Ocean and polar region atmosphere weighted average temperature estimation method fusing GNSS-RO / RS / ERA5 multi-source data

Abstract

The invention relates to an ocean and polar region atmospheric weighted average temperature estimation method fusing GNSS-RO / RS / ERA5 multi-source data, and the method comprises the steps: obtaining GNSS radio occultation data, RS data and ERA5 data, and carrying out the data fusion; performing grid division on the fused data, calculating the atmospheric weighted average temperature of all window grid centers, and determining the weight of different data in the fused data, so as to calculate the atmospheric weighted average temperature of all window grid points; constructing a Tm fusion model, obtaining a model coefficient and an elevation correction coefficient of each grid point in the Tm fusion model, querying the model coefficients and the elevation correction coefficients of a plurality of grid points closest to the target point according to the position, the height and the time of the target point, inputting the model coefficients and the elevation correction coefficients of the plurality of grid points into the Tm fusion model, and acquiring the atmospheric weighted average temperature of the plurality of grid points, and further acquiring the atmospheric weighted average temperature of the target point.

Description

Technical Field

[0001] The present invention relates to the technical fields of the integration of GNSS-RO and RS technologies in meteorological applications and remote sensing technologies for the ocean and polar atmosphere, and particularly to a method for estimating the weighted mean temperature of the atmosphere in the ocean and polar regions by fusing multi-source data of GNSS-RO / RS / ERA5. Background Art

[0002] The weighted mean temperature of the atmosphere (T m ) is a key parameter for retrieving the atmospheric water vapor content by GNSS or GNSS-R, and its accuracy will directly affect the result of water vapor retrieval. The change of the weighted mean temperature T m of the atmosphere can indirectly affect the vertical stability of the atmosphere and the formation and evolution of weather systems. Therefore, high-precision T m information is crucial for meteorological research. Constructing a weighted mean temperature model of the atmosphere with a single data source has a small coverage range, low timeliness, and may have errors or biases, thus affecting the accuracy of constructing the weighted mean temperature model of the atmosphere. Fusing multi-source data to construct the weighted mean temperature of the atmosphere can improve the accuracy and reliability of the data, enhance the data coverage range and time resolution, improve the comprehensiveness and integrity of the data, reduce the uncertainty of data processing and analysis, and support a wider range of applications and research. These benefits contribute to a better understanding of climate change, improve the accuracy of meteorological forecasts, and provide a more reliable basis for decision-making. Therefore, it is of great significance to construct a weighted mean temperature fusion model of the atmosphere by fusing multi-source data.

[0003] The commonly used method for obtaining the weighted mean temperature T m of the atmosphere is to perform numerical integration on the temperature and humidity profiles obtained by radiosonde. However, in actual work, it is often difficult to obtain these profiles, and this method is easily restricted by the observation time and site distribution, resulting in the inability to obtain the weighted mean temperature T m of the atmosphere at any location in real time. Currently, according to whether ground meteorological data is used, two T m modeling methods have been developed: (1) establishing a T m model by finding the linear or non-linear relationship between the weighted mean temperature T s of the atmosphere and other ground meteorological data (such as surface temperature (T m ))). (2) Without using on-site meteorological data, but directly modeling and predicting T m using information such as time, latitude, and longitude. This model has good spatial and temporal resolutions and is more suitable for areas where in-situ surface data cannot be obtained. In the field of GNSS meteorology, many regional T m models have been developed to obtain better performance than global models, but mainly focus on land T mModeling. However, for ocean or polar regions, there is currently no publicly available high-precision T m model that can meet the requirements of high-precision and high spatio-temporal resolution atmospheric and water vapor inversion, and there is no literature reporting the construction of an atmospheric weighted mean temperature fusion model for ocean or polar regions by fusing multi-source data such as COSMIC-1 / 2 GNSS-RO / RS (FY-4A / 4B) and ERA5.

[0004] Therefore, there is an urgent need for a method for estimating the atmospheric weighted mean temperature in ocean and polar regions by fusing multi-source data of GNSS-RO / RS / ERA5 to solve the problems existing in the above-mentioned prior art. Summary of the Invention

[0005] To solve the problems existing in the above-mentioned prior art, the object of the present invention is to propose a method for estimating the atmospheric weighted mean temperature in ocean and polar regions by fusing multi-source data of GNSS-RO / RS / ERA5. Among them, for polar regions (Arctic and Antarctic), since COSMIC-2 GNSS-RO and FY-4A / 4B data do not have good coverage, COSMIC-1 GNSS-RO and ERA5 data are mainly used to construct a fusion model. For ocean regions, a fusion model is constructed using COSMIC-1 / 2 GNSS-RO / RS (FY-4A / 4B) and ERA5 data, showing high-precision T m estimation ability in ocean and polar regions (Antarctic and Arctic).

[0006] To achieve the above object, the present invention provides the following solutions:

[0007] A method for estimating the atmospheric weighted mean temperature in ocean and polar regions by fusing multi-source data of GNSS-RO / RS / ERA5, comprising:

[0008] Obtain GNSS radio occultation data, RS data and ERA5 data, and perform data fusion;

[0009] Perform grid division on the fusion data, and calculate the atmospheric weighted mean temperature at the center of all window grids and the atmospheric weighted mean temperature at all window grid points;

[0010] Construct a T m fusion model considering fine spatio-temporal variation characteristics, and obtain the model coefficients and elevation correction coefficients of each grid point in the T m fusion model. According to the position, height and time of the target point, query the model coefficients and elevation correction coefficients of multiple grid points closest to the target point, and input the model coefficients and elevation correction coefficients of the multiple grid points into the T m fusion model to obtain the atmospheric weighted mean temperature of multiple grid points;

[0011] Obtain the atmospheric weighted mean temperature of the target point based on the atmospheric weighted mean temperatures of the multiple grid points.

[0012] Optionally, data fusion of the RS data and the ERA5 data includes:

[0013] On the field of view of the RS data, interpolate the water vapor pressure for each pressure layer height according to the multiple ERA5 grid point data closest in distance, and interpolate the relative humidity and specific humidity values of the multiple ERA5 grid point data to the pressure layer height corresponding to the field of view temperature profile.

[0014] Optionally, obtaining the interpolation of the relative humidity and specific humidity values of the multiple ERA5 grid point data includes:

[0015]

[0016] In the formula, X S represents the relative humidity or specific humidity to be inserted at the pressure layer height corresponding to the field of view temperature profile; P S represents the air pressure at the pressure layer height corresponding to the field of view temperature profile; X 2i , X 1i represent the relative humidity or specific humidity corresponding to the adjacent two pressure layer heights of the ERA5 grid point; P 2i , P 1i represent the air pressures of the adjacent two layers of the ERA5 grid point, and i represents 4 ERA5 grid points closest to the field of view.

[0017] Optionally, interpolating the water vapor pressure for each pressure layer height includes:

[0018]

[0019] In the formula, P w represents the water vapor pressure; q represents the specific humidity; P represents the air pressure.

[0020] Optionally, calculating the height corresponding to each pressure layer includes:

[0021]

[0022] Among them, ΔH represents the height between adjacent two pressure layers; R d represents the specific gas constant of dry air; G represents the acceleration of gravity; t 1 and t 2 respectively represent the air temperatures of adjacent two pressure layers; represents the average air temperature between adjacent two pressure layers; U 1 and U 2 respectively represent the relative humidities of adjacent two pressure layers; P 1 , P2 Respectively represent the air pressure at two adjacent heights.

[0023] Optionally, calculating the atmospheric weighted average temperature at all window grid centers and the atmospheric weighted average temperature at all window grid points includes:

[0024] S1, calculate the atmospheric weighted average temperature of the field of view:

[0025]

[0026] Where, T m represents the weighted average temperature of the atmosphere; P wi represents the average water vapor pressure of the i-th layer; T i represents the average temperature of the i-th layer; Δh i represents the thickness of the i-th layer; N represents the number of pressure layers in the AVP field of view;

[0027] S2, gridding the fused data at a target spatial resolution according to the atmospheric weighted average temperature of the field of view;

[0028] S3, determine the geographical location of the target area. If the target area is an ocean area, slide the window in sequence, and calculate the T of all data points in the window while sliding the window. m The average is calculated and used as the weighted average temperature of the atmosphere at the center of the window grid. The GNSS radio occultation data is further used in combination with the Ikhonov regularization estimation method to determine the weights of different data and calculate the weighted average temperature of the atmosphere at the window grid point. If the target area is a polar region, the window is slid in sequence, and the T of all data points in the window is calculated while sliding the window. m The average is calculated and used as the weighted average temperature of the atmosphere at the center of the window grid, and the weighted average temperature of the atmosphere at the window grid points is further calculated.

[0029] Optionally, the T that takes into account the fine spatiotemporal variation characteristics m The fusion model includes: T m T at the mean elevation of the model and window m Model;

[0030] The T that takes into account the fine spatiotemporal variation characteristics m The fusion model expression is:

[0031]

[0032] In the formula, To take into account the nonlinear vertical variation of T m Model, is T at the average height of the window m Model.

[0033] Optionally, construct the T that takes into account the fine spatio-temporal variation characteristics m The fusion model includes:

[0034] Taking into account the time parameter and the spatial variation, T at the window-averaged elevation m The expression of the model is:

[0035]

[0036] In the formula, λ represents longitude; represents latitude; JD represents Julian day; HOD is the time of day; A 0 represents the constant term; A 1 、A 2 represent the longitude and latitude coefficients respectively; A 3 、A 4 represent the annual cycle coefficients; A 5 、A 6 represent the semi-annual cycle coefficients, A 7 、A 8 represent the daily cycle coefficients; A 9 、A 10 represent the semi-daily cycle coefficients;

[0037] Use the third-order polynomial function with non-linear variation between the T m value and the elevation to perform elevation correction. The relationship expression between the T m fusion model taking into account the fine spatio-temporal variation characteristics and the elevation is:

[0038] T m = a 1 + a 2 h + a 3 h 2 + a 4 h 3

[0039] In the formula, h represents the height at the target point; a 1 、a 2 、a 3 and a 4 are the 4 unknown coefficient parameters of the fitting function;

[0040] The expression of the T m model taking into account the non-linear vertical variation is:

[0041]

[0042] In the formula, h represents the height at the target point; b 2 、b 3 and b 4Four unknown coefficient parameters of the fitting function, H m is the window-averaged elevation;

[0043] Based on the T m model considering the non-linear vertical variation, and the T m model at the window-averaged elevation, the T m fusion model considering the fine spatio-temporal variation characteristics is obtained.

[0044] Optionally, the T m fusion model considering the fine spatio-temporal variation characteristics includes:

[0045]

[0046] where λ represents longitude; represents latitude; JD represents Julian day; HOD is the time of day; A 0 represents the constant term; A 1 , A 2 respectively represent the longitude and latitude coefficients; A 3 , A 4 represent the annual cycle coefficients; A 5 , A 6 represent the semi-annual cycle coefficients, A 7 , A 8 represent the daily cycle coefficients; A 9 , A 10 represent the semi-daily cycle coefficients, A 11 ~A 13 are the elevation correction coefficients.

[0047] The beneficial effects of the present invention are:

[0048] The present invention obtains GNSS radio occultation (COSMIC-1 / 2 GNSS-RO) data, RS (FY-4A / 4B) data and ERA5 data; fuses the FY-4A / 4B, COSMIC-1 / 2 GNSS-RO and ERA5 data; introduces a sliding window algorithm to calculate the T m at the center of all window grids for different data; uses the Tikhonov regularization estimation method to determine the weights of different data, and performs weighted fusion on different data to obtain the T m at the fused grid points; constructs a T m fusion model considering the fine spatio-temporal variation characteristics based on the GPT series model; uses the least squares algorithm to solve the model coefficients and elevation correction coefficients of each grid point in the T m fusion model; queries the four model grid point coefficients and elevation correction closest to the target point according to the position, height and time of the target point; utilizes the model coefficients and elevation correction provided by the grid points to calculate the Tm The fusion model calculates the T of each grid point m ; and obtains the T of the target point through bilinear interpolation m . The estimation method of the weighted average temperature (T m ) of the atmosphere in the ocean and polar regions that fuses multi-source data of GNSS-RO / RS / ERA5 can exhibit high-precision T m estimation ability in the ocean and polar regions (Antarctica and the Arctic), filling the T m data gap caused by sparse stations, inaccessible terrain, and adverse environmental conditions in the ocean and polar regions. It provides important data support for GNSS or GNSS-R meteorological research and atmospheric water vapor retrieval BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts

[0050] Figure 1 is a flowchart of the method for estimating the weighted average temperature of the atmosphere in the ocean and polar regions that fuses multi-source data of GNSS-RO / RS / ERA5 according to an embodiment of the present invention

[0051] Figure 2 is a schematic diagram of the principle of the sliding window algorithm introduced in an embodiment of the present invention

[0052] Figure 3 is a schematic diagram of the modeling calculation of the weighted average temperature of the atmosphere in the polar region based on multi-source data fusion according to an embodiment of the present invention; (a) is the T m of the modeling calculation of the weighted average temperature of the atmosphere in the Arctic region based on multi-source data fusion, and (b) is the T m, (c) is the annual cycle coefficient calculated by the weighted average temperature modeling of the Arctic region atmosphere based on multi-source data fusion, (d) is the annual cycle coefficient calculated by the weighted average temperature modeling of the Antarctic region atmosphere based on multi-source data fusion, (e) is the semi-annual cycle coefficient calculated by the weighted average temperature modeling of the Arctic region atmosphere based on multi-source data fusion, (f) is the semi-annual cycle coefficient calculated by the weighted average temperature modeling of the Antarctic region atmosphere based on multi-source data fusion, (g) is the daily cycle coefficient calculated by the weighted average temperature modeling of the Arctic region atmosphere based on multi-source data fusion, (h) is the daily cycle coefficient calculated by the weighted average temperature modeling of the Antarctic region atmosphere based on multi-source data fusion, (i) is the semi-daily cycle coefficient calculated by the weighted average temperature modeling of the Arctic region atmosphere based on multi-source data fusion, (j) is the semi-daily cycle coefficient calculated by the weighted average temperature modeling of the Antarctic region atmosphere based on multi-source data fusion;

[0053] Figure 4 is the weighted average temperature modeling calculation of the ocean region (between 50° south latitude and 50° north latitude) based on multi-source data fusion in the embodiment of the present invention; (a) is T of the weighted average temperature modeling calculation of the ocean region atmosphere based on multi-source data fusion m , (b) is the annual cycle coefficient calculated by the weighted average temperature modeling of the ocean region atmosphere based on multi-source data fusion, (c) is the semi-annual cycle coefficient calculated by the weighted average temperature modeling of the ocean region atmosphere based on multi-source data fusion, (d) is the daily cycle coefficient calculated by the weighted average temperature modeling of the ocean region atmosphere based on multi-source data fusion, (e) is the semi-daily cycle coefficient calculated by the weighted average temperature modeling of the ocean region atmosphere based on multi-source data fusion. Detailed implementation manners

[0054] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0055] To make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below with reference to the drawings and specific implementation manners.

[0056] Such as Figure 1As shown in the figure, this embodiment discloses a method for estimating the weighted average temperature of the atmosphere in the ocean and polar regions by fusing GNSS-RO / RS / ERA5 multi-source data, including: obtaining GNSS radio occultation data, RS data, and ERA5 data, and performing data fusion; dividing the fused data into grids, calculating the weighted average temperature of the atmosphere at the center of all window grids, determining the weights of different data in the fused data, and thus calculating the weighted average temperature of the atmosphere at all window grid points; constructing a T m fusion model that takes into account the fine spatio-temporal variation characteristics, and obtaining the model coefficients and elevation correction coefficients of each grid point in the T m fusion model. According to the position, altitude, and time of the target point, query the model coefficients and elevation correction coefficients of multiple grid points closest to the target point, and input the model coefficients and elevation correction coefficients of multiple grid points into the T m fusion model to obtain the weighted average temperature of the atmosphere at multiple grid points; based on the weighted average temperature of the atmosphere at multiple grid points, obtain the weighted average temperature of the atmosphere at the target point.

[0057] Specifically: This embodiment discloses a method for estimating the weighted average temperature of the atmosphere in the ocean and polar regions by fusing GNSS-RO / RS / ERA5 multi-source data, including the following steps:

[0058] Step S1, obtain GNSS radio occultation (COSMIC-1 / 2 GNSS-RO) data, RS (FY-4A / 4B) data, and ERA5 data;

[0059] Step S2, perform fusion processing on FY-4A / 4B, COSMIC-1 / 2 GNSS-RO, and ERA5 data;

[0060] Step S3, introduce a sliding window algorithm to calculate the T m at the center of all window grids for different data;

[0061] Step S4, use the Tikhonov regularization estimation method to determine the weights of different data, and perform weighted fusion on different data to obtain the T m at the fused grid points;

[0062] Step S5, based on the GPT series model, construct a T m fusion model that takes into account the fine spatio-temporal variation characteristics;

[0063] Step S6, use the least squares algorithm to solve the model coefficients and elevation correction coefficients of each grid point in the T m fusion model;

[0064] Step S7, query the coefficients of the 4 model grid points closest to the target point and the elevation correction according to the position, altitude, and time of the target point;

[0065] Step S8, according to the model coefficients and elevation corrections provided by the grid points, use the T m fusion model to calculate the T of each grid point m ;

[0066] Step S9, obtain the T of the target point through bilinear interpolation m .

[0067] Furthermore, the data fusion of GNSS radio occultation data, RS data, and ERA5 data includes: on the field of view of the RS data, according to multiple ERA5 grid point data with the closest distance, interpolate and calculate the water vapor pressure for each pressure layer height, and interpolate the relative humidity and specific humidity values of multiple ERA5 grid point data to the pressure layer height corresponding to the field of view temperature profile.

[0068] Furthermore, obtaining the interpolation of the relative humidity and specific humidity values of multiple ERA5 grid point data includes:

[0069]

[0070] In the formula, X S represents the relative humidity or specific humidity to be inserted at the pressure layer height corresponding to the field of view temperature profile; P S represents the air pressure at the pressure layer height corresponding to the field of view temperature profile; X 2i , X 1i represent the relative humidity or specific humidity corresponding to the adjacent two pressure layer heights of the ERA5 grid point; P 2i , P 1i represent the air pressures of the adjacent two layers of the ERA5 grid point, and i represents 4 ERA5 grid points closest to the field of view.

[0071] Furthermore, interpolating and calculating the water vapor pressure for each pressure layer height includes:

[0072]

[0073] In the formula, P w represents the water vapor pressure; q represents the specific humidity; P represents the air pressure.

[0074] Furthermore, calculating the height corresponding to each pressure layer includes:

[0075]

[0076] where, ΔH represents the height between adjacent two pressure layers; R d represents the specific gas constant of dry air; G represents the acceleration due to gravity; t 1 and t 2 respectively represent the air temperatures of adjacent two pressure layers; represents the average temperature between two adjacent pressure levels; U 1 and U 2 respectively represent the relative humidity between two adjacent pressure levels; P 1 、P 2 respectively represent the air pressures at two adjacent heights.

[0077] Specifically: Preferably, the fusion process of FY-4A / 4B, COSMIC-1 / 2 GNSS-RO and ERA5 data in step S2 is as follows:

[0078] S2.1, calculate the water vapor pressure, relative humidity and specific humidity at each pressure level height. On the FY-4A / 4B data field, according to the data of the four closest ERA5 grid points, interpolate and calculate the water vapor pressure for each pressure level height. Interpolate the relative humidity and specific humidity values of these grid points to the pressure level height corresponding to the field temperature profile. The calculation formulas are as follows:

[0079]

[0080] In the formula, X S represents the relative humidity or specific humidity to be inserted at the pressure level height corresponding to the field temperature profile; P S represents the air pressure at the pressure level height corresponding to the field temperature profile; X 2i 、X 1i represent the relative humidity or specific humidity corresponding to the adjacent two pressure level heights of the ERA5 grid points; P 2i 、P 1i represent the air pressures of the adjacent two layers of the ERA5 grid points, where i represents the 4 ERA5 grid points closest to the field. From formula (1), the relative humidity and specific humidity of the 4 ERA5 grid points at the pressure level height corresponding to the field temperature profile can be obtained, and then the relative humidity and specific humidity at each pressure level height of the field can be obtained by bilinear interpolation. Finally, the water vapor pressure at each pressure level height of the field is calculated according to the following formula:

[0081]

[0082] In the formula, P w represents the water vapor pressure, with the unit of hPa; q represents the specific humidity; P represents the air pressure, with the unit of hPa. S2.2, calculate the height corresponding to each pressure level of the field by cumulative calculation from bottom to top. The calculation formula is as follows

[0083]

[0084] In the formula, ΔH represents the height between two adjacent pressure levels; R d represents the specific gas constant of dry air, taking 287.05 J / (kg·K); G represents the acceleration of gravity, taking 9.80665 m / s2; t1 and t 2 Respectively represent the air temperature of two adjacent pressure layers (in °C); It represents the average temperature between two adjacent pressure layers (in °C); U 1 and U 2 Respectively represent the relative humidity of two adjacent pressure layers (%); P 1 , P 2 Respectively represent the air pressure at two adjacent heights (in hPa).

[0085] Further, calculating the atmospheric weighted average temperature at the centers of all window grids and the atmospheric weighted average temperature at all window grid points includes:

[0086] S1, calculate the atmospheric weighted average temperature of the field of view:

[0087]

[0088] Where, T m represents the weighted average temperature of the atmosphere; P wi represents the average water vapor pressure of the i-th layer; T i represents the average temperature of the i-th layer; Δh i represents the thickness of the i-th layer; N represents the number of pressure layers in the AVP field of view;

[0089] S2, gridding the fused data at the target spatial resolution according to the atmospheric weighted mean temperature of the field of view;

[0090] S3, determine the geographical location of the target area. If the target area is an ocean area, slide the window in sequence, and calculate the T of all data points in the window while sliding the window. m The average is calculated and used as the weighted average temperature of the atmosphere at the center of the window grid. The GNSS radio occultation data is further used in combination with the Ikhonov regularization estimation method to determine the weights of different data and calculate the weighted average temperature of the atmosphere at the window grid points. If the target area is the polar region, the window is slid in sequence, and the T of all data points in the window is calculated while sliding the window. m The average is calculated and used as the weighted average temperature of the atmosphere at the center of the window grid, and the weighted average temperature of the atmosphere at the window grid points is further calculated.

[0091] Specifically:

[0092] Step S3 introduces a sliding window algorithm to calculate the T of different data at the center of all window grids. m , is to solve the problem of T constructed by relying only on single grid point data. m The model estimation accuracy is low. The specific process of the algorithm is as follows: Figure 2 As shown, the implementation steps are as follows:

[0093] S3.1, grid the fusion data at a spatial resolution of 0.5°×0.5°, and the gridding result is i×j solid-line windows.

[0094] S3.2, calculate the mean value of T for all data points within the first window N 1,1 (regarded as a sliding window), and use it as the T of the blue grid point at the center of window N m . 1,1 m .

[0095] S3.3, move the window 1 grid point eastward in longitude, and solve the mean value of T for all data points within the new window N 1,2 , and use it as the T of the blue grid point at the center of window N m . By analogy, calculate the T of the center of each window 1,2 . m . m .

[0096] S3.4, form the grid of the T fusion model with the red solid points and dashed lines in each window, and the spatial resolution of the final model grid is 0.5°×0.5°. m

[0097] S3.5, calculate the T of the grid points using COSMIC-1 / 2 GNSS-RO data according to the methods in the above steps 2 and 3 m .

[0098] Preferably, step S4 uses the Tikhonov regularization estimation method to determine the weights of different data, and performs weighted fusion on different data to obtain the T of the fused grid points m to solve the problem of inconsistent accuracy of T calculated from different data m .

[0099] Furthermore, the T fusion model considering the fine spatio-temporal variation characteristics includes: the T m model considering the non-linear vertical variation and the T m model at the window average elevation; m

[0100] The expression of the T fusion model considering the fine spatio-temporal variation characteristics is: m

[0101]

[0102] In the formula, is the T m model considering the non-linear vertical variation, is the T m model at the window average elevation.

[0103] Furthermore, construct a T that takes into account fine spatio-temporal variation characteristics m The fusion model includes:

[0104] Taking into account the time parameter and also the variation in space, the T at the window-averaged elevation m The expression of the model is:

[0105]

[0106] In the formula, λ represents longitude; represents latitude; JD represents the Julian day; HOD is the time of day; A 0 represents the constant term; A 1 and A 2 respectively represent the longitude and latitude coefficients; A 3 and A 4 represent the annual cycle coefficients; A 5 and A 6 represent the semi-annual cycle coefficients. A 7 and A 8 represent the diurnal cycle coefficients; A 9 and A 10 represent the semi-diurnal cycle coefficients;

[0107] Use a third-order polynomial function with a non-linear variation between the T m value and the elevation for elevation correction. The relationship expression between the T m fusion model taking into account fine spatio-temporal variation characteristics and the elevation is:

[0108] T m = a 1 + a 2 h + a 3 h 2 + a 4 h 3

[0109] In the formula, h represents the height at the target point; a 1 and a 2 and a 3 and a 4 are the 4 unknown coefficient parameters of the fitting function;

[0110] The expression of the T m model taking into account non-linear vertical variation is:

[0111]

[0112] In the formula, h represents the height at the target point; b 2 and b 3 and b 4 are the 4 unknown coefficient parameters of the fitting function, Hm is the window average elevation;

[0113] Based on the T m model considering the non - linear vertical change and the T m model at the window average elevation, a T m fusion model that takes into account the fine spatio - temporal variation characteristics is obtained.

[0114] Furthermore, the T m fusion model includes:

[0115]

[0116] Among them, λ represents longitude; represents latitude; JD represents Julian day; HOD is the time of day; A 0 represents the constant term; A 1 and A 2 respectively represent the longitude - latitude coefficients; A 3 and A 4 represent the annual cycle coefficients; A 5 and A 6 represent the semi - annual cycle coefficients. A 7 and A 8 represent the diurnal cycle coefficients; A 9 and A 10 represent the semi - diurnal cycle coefficients, A 11 ~A 13 are the elevation correction coefficients.

[0117] Specifically: Step S5 includes the following sub - steps:

[0118] Considering the time parameters and also considering the changes in space (longitude - latitude), the expression of the model is:

[0119]

[0120] In the formula, λ represents longitude; represents latitude; JD represents Julian day; HOD is the time of day; A 0 represents the constant term; A 1 and A 2 respectively represent the longitude - latitude coefficients; A 3 and A 4 represent the annual cycle coefficients; A 5 and A 6 represent the semi - annual cycle coefficients. A 7 and A 8 represent the diurnal cycle coefficients; A 9 and A 10 represent the semi - diurnal cycle coefficients.

[0121] S5.2, for T in the vertical direction m The non-linear characteristics are very strong. Using a linear function for elevation correction results in a large error in the calculated T m problem. Using a third-order polynomial function with a significant non-linear variation between the T m value and elevation for elevation correction, the relationship expression between T m and elevation is

[0122] T m = a 1 + a 2 h + a 3 h 2 + a 4 h 3 (6)

[0123] In the formula, h represents the height at the target point; a 1 , a 2 , a 3 and a 4 are the 4 unknown coefficient parameters of the fitting function.

[0124] S5.3, therefore, the expression of the model is:

[0125]

[0126] S5.4, finally, the T m fusion model considering the fine spatio-temporal variation characteristics is:

[0127]

[0128] In the formula, A 11 ~A 13 are the elevation correction coefficients, H m is the window average elevation, and h is the target height.

[0129] Figure 3 This is the schematic diagram of the modeling calculation of the polar region atmospheric weighted average temperature based on multi-source data fusion in this embodiment;

[0130] Among them Figure 3 (a) is the T m for the modeling calculation of the Arctic region atmospheric weighted average temperature based on multi-source data fusion, Figure 3 (b) is the T m for the modeling calculation of the Antarctic region atmospheric weighted average temperature based on multi-source data fusion, Figure 3 (c) is the annual cycle coefficient for the modeling calculation of the Arctic region atmospheric weighted average temperature based on multi-source data fusion, Figure 3(d) is the annual cycle coefficient for the modeling calculation of the weighted average temperature of the Antarctic region's atmosphere based on multi-source data fusion, Figure 3 (e) is the semi-annual cycle coefficient for the modeling calculation of the weighted average temperature of the Arctic region's atmosphere based on multi-source data fusion, Figure 3 (f) is the semi-annual cycle coefficient for the modeling calculation of the weighted average temperature of the Antarctic region's atmosphere based on multi-source data fusion, Figure 3 (g) is the daily cycle coefficient for the modeling calculation of the weighted average temperature of the Arctic region's atmosphere based on multi-source data fusion, Figure 3 (h) is the daily cycle coefficient for the modeling calculation of the weighted average temperature of the Antarctic region's atmosphere based on multi-source data fusion, Figure 3 (i) is the semi-daily cycle coefficient for the modeling calculation of the weighted average temperature of the Arctic region's atmosphere based on multi-source data fusion, Figure 3 (j) is the semi-daily cycle coefficient for the modeling calculation of the weighted average temperature of the Antarctic region's atmosphere based on multi-source data fusion;

[0131] Figure 4 is the modeling calculation of the weighted average temperature of the ocean region (between 50° south and north latitudes) based on multi-source data fusion in this embodiment; Figure 4 (a) is the T for the modeling calculation of the weighted average temperature of the ocean region's atmosphere based on multi-source data fusion m , Figure 4 (b) is the annual cycle coefficient for the modeling calculation of the weighted average temperature of the ocean region's atmosphere based on multi-source data fusion, Figure 4 (c) is the semi-annual cycle coefficient for the modeling calculation of the weighted average temperature of the ocean region's atmosphere based on multi-source data fusion, Figure 4 (d) is the daily cycle coefficient for the modeling calculation of the weighted average temperature of the ocean region's atmosphere based on multi-source data fusion, Figure 4 (e) is the semi-daily cycle coefficient for the modeling calculation of the weighted average temperature of the ocean region's atmosphere based on multi-source data fusion.

[0132] The above embodiments are only descriptions of the preferred embodiments of the present invention and do not limit the scope of the present invention. Without departing from the design spirit of the present invention, various deformations and improvements made by those of ordinary skill in the art to the technical solutions of the present invention shall fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for estimating atmospheric weighted average temperature in ocean and polar regions by integrating GNSS-RO / RS / ERA5 multi-source data, characterized in that: include: Acquire GNSS radio occultation data, RS data and ERA5 data, and perform data fusion; Grid the fused data and calculate the atmospheric weighted average temperature at the center of all window grids and the atmospheric weighted average temperature at all window grid points; Constructing T that takes into account the fine spatiotemporal variation characteristics m Fusion model, and obtain each grid point in the T m The model coefficients and elevation correction coefficients in the fusion model are queried according to the position, height and time of the target point, and the model coefficients and elevation correction coefficients of multiple grid points closest to the target point are input into the T m Fusion model to obtain the atmospheric weighted average temperature of multiple grid points; Based on the weighted average temperatures of the atmosphere at the plurality of grid points, the weighted average temperature of the atmosphere at the target point is obtained.

2. The method for estimating the atmospheric weighted average temperature of ocean and polar regions by integrating GNSS-RO / RS / ERA5 multi-source data according to claim 1, characterized in that: The RS data and the ERA5 data are fused together, including: In the field of view of the RS data, water vapor pressure is interpolated and calculated for each pressure layer height according to the multiple ERA5 grid point data closest to it, and the relative humidity and specific humidity values ​​of the multiple ERA5 grid point data are interpolated to the pressure layer height corresponding to the field of view temperature profile.

3. The method for estimating the atmospheric weighted average temperature of ocean and polar regions by integrating GNSS-RO / RS / ERA5 multi-source data according to claim 2, characterized in that: Obtaining the relative humidity and specific humidity value interpolation of the multiple ERA5 grid point data includes: Where, X S Indicates the relative humidity or specific humidity that needs to be inserted at the pressure layer height corresponding to the field of view temperature profile; P S Indicates the air pressure at the pressure layer height corresponding to the field of view temperature profile; X 2i , X 1i It indicates the relative humidity or specific humidity at the two adjacent layers of air pressure height of ERA5 grid point; P 2i , P 1i It represents the air pressure of two adjacent layers of ERA5 grid points, and i represents the four ERA5 grid points closest to the field of view.

4. The method for estimating the atmospheric weighted average temperature of ocean and polar regions by integrating GNSS-RO / RS / ERA5 multi-source data according to claim 2, characterized in that: Interpolation of water vapor pressure for each pressure layer height includes: Where P w represents water vapor pressure; q represents specific humidity; P represents air pressure.

5. The method for estimating the atmospheric weighted average temperature of ocean and polar regions by integrating GNSS-RO / RS / ERA5 multi-source data according to claim 2, characterized in that: Calculating the height corresponding to each pressure layer includes: Where ΔH represents the height between two adjacent pressure layers; R d represents the dry air specific gas constant; G represents the gravitational acceleration; t1 and t2 represent the temperatures of two adjacent pressure layers respectively; It represents the average temperature between two adjacent pressure layers; U1 and U2 represent the relative humidity between two adjacent pressure layers; P1 and P2 represent the air pressure at two adjacent heights.

6. The method for estimating the atmospheric weighted average temperature of ocean and polar regions by integrating GNSS-RO / RS / ERA5 multi-source data according to claim 1, characterized in that: Calculating the atmospheric weighted average temperature at all window grid centers and the atmospheric weighted average temperature at all window grid points includes: S1, calculate the atmospheric weighted average temperature of the field of view: Where, T m represents the weighted average temperature of the atmosphere; P wi represents the average water vapor pressure of the i-th layer; T i represents the average temperature of the i-th layer; Δh i represents the thickness of the i-th layer; N represents the number of pressure layers in the AVP field of view; S2, gridding the fused data at a target spatial resolution according to the atmospheric weighted average temperature of the field of view; S3, determine the geographical location of the target area. If the target area is an ocean area, slide the window in sequence, and calculate the T of all data points in the window while sliding the window. m The average is calculated and used as the weighted average temperature of the atmosphere at the center of the window grid. The GNSS radio occultation data is further used in combination with the Ikhonov regularization estimation method to determine the weights of different data and calculate the weighted average temperature of the atmosphere at the window grid point. If the target area is a polar region, the window is slid in sequence, and the T of all data points in the window is calculated while sliding the window. m The average is calculated and used as the weighted average temperature of the atmosphere at the center of the window grid, and the weighted average temperature of the atmosphere at the window grid points is further calculated.

7. The method for estimating the atmospheric weighted average temperature of ocean and polar regions by integrating GNSS-RO / RS / ERA5 multi-source data according to claim 1, characterized in that: The T that takes into account the fine spatiotemporal variation characteristics m The fusion model includes: T m T at the mean elevation of the model and window m Model; The T that takes into account the fine spatiotemporal variation characteristics m The fusion model expression is: In the formula, To take into account the nonlinear vertical variation of T m Model, is T at the average height of the window m Model.

8. The method for estimating the atmospheric weighted average temperature of ocean and polar regions by integrating GNSS-RO / RS / ERA5 multi-source data according to claim 7, characterized in that: Constructing the T that takes into account the fine spatiotemporal variation characteristics m The fusion model includes: Taking into account the time parameter and the spatial variation, T at the average elevation of the window m The expression of the model is: Where λ represents longitude; represents latitude; JD represents Julian day; HOD represents the time of day; A0 represents a constant term; A1 and A2 represent longitude and latitude coefficients respectively; A3 and A4 represent annual cycle coefficients; A5 and A6 represent semi-annual cycle coefficients; A7 and A8 represent daily cycle coefficients; A9 and A 10 represents the semi-diurnal period coefficient; Adopt T m The elevation correction is performed using a third-order polynomial function that has nonlinear changes in the value and elevation. The T m The relationship between the fusion model and elevation is expressed as: T m =a1+a2h+a3h 2 +a4h 3 In the formula, h represents the height of the target point; a1, a2, a3 and a4 are the four unknown coefficient parameters of the fitting function; The nonlinear vertical variation of T m The expression of the model is: Where h represents the height of the target point; b2, b3 and b4 are the four unknown coefficient parameters of the fitting function, H m is the average elevation of the window; Based on T considering nonlinear vertical changes m Model, T at the average elevation of the window m The model is combined to obtain the T that takes into account the fine spatiotemporal variation characteristics. m Fusion model.

9. The method for estimating the atmospheric weighted average temperature of ocean and polar regions by integrating GNSS-RO / RS / ERA5 multi-source data according to claim 8, characterized in that: Obtain the T that takes into account the fine spatiotemporal variation characteristics m The fusion model includes: Where λ represents longitude; represents latitude; JD represents Julian day; HOD represents the time of day; A0 represents a constant term; A1 and A2 represent longitude and latitude coefficients respectively; A3 and A4 represent annual cycle coefficients; A5 and A6 represent semi-annual cycle coefficients; A7 and A8 represent daily cycle coefficients; A9 and A 10 represents the semi-diurnal period coefficient, A 11 ~A 13 is the elevation correction factor.

Images