be a constant of 2.7K. Indicates atmospheric optical thickness, and has , This indicates the height of the ice sheet surface relative to sea level, and its unit is kilometers. Indicates the first i One observation angle.

[0043] The Kirchhoff coefficient for the atmosphere can be obtained directly from existing models in this field.

[0044] S43. For dry snow layers, in the L-band, dry snow layers can be considered as fully transparent, so their brightness temperature can be set to null, and their Kirchhoff coefficient can be approximated as 0.

[0045] S44. For the brightness temperature of wet snow layers, since snow usually approaches its freezing point when melting, its brightness temperature can be set to a constant value, i.e. Its Kirchhoff coefficients can be directly obtained from existing models in this field.

[0046] S45. Based on the Kirchhoff coefficient, the brightness temperatures of each layer are weighted and summed to obtain the brightness temperature of the radiative transfer at the bottom of the atmosphere, which is expressed as follows: , in, Indicates the brightness temperature of radiative transfer at the bottom of the atmosphere; Denotes the Kirchhoff coefficient of the j-th layer, and when Time represents the basis, when When indicates the dry snow layer, Time indicates the wet snow layer; Indicates the brightness temperature of the j-th layer; The Kirchhoff coefficient representing the sky; This indicates that the sky is at an observation angle of 0°. Brightness temperature; Indicates snow density; Indicates the liquid water content; Indicates the polarization mode, when When indicates vertical polarization, The time indicates horizontal polarization.

[0047] S5. A fitness function is constructed based on the daily average value and the brightness temperature of the radiative transfer at the bottom of the atmosphere. Based on the fitness function, a genetic algorithm is used to solve for the optimal snow density and liquid water content.

[0048] In this invention, the difference between the daily average brightness temperature calculated by S2 from multiple angles and the radiative transfer brightness temperature simulated by S4 is used as the fitness function of the genetic algorithm, and its expression is: , in, Indicates snow density; Indicates the liquid water content; Indicates the number of observed angles; Indicates the observation angle as of The daily average brightness temperature observed by the satellite; This indicates the brightness temperature of radiation transmission at the bottom of the atmosphere.

[0049] The detailed solution process is as follows: S51. Initialize the genetic algorithm parameters. Within the range of values, encode the snow density and liquid water content into two-dimensional vectors using real number encoding, and use them as initial candidate solutions to construct the population.

[0050] In this invention, a standard binary encoding strategy is employed to map the parameters to be optimized (i.e., snow density and liquid water content) within their value range to a string of length [length missing]. m The binary string. Specifically, the range of values ​​for the given parameter is... Its encoding process can be represented as: , in, This represents the minimum value of the parameter; Indicates the maximum value of the parameter; Represents a binary factor; This indicates the length of the binary string.

[0051] After sampling snow density and liquid water content using the above encoding method, they are jointly represented as a two-dimensional vector. ), This represents the snow density and its value range is... , This represents the liquid water content, and its value range is... .

[0052] The initial population size was set to 50 candidate solutions.

[0053] S52. Calculate the fitness function value of each initial candidate solution individual, and take the candidate solution individual corresponding to the minimum fitness function value as the global optimal solution.

[0054] Iterative execution steps S53~S5: S53. Select several candidate solutions from the population using the roulette wheel betting method, calculate the fitness function value of each candidate solution, and put the candidate solution corresponding to the minimum fitness function value into the mating pool. Repeat this process until the mating pool is full.

[0055] For each candidate solution, the probability of it being selected is: , in, Let represent the probability that the i-th candidate solution is selected. Let be the fitness value of the i-th candidate solution individual. The larger the fitness value of a candidate solution individual, the greater its probability of being selected. N is the population size.

[0056] S54. Randomly select two parent candidate solution individuals from the mating pool and perform gene crossover and gene mutation operations to generate new candidate solution individuals.

[0057] To generate the next generation of individuals, this study introduces standard genetic manipulation mechanisms, including single-point crossover and gene mutation, based on roulette wheel selection. In the crossover operation, one or more crossover points are randomly selected, and the chromosomes of two parent individuals are recombinated to produce two new individuals. In this invention, the crossover probability is set to 0.9 to maintain the dominant evolutionary driver of the population. Subsequently, a mutation operation is performed, randomly selecting several bits on the chromosome for bit flipping (0→1 or 1→0), introducing perturbation with a mutation probability of 0.01, thereby improving local search capabilities and breaking potential convergence stagnation.

[0058] S55. For the population in the current iteration, an elite retention strategy is adopted. The candidate solutions of the candidate solutions corresponding to the top 5% of the optimal fitness function values ​​in the population are copied to the next generation and formed into a new population with the new candidate solutions.

[0059] S56. Calculate the fitness function of each candidate solution individual in the new population, and select the candidate solution individual corresponding to the minimum fitness function value to update the global optimal solution.

[0060] S57. Determine whether the iteration termination condition is met. If yes, end the iteration and output the global optimal solution for the current iteration. Decode the solution to obtain the optimal snow density and liquid water content. Otherwise, return to execute S53.

[0061] Specifically, the termination conditions include: (1) reaching the maximum number of generations G=100; (2) the change in optimal fitness over 10 consecutive generations being less than 100. (3) The optimal fitness function value drops to below 0.5K.

[0062] The optimal two-dimensional vector is obtained after satisfying the condition. The inverted liquid water content is obtained by decoding in the following way: .

[0063] In addition to outputting the retrieved liquid water content, the simulation results can be analyzed to generate corresponding visualizations, which can then be output and saved. These visualizations can include spatial distribution maps of observed brightness temperature, optimized multi-angle model brightness temperature distribution maps, and retrieved surface water content distribution maps, visually demonstrating the spatial distribution characteristics and trends of shallow liquid water content in the Antarctic ice sheet. Furthermore, the simulation results can be visualized over time to further illustrate the spatial distribution of water content in the Antarctic ice sheet. Saving these visualizations facilitates subsequent data analysis and scientific research, and also provides reference and guidance for researchers in related fields.

[0064] Example 2 To verify the superiority of the radiative transfer model constructed in this invention, this embodiment introduces observed values ​​and their standard deviations to verify the optimized radiative transfer model proposed in this invention. After the model is optimized by a genetic algorithm, the convergence curve of the corresponding genetic algorithm is shown below. Figure 4 As shown, the comparison curves are as follows: Figure 5 As shown, the upward trend represents the vertically polarized brightness temperature, while the downward trend represents the horizontally polarized brightness temperature. Figure 5 The dashed line corresponds to the simulated value of the radiative transfer model optimized by the genetic algorithm, the solid line with the triangular symbol represents the observed value, and the other line represents the standard deviation of the observed value. It can be seen that the model brightness temperature after optimization algorithm correction is well fitted to the observed brightness temperature.

[0065] Furthermore, to verify the feasibility of the method provided by this invention, this embodiment inverts the annual average liquid water content distribution across Antarctica and the liquid water content distribution from 2018 to 2020 in areas including the Antarctic Peninsula, Amundsen Ice Shelf, Amery Ice Shelf, and Dronning Maud region, obtaining the following results: Figure 6 The results shown in the figure illustrate the spatial distribution of liquid water content more clearly. In this embodiment, the natural segment method is used for classification. Figure 5 As shown in (a) and (b), the distribution of liquid water on the Antarctic surface was similar in 2018 and 2019, mainly concentrated at the edge of the Antarctic continent, while the liquid water content was lower in the interior and other parts of the continent. However, (c) shows that in 2020, the coverage of liquid water in the West Antarctic coastal area and the Antarctic Peninsula expanded significantly, and the liquid water content further increased, which may be related to the rare heat wave event in Antarctica that year. Pixels with an annual average liquid water content exceeding 0.3 are extremely rare, mainly distributed in a few areas near the coast of the Antarctic Peninsula and coastal ice shelves. This phenomenon may be related to the instability of the ice shelves; for example, Google Earth imagery shows that the Amundsen Ice Shelf even partially collapsed during this period.

[0066] Furthermore, additional validation data was selected for inversion, and the results are as follows: Figure 7 As shown, the inversion data comes from the number of melting days, annual average surface temperature, and surface mass balance (SMB) in Antarctica detected by a threshold-based algorithm. The detailed threshold algorithm analyzes the annual brightness-temperature variation of each pixel on the 1.4 GHz SMOS satellite, determines the optimal threshold based on statistical analysis or adaptive methods using historical observation data, and considers any pixel exceeding this threshold as having melted. Figure 7 As shown in (b), the annual average surface temperature was obtained by averaging the monthly surface temperature data from 2018 to 2020 obtained from the ECMWF database, as shown in (b). Figure 7 As shown in (c); the estimated value of surface quality balance is estimated using the net value of precipitation (P) minus evaporation (E), i.e. ,like Figure 7 As shown in (d); the inversion result of the method provided by this invention is as follows Figure 7 As shown in (a), it can be seen that the inverted liquid water content is highly consistent with the spatial distribution of surface temperature, melting days, and SMB. The areas where liquid water exists are all distributed in areas with high surface temperature, long melting days, and high SMB.

[0067] Example 3 The present invention also provides an electronic device including a central processing unit (CPU) that can perform various appropriate actions and processes according to computer program instructions stored in read-only memory (ROM) or loaded from a storage unit into random access memory (RAM). The RAM may also store various programs and data required for device operation. The CPU, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0068] Multiple components in the device are connected to the I / O interface, including: input units such as keyboards and mice; output units such as various types of displays and speakers; storage units such as disks and optical discs; and communication units such as network interface cards (NICs), modems, and wireless transceivers. The communication unit allows the device to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0069] The processing unit executes the various methods and processes described above, such as methods S1 to S5. For example, in some embodiments, methods S1 to S5 may be implemented as computer software programs tangibly contained in a machine-readable medium, such as a storage unit. In some embodiments, part or all of the computer program may be loaded and / or installed on the device via ROM and / or a communication unit. When the computer program is loaded into RAM and executed by the CPU, one or more steps of methods S1 to S5 described above may be performed. Alternatively, in other embodiments, the CPU may be configured to execute methods S1 to S5 by any other suitable means (e.g., by means of firmware).

[0070] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: Field Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), Application Standard Products (ASSPs), System-on-Chip (SoCs), Complex Programmable Logic Devices (CPLDs), and so on.

[0071] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0072] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0073] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for inverting the surface liquid water content of the Antarctic ice sheet based on a genetic algorithm-optimized radiative transfer model, characterized in that, include: S1. Obtain L-band radiation brightness temperature data, and construct a brightness temperature distribution map of the Antarctic ice sheet from different observation angles, including time information, based on the radiation brightness temperature data. S2. Based on the aforementioned Antarctic ice sheet brightness temperature distribution map, obtain the observed brightness temperature. After preprocessing the observed brightness temperature, select V-polarized and H-polarized brightness temperature data with incident angles from 22.5° to 62.5° and calculate the daily average value. S3. Construct a radiative transfer model of the Antarctic ice sheet, including the dry snow layer, wet snow layer, atmosphere, and basement, and model the key parameters. S4. Calculate the brightness temperature of each layer and obtain the Kirchhoff coefficient of each layer. Calculate the brightness temperature of atmospheric bottom radiation transmission based on the Kirchhoff coefficient and brightness temperature. S5. Construct a fitness function based on the daily average value and the brightness temperature of the radiative transfer at the bottom of the atmosphere, and use a genetic algorithm to solve for the optimal snow density and liquid water content based on the fitness function. The methods for determining the optimal snow density and liquid water content include: S51. Initialize the genetic algorithm parameters. Within the range of values, encode the snow density and liquid water content into two-dimensional vectors using real number encoding, and use them as initial candidate solution individuals to construct a population. S52. Calculate the fitness function value of each initial candidate solution individual, and take the candidate solution individual corresponding to the minimum fitness function value as the global optimal solution; Iterative execution steps S53~S5: S53. Select several candidate solution individuals from the population using the roulette wheel betting method, calculate the fitness function value of each candidate solution individual, put the candidate solution individual corresponding to the minimum fitness function value into the mating pool, and repeat the process until the mating pool is full. S54. Randomly select two parent candidate solution individuals from the mating pool and perform gene crossover and gene mutation operations to generate new candidate solution individuals; S55. For the population in the current iteration, an elite retention strategy is adopted to copy a preset number of candidate solutions in the population to the next generation, and form a new population with the new candidate solution individuals. S56. Calculate the fitness function of each candidate solution individual in the new population, and select the candidate solution individual corresponding to the minimum fitness function value to update the global optimal solution; S57. Determine whether the iteration termination condition is met. If yes, end the iteration and output the global optimal solution for the current iteration. Decode the solution to obtain the optimal snow density and liquid water content. Otherwise, return to execute S53.

2. The method for inverting the liquid water content on the surface of the Antarctic ice sheet based on a genetic algorithm-optimized radiative transfer model, as described in claim 1, is characterized in that... The preprocessing includes: Based on satellite orbital characteristics, brightness temperature data of V-polarization and H-polarization with an incident angle range of 22.5° to 62.5° in the L-band were screened and the standard deviation was calculated. Brightness temperature data with a standard deviation greater than or equal to the preset value were screened out. Based on the 5° observation angle interval of the SMOS satellite, the aforementioned incident angle range is divided into 8 segments, resulting in... For each observation angle, the time interval of brightness temperature data is less than the preset time interval, and the gap is filled by linear interpolation. The gap is discarded if it is greater than or equal to the preset time interval. Based on satellite orbit characteristics and polar observation coverage limitations, the latitude limit was selected as 83°S, and the pixel values ​​of the region south of the latitude limit in the Antarctic ice sheet brightness temperature distribution map were set to empty. The brightness temperature from the top of the atmosphere to the bottom of the atmosphere is corrected using an empirical atmospheric correction method.

3. The method for inverting the surface liquid water content of the Antarctic ice sheet based on a genetic algorithm-optimized radiative transfer model, as described in claim 2, is characterized in that... The empirical atmospheric correction method is as follows: ， in, This indicates the corrected brightness temperature value; Indicates the observed brightness temperature; Indicates the polarization mode, when When indicates vertical polarization, Time indicates horizontal polarization; Indicates the i-th observation angle; Indicates the equivalent atmospheric temperature; This indicates atmospheric transmittance.

4. The method for inverting the surface liquid water content of the Antarctic ice sheet based on a genetic algorithm-optimized radiative transfer model, as described in claim 1, is characterized in that... The key parameters include the complex dielectric constant of the dry snow layer and the wet snow layer, the liquid water content, and the snow density. The method for modeling these key parameters is as follows: The wet snow layer includes dry snow matrix and liquid water. A model of the relationship between liquid water content and complex permittivity of wet snow layer is constructed using the Maxwell-Garnett mixing model. A model of snow density and complex permittivity of dry snow layer was constructed using empirical formulas.

5. The method for inverting the surface liquid water content of the Antarctic ice sheet based on a genetic algorithm-optimized radiative transfer model, as described in claim 4, is characterized in that... The method for constructing a model of liquid water content and complex permittivity of wet snow layer is as follows: ， in, This represents the liquid water content in the wet snow layer, and is an unknown value. This represents the complex permittivity of the dry snow matrix. The complex dielectric constant of liquid water is represented by a known value. The factor is represented, and its calculation method is as follows: ， It is a constant, and =0.005, = =0.4975.

6. The method for inverting the surface liquid water content of the Antarctic ice sheet based on a genetic algorithm-optimized radiative transfer model, as described in claim 4, is characterized in that... The method for constructing a model of snow density and the complex permittivity of dry snow layer is as follows: ， in, Indicates the complex permittivity of the dry snow layer; This represents the snow density, which is an unknown value. This represents the density ratio of accumulated snow to theoretical dry snow, and its calculation method is as follows: .

7. The method for inverting the surface liquid water content of the Antarctic ice sheet based on a genetic algorithm-optimized radiative transfer model, as described in claim 1, is characterized in that... Methods for calculating the brightness temperature of each layer include: For the substrate, obtain the vertical polarization brightness temperature within a preset range of the Brewster angle, and calculate the average value of the vertical polarization brightness temperature, which is the brightness temperature of the substrate. For the aforementioned atmosphere, the sky brightness temperature is calculated as follows: ， in, Indicates the brightness temperature of the sky; Represents the equivalent atmospheric temperature, and , This indicates the air temperature at a height of 2 meters above the ground. Represents the cosmic microwave background radiation; Indicates atmospheric optical thickness, and has , This indicates the height of the ice sheet surface relative to sea level; Indicates the first i One observation angle; For the dry snow layer, its brightness temperature is set to null based on the properties of the dry snow layer; For the aforementioned wet snow layer, its brightness temperature is set to a constant value based on the properties of the wet snow layer, i.e. .

8. The method for inverting the liquid water content on the surface of the Antarctic ice sheet based on a genetic algorithm-optimized radiative transfer model, as described in claim 1, is characterized in that... The method for calculating the brightness temperature of atmospheric bottom radiative transfer is as follows: ， in, Indicates the brightness temperature of radiation transmission at the bottom of the atmosphere; Denotes the Kirchhoff coefficient of the j-th layer, and when Time represents the basis, when When indicates the dry snow layer, Time indicates the wet snow layer; Indicates the brightness temperature of the j-th layer; The Kirchhoff coefficient representing the sky; Indicates the sky at an observation angle of Brightness temperature; Indicates snow density; Indicates the liquid water content; Indicates the polarization mode, when When indicates vertical polarization, The time indicates horizontal polarization.

9. The method for inverting the surface liquid water content of the Antarctic ice sheet based on a genetic algorithm-optimized radiative transfer model, as described in claim 1, is characterized in that... The fitness function is: ， in, Indicates snow density; Indicates the liquid water content; Indicates the number of observed angles; Indicates the observation angle as of The daily average brightness temperature observed by the satellite; This indicates the brightness temperature of radiation transmission at the bottom of the atmosphere.

Images