An evaluation method for quantifying crop drought vulnerability and yield loss thresholds

Abstract

The application discloses an evaluation method for quantifying crop drought vulnerability and yield loss threshold, comprising S1: data preparation and preprocessing, S2: multi-scale drought index calculation and optimal index screening, S3: edge distribution fitting and Copula joint distribution modeling, S4: drought vulnerability probabilistic evaluation and S5: two threshold quantitative extraction based on the joint probability distribution model. The application effectively removes the trend influence of non-climatic factors such as agricultural technology and management on yield formation by detrending the historical yield per unit of crops, and the obtained "relative meteorological yield" more purely reflects the role of climate fluctuation, especially drought stress, fundamentally solving the problem that the climate signal is distorted and the evaluation benchmark is not unified due to the use of original yield data in the traditional method.

Description

Technical Field

[0001] This invention relates to the field of agricultural meteorological disaster risk assessment technology, and in particular to an assessment method for quantifying crop drought vulnerability and yield loss thresholds. Background Technology

[0002] Drought is a major meteorological disaster affecting crop yields and food security globally. To assess the impact of drought, a series of methods have been developed, including: empirical assessment methods based on historical disaster data, statistical models based on simple correlation analysis between meteorological indicators and yield, and simulation models utilizing crop growth mechanisms. Among these, regression or correlation analysis between drought indicators such as the Standardized Precipitation Index (SPI) and yield is a common method for assessing agricultural drought. Furthermore, statistical methods such as the Copula function are also applied in the joint analysis of multiple hydrological and meteorological variables.

[0003] However, the aforementioned existing technologies have significant shortcomings when applied to the quantitative assessment and threshold determination of crop drought vulnerability: First, most methods directly use raw yield sequences that include factors such as technological advancements, failing to isolate the influence of non-climate factors. This masks the response signals to climate stress, leading to biased assessment results. Second, assessment conclusions are mostly qualitative or based on deterministic relationships, such as simply defining the yield reduction range corresponding to drought levels, failing to provide probabilistic information needed for risk decision-making, such as "the probability of a specific loss occurring under given drought conditions." Third, for the two key management questions of "how much drought will lead to a specific loss" or "what is the maximum possible loss under a specific drought condition," existing methods lack objective and quantitative threshold extraction methods based on rigorous probability theory, relying heavily on subjective experience, resulting in poor spatial comparability and operability of threshold results. Fourth, existing methods typically employ uniform models or indicators, making it difficult to finely characterize the heterogeneity of drought response mechanisms in different regions and crop types. Summary of the Invention

[0004] The purpose of this invention is to overcome the aforementioned deficiencies in existing technologies and provide a method for quantifying crop drought vulnerability and yield loss thresholds. This method aims to achieve an objective and quantitative assessment of crop drought risk through precise data preprocessing, optimal indicator selection, joint probability modeling, and inversion of drought-induced disaster thresholds and potential loss thresholds, thus providing accurate scientific basis for disaster prevention and mitigation.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: An assessment method for quantifying crop drought vulnerability and yield loss threshold includes the following steps: S1: Data Preparation and Preprocessing Acquire meteorological and crop yield data for the target area. Perform detrending processing on historical crop yield data and calculate the relative meteorological yield reflecting the impact of climate.

[0006] S2: Calculation of Multiscale Drought Indicators and Screening of Optimal Indicators Standardized precipitation indices (SPIs) were calculated for four time scales: 1 month, 3 months, 6 months, and 12 months, to construct a multi-scale drought index system. Correlation analysis was used to select the SPI index with the highest correlation to the relative meteorological yield as the optimal drought index for this crop in the region.

[0007] S3: Marginal Distribution Fitting and Copula Joint Distribution Modeling Marginal distributions are fitted to the relative meteorological yield sequence and the optimal drought index sequence, respectively, and the optimal marginal distribution is selected from a preset distribution family. Based on the fitted marginal distribution, a two-dimensional joint probability distribution model of the relative meteorological yield and the optimal drought index is constructed using a Copula function, and the optimal model is selected from multiple Copula functions using the Akaike information content criterion.

[0008] S4: Probabilistic Assessment of Drought Vulnerability Based on the aforementioned joint probability distribution model, the conditional probabilities of different preset yield reduction rates under different preset drought levels are calculated. These conditional probabilities are then statistically analyzed to create spatial zoning maps and vulnerability curves for crops under different drought levels, achieving a probabilistic and spatial assessment of crop drought vulnerability.

[0009] S5: Quantitative extraction of drought-induced disaster threshold and yield loss threshold; S5.1: Drought-induced disaster threshold extraction: For each preset yield reduction rate, in the joint probability distribution model, find the drought index corresponding to the conditional probability of the yield reduction rate reaching a preset probability value (such as 0.9), and define this value as the drought-induced disaster threshold that triggers the yield reduction rate under this probability.

[0010] S5.2: Yield loss threshold extraction: For each preset drought scenario (i.e., a fixed drought index), in the joint probability distribution model, find the maximum yield reduction rate corresponding to the conditional probability reaching a preset probability value (e.g., 0.9), and define this value as the maximum yield loss threshold that the drought scenario may cause under this probability.

[0011] Preferably, the detrending process uses a moving average method to calculate the trend output.

[0012] Preferably, the candidate distribution family for fitting the marginal distribution includes extreme value distribution, generalized extreme value distribution, normal distribution, logistic distribution, and... Student's t distributed.

[0013] Preferably, the Copula function includes Gaussian Copula, Student's t Copula, Frank Copula, Gumbel Copula and Clayton Copula.

[0014] Preferably, the drought level is divided into extreme drought, severe drought, moderate drought, and mild drought based on the SPI value.

[0015] The present invention has the following beneficial effects: 1. This invention effectively removes the trend influence of non-climate factors such as agricultural technology and management on yield formation by detrending the historical crop yield data sequence. The resulting "relative meteorological yield" more purely reflects the effect of climate fluctuations, especially drought stress, and fundamentally solves the problems of climate signal distortion and inconsistent evaluation benchmarks caused by the use of raw yield data in traditional methods.

[0016] 2. This invention utilizes the Copula function to construct a joint probability distribution of relative meteorological yield and optimal drought index, and calculates conditional probabilities based on this distribution, thus achieving a leap from traditional deterministic assessment to probabilistic risk assessment. The assessment results can clearly answer "how likely it is that crop yield reduction will reach a certain level under specific drought conditions." This probabilistic information is more in line with the uncertain nature of natural disasters, greatly enhancing the scientific validity of the assessment conclusions and their application value in risk pricing and decision optimization.

[0017] 3. This invention creatively proposes a technical approach for quantitatively extracting drought-induced disaster thresholds and potential loss thresholds based on a joint probability model. On the one hand, for a specific yield reduction rate, the drought intensity threshold required to reach a preset occurrence probability can be derived; on the other hand, for a specific drought scenario, the maximum possible loss threshold when the preset occurrence probability is reached can be derived. This process is entirely based on objective probability calculations, replacing subjective experience-based judgments, resulting in threshold results with clear statistical confidence and strong spatial comparability. It can be directly used to guide the formulation of drought relief plan activation standards and disaster loss prediction.

[0018] 4. This invention, through a special screening strategy and by fitting marginal distributions and Copula functions to different sequences, fully respects and characterizes the uniqueness and nonlinearity of drought responses in different regions and crop types. This method overcomes the limitations of "one-size-fits-all" models, significantly improving the refinement and accuracy of the evaluation results, and is able to generate vulnerability zoning maps and threshold distribution maps that truly reflect spatial heterogeneity. Attached Figure Description

[0019] Figure 1 This is a flowchart of the evaluation method of the present invention; Figure 2 This is a schematic diagram illustrating the joint probability distribution relationship between the optimal drought index and the relative meteorological yield of crops. Figure 3 This is a schematic diagram of the spatial distribution of vulnerabilities. Figure 4 The trend of crop yield loss probability under different drought scenarios. Detailed Implementation

[0020] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0021] Example: A method for assessing crop drought vulnerability and yield loss threshold in a certain region. like Figure 1 As shown, the quantitative assessment method for crop drought vulnerability and yield loss threshold described in this embodiment includes, in sequence, data preparation and preprocessing, calculation of drought indicators at multiple time scales and screening of optimal drought indicators, joint probability distribution modeling, probabilistic assessment of drought vulnerability, and quantitative extraction of two thresholds.

[0022] S1: Data Preparation and Preprocessing Multi-year daily or monthly precipitation data from multiple meteorological stations within the study area, along with historical yield data for the target crop within the corresponding area, are obtained. The historical yield data is detrended using a moving average method to obtain the crop trend yield. Based on the relationship between actual yield and trend yield, a relative meteorological yield sequence reflecting the influence of meteorological conditions is calculated.

[0023] S2: Calculation of drought indices at multiple timescales and selection of optimal drought indices Based on the precipitation data, standardized precipitation indices at multiple time scales are calculated to construct a multi-time-scale drought index set. By calculating the correlation coefficient between each drought index and the relative meteorological yield, the drought index with the highest correlation to the relative meteorological yield is selected as the optimal drought index for this crop in the region.

[0024] S3: Marginal Distribution Fitting and Joint Probability Distribution Model Construction Marginal distribution fitting is performed on the relative meteorological yield sequence and the optimal drought index sequence respectively to determine their respective marginal distribution functions. Based on this, a joint probability distribution model of the two is constructed to describe the probabilistic dependence between the drought index and crop yield loss. The joint probability distribution relationship is as follows: Figure 2 As shown, Figure 2This diagram illustrates the joint probability distribution relationship, used to demonstrate the dependency structure between drought indicators and crop yield loss through a joint probability model, and to provide a basis for subsequent conditional probability calculations.

[0025] S4: Probabilistic Assessment of Drought Vulnerability Based on the joint probability distribution model, the conditional probabilities of crop yield reduction at different preset drought levels are calculated. Statistical processing of these conditional probabilities yields crop drought vulnerability assessment results under corresponding drought scenarios, and a vulnerability spatial distribution diagram can be generated based on these assessment results, such as... Figure 3 As shown.

[0026] S5: Quantitative Extraction of Drought-Induced Thresholds and Potential Loss Thresholds S5.1 Drought-induced disaster threshold extraction For a preset crop yield reduction rate, in the joint probability distribution model, under the constraint that the conditional probability changes monotonically with the drought index, the drought index is inverted and searched. When the conditional probability of the crop experiencing the yield reduction rate reaches a preset probability threshold, the corresponding drought index value is determined as the drought disaster threshold corresponding to the yield reduction rate.

[0027] S5.2 Extraction of Yield Loss Threshold For a pre-defined drought scenario, in the joint probability distribution model, under the constraint that the conditional probability changes monotonically with the yield reduction rate, the yield reduction rate is inverted and searched. When the conditional probability reaches a pre-defined probability threshold, the corresponding yield reduction rate is determined as the maximum yield loss threshold that may be caused under the drought scenario.

[0028] In some implementations, a schematic diagram of crop vulnerability change curves can also be generated based on the conditional probability calculation results, such as... Figure 4 As shown, this helps to understand the changing trends of crop yield loss probability under different drought scenarios.

[0029] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for quantifying crop drought vulnerability and yield loss threshold, characterized in that, Includes the following steps: S1: Detrend the historical crop yield data to obtain the relative meteorological yield; S2: Calculate the standardized precipitation index (SPI) at multiple time scales and select the optimal drought index that has the highest correlation with the relative meteorological yield. S3: Perform marginal distribution fitting on the relative meteorological yield sequence and the optimal drought index sequence respectively, and construct a two-dimensional joint probability distribution model of the two using the Copula function; S4: Based on the joint probability distribution model, calculate the conditional probability of different preset yield reduction rates under different drought levels, and assess crop drought vulnerability based on the conditional probability; S5: Based on the joint probability distribution model, two thresholds are extracted, the two thresholds being: S5.1: For the preset yield reduction rate, determine the drought index value corresponding to the conditional probability of its occurrence reaching the first preset probability, and use it as the drought disaster threshold. S5.2: For a preset drought scenario, determine the maximum yield reduction rate corresponding to the conditional probability reaching the second preset probability, and use it as the yield loss threshold.

2. The method according to claim 1, characterized in that, In step S1, the detrending process specifically involves: using the moving average method to separate the trend yield from the historical crop yield data, then calculating the difference between the actual yield and the trend yield to obtain the meteorological yield, and dividing the meteorological yield by the trend yield to obtain the relative meteorological yield.

3. The method according to claim 1, characterized in that, In step S2, the multiple time scales include 1 month, 3 months, 6 months, and 12 months.

4. The method according to claim 1 or 3, characterized in that, In step S2, the screening specifically involves: calculating the correlation coefficient between the relative meteorological yield and the SPI values ​​of each month at each time scale, and determining the SPI index with the highest positive correlation coefficient with the relative meteorological yield as the optimal drought index.

5. The method according to claim 1, characterized in that, In step S3, when fitting the marginal distribution, a selection is made from a family of candidate distributions including extreme value distribution, generalized extreme value distribution, normal distribution, logistic distribution and Student's distribution, and the optimal marginal distribution is determined by the Kolmogorov-Smirnov test.

6. The method according to claim 1, characterized in that, In step S3, the Copula function is selected from GaussianCopula, Student's t The set consists of Copula, Frank Copula, Gumbel Copula, and Clayton Copula, and the optimal Copula function is selected from the set using the Akaike Information Criterion to construct the joint probability distribution model.

7. The method according to claim 1, characterized in that, In step S4, the drought level is divided according to the value of the optimal drought index, including: mild drought (-1.0 < SPI ≤ -0.5), moderate drought (-1.5 < SPI ≤ -1.0), severe drought (-2.0 < SPI ≤ -1.5), and extreme drought (SPI ≤ -2.0).

8. The method according to claim 1, characterized in that, In step S4, the assessment of crop drought vulnerability specifically includes: calculating the average value of the conditional probability of each preset yield reduction rate under a specific drought level, and performing spatial classification mapping based on the average value to generate a crop vulnerability zoning map.

9. The method according to claim 1, characterized in that, In step S5.1, the first preset probability is 0.9; in step S5.2, the second preset probability is 0.

9.

10. The method according to claim 1, characterized in that, In step S5.1, the method for determining the drought-induced disaster threshold is as follows: fix the yield reduction rate, iterate through a series of drought index values, calculate the corresponding conditional probability, and take the drought index value corresponding to the first time the conditional probability reaches or exceeds the first preset probability as the drought-induced disaster threshold under the yield reduction rate; In step S5.2, the method for determining the yield loss threshold is as follows: fix the drought index value, iterate through a series of yield reduction rates, calculate the corresponding conditional probability, and take the yield reduction rate corresponding to the first time the conditional probability reaches or exceeds the second preset probability as the maximum yield loss threshold under the drought scenario.

Images