GNSS-aided retrieval of precipitable water vapor from FY-3 satellite

Abstract

The application discloses a GNSS-assisted FY3 satellite PWV inversion method, relates to the technical field of GNSS meteorology and remote sensing water vapor inversion, and aims at the current situation that the atmospheric transmittance parameter is underestimated and the regression coefficient of water vapor and the atmospheric transmittance is selected empirically, which leads to low accuracy of the PWV product when the FY3 satellite MERSI sensor is used to invert the PWV.The application provides the following scheme, which comprises the following steps: S1, preprocessing FY3 satellite (FY3) MERSI L1 data and GNSS data; S2, calculating the atmospheric transmittance; S3, estimating the regression coefficient and calculating the channel water vapor; S4, performing weighted average summation on the multi-channel PWV; and S5, introducing a digital elevation model (DEM) to further correct the PWV inversion deviation of the FY satellite.The application has the advantages of reasonable design, high-precision PWV provided by the GNSS, PWV inversion of high-spatial-resolution FY3 / MERSI L1 data, and fast acquisition of high-precision high-spatial-temporal-resolution PWV of the GNSS-assisted FY3 satellite.

Description

Technical Field

[0001] This invention relates to the fields of GNSS meteorology and remote sensing water vapor inversion technology, and in particular to a GNSS-assisted PWV inversion method for the Fengyun-3 satellite. Background Technology

[0002] Water vapor is a crucial component of the terrestrial atmosphere, primarily found in the lower troposphere. Although it constitutes only a small portion of the atmosphere, it varies significantly in both space and time. During phase transitions, it releases or absorbs substantial amounts of latent heat, thus playing a vital role in global energy balance, climate change processes, and the formation and evolution of severe weather events. Precipitable water vapor (PWV) refers to the amount of precipitation formed when all water vapor in a unit cross-sectional area of ​​air column from the ground to the top of the atmosphere condenses into rain. This indicator is commonly used to quantify the water vapor content in the atmosphere. PWV is an important research area in GNSS meteorology and has been widely applied in short-term weather warnings and long-term climate monitoring. Currently, various PWV retrieval techniques have been developed, including numerical weather prediction, satellite remote sensing, and radio sounding. However, these techniques still suffer from low spatial coverage, missing time series data, and poor product accuracy. Therefore, utilizing a multi-technology fusion method for water vapor retrieval is of great significance for obtaining high-precision water vapor data. High-precision PWV (partial water vapor) acquired by GNSS is often used to validate PWV acquisition by other technologies. However, due to the uneven distribution of GNSS stations and their low spatial resolution, it is difficult to obtain the distribution of PWV across the entire area. Compared with GNSS technology represented by stations, satellite remote sensing technology has advantages in acquiring PWV, such as wider coverage and higher temporal resolution. For example, the Medium Resolution Spectral Imager (MERSI) on the Fengyun-3 satellite can provide atmospheric water vapor information with a spatial resolution of up to 1×1 km. However, when inverting PWV, there are defects such as underestimation of atmospheric transmittance parameters and empirical selection of regression coefficients between water vapor and atmospheric transmittance. This makes it impossible to achieve rapid acquisition of high-precision, high-temporal-resolution PWV, and thus difficult to meet the application requirements of high-precision PWV. Summary of the Invention

[0003] The purpose of this invention is to address the current situation where the atmospheric transmittance parameter is underestimated and the regression coefficients of water vapor and atmospheric transmittance are selected empirically when inverting PWV using the MERSI sensor of the Fengyun-3 satellite, resulting in low accuracy of PWV products. This invention proposes a GNSS-assisted PWV inversion method for FY3.

[0004] To achieve the above objectives, the present invention adopts the following technical solution:

[0005] The GNSS-assisted PWV inversion method for Fengyun-3 satellite includes the following steps:

[0006] S1: Preprocessing MERSI L1 data and GNSS data from Fengyun-3 (FY3) satellite;

[0007] S2: Calculation of atmospheric transmittance;

[0008] S3: Estimation of regression coefficients and calculation of channel water vapor;

[0009] S4: Weighted average summation of PWV across multiple channels;

[0010] S5: Introduce a Digital Elevation Model (DEM) to further correct the bias of the PWV retrieved from the Fengyun satellite.

[0011] In a preferred embodiment, S1 includes:

[0012] S1.1: Cloud removal processing of FY3 / MERSI L1 channel data: Cloud removal is achieved using the Cloud_Mask attribute in the FY3-L2 PWV product. The cloud mask data is a six-byte integer array whose contents are stored bit by bit. The integer array is converted into a binary array, and then a cloud mask is generated based on the relationship between the bits and the corresponding clouds to achieve cloud removal. Refer to the attached manual. Figure 2 Instruction manual attached Figure 2 The contents of the FY3 / MERSI cloud detection array bit are given.

[0013] S1.2: Spatiotemporal matching of GNSS stations and FY3 / MERSI L1 data: The average value of the latitude and longitude of the GNSS stations corresponding to a certain range of FY3 / MERSI L1 data is taken as the matching data between the stations and FY3 / MERSI L1 data; in addition, since the transit times of GNSS data and FY3 / MERSI L1 data are different, time matching needs to be performed on them according to a certain time range; the horizontal distance range and time range are 0.15°. 0.15° and 30 minutes.

[0014] In a preferred embodiment, the near-infrared channel transmittance calculation in S2 is performed using the two-channel ratio method and the three-channel ratio method. The difference between the two methods lies in whether the remote sensing ground object is a single type. For a single ground object, the two-channel ratio method is used as shown in equation (1), while for a complex ground object, the three-channel ratio method is used as shown in equation (2).

[0015]

[0016]

[0017] In the formula, Atmospheric transmittance, Three water vapor absorption channels for MERSI L1 data ( Reflectivity of (=17, 18, 19) and The reflectance of the two window channels of MERSI. and The values ​​are 0.8 and 0.2 respectively.

[0018] In a preferred embodiment, the regression fitting relationship between atmospheric transmittance and PWV in S3 is as shown in equation (3):

[0019] (3)

[0020] In the formula, This refers to the amount of atmospheric precipitable water calculated based on each channel. and These are the fitting coefficients;

[0021] Based on the above exponential relationship between atmospheric transmittance and atmospheric precipitable water, this invention introduces high-precision GNSS PWV as a fitting model parameter and takes into account seasonal factors to calculate the coefficients. The specific formula is as shown in equation (4):

[0022]

[0023] In the formula, PWV provided by GNSS stations for different seasons , These are the fitting coefficients obtained through the polynomial fitting function;

[0024] The channel water vapor values ​​of different channels on each grid point of FY3 / MERSI L1 are calculated using the obtained fitting coefficients, and the specific formula is shown in equation (5):

[0025]

[0026] In the formula, For the first Channel moisture value, , .

[0027] In a preferred embodiment, the three near-infrared absorption channels of the MERSI sensor have different sensitivities to water vapor. Absorption channel 18 is more sensitive to dry atmosphere, while the sensitivity of absorption channels 17 and 19 changes gradually with the increase of water vapor content. However, absorption channel 17 is slightly more sensitive to water vapor. Therefore, the transmittance of the three absorption channels can represent the magnitude of radiation attenuation caused by water vapor and be used to calculate the weights. Then, the weighted summation is performed to calculate the weighted average value of water vapor in the three absorption channels to obtain a more accurate water vapor inversion result.

[0028] The formula for calculating the weight of each channel is shown in equation (6):

[0029]

[0030] In the formula, For the first The water vapor weights corresponding to each channel need to be normalized, and the normalized weights are as follows:

[0031]

[0032] In the formula, For the first The normalized water vapor weights for each channel. Therefore, the final PWV obtained by inverting the water vapor values ​​and water vapor weights obtained above is:

[0033] PWV L1 =f 17 PWV 17 +f 18 PWV 18 +f 19 PWV 19 (8)

[0034] In the formula, FY3-L1 PWV for GNSS-assisted FY3 / MERSI L1 inversion.

[0035] In a preferred embodiment, the bias of the PWV corresponding to the GNSS station and the FY3-L1 PWV is first calculated, and a multivariate quadratic polynomial is constructed based on the GNSS station location and elevation. The calculation formula is as shown in equation (9):

[0036]

[0037] in, For deviation, These are the polynomial fitting coefficients. , and These are the latitude, longitude, and elevation of the GNSS station, respectively. Number of GNSS stations;

[0038] Then, using the grid latitude and longitude data and DEM grid elevation data provided by FY3 / MERSI L1, based on the above... The fitting coefficients are used to calculate the grid deviation value, and finally the deviation is compared with the value obtained above. The sums are used to obtain the final PWV.

[0039] In this invention, the GNSS-assisted PWV inversion method for Fengyun-3 satellite introduces model regression coefficients for accurate estimation of PWV and atmospheric transmittance by GNSS. In addition, considering the influence of seasonal factors on PWV, more accurate model regression coefficients are constructed for each season.

[0040] In this invention, the GNSS-assisted PWV inversion method for Fengyun-3 satellite uses regression coefficients to calculate the water vapor value of each absorption channel and performs a weighted summation of the PWV values ​​of different channels to obtain the inverted PWV. Finally, considering the strong correlation between surface elevation and atmospheric water vapor, a digital elevation model is introduced to correct the deviation of the inverted PWV, resulting in a more accurate PWV inversion result.

[0041] This invention is rationally designed and uses a multi-technology combined PWV inversion mode. By leveraging the complementary advantages of the high spatial resolution of FY3 / MERSI L1 data and the high precision and high temporal resolution of GNSS PWV data, it achieves rapid acquisition of high-precision, high-temporal-resolution PWV data from the FY3 satellite with the assistance of GNSS. Attached Figure Description

[0042] Figure 1 This is a flowchart illustrating the steps of the GNSS-assisted PWV inversion method for Fengyun-3 satellite proposed in this invention.

[0043] Figure 2 This diagram shows the storage content of the FY3 / MERSI cloud detection array bits in the GNSS-assisted large PWV inversion method for Fengyun-3 satellite proposed in this invention. Detailed Implementation

[0044] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0045] Reference Figure 1-2 This solution provides one embodiment: a GNSS-assisted PWV inversion method for Fengyun-3 satellites, comprising the following steps:

[0046] S1: Preprocessing MERSI L1 data and GNSS data from Fengyun-3 (FY3) satellite;

[0047] S2: Calculation of atmospheric transmittance;

[0048] S3: Estimation of regression coefficients and calculation of channel water vapor;

[0049] S4: Weighted average summation of PWV across multiple channels;

[0050] S5: Introduce a Digital Elevation Model (DEM) to further correct the bias of the PWV retrieved from the Fengyun satellite.

[0051] In this embodiment, S1 includes:

[0052] S1.1: Cloud removal processing of FY3 / MERSI L1 channel data: Cloud removal is achieved using the Cloud_Mask attribute in the FY3-L2 PWV product. The cloud mask data is a six-byte integer array whose contents are stored bit by bit. The integer array is converted into a binary array, and then a cloud mask is generated based on the relationship between the bits and the corresponding clouds to achieve cloud removal. Refer to the attached manual. Figure 2 Instruction manual attached Figure 2 A diagram showing the contents of the FY3 / MERSI cloud detection array's bit positions is provided.

[0053] S1.2: Spatiotemporal matching of GNSS stations and FY3 / MERSI L1 data: The average value of the latitude and longitude of the GNSS stations corresponding to a certain range of FY3 / MERSI L1 data is taken as the matching data between the stations and FY3 / MERSI L1 data; in addition, since the transit times of GNSS data and FY3 / MERSI L1 data are different, time matching needs to be performed on them according to a certain time range; the horizontal distance range and time range are 0.15°. 0.15° and 30 minutes.

[0054] In this embodiment, the near-infrared channel transmittance calculation in S2 is performed using the two-channel ratio method and the three-channel ratio method. The difference between the two methods lies in whether the remote sensing ground object is a single type. For a single ground object, the two-channel ratio method is used as shown in equation (1), while for a complex ground object, the three-channel ratio method is used as shown in equation (2).

[0055]

[0056]

[0057] In the formula, Atmospheric transmittance, Three water vapor absorption channels for MERSI L1 data ( Reflectivity of (=17, 18, 19) and The reflectance of the two window channels of MERSI. and The values ​​are 0.8 and 0.2 respectively.

[0058] In this embodiment, the regression fitting relationship between atmospheric transmittance and PWV in S3 is as shown in equation (3):

[0059] (3)

[0060] In the formula, This refers to the amount of atmospheric precipitable water calculated based on each channel. and These are the fitting coefficients;

[0061] Based on the above exponential relationship between atmospheric transmittance and atmospheric precipitable water, this invention introduces high-precision GNSS PWV as a fitting model parameter and takes into account seasonal factors to calculate the coefficients. The specific formula is as shown in equation (4):

[0062]

[0063] In the formula, PWV provided by GNSS stations for different seasons , These are the fitting coefficients obtained through the polynomial fitting function;

[0064] The channel water vapor values ​​of different channels on each grid point of FY3 / MERSI L1 are calculated using the obtained fitting coefficients, and the specific formula is shown in equation (5):

[0065]

[0066] In the formula, For the first Channel moisture value, , .

[0067] In this embodiment, the three near-infrared absorption channels of the MERSI sensor have different sensitivities to water vapor. Absorption channel 18 is more sensitive to dry atmosphere, while the sensitivity of absorption channels 17 and 19 changes slowly with the increase of water vapor. However, absorption channel 17 is slightly more sensitive to water vapor. Therefore, the transmittance of the three absorption channels can represent the magnitude of radiation attenuation caused by water vapor and be used to calculate the weights. Then, the weighted average value of water vapor in the three absorption channels is calculated by weighted summation to obtain a more accurate water vapor inversion result.

[0068] The formula for calculating the weight of each channel is shown in equation (6):

[0069]

[0070] In the formula, For the first The water vapor weights corresponding to each channel need to be normalized, and the normalized weights are as follows:

[0071]

[0072] In the formula, For the first The normalized water vapor weights for each channel. Therefore, the final PWV obtained by inverting the water vapor values ​​and water vapor weights obtained above is:

[0073] PWV L1 =f 17 PWV 17 +f 18 PWV 18 +f 19 PWV 19 (8)

[0074] In the formula, FY3-L1 PWV for GNSS-assisted FY3 / MERSI L1 inversion.

[0075] In this embodiment, the bias of the PWV corresponding to the GNSS station and the FY3-L1 PWV is first calculated, and a multivariate quadratic polynomial is constructed based on the GNSS station location and elevation. The calculation formula is as shown in equation (9):

[0076]

[0077] in, For deviation, These are the polynomial fitting coefficients. , and These are the latitude, longitude, and elevation of the GNSS station, respectively. Number of GNSS stations;

[0078] Then, using the grid latitude and longitude data and DEM grid elevation data provided by FY3 / MERSI L1, based on the above... The fitting coefficients are used to calculate the grid deviation value, and finally the deviation is compared with the value obtained above. The sums are used to obtain the final PWV.

[0079] The goal of the Fengyun-3 (FY3) satellite is to acquire three-dimensional, global, all-weather, quantitative, and high-precision data on the Earth's atmospheric environment. Its main tasks are: (1) to provide global meteorological parameters such as temperature, humidity, cloud radiation, etc. for weather forecasting, especially medium-range numerical weather prediction; (2) to monitor large-scale natural disasters and ecological environment; (3) to study global environmental changes, explore the laws of global climate change, and provide the geophysical parameters required for climate diagnosis and prediction; and (4) to provide global and regional meteorological information for military meteorology and professional meteorological services such as aviation and navigation.

[0080] The working principle is as follows: First, the FY3 / MERSI L1 data and GNSS data of the study area are preprocessed; then the atmospheric transmittance is calculated; then the regression coefficients are estimated and the channel water vapor is calculated; next, the weighted average summation of the multi-channel PWV is performed; finally, the PWV deviation correction of the DEM is introduced.

[0081] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

[0082] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A GNSS-assisted PWV inversion method for Fengyun-3 satellite, characterized in that, Includes the following steps: S1: Preprocessing MERSI L1 data and GNSS data from Fengyun-3 (FY3) satellite; S2: Calculation of atmospheric transmittance; S3: Estimation of regression coefficients and calculation of channel water vapor; S4: Weighted average summation of PWV across multiple channels; S5: Introduce a Digital Elevation Model (DEM) to further correct the bias in the PWV retrieved from the Fengyun satellite; The regression fitting relationship between atmospheric transmittance and PWV in S3 is shown in equation (3): In the formula, This refers to the amount of atmospheric precipitable water calculated based on each channel. and These are the fitting coefficients; Based on the above exponential relationship between atmospheric transmittance and atmospheric precipitable water, a high-precision GNSS PWV is introduced as a fitting model parameter, and the coefficients are calculated taking into account seasonal factors. The specific formula is shown in equation (4): In the formula, PWV provided by GNSS stations for different seasons , These are the fitting coefficients obtained through the polynomial fitting function; The channel water vapor values ​​of different channels on each grid point of FY3 / MERSI L1 are calculated using the obtained fitting coefficients, and the specific formula is shown in equation (5): In the formula, For the first Channel moisture value, , ; First, calculate the bias between the PWV and FY3-L1 PWV corresponding to the GNSS station, and then construct a multivariate quadratic polynomial based on the GNSS station location and elevation. The calculation formula is shown in equation (9): in, For deviation, These are the polynomial fitting coefficients. , and These are the latitude, longitude, and elevation of the GNSS station, respectively. Number of GNSS stations; Then, using the grid latitude and longitude data and DEM grid elevation data provided by FY3 / MERSI L1, based on the above... The fitting coefficients are used to calculate the grid deviation value, and finally the deviation is compared with the value obtained above. The sums are used to obtain the final PWV.

2. The GNSS-assisted PWV inversion method for Fengyun-3 satellite according to claim 1, characterized in that, S1 includes: S1.1: Cloud removal processing of FY3 / MERSI L1 channel data: Cloud removal processing is achieved by using the Cloud_Mask attribute in the FY3-L2 PWV product. The cloud mask data is a six-byte integer array whose contents are stored bit by bit. The integer array is converted into a binary array and then a cloud mask is generated according to the relationship between the bit and the corresponding cloud to achieve cloud removal. S1.2: Spatiotemporal matching of GNSS stations and FY3 / MERSI L1 data: The average value of the latitude and longitude of the GNSS stations corresponding to a certain range of FY3 / MERSI L1 data is taken as the matching data between the stations and FY3 / MERSI L1 data; in addition, since the transit times of GNSS data and FY3 / MERSI L1 data are different, time matching needs to be performed on them according to a certain time range; the horizontal distance range and time range are 0.15°. 0.15° and 30 minutes.

3. The GNSS-assisted PWV inversion method for Fengyun-3 satellite according to claim 1, characterized in that: The near-infrared channel transmittance in S2 is calculated using the two-channel ratio method and the three-channel ratio method. The difference between the two methods lies in whether the remote sensing ground object is a single type. For a single ground object, the two-channel ratio method is used as shown in equation (1), while for a complex ground object, the three-channel ratio method is used as shown in equation (2). In the formula, Atmospheric transmittance, Three water vapor absorption channels for MERSI L1 data ( Reflectivity of (=17, 18, 19) and The reflectance of the two window channels of MERSI. and The values ​​are 0.8 and 0.2 respectively.

4. The GNSS-assisted PWV inversion method for Fengyun-3 satellite according to claim 1, characterized in that: The three near-infrared absorption channels of the MERSI sensor have different sensitivities to water vapor. Absorption channel 18 is more sensitive to dry atmospheres, while the sensitivity of absorption channels 17 and 19 changes slowly with the increase of water vapor content. However, absorption channel 17 is slightly more sensitive to water vapor. Therefore, the transmittance of the three absorption channels can represent the magnitude of radiation attenuation caused by water vapor and be used to calculate the weights. Then, the weighted summation is performed to calculate the weighted average of water vapor in the three absorption channels to obtain a more accurate water vapor inversion result. The formula for calculating the weight of each channel is shown in equation (6): In the formula, For the first The water vapor weights corresponding to each channel need to be normalized, and the normalized weights are as follows: In the formula, For the first The normalized water vapor weights for each channel are used to calculate the final PWV obtained by inverting the channel water vapor values ​​and water vapor weights as described above: PWV L1 =f 17 PWV 17 +f 18 PWV 18 +f 19 PWV 19 (8) In the formula, FY3-L1 PWV for GNSS-assisted FY3 / MERSI L1 inversion.

Images