CN-121996897-A - Uncertainty analysis method for earth surface ecological parameter measurement and calculation

CN121996897ACN 121996897 ACN121996897 ACN 121996897ACN-121996897-A

Abstract

The disclosure provides an uncertainty analysis method for surface ecological parameter measurement and calculation. The method comprises the steps of obtaining an observation value of a target index aiming at a target area, sampling from prior probability distribution of a measuring and calculating factor contained in a target index measuring and calculating model to obtain a plurality of candidate prior parameter combinations, wherein the measuring and calculating factor comprises a model input variable and an error parameter, the model input variable comprises prior earth surface ecological parameters, the plurality of candidate prior parameter combinations are respectively input into the target index measuring and calculating model, the index measuring and calculating results of the candidate prior parameter combinations comprise measuring and calculating values of target index values, the target prior parameter combinations are selected from the candidate prior parameter combinations according to the measuring and calculating values of the target index and the observation value of the target index, the target prior parameter combinations comprise target prior earth surface ecological parameters, and uncertainty aiming at the earth surface ecological parameters is calculated according to probability distribution determined based on the target prior earth surface ecological parameters.

Inventors

WANG YAXIN
Zhang Wangfei
MA HUIJING
ZHAO HAN
JIANG YING
SONG CI
ZHANG ZHUOLI
YA NUYI
JIA YUNQIAN

Assignees

中国林业科学研究院资源信息研究所

Dates

Publication Date: 20260508
Application Date: 20260120

Claims (10)

1. An uncertainty analysis method for surface ecological parameter measurement and calculation comprises the following steps: Obtaining an observation value of a target index aiming at a target area; sampling from prior probability distribution of measuring factors contained in a target index measuring model to obtain a plurality of candidate prior parameter combinations, wherein the measuring factors comprise model input variables and error parameters, and the model input variables comprise prior earth surface ecological parameters; Inputting the candidate prior parameter combinations into the target index measurement model respectively, and outputting respective index measurement results of the candidate prior parameter combinations, wherein the index measurement results comprise measurement values of the target index values; Selecting a target prior parameter combination from the plurality of candidate prior parameter combinations according to the measuring and calculating value of the target index and the observed value of the target index, wherein the target prior parameter combination comprises a target prior surface ecological parameter; and calculating uncertainty aiming at the surface ecological parameters according to probability distribution determined based on the target priori surface ecological parameters.
2. The method of claim 1, wherein selecting a target a priori parameter combination from the plurality of candidate a priori parameter combinations based on the measured value of the target indicator and the observed value of the target indicator comprises: constructing a likelihood function according to the measuring and calculating value of the target index and the observed value of the target index; calculating the joint posterior probability of any group of candidate prior parameter combinations based on the likelihood function and the prior probability distribution; And comparing the joint posterior probability with the joint posterior probability obtained by the previous iteration calculation, and determining at least one group of target prior parameter combinations according to the comparison result.
3. The method of claim 1, wherein the prior probability distribution of the error parameter is determined from: obtaining an observation data set of the target index obtained by carrying out multiple observations on the target area; determining a super prior probability distribution based on the probability distribution type of the observation data set; determining a numerical value obtained by sampling from the super prior probability distribution as the variance of the prior probability distribution of the error parameter; And determining the prior probability distribution of the error parameter according to the variance.
4. A method according to claim 3, wherein determining a super prior probability distribution based on the probability distribution type of the observed dataset comprises: Determining a super prior probability distribution type according to the probability distribution type, the distribution parameters and the conjugate distribution of the observation data set; And determining the super prior probability distribution according to a preset distribution parameter and the prior probability distribution type.
5. The method of claim 4, wherein the probability distribution type of the observed dataset is a normal distribution or a uniform distribution; the determining the super prior probability distribution type according to the probability distribution type, the distribution parameters and the conjugate prior distribution of the observation data set comprises the following steps: under the condition that the probability distribution of the observation data set is normal distribution, determining the inverse gamma distribution of the conjugate distribution as the distribution type of the super prior probability distribution by the distribution parameter variance; And under the condition that the probability distribution of the observation data set is uniform distribution, determining the pareto distribution of the conjugate distribution as the distribution type of the super prior probability distribution by the upper bound of the distribution parameters.
6. The method of claim 1, wherein, The surface ecological parameter includes vegetation coverage or forest biomass.
7. The method of claim 6, wherein, In the case where the surface ecological parameter is vegetation coverage, the model input variables of the target index measurement model include: A vegetation coverage, the vegetation coverage characterizing an area ratio of vegetation in the pixel; Normalized vegetation index values for the pure vegetation pixels; normalized vegetation index values for pure bare soil pixels; a first error parameter; The target index value output by the target index measuring and calculating model comprises a normalized vegetation index.
8. The method according to any one of claims 6 or 7, wherein, Under the condition that the surface ecological parameters are vegetation coverage, respectively inputting the multiple candidate prior parameter combinations into the target index measurement model, and outputting index measurement results of the multiple candidate prior parameter combinations, wherein the method comprises the following steps: Calculating the product of the prior vegetation coverage and the prior normalized vegetation index value of the prior pure vegetation pixels to obtain a vegetation index contribution value; calculating the product of the prior bare soil coverage and the prior normalized vegetation index value of the prior pure bare soil pixel to obtain a bare soil index contribution value, wherein the prior bare soil coverage is determined according to the prior vegetation coverage; And carrying out summation calculation on the vegetation index contribution value, the bare soil index contribution value and the first error parameter, and outputting the calculated value of the normalized vegetation index.
9. The method of claim 6, wherein, In the case where the surface ecological parameter is forest biomass, the model input variables of the target index measurement model include: Forest biomass; Tree height; a second error parameter; the target index value output by the target index measurement model comprises a chest diameter.
10. The method according to any one of claims 6 or 9, wherein, Under the condition that the surface ecological parameters are forest biomass, respectively inputting the candidate prior parameter combinations into the target index measurement model, and outputting index measurement results of the candidate prior parameter combinations, wherein the method comprises the following steps: calculating the difference value between the prior forest biomass and the second error parameter to obtain an estimated value of the forest biomass; Calculating the product of the model priori coefficient, the model priori correction coefficient and the different-speed growth regulating factor of the tree height to obtain a combined correction factor, wherein the different-speed growth regulating factor of the tree height is obtained by calculating according to the priori tree height; Calculating the ratio of the estimated value of the forest biomass and the combined correction factor to obtain a chest diameter candidate measuring and calculating base value; Calculating and outputting the measuring and calculating value of the chest diameter according to the prior influence index of the chest diameter on forest biomass and the chest diameter candidate measuring and calculating base value.

Description

Uncertainty analysis method for earth surface ecological parameter measurement and calculation Technical Field The disclosure relates to the technical field of earth surface ecological parameter measurement and calculation, in particular to an uncertainty analysis method for earth surface ecological parameter measurement and calculation. Background The surface ecological parameters are key indexes for reflecting vegetation growth conditions and evaluating ecological system functions and carbon reserves, and are widely applied to the fields of ecological environment monitoring, natural resource management, global climate change research and the like. The measurement accuracy of the surface ecological parameters can provide reliable data support for national land greening achievement evaluation, forest carbon sink metering, ecological protection and restoration engineering and the like. The traditional uncertainty evaluation method is difficult to simultaneously process nonlinear coupling relation of multi-source errors, so that the uncertainty estimation finally given is often too rough or deviated from practice, and the uncertainty reliability is not high. Disclosure of Invention In view of the above, to at least partially solve at least one of the above-mentioned technical problems, the present disclosure provides a method for uncertainty analysis of surface ecological parameter measurements, The uncertainty analysis method for the surface ecological parameter measurement comprises the steps of obtaining an observation value of a target index of a target area, sampling from prior probability distribution of measurement factors contained in a target index measurement model to obtain a plurality of candidate prior parameter combinations, wherein the measurement factors comprise model input variables and error parameters, the model input variables comprise prior surface ecological parameters, respectively inputting the plurality of candidate prior parameter combinations into the target index measurement model, outputting index measurement results of the candidate prior parameter combinations, wherein the index measurement results comprise measurement values of target index values, selecting the target prior parameter combination from the candidate prior parameter combinations according to the measurement values of the target index and the observation values of the target index, wherein the target prior parameter combination comprises the target prior surface ecological parameters, and calculating uncertainty for the surface ecological parameters according to probability distribution determined based on the target prior surface ecological parameters. According to the embodiment of the disclosure, a target prior parameter combination is selected from a plurality of candidate prior parameter combinations according to a measuring and calculating value of a target index and an observed value of the target index, wherein the method comprises the steps of constructing a likelihood function according to the measuring and calculating value of the target index and the observed value of the target index, calculating joint posterior probability of any group of candidate prior parameter combinations based on the likelihood function and prior probability distribution, comparing the joint posterior probability with joint posterior probability obtained by previous iteration calculation, and determining at least one group of target prior parameter combinations according to a comparison result. According to the embodiment of the disclosure, the prior probability distribution of the error parameter is determined according to the following operation, namely an observation data set of a target index obtained by carrying out multiple observations on a target area is obtained, the super prior probability distribution is determined based on the probability distribution type of the observation data set, the numerical value obtained by sampling from the super prior probability distribution is determined as the variance of the prior probability distribution of the error parameter, and the prior probability distribution of the error parameter is determined according to the variance. According to the embodiment of the disclosure, the super prior probability distribution is determined based on the probability distribution type of the observation data set, wherein the super prior probability distribution is determined according to the probability distribution type of the observation data set, the distribution parameters and the conjugate distribution, and the super prior probability distribution is determined according to the preset distribution parameters and the prior probability distribution type. According to the embodiment of the disclosure, the probability distribution type of the observation data set is normal distribution or uniform distribution, the super prior probability distribution type is determined according to the probabili