A method for time-varying risk assessment of terrestrial ecosystems under drought stress

Abstract

The present invention discloses a method for assessing time-varying risk of a terrestrial ecosystem under drought stress, comprising the following steps: S1. calculating the exceedance probability of two variables, drought duration and drought intensity, and using the calculation result of the exceedance probability to characterize the drought hazard of the terrestrial ecosystem; S2. constructing a two-dimensional terrestrial ecosystem exposure index formula based on two indexes, namely, vegetation cover and leaf area index, and using the calculation result of the exposure index to characterize the exposure of the terrestrial ecosystem; S3. constructing a three-vegetation index loss conditional probability formula based on three vegetation indices, namely, vegetation difference index, leaf area index and total primary productivity index, and using the calculation result of the three-vegetation index loss conditional probability to characterize the vulnerability of the terrestrial ecosystem; S4. calculating the time-varying risk of the terrestrial ecosystem under drought stress according to a preset time-varying risk assessment algorithm and combining the three major risk factors of drought hazard, exposure and vulnerability.

Description

Technical Field

[0001] The invention relates to the technical field of water conservancy engineering, and in particular to a method for assessing time-varying risks of terrestrial ecosystems under drought stress. Background Art

[0002] The possibility of drastic changes in the normal operation of natural and artificial systems due to extreme climate events such as drought is usually defined as risk. Terrestrial ecosystems cover 28.26% of the total land area and are an important part of the biosphere. After a drought occurs, water shortage often changes the appearance, structure and service functions of vegetation by inhibiting metabolic processes such as photosynthesis, respiration, transpiration, and nutrient transport, thereby inducing the risk of terrestrial ecosystems under drought stress. After a drought occurs, humans are prone to perceive the objective existence of terrestrial ecosystem risks from various phenomena such as reduced food production, deteriorating water quality, and obstructed leisure and entertainment activities. In addition, when the service functions of terrestrial ecosystems cannot meet human needs, terrestrial ecosystem risks will also be transmitted to human society, exacerbating human social risks. Therefore, assessing the risk of terrestrial ecosystems under drought stress is an important topic in drought crisis management.

[0003] However, the traditional risk assessment model for terrestrial ecosystems under drought stress has not yet considered the time-varying nature of risk factors, and therefore cannot reveal the time-varying laws of risk. Ignoring the time-varying laws of risk will make it difficult to effectively manage the terrestrial ecosystem risks that may intensify in the future. Therefore, how to accurately assess the time-varying risks of terrestrial ecosystems under drought stress so that drought prevention planning can be more targeted at high-risk hotspots and have reliable effectiveness is an urgent problem to be solved in the current research on terrestrial ecosystem risks under drought stress. Summary of the invention

[0004] Based on this, it is necessary to provide a time-varying risk assessment method for terrestrial ecosystems under drought stress that can accurately assess the time-varying laws of risks of terrestrial ecosystems under drought stress.

[0005] A method for assessing time-varying risk of terrestrial ecosystems under drought stress comprises the following steps:

[0006] S1. Calculate the exceedance probability of the two variables of drought duration and drought intensity, and characterize the drought risk of terrestrial ecosystems with the calculated results of the exceedance probability;

[0007] S2. constructing a two-dimensional terrestrial ecosystem exposure index formula based on the vegetation cover and leaf area index, solving the exposure index formula to obtain the calculation result of the exposure index, and using the calculation result of the exposure index to characterize the exposure of the terrestrial ecosystem;

[0008] S3. construct a three-vegetation index loss conditional probability formula based on the three vegetation indices of vegetation difference index, leaf area index and total primary productivity index, solve the three-vegetation index loss conditional probability formula, obtain the calculation results of the three-vegetation index loss conditional probability, and use the calculation results of the three-vegetation index loss conditional probability to characterize the vulnerability of terrestrial ecosystems;

[0009] S4. The time-varying risk of terrestrial ecosystems under drought stress is calculated based on a preset time-varying risk assessment algorithm and combined with the three major risk factors of drought hazard, exposure and vulnerability.

[0010] In one embodiment, the step S1 comprises:

[0011] S11. Extract drought events and their drought attributes from the standardized precipitation index sequence by using the run theory, the drought attributes including drought duration, drought intensity, drought start time and drought end time; combine the interval time and exceedance criterion theory, merge the interrelated drought events and eliminate the micro-drought events to obtain independent drought event samples;

[0012] S12. Use the KPSS test method to test whether the drought duration and drought intensity univariate sequences of the independent drought events maintain stability; when the drought duration and drought intensity univariate sequences are stable, use a variety of probability distribution functions with constant parameters to fit the drought duration and drought intensity, and evaluate the goodness of fit of the probability distribution functions based on the BIC criterion to obtain the optimal marginal distribution function of drought duration and drought intensity; when the drought duration and drought intensity univariate sequences are not stable, use a variety of probability distribution functions with time-varying parameters to fit the drought duration and drought intensity, and estimate the time-varying parameters of the probability distribution functions based on the generalized additive model and evaluate the goodness of fit of the time-varying parameters of the probability distribution functions based on the BIC criterion to obtain the optimal marginal distribution function of drought duration and drought intensity;

[0013] S13. Use the Copula function likelihood ratio method to test whether the stationarity assumption between the marginal distribution function of drought duration and the marginal distribution function of drought intensity holds true; when the stationarity assumption holds true, use multiple Copula joint distribution functions with constant parameters to fit the joint distribution function of drought duration and drought intensity, and evaluate the goodness of fit of the Copula joint distribution function based on the BIC criterion to obtain the optimal joint distribution function of drought duration and drought intensity; when the stationarity assumption does not hold true, use multiple Copula joint distribution functions with time-varying parameters to fit the joint distribution function of drought duration and drought intensity, and evaluate the goodness of fit of the Copula joint distribution function based on the BIC criterion to obtain the optimal joint distribution function of drought duration and drought intensity;

[0014] S14. According to the preset exceedance probability calculation formula and in combination with the drought duration marginal distribution function, the drought intensity marginal distribution function and the joint distribution function of the drought duration and drought intensity, the exceedance probabilities of the two variables of the drought duration and drought intensity are calculated.

[0015] In one embodiment, in step S12, when the drought duration and drought intensity univariate series stationarity is established, multiple probability distribution functions with constant parameters are used to fit the drought duration and drought intensity respectively, and the goodness of fit of the probability distribution function is evaluated based on the BIC criterion, and the content of obtaining the optimal drought duration marginal distribution function and drought intensity marginal distribution function includes:

[0016] Seven probability distribution functions with constant parameters, including exponential, normal, Gumbel, logistic, lognormal, gamma and Weibull, were selected as alternatives. Drought duration and drought intensity were fitted by the seven probability distribution functions, and the goodness of fit of all the alternative probability distribution functions was evaluated by the BIC criterion.

[0017] The calculation formula of BIC criterion is as follows:

[0018]

[0019] Where L(·) represents the log-likelihood function; and m represent the estimated value and number of parameters of probability distribution, respectively; num represents the total number of drought times;

[0020] Calculate the BIC values ​​of all candidate probability distribution functions, then select the probability distribution function that minimizes the calculated value of BIC among all the candidate probability distribution functions as the optimal probability distribution function, and use the optimal probability distribution function as the optimal drought duration marginal distribution function and drought intensity marginal distribution function;

[0021] In step S12, when the drought duration and drought intensity univariate series stationarity does not hold, multiple probability distribution functions with time-varying parameters are used to fit the drought duration and drought intensity respectively, and the time-varying parameters of the probability distribution function are estimated based on the generalized additive model and the goodness of fit of the probability distribution function is evaluated based on the BIC criterion, and the optimal marginal distribution function of drought duration and the marginal distribution function of drought intensity are obtained, including:

[0022] Seven probability distribution functions with time-varying parameters, including exponential, normal, Gumbel, logistic, lognormal, gamma and Weibull, were selected as candidates. The drought duration and drought intensity were fitted by the seven probability distribution functions respectively. The generalized additive model was introduced to estimate the time-varying parameters of the probability distribution function. The drought onset time was taken as the explanatory variable, and the connection function between the time-varying parameters of the probability distribution function and the explanatory variables was established. The expression of the connection function is as follows:

[0023]

[0024] Where θ is a column vector of time-varying parameters of the probability distribution function, with a length of n; g() is a link function; X is a known matrix of n×J, corresponding to the explanatory variable values ​​of n time periods; β=(β1,β2,K,β J ) T is the parameter column vector to be solved, with dimension J; Z j is n×q j The known design matrix of γ j is a column vector of random variables with dimension equal to q j ;

[0025] The CG and RS algorithms are used to solve the generalized additive model and calculate the probability distribution parameters corresponding to all alternative probability distribution functions that change over time;

[0026] The goodness of fit of the probability distribution parameters corresponding to all the above alternative probability distribution functions is evaluated by the BIC criterion;

[0027] Calculate the BIC values ​​of the probability distribution parameters corresponding to all alternative probability distribution functions, and then select the probability distribution function that minimizes the calculated value of BIC among all the alternative probability distribution functions as the optimal probability distribution function, and use the optimal probability distribution function as the optimal drought duration marginal distribution function and drought intensity marginal distribution function.

[0028] In one embodiment, in step S14, the calculation formula of the exceedance probability is as follows:

[0029] P AND (d>D con. ,in>IN con.)

[0030] =1-P(d≤D con. )-P(in≤IN con. )+P(d≤D con. ,in≤IN con. )

[0031] =1-F D (D con. )-F IN (IN con. )+F D,IN (D con. ,IN con. )

[0032] Among them, P AND (d>D con. ,in>.IN con. ) represents the exceedance probability of the two variables of drought duration and drought intensity; D con. and IN con. Represent the given drought duration and drought intensity respectively; F D () and F IN () represent the cumulative probability of the marginal distribution function of drought duration and the marginal distribution function of drought intensity respectively; F D,IN () represents the cumulative probability of the joint distribution function of drought duration and drought intensity.

[0033] In one embodiment, step S1 further includes:

[0034] S15, normalizing the calculated exceedance probability, and using the normalized exceedance probability result to characterize the drought risk of the terrestrial ecosystem;

[0035] In step S15, the expression of drought risk based on the normalized exceedance probability is as follows:

[0036]

[0037] Among them, H represents drought risk; P AND represents the exceedance probability of the two variables of drought duration and drought intensity; and They represent the maximum and minimum values ​​of the exceedance probability of the two variables, drought duration and drought intensity, respectively.

[0038] In one embodiment, in step S2, the exposure expression based on the exposure index is as follows:

[0039] E=max(FVC,nLAI)

[0040] Among them, E represents exposure; FVC represents annual vegetation cover; nLAI represents annual normalized leaf area index.

[0041] In one embodiment, in step S3, the three vegetation index loss conditional probability formula is as follows:

[0042]

[0043] Among them, P(VI lower ≤vi≤VI upper |d=D,in=IN) represents the conditional probability of loss of three vegetation indices; vi represents the vegetation index; VI lower Represents the lower bound of the vegetation index; VI upper represents the upper bound of the vegetation index; f(vi|d,in) represents the conditional probability density function of the loss of the three vegetation indices; NDVI represents the vegetation difference index; LAI represents the leaf area index; GPP represents the gross primary productivity index;

[0044] The vulnerability expression based on the conditional probability of loss of different types of vegetation indices is as follows:

[0045]

[0046] Among them, vul NDVI 、vul LAI and vul GPP They represent the conditional probability of loss based on vegetation difference index NDVI, leaf area index LAI and gross primary productivity index GPP respectively; and They represent the minimum and maximum values ​​of the conditional probability of vegetation loss based on the vegetation difference index NDVI; and They represent the minimum and maximum values ​​of the conditional probability of vegetation loss based on the leaf area index LAI; and They represent the minimum and maximum values ​​of the conditional probability of vegetation loss based on the gross primary productivity index GPP, respectively.

[0047] In one embodiment, step S3 further includes:

[0048] S31, calculating the arithmetic mean of the conditional probabilities of loss of three vegetation indices, and using the calculation result of the arithmetic mean to characterize the vulnerability of the terrestrial ecosystem;

[0049] In step S31, the vulnerability expression based on the arithmetic mean of the conditional probabilities of the three vegetation index losses is as follows:

[0050] V=mean(vul NDVI ,vulLAI ,vul GPP )

[0051] Among them, V represents vulnerability; mean(vul NDVI , vul LAI , vul GPP ) represents the arithmetic mean of the conditional probability of loss of three vegetation indices: vegetation difference index NDVI, leaf area index LAI and gross primary productivity index GPP.

[0052] In one embodiment, in step S4, the time-varying risk of the terrestrial ecosystem is measured in years, and the expression of the time-varying risk assessment algorithm is as follows:

[0053] Risk i =H i ×E i ×V i

[0054] Among them, Risk i represents the time-varying risk of the terrestrial ecosystem in year i; H i 、E i and V i They represent the drought hazard, exposure and vulnerability of the terrestrial ecosystem in year i, respectively.

[0055] In one embodiment, it also includes:

[0056] S5, performing a linear fit between the calculated time-varying risk of the terrestrial ecosystem in the target year and the corresponding target year to simulate a linear relationship between the two.

[0057] The method for assessing the time-varying risk of terrestrial ecosystems under drought stress provided by the present invention takes terrestrial ecosystems with wider spatial distribution and longer impact links as drought disaster-prone objects, and calculates the time-varying risk of terrestrial ecosystems under drought stress according to a preset time-varying risk assessment algorithm and in combination with three major risk factors of drought hazard, exposure and vulnerability of terrestrial ecosystems with time-varying characteristics. Therefore, the time-varying law of the risk of terrestrial ecosystems under drought stress can be accurately assessed based on the calculation results of the time-varying risk, which can provide technical support for the scientific and efficient construction of a drought disaster prevention and control system. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying creative work.

[0059] Figure 1 The flowchart is a method for assessing time-varying risk of terrestrial ecosystems under drought stress in one embodiment. DETAILED DESCRIPTION

[0060] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0061] It should be noted that all directional indications in the embodiments of the present invention (such as up, down, left, right, front, back, etc.) are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0062] In addition, the descriptions of "first", "second", etc. in the present invention are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, "and / or" in the full text includes three solutions. Taking A and / or B as an example, it includes technical solution A, technical solution B, and technical solution that satisfies both A and B. In addition, the technical solutions between the various embodiments can be combined with each other, but it must be based on the ability of ordinary technicians in this field to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0063] like Figure 1 As shown, the present application provides a method for assessing time-varying risk of terrestrial ecosystems under drought stress, the method comprising the following steps:

[0064] S1. Calculate the exceedance probability of the two variables of drought duration and drought intensity, and use the calculated exceedance probability to characterize the drought hazard of terrestrial ecosystems.

[0065] S2. Construct a two-dimensional terrestrial ecosystem exposure index formula based on vegetation cover and leaf area index, solve the exposure index formula, obtain the calculation result of the exposure index, and use the calculation result of the exposure index to characterize the exposure of the terrestrial ecosystem.

[0066] S3. Based on the three vegetation indices of vegetation difference index, leaf area index and total primary productivity index, the conditional probability formula of the loss of three vegetation indices is constructed, and the conditional probability formula of the loss of three vegetation indices is solved to obtain the calculation results of the conditional probability of the loss of three vegetation indices, and the calculation results of the conditional probability of the loss of three vegetation indices are used to characterize the vulnerability of terrestrial ecosystems.

[0067] S4. Based on the preset time-varying risk assessment algorithm and combined with the three major risk factors of drought hazard, exposure and vulnerability, the time-varying risk of terrestrial ecosystems under drought stress can be calculated.

[0068] The method for assessing the time-varying risk of terrestrial ecosystems under drought stress provided by the present invention takes terrestrial ecosystems with wider spatial distribution and longer impact links as drought disaster-prone objects, and calculates the time-varying risk of terrestrial ecosystems under drought stress according to a preset time-varying risk assessment algorithm and in combination with three major risk factors of drought hazard, exposure and vulnerability of terrestrial ecosystems with time-varying characteristics. Therefore, the time-varying law of the risk of terrestrial ecosystems under drought stress can be accurately assessed based on the calculation results of the time-varying risk, which can provide technical support for the scientific and efficient construction of a drought disaster prevention and control system.

[0069] Drought hazard refers to the severity of the damage that may be caused by the intensity and frequency of drought disasters; drought hazard mainly includes indicators such as precipitation, topography, slope, river network density, and water conservation; exposure refers to the degree of contact between the disaster-prone body and disaster-causing factors during the occurrence of drought disasters; exposure mainly includes indicators such as population, GDP and cultivated land; vulnerability refers to the ability of the disaster-prone body to resist drought disasters; vulnerability mainly includes indicators such as vegetation drought resistance.

[0070] Against the backdrop of global warming, high-intensity droughts occur frequently and the risk of drought in terrestrial ecosystems continues to increase. Human activities such as land use conversion and afforestation cause the exposure of terrestrial ecosystems to change dynamically over time. The vulnerability of terrestrial ecosystems is also time-varying due to changes in water demand at different growth stages, the number of drought-resistant facilities, and the drought tolerance of vegetation.

[0071] In one embodiment, step S1 includes:

[0072] S11. Drought events and their drought attributes are extracted from the standardized precipitation index sequence through the run theory. The drought attributes include drought duration D, drought intensity IN, drought start time O and drought end time. The interval time and exceedance criterion theory are comprehensively used to merge interrelated drought events and eliminate micro-drought events to obtain independent drought event samples and reduce frequency analysis errors.

[0073] S12. The KPSS (Kwiatkowski–Phillips–Schmidt–Shin) test method was used to test whether the drought duration and drought intensity univariate series of independent drought events maintained stationarity. When the stationarity of drought duration and drought intensity univariate series was established, multiple probability distribution functions with constant parameters were used to fit drought duration and drought intensity respectively, and the goodness of fit of probability distribution functions was evaluated based on the BIC (Schwarz Bayesian information criterion) criterion to obtain the optimal marginal distribution function of drought duration and drought intensity. When the stationarity of drought duration and drought intensity univariate series was not established, multiple probability distribution functions with time-varying parameters were used to fit drought duration and drought intensity respectively, and the time-varying parameters of probability distribution functions were estimated based on the generalized additive model and the goodness of fit of probability distribution functions was evaluated based on the BIC criterion to obtain the optimal marginal distribution function of drought duration and drought intensity.

[0074] S13. Use the Copula function likelihood ratio method to test whether the stationarity assumption between the marginal distribution function of drought duration and the marginal distribution function of drought intensity is valid; when the stationarity assumption is valid, use multiple Copula joint distribution functions with constant parameters to fit the joint distribution function of drought duration and drought intensity, and evaluate the goodness of fit of the Copula joint distribution function based on the BIC criterion to obtain the optimal joint distribution function of drought duration and drought intensity; when the stationarity assumption is not valid, use multiple Copula joint distribution functions with time-varying parameters to fit the joint distribution function of drought duration and drought intensity, and evaluate the goodness of fit of the Copula joint distribution function based on the BIC criterion to obtain the optimal joint distribution function of drought duration and drought intensity;

[0075] S14. According to the preset exceedance probability calculation formula and in combination with the marginal distribution function of drought duration, the marginal distribution function of drought intensity and the joint distribution function of drought duration and drought intensity, the exceedance probabilities of the two variables of drought duration and drought intensity are calculated.

[0076] In step S11, the contents of drought events and drought attributes extracted from the standardized precipitation index sequence by run theory include:

[0077] Through the run theory, with one year as the calculation period, the drought duration and drought intensity data of each month in the target year are identified and calculated from the standardized precipitation index SPI series, and then the drought duration and drought intensity are determined based on the drought duration and drought intensity data of each month in the target year.

[0078] Specifically, drought duration D refers to the duration of the drought state from the beginning of a drought event (i.e., the drought start time) to the end of the drought event (i.e., the drought end time).

[0079] According to the principle of run theory, if an evaluation unit (pixel) is in a non-drought state at time n-1, but in a drought state at time n, then time n is considered to be the starting time of the drought event; if the drought state continues until time n+L, and at time n+L+1, the evaluation unit is in a non-drought state again, then time n+L can be considered to be the end time of the drought event, thereby obtaining the drought duration D=L of this drought event.

[0080] Specifically, a standardized precipitation index is obtained, and drought is characterized by the standardized precipitation index, wherein the threshold of the standardized precipitation index is -1. When it is identified that the standardized precipitation index value of a certain month in the target year is less than the threshold value, it is determined that drought has occurred in that month of the target year, and the drought duration is recorded as 1. Otherwise, the drought duration is recorded as 0. For example, if the standardized precipitation index value of a certain month in the target year is -1.2, the drought duration of that month in the target year is 1. Specifically in this embodiment, the drought duration of the target year is the sum of the drought durations of all months in which drought occurs in the target year.

[0081] Drought intensity IN refers to the sum of drought intensities at each moment during the duration of a drought event.

[0082] In step S11, the interval time and exceedance criterion theory are comprehensively applied to merge the interrelated drought events and eliminate the micro-drought events to obtain the content of independent drought event samples, including:

[0083] The interval time criterion theory is: when the interval time between adjacent drought events is t i Less than a given threshold t c When , it is considered that adjacent drought events are correlated;

[0084] The exceedance criterion theory is: when the precipitation of adjacent drought events exceeds v i and precipitation deficit w i The ratio is less than a given threshold ρ c , it is believed that adjacent drought events are correlated; when the standardized precipitation index SPI is greater than -1, it is considered that precipitation exceeds, and when the standardized precipitation index SPI is less than -1, it is considered that precipitation deficit occurs.i .

[0085] Based on the above interval time and exceedance criterion theory, the correlation of drought events includes two necessary conditions: 1) the interval time t between adjacent drought events i Less than a given threshold t c ; 2) Precipitation exceedance v of adjacent drought events i and precipitation deficit w i The ratio is less than a given threshold ρ c .

[0086] When the above two conditions are met at the same time, adjacent drought events are merged, and the combined drought duration is D pooled and drought intensity IN pooled Updated to:

[0087]

[0088] Among them, D i represents the drought duration in the i-th period; D i+1 represents the drought duration in the i+1th period; t i Indicates the interval between adjacent drought events; IN i Indicates the drought intensity in the i-th period; IN i+1 represents the drought intensity in the i+1th period; v i represents the precipitation exceedance of adjacent drought events;

[0089] The basis for judging a micro-drought event is as follows: when the ratio of the drought duration (or drought intensity) at a certain time to the average drought duration (or average drought intensity) of the drought event sequence is less than a given threshold, the drought event at that time is considered to be a micro-drought.

[0090] Specifically, in the process of eliminating micro-drought events, the threshold r is set d and r in , when the ratio of the drought duration at a certain time to the average drought duration of the drought event sequence is less than the threshold r d Or the ratio of the drought intensity at a certain time to the average drought intensity of the drought event sequence is less than the threshold r in When the drought event is considered to be a micro-drought, it is eliminated.

[0091] In step S12, when the drought duration and drought intensity univariate sequence stationarity is established, the drought duration and drought intensity univariate probability distributions remain constant, and multiple probability distribution functions with constant parameters are used to fit the drought duration and drought intensity respectively. The goodness of fit of the probability distribution function is evaluated based on the BIC criterion, and the optimal marginal distribution function of drought duration and the marginal distribution function of drought intensity are obtained, including:

[0092] Seven probability distribution functions with constant parameters, including exponential, normal, Gumbel, logistic, lognormal, gamma and Weibull, were selected as alternatives. Drought duration and drought intensity were fitted by the seven probability distribution functions, and the goodness of fit of all the alternative probability distribution functions was evaluated by the BIC criterion.

[0093] The calculation formula of BIC criterion is as follows:

[0094]

[0095] Where L(·) represents the log-likelihood function; and m represent the estimated value and number of probability distribution parameters (specifically the probability distribution parameters of drought duration and drought intensity), respectively; num represents the total number of drought times;

[0096] Calculate the BIC values ​​of all alternative probability distribution functions, and then select the probability distribution function that minimizes the calculated BIC value among all the alternative probability distribution functions as the optimal probability distribution function, and use the optimal probability distribution function as the optimal drought duration marginal distribution function and drought intensity marginal distribution function.

[0097] Further, in step S12, when the drought duration and drought intensity univariate sequence stationarity does not hold, the drought duration and drought intensity univariate probability distributions vary with time, and multiple probability distribution functions with time-varying parameters are used to fit the drought duration and drought intensity respectively, and the time-varying parameters of the probability distribution function are estimated based on the generalized additive model and the goodness of fit of the probability distribution function is evaluated based on the BIC criterion, and the optimal marginal distribution function of drought duration and the marginal distribution function of drought intensity are obtained, including:

[0098] Seven probability distribution functions with time-varying parameters, including exponential, normal, Gumbel, logistic, lognormal, gamma and Weibull, were selected as candidates. Drought duration and drought intensity were fitted by the seven probability distribution functions respectively. The generalized additive models for location, scale and shape (GAMLSS) were introduced to estimate the time-varying parameters of the probability distribution function. The drought onset time was taken as the explanatory variable, and the connection function between the time-varying parameters of the probability distribution function and the explanatory variables was established. The expression of the connection function is as follows:

[0099]

[0100] Where θ is a column vector of time-varying parameters of the probability distribution function, with a length of n; g() is a link function; X is a known matrix of n×J, corresponding to the explanatory variable values ​​of n time periods; β=(β1,β2,K,β J ) T is the parameter column vector to be solved, with dimension J; Z j is n×q j The known design matrix of γ j is a column vector of random variables with dimension equal to q j ;

[0101] The CG and RS algorithms are used to solve the generalized additive model and calculate the time-varying parameters corresponding to all the alternative probability distribution functions that change with time;

[0102] The goodness of fit of the time-varying parameters corresponding to all the above alternative probability distribution functions is evaluated by the BIC criterion;

[0103] The BIC values ​​of the time-varying parameters corresponding to all the alternative probability distribution functions are calculated, and then among all the alternative probability distribution functions, the probability distribution function that minimizes the calculated value of BIC is selected as the optimal probability distribution function, and the optimal probability distribution function is used as the optimal drought duration marginal distribution function and drought intensity marginal distribution function.

[0104] In the present invention, the identified drought onset time is used as the explanatory variable, the connection function between the time-varying parameter of the probability distribution function and the explanatory variable is established, and the semi-parametric cumulative form of the generalized additive model is adopted, which is conducive to capturing possible nonlinear relationships.

[0105] In step S13, the contents of using the Copula function likelihood ratio method to test whether the stationarity assumption between the drought duration marginal distribution function and the drought intensity marginal distribution function is valid include:

[0106] In the Copula function likelihood ratio method, the statistical test quantity Z based on the likelihood ratio is n and the likelihood ratio Λ k The expressions are as follows:

[0107]

[0108] Among them, Λ k is the likelihood ratio; L n () represents the likelihood function of the two variable series of drought duration and drought intensity of the entire sample; L k ()and are the likelihood functions before and after the test point k respectively; c() represents the Copula probability density function; and are the maximum likelihood estimates of the Copula function parameters η0, η1, and η2 respectively; when the statistical test quantity Z n When the value is less than the threshold corresponding to the 5% significance level, the stationarity hypothesis between the marginal distribution function of drought duration and the marginal distribution function of drought intensity is considered to be established; otherwise, when the statistical test quantity Z n When it is greater than or equal to the threshold corresponding to the 5% significance level, it is considered that the stationarity assumption between the marginal distribution function of drought duration and the marginal distribution function of drought intensity does not hold.

[0109] Further, in step S13, when the stationarity assumption holds, the joint distribution of drought duration and drought intensity remains constant, and multiple copula joint distribution functions with constant parameters are used to fit the joint distribution function of drought duration and drought intensity, and the goodness of fit of the copula joint distribution function is evaluated based on the BIC criterion, and the content of the optimal joint distribution function of drought duration and drought intensity is obtained, including:

[0110] Five Copula joint distribution functions with constant parameters, namely Gaussian, Clayton, Gumbel, Frank and Joe, were selected as alternatives. The joint distribution functions of drought duration and drought intensity were fitted by the five Copula joint distribution functions, and the goodness of fit of all the alternative Copula joint distribution functions was evaluated by the BIC criterion.

[0111] Calculate the BIC values ​​of all the alternative Copula joint distribution functions, and then select the Copula joint distribution function that minimizes the calculated BIC value among all the alternative Copula joint distribution functions as the optimal Copula joint distribution function, and use the optimal Copula joint distribution function as the optimal joint distribution function of drought duration and drought intensity.

[0112] Further, in step S13, when the stationarity assumption does not hold, the joint distribution of drought duration and drought intensity changes with time, and multiple Copula joint distribution functions with time-varying parameters are used to fit the joint distribution function of drought duration and drought intensity, and the goodness of fit of the Copula joint distribution function is evaluated based on the BIC criterion, and the content of the optimal joint distribution function of drought duration and drought intensity is obtained, including:

[0113] Five Copula joint distribution functions with time-varying parameters, namely Gaussian, Clayton, Gumbel, Frank and Joe, were selected as candidates. The joint distribution functions of drought duration and drought intensity were fitted by the above five Copula joint distribution functions. Similar to the introduction principle of the generalized additive model, in order to estimate the time-varying parameters of the Copula joint distribution function, the drought start time was taken as the explanatory variable, and the fourth-power connection function between the time-varying parameters of the Copula joint distribution function and the explanatory variables was established. The expression of the connection function is as follows:

[0114]

[0115] Among them, η t Represents the time-varying parameter of the Copula joint distribution function; O t represents the drought start time; α0, α1, α2, α3 and α4 represent the parameters to be solved;

[0116] The marginal function inference method and Nelder-Mead algorithm are used to solve the parameters α0, α1, α2, α3 and α4, and then the time-varying parameters of the Copula joint distribution function are obtained. The solution formulas for the parameters α0, α1, α2, α3 and α4 are as follows:

[0117]

[0118] in, represents the marginal distribution probability of drought duration in period t; represents the marginal distribution probability of drought intensity in the tth period; D t represents the drought duration in period t; IN t represents the drought intensity in period t; Parameters representing the marginal distribution of drought duration in period t; represents the parameter of the marginal distribution of drought intensity in the tth period; c() represents the Copula joint distribution function; f D () represents the marginal distribution function of drought duration; f IN () represents the marginal distribution function of drought intensity;

[0119] Calculate the BIC values ​​of the time-varying parameters of all the alternative Copula joint distribution functions, and then select the Copula joint distribution function that minimizes the calculated BIC value among all the alternative Copula joint distribution functions as the optimal Copula joint distribution function, and use the optimal Copula joint distribution function as the optimal joint distribution function of drought duration and drought intensity.

[0120] Furthermore, in step S14, the dual non-stationarity of the single variables of drought duration and drought intensity and the interdependent structure of the two variables of drought duration and drought intensity are considered, the drought duration and drought intensity are comprehensively considered, and the exceedance probability of the two variables of drought duration and drought intensity is calculated. The drought risk of terrestrial ecosystems is evaluated through the exceedance probability of the two variables of drought duration and drought intensity.

[0121] Among them, the exceedance probability mainly refers to the probability of reaching or exceeding a certain threshold, and the exceedance probability of the two variables of drought duration and drought intensity is the combination of the exceedance probabilities of the two variables of drought duration and drought intensity, that is, the probability that the two variables of drought duration and drought intensity do not exceed a certain threshold. The calculation formula of the exceedance probability of the two variables of drought duration and drought intensity is as follows:

[0122] P AND (d>D con. ,in>IN con. )

[0123] =1-P(d≤D con. )-P(in≤IN con. )+P(d≤D con. ,in≤IN con. )

[0124] =1-F D (D con. )-F IN (IN con. )+F D,IN (D con. ,IN con. )

[0125] Among them, P AND (d>D con. ,in>.IN con. ) represents the exceedance probability of the two variables of drought duration and drought intensity; D con. and IN con. Represent the given drought duration and drought intensity respectively; F D () and F IN () represent the cumulative probability of the marginal distribution function of drought duration and the marginal distribution function of drought intensity respectively; F D,IN () represents the cumulative probability of the joint distribution function of drought duration and drought intensity.

[0126] When the stationarity assumption between the drought duration and drought intensity univariate series or the drought duration marginal distribution function and the drought intensity marginal distribution function does not hold, the exceedance probability of the two variables of drought duration and drought intensity will change over time. In order to make the exceedance probability under this condition more accurately represent the drought risk of terrestrial ecosystems, step S1 also includes:

[0127] S15, normalizing the calculated exceedance probability, and using the normalized exceedance probability result to characterize the drought risk of the terrestrial ecosystem;

[0128] In step S15, the expression of drought risk based on the normalized exceedance probability is as follows:

[0129]

[0130] Among them, H represents drought risk; P AND represents the exceedance probability of the two variables of drought duration and drought intensity; and They represent the maximum and minimum values ​​of the exceedance probability of the two variables, drought duration and drought intensity, respectively.

[0131] In step S2, the three-dimensional structure of the terrestrial ecosystem is considered. Based on the annual value series of vegetation cover and leaf area index obtained by satellite remote sensing, the exposure components of the terrestrial ecosystem in the horizontal and vertical directions are characterized respectively. At the same time, a non-compensatory method is used to integrate the exposure components of the terrestrial ecosystem in the horizontal and vertical directions, and the exposure of the terrestrial ecosystem is evaluated year by year. Specifically, the exposure expression based on the exposure index is as follows:

[0132] E=max(FVC,nLAI)

[0133] Among them, E represents exposure; FVC represents the annual vegetation cover; nLAI represents the annual normalized leaf area index. It should be pointed out that the non-compensatory method can retain the dominant features in the index and ensure that it is not eliminated by the disadvantageous features. That is, when either the vegetation cover or the leaf area index is large enough, the exposure of the terrestrial ecosystem is considered to be high.

[0134] It should be noted that, since the dimensions of the risk factor indicators are different, in order to make the evaluation uniform and comparable, the indicators need to be normalized. Therefore, in this embodiment, the leaf area index is preferably the normalized leaf area index.

[0135] Specifically, the calculation formula of normalized leaf area index is as follows:

[0136]

[0137] Among them, LAI max and LAI min They represent the maximum and minimum values ​​of the annual series of leaf area index respectively.

[0138] Furthermore, in step S3, considering the drought duration-drought intensity combinations with different return periods, the conditional probability of loss of the three vegetation indices under the influence of drought is calculated. The conditional probability formula of loss of the three vegetation indices is as follows:

[0139]

[0140] Among them, P(VI lower ≤vi≤VI upper |d=D,in=IN) represents the conditional probability of loss of three vegetation indices; vi represents the vegetation index; VI lower Represents the lower bound of the vegetation index; VI upper represents the upper bound of the vegetation index; f(vi|d,in) represents the conditional probability density function of the loss of the three vegetation indices; NDVI represents the vegetation difference index; LAI represents the leaf area index; GPP represents the gross primary productivity index;

[0141] The solution formula for the three vegetation index loss conditional probability density function f(vi|d,in) is as follows:

[0142]

[0143] Where c[] represents the Copula density function; f VI (vi) is the probability density function of the vegetation index; f(d,in) is the probability density function of the two variables of drought duration and drought intensity; f(vi,d,in) is the probability density function of the three variables of drought duration, drought intensity and vegetation index; F VI (vi) is the cumulative probability function of the vegetation index; F D (d) is the cumulative probability function of drought duration; F IN (in) is the cumulative probability function of drought intensity.

[0144] After f(vi|d,in) is calculated by the above solution formula, then in [VI lower , VI uppper By integrating f(vi|d,in) over the interval ], we can get the conditional probability of loss of the three vegetation indices P(VI lower ≤vi≤VI upper |d=D,in=IN).

[0145] It should be noted that, due to the different dimensions of the risk factor indicators, in order to make the evaluation uniform and comparable, the indicators need to be normalized. Therefore, in this embodiment, the vegetation difference index is preferably the normalized vegetation difference index, the leaf area index is preferably the normalized leaf area index, and the gross primary productivity index is preferably the normalized gross primary productivity index; the lower bounds of the vegetation indices corresponding to the vegetation difference index NDVI, the leaf area index LAI and the gross primary productivity index GPP are VI: lower Can be set to -1, 0 and 1 respectively, the upper limit of the vegetation index of each type of vegetation index VI upper It can be set to the 50th percentile (ie median) of the long-term observation value of each type of vegetation index.

[0146] The vulnerability of terrestrial ecosystems is quantitatively characterized by the conditional probability of loss of three vegetation indices. The larger the conditional probability of loss of the vegetation index, the higher the vulnerability. Specifically, the vulnerability expression based on the conditional probability of loss of different types of vegetation indices is as follows:

[0147]

[0148] Among them, vul NDVI 、vul LAI and vul GPP They represent the conditional probability of loss based on vegetation difference index NDVI, leaf area index LAI and gross primary productivity index GPP respectively; and They represent the minimum and maximum values ​​of the conditional probability of vegetation loss based on the vegetation difference index NDVI; and They represent the minimum and maximum values ​​of the conditional probability of vegetation loss based on the leaf area index LAI; and They represent the minimum and maximum values ​​of the conditional probability of vegetation loss based on the gross primary productivity index GPP, respectively.

[0149] Furthermore, in order to enable the calculation result of the vegetation index loss conditional probability to more accurately characterize the vulnerability of the terrestrial ecosystem, step S3 further includes:

[0150] S31, calculate the arithmetic mean of the conditional probability of loss of the three vegetation indices, and use the calculation results of the arithmetic mean to characterize the vulnerability of terrestrial ecosystems;

[0151] In step S31, the expression of vulnerability based on the arithmetic mean of the conditional probability of loss of three vegetation indices is as follows:

[0152] V=mean(vul NDVI ,vul LAI ,vulGPP )

[0153] Among them, V represents vulnerability; mean(vul NDVI ,vul LAI ,vul GPP ) represents the arithmetic mean of the conditional probability of loss of the three vegetation indices: NDVI, LAI and GPP. It should be noted that for ease of understanding, the mean (vul NDVI ,vul LAI ,vul GPP ) can also be written as (vul NDVI +vul LAI +vul GPP ) / 3.

[0154] In step S4, the time-varying risk of terrestrial ecosystems is measured in years, and the expression of the time-varying risk assessment algorithm is as follows:

[0155] Risk i =H i ×E i ×V i

[0156] Among them, Risk i represents the time-varying risk of the terrestrial ecosystem in year i; H i 、E i and V i represent the drought risk, exposure and vulnerability of the terrestrial ecosystem in the i-th year, respectively. i 、E i and V i Take the index calculation results corresponding to the i-th year in steps S1, S2 and S3 respectively.

[0157] Furthermore, in order to more intuitively analyze the relationship between the time-varying risk of terrestrial ecosystems in the target year and the corresponding target year, the method also includes:

[0158] Step S5, linearly fitting the calculated time-varying risk of the terrestrial ecosystem in the target year and its corresponding target year to simulate the linear relationship between the two.

[0159] Specifically, the calculated time-varying risk of terrestrial ecosystems in the target year and the corresponding target year are used as the dependent variable and independent variable, respectively, to simulate the linear relationship between the two. The linear relationship expression is as follows:

[0160] Risk i =at+b

[0161] Where t is time; a and b represent the slope and intercept respectively. a is used to intuitively represent the annual change rate of the time-varying risk of terrestrial ecosystems. When a>0, it indicates that the time-varying risk of terrestrial ecosystems increases with the increase of years; when a<0, it indicates that the time-varying risk of terrestrial ecosystems decreases with the increase of years.

[0162] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. All equivalent structural changes made using the contents of the present invention's specification and drawings, or directly / indirectly applied in other related technical fields, are included in the patent protection scope of the present invention.

Claims

1. A method for assessing time-varying risk of terrestrial ecosystems under drought stress, characterized in that: The steps include: S1. Calculate the exceedance probability of the two variables of drought duration and drought intensity, and characterize the drought risk of terrestrial ecosystems with the calculated results of the exceedance probability; S2. constructing a two-dimensional terrestrial ecosystem exposure index formula based on the vegetation cover and leaf area index, solving the exposure index formula to obtain the calculation result of the exposure index, and using the calculation result of the exposure index to characterize the exposure of the terrestrial ecosystem; S3. construct a three-vegetation index loss conditional probability formula based on the three vegetation indices of vegetation difference index, leaf area index and total primary productivity index, solve the three-vegetation index loss conditional probability formula, obtain the calculation results of the three-vegetation index loss conditional probability, and use the calculation results of the three-vegetation index loss conditional probability to characterize the vulnerability of terrestrial ecosystems; S4. Calculate the time-varying risk of the terrestrial ecosystem under drought stress according to a preset time-varying risk assessment algorithm and in combination with the three risk factors of drought hazard, exposure and vulnerability; The vulnerability expression based on the conditional probability of loss of different types of vegetation indices is as follows: Among them, vul NDVI 、vul LAI and vul GPP They represent the conditional probability of loss based on vegetation difference index NDVI, leaf area index LAI and gross primary productivity index GPP respectively; and They represent the minimum and maximum values ​​of the conditional probability of vegetation loss based on the vegetation difference index NDVI; and They represent the minimum and maximum values ​​of the conditional probability of vegetation loss based on the leaf area index LAI; and They represent the minimum and maximum values ​​of the conditional probability of vegetation loss based on the gross primary productivity index GPP; Step S3 also includes: S31. Calculate the arithmetic mean of the conditional probability of loss of the three vegetation indices, and use the calculation result of the arithmetic mean to characterize the vulnerability of the terrestrial ecosystem; In step S31, the vulnerability expression based on the arithmetic mean of the conditional probabilities of the three vegetation index losses is as follows: V=mean(vul NDVI I want LAI I want GPP ) Among them, V represents vulnerability; mean(vul NDVI ,vul LAI , vul GPP ) represents the arithmetic mean of the conditional probability of loss of three vegetation indices: vegetation difference index NDVI, leaf area index LAI and gross primary productivity index GPP.

2. The method for time-varying risk assessment of terrestrial ecosystems under drought stress according to claim 1, characterized in that: The step S1 comprises: S11. Extract drought events and their drought attributes from the standardized precipitation index sequence by using the run theory, the drought attributes including drought duration, drought intensity, drought start time and drought end time; combine the interval time and exceedance criterion theory, merge the interrelated drought events and eliminate the micro-drought events to obtain independent drought event samples; S12. Use the KPSS test method to test whether the drought duration and drought intensity univariate sequences of the independent drought events maintain stability; when the drought duration and drought intensity univariate sequences are stable, use a variety of probability distribution functions with constant parameters to fit the drought duration and drought intensity, and evaluate the goodness of fit of the probability distribution functions based on the BIC criterion to obtain the optimal marginal distribution function of drought duration and drought intensity; when the drought duration and drought intensity univariate sequences are not stable, use a variety of probability distribution functions with time-varying parameters to fit the drought duration and drought intensity, and estimate the time-varying parameters of the probability distribution functions based on the generalized additive model and evaluate the goodness of fit of the time-varying parameters of the probability distribution functions based on the BIC criterion to obtain the optimal marginal distribution function of drought duration and drought intensity; S13. Use the Copula function likelihood ratio method to test whether the stationarity assumption between the marginal distribution function of drought duration and the marginal distribution function of drought intensity holds true; when the stationarity assumption holds true, use multiple Copula joint distribution functions with constant parameters to fit the joint distribution function of drought duration and drought intensity, and evaluate the goodness of fit of the Copula joint distribution function based on the BIC criterion to obtain the optimal joint distribution function of drought duration and drought intensity; when the stationarity assumption does not hold true, use multiple Copula joint distribution functions with time-varying parameters to fit the joint distribution function of drought duration and drought intensity, and evaluate the goodness of fit of the Copula joint distribution function based on the BIC criterion to obtain the optimal joint distribution function of drought duration and drought intensity; S14. According to the preset exceedance probability calculation formula and in combination with the drought duration marginal distribution function, the drought intensity marginal distribution function and the joint distribution function of the drought duration and drought intensity, the exceedance probabilities of the two variables of the drought duration and drought intensity are calculated.

3. The method for time-varying risk assessment of terrestrial ecosystems under drought stress according to claim 2, characterized in that: In step S12, when the drought duration and drought intensity univariate series stationarity is established, multiple probability distribution functions with constant parameters are used to fit the drought duration and drought intensity respectively, and the goodness of fit of the probability distribution function is evaluated based on the BIC criterion, and the optimal marginal distribution function of drought duration and the marginal distribution function of drought intensity are obtained, including: Seven probability distribution functions with constant parameters, including exponential, normal, Gumbel, logistic, lognormal, gamma and Weibull, were selected as alternatives. Drought duration and drought intensity were fitted by the seven probability distribution functions, and the goodness of fit of all the alternative probability distribution functions was evaluated by the BIC criterion. The calculation formula of BIC criterion is as follows: Where L(·) represents the log-likelihood function; and m represent the estimated value and number of parameters of probability distribution, respectively; num represents the total number of drought times; Calculate the BIC values ​​of all candidate probability distribution functions, then select the probability distribution function that minimizes the calculated value of BIC among all the candidate probability distribution functions as the optimal probability distribution function, and use the optimal probability distribution function as the optimal drought duration marginal distribution function and drought intensity marginal distribution function; In step S12, when the drought duration and drought intensity univariate series stationarity does not hold, multiple probability distribution functions with time-varying parameters are used to fit the drought duration and drought intensity respectively, and the time-varying parameters of the probability distribution function are estimated based on the generalized additive model and the goodness of fit of the probability distribution function is evaluated based on the BIC criterion, and the optimal marginal distribution function of drought duration and the marginal distribution function of drought intensity are obtained, including: Seven probability distribution functions with time-varying parameters, including exponential, normal, Gumbel, logistic, lognormal, gamma and Weibull, were selected as candidates. The drought duration and drought intensity were fitted by the seven probability distribution functions respectively. The generalized additive model was introduced to estimate the time-varying parameters of the probability distribution function. The drought onset time was taken as the explanatory variable, and the connection function between the time-varying parameters of the probability distribution function and the explanatory variables was established. The expression of the connection function is as follows: Where θ is a column vector of time-varying parameters of the probability distribution function, with a length of n; g() is a link function; X is a known matrix of n×J, corresponding to the explanatory variable values ​​of n time periods; β=(β1,β2,K,β J ) T is the parameter column vector to be solved, with dimension J; Z j is n×q j The known design matrix of γ j is a column vector of random variables with dimension equal to q j ; The CG and RS algorithms are used to solve the generalized additive model and calculate the probability distribution parameters corresponding to all alternative probability distribution functions that change over time; The goodness of fit of the probability distribution parameters corresponding to all the above alternative probability distribution functions is evaluated by the BIC criterion; Calculate the BIC values ​​of the probability distribution parameters corresponding to all alternative probability distribution functions, and then select the probability distribution function that minimizes the calculated value of BIC among all the alternative probability distribution functions as the optimal probability distribution function, and use the optimal probability distribution function as the optimal drought duration marginal distribution function and drought intensity marginal distribution function.

4. The method for time-varying risk assessment of terrestrial ecosystems under drought stress according to claim 2, characterized in that: In step S14, the calculation formula of the exceedance probability is as follows: P AND (d>D con. ,in>IN con. ) =1-P(d≤D con. )-P(in≤IN con. )+P(d≤D con. ,in≤IN con. ) =1-F D (D con. )-F IN (IN con. )+F D,IN (D con. ,IN con. ) Among them, P AND (d>D con. ,in>.IN con. ) represents the exceedance probability of the two variables of drought duration and drought intensity; D con. and IN con. Represent the given drought duration and drought intensity respectively; F D () and F IN () represent the cumulative probability of the marginal distribution function of drought duration and the marginal distribution function of drought intensity respectively; F D,IN () represents the cumulative probability of the joint distribution function of drought duration and drought intensity.

5. The method for time-varying risk assessment of terrestrial ecosystems under drought stress according to claim 4, characterized in that: Step S1 also includes: S15. normalizing the calculated exceedance probability, and using the normalized exceedance probability result to characterize the drought risk of the terrestrial ecosystem; In step S15, the expression of drought risk based on the normalized exceedance probability is as follows: Among them, H represents drought risk; P AND represents the exceedance probability of the two variables of drought duration and drought intensity; and They represent the maximum and minimum values ​​of the exceedance probability of the two variables, drought duration and drought intensity, respectively.

6. The method for time-varying risk assessment of terrestrial ecosystems under drought stress according to claim 1, characterized in that: In step S2, the exposure expression based on the exposure index is as follows: E=max(FVC,nLAI) Among them, E represents exposure; FVC represents annual vegetation cover; nLAI represents annual normalized leaf area index.

7. The method for time-varying risk assessment of terrestrial ecosystems under drought stress according to claim 1, characterized in that: In step S3, the three vegetation index loss conditional probability formula is as follows: Among them, P(VI lower ≤vi≤VI upper |d=D,in=IN represents the conditional probability of loss of three vegetation indices; vi represents the vegetation index; VI lower Represents the lower bound of the vegetation index; VI upper represents the upper limit of the vegetation index; f(vi|d,in) represents the conditional probability density function of the loss of the three vegetation indices; NDVI represents the vegetation difference index; LAI represents the leaf area index; GPP represents the gross primary productivity index.

8. The method for time-varying risk assessment of terrestrial ecosystems under drought stress according to claim 1, characterized in that: In step S4, the time-varying risk of the terrestrial ecosystem is measured in years, and the expression of the time-varying risk assessment algorithm is as follows: Risk i =H i ×E i ×V i Among them, Risk i represents the time-varying risk of the terrestrial ecosystem in year i; H i 、E i and V i They represent the drought hazard, exposure and vulnerability of the terrestrial ecosystem in year i, respectively.

9. The method for time-varying risk assessment of terrestrial ecosystems under drought stress according to claim 8, characterized in that: Also includes: S5. Perform a linear fit between the calculated time-varying risk of the terrestrial ecosystem in the target year and the corresponding target year to simulate the linear relationship between the two.

Images