Calibration performance evaluation and analysis method for geostationary satellite hyperspectral infrared sounder

Abstract

The present invention discloses a calibration performance evaluation and analysis method for a geostationary satellite hyperspectral infrared detector. First, the ERA5 data set and the L1-level observation data of the geostationary satellite hyperspectral infrared detector are obtained and input into a rapid transmission radiation model to obtain simulated radiation data. Then, the simulated radiation data is compared with the L1-level observation data of the geostationary satellite hyperspectral infrared detector, and the systematic deviation is calculated and the deviation analysis is performed. Finally, the L1-level observation data of the geostationary satellite hyperspectral infrared detector is cross-matched and verified with the observation data of the polar-orbiting satellite infrared atmospheric sounding interferometer. Based on the atmospheric radiation model, the present invention uses the latest high-quality data and combines multi-source reference data to provide a more accurate and reliable calibration performance daily variation characteristic alternative calibration technology, thereby effectively overcoming the shortcomings of the existing technology and significantly improving the quality of the detection data and the application effect.

Description

Technical Field

[0001] The present invention belongs to the technical field of satellite remote sensing, and in particular relates to a calibration performance evaluation and analysis method for a geostationary satellite hyperspectral infrared detector. Background Art

[0002] Liu Juanjuan et al. from the Institute of Atmospheric Physics, Chinese Academy of Sciences (Liu Juanjuan, Xu Lan, Cheng Wei, et al. Analysis of bias characteristics and bias correction of the GIIRS detector on the FY-4A satellite for data assimilation [J]. Journal of Atmospheric Sciences, 2022, 46(02): 275-292.) conducted a comprehensive analysis of the observation bias (O−B) of the Interferometric Atmospheric Vertical Sounder (GIIRS). The team used the high-resolution regional model WRF and its assimilation system WRFDA to study the bias distribution characteristics of different channels. The study found that the bias and standard deviation of the longwave channel are generally smaller than those of the mediumwave channel, but both channels are contaminated. The diurnal variation of the bias is weakly related to the satellite zenith angle, but is related to the brightness temperature and the position of the satellite scanning array, showing the characteristics of "array bias". After recalibration in 2020, the quality of GIIRS observation data has been significantly improved. Further bias correction experiments have shown that selecting the scanning array as the main factor can effectively improve the bias.

[0003] Niu, Zeyi, et al. "Performances between the FY-4A / GIIRS and FY-4B / GIIRS long-wave infrared (LWIR) channels under clear-sky and all-sky conditions." Quarterly Journal of the Royal Meteorological Society 149.754(2023): (1612-1628) compared the observed and background deviation characteristics of the longwave infrared channel of the Geostationary Interferometric Infrared Sounder (GIIRS) on the Fengyun-4A (FY-4A) and Fengyun-4B (FY-4B) satellites. The results showed that under clear sky conditions, the FY-4B / GIIRS data quality was superior to that of the FY-4A / GIIRS, especially in the carbon dioxide and ozone absorption bands, where the standard deviation of the FY-4B was less than 1 K, while that of the FY-4A was greater than 1 K. Furthermore, the diurnal deviation of the ozone absorption band of the FY-4B was smaller than that of the FY-4A, while the diurnal deviation of the carbon dioxide absorption band exhibited a significant phase shift, likely due to diurnal variations in the environmental field caused by the different geostationary orbital platforms. Furthermore, the FY-4B / GIIRS observations were more consistent with the cloud simulations under cloudy conditions, showing a smaller root mean square error, further demonstrating that its observation quality is superior to that of the FY-4A / GIIRS in cloudy areas.

[0004] Lu Qifeng and others from the National Satellite Meteorological Center invented a satellite data quality monitoring method based on the intersection dual reference source deviation (application number: CN202111576921.9). This method combines the simulation reference source and the observation reference source to quickly distinguish between background field problems caused by numerical forecast models and problems caused by satellite or instrument anomalies, thereby effectively monitoring changes in the instrument L1 data quality, quickly locating the causes that affect the quality of the observation data, and providing the necessary information for analyzing and correcting instrument performance. Although this invention effectively solves the data accumulation problem of the observation reference source and the accuracy problem of the simulation reference source, the polar orbit and geostationary satellite observation system methods it covers tend to be broadly statistical, ignoring the difference between the two. For geostationary satellite hyperspectral infrared sounders, special data processing methods are needed to improve the quality of statistical data, and incorporate time dimension information into the statistical results to better reflect its high temporal resolution characteristics.

[0005] With the continuous development of China's geostationary infrared hyperspectral detection technology, the quality and resolution of detection data have significantly improved. However, existing research and technologies are mostly based on data from before 2022, which is relatively old and does not reflect the data quality of the latest technological improvements. Therefore, the use of outdated data may reduce the precision of research results and applications and fail to fully utilize the advantages of the latest detectors. Furthermore, many existing studies have limited data sets and sample coverage, resulting in poor generalizability of their conclusions. Excessive data may not represent a wider range of climate and atmospheric conditions, limiting the reliability and generalizability of research results in practical applications. Larger and more extensive data collection and analysis are important ways to improve the confidence of the results. Furthermore, current technologies often rely on atmospheric radiation models as a reference for radiation correction. However, atmospheric radiation models themselves may have errors, especially under complex meteorological conditions. This over-reliance on models can introduce systematic errors, affecting calibration accuracy and, in turn, the interpretation and application of detection data. Therefore, relying solely on atmospheric radiation models may not be sufficient to meet the needs of high-precision radiation calibration. More field data and advanced algorithms are needed to improve accuracy. In addition, current technology ignores the high temporal resolution of geostationary infrared hyperspectral observation data. Therefore, it is necessary to incorporate information of the time dimension into the statistical results to improve the accuracy and information richness of the calibration performance evaluation, thereby providing a more valuable reference for meteorological products. Summary of the Invention

[0006] The main purpose of the present invention is to overcome the shortcomings and deficiencies of the existing technology and provide a calibration performance evaluation and analysis method for a geostationary satellite hyperspectral infrared detector. Based on the atmospheric radiation model, the method combines multi-source data to conduct a comprehensive analysis of the diurnal variation of the observation data deviation, the correlation of brightness temperature deviation, and the wavelength and time variation over the year, thereby significantly improving the accuracy and reliability of the calibration technology of the diurnal variation characteristics of the radiation calibration performance.

[0007] In order to achieve the above object, the present invention adopts the following technical solutions:

[0008] The calibration performance evaluation and analysis method of a geostationary satellite hyperspectral infrared sounder includes the following steps:

[0009] The ERA5 dataset and L1-level observation data from a geostationary satellite hyperspectral infrared sounder were obtained and input into a rapid transfer radiation model to obtain simulated radiation data. The ERA5 dataset includes ERA5 temperature and humidity profiles and ERA5 surface temperature. The geostationary satellite hyperspectral infrared sounder observation data includes ground infrared emissivity, land and sea markers, and a digital elevation model. The rapid transfer radiation model was simulated over land and sea areas under clear sky conditions.

[0010] The simulated radiation data were compared with the L1 observation data of the geostationary satellite hyperspectral infrared sounder, and the systematic deviation was calculated and analyzed.

[0011] The L1 level observation data of the geostationary satellite hyperspectral infrared sounder and the observation data of the polar-orbiting satellite infrared atmospheric sounding interferometer were cross-matched and verified.

[0012] As a preferred technical solution, the method for determining the land and sea areas under clear sky conditions is:

[0013] Cloud images are obtained using geostationary radiation imagers carried by geostationary satellites;

[0014] Divide the cloud image into multiple sub-image areas of x*x pixels;

[0015] Determine whether each pixel in the sub-image area is covered by clouds. If not, mark the center point of the sub-image area as clear sky.

[0016] As a preferred technical solution, during the simulation of the rapid transmission radiation model, a data preprocessing operation is performed, specifically:

[0017] Exclude L1-level observation data with a time difference of more than m1 minutes compared with the ERA5 dataset, and exclude L1-level observation data with a satellite zenith angle of more than 30°;

[0018] L1-level observation data whose deviation between simulated radiation data and L1-level observation data exceeds n times the standard deviation are regarded as outliers and are eliminated.

[0019] As a preferred technical solution, the calculation of systematic deviation is specifically as follows:

[0020] Use the simulated radiation data as reference data;

[0021] The correspondence between the reference data and the L1-level observation data is determined through spatiotemporal matching: the reference data is interpolated based on the spatiotemporal information of the L1-level observation data to obtain the first fused data; during the interpolation process, linear interpolation is used for time, and spline interpolation is used for spatial latitude and longitude;

[0022] The first fusion data is grouped according to the spectral channel based on the spectral center wavenumber to obtain multiple groups of sub-data; each group of sub-data contains L1 observation data and reference data under the corresponding spectral channel;

[0023] The overall mean and standard deviation of each group of sub-data in each spectral channel are calculated, and the mean and standard deviation of the deviation between the L1-level observation data and the reference data in each group of sub-data are also calculated.

[0024] As a preferred technical solution, the deviation analysis refers to performing statistical analysis in each spectral channel;

[0025] The statistical analysis includes the following four parts:

[0026] Annual analysis: With a time range of one year, the mean and standard deviation of the deviation between the brightness temperature of the L1 observation data and the reference data are calculated each year;

[0027] Daily analysis: Taking the day as the time range, calculate the average value of the deviation between the brightness temperature of the daily L1 observation data and the brightness temperature of the reference data, and analyze the change pattern of the average value with the date;

[0028] Hourly analysis: Using the hourly time range, calculate the average deviation between the brightness temperature of the L1 observation data and the reference data every hour. Group the statistical deviation results at the same time on different days into one group, and calculate their average and standard deviation.

[0029] Overall deviation analysis: Taking the year as the time range, calculate the deviation between the brightness temperature of the L1-level observation data of all spectral channels and the brightness temperature of the reference data in each year, make a deviation scatter plot, and calculate the correlation coefficient between the brightness temperature of the L1-level observation data of all channels and the brightness temperature of the reference data.

[0030] As a preferred technical solution, before the cross-matching verification, the L1-level observation data of the geostationary satellite hyperspectral infrared sounder and the observation data of the polar-orbiting satellite infrared atmospheric sounding interferometer are screened to select sub-satellite point cross samples that meet the standards, specifically:

[0031] Select the L1-level observation data of the geostationary satellite hyperspectral infrared sounder and the observation data of the polar-orbiting satellite infrared atmospheric sounding interferometer with a measurement time difference of less than m2 minutes and a relative difference between the zenith angles of less than a%;

[0032] Use the geostationary orbit radiation imager carried by the geostationary satellite to obtain data in the central wavelength range of 10.30 to 11.30 μm to select the spatially uniform observation area;

[0033] The criteria for determining a spatially uniform observation area are as follows: a circle is drawn in the observation area with the center of the pixel of the observation data of the geostationary satellite hyperspectral infrared detector as the center and half of the spatial resolution distance as the radius. All the time-space-matched geostationary orbit radiation imager observation data within the circle constitute a determination data group. If the ratio of the standard deviation to the mean of the determination data group is less than 0.01, the circle is considered to be a spatially uniform observation area.

[0034] As a preferred technical solution, when performing the cross-matching verification, the observation data of the polar-orbiting satellite infrared atmospheric sounding interferometer is processed in the following steps:

[0035] Perform inverse Fourier transform on the raw spectral data of the polar-orbiting satellite infrared atmospheric sounding interferometer to convert the frequency domain data into the first interferogram;

[0036] The de-Gaussian apodization technique is applied to reduce the spectral distortion in the first interferogram, and the first interferogram is truncated to remove the high noise part in the signal and reduce the spectral resolution to obtain the second interferogram;

[0037] Performing Fourier transform on the second interferogram and reconverting it back to frequency domain data to obtain processed spectral data;

[0038] The spectral data is smoothed using a Hamming window to reduce the sidelobe effect.

[0039] As a preferred technical solution, the first interference pattern is truncated to an optical path difference of 0.8 cm;

[0040] The spectral resolution is reduced to the spectral resolution of a geostationary satellite hyperspectral infrared detector.

[0041] As a preferred technical solution, the cross-matching verification is specifically as follows:

[0042] The observation data of the polar-orbiting satellite infrared atmospheric sounding interferometer is used as reference data;

[0043] The correspondence between the reference data and the L1-level observation data is determined through spatiotemporal matching: the reference data is interpolated based on the spatiotemporal information of the L1-level observation data to obtain the second fused data; during the interpolation process, linear interpolation is used for time, and spline interpolation is used for spatial latitude and longitude;

[0044] The second fusion data is grouped according to the spectral channel based on the spectral center wavenumber to obtain multiple groups of channel data; each group of channel data contains L1 observation data and reference data under the corresponding spectral channel;

[0045] Calculate the overall mean and overall standard deviation of each set of channel data in each spectral channel, and calculate the mean and standard deviation of the deviation between the L1 observation data and the reference data in each set of channel data.

[0046] On the other hand, it provides a calibration performance evaluation and analysis system for geostationary satellite hyperspectral infrared detectors, including a radiation data simulation module, a deviation calculation and analysis module, and a cross-matching verification module;

[0047] The radiation data simulation module is used to obtain the ERA5 dataset and L1-level observation data from the geostationary satellite hyperspectral infrared sounder, and input them into the rapid transmission radiation model to obtain simulated radiation data; the ERA5 dataset includes the ERA5 temperature and humidity profile and the ERA5 surface temperature; the observation data from the geostationary satellite hyperspectral infrared sounder includes infrared channel observation brightness temperature, observation angle, ground infrared emissivity, sea and land identification, and digital elevation model; the rapid transmission radiation model is simulated in land and sea areas under clear sky conditions;

[0048] The deviation calculation and analysis module is used to compare the simulated radiation data with the L1 level observation data of the geostationary satellite hyperspectral infrared sounder, calculate the systematic deviation and perform deviation analysis;

[0049] The cross-matching verification module is used to cross-match and verify the L1 level observation data of the geostationary satellite hyperspectral infrared sounder with the observation data of the polar-orbiting satellite infrared atmospheric sounding interferometer.

[0050] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0051] 1. Based on the atmospheric radiation model, this invention also references observational data from the IASI, a similar instrument carried on polar-orbiting satellites. By combining multi-source data, this method comprehensively analyzes the diurnal variation of observational data deviations, the correlation of brightness temperature deviations, and the annual changes in wavelength and time. Existing technologies typically rely on a single atmospheric radiation model, which is prone to model errors. The multi-source data reference method of the present invention significantly improves the accuracy and reliability of the diurnal variation characteristics of radiometric calibration performance, replacing calibration techniques.

[0052] 2. This invention comprehensively analyzes the deviation changes in observation data over a year, particularly under different wavelength and time conditions, to more meticulously characterize the characteristics of the observation data. Compared with existing technologies, this technology is more sophisticated and comprehensive in its implementation, avoiding the errors that may be introduced by over-reliance on atmospheric radiation models.

[0053] 3. Existing technologies primarily rely on outdated data and a single atmospheric radiation model for radiometric calibration, ignoring the unique characteristics of geostationary satellite infrared hyperspectral instruments (low cloud detection accuracy, high temporal resolution, multiple channels, and high spectral resolution). This results in insufficient timeliness and universality of the data and may introduce systematic errors. In contrast, the present invention uses the latest, high-quality data combined with multi-source reference data to provide a more accurate and reliable calibration technique based on diurnal variation characteristics. This effectively overcomes the shortcomings of existing technologies and significantly improves the quality of detection data and its application effectiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0055] Figure 1 This is an overall flow chart of the calibration performance evaluation and analysis method of a geostationary satellite hyperspectral infrared detector in an embodiment of the present invention.

[0056] Figure 2 Schematic diagram of the process of comparison and deviation analysis in an embodiment of the present invention.

[0057] Figure 3 Schematic diagram of the cross-matching verification process in an embodiment of the present invention.

[0058] Figure 4 Schematic diagram of the annual average statistical deviation of each channel in an embodiment of the present invention.

[0059] Figure 5 Schematic diagram of daily variation statistics of the calibration performance of a stationary hyperspectral infrared detector in an embodiment of the present invention.

[0060] Figure 6 Schematic diagram of the statistical characteristics of the momentary change of the calibration performance of the stationary hyperspectral infrared detector in an embodiment of the present invention.

[0061] Figure 7 This is a statistical diagram of the scatter points of each sample deviation of the calibration performance of the stationary hyperspectral infrared detector in an embodiment of the present invention.

[0062] Figure 8 4 is a structural diagram of a calibration performance evaluation and analysis system for a geostationary satellite hyperspectral infrared detector according to an embodiment of the present invention. DETAILED DESCRIPTION

[0063] In order to enable those skilled in the art to better understand the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0064] References to "embodiments" in this application mean that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute an independent or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described in this application may be combined with other embodiments.

[0065] The following are definitions of the technical terms used in this application. Other undefined terms should be understood according to the common technical knowledge of those skilled in the art:

[0066] Geostationary satellite: A satellite that orbits at an altitude of approximately 36,000 kilometers above the Earth's equator and is able to keep pace with the Earth's rotation, thereby maintaining a fixed position relative to the ground.

[0067] Polar-orbiting satellite: A near-geosynchronous satellite that uses a near-polar sun-synchronous orbit and maintains a fixed angle between the satellite orbit plane and the sun's rays. It can provide effective observation data in medium-term numerical weather forecasting, climate diagnosis and prediction, natural disasters and environmental monitoring.

[0068] Geosynchronous Interferometric Infrared Sounder / GIIRS: The world's first precision remote sensing instrument that uses infrared hyperspectral interferometry to detect the three-dimensional vertical structure of the atmosphere in geostationary orbit. It can detect and record the radiation characteristics of target objects with high spectral resolution in the infrared band. It observes infrared radiation in different spectral bands through Michelson interferometry to obtain the vertical distribution of atmospheric temperature and humidity, providing large-scale, continuous, rapid and accurate remote sensing information for weather forecasting. It is used for detecting vertical profiles of atmospheric temperature and humidity and assimilating numerical forecast data.

[0069] Infrared Atmospheric Sounding Interferometer (IASI): A new generation of ultra-high-spectral infrared atmospheric sounding instrument carried on European polar-orbiting meteorological satellites. It uses interferometric spectroscopy technology. Its spectral measurement range covers multiple absorption bands and can be used to invert the atmosphere, ocean, clouds and atmospheric composition, providing rich remote sensing data for Earth's atmospheric remote sensing, meteorological operations and scientific research.

[0070] Advanced Geostationary Radiation Imager (AGRI): One of the main payloads of the Fengyun-4 geostationary meteorological satellite, it achieves precise and flexible two-dimensional pointing through a sophisticated dual-scanning mirror mechanism, enabling rapid regional scanning at the minute level. It uses an off-axis three-mirror main optical system to acquire Earth cloud images in more than 14 bands at high frequency, and uses an on-board blackbody for high-frequency infrared calibration to ensure the accuracy of the observation data. It is primarily responsible for acquiring cloud images.

[0071] Calibration: Adjusting a measuring instrument using a known standard or reference object to ensure the accuracy and consistency of its output.

[0072] Rapid Transfer for TOVS (RTTOV): A fast radiative transfer model developed by the European Centre for Medium-Range Weather Forecasts (ECMWF) that simulates infrared and microwave radiation from Earth observations by meteorological satellite sensors.

[0073] The fifth generation of European Centre for Medium-Range Weather Forecasts climate reanalysis data / ERA5: A global atmospheric reanalysis dataset provided by the European Centre for Medium-Range Weather Forecasts (ECMWF), which provides high-resolution atmospheric, land, and ocean meteorological data and is widely used as input to radiative transfer models.

[0074] This embodiment takes the hyperspectral infrared detector (i.e., GIIRS) carried by Fengyun-4B satellite (FY-4B) as an example to elaborate on the calibration performance evaluation and analysis method of the geostationary satellite hyperspectral infrared detector proposed in this application. Figure 1 As shown, the following steps are included:

[0075] S1. Obtain the ERA5 dataset and FY-4B / GIIRS L1-level observation data and input them into the Rapid Transmission Radiation Model (RTTOV) to generate simulated radiation data. The ERA5 dataset includes ERA5 temperature and humidity profiles and ERA5 surface temperature. The geostationary satellite hyperspectral infrared sounder observation data includes infrared channel brightness temperature, observation angle, ground infrared emissivity, land and sea markers, and a digital elevation model (DEM). The RTTOV simulates land and sea areas under clear-sky conditions to generate simulated radiation data.

[0076] S2. Compare the simulated radiation data with the L1 observation data of the geostationary satellite hyperspectral infrared sounder, calculate the systematic deviation and perform deviation analysis.

[0077] S3. Cross-match and verify the L1 observation data of the geostationary satellite hyperspectral infrared sounder with the observation data of the polar-orbiting satellite infrared atmospheric sounding interferometer.

[0078] Furthermore, to ensure simulation accuracy, this application focuses on radiative transfer simulation over land and sea areas under clear-sky conditions, as the presence of clouds introduces significant uncertainty. Under clear-sky conditions, RTTOV simulation accuracy is generally high, with simulation errors over land areas less than 0.01 K. For combined land and sea areas under clear-sky conditions, the average deviation between RTTOV and line-by-line simulation is also less than 0.01 K. Therefore, clear-sky land and sea areas are selected for simulation, and the method for determining clear-sky land and sea areas is as follows:

[0079] Cloud images were acquired using the AGRI instrument carried by FY4B. The cloud images were then divided into multiple x*x pixel sub-image regions. A check was performed to determine whether each sub-image region was cloud-covered. If no pixel in the sub-image region was cloud-free, the center point of the sub-image region was marked as clear sky. This method ensured that the radiation transfer characteristics of the selected region met the clear sky condition, thereby minimizing the impact of clouds on the bias analysis.

[0080] In this embodiment, the cloud image is divided into multiple 5*5 pixel sub-image areas. If 25 pixels in a sub-image area are not covered by clouds, the center point of the sub-image area is marked as clear sky, which meets the clear sky condition.

[0081] Furthermore, to improve data quality, data preprocessing was performed during the rapid transmission radiation model simulation to ensure the reliability of the analysis results. First, FY-4B / GIIRSL1 observations with a time difference of more than m1 minutes compared to the ERA5 dataset were excluded. Second, L1 observations with satellite zenith angles exceeding 30° were excluded to help reduce uncertainty caused by observation geometry. Finally, L1 observations with deviations between the simulated radiation data and the L1 observations exceeding n times the standard deviation were considered outliers and removed. These operations ensured high data quality during the deviation analysis.

[0082] In this embodiment, observation data with a time difference of more than 10 minutes are excluded, and L1-level observation data with a deviation exceeding 3 times the standard deviation are regarded as outliers and removed.

[0083] In this example, RTTOV utilizes the TOVS 13.1 radiative transfer model. Although the RTTOV model has uncertainties (especially in terms of input profiles), it can simulate GIIRS observations throughout the day, thus providing a large sample size. This advantage enables a more detailed analysis, including the diurnal variation of the bias, the correlation of brightness temperature bias, and the variation of the bias with wavelength and time over the year. This large sample size makes up for the shortcomings of the limited cross-matching data.

[0084] Furthermore, after obtaining the simulated radiation data, we compared it with the FY-4B / GIIRS L1 observation data to calculate the systematic deviation, specifically:

[0085] The simulated radiation data is used as the reference data. First, the correspondence between the reference data and the L1-level observation data is determined through spatiotemporal matching. Based on the spatiotemporal information of the L1-level observation data, the reference data is interpolated to obtain the first fused data. Linear interpolation is used for time, and spline interpolation is used for spatial latitude and longitude. Next, the first fused data is grouped according to spectral channel based on the spectral center wavenumber. The data for each spectral channel is independently divided into a group, resulting in multiple sub-data sets. Each sub-data set contains two data sets for the corresponding spectral channel: the L1-level observation data and the reference data. Each element in these two data sets maintains consistency in spatiotemporal information. Finally, for each sub-data set, the overall mean and overall standard deviation of each sub-data set in each spectral channel are calculated. The mean and standard deviation of the deviation between the L1-level observation data and the reference data in each sub-data set are also calculated, providing a basis for analyzing systematic deviations.

[0086] Furthermore, the geostationary satellite hyperspectral infrared detector is divided into several spectral channels according to the spectral center wave number, and statistical analysis is performed independently on each spectral channel. The statistical analysis mainly includes the following four parts:

[0087] Annual analysis: With a time range of one year, the mean and standard deviation of the deviation between the brightness temperature of the L1 observation data and the brightness temperature of the reference data are calculated each year.

[0088] Daily analysis: Taking the day as the time range, the average value of the deviation between the brightness temperature of the daily L1 observation data and the brightness temperature of the reference data is calculated, and the change pattern of the average value with the date is analyzed.

[0089] Hourly analysis: Using the hourly time range, the average deviation between the brightness temperature of the L1 observation data and the reference data is calculated every hour. The statistical deviation results at the same time on different dates are grouped together, and their average and standard deviation are calculated.

[0090] Overall deviation analysis: Taking the year as the time range, calculate the deviation between the brightness temperature of the L1-level observation data points of all spectral channels and the brightness temperature of the reference data in each year, make a deviation scatter plot, and calculate the correlation coefficient between the brightness temperature of the L1-level observation data of all spectral channels and the brightness temperature of the reference data.

[0091] Specifically, the correlation coefficient calculation formula is:

[0092] ,

[0093] in, X and Yare vectors composed of observation data and reference data respectively, Cov ( X , Y )for X and Y The covariance of Var ( X )and Var ( Y ) are X and Y The variance of .

[0094] By analyzing GIIRS-RTTOV data (i.e., L1-level observational data and simulated radiation data), we can observe diurnal variations in biases. This is crucial for understanding the performance of FY-4B / GIIRS, as the radiation properties of the atmosphere and surface can differ significantly between day and night. These analysis results will provide a scientific basis for future calibration and data applications. In addition, this application also analyzes how biases vary with wavelength and time. This analysis helps identify systematic biases in different spectral bands and over different time periods, and further understands the long-term stability of FY-4B / GIIRS. These findings are crucial for improving the quality and reliability of satellite data.

[0095] Furthermore, to more comprehensively evaluate the performance of FY-4B / GIIRS, this application also compared observation data with that of IASI, a similar polar-orbiting satellite. To ensure that the observations between GIIRS and IASI were conducted simultaneously, a sub-satellite point intersection (SNO) sample was selected that met strict criteria, specifically:

[0096] The GIIRS L1-level observational data and the IASI observational data were selected with a measurement time difference of less than m2 minutes and a relative difference of less than a% between the zenith angles of the two instruments. These conditions ensured that the two instruments collected data under similar observing conditions, thus enhancing the reliability of the comparison results. Due to the differences in the fields of view between the GIIRS and IASI, data acquired by the AGRI onboard the FY-4B within a central wavelength range of 10.30 to 11.30 μm were used to select a spatially uniform observation area. This selection helped minimize the impact of field of view differences on the comparison results, resulting in more accurate analysis.

[0097] More specifically, the criteria for determining a spatially uniform observation area are as follows: a circle is drawn in the observation area with the center of the pixel of the geostationary satellite hyperspectral infrared detector observation data as the center and half of the spatial resolution distance as the radius. All the time-space-matched geostationary orbit radiation imager observation data within the circle constitute a judgment data group. If the ratio of the standard deviation to the mean of the judgment data group is less than 0.01, then the circle is considered to be a spatially uniform observation area.

[0098] In this embodiment, the measurement time difference is less than 10 minutes, and the relative difference between the zenith angles is less than 1%.

[0099] Furthermore, in the GIIRS-IASI comparison process, this application uses a series of data processing to process the IASI observation data to ensure the accuracy and consistency of the results, such as Figure 3 As shown, the steps include:

[0100] Inverse Fourier transform: First, the raw spectral data of IASI are subjected to inverse Fourier transform to convert the frequency domain data back into the first interferogram.

[0101] De-Gaussian apodization: In the first interferogram, de-Gaussian apodization is applied to reduce the spectral distortion caused by the spectrometer system response function.

[0102] Truncated interferogram: The first interferogram is truncated to remove high-noise components from the signal, retaining the primary signal information while reducing the spectral resolution to produce a second interferogram. In this embodiment, the first interferogram is truncated to a path length difference of 0.8 cm, and the spectral resolution is reduced to that of a geostationary satellite hyperspectral infrared detector.

[0103] Fourier transform: Then, Fourier transform is performed to convert the second interferogram back into frequency domain data to obtain processed spectral data.

[0104] Hamming window apodization: Finally, the spectral data is further smoothed using Hamming window apodization to reduce sidelobe effects.

[0105] Furthermore, the processed IASI observation data and GIIRS observation data were cross-matched and verified, specifically:

[0106] The processed IASI observation data with higher spectral resolution were used as reference data and compared with the FY-4B / GIIRS observation data. The systematic deviations were calculated and analyzed after obtaining the systematic deviations. This process is consistent with the comparison process of the GIIRS-RTTOV data mentioned above. The difference is that the reference dataset is replaced, so it will not be repeated here.

[0107] Figure 4The results of the first year after the launch of FY-4B (July 2022-June 2023) for each FY-4B / GIIRS channel average noise equivalent temperature difference (GIIRS NEdT), statistical mean bias with reference to IASI observation data (GIIRS-IASIBias), statistical standard deviation with reference to IASI observation data (GIIRS-IASI STD), statistical mean bias with reference to simulated radiation data (GIIRS-RT Bias), and statistical standard deviation with reference to simulated radiation data (GIIRS-RT STD) are shown. The ordinate represents the brightness temperature bias / standard deviation (unit: K), the abscissa represents the spectral center wavenumber of each group of spectral channels, "Bias" represents bias, "STD" represents standard deviation, "Wave Length" represents wavelength, "Wave Number" represents wavenumber, and "NedT" represents equivalent noise temperature difference.

[0108] Figure 5 The following figure shows the time series of the daily average deviation between the observed brightness temperature (BT) and the simulated brightness temperature (OB) during the first year after the FY-4B launch, as well as the performance of the corresponding fitting line under different channels: Figure 5 Middle (a) 703.125 cm -1 , (b) 775.0cm -1 , (c) 1807.5 cm -1 , and (d) 2022.5 cm -1 aisle, Figure 5 (e) shows the slope and intercept of the bias-time fitting line for all FY-4B / GIIRS channels. These linear fitting results provide a systematic bias correction formula for each channel. Figure 5 In (a), (b), (c), and (d), the horizontal axis represents the number of days, and the vertical axis represents the daily average deviation; Figure 5 In (e), the horizontal axis represents the central wavenumber of the spectrum of each channel group, and the vertical axis represents the deviation slope and deviation intercept.

[0109] Figure 6 An error bar graph is shown, where the horizontal axis represents the hour and the vertical axis represents the bias, illustrating the daytime GIIRS observation and background bias. In the four selected spectral channels, the average bias at 1807.5 cm⁻¹ ranges from -0.41 to 0.75 K ( Figure 6 (c)), while the range at 2022.5 cm⁻¹ is -0.57 to 0.63 K ( Figure 6 (d) shows that the daily variation of the deviation is relatively small. However, in the two spectral channels of 703.125 cm⁻¹ and 775.0 cm⁻¹ ( Figure 6 In (a) and (b), the mean biases range from -1.48 to 0.77 K and -1.19 to -0.11 K, respectively, indicating relatively large biases or uncertainties in these two spectral channels. Furthermore, the diurnal variation of the biases in the two longwave infrared (LWIR) sample channels of the FY-4B / GIIRS is not significant. In contrast, the mean biases in the two midwave infrared (MWIR) channels exhibit a slow diurnal variation, gradually increasing from 18:00 to 05:00 UTC and then decreasing from 06:00 to 17:00 UTC.

[0110] Figure 7 The relationship between the GIIRS bias and the observed brightness temperature of four selected spectral channels and the correlation coefficient are shown. The distribution of the GIIRS-RTTOV bias and the observed brightness temperature in the range of 0.1K and 0.5K is analyzed in detail, covering Figure 7 The four spectral channels are (a) 703.125 cm⁻¹, (b) 775.0 cm⁻¹, (c) 1807.5 cm⁻¹, and (d) 2022.5 cm⁻¹. In addition, Figure 7 Panel (e) shows the correlation coefficients between the GIIRS biases and the observed brightness temperature for all FY-4B / GIIRS channels. The results show that the GIIRS observed biases remain relatively stable across the dynamic range of -4 to 4 K across different spectral channels. Furthermore, the correlation coefficient between the BT and the biases of the FY-4B / GIIRS is less than 0.83. With the exception of spectral channels in the wavenumber ranges of 691.25 cm⁻¹ to 708.125 cm⁻¹ and 2228.125 cm⁻¹ to 2250 cm⁻¹, the calibration stability of the FY-4B / GIIRS is relatively unaffected by the observed brightness temperature.

[0111] In practice, these statistical results can be used to linearly compensate the L1 observation data, thereby achieving calibration of the FY-4B / GIIRS observation data. Specifically, by linearly compensating the deviations of each channel, systematic errors caused by variations in detector performance and environmental factors can be effectively reduced. This method is not only simple and efficient, but also significantly improves data accuracy, making it more suitable for subsequent quantitative analysis and applied research. In combination with the latest high-quality data and multi-source reference information, the method of the present invention ensures the accuracy and reliability of the calibration process, providing a scientific basis for further optimizing detector performance.

[0112] This paper analyzes the bias characteristics of the FY-4B / GIIRS satellite during its first year using the GIIRS-RTTOV and GIIRS-IASI datasets (GIIRS-RTTOV represents the statistical analysis of GIIRS observational data errors using simulated radiance data as a reference, and GIIRS-IASI represents the statistical analysis of GIIRS observational data errors using IASI observation data as a reference). The GIIRS-IASI dataset is primarily used to analyze the variations of biases with wavelength and time. However, the limited availability of IASI data, with only one or two tracks intersecting GIIRS daily, results in an insufficient sample size for analyzing the BT-Bias (brightness temperature-bias) relationship and diurnal variations. This limitation makes comprehensive analysis based solely on IASI data difficult. This paper comprehensively evaluates the bias characteristics of the FY-4B / GIIRS satellite by comparing L1-level observational data with reference data (including RTTOV simulated radiance data and IASI observational data). The use of clear-sky samples ensures high accuracy in RT simulations, and rigorous data quality control measures guarantee the reliability of the analysis results. Despite the limited availability of IASI data, the large sample size provided by the RTTOV simulation enables a more detailed analysis, including the diurnal and temporal variations of the biases. These results provide important scientific support for the on-orbit radiometric calibration performance of the FY-4B / GIIRS and lay the foundation for future quantitative data applications.

[0113] It should be noted that, for the sake of convenience, the aforementioned method embodiments are all expressed as a series of action combinations, but those skilled in the art should know that the present invention is not limited to the described order of actions, because according to the present invention, certain steps can be performed in other orders or simultaneously.

[0114] Based on the same concept as the calibration performance evaluation and analysis method for a geostationary satellite hyperspectral infrared detector in the above-mentioned embodiment, the present invention also provides a calibration performance evaluation and analysis system for a geostationary satellite hyperspectral infrared detector, which can be used to execute the calibration performance evaluation and analysis method for a geostationary satellite hyperspectral infrared detector. For ease of explanation, the structural diagram of the embodiment of the calibration performance evaluation and analysis system for a geostationary satellite hyperspectral infrared detector only shows the parts related to the embodiment of the present invention. Those skilled in the art will understand that the illustrated structure does not constitute a limitation of the device, and may include more or fewer components than shown, or combine certain components, or arrange the components differently.

[0115] like Figure 8 As shown, another embodiment of the present invention provides a calibration performance evaluation and analysis system for a geostationary satellite hyperspectral infrared detector, including a radiation data simulation module, a deviation calculation and analysis module, and a cross-matching verification module;

[0116] Among them, the radiation data simulation module is used to obtain the ERA5 dataset and the L1-level observation data of the geostationary satellite hyperspectral infrared sounder, and input them into the rapid transmission radiation model to obtain simulated radiation data; the ERA5 dataset includes the ERA5 temperature and humidity profiles and the ERA5 surface temperature; the observation data of the geostationary satellite hyperspectral infrared sounder includes the ground infrared emissivity, sea and land identification and digital elevation model; RTTOV simulates the land and sea areas under clear sky conditions to obtain simulated radiation data.

[0117] The deviation calculation and analysis module is used to compare the simulated radiation data with the L1-level observation data of the geostationary satellite hyperspectral infrared sounder, calculate the systematic deviation and perform deviation analysis;

[0118] The cross-matching verification module is used to cross-match and verify the L1 level observation data of the geostationary satellite hyperspectral infrared sounder with the observation data of the polar-orbiting satellite infrared atmospheric sounding interferometer.

[0119] It should be noted that the calibration performance evaluation and analysis system of the geostationary satellite hyperspectral infrared detector of the present invention corresponds one-to-one to the calibration performance evaluation and analysis method of the geostationary satellite hyperspectral infrared detector of the present invention. The technical features and beneficial effects described in the embodiment of the calibration performance evaluation and analysis method of the above-mentioned geostationary satellite hyperspectral infrared detector are applicable to the embodiment of the calibration performance evaluation and analysis system of the geostationary satellite hyperspectral infrared detector. For specific contents, please refer to the description in the embodiment of the method of the present invention. No further details will be given here. This is hereby declared.

[0120] In addition, in the implementation of the calibration performance evaluation and analysis system of the geostationary satellite hyperspectral infrared detector in the above-mentioned embodiment, the logical division of each program module is only an example. In actual application, the above-mentioned functions can be assigned to different program modules as needed, for example, for the configuration requirements of the corresponding hardware or the convenience of software implementation. That is, the internal structure of the calibration performance evaluation and analysis system of the geostationary satellite hyperspectral infrared detector is divided into different program modules to complete all or part of the functions described above.

[0121] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0122] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be considered as equivalent replacement methods and are included in the scope of protection of the present invention.

Claims

1. A calibration performance evaluation and analysis method for a geostationary satellite hyperspectral infrared detector, characterized in that: The steps include: The ERA5 dataset and L1-level observation data from a geostationary satellite hyperspectral infrared sounder were obtained and input into a rapid transfer radiation model to obtain simulated radiation data. The ERA5 dataset includes ERA5 temperature and humidity profiles and ERA5 surface temperature. The geostationary satellite hyperspectral infrared sounder observation data includes infrared channel observation brightness temperature, observation angle, ground infrared emissivity, land and sea identification, and digital elevation model. The rapid transfer radiation model was simulated over land and sea areas under clear sky conditions. The simulated radiation data were compared with the L1 observation data of the geostationary satellite hyperspectral infrared sounder, and the systematic deviation was calculated and analyzed. Cross-match and verify the L1 observation data of the geostationary satellite hyperspectral infrared sounder with the observation data of the polar-orbiting satellite infrared atmospheric sounding interferometer; When performing the cross-matching verification, the observation data of the polar-orbiting satellite infrared atmospheric sounding interferometer are processed in the following steps: Perform inverse Fourier transform on the raw spectral data of the polar-orbiting satellite infrared atmospheric sounding interferometer to convert the frequency domain data into the first interferogram; The de-Gaussian apodization technique is applied to reduce the spectral distortion in the first interferogram, and the first interferogram is truncated to remove the high noise part in the signal and reduce the spectral resolution to obtain the second interferogram; Performing Fourier transform on the second interferogram and reconverting it back to frequency domain data to obtain processed spectral data; The spectral data is smoothed using a Hamming window to reduce the sidelobe effect.

2. The calibration performance evaluation and analysis method for a geostationary satellite hyperspectral infrared detector according to claim 1 is characterized in that: The method for determining the land and sea areas under the clear sky conditions is as follows: Cloud images are obtained using geostationary radiation imagers carried by geostationary satellites; Divide the cloud image into multiple sub-image areas of x*x pixels; Determine whether each pixel in the sub-image area is covered by clouds. If not, mark the center point of the sub-image area as clear sky.

3. The calibration performance evaluation and analysis method for a geostationary satellite hyperspectral infrared detector according to claim 1 is characterized in that: During the simulation of the rapid transmission radiation model, data preprocessing operations are performed, specifically: Exclude L1-level observation data with a time difference of more than m1 minutes compared with the ERA5 dataset, and exclude L1-level observation data with a satellite zenith angle of more than 30°; L1-level observation data whose deviation between simulated radiation data and L1-level observation data exceeds n times the standard deviation are regarded as outliers and are eliminated.

4. The calibration performance evaluation and analysis method for a geostationary satellite hyperspectral infrared detector according to claim 1 is characterized in that: The calculation of systematic deviation is specifically: Use the simulated radiation data as reference data; The correspondence between the reference data and the L1-level observation data is determined through spatiotemporal matching: the reference data is interpolated based on the spatiotemporal information of the L1-level observation data to obtain the first fused data; during the interpolation process, linear interpolation is used for time, and spline interpolation is used for spatial latitude and longitude; The first fusion data is grouped according to the spectral channel based on the spectral center wavenumber to obtain multiple groups of sub-data; each group of sub-data contains L1 observation data and reference data under the corresponding spectral channel; The overall mean and standard deviation of each group of sub-data in each spectral channel are calculated, and the mean and standard deviation of the deviation between the L1-level observation data and the reference data in each group of sub-data are also calculated.

5. The calibration performance evaluation and analysis method for a geostationary satellite hyperspectral infrared detector according to claim 4 is characterized in that: The deviation analysis refers to performing statistical analysis on each spectral channel; The statistical analysis includes the following four parts: Annual analysis: With a time range of one year, the mean and standard deviation of the deviation between the brightness temperature of the L1 observation data and the reference data are calculated each year; Daily analysis: Taking the day as the time range, calculate the average value of the deviation between the brightness temperature of the daily L1 observation data and the brightness temperature of the reference data, and analyze the change pattern of the average value with the date; Hourly analysis: Using the hourly time range, calculate the average deviation between the brightness temperature of the L1 observation data and the reference data every hour. Group the statistical deviation results at the same time on different days into one group, and calculate their average and standard deviation. Overall deviation analysis: Taking the year as the time range, calculate the deviation between the brightness temperature of the L1-level observation data of all spectral channels and the brightness temperature of the reference data in each year, make a deviation scatter plot, and calculate the correlation coefficient between the brightness temperature of the L1-level observation data of all channels and the brightness temperature of the reference data.

6. The calibration performance evaluation and analysis method for a geostationary satellite hyperspectral infrared detector according to claim 1 is characterized in that: Before the cross-matching verification, the L1 level observation data of the geostationary satellite hyperspectral infrared sounder and the observation data of the polar-orbiting satellite infrared atmospheric sounding interferometer are screened to select sub-satellite point cross samples that meet the standards, specifically: Select the L1-level observation data of the geostationary satellite hyperspectral infrared sounder and the observation data of the polar-orbiting satellite infrared atmospheric sounding interferometer with a measurement time difference of less than m2 minutes and a relative difference between the zenith angles of less than a%; Use the geostationary orbit radiation imager carried by the geostationary satellite to obtain data in the central wavelength range of 10.30 to 11.30 μm to select the spatially uniform observation area; The criteria for determining a spatially uniform observation area are as follows: a circle is drawn in the observation area with the center of the pixel of the observation data of the geostationary satellite hyperspectral infrared detector as the center and half of the spatial resolution distance as the radius. All the time-space-matched geostationary orbit radiation imager observation data within the circle constitute a determination data group. If the ratio of the standard deviation to the mean of the determination data group is less than 0.01, the circle is considered to be a spatially uniform observation area.

7. The calibration performance evaluation and analysis method for a geostationary satellite hyperspectral infrared detector according to claim 1 is characterized in that: The first interference pattern is truncated to an optical path difference of 0.8 cm; The spectral resolution is reduced to the spectral resolution of a geostationary satellite hyperspectral infrared detector.

8. The calibration performance evaluation and analysis method for a geostationary satellite hyperspectral infrared detector according to claim 1 is characterized in that: The cross-matching verification is specifically as follows: The observation data of the polar-orbiting satellite infrared atmospheric sounding interferometer is used as reference data; The correspondence between the reference data and the L1-level observation data is determined through spatiotemporal matching: the reference data is interpolated based on the spatiotemporal information of the L1-level observation data to obtain the second fused data; during the interpolation process, linear interpolation is used for time, and spline interpolation is used for spatial latitude and longitude; The second fusion data is grouped according to the spectral channel based on the spectral center wavenumber to obtain multiple groups of channel data; each group of channel data contains L1 observation data and reference data under the corresponding spectral channel; Calculate the overall mean and overall standard deviation of each set of channel data in each spectral channel, and calculate the mean and standard deviation of the deviation between the L1 observation data and the reference data in each set of channel data.

9. The calibration performance evaluation and analysis system for geostationary satellite hyperspectral infrared detectors is characterized by: It includes radiation data simulation module, deviation calculation and analysis module and cross-matching verification module; The radiation data simulation module is used to obtain the ERA5 dataset and L1-level observation data from the geostationary satellite hyperspectral infrared sounder, and input them into the rapid transmission radiation model to obtain simulated radiation data; the ERA5 dataset includes the ERA5 temperature and humidity profile and the ERA5 surface temperature; the observation data from the geostationary satellite hyperspectral infrared sounder includes infrared channel observation brightness temperature, observation angle, ground infrared emissivity, sea and land identification, and digital elevation model; the rapid transmission radiation model is simulated in land and sea areas under clear sky conditions; The deviation calculation and analysis module is used to compare the simulated radiation data with the L1 level observation data of the geostationary satellite hyperspectral infrared sounder, calculate the systematic deviation and perform deviation analysis; The cross-matching verification module is used to cross-match and verify the L1 level observation data of the geostationary satellite hyperspectral infrared sounder with the observation data of the polar-orbiting satellite infrared atmospheric sounding interferometer; When performing the cross-matching verification, the observation data of the polar-orbiting satellite infrared atmospheric sounding interferometer are processed in the following steps: Perform inverse Fourier transform on the raw spectral data of the polar-orbiting satellite infrared atmospheric sounding interferometer to convert the frequency domain data into the first interferogram; The de-Gaussian apodization technique is applied to reduce the spectral distortion in the first interferogram, and the first interferogram is truncated to remove the high noise part in the signal and reduce the spectral resolution to obtain the second interferogram; Performing Fourier transform on the second interferogram and reconverting it back to frequency domain data to obtain processed spectral data; The spectral data is smoothed using a Hamming window to reduce the sidelobe effect.

Images