Method for estimating uncertainty of earth radiation energy in lunar-based observation

Abstract

The invention relates to a lunar-based observation earth radiation energy uncertainty estimation method, which comprises the following steps of: respectively giving a plurality of options which are related to each error source and influence uncertainty for each error source related to a lunar-based observation earth radiation energy estimation model, and constructing a data set in all combination modes of all options among the error sources; and implementing or not implementing simplification of the data set, and implementing lunar-based observation earth radiation energy estimation based on the lunar-based observation earth radiation energy estimation model by adopting each group of data in the data set or each group of data in the simplified data set, the standard deviation of each estimation result is used as the lunar-based observation earth radiation energy uncertainty, and error sources comprise earth attitude errors, moon position attitude errors, earth shape errors, angle anisotropy errors and unit infinitesimal subdivision errors. According to the method, the uncertainty of the lunar-based observation earth radiation energy caused by various error sources can be estimated, and then the lunar-based observation earth radiation energy with the additional uncertainty is obtained.

Description

Technical Field

[0001] The invention relates to a method for estimating uncertainty of lunar-based observation of earth radiation energy and a computer system for implementing the method, belonging to the technical field of earth observation. Background Art

[0002] The Earth's radiation energy balance reflects how much energy the Earth system accumulates or releases, and is the most direct indicator for measuring the extent of global climate change. Its estimation requires the measurement of both the "income" and "payment" parts of energy. Among them, the "payment" part requires the measurement of the Earth's overall radiation energy in all directions, which is a difficult point to observe. The current observation data mainly comes from low-Earth orbit satellites, and the overall uncertainty is large, making it difficult to provide conclusive support for global climate change. The lunar-based platform can conduct multi-angle, holistic, and long-term observations of the Earth's radiation energy, which is expected to effectively supplement the existing observation system and achieve a more accurate estimate of the Earth's overall radiation energy.

[0003] The uncertainty of lunar-based observation of the Earth's hemispherical-scale radiation energy mainly comes from the quantification errors of its observation geometry, sensor parameters, Earth's radiation energy and its radiation anisotropy. The estimation of lunar-based observation energy is a relatively cumbersome process. Starting from the initial platform, lunar attitude and orbit, Earth's rotation, and Earth's radiation energy data, it is necessary to go through multiple core technical links to estimate the lunar-based observation energy. The error sources involved in each of the key steps will lead to the estimation error of the final observation energy, thereby affecting the design of key sensor parameters. Lunar-based Earth observation is a completely new observation method. There are no similar scenes and analyses as references in satellite observations. In addition, due to the unique geometry of lunar-based Earth observations, it is impossible to use the methods used by satellite platforms for analog analysis. Therefore, it is necessary to quantify the impact of observation geometry, microelement decomposition, and anisotropic radiation characteristic errors on the accuracy of lunar-based observation energy estimation.

[0004] At present, the research progress of lunar-based observation of radiation energy lies in observation geometry, simulation methods, etc., and a preliminary understanding of the magnitude and variation characteristics of the observed energy has been obtained. However, related research at home and abroad is still insufficient in terms of uncertainty estimation. Among them, the error source only considers the anisotropic characteristics, and has not yet realized the quantitative analysis and uncertainty estimation algorithm optimization of lunar-based observation of the earth's hemispherical-scale radiation energy for error sources such as observation geometry error (including lunar position, attitude error, earth attitude error, earth ellipsoid error), observation field segmentation error, and radiation anisotropy model error. It is difficult to effectively support the realization of the later engineering and data application of lunar-based earth observation. Since scientific data lacks practical significance if the uncertainty is not quantified, data without proper uncertainty evaluation will seriously restrict its effective application in scientific research. Therefore, it is very important to propose a method for estimating the uncertainty of lunar-based observation of earth radiation energy, which will be conducive to revealing the long-term variation law of earth radiation energy, exploring the scientific application potential of lunar-based data, and laying a scientific foundation for the realization of subsequent engineering. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, the present invention provides a method for estimating the uncertainty of lunar-based observation of earth radiation energy and a system for estimating lunar-based observation of earth radiation energy using this method, so as to realize the estimation of the uncertainty of lunar-based observation of earth radiation energy caused by multiple error sources.

[0006] The technical solution of the present invention is: a method for estimating the uncertainty of lunar-based observation of earth radiation energy, for each error source involved in the lunar-based observation of earth radiation energy estimation model, several options / factors (data items, discrete value ranges of data items, or factors) affecting the uncertainty involved in each error source are respectively given, a data (option combination) set (data set) is constructed in the form of all combinations of all options between the error sources, simplification of the data set is implemented or not, and each group of data in the data set or each group of data in the simplified data set is used to implement lunar-based observation of earth radiation energy estimation based on the lunar-based observation of earth radiation energy estimation model (obtaining corresponding estimation results), and the standard deviation of each estimation result is used as the uncertainty of lunar-based observation of earth radiation energy.

[0007] Accordingly, the mean of the estimates is the central estimate of the Earth's radiated energy.

[0008] Accordingly, the covariance of each estimation result can be analyzed to obtain the spatiotemporal structure / spatial characteristics of the error.

[0009] Furthermore, the lunar-based observation of the Earth's radiation energy estimation model is:

[0010]

[0011] in, The position of the moon-based sensor is The lunar-based observation of the Earth's radiation energy, is any point on the earth in the corresponding field of view (hemispherical field of view) The radiant brightness of the moon-based sensor (or position ) The line of sight vector pointing to the Earth's center of mass and any point on the Earth and lunar-based sensors (position ) The angle between the lines, is the corresponding solid angle for Earth observation, dΩ is the solid angle for Earth observation The differential of .

[0012] Furthermore, the error sources include earth attitude error, moon position attitude error, earth shape (earth ellipsoid) error, angular anisotropy error and unit element subdivision error.

[0013] Preferably, the several options for influencing uncertainty involved in the earth attitude error are several groups of random perturbations on five parameters of the earth's rotation parameters, and the random perturbation applied to any parameter conforms to a normal distribution with a mean of 0. The five parameters of the earth's rotation parameters are polar drift in the x-direction, polar drift in the y-direction, UT1-UTC, nutation in the x-direction, and nutation in the y-direction.

[0014] Preferably, several options affecting the uncertainty involved in the lunar position attitude error are several planetary ephemeris used, including DE of NASA of the United States, INPOP developed by the Paris Observatory of France, and EPM developed by the Institute of Applied Astronomy of the Russian Academy of Sciences.

[0015] Preferably, several options affecting uncertainty involved in the earth shape error are several earth ellipsoid parameters used, including CGCS2000 earth ellipsoid parameters, WGS84 earth ellipsoid parameters and GRS80 earth ellipsoid parameters.

[0016] Preferably, the several options affecting the uncertainty involved in the angular anisotropy error are various combinations of several atmospheric parameter dataset options and several angular anisotropy model options, wherein the several atmospheric parameter dataset options include the ERA5 dataset, the GEOS-5FP dataset and the CERES dataset, and / or other datasets represented by these datasets, or other suitable datasets; the several angular anisotropy model options include the ERBE model, the CERES Terra / AquaEdition4 model and the CERES / TRMM SSF Edition2B model.

[0017] Preferably, the several options affecting the uncertainty involved in the unit element segmentation error are several spatial interpolation methods used for the unit element segmentation, including the inverse distance weighted method, the Kriging interpolation method and the spline interpolation method.

[0018] Preferably, the data set is merged and simplified in the following manner:

[0019] 1) Merge the process error source data;

[0020] 2) Merge the error source data with large correlation.

[0021] A system for estimating lunar-based earth radiation energy observation adopts any one of the lunar-based earth radiation energy uncertainty estimation methods disclosed in the present invention to implement lunar-based earth radiation energy estimation (or lunar-based earth radiation energy uncertainty estimation), and obtains lunar-based earth (hemispherical scale) radiation energy with additional uncertainty, wherein the mean of each estimation result is the lunar-based earth radiation energy estimation value (or lunar-based earth radiation energy observation), and the standard deviation of each estimation result is the lunar-based earth radiation energy uncertainty.

[0022] The beneficial effect of the present invention is that a method for estimating the uncertainty of lunar-based observations of the Earth's radiation energy is proposed, which provides conditions for more accurately obtaining the lunar-based observations of the Earth's hemispherical-scale radiation energy with additional uncertainty, and is conducive to providing more accurate input for the reconstruction of the Earth's overall radiation energy based on lunar-based observations.

[0023] The present invention summarizes the energy error sources of lunar-based observations and calculates the uncertainty, optimizes the construction of ensemble data to obtain the uncertainty of hemispherical-scale radiation energy of lunar-based observations, enables researchers to analyze the spatiotemporal characteristics of errors, greatly promotes the understanding of lunar-based observations of the Earth's radiation energy, and is beneficial to support the scientific configuration of sensors and the rational design of the system from the perspective of scientific phenomena, and has guiding significance for engineering implementation and subsequent data application. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a schematic diagram of the present invention relating to the construction of set data;

[0025] Figure 2 The present invention relates to 100 sets of collective data examples of the monthly average long-wave radiation energy observation values ​​of the earth in 2023;

[0026] Figure 3 This is an example of the effect of the grid size on the field of view of lunar-based observations involved in the present invention. DETAILED DESCRIPTION

[0027] On the basis of considering the differences in characteristics of different error sources at the same time, the present invention constructs "aggregate data" according to the influence mechanism of each error source on the lunar-based observation energy, and estimates the uncertainty of lunar-based observation of Earth radiation energy, so that related observation research can obtain lunar-based observation of Earth hemispherical-scale radiation energy with additional uncertainty, clarify its temporal and spatial variation law, iteratively realize the system design of parameters such as the dynamic range, observation sensitivity, system calibration accuracy and other parameters of the corresponding channels of lunar-based sensors, and propose an optimization scheme for key sensor parameters.

[0028] Placing sensors / facilities such as cavity radiometers on the lunar surface can realize the observation of the radiation energy on the scale of the Earth's hemisphere, which is the most important observation data for reconstructing the overall radiation energy of the Earth. Set the location of the sensors deployed on the lunar surface and the observation time, obtain the position of the sun, the Earth, the moon and the attitude of the moon through the planetary ephemeris, obtain the attitude of the Earth through the Earth's rotation parameters, and establish a lunar-based observation of the Earth's radiation energy estimation model:

[0029]

[0030] Among them, the Earth radiation energy observed by the lunar-based sensor is recorded as The position of the lunar-based sensor on the lunar surface is denoted by The sensor location The latitude θ of the lunar space coordinate system can be used p ,longitude and height h p It can also be represented by any other suitable coordinate or position marking method. Represents any point on the Earth (hemisphere field of view) The radiance of the point The location can be expressed in terms of latitude θ in the geographic coordinate system. d ,longitude Height h d It can also be represented by any other suitable coordinate or position marking method. β is the line of sight vector from the lunar-based sensor to the Earth's center of mass and the point on the Earth. The angle between the line connecting the moon-based sensor and the moon-based sensor, dΩ represents the moon-based sensor / position Earth Observation Solid Angle The differential of .

[0031] The error sources involved in the above-mentioned lunar-based observation of the Earth's radiation energy estimation model mainly include: the estimation error of the Sun-Earth-Moon position caused by the planetary ephemeris error and the lunar attitude error, and the Earth attitude error caused by the Earth's rotation parameter estimation error. These error sources will cause Ω, β, and dΩ in the above formula to change; the Earth shape error, this error source will cause Ω, β, and dΩ to change; the angular anisotropy model used is different, and the use of different models will cause the radiation angle anisotropy estimation to change; the error introduced by the unit microelement decomposition method, this error source will affect dΩ. The above factors / error sources will ultimately affect the accuracy of the lunar-based observation of the Earth's radiation energy estimation.

[0032] The present invention adopts the following method to quantify the above errors.

[0033] 1. Quantify the Earth attitude error using multiple random perturbations

[0034] The estimation of Earth attitude calculation requires the use of Earth rotation parameters, which have measurement uncertainties. The uncertainty in the measurement of Earth rotation parameters will cause errors in the estimation of Earth attitude, and further cause errors in the estimation of Earth radiation energy in lunar-based observations.

[0035] The uncertainty of the Earth's rotation parameters at different times in the Earth's attitude calculation is different. This type of error can be regarded as a normal distribution with a mean of the mean and a standard deviation of the uncertainty. All parameters of the Earth's rotation parameters are randomly perturbed to quantify the impact of the Earth's rotation parameters on the estimation of the Earth's radiation energy from lunar-based observations, that is, a random perturbation (random perturbation amount) is added to each of the five parameters of the Earth's rotation parameters (polar motion in the x and y directions, UT1-UTC, and nutation in the x and y directions) to form a set (or group) of random perturbation data for these five parameters. All parameters of the Earth's rotation parameters are randomly perturbed 200 times (or other times). The random perturbation conforms to the normal distribution with a mean of 0 and a standard deviation of the uncertainty of each parameter of the Earth's rotation parameters, and then 200 sets (groups) of random perturbation data sets are constructed. Each of the 200 sets of randomly perturbed data is used as the input data in formula (1) one by one to obtain the calculation results after 200 perturbations. The standard deviation of these results is the uncertainty of the Earth attitude error in estimating the lunar-based observation energy.

[0036] 2. Quantify the lunar position attitude error using multiple ephemeris version data sets

[0037] There are three commonly used / recognized planetary ephemeris: JPL Development Ephemeris (DE) of NASA, Intégration Numérique Planétaire de l′Observatoire de Paris (INPOP) developed by Paris Observatory, France, and Ephemerides of Planets and the Moon (EPM) developed by Institute of Applied Astronomy of Russian Academy of Sciences. There are inconsistencies in the quantification of the position and attitude of the moon among these three ephemeris. Since all measurements are estimates of the real state, it is necessary to use the three ephemeris to estimate the position and attitude of the moon in combination, so as to quantify the uncertainty of the estimation caused by the position and attitude of the moon.

[0038] The quantification method is: respectively use the above three ephemeris to implement the calculation of formula (1) to calculate three different results of the lunar-based observation of the earth's radiation energy estimation. The standard deviation of these three different results is the uncertainty of the lunar position and attitude for the lunar-based observation energy estimation.

[0039] 3. Quantify the simulation error caused by the earth shape model using multiple earth ellipsoid models (earth shape error)

[0040] At present, the commonly used / recognized earth ellipsoids mainly include: World Geodetic System 1984 (WGS84), Chinese Geodetic Coordinate System 2000 (CGCS2000) and Geodetic Reference System 1980 (GRS80). The above three earth ellipsoid parameters are used to implement the calculation of formula (1) to calculate the results of the estimation of the earth radiation energy of the lunar-based observation under three different earth ellipsoids. The standard deviation of these three different results is the uncertainty of the earth shape model in estimating the energy of the lunar-based observation.

[0041] 4. Quantify the radiation anisotropy (angular anisotropy error) using different atmospheric parameter datasets and different angular anisotropy models

[0042] The analysis of radiation anisotropy is not only affected by the angular anisotropy model used, but also by related atmospheric parameters such as clouds and aerosols. ERA5 (ECMWF atmospheric reanalysis of the global climate, the fifth generation of the European Center for Medium-Range Weather Forecasts Global Climate Atmospheric Reanalysis Tool) reanalysis dataset, GEOS-5FP (Goddard Earth Observing System Model, Version 5 Forward Processing) dataset and CERES (Clouds and the Earth's Radiant Energy System) dataset are three representative / recognized data sources involving atmospheric parameters. ERBE (Earth Radiation Budget Experiment, an angular anisotropy model built using ERBE measurement data), CERES Terra / Aqua Edition4 (an angular anisotropy model built using CERES Terra and Aqua satellite data) and CERES / TRMM SSF Edition2B (an angular anisotropy model built using SSF data) are three representative / recognized different angular anisotropy models. There are 9 combinations of the above three atmospheric parameter data sources and angular anisotropy models. The above 9 combinations are used to implement the calculation of formula (1) and 9 different results are calculated. The results of the estimation of the lunar-based Earth radiation energy under the 9 different atmospheric parameter data sources and angular anisotropy model combinations are calculated. The standard deviation of these 9 different results is the uncertainty of the angular anisotropy model for the estimation of lunar-based energy.

[0043] 5. Use decomposition quantization method to quantify unit element subdivision (unit element subdivision error)

[0044] The Earth radiation energy dataset is grid data. When using it to estimate the lunar-based field of view energy, we will face the question of whether the grid is within the lunar-based field of view. The grid size will affect the lunar-based observation energy estimation (see Figure 3 ), when the grid spatial resolution is greater than 1°, it will seriously affect the estimation of lunar-based observation energy, which is mainly reflected in the fact that the area integral of the hemispherical field of view can no longer show the change in the distance between the earth and the moon. Therefore, the grid resolution is determined to be 1°, and spatial interpolation is performed to quantify the impact of unit microelement subdivision on the estimation of lunar-based field of view energy.

[0045] The inverse distance weighted method, Kriging interpolation method and spline interpolation method are used for interpolation respectively. The standard deviation of the lunar-based observation energy obtained by the three interpolation methods is calculated, and the influence of unit microelement decomposition on the estimation of lunar-based observation energy is quantified.

[0046] Based on the above analysis, the uncertainty estimation of lunar-based observation of earth hemispheric radiation energy is implemented. Since various error sources are not independent of each other, they cannot be simply added together to obtain the total uncertainty. This paper adopts the method of aggregate data to comprehensively consider all error sources and construct a lunar-based observation energy data set composed of several sets of data (see Figure 1 ), that is, the aggregate data, all the groups or representative parts of the group data in the aggregate data are substituted into formula (1) for calculation, and the total uncertainty is obtained based on the standard deviation of each calculation result.

[0047] In order to reduce the number of ensemble members while ensuring the accuracy of uncertainty estimation, and control the data processing volume within the appropriate computing capacity that the computer can bear, the error sources / error items can be sorted out and merged. The merging principle / method can be: 1) merging process error sources (such as merging the errors of various atmospheric parameters); 2) merging error sources with large correlation (such as merging all observation geometry errors into line element and angle element errors). Any other appropriate merging and simplification methods or other simplification methods can also be used.

[0048] The ensemble data is obtained by integrating the three major factors of observation geometry, angular anisotropy, and spatial resolution interpolation, and finally 100 sets (or other appropriate sets) of representative ensemble data are generated (see Figure 2 ). Input each group of data in the ensemble data into formula (1) respectively. The mean of the calculated results is the central estimate of the earth's radiation energy. The difference between samples in the same ensemble (usually the standard deviation of all calculated results) can be defined as the total uncertainty. The covariance analysis between samples in different ensembles can obtain the spatiotemporal structure / spatial characteristics of the error.

[0049] A computer / data processing system capable of implementing the above method can be established based on existing technologies. For example, a processor and a memory are provided, and executable code (software) is stored in the memory. When the executable code is executed by the processor, the processor executes any one of the methods for estimating the uncertainty of lunar-based observation of Earth radiation energy disclosed in the present invention.

[0050] A data storage device / database may be provided for storing various data / models required for estimation, and real-time detection data from relevant moon-based sensors may be obtained through a communication unit, and / or a website / server related to the data / models required for estimation may be accessed to obtain real-time detection data and other relevant data.

[0051] Random disturbances and other similar data can be automatically generated by a computer system under the constraints of set rules / parameters.

[0052] The corresponding software can be loaded into the existing computer system to form the moon-based earth radiation energy estimation system of the present invention.

[0053] Unless otherwise specified or when one preferred or optional technical means is a further limitation of another technical means, the preferred and optional technical means disclosed in the present invention can be arbitrarily combined to form several different specific implementation methods.

Claims

1. A method for estimating the uncertainty of lunar-based earth radiation energy observation, for each error source involved in the lunar-based earth radiation energy estimation model, several options of the uncertainty involved in each error source are given respectively, a data set is constructed in the form of all combinations of all options between the error sources, the data set is simplified or not, and the lunar-based earth radiation energy estimation based on the lunar-based earth radiation energy estimation model is estimated by using each group of data in the data set or each group of data in the simplified data set, and the standard deviation of each estimation result is used as the uncertainty of the lunar-based earth radiation energy observation.

2. The uncertainty estimation method of lunar-based observation of earth radiation energy according to claim 1, characterized in that The lunar-based observation of the Earth's radiation energy estimation model is: in, The position of the moon-based sensor is The lunar-based observation of the Earth's radiation energy, is any point on the earth in the corresponding field of view The radiant brightness of the moon is β, which is the line of sight between the lunar sensor pointing to the Earth's center of mass and any point on the Earth. The angle between the line connecting the moon-based sensor and is the corresponding solid angle for Earth observation, dΩ is the solid angle for Earth observation The differential of .

3. The uncertainty estimation method of lunar-based observation of earth radiation energy according to claim 2, characterized in that The error sources include earth attitude error, moon position attitude error, earth shape error, angular anisotropy error and unit element segmentation error.

4. The uncertainty estimation method for lunar-based observation of earth radiation energy according to any one of claims 1 to 3, characterized in that Several options for the uncertainty affecting the earth attitude error are several groups of random perturbations on five parameters of the earth's rotation parameters. The random perturbation applied to any parameter conforms to a normal distribution with a mean of 0. The five parameters of the earth's rotation parameters are polar excursion in the x-direction, polar excursion in the y-direction, UT1-UTC, nutation in the x-direction, and nutation in the y-direction.

5. The uncertainty estimation method of lunar-based observation of earth radiation energy according to claim 3, characterized in that Several options for the uncertainty involved in the lunar position attitude error are the several planetary ephemeris used, including DE from NASA, INPOP developed by the Paris Observatory in France, and EPM developed by the Institute of Applied Astronomy of the Russian Academy of Sciences.

6. The uncertainty estimation method of lunar-based observation of earth radiation energy according to claim 3, characterized in that Several options affecting the uncertainty involved in the earth shape error are the earth ellipsoid parameters used, including CGCS2000 earth ellipsoid parameters, WGS84 earth ellipsoid parameters and GRS80 earth ellipsoid parameters.

7. The uncertainty estimation method of lunar-based observation of earth radiation energy according to claim 3, characterized in that The several options that affect the uncertainty involved in the angular anisotropy error are various combinations of several atmospheric parameter dataset options and several angular anisotropy model options. The several atmospheric parameter dataset options include the ERA5 dataset, the GEOS-5FP dataset and the CERES dataset. The several angular anisotropy model options include the ERBE model, the CERES Terra / Aqua Edition4 model and the CERES / TRMM SSF Edition2B model.

8. The uncertainty estimation method of lunar-based observation of earth radiation energy according to claim 3, characterized in that Several options that affect the uncertainty involved in the unit element segmentation error are the several spatial interpolation methods used in the unit element segmentation, including the inverse distance weighted method, kriging interpolation method, and spline interpolation method.

9. The uncertainty estimation method for lunar-based observation of earth radiation energy according to any one of claims 1 to 8, characterized in that The data set is merged and simplified in the following ways: 1) Merge the process error source data; 2) Merge the error source data with large correlation.

10. A lunar-based Earth radiation energy estimation system, characterized in that The uncertainty estimation method for lunar-based observation of earth radiation energy according to any one of claims 1 to 9 is used to implement the estimation of lunar-based observation of earth radiation energy to obtain lunar-based observation of earth radiation energy with additional uncertainty, wherein the mean of each estimation result is the estimated value of lunar-based observation of earth radiation energy, and the standard deviation of each estimation result is the uncertainty of lunar-based observation of earth radiation energy.

Images