Method for determining drought disaster-causing threshold value of vegetation

Abstract

The invention belongs to the technical field of ecological disaster risk assessment and early warning, and discloses a method for determining a vegetation drought disaster-causing threshold value, and the method comprises the steps: constructing a vegetation index sequence according to the vegetation data of each sub-region of a to-be-researched region; constructing a plurality of drought index sequences under different time scales according to the meteorological data of each sub-region of the to-be-researched region; determining a first drought index sequence corresponding to the vegetation index sequence from the plurality of drought index sequences according to correlation coefficients of the vegetation index sequence and the plurality of drought index sequences; constructing a joint distribution model between the vegetation index sequence and the first drought index sequence; and constructing a vegetation loss probability evaluation model under different drought grades based on the joint distribution model, and determining a disaster-causing threshold value of vegetation drought according to the vegetation loss probability evaluation model. According to the method, the vegetation loss probability under different drought levels can be quantified, the vegetation drought threshold value can be determined, and more accurate data support is provided for drought risk management.

Description

Technical Field

[0001] The invention discloses a method for determining a vegetation drought disaster threshold, and belongs to the technical field of ecological disaster risk assessment and early warning. Background Art

[0002] In the field of ecological and environmental science, the impact of drought on ecosystems has always been a research focus. In recent years, with the continuous changes in global climate patterns, the frequent occurrence of drought events has posed a severe challenge to natural ecosystems. Currently, research on drought thresholds has achieved a series of important results. For example, some scholars have successfully quantified the propagation threshold of meteorological drought to hydrological drought using the Copula model, providing a scientific basis for revealing the propagation mechanism of drought in different ecosystems. In addition, some studies have effectively identified the triggering conditions of hydrological drought scenarios by improving drought propagation models, providing strong support for drought early warning and risk management. At the same time, research on the aridity index of global dryland ecosystems has also provided an important reference for assessing the impact of drought on ecosystems.

[0003] However, existing research still has limitations. On the one hand, most studies focus on drought propagation thresholds or spatial gradient changes in vegetation attributes under drought conditions. However, there is a lack of in-depth research on the overall response mechanism of ecosystems under drought stress, especially the quantitative analysis of vegetation vulnerability and disaster thresholds. On the other hand, in ecosystem research, regional climate differences and the diversity of vegetation types are often not fully considered, which leads to insufficient precision in drought impact assessments. Especially in areas with complex climate conditions and diverse vegetation types, existing research methods often find it difficult to fully reveal the impact mechanism of drought on ecosystems, and thus cannot provide a strong scientific basis for ecological protection and restoration. Summary of the Invention

[0004] The present invention aims to provide a method for determining the threshold for vegetation drought disasters, addressing the existing technical difficulties in fully revealing the response mechanism of vegetation to drought and subsequently accurately determining the threshold for vegetation drought disasters. To achieve this objective, the present invention proposes a method for determining the threshold for vegetation drought disasters, the specific scheme of which is as follows:

[0005] A method for determining a vegetation drought disaster threshold comprises the following steps:

[0006] Step 1: constructing a vegetation index sequence based on vegetation data of each sub-region of the area to be studied; constructing multiple drought index sequences at different time scales based on meteorological data of each sub-region of the area to be studied;

[0007] Step 2: determining a first drought index sequence corresponding to the vegetation index sequence from the multiple drought index sequences based on the correlation coefficients between the vegetation index sequence and the multiple drought index sequences;

[0008] Step 3: constructing a joint distribution model between the vegetation index sequence and the first drought index sequence;

[0009] Step 4: construct a vegetation loss probability assessment model under different drought levels based on the joint distribution model, and determine the disaster threshold of vegetation drought according to the vegetation loss probability assessment model.

[0010] Preferably, after step 1 and before step 2, the following steps are further included:

[0011] By using the wavelet analysis method, the periodicity test of multiple drought index series at different time scales was carried out, and multiple drought index series that met the preset period requirements were screened out.

[0012] A trend test is performed on the vegetation index sequence and the multiple drought index sequences that meet the preset period requirement, respectively, to screen out the vegetation index sequence and the multiple drought index sequences that meet the preset trend.

[0013] Preferably, trend testing is performed on the vegetation index sequence and the plurality of drought index sequences meeting the preset period requirements, specifically including:

[0014] Calculating the statistics of the vegetation index sequence and the selected drought index sequences respectively according to the Mann-Kendall trend test method;

[0015] The trends of the vegetation index sequence and the selected drought index sequences are determined based on the statistic and a preset significance level.

[0016] Preferably, the step 2 specifically includes:

[0017] Determining multiple groups of correlation coefficients between the selected vegetation index sequence and the multiple drought index sequences using a Pearson correlation coefficient method;

[0018] A first drought index sequence corresponding to the vegetation index sequence is determined according to the multiple groups of correlation coefficients.

[0019] Preferably, determining the relevant drought index sequence corresponding to the vegetation index sequence according to the multiple sets of correlation coefficients specifically includes:

[0020] Selecting multiple correlation coefficients that meet the first preset requirement in each group of correlation coefficients;

[0021] The drought index sequence corresponding to the maximum correlation coefficient among the multiple correlation coefficients is recorded as a first drought index sequence corresponding to the vegetation index sequence.

[0022] Preferably, the step 3 specifically includes:

[0023] respectively constructing a first marginal distribution function of the vegetation index sequence and a second marginal distribution function of the first drought index sequence;

[0024] A joint distribution model of the first marginal distribution function and the second marginal distribution function is constructed by using a copula function.

[0025] Preferably, a vegetation loss probability assessment model under different drought levels is constructed based on the joint distribution model, specifically including:

[0026] According to the drought grade classification standard, the relevant drought index sequence is divided into a plurality of different drought grades;

[0027] According to the joint distribution model, vegetation loss probability assessment models under different drought levels are constructed based on conditional probability.

[0028] Preferably, determining the disaster threshold of vegetation drought according to the vegetation loss probability assessment model specifically includes:

[0029] Setting standards for grading vegetation loss;

[0030] Based on the vegetation loss probability assessment model and Bayesian theory, reversely inferring and iteratively calculating the drought index corresponding to the predetermined probability level in the vegetation loss classification standard;

[0031] The drought level interval to which the drought index belongs is recorded as the disaster threshold of vegetation drought.

[0032] Preferably, the step 4 further includes:

[0033] A spatial distribution map of the disaster threshold value of the area to be studied is generated according to the disaster threshold value of vegetation drought.

[0034] Preferably, the step 1 specifically includes:

[0035] Obtain long-term vegetation and meteorological data for the study area;

[0036] Constructing a vegetation index sequence based on vegetation data of each sub-region of the area to be studied; constructing multiple drought index sequences at different time scales based on meteorological data of each sub-region of the area to be studied;

[0037] The vegetation index sequence and the multiple drought index sequences are temporally and spatially aligned.

[0038] Beneficial Effects: By comprehensively considering the vegetation index and drought index, and applying statistical analysis and probabilistic assessment techniques, this method can more scientifically and accurately define the risk threshold for drought-related vegetation damage. Compared to traditional single-index monitoring methods, this method can more comprehensively reflect the complex response of vegetation to drought, thereby improving the accuracy and reliability of drought warnings.

[0039] By combining time series analysis techniques, such as wavelet analysis and the Mann-Kendall trend test, we can effectively screen drought index series that match the climate characteristics of the study area and determine whether there is a significant trend in the series. This helps us more accurately understand the development patterns and trends of drought events, providing a strong scientific basis for drought early warning and response.

[0040] Furthermore, by constructing a joint distribution model and a vegetation loss probability assessment model, the method can quantify the probability of vegetation loss under different drought levels, providing more accurate data support for drought risk management. Furthermore, the use of Bayesian theory for reverse deduction and iterative calculation of the disaster threshold makes the determined disaster threshold more scientific and reasonable.

[0041] Finally, the proposed method generates a spatial distribution map of disaster thresholds for the study area, visually demonstrating the drought risk levels and spatial distribution characteristics of different regions. This helps decision-makers better understand the distribution of drought risk and develop more effective drought response strategies and measures, thereby protecting ecosystems and promoting sustainable development. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 Flowchart of an embodiment of the present invention. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and do not limit the scope of protection of the present invention.

[0044] like Figure 1 As shown, the present invention takes Inner Mongolia as an example to specifically illustrate a specific scheme of a method for determining a vegetation drought disaster threshold:

[0045] It should be noted that due to the diverse vegetation types in Inner Mongolia, including grasslands, forests, and shrublands, these different vegetation types may have significantly different responses to drought. Therefore, this example divides Inner Mongolia into different subregions based on vegetation type. In the following steps, the area to be studied is the method for determining the disaster threshold for each subregion.

[0046] Step 1: construct a vegetation index sequence based on vegetation data of the area to be studied; construct multiple drought index sequences at different time scales based on meteorological data of the area to be studied;

[0047] First, obtain long-term vegetation and meteorological data for the study area. Meteorological data includes precipitation, temperature, and soil moisture. Specifically, in this example, monthly precipitation and soil moisture data at a spatial resolution of 0.25° × 0.25° were obtained from the Global Land Data Assimilation System (GLDAS). The corresponding vegetation data was also obtained using NDVI data from 1982 to 2015.

[0048] Subsequently, the meteorological and vegetation data were inspected and preprocessed, including cloud masking, radiometric correction, and outlier correction, to ensure data accuracy and completeness. Next, the collected data was organized into monthly sequences and gridded. In this example, all data was converted to the same 0.25° × 0.25° spatial resolution as the GLDAS data, and spatial matching and temporal alignment were performed to ensure that the GLDAS and NDVI data were synchronized in both time and space.

[0049] A vegetation index sequence is constructed based on vegetation data. In this embodiment, a normalized difference vegetation index sequence (NDVI sequence) is used to reflect changes in vegetation coverage in the study area. Multiple drought index sequences covering different time scales are constructed based on meteorological data. In this embodiment, the standardized precipitation evapotranspiration index (SPEI) is selected as the drought index, and a 1-24 month SPEI sequence is constructed. These sequences are used to assess the drought conditions in the study area.

[0050] In a further embodiment, after step 1 and before step 2, the method further includes:

[0051] Through the wavelet analysis method, the periodicity of multiple drought index series at different time scales was tested, and multiple drought index series that met the preset period requirements were screened out; specifically, a suitable wavelet function was selected, and the annual precipitation, temperature, and SPEI series in the Inner Mongolia region were analyzed on multiple time scales through wavelet function, wavelet transform, and wavelet variance to find the periodicity and screen out the SPEI series that met the periodicity requirements.

[0052] Furthermore, trend tests are performed on the vegetation index sequence and the multiple drought index sequences that meet the preset period requirements, respectively, to screen out the vegetation index sequence and the multiple drought index sequences that meet the preset trend.

[0053] Specifically, the statistics of the vegetation index sequence and the multiple selected drought index sequences are respectively calculated according to the Mann-Kendall trend test method; and the trends of the vegetation index sequence and the multiple selected drought index sequences are judged according to the statistics and a preset significance level value.

[0054] In this embodiment, the calculation formula of the statistic is as follows:

[0055]

[0056] in:

[0057]

[0058] Where:

[0059] U represents the standardized statistic;

[0060] τ represents the Mann-Kendall trend statistic, which reflects the relative strength of the upward and downward trends in the series;

[0061] ∑d i represents all x i <x j , the number of ordered pairs with j>i;

[0062] d i represents x of the i-th data point in the sequence i ;

[0063] Var(τ) represents the variance of the trend statistic τ, which is used to standardize τ to obtain U;

[0064] x i represents the observation value of the i-th data point in the sequence;

[0065] x j represents the observed value of the jth data point in the sequence, and j>i;

[0066] n represents the total length of the sequence;

[0067] P I Indicates the relative trend of the data points in a series.

[0068] In this embodiment, the preset significance level value α is specifically 0.05, and the trend of the SPEI series and the NDVI series is judged based on the significance level value and the statistic. Specifically, under the significance level value α, if |U|<U α / 2 When ∣U∣>U α / 2When , the sequence trend is significant (U>0 for an increase, U<0 for a decrease). Sequences with insignificant trends are screened out to select SPEI and NDVI series that meet the preset trend requirements. The above steps are used to screen out sequences with periodicity and trend to avoid noise interference in subsequent analysis.

[0069] Step 2: determining a first drought index sequence corresponding to the vegetation index sequence from the multiple drought index sequences based on the correlation coefficients between the vegetation index sequence and the multiple drought index sequences;

[0070] Furthermore, step 2 specifically includes:

[0071] A Pearson correlation coefficient method is used to determine multiple groups of correlation coefficients between the selected vegetation index sequence and the multiple drought index sequences; and a first drought index sequence corresponding to the vegetation index sequence is determined according to the multiple groups of correlation coefficients.

[0072] Among them, determining the first drought index sequence corresponding to the vegetation index sequence based on the multiple groups of correlation coefficients specifically includes: selecting multiple correlation coefficients that meet a first preset requirement in each group of correlation coefficients; and recording the drought index sequence corresponding to the maximum correlation coefficient among the multiple correlation coefficients as the first drought index sequence corresponding to the vegetation index sequence.

[0073] Specifically, in this embodiment, based on the screened SPEI sequence and NDVI sequence with periodicity and trend, the correlation coefficient between the NDVI sequence and its corresponding multiple groups of SPEI sequences at different time scales is calculated to obtain a correlation coefficient matrix at different time scales. All correlation coefficients that meet the first preset requirement in each group of relative relationships are selected, and the largest correlation coefficient is selected from them. The SPEI sequence corresponding to the correlation coefficient is extracted and recorded as the first drought index sequence, which represents the SPEI sequence with the highest correlation with the NDVI sequence, reflecting the response time (RT) of vegetation to drought. According to the above steps, the difference in the corresponding time of each sub-region, that is, the region with different vegetation types in Inner Mongolia, is obtained to further reveal the response mechanism of different types of vegetation to drought.

[0074] Step 3: constructing a joint distribution model between the vegetation index sequence and the first drought index sequence;

[0075] Furthermore, step 3 specifically includes: constructing a first marginal distribution function of the vegetation index sequence and a second marginal distribution function of the first drought index sequence respectively; and constructing a joint distribution model of the first marginal distribution function and the second marginal distribution function through a Copula function.

[0076] Specifically, after determining the first drought index sequence of different vegetation type areas in Inner Mongolia (i.e., the SPEI sequence with the highest correlation with the NDVI sequence), this embodiment further constructs a joint distribution model between the NDVI sequence and the first drought index sequence. Specifically, the NDVI sequence is first fitted with a marginal distribution. Various distribution types such as normal distribution, gamma distribution, lognormal distribution, Weibull distribution and generalized extreme value distribution are selected for fitting. Parameters of these distributions are estimated using sample data, and then the optimal marginal distribution function is selected as the first marginal distribution function of the NDVI sequence. Similarly, the second marginal distribution function of the first drought index sequence is obtained.

[0077] Then, according to Sklar's theorem, a suitable copula function (such as Gaussian copula, t-copula, etc.) is used to connect the first marginal distribution function and the second marginal distribution function to construct a joint distribution model. The joint distribution model is as follows:

[0078]

[0079] Where:

[0080] F(x1, x2, ..., x n ) represents the joint distribution function, representing the random variables x1, x2, ..., x n At the same time, X1≤x1, X2≤x2, X3≤x3…, X n ≤x n The cumulative probability of

[0081] Represents the joint distribution function, which is used to transform the marginal distribution Mapping to joint distribution;

[0082] represents the nth random variable x n The marginal cumulative distribution function of u n Represents the output value of the marginal cumulative distribution function;

[0083] n represents the number of random variables in the joint distribution.

[0084] Step 4: construct a vegetation loss probability assessment model under different drought levels based on the joint distribution model, and determine the disaster threshold of vegetation drought according to the vegetation loss probability assessment model.

[0085] Furthermore, a vegetation loss probability assessment model under different drought levels is constructed based on the joint distribution model, specifically including:

[0086] According to the drought level division standard, the relevant drought index sequence is divided into multiple different drought levels; according to the joint distribution model, vegetation loss probability assessment models for different drought levels are constructed based on conditional probabilities respectively.

[0087] Furthermore, the disaster-causing threshold of vegetation drought is determined according to the vegetation loss probability assessment model, specifically including: setting the vegetation loss grading standard; based on the vegetation loss probability assessment model and Bayesian theory, inversely inferring and iteratively calculating the drought index corresponding to the preset probability level in the vegetation loss grading standard; and recording the interval of the drought level to which the drought index belongs as the disaster-causing threshold of vegetation drought.

[0088] Specifically, in this embodiment, according to the SPEI drought level division standard, the SPEI sequence is divided into mild drought, moderate drought, severe drought and extreme drought, and corresponding SPEI value ranges are set for each drought level. Specifically, mild drought (-1 < SPEI ≤ -0.5), moderate drought (-1.5 < SPEI ≤ -1), severe drought (-2 < SPEI ≤ -1.5) and extreme drought (SPEI ≤ -2). Using the joint distribution model, the probability that the NDVI value is lower than a certain threshold under different drought levels is calculated according to the conditional probability formula, that is, the vegetation loss probability.

[0089] Specifically, the vegetation loss probability under mild drought stress is:

[0090]

[0091] In the formula:

[0092] P(NDVI ≤ ndvi|-1 < SPEI ≤ -0.5) represents the conditional probability that the normalized difference vegetation index (NDVI) is less than or equal to the set ndvi value (vegetation loss threshold) at which vegetation loss occurs under the condition that the standardized precipitation evapotranspiration index is in the mild drought stress range (-1 < SPEI ≤ -0.5);

[0093] C(x) represents the joint distribution function of SPEI and NDVI;

[0094] F SPEI (x) represents the marginal cumulative distribution function of SPEI;

[0095] F NDVI (ndvi) is the marginal cumulative distribution function of the NDVI variable, representing the probability that NDVI is less than or equal to the threshold ndvi;

[0096] ndvi represents the set vegetation loss threshold.

[0097] The vegetation loss probability under moderate drought stress is:

[0098]

[0099]

[0100] Where: P(NDVI≤ndvi|-1.5<SPEI≤-1) represents the conditional probability that the Normalized Difference Vegetation Index (NDVI) is less than or equal to the ndvi value (vegetation loss threshold) at which vegetation loss occurs under the condition that the Standardized Precipitation Evapotranspiration Index is in the moderate drought stress range (-1.5 < SPEI ≤ -1).

[0101] The probability of vegetation loss under severe drought stress is:

[0102]

[0103] Where: P(NDVI≤ndvi|-2<SPEI≤-1.5) represents the conditional probability that the Normalized Difference Vegetation Index (NDVI) is less than or equal to the ndvi value (vegetation loss threshold) at which vegetation loss occurs under the condition that the Standardized Precipitation Evapotranspiration Index is in the severe drought stress range (-2 < SPEI ≤ -1.5). The probability of vegetation loss under extreme drought stress is:

[0104]

[0105] Where: P(NDVI≤ndvi|SPEU≤-2) represents the conditional probability that the Normalized Difference Vegetation Index (NDVI) is less than or equal to the ndvi value (vegetation loss threshold) at which vegetation loss occurs under the condition that the Standardized Precipitation Evapotranspiration Index is in the extreme drought stress range (SPEI ≤ -2). Specifically, for each drought level, a vegetation loss classification standard is set, which is determined based on the ndvi value, and the ndvi value is recorded as the vegetation loss threshold. The classification standard for each drought level includes NDVI < NDVI40th, NDVI < NDVI30th, NDVI < NDVI20th, NDVI < NDVI10th, etc. It should be noted that this standard can be adjusted according to the actual situation.

[0106] Calculate the probability that the NDVI value is lower than its classification standard (i.e., lower than the corresponding threshold) within the range of SPEI values for any drought level. Based on the calculation results of the conditional probability, an evaluation model for the probability of vegetation loss under different drought levels is constructed. This model can input specific SPEI values and NDVI thresholds and output the corresponding probability of vegetation loss.

[0107] After obtaining the vegetation loss probability assessment model, the team also constructs a trigger threshold framework based on Bayesian theory to reverse engineer the disaster threshold. Specifically, the SPEI value, starting from -0.5 and varying by 0.1, is substituted into the vegetation loss assessment model to calculate the probability of vegetation loss under different SPEI value ranges. When the vegetation loss probability is greater than or equal to a preset probability level (e.g., 0.7), the corresponding SPEI value range is recorded, marking the disaster threshold under that vegetation loss classification standard.

[0108] Furthermore, after step 4, the following steps are further included:

[0109] A spatial distribution map of the disaster threshold value of the area to be studied is generated according to the disaster threshold value of vegetation drought.

[0110] Disaster thresholds were calculated for each subregion. Spatially interpolated across all subregions, generating a high-resolution spatial distribution map of drought thresholds for the study area. This map visually illustrates drought thresholds for different vegetation types in Inner Mongolia under different vegetation loss classification standards.

[0111] By comprehensively considering both the vegetation index and the drought index, and applying statistical analysis and probabilistic assessment techniques, this method can more scientifically and accurately define the risk threshold for drought-related vegetation damage. Compared to traditional monitoring methods using a single drought index, this method more comprehensively reflects the complex response of vegetation to drought, thereby improving the accuracy and reliability of drought warnings.

[0112] By combining time series analysis techniques, such as wavelet analysis and the Mann-Kendall trend test, we can effectively screen drought index series that match the climate characteristics of the study area and determine whether there is a significant trend in the series. This helps us more accurately understand the development patterns and trends of drought events, providing a strong scientific basis for drought early warning and response.

[0113] Furthermore, by constructing a joint distribution model and a vegetation loss probability assessment model, the method can quantify the probability of vegetation loss under different drought levels, providing more accurate data support for drought risk management. Furthermore, the use of Bayesian theory for reverse deduction and iterative calculation of the disaster threshold makes the determined disaster threshold more scientific and reasonable.

[0114] Finally, the proposed method generates a spatial distribution map of disaster thresholds for the study area, visually demonstrating the drought risk levels and spatial distribution characteristics of different regions. This helps decision-makers better understand the distribution of drought risk and develop more effective drought response strategies and measures, thereby protecting ecosystems and promoting sustainable development.

[0115] The above descriptions are merely several embodiments of the present invention and do not constitute any form of limitation to the present invention. Although the present invention is disclosed as above in terms of preferred embodiments, they are not intended to limit the present invention. Any technician familiar with the present profession who, without departing from the scope of the technical solution of the present invention, makes slight changes or modifications using the technical contents disclosed above are equivalent to equivalent implementation cases and fall within the scope of the technical solution.

Claims

1. A method for determining a vegetation drought disaster threshold, characterized in that: The following steps are involved: Step 1: constructing a vegetation index sequence based on vegetation data of each sub-region of the area to be studied; constructing multiple drought index sequences at different time scales based on meteorological data of each sub-region of the area to be studied; Step 2: determining a first drought index sequence corresponding to the vegetation index sequence from the multiple drought index sequences based on the correlation coefficients between the vegetation index sequence and the multiple drought index sequences; Step 3: constructing a joint distribution model between the vegetation index sequence and the first drought index sequence; Step 4: construct a vegetation loss probability assessment model under different drought levels based on the joint distribution model, and determine the disaster threshold of vegetation drought according to the vegetation loss probability assessment model.

2. The determination method according to claim 1, characterized in that After step 1 and before step 2, the following steps are also included: By using the wavelet analysis method, the periodicity test of multiple drought index series at different time scales was carried out, and multiple drought index series that met the preset period requirements were screened out. A trend test is performed on the vegetation index sequence and the multiple drought index sequences that meet the preset period requirement, respectively, to screen out the vegetation index sequence and the multiple drought index sequences that meet the preset trend.

3. The determination method according to claim 2, characterized in that: Performing trend tests on the vegetation index sequence and the multiple drought index sequences that meet the preset period requirements respectively, specifically including: Calculating the statistics of the vegetation index sequence and the selected drought index sequences respectively according to the Mann-Kendall trend test method; The trends of the vegetation index sequence and the selected drought index sequences are determined based on the statistic and a preset significance level.

4. The determination method according to claim 2, characterized in that: The step 2 specifically includes: Determining multiple groups of correlation coefficients between the selected vegetation index sequence and the multiple drought index sequences using a Pearson correlation coefficient method; A first drought index sequence corresponding to the vegetation index sequence is determined according to the multiple groups of correlation coefficients.

5. The determination method according to claim 4, characterized in that: Determining a related drought index sequence corresponding to the vegetation index sequence according to the multiple sets of correlation coefficients specifically includes: Selecting multiple correlation coefficients that meet the first preset requirement in each group of correlation coefficients; The drought index sequence corresponding to the maximum correlation coefficient among the multiple correlation coefficients is recorded as a first drought index sequence corresponding to the vegetation index sequence.

6. The determination method according to claim 1, characterized in that: The step 3 specifically includes: respectively constructing a first marginal distribution function of the vegetation index sequence and a second marginal distribution function of the first drought index sequence; A joint distribution model of the first marginal distribution function and the second marginal distribution function is constructed by using a copula function.

7. The determination method according to claim 1, characterized in that: Based on the joint distribution model, a vegetation loss probability assessment model under different drought levels is constructed, specifically including: According to the drought grade classification standard, the relevant drought index sequence is divided into a plurality of different drought grades; According to the joint distribution model, vegetation loss probability assessment models under different drought levels are constructed based on conditional probability.

8. The determination method according to claim 7, characterized in that: Determining the disaster threshold of vegetation drought based on the vegetation loss probability assessment model specifically includes: Setting standards for grading vegetation loss; Based on the vegetation loss probability assessment model and Bayesian theory, reversely inferring and iteratively calculating the drought index corresponding to the predetermined probability level in the vegetation loss classification standard; The drought level interval to which the drought index belongs is recorded as the disaster threshold of vegetation drought.

9. The determination method according to claim 8, characterized in that: After step 4, the following steps are also included: A spatial distribution map of the disaster threshold value of the area to be studied is generated according to the disaster threshold value of vegetation drought.

10. The determination method according to claim 1, characterized in that: The step 1 specifically includes: Obtain long-term vegetation and meteorological data for the study area; Constructing a vegetation index sequence based on vegetation data of each sub-region of the area to be studied, and constructing a plurality of drought index sequences at different time scales based on meteorological data of each sub-region of the area to be studied; The vegetation index sequence and the multiple drought index sequences are temporally and spatially aligned.

Images