A method for assessing the climate risk resilience of dryland farmland based on the superposition of three factors

Abstract

This invention relates to the field of farmland climate assessment technology, specifically a method for calculating the climate risk resilience of dryland farmland based on a three-factor overlay. The method includes: assembling a multi-source basic dataset composed of long-term meteorological, high-resolution remote sensing, soil physicochemical, agricultural management, and location monitoring experimental data; designing a disaster hazard quantification algorithm based on the spatiotemporal characteristics of climate events; calculating the hazard index of disaster-causing factors; and separately calculating the exposure index and vulnerability index of disaster-bearing bodies. The three indices are normalized to unify data calculation standards, and the multi-factor data are integrated through a weighted overlay fusion method to generate a comprehensive climate risk resilience index for dryland farmland. This method integrates multi-source data to complete the quantification of multi-dimensional indicators, optimizes the climate risk quantification algorithm structure, and realizes a systematic calculation of the climate risk resilience of dryland farmland.

Description

Technical Field

[0001] This invention relates to the field of farmland climate assessment technology, and in particular to a method for calculating the climate risk resilience of dryland farmland based on the superposition of three factors. Background Technology

[0002] Dryland farmland has long faced the stresses of climate fluctuations, making quantitative assessment of climate risk resilience a crucial aspect of agricultural regional environmental management. Currently, conventional farmland climate risk assessments generally rely on single-type monitoring data as the basis for calculations, resulting in narrow data coverage and an inability to integrate multi-source information from meteorological remote sensing, soil physicochemical analysis, agricultural production management, and on-site experiments for collaborative analysis. Conventional disaster quantification methods employ fixed algorithmic frameworks with standardized operational logic, failing to adapt to the temporal evolution and spatial distribution characteristics of climate events. Consequently, regional differences and temporal variations in climate risk cannot be reflected in the calculation process.

[0003] The existing accounting system has limitations in its indicator settings, rarely conducting independent quantification of the exposure attributes of farmland systems, making it difficult to objectively characterize the spatial distribution and temporal coverage of climate stress effects. The sensitivity of farmland ecosystems to external climate disturbances and their self-regulating and recovering attributes lack standardized quantitative accounting methods, and related intrinsic attributes cannot be incorporated into the overall evaluation process. Different evaluation indicators have objective differences in measurement units and numerical magnitudes, and conventional processing procedures lack standardized correction steps, making it difficult to achieve unified calculation of various risk factors. Single-dimensional evaluation models cannot meet the implementation requirements of refined resilience assessment of dryland farmland. It is necessary to construct a complete process of multi-factor collaborative accounting, improve the technical framework for multi-level indicator collaborative calculation, and adapt to the implementation requirements of quantitative accounting of climate risk resilience in the complex environment of dryland farmland. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a method for calculating the climate risk resilience of dryland farmland based on the superposition of three factors.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a method for calculating the climate risk resilience of dryland farmland based on the superposition of three factors, comprising: The target dryland farmland area was acquired by acquiring a multi-source basic dataset including long-term meteorological observation data, high-resolution remote sensing image data, soil profile physicochemical property data, agricultural production management record data, and multi-year multi-point positioning monitoring test data. Based on the long-term meteorological observation data, a disaster risk quantification algorithm is used to calculate the disaster risk index. The disaster risk quantification algorithm is designed based on the temporal and frequency evolution patterns and spatial heterogeneity characteristics of climate events. Based on the high-resolution remote sensing image data, soil profile physicochemical property data, and agricultural production management record data, the exposure index of disaster-bearing bodies is calculated. The exposure index of disaster-bearing bodies characterizes the spatial exposure and temporal coverage of farmland systems under climate stress. Based on the multi-year, multi-point location monitoring test data and some of the agricultural production management record data, the vulnerability index of the disaster-bearing body is calculated. The vulnerability index of the disaster-bearing body describes the sensitivity and recovery capacity of the farmland ecosystem to yield loss under different climate stresses. The calculated hazard index of the disaster-causing factor, the exposure index of the disaster-bearing body, and the vulnerability index of the disaster-bearing body are normalized to eliminate dimensional differences. The disaster-causing factor hazard index, disaster-bearing body exposure index, and disaster-bearing body vulnerability index, after normalization, are weighted, superimposed, and fused to generate a comprehensive climate risk resilience index for dryland farmland.

[0006] As a further aspect of the present invention, the step of calculating the hazard index of disaster-causing factors using a disaster risk quantification algorithm includes: From the long-term meteorological observation data, daily value sequences of meteorological elements related to drought, high temperature, low temperature, and rainstorm are extracted; For each type of meteorological disaster, identify disaster events in the daily value sequence of the meteorological elements that exceed or fall below a preset threshold, and record the start time, end time, duration, and extreme intensity of each disaster event; The total frequency of each type of meteorological disaster event within a preset historical time period is statistically analyzed, and its average frequency per unit time is calculated. Calculate the extreme value of the process intensity for each disaster event, and calculate the average process intensity of all events for each type of meteorological disaster; Using spatial interpolation technology, the point-like long-term meteorological observation data is generated into a gridded meteorological data field covering the target dryland farmland area. Based on the gridded meteorological data field, the spatial distribution concentration of each disaster type is analyzed. By integrating the average occurrence frequency, the average process intensity, and the distribution concentration, and introducing a time trend correction coefficient, a single-hazard risk score is calculated for each type of meteorological disaster. The single-hazard risk scores of various meteorological disaster types faced by the target dryland farmland area are calculated using multi-hazard coupling to generate a comprehensive disaster-causing factor risk index.

[0007] As a further aspect of the present invention, based on the aforementioned high-resolution remote sensing image data, soil profile physicochemical property data, and agricultural production management record data, the exposure index of the disaster-bearing body is calculated, including: The high-resolution remote sensing image data is interpreted and classified to identify the spatial distribution of different land use types, including farmland, forest land, water area, and construction land, within the target dryland farmland area, and to extract the boundary, area, and spatial connectivity information of farmland plots. The available water content, soil texture, and soil layer thickness of each farmland plot are extracted from the soil profile physicochemical property data to evaluate and generate drought resistance indicators that characterize its basic water retention and drought resistance capabilities. Extract planting structure, sowing date, harvest date, and irrigation and fertilization records from the agricultural production management records, and calculate the time windows of different crops during their growth period and their consumption of climate resources; The spatial distribution, area, and spatial connectivity information of the farmland plots are overlaid and analyzed with the drought resistance index to assess the extent and degree of damage to farmland spatial entities during climate disasters and generate spatial exposure components. By matching the crop growth period window with the historical timing patterns of climate disasters, the degree of overlap between the crop growth process and the high-risk period of climate disasters in the time dimension is assessed, and a time exposure component is generated. The spatial exposure component and the temporal exposure component are weighted and combined to obtain the disaster-bearing body exposure index, which characterizes the spatiotemporal exposure features of the farmland system.

[0008] As a further aspect of the present invention, the disaster risk quantification algorithm is designed based on the temporal and frequency evolution patterns and spatial heterogeneity characteristics of climate events, and its working principle includes: A sliding time window is established to dynamically calculate the statistical characteristic values ​​of meteorological elements on the time series of the long-term meteorological observation data in order to capture the time-varying characteristics of climate state. Within each sliding time window, the intensity of meteorological disasters is fitted using extreme value distribution theory, and the disaster intensity under different return periods is estimated to quantify the extreme nature and frequency characteristics of disasters. Spatial autocorrelation analysis is introduced to calculate the spatial clustering and heterogeneity indicators of disaster intensity based on gridded meteorological data fields, so as to quantify the spatial pattern characteristics of disasters; Define a dynamic correction factor for disaster risk, which is a function of the temporal variation trend coefficient of disaster event frequency and the spatial variation coefficient of disaster intensity; The estimated disaster intensity and frequency, spatial clustering index, and the dynamic correction factor for disaster risk within the sliding window are nonlinearly combined to form a spatiotemporal dynamic risk assessment model for each grid cell. The calculation is performed by traversing all grid cells and applying the spatiotemporal dynamic hazard assessment model. The calculation results are then spatially aggregated to finally generate the hazard index of the disaster-causing factor.

[0009] As a further aspect of the present invention, based on the multi-year, multi-point location monitoring test data and some of the agricultural production management record data, a vulnerability index of the disaster-bearing body is calculated, including: From the multi-year, multi-point location monitoring test data, actual crop yield data and corresponding meteorological data for different growth stages under different climate year types and management measures were extracted; The quantitative relationship between actual crop yield data and meteorological stress factors during key growth periods was analyzed using statistical models, and a yield-meteorological stress response function was constructed. Based on the yield-meteorological stress response function, the theoretical yield loss rate of the target farmland system is calculated under the preset standard meteorological stress scenario; From the agricultural production management record data, information on irrigation facility coverage, application of soil improvement measures, adoption rate of stress-resistant varieties, and adaptive measures is extracted, and their effectiveness in mitigating yield loss is evaluated. The speed and extent of crop recovery and growth after disasters are analyzed from the multi-year, multi-point location monitoring test data, and the intrinsic resilience of the farmland system is quantified. By coupling the theoretical yield loss rate, the mitigation effectiveness of adaptive measures, and the intrinsic resilience value, the vulnerability index of the disaster-bearing body, which reflects the overall situation of the system's sensitivity to climate stress and its self-adjustment capacity, is calculated.

[0010] As a further aspect of the present invention, the step of using a statistical model to analyze the quantitative relationship between the actual crop yield data and meteorological stress factors during the key growth period, and constructing a yield-meteorological stress response function, includes: Identify one or more critical growth stages throughout the crop's entire growth period that are most sensitive to yield formation; For each of the key reproductive stages, one or more meteorological stress factors are extracted from the meteorological data of the corresponding time period. The meteorological stress factors include the water stress index, the temperature stress index, and the sunshine stress index. Establish a multiple regression relationship between the actual crop yield data from multiple years and locations and all meteorological stress factors at all key growth stages, or use a machine learning model to train a nonlinear mapping relationship. Sensitivity analysis was used to determine the most significant functional forms and parameters in response to changes in various meteorological stress factors from the multiple regression or nonlinear mapping relationships. Based on a defined function form and parameters, a mathematical model is constructed to quantitatively describe the changes in yield caused by changes in the combination of meteorological stress factors, namely the yield-meteorological stress response function.

[0011] As a further aspect of the present invention, the step of weightedly superimposing and fusing the normalized disaster-causing factor hazard index, disaster-bearing body exposure index, and disaster-bearing body vulnerability index to generate a comprehensive climate risk resilience index for dryland farmland includes: The weight coefficients of the disaster-causing factor hazard index, disaster-bearing body exposure index, and disaster-bearing body vulnerability index in the comprehensive resilience evaluation are determined by using the analytic hierarchy process or the entropy weight method, respectively. The weighted risk value is obtained by multiplying the normalized hazard index of the disaster-causing factor by its corresponding weight coefficient. The normalized exposure index of the disaster-bearing body is multiplied by its corresponding weighting coefficient to obtain the weighted exposure value; The normalized vulnerability index of the disaster-bearing body is multiplied by its corresponding weight coefficient to obtain the weighted vulnerability value. The weighted hazard value, weighted exposure value, and weighted vulnerability value are substituted into a preset superposition and fusion model for calculation. The superposition and fusion model is a nonlinear function, whose input is the weighted value of the three factors and whose output is a scalar value between zero and one. The scalar value calculated by the superposition and fusion model is defined as the comprehensive resilience index of climate risk in dryland farmland. The higher the comprehensive resilience index of climate risk in dryland farmland, the stronger the resilience to climate risk.

[0012] As a further aspect of the present invention, the superposition and fusion model is a nonlinear function, and its construction and calculation process includes: The theoretical maximum and minimum values ​​of the comprehensive resilience index of climate risk in dry farmland are set, corresponding to the ideal states of complete resilience and complete lack of resilience, respectively. A contribution function is established for the weighted hazard value, weighted exposure value, and weighted vulnerability value to the comprehensive resilience index of climate risk in dryland farmland. The contribution function characterizes the nonlinear influence relationship of a single factor on the final index. Define interaction terms among factors to characterize the synergistic or antagonistic effects of the weighted hazard value, weighted exposure value, and weighted vulnerability value on resilience. The contribution function of a single factor is combined with the interaction term between factors to construct a nonlinear equation containing linear, quadratic, and interaction terms as the superposition and fusion model. The parameters of the overlay and fusion model are calibrated and verified using historical case data or expert knowledge; The weighted hazard value, weighted exposure value, and weighted vulnerability value of the area to be accounted for are input into the calibrated superposition and fusion model to directly calculate the comprehensive resilience index of the dryland farmland climate risk of the area to be accounted for.

[0013] As a further aspect of the present invention, the method further includes the steps of spatial visualization and causal analysis of the calculation results: The calculated hazard index of disaster-causing factors, exposure index of disaster-bearing bodies, vulnerability index of disaster-bearing bodies, and the final comprehensive resilience index of climate risk for dryland farmland for each evaluation unit are associated with their corresponding geographic spatial locations. Using the mapping function of the Geographic Information System, spatial distribution thematic maps of the disaster-causing factor hazard index, disaster-bearing body exposure index, disaster-bearing body vulnerability index, and dryland farmland climate risk comprehensive resilience index were generated respectively. The main river systems, administrative boundaries, and key infrastructure locations of the target dryland farmland area are overlaid on the spatial distribution thematic map. For areas where the comprehensive resilience index of climate risk in dryland farmland is low, the specific values ​​and spatial distribution characteristics of the corresponding disaster-causing factor hazard index, disaster-bearing body exposure index, and disaster-bearing body vulnerability index are traced back. By combining the backtested values ​​and characteristics, we analyze whether the main limiting factors leading to the low regional comprehensive resilience index are excessive risk, excessive exposure, or excessive vulnerability, and generate a report on the causes of spatial differentiation.

[0014] As a further aspect of the present invention, the method further includes: classifying the comprehensive climate risk resilience index of the dryland farmland according to a preset resilience level threshold range, determining and outputting the climate risk resilience level of the target dryland farmland area, including: Establish an evaluation standard that includes multiple resilience levels, set clear and mutually exclusive numerical ranges for the comprehensive resilience index of climate risk in dryland farmland for each resilience level, and form a reference table of resilience level threshold ranges. The calculated comprehensive resilience index of climate risk of the target dryland farmland area is compared one by one with each value range in the resilience level threshold range comparison table. When the value of the comprehensive resilience index of climate risk of dryland farmland falls within a preset value range corresponding to a certain resilience level, it is determined that the climate risk resilience of the target dryland farmland area belongs to the resilience level. The climate risk resilience level of the target dryland farmland area is determined and associated with the geographical location information of the target dryland farmland area and the specific value of the comprehensive climate risk resilience index of the dryland farmland area. Output a final report in the form of structured data files or visual charts, which includes geographic location information, the comprehensive resilience index of dryland farmland climate risk, and the results of climate risk resilience level determination.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: Based on the temporal and frequency evolution patterns and spatial heterogeneity of climate events, a disaster risk quantification algorithm is established. The fixed computational structure is adjusted, and the index calculation rules are improved according to the natural variation attributes of climate elements. Using long-term meteorological data as a carrier, a complete calculation of the disaster-causing factor risk index is completed, mitigating the problem of insufficient adaptability of general calculation models. This ensures that the quantification process of climate risk source-related attributes aligns with the regional climate's own variation patterns, expands the analytical levels of climate risk indicators, and strengthens the integration of dynamic climate change characteristics into index accounting.

[0016] This study utilizes diverse and differentiated basic data to calculate the exposure and vulnerability indices of disaster-bearing bodies. Supported by remote sensing imagery of soil physicochemical properties and production management data, it quantifies the spatiotemporal cover status of farmland systems under climate stress. Combined with location-based monitoring experimental data, it characterizes the yield fluctuation response and self-recovery attributes of ecosystems after climate disturbance. It breaks down the evaluation units of the comprehensive attributes of farmland systems and eliminates singular evaluation models, achieving a hierarchical quantitative expression of the external stress contact status and internal response attributes of farmland.

[0017] Normalization was performed on the three core indices to eliminate objective differences in units and magnitudes among different indicators, and to unify the calculation standards for multiple indices. A weighted, superimposed, and fused calculation method was adopted to integrate multiple index data, balance the proportion of different factors in resilience accounting, and link the three evaluation dimensions of hazard exposure and vulnerability. The execution process of multi-indicator collaborative calculation was improved, the quantitative composition dimensions of climate risk resilience in dryland farmland were enriched, a standardized accounting form for multi-factor collaborative calculation was formed, and the overall framework for quantitative evaluation of farmland climate resilience was expanded. Attached Figure Description

[0018] Figure 1 This is a state diagram of a method for calculating the climate risk resilience of dryland farmland based on the superposition of three factors, as described in this invention. Figure 2 A flowchart for calculating the hazard index of disaster-causing factors; Figure 3 The flowchart for calculating the exposure index of disaster-bearing bodies. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0020] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0021] See Figure 1 This invention provides a method for assessing the climate risk resilience of dryland farmland based on the superposition of three factors. This method integrates multi-source data to quantify the hazard of disaster-causing factors, the exposure of disaster-bearing bodies, and their vulnerability, thereby fusing them to form a comprehensive resilience index. The overall implementation scheme is as follows: A multi-source basic dataset of the target dryland farmland area is acquired. This dataset includes long-term meteorological observation data, high-resolution remote sensing image data, soil profile physicochemical property data, agricultural production management records, and multi-year, multi-point location monitoring experimental data. Based on the long-term meteorological observation data, a disaster hazard quantification algorithm designed according to the temporal and frequency evolution patterns and spatial heterogeneity characteristics of climate events is used to calculate the hazard index of disaster-causing factors. Based on the high-resolution remote sensing image data, soil profile physicochemical property data, and agricultural production management records, the exposure index of disaster-bearing bodies, characterizing the spatial and temporal exposure of the farmland system under climate stress, is calculated. Based on multi-year, multi-point location monitoring experimental data and some agricultural production management records, the vulnerability index of disaster-bearing bodies, characterizing the sensitivity and recovery capacity of the farmland ecosystem to yield loss under different climate stresses, is calculated. The calculated hazard index of disaster-causing factors, exposure index of disaster-bearing bodies, and vulnerability index of disaster-bearing bodies were normalized to eliminate dimensional differences. The normalized hazard index of disaster-causing factors, exposure index of disaster-bearing bodies, and vulnerability index of disaster-bearing bodies were then weighted, superimposed, and merged to generate a comprehensive climate risk resilience index for dryland farmland.

[0022] In one embodiment of the present invention, the calculation process of the hazard index of disaster-causing factors is described in detail. (See also...) Figure 2This study extracts daily value sequences of meteorological elements related to drought, high temperature, low temperature, and heavy rainfall from long-term meteorological observation data. For each type of meteorological disaster, disaster events exceeding or falling below preset thresholds in the daily value sequences of meteorological elements are identified, and the start time, end time, duration, and extreme intensity of each disaster event are recorded. The total frequency of events for each type of meteorological disaster within a preset historical time period is statistically analyzed, and the average frequency per unit time is calculated. The extreme intensity of each disaster event is calculated, and the average intensity of all events for each meteorological disaster type is also calculated. Using spatial interpolation techniques, the point-like long-term meteorological observation data is transformed into a gridded meteorological data field covering the target arid farmland area. Based on the gridded meteorological data field, the spatial distribution concentration of each disaster type is analyzed. By fusing the average frequency, average intensity, and distribution concentration, and introducing a time trend correction coefficient, a single-disaster hazard score is calculated for each meteorological disaster type. The single-hazard risk scores of various meteorological disaster types faced by the target dryland farmland area are calculated by multi-hazard coupling to generate a comprehensive disaster-causing factor risk index.

[0023] In the specific implementation, a dryland farming area in northern China was selected as the target dryland farmland region. Long-term meteorological observation data from the past 40 years was obtained, sourced from 25 meteorological observation stations within and around the region. Daily value sequences of meteorological elements related to drought, high temperature, low temperature, and heavy rainfall were extracted from the long-term meteorological observation data, specifically including daily precipitation sequences, daily maximum temperature sequences, and daily minimum temperature sequences. For high-temperature disasters, a daily maximum temperature greater than or equal to 35 degrees Celsius was set as the threshold for high-temperature events. All meteorological events exceeding this threshold in the daily maximum temperature sequence were identified, and the start date, end date, duration, and extreme daily maximum temperature during each high-temperature event were recorded. The total frequency of high-temperature disaster events over the past 40 years was statistically analyzed, totaling 180 events, resulting in an average frequency of 4.5 events per year. The extreme intensity of each high-temperature event, i.e., the specific value of the daily maximum temperature exceeding 35 degrees Celsius, was calculated. The arithmetic mean of all 180 extreme intensity values ​​of high-temperature events was calculated, yielding an average intensity of 3.2 degrees Celsius. Using inverse distance weighted spatial interpolation, the daily maximum temperature sequence from 25 station point-like long-term meteorological observation data was generated into a gridded meteorological data field with a spatial resolution of 1 km x 1 km covering the target dryland farmland area. Based on this gridded meteorological data field, the spatial Gini coefficient was used to analyze the spatial distribution concentration of high temperature disaster intensity, and the calculated distribution concentration index value was 0.15.

[0024] In some embodiments, the average occurrence frequency, average process intensity, and distribution concentration are integrated, and a time trend correction coefficient for the annual occurrence frequency of high-temperature events, calculated based on linear regression (with a coefficient of 1.1), is introduced to calculate the single-hazard hazard score for high-temperature disaster types. It can be understood that the single-hazard hazard score can be calculated using a specific fusion formula, as shown below:

[0025] in: This represents a single-hazard risk score for a specific type of disaster. This represents the original value of the calculated average occurrence frequency. and The minimum and maximum values ​​represent the average occurrence frequency of all grid points within the study area. The weight representing the average frequency of occurrence is set to 0.4 here. This represents the original value of the calculated average process intensity. and The minimum and maximum values ​​represent the average process intensity at all grid points within the study area. The weight representing the average process intensity is set to 0.4 here. This represents the original value of the calculated distribution concentration. and These represent the minimum and maximum values ​​of the distribution concentration of all grid points within the study area. The weight representing the concentration of the distribution is set to 0.2 here. This represents the time trend correction coefficient. Following the same process, the remaining three meteorological disaster types—drought, low temperature, and torrential rain—are processed in parallel, and their individual hazard scores are calculated separately. In practice, the individual hazard scores for various meteorological disaster types, including drought, high temperature, low temperature, and torrential rain, faced by the target dryland farmland area are subjected to multi-hazard coupling calculation. The coupling calculation uses the method of taking the maximum score of each hazard type to generate a comprehensive disaster-causing factor hazard index. Optionally, the multi-hazard coupling calculation can also use a weighted summation method, where the weights are determined based on historical disaster loss data.

[0026] In practice, identifying disaster events involves daily scanning and analysis of the original meteorological sequence. For drought disasters, a standardized precipitation index (SPI) sequence is calculated from the daily precipitation sequence. An event with a SPI less than -1 for 30 consecutive days is identified as a drought event. The start time of this event is recorded as the date the SPI first falls below -1, the end time as the date the SPI first returns to a value greater than or equal to -1, the duration as the number of days from the start to the end, and the extreme value of the event intensity as the minimum value of the SPI during the event period. Similarly, for low-temperature disasters, a daily minimum temperature less than or equal to 0 degrees Celsius is set as the threshold for low-temperature events, and all meteorological events in the daily minimum temperature sequence that fall below this threshold are identified. For rainstorm disasters, a daily precipitation greater than or equal to 50 mm is set as the threshold for rainstorm events, and all meteorological events in the daily precipitation sequence that exceed this threshold are identified.

[0027] In some embodiments, when calculating the average occurrence frequency, the unit time can be selected as a year, or a more specific time period such as the crop growth period. When calculating the average process intensity, for high temperature, low temperature, and rainstorm disasters, the extreme values ​​of process intensity can be directly used as specific meteorological values ​​exceeding or falling below the threshold. For drought disasters, the extreme values ​​of process intensity use the negative deviation of the standardized precipitation index. In specific implementations, generating a gridded meteorological data field using spatial interpolation techniques is the basis for analyzing spatial distribution concentration. In addition to the inverse distance weighting method, interpolation methods such as Kriging interpolation or spline function interpolation can be selected. When analyzing distribution concentration, in addition to the spatial Gini coefficient, spatial autocorrelation statistics such as the Moran index can also be used.

[0028] Optionally, the time trend correction coefficient is calculated based on the time series of annual occurrence frequencies of high-temperature events over the past 40 years. A linear trend line is fitted, and the correction coefficient is calculated by adding 1 to the ratio of the absolute value of the slope to the frequency of occurrence in the base year. In practice, multi-hazard coupling calculation is the final step in generating the final hazard factor risk index. The choice of method depends on the assumptions about the superposition relationship of disasters. The method of taking the maximum value emphasizes the role of the dominant disaster, while the weighted summation method considers the cumulative effect of multiple disasters.

[0029] In one embodiment of the present invention, the calculation of the disaster-bearing body exposure index and the working principle of the disaster risk quantification algorithm are described in detail. See also... Figure 3This study interprets and classifies high-resolution remote sensing image data to identify the spatial distribution of different land use types, including farmland, woodland, water areas, and construction land, within the target dryland farmland area. It also extracts the boundaries, area, and spatial connectivity information of farmland plots. From soil profile physicochemical property data, it extracts the available water content, soil texture, and soil layer thickness information for each farmland plot, assessing and generating drought resistance indicators characterizing its basic water retention and drought resistance capabilities. From agricultural production management records, it extracts planting structure, sowing date, harvest date, and irrigation and fertilization records, calculating the time windows of different crops during their growth period and their occupation of climate resources. The spatial distribution, area, and spatial connectivity information of farmland plots are overlaid with drought resistance indicators to assess the extent and severity of damage to farmland spatial entities during climate disasters, generating a spatial exposure component. Finally, it matches the crop growth period time windows with the temporal patterns of historical climate disasters to assess the degree of overlap between the crop growth process and high-risk periods of climate disasters in the temporal dimension, generating a temporal exposure component. By weighting and combining the spatial and temporal exposure components, a disaster-bearing body exposure index that characterizes the spatiotemporal exposure features of the farmland system is obtained.

[0030] The disaster hazard quantification algorithm is designed based on the temporal and frequency evolution patterns and spatial heterogeneity of climate events. Its working principle includes: establishing a sliding time window to dynamically calculate the statistical characteristic values ​​of meteorological elements on the time series of long-term meteorological observation data to capture the time-varying characteristics of climate states; within each sliding time window, using extreme value distribution theory to fit the intensity of meteorological disasters and estimating the disaster intensity under different return periods to quantify the extremeness and frequency characteristics of disasters; introducing spatial autocorrelation analysis to calculate the spatial clustering and heterogeneity indicators of disaster intensity based on a gridded meteorological data field to quantify the spatial pattern characteristics of disasters; defining a dynamic correction factor for disaster hazard, which is a function of the temporal variation trend coefficient of disaster event frequency and the spatial variation coefficient of disaster intensity; nonlinearly combining the estimated disaster intensity and frequency, spatial clustering indicators, and dynamic correction factor within the sliding window to construct a temporal-spatial dynamic hazard assessment model for each grid cell; traversing all grid cells and applying the temporal-spatial dynamic hazard assessment model for calculation, and spatially aggregating the calculation results to generate a hazard index for disaster-causing factors.

[0031] In the specific implementation, high-resolution remote sensing imagery covering the same target dryland farmland area was acquired, with a spatial resolution of 2 meters, during the 2024 crop growing season. The high-resolution remote sensing imagery was interpreted and classified using an object-oriented classification method to identify the spatial distribution of different land use types within the target dryland farmland area, including farmland, woodland, water bodies, and construction land. Image segmentation algorithms were then used to extract the vector boundaries, area, and spatial connectivity information with adjacent farmland plots for each independent plot. From the accompanying soil profile physicochemical property data, the available soil water content, soil texture classification, and soil layer thickness information for each farmland plot in the 0-100 cm soil layer were extracted. A weighted scoring model based on available soil water content, soil clay content, and soil layer thickness was established to evaluate and generate a drought resistance index ranging from 0 to 1, characterizing the basic water retention and drought resistance capacity.

[0032] In practice, the planting structure, sowing dates, harvest dates, and irrigation and fertilization records of major crops for the past five years are extracted from agricultural production management records for the target area. The time windows for different crops during their growth period from sowing to harvest are calculated, and the accumulated water demand and accumulated temperature within these time windows are statistically analyzed to quantify the crops' consumption of climate resources. It can be understood that the spatial distribution, area, and spatial connectivity information of farmland plots are overlaid with drought resistance indicators. This is achieved through spatial overlay operations using a geographic information system. For each farmland plot, its area is multiplied by the corresponding drought resistance indicator value to obtain a product representing the potential scale of disaster loss for that plot. This product of all farmland plots in the region is then aggregated and divided by the total farmland area to generate a standardized spatial exposure component, with a value ranging from 0 to 1. In some embodiments, matching the crop's growth period window with the historical patterns of climate disaster occurrence involves analyzing historical meteorological data to determine the high-incidence periods of various types of climate disasters. This includes calculating the overlap ratio of calendar days between each crop growth stage and the high-incidence period, and using the yield sensitivity coefficient of each growth stage as a weight to calculate a weighted average temporal overlap degree, which is the temporal exposure component. Optionally, the spatial exposure component and the temporal exposure component are weighted and combined using a weighted geometric mean method to calculate the disaster-bearing body exposure index. It can be understood that the disaster-bearing body exposure index... The calculation formula is expressed as follows:

[0033] in: Represents the exposure index of the disaster-bearing body. Represents the spatial exposure component. Represents the time exposure component. The value represents the weighting coefficient assigned to the spatial exposure component, which is set to 0.6 here. The weighted geometric mean method can simultaneously reflect the synergistic effect of exposure features in both spatial and temporal dimensions.

[0034] In its implementation, the disaster risk quantification algorithm is designed based on the temporal and frequency evolution patterns and spatial heterogeneity of climate events. Its working principle includes the following steps: First, a 15-year sliding time window is established. Starting from the initial year, the window slides annually across the time series of long-term meteorological observation data, dynamically calculating the average and standard deviation of high-temperature days within each sliding window to capture the temporal variation of climate conditions. Second, within each sliding time window, for high-temperature disasters, the distribution of the annual extreme maximum temperature is fitted using generalized extreme value distribution theory, and the intensity of extreme high temperatures with return periods of 20 and 50 years is estimated to quantify the extremeness and frequency characteristics of the disaster. Third, spatial autocorrelation analysis is introduced. Based on the generated gridded meteorological data field, the local Moran's index between the high-temperature intensity of each grid cell and the high-temperature intensity of surrounding grid cells is calculated as an indicator of the spatial clustering and heterogeneity of disaster intensity, quantifying the spatial pattern characteristics of the disaster. Finally, a dynamic correction factor for disaster risk is defined, which is a function of the temporal trend coefficient of disaster event frequency and the spatial variation coefficient of disaster intensity. In some embodiments, the estimated disaster intensity and frequency, spatial clustering index, and dynamic correction factor for disaster hazard within the sliding window are nonlinearly combined to form a spatiotemporal dynamic hazard assessment model for each grid cell. This nonlinear combination is achieved by multiplying each normalized index with the correction factor and adding an interaction term. The spatiotemporal dynamic hazard assessment model is applied to all grid cells, and the calculation results for each grid cell are spatially aggregated. The calculated values ​​of all farmland-covered grids within the target dryland farmland area are then averaged by area to generate the disaster hazard index. In specific implementations, the length of the sliding time window, the type of extreme value distribution, the selection of the return period, the specific indicators for spatial autocorrelation analysis, and the functional form of the nonlinear combination can all be adjusted according to actual data and disaster characteristics. Optionally, the temporal trend coefficient can be calculated using the linear regression slope of the disaster frequency sequence within the sliding window, and the spatial variation coefficient can be calculated using the ratio of the standard deviation to the mean of the gridded disaster intensity field.

[0035] In one embodiment of the present invention, the calculation of the vulnerability index of the disaster-bearing body and the construction of the yield-meteorological stress response function are described in detail. Actual crop yield data and corresponding meteorological data for different growth stages under different climate years and management measures are extracted from multi-year, multi-location monitoring experimental data. A statistical model is used to analyze the quantitative relationship between actual crop yield data and meteorological stress factors during key growth stages, constructing the yield-meteorological stress response function. Based on the yield-meteorological stress response function, the theoretical yield loss rate of the target farmland system under a preset standard meteorological stress scenario is calculated. Information on irrigation facility coverage, soil improvement measures application, stress-resistant variety adoption rate, and adaptive measures is extracted from agricultural production management records, and their effectiveness in mitigating yield loss is evaluated. The speed and extent of post-disaster crop recovery growth are analyzed from multi-year, multi-location monitoring experimental data, quantifying the intrinsic resilience of the farmland system. The theoretical yield loss rate, the mitigation effectiveness of adaptive measures, and the quantitative value of intrinsic resilience are coupled to calculate the vulnerability index of the disaster-bearing body, reflecting the system's comprehensive sensitivity to climate stress and its self-adjustment capacity.

[0036] The specific process of constructing the yield-meteorological stress response function includes: identifying one or more key growth stages during the entire crop growth period that are most sensitive to yield formation; for each key growth stage, extracting one or more meteorological stress factors from the corresponding meteorological data, including water stress index, temperature stress index, and sunshine stress index; establishing multiple regression relationships between actual crop yield data from multiple years and locations and all meteorological stress factors for all key growth stages, or training nonlinear mapping relationships using machine learning models; and determining the functional form and parameters that most significantly respond to changes in each meteorological stress factor from the multiple regression or nonlinear mapping relationships through sensitivity analysis. Based on the determined functional form and parameters, constructing a mathematical model that quantitatively describes the yield change caused by changes in the combination of meteorological stress factors, i.e., the yield-meteorological stress response function.

[0037] In practical implementation, the calculation of the vulnerability index of the disaster-bearing body and the construction of the yield-meteorological stress response function are described and implemented in the following way. In the specific implementation, from the multi-year, multi-point location monitoring experiment data of the target dryland farmland area, actual yield data of spring maize and daily meteorological data for the corresponding growth stages were extracted from 15 location observation points over eight consecutive years. The meteorological data included daily precipitation, daily average temperature, daily maximum temperature, daily minimum temperature, and sunshine duration. Analyzing the quantitative relationship between actual crop yield data and meteorological stress factors during key growth stages using statistical models requires identifying the key growth stages most sensitive to yield formation throughout the crop's entire growth period. For spring maize, these three stages were identified as the jointing stage, the tasseling and silking stage, and the grain-filling stage. For each key growth stage, one or more meteorological stress factors were extracted from the meteorological data of the corresponding time period. For example, a water stress index was extracted for the jointing stage, a high-temperature stress index for the tasseling and silking stage, and a water stress index and a low-temperature stress index for the grain-filling stage.

[0038] It is understandable that multiple regression relationships are established between multi-year, multi-location spring maize yield data and all meteorological stress factors at all key growth stages. A stepwise regression method is used to screen out significant meteorological stress factors and determine their coefficients. Through sensitivity analysis, the function form that most significantly responds to changes in each meteorological stress factor is determined to be a linear summation form from the multiple regression relationships. Based on the determined function form and parameters, a mathematical model is constructed to quantitatively describe the yield change caused by changes in the combination of meteorological stress factors, namely the yield-meteorological stress response function. The form of the yield-meteorological stress response function is:

[0039] in: Represents the predicted actual output. This represents potential output (i.e., theoretical output under conditions of no weather stress). , , These represent water stress factors during the jointing stage. High temperature stress factors during the male elongation and silking period Low temperature stress factors during grouting period The regression coefficient (whose value is negative). , , These represent the quantitative values ​​of the corresponding stress factors, such as water deficit, accumulated temperature above the threshold temperature, and accumulated temperature below the threshold temperature.

[0040] In some embodiments, the theoretical yield loss rate of the target farmland system under a preset standard meteorological stress scenario is calculated based on the yield-meteorological stress response function. The preset standard meteorological stress scenario is defined as the main stress at a moderate level occurring at all key growth stages of the crop. The values ​​of each stress factor under the standard stress scenario are substituted into the yield-meteorological stress response function to calculate the predicted yield. The theoretical yield loss rate is calculated by subtracting the predicted yield from the potential yield and then dividing by the potential yield. In specific implementations, information on irrigation facility coverage, soil improvement measures application, stress-resistant variety adoption rate, and adaptive measures is extracted from agricultural production management records, and their effectiveness in mitigating yield loss is evaluated. These indicators are standardized and weighted to obtain a comprehensive score for the mitigation effectiveness of adaptive measures between 0 and 1, with a higher score indicating stronger mitigation effectiveness. The speed and extent of crop recovery after disasters are analyzed from multi-year, multi-point location monitoring experimental data to quantify the intrinsic resilience of the farmland system.

[0041] The specific method involves selecting observational data from the experiment after re-irrigation following a specific disaster or after weather improvement, calculating the reciprocal of the number of days required for crop growth rate to recover to pre-disaster levels, or using the ratio of the final yield to the yield of the unaffected control area as a quantitative indicator of intrinsic resilience, and normalizing it to between 0 and 1. This can be understood as coupling the theoretical yield loss rate, the comprehensive score of the mitigation effectiveness of adaptive measures, and the quantitative value of intrinsic resilience to calculate a vulnerability index reflecting the system's overall sensitivity to climate stress and its self-adjustment capacity. The coupled calculation is achieved by multiplying the theoretical yield loss rate by a reduction factor determined by the effectiveness of adaptive measures, and then dividing by an amplification factor determined by intrinsic resilience. The numerical result is the vulnerability index; a higher index value indicates a more vulnerable system. Table 1 shows a correspondence between key growth stages of spring maize and major meteorological stress factors.

[0042] Table 1: Correspondence between key growth stages of spring maize and major meteorological stress factors

[0043] In practical implementation, identifying critical growth stages relies on a combination of agronomic knowledge and historical experimental data. The construction of the yield-meteorological stress response function can be based on multiple linear regression, or machine learning models such as random forests and support vector machines can be used to train nonlinear mapping relationships. Optionally, the meteorological stress factor water stress index can be the crop water deficit index, the temperature stress index can be the high temperature accumulation or cold accumulated temperature, and the sunshine stress index can be the sunshine duration anomaly. In some embodiments, the standard meteorological stress scenario can be defined based on the characteristics of historical disaster events, such as using the combination of moderate disaster intensities that occurred most frequently in the past 20 years. The extraction and effectiveness evaluation of adaptive measures information requires converting qualitative records into quantitative indicators and determining the relative weights of different measures based on expert scoring or literature results. Optionally, the calculation of intrinsic resilience quantification values ​​can also focus on the recovery of physiological indicators, such as the recovery rate of leaf water potential or photosynthetic rate.

[0044] In one embodiment of the present invention, the process of generating a comprehensive resilience index through weighted superposition and fusion, and the construction of the superposition and fusion model are described in detail. The weight coefficients of the hazard index of the disaster-causing factor, the exposure index of the disaster-bearing body, and the vulnerability index of the disaster-bearing body in the comprehensive resilience evaluation are determined using either the analytic hierarchy process (AHP) or the entropy weight method. The normalized hazard index of the disaster-causing factor is multiplied by its corresponding weight coefficient to obtain a weighted hazard value. The normalized exposure index of the disaster-bearing body is multiplied by its corresponding weight coefficient to obtain a weighted exposure value. The normalized vulnerability index of the disaster-bearing body is multiplied by its corresponding weight coefficient to obtain a weighted vulnerability value. The weighted hazard value, weighted exposure value, and weighted vulnerability value are substituted into a preset superposition and fusion model for calculation. This model is a nonlinear function; its input is the weighted value of the three factors, and its output is a scalar value between zero and one. The scalar value calculated by the superposition and fusion model is defined as the comprehensive resilience index of dryland farmland climate risk; the higher the index value, the stronger the climate risk resilience.

[0045] The construction and calculation process of the superposition and fusion model includes: setting the theoretical maximum and minimum values ​​of the comprehensive resilience index of dryland farmland climate risk, corresponding to the ideal states of complete resilience and complete inresilience, respectively. Establishing contribution functions for the weighted hazard value, weighted exposure value, and weighted vulnerability value to the comprehensive resilience index of dryland farmland climate risk; these contribution functions characterize the nonlinear influence of individual factors on the final index. Defining interaction terms between factors to characterize the synergistic or antagonistic effects of each pair of weighted hazard value, weighted exposure value, and weighted vulnerability value on resilience. Combining the contribution functions of individual factors with the interaction terms between factors to construct a nonlinear equation containing linear, quadratic, and interaction terms as the superposition and fusion model. Calibrating and validating the parameters of the superposition and fusion model using historical case data or expert knowledge. Inputting the weighted hazard value, weighted exposure value, and weighted vulnerability value of the area to be assessed into the calibrated superposition and fusion model to directly calculate the comprehensive resilience index of dryland farmland climate risk for the area to be assessed.

[0046] In practical implementation, the construction of the weighted overlay and fusion model for generating the comprehensive resilience index of dryland farmland climate risk is described and implemented in the following way. In practical implementation, it is assumed that the normalized values ​​of the hazard index of disaster-causing factors, the exposure index of disaster-bearing bodies, and the vulnerability index of disaster-bearing bodies for a certain evaluation unit within the target area have been calculated through the aforementioned steps, and are denoted as follows: , , The analytic hierarchy process (AHP) was used to determine the weight coefficients of the hazard index of disaster-causing factors, the exposure index of disaster-bearing bodies, and the vulnerability index of disaster-bearing bodies in the comprehensive resilience assessment. Specifically, several experts in agrometeorology and risk management were invited to compare and judge the relative importance of each index pairwise, constructing a judgment matrix. Then, the eigenvectors of the matrix were calculated and a consistency test was performed. The eigenvectors obtained after passing the test are the weight coefficients of each index. The calculated weight coefficients are as follows: Hazard index weight of disaster-causing factors The weight of the disaster-bearing body exposure index is 0.4. The weight of the vulnerability index of the disaster-bearing body is 0.3. The value is 0.3. This is understandable, as it represents the normalized hazard index of the disaster-causing factor. Its corresponding weighting coefficient Multiply by each other to obtain the weighted risk value. The normalized exposure index of the disaster-bearing body. Its corresponding weighting coefficient Multiply to obtain the weighted exposure value. The normalized vulnerability index of the disaster-bearing body. Its corresponding weighting coefficient Multiply by each other to obtain the weighted vulnerability value. .

[0047] In some embodiments, the weighted risk value Weighted exposure value and weighted vulnerability value The data is substituted into a pre-defined superposition and fusion model for calculation. This model is a nonlinear function. The construction and calculation process of the superposition and fusion model includes: setting the theoretical maximum and minimum values ​​of the comprehensive resilience index for climate risk in dryland farmland, corresponding to ideal states of complete resilience and complete inflexibility, respectively; for example, setting the theoretical maximum value to 1 and the theoretical minimum value to 0. A weighted hazard value is then established. Weighted exposure value and weighted vulnerability value The contribution function to the comprehensive resilience index of climate risk in dryland farmland is defined. This function characterizes the nonlinear influence of individual factors on the final index; for example, a negative exponential function can be used to represent the weakening effect of hazard and exposure on the resilience index. An interaction term between factors is defined to characterize the weighted hazard value. Weighted exposure value and weighted vulnerability value The synergistic or antagonistic effects on resilience between factors can be amplified, for example, the impact of exposure may be magnified when the risk is high. In practice, the contribution functions of individual factors are combined with the interaction terms between factors to construct a nonlinear equation containing linear, quadratic, and interaction terms as a superposition fusion model. A specific example of a superposition fusion model is as follows:

[0048] in: The comprehensive resilience index for climate risk in dryland farmland ranges from 0 to 1. , , These represent the weighted hazard value, weighted exposure value, and weighted vulnerability value, respectively. These are the model parameters to be calibrated. The formula calculates the final resilience index by subtracting the resilience loss caused by hazard, exposure, vulnerability, and their interactions from the fully resilient state (numerical 1). This can be understood as using historical case data or expert knowledge to calibrate and validate the parameters of the superimposed and fused model. For example, multiple historical regional cases with known climate risk and resilience status can be collected, their three-factor weighted values ​​used as input, and their expert assessment level of overall resilience status used as the target. The model parameters are then determined through nonlinear fitting. The value of the weighted hazard value of the area to be calculated. Weighted exposure value and weighted vulnerability value By inputting a calibrated superposition and fusion model, the comprehensive resilience index of dryland farmland in the area to be assessed is directly calculated. , The higher the value, the stronger the resilience to climate risks.

[0049] In practical implementation, constructing the judgment matrix in the analytic hierarchy process (AHP) is a crucial step in determining the weight coefficients. An example judgment matrix comparing the pairwise importance of the hazard index of disaster-causing factors, the exposure index of disaster-bearing bodies, and the vulnerability index of disaster-bearing bodies is shown in the table below. Optionally, the entropy weight method can also be used to determine the weight coefficients. This method calculates the weight of each index based on the dispersion of the three-factor index values ​​for each evaluation unit using information entropy. This method relies on the distribution of objective data. The functional form of the superposition and fusion model is not limited to the above-mentioned combination of linear and quadratic terms; it can also be other nonlinear forms, such as those containing logarithmic or exponential operations. In some embodiments, the contribution function can be designed as an S-shaped curve to reflect the threshold effect of the factor's influence from quantitative to qualitative change. The calibration of the model parameters can be based on regression analysis using the statistical relationship between historical disaster loss data and the three-factor indices. Optionally, the interaction term can include higher-order interactions, such as the product term of the combined effect of the three factors. See Table 2, which shows the judgment matrix for the analytic hierarchy process.

[0050] Table 2: Matrix for Judging the Hazard, Exposure, and Vulnerability of Disaster-Causing Factors

[0051] In one embodiment of the present invention, the steps of spatial visualization, causal analysis, and resilience level classification of the calculation results are detailed. The calculated hazard index of disaster-causing factors, exposure index of disaster-bearing bodies, vulnerability index of disaster-bearing bodies, and the final comprehensive resilience index of dryland farmland climate risk for each evaluation unit are associated with their corresponding geographic spatial locations. Using the mapping function of a geographic information system, thematic maps of the spatial distribution of the hazard index of disaster-causing factors, exposure index of disaster-bearing bodies, vulnerability index of disaster-bearing bodies, and comprehensive resilience index of dryland farmland climate risk are generated. The main river systems, administrative boundaries, and locations of critical infrastructure in the target dryland farmland area are overlaid on the thematic maps. For areas where the comprehensive resilience index of dryland farmland climate risk shows a low value, the specific values ​​and spatial distribution characteristics of the corresponding hazard index of disaster-causing factors, exposure index of disaster-bearing bodies, and vulnerability index of disaster-bearing bodies are traced back. Combining the traced values ​​and characteristics, the main limiting factors leading to the low comprehensive resilience index of the region are analyzed as either excessively high risk, excessive exposure, or excessive vulnerability, and a causal analysis report of spatial differentiation is generated.

[0052] The process of resilience level classification includes: establishing evaluation criteria encompassing multiple resilience levels; defining clear and mutually exclusive numerical ranges for the comprehensive resilience index of dryland farmland climate risk for each resilience level; and creating a resilience level threshold range comparison table. The calculated comprehensive resilience index of the target dryland farmland area is compared one by one with each numerical range in the resilience level threshold range comparison table. When the value of the comprehensive resilience index falls within the preset numerical range corresponding to a certain resilience level, the target dryland farmland area is determined to belong to that resilience level. The determined climate risk resilience level of the target dryland farmland area is then associated and encapsulated with the geographical location information of the target dryland farmland area and the specific value of the comprehensive resilience index. A final report containing geographical location information, the comprehensive resilience index value of dryland farmland climate risk, and the climate risk resilience level determination result is output in the form of a structured data file or a visual chart.

[0053] In the specific implementation, the steps of spatial visualization, causal analysis, and resilience level classification of the calculation results are described and implemented in the following way. In the specific implementation, the calculated hazard index of disaster-causing factors, exposure index of disaster-bearing bodies, vulnerability index of disaster-bearing bodies, and the final comprehensive resilience index of dryland farmland climate risk for each evaluation unit are linked to the corresponding geospatial location vector surface data through the unique geographic code of the evaluation unit. The attribute table of each surface data records the specific values ​​of the four indices. Using the mapping function of the Geographic Information System (GIS), spatial distribution thematic maps of the hazard index of disaster-causing factors, exposure index of disaster-bearing bodies, vulnerability index of disaster-bearing bodies, and comprehensive resilience index of dryland farmland climate risk are generated respectively. The thematic maps use a segmented color scheme, dividing the index values ​​from low to high into multiple continuous intervals and rendering them with a gradient color system from warm to cool tones.

[0054] On the spatial distribution thematic map, the main river systems, administrative boundaries, and key infrastructure locations of the target dryland farmland area are overlaid. These base map elements are represented by semi-transparent lines or dots to clearly identify the relationship between the spatial pattern of the resilience index and the geographical background. For areas where the comprehensive climate risk resilience index of dryland farmland is low, the specific values ​​and spatial distribution characteristics of the corresponding disaster-causing factor hazard index, disaster-bearing body exposure index, and disaster-bearing body vulnerability index are traced back. Through the spatial query and statistical functions of the geographic information system, the three index values ​​of all low-resilience units are extracted and listed. Combining the traced values ​​and characteristics, the main limiting factors leading to the low comprehensive resilience index of the region are analyzed: is it excessively high hazard, excessively high exposure, or excessively high vulnerability? The specific method is to set a threshold. If the hazard index value of a unit is higher than 0.8 and significantly higher than its exposure and vulnerability index values, then excessively high hazard is judged to be its main limiting factor. A report analyzing the causes of spatial differentiation is generated. The report explains the dominant causes of different low-resilience areas in the form of text combined with thematic map screenshots.

[0055] In some embodiments, the comprehensive resilience index of dryland farmland to climate risk is classified according to a preset resilience level threshold range, and the climate risk resilience level of the target dryland farmland area is determined and output. This process includes establishing an evaluation standard that includes multiple resilience levels. For each resilience level, a clear and mutually exclusive numerical range of the comprehensive resilience index of dryland farmland to climate risk is set, forming a resilience level threshold range reference table. For example, five levels are set: low resilience, relatively low resilience, medium resilience, relatively high resilience, and high resilience, with corresponding numerical ranges of the comprehensive resilience index of dryland farmland to climate risk of [0, 0.2), [0.2, 0.4), [0.4, 0.6), [0.6, 0.8), and [0.8, 1.0], respectively. It can be understood that the calculated comprehensive resilience index of dryland farmland to climate risk of the target dryland farmland area is compared one by one with each numerical range in the resilience level threshold range reference table, and the judgment logic is to check which preset numerical range the index value falls into.

[0056] When the value of the comprehensive climate risk resilience index of dryland farmland falls within a preset range corresponding to a certain resilience level, the climate risk resilience of the target dryland farmland area is determined to belong to that resilience level. For example, if the comprehensive climate risk resilience index value of a certain unit is 0.65, and it falls within the range [0.6, 0.8), then its climate risk resilience level is determined to be "highly resilient". In specific implementation, the determined climate risk resilience level of the target dryland farmland area is associated and encapsulated with the geographical location information of the target dryland farmland area and the specific value of the comprehensive climate risk resilience index, and stored in a structured database table or GeoJSON geographic data format. Each record contains a geographic unit identifier, geographic coordinate range, comprehensive climate risk resilience index value, and climate risk resilience level field. The final report, containing geographical location information, comprehensive climate risk resilience index value, and climate risk resilience level determination results, is output in the form of structured data files or visualization charts. For example, a PDF document report is generated, which includes a level distribution map, tables of detailed data for each unit, and a brief text summary.

[0057] In practical implementation, the threshold range for resilience levels can be determined based on the statistical distribution of the comprehensive resilience index of climate risk in all evaluation units of dryland farmland within the study area, for example, by using the natural breakpoint method or the equal interval method. The formulaic expression for the level classification can be:

[0058] in: Represents the toughness level. This represents the comprehensive resilience index against climate risks in dryland farmland. Representative division Thresholds for each level. Optionally, the number and specific naming of resilience levels can be adjusted according to management or assessment needs, for example, simplified to three levels: "high," "medium," and "low." In some embodiments, the generated spatial distribution thematic map can be further aggregated with administrative division units to calculate the average resilience index at the township or county level and perform level mapping. Optionally, the causal analysis report can be automatically generated by automatically filling the corresponding positions in the report template with the keywords of the dominant limiting factors obtained from the retrospective analysis, specific index values, and spatial distribution descriptive text through preset rule templates. It is understood that the final report output format can be in addition to PDF documents, or it can be an interactive web map format, where users can click on map units to view detailed index values ​​and level information.

[0059] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for calculating the climate risk resilience of dryland farmland based on the superposition of three factors, characterized in that, The method includes: The target dryland farmland area was acquired by acquiring a multi-source basic dataset including long-term meteorological observation data, high-resolution remote sensing image data, soil profile physicochemical property data, agricultural production management record data, and multi-year multi-point positioning monitoring test data. Based on the long-term meteorological observation data, a disaster risk quantification algorithm is used to calculate the disaster risk index. The disaster risk quantification algorithm is designed based on the temporal and frequency evolution patterns and spatial heterogeneity characteristics of climate events. Based on the high-resolution remote sensing image data, soil profile physicochemical property data, and agricultural production management record data, the exposure index of disaster-bearing bodies is calculated. The exposure index of disaster-bearing bodies characterizes the spatial exposure and temporal coverage of farmland systems under climate stress. Based on the multi-year, multi-point location monitoring test data and some of the agricultural production management record data, the vulnerability index of the disaster-bearing body is calculated. The vulnerability index of the disaster-bearing body describes the sensitivity and recovery capacity of the farmland ecosystem to yield loss under different climate stresses. The calculated hazard index of the disaster-causing factor, the exposure index of the disaster-bearing body, and the vulnerability index of the disaster-bearing body are normalized to eliminate dimensional differences. The disaster-causing factor hazard index, disaster-bearing body exposure index, and disaster-bearing body vulnerability index, after normalization, are weighted, superimposed, and fused to generate a comprehensive climate risk resilience index for dryland farmland.

2. The method for calculating the climate risk resilience of dryland farmland based on three-factor superposition as described in claim 1, characterized in that, The calculation of the hazard index of disaster-causing factors using a disaster risk quantification algorithm includes: From the long-term meteorological observation data, daily value sequences of meteorological elements related to drought, high temperature, low temperature, and rainstorm are extracted; For each type of meteorological disaster, identify disaster events in the daily value sequence of the meteorological elements that exceed or fall below a preset threshold, and record the start time, end time, duration, and extreme intensity of each disaster event; The total frequency of each type of meteorological disaster event within a preset historical time period is statistically analyzed, and its average frequency per unit time is calculated. Calculate the extreme value of the process intensity for each disaster event, and calculate the average process intensity of all events for each type of meteorological disaster; Using spatial interpolation technology, the point-like long-term meteorological observation data is generated into a gridded meteorological data field covering the target dryland farmland area. Based on the gridded meteorological data field, the spatial distribution concentration of each disaster type is analyzed. By integrating the average occurrence frequency, the average process intensity, and the distribution concentration, and introducing a time trend correction coefficient, a single-hazard risk score is calculated for each type of meteorological disaster. The single-hazard risk scores of various meteorological disaster types faced by the target dryland farmland area are calculated using multi-hazard coupling to generate a comprehensive disaster-causing factor risk index.

3. The method for calculating the climate risk resilience of dryland farmland based on three-factor superposition as described in claim 1, characterized in that, Based on the aforementioned high-resolution remote sensing image data, soil profile physicochemical property data, and agricultural production management records, the exposure index of the disaster-bearing body is calculated, including: The high-resolution remote sensing image data is interpreted and classified to identify the spatial distribution of different land use types, including farmland, forest land, water area, and construction land, within the target dryland farmland area, and to extract the boundary, area, and spatial connectivity information of farmland plots. The available water content, soil texture, and soil layer thickness of each farmland plot are extracted from the soil profile physicochemical property data to evaluate and generate drought resistance indicators that characterize its basic water retention and drought resistance capabilities. Extract planting structure, sowing date, harvest date, and irrigation and fertilization records from the agricultural production management records, and calculate the time windows of different crops during their growth period and their consumption of climate resources. The spatial distribution, area, and spatial connectivity information of the farmland plots are overlaid and analyzed with the drought resistance index to assess the extent and degree of damage to farmland spatial entities during climate disasters and generate spatial exposure components. By matching the crop growth period window with the historical timing patterns of climate disasters, the degree of overlap between the crop growth process and the high-risk period of climate disasters in the time dimension is assessed, and a time exposure component is generated. The spatial exposure component and the temporal exposure component are weighted and combined to obtain the disaster-bearing body exposure index, which characterizes the spatiotemporal exposure features of the farmland system.

4. The method for calculating the climate risk resilience of dryland farmland based on three-factor superposition as described in claim 1, characterized in that, The disaster risk quantification algorithm is designed based on the temporal and frequency evolution patterns and spatial heterogeneity characteristics of climate events. Its working principle includes: A sliding time window is established to dynamically calculate the statistical characteristic values ​​of meteorological elements on the time series of the long-term meteorological observation data in order to capture the time-varying characteristics of climate state. Within each sliding time window, the intensity of meteorological disasters is fitted using extreme value distribution theory, and the disaster intensity under different return periods is estimated to quantify the extreme nature and frequency characteristics of disasters. Spatial autocorrelation analysis is introduced to calculate the spatial clustering and heterogeneity indicators of disaster intensity based on gridded meteorological data fields, so as to quantify the spatial pattern characteristics of disasters; Define a dynamic correction factor for disaster risk, which is a function of the temporal variation trend coefficient of disaster event frequency and the spatial variation coefficient of disaster intensity; The estimated disaster intensity and frequency, spatial clustering index, and the dynamic correction factor for disaster risk within the sliding window are nonlinearly combined to form a spatiotemporal dynamic risk assessment model for each grid cell. The calculation is performed by traversing all grid cells and applying the spatiotemporal dynamic hazard assessment model. The calculation results are then spatially aggregated to finally generate the hazard index of the disaster-causing factor.

5. The method for calculating the climate risk resilience of dryland farmland based on three-factor superposition as described in claim 1, characterized in that, Based on the multi-year, multi-point location monitoring test data and some of the agricultural production management record data, the vulnerability index of the disaster-bearing body was calculated, including: From the multi-year, multi-point location monitoring test data, actual crop yield data and corresponding meteorological data for different growth stages under different climate year types and management measures were extracted; The quantitative relationship between actual crop yield data and meteorological stress factors during key growth periods was analyzed using statistical models, and a yield-meteorological stress response function was constructed. Based on the yield-meteorological stress response function, the theoretical yield loss rate of the target farmland system is calculated under the preset standard meteorological stress scenario; From the agricultural production management record data, information on irrigation facility coverage, application of soil improvement measures, adoption rate of stress-resistant varieties, and adaptive measures is extracted, and their effectiveness in mitigating yield loss is evaluated. The speed and extent of crop recovery and growth after disasters are analyzed from the multi-year, multi-point location monitoring test data, and the intrinsic resilience of the farmland system is quantified. By coupling the theoretical yield loss rate, the mitigation effectiveness of adaptive measures, and the intrinsic resilience value, the vulnerability index of the disaster-bearing body, which reflects the overall situation of the system's sensitivity to climate stress and its self-adjustment capacity, is calculated.

6. The method for calculating the climate risk resilience of dryland farmland based on three-factor superposition as described in claim 5, characterized in that, The method utilizes statistical models to analyze the quantitative relationship between actual crop yield data and meteorological stress factors during key growth periods, constructing a yield-meteorological stress response function, including: Identify one or more critical growth stages throughout the crop's entire growth period that are most sensitive to yield formation; For each of the key reproductive stages, one or more meteorological stress factors are extracted from the meteorological data of the corresponding time period. The meteorological stress factors include the water stress index, the temperature stress index, and the sunshine stress index. Establish a multiple regression relationship between the actual crop yield data from multiple years and locations and all meteorological stress factors at all key growth stages, or use a machine learning model to train a nonlinear mapping relationship. Sensitivity analysis was used to determine the most significant functional forms and parameters in response to changes in various meteorological stress factors from the multiple regression or nonlinear mapping relationships. Based on a defined function form and parameters, a mathematical model is constructed to quantitatively describe the changes in yield caused by changes in the combination of meteorological stress factors, namely the yield-meteorological stress response function.

7. The method for calculating the climate risk resilience of dryland farmland based on three-factor superposition as described in claim 1, characterized in that, The normalized hazard index of the disaster-causing factor, the exposure index of the disaster-bearing body, and the vulnerability index of the disaster-bearing body are weighted, superimposed, and fused to generate a comprehensive climate risk resilience index for dryland farmland, including: The weight coefficients of the disaster-causing factor hazard index, disaster-bearing body exposure index, and disaster-bearing body vulnerability index in the comprehensive resilience evaluation are determined by using the analytic hierarchy process or the entropy weight method, respectively. The weighted risk value is obtained by multiplying the normalized hazard index of the disaster-causing factor by its corresponding weight coefficient. The normalized exposure index of the disaster-bearing body is multiplied by its corresponding weighting coefficient to obtain the weighted exposure value; The normalized vulnerability index of the disaster-bearing body is multiplied by its corresponding weight coefficient to obtain the weighted vulnerability value. The weighted hazard value, weighted exposure value, and weighted vulnerability value are substituted into a preset superposition and fusion model for calculation. The superposition and fusion model is a nonlinear function, whose input is the weighted value of the three factors and whose output is a scalar value between zero and one. The scalar value calculated by the superposition and fusion model is defined as the comprehensive resilience index of climate risk in dryland farmland. The higher the comprehensive resilience index of climate risk in dryland farmland, the stronger the resilience to climate risk.

8. The method for calculating the climate risk resilience of dryland farmland based on three-factor superposition as described in claim 7, characterized in that, The superposition and fusion model is a nonlinear function, and its construction and calculation process includes: The theoretical maximum and minimum values ​​of the comprehensive resilience index of climate risk in dry farmland are set, corresponding to the ideal states of complete resilience and complete lack of resilience, respectively. A contribution function is established for the weighted hazard value, weighted exposure value, and weighted vulnerability value to the comprehensive resilience index of climate risk in dryland farmland. The contribution function characterizes the nonlinear influence relationship of a single factor on the final index. Define interaction terms among factors to characterize the synergistic or antagonistic effects of the weighted hazard value, weighted exposure value, and weighted vulnerability value on resilience. The contribution function of a single factor is combined with the interaction term between factors to construct a nonlinear equation containing linear, quadratic, and interaction terms as the superposition and fusion model. The parameters of the overlay and fusion model are calibrated and verified using historical case data or expert knowledge; The weighted hazard value, weighted exposure value, and weighted vulnerability value of the area to be assessed are input into the calibrated superposition and fusion model to directly calculate the comprehensive resilience index of the dryland farmland climate risk of the area to be assessed.

9. The method for calculating the climate risk resilience of dryland farmland based on three-factor superposition as described in claim 3, characterized in that, The method also includes steps for spatial visualization and causal analysis of the calculation results: The calculated hazard index of disaster-causing factors, exposure index of disaster-bearing bodies, vulnerability index of disaster-bearing bodies, and the final comprehensive resilience index of climate risk for dryland farmland for each evaluation unit are associated with their corresponding geographic spatial locations. Using the mapping function of the Geographic Information System, spatial distribution thematic maps of the disaster-causing factor hazard index, disaster-bearing body exposure index, disaster-bearing body vulnerability index, and dryland farmland climate risk comprehensive resilience index were generated respectively. The main river systems, administrative boundaries, and key infrastructure locations of the target dryland farmland area are overlaid on the spatial distribution thematic map. For areas where the comprehensive resilience index of climate risk in dry farmland is low, the specific values ​​and spatial distribution characteristics of the corresponding disaster-causing factor hazard index, disaster-bearing body exposure index, and disaster-bearing body vulnerability index are traced back. By combining the backtested values ​​and characteristics, we analyze whether the main limiting factors leading to the low regional comprehensive resilience index are excessive risk, excessive exposure, or excessive vulnerability, and generate a report on the causes of spatial differentiation.

10. The method for calculating the climate risk resilience of dryland farmland based on the superposition of three factors according to claim 1, characterized in that, The method further includes: classifying the comprehensive climate risk resilience index of the dryland farmland according to a preset resilience level threshold range, determining and outputting the climate risk resilience level of the target dryland farmland area, including: Establish an evaluation standard that includes multiple resilience levels, set clear and mutually exclusive numerical ranges for the comprehensive resilience index of climate risk in dryland farmland for each resilience level, and form a reference table of resilience level threshold ranges. The calculated comprehensive resilience index of climate risk of the target dryland farmland area is compared one by one with each value range in the resilience level threshold range comparison table. When the value of the comprehensive resilience index of climate risk of dryland farmland falls within a preset value range corresponding to a certain resilience level, it is determined that the climate risk resilience of the target dryland farmland area belongs to the resilience level. The climate risk resilience level of the target dryland farmland area is determined and associated with the geographical location information of the target dryland farmland area and the specific value of the comprehensive climate risk resilience index of the dryland farmland area. Output a final report in the form of structured data files or visual charts, which includes geographic location information, the comprehensive resilience index of dryland farmland climate risk, and the results of climate risk resilience level determination.

Images