An integrated method for constructing drought-resistant plant communities in desert areas

Abstract

This application discloses an integrated method for constructing drought-resistant plant communities in desert areas. The method includes: acquiring soil desertification distribution data and environmental dynamic data of the target desert area, and dividing the area into regions; constructing spatial configuration models of drought-resistant plant communities in each region; conducting multi-dimensional ecological data monitoring and constructing a dynamic change matrix of soil desertification; predicting future desertification trends in each region; and dynamically adjusting plant community configuration strategies. This application significantly improves the success rate and sustainability of vegetation restoration in desert areas, accelerates the self-repair process of ecosystems, and reduces ecological restoration costs.

Description

Technical Field

[0001] This application relates to the field of ecological restoration technology, and more specifically, to a method for the integrated construction of drought-resistant plant communities in desert areas. Background Technology

[0002] Desertification is one of the most serious ecological and environmental problems facing the world. Currently, desert ecological restoration mainly adopts vegetation restoration methods, which involve planting drought-resistant plants to slow down wind erosion, stabilize sand dunes, and improve the soil environment. Traditional desert vegetation restoration technologies mainly include three methods: mechanical sand fixation, chemical sand fixation, and biological sand fixation. Mechanical sand fixation uses physical means such as setting up sand barriers and laying straw checkerboards to slow down wind speed and create conditions for plant growth; chemical sand fixation uses polymers and other chemical substances to spray and form a protective layer to temporarily stabilize the sand surface; biological sand fixation uses drought-resistant plants to achieve long-term and stable ecological restoration effects by utilizing the sand-fixing effect of plant roots and the windproofing effect of above-ground parts.

[0003] However, traditional desertification control methods generally employ single-function plant configuration strategies, lacking a holistic ecosystem design approach. This results in monotonous plant community structures and one-sided ecological functions, making it difficult to form a stable windbreak and sand-fixing system. For example, in desertification control, while large-scale planting of a single tree species may increase vegetation cover in the short term, neglecting water balance and soil improvement leads to severe vegetation decline in the later stages. Plant configuration schemes often rely on empiricism, lacking precise assessment and quantitative analysis of the differences in desert microenvironments. They fail to develop differentiated configuration strategies for areas with different degrees of desertification. For instance, in the edge of the Taklamakan Desert, the same vegetation restoration scheme shows significantly different effects in areas with different degrees of desertification. Furthermore, existing technologies adopt a "one-time design, static implementation" model, which cannot adapt to the dynamic changes in the desert environment. Especially against the backdrop of climate change, desert ecosystems face frequent extreme weather events, making it difficult for static restoration strategies to maintain long-term restoration results. Furthermore, the utilization of ecological monitoring data is severely insufficient. A large amount of valuable monitoring data has failed to be transformed into decision-making basis for ecological restoration, resulting in an "information silo" phenomenon between monitoring and restoration. Insufficient predictive capabilities lead to reactive governance measures, failing to proactively deploy prevention and control strategies against future desertification trends. For example, in the governance of the Mu Us Desert, the lack of effective predictive models made it difficult to cope with the damage of seasonal sandstorms to newly formed vegetation. Inadequate research on the synergistic effects of plants and soil microorganisms has overlooked the crucial role of microorganisms in desert ecosystems, leading to difficulties in plant establishment, slow growth, and the inability to form self-sustaining ecosystems. Ultimately, continuous human intervention is required to maintain these ecosystems, significantly increasing the cost of ecological restoration and reducing sustainability. Summary of the Invention

[0004] The main objective of this application is to provide an integrated method for constructing drought-resistant plant communities in desert areas, in order to overcome the aforementioned deficiencies of the prior art.

[0005] To achieve the above-mentioned objectives, this application provides the following technical solution.

[0006] Some embodiments of this application provide an integrated method for constructing drought-resistant plant communities in desert areas, including:

[0007] Acquire soil desertification distribution data and environmental dynamic data of the target desert area, and divide the target desert area into regions based on the soil desertification distribution data to obtain lightly desertified areas, moderately desertified areas and severely desertified areas;

[0008] Based on the regional division results and the environmental dynamic data, a spatial configuration model of drought-resistant plant communities in each region is constructed. The spatial configuration model includes an outer windbreak layer, a middle water-locking layer, and an inner soil amendment layer.

[0009] Multidimensional ecological data monitoring was performed on the spatial configuration model of drought-resistant plant communities in each region to obtain the time series sequence of multidimensional ecological data for each region, and a dynamic change matrix of soil desertification in each region was constructed based on the time series sequence of multidimensional ecological data.

[0010] Based on the aforementioned dynamic change matrix of soil desertification, the future desertification trend of each region is predicted, and regional desertification prediction results are generated.

[0011] Based on the regional desertification prediction results, the plant species distribution and spatial layout strategy of the drought-resistant plant community spatial configuration model are dynamically adjusted to achieve the integrated construction of drought-resistant plant communities in desert areas.

[0012] Furthermore, the step of dividing the target desert area into regions based on the soil desertification distribution data includes:

[0013] Acquire soil particle composition data, soil moisture content data, and soil organic matter content data for the target desert region;

[0014] Based on the proportion of fine particles with a diameter less than 0.05 mm in the soil particle composition data, the soil desertification index of each sampling point was calculated.

[0015] The soil desertification index is weighted, and the weighting process includes using the soil moisture content data as a negative correlation weight and the soil organic matter content data as a negative correlation weight to obtain the comprehensive desertification degree value of each sampling point.

[0016] Based on a preset desertification degree threshold range, the comprehensive desertification degree value is divided into a light desertification range, a moderate desertification range, and a severe desertification range, which correspond to the light desertification area, the moderate desertification area, and the severe desertification area, respectively.

[0017] The boundaries of the divided regions are smoothed to obtain continuous region division results.

[0018] In one embodiment, constructing a spatial configuration model of drought-resistant plant communities in each region includes:

[0019] Based on the soil desertification degree and wind speed distribution data of each region, the plant species and planting density of the outer windbreak layer are determined. The plant species include trees and shrubs with high wind resistance. The planting density is adaptively adjusted based on the gradient changes of wind speed distribution data.

[0020] Based on soil moisture evaporation rate and surface runoff data for each region, the plant species and root distribution characteristics of the middle water-locking layer are determined. The plant species include herbaceous plants with deep roots and high water-locking capacity. The root distribution characteristics are characterized by the ratio between root depth and root density.

[0021] Based on the soil organic matter content and microbial activity data of each region, the plant species and symbiotic microbial community configuration of the inner soil amendment layer are determined. The plant species include leguminous plants with nitrogen-fixing capabilities, and the symbiotic microbial community configuration is optimized using the soil microbial diversity index.

[0022] Based on topographic relief data and hydrological connectivity data of each region, the spatial stratification width and transition zone width of the outer windbreak layer, the middle water-locking layer and the inner soil improvement layer are determined. The transition zone width is calculated by normalizing the product of the standard deviation of the topographic relief data and the watershed convergence index of the hydrological connectivity data.

[0023] In one embodiment, the step of performing multidimensional ecological data monitoring on the spatial configuration model of the drought-resistant plant community in each region to obtain the time series sequence of multidimensional ecological data for each region includes:

[0024] Soil moisture sensors, wind speed sensors, and soil organic matter sensors are deployed in the outer windbreak layer, the middle water-locking layer, and the inner soil improvement layer in each region to obtain time-series data of soil moisture, wind speed, and soil organic matter in each layer.

[0025] Time series data of surface vegetation coverage and surface temperature were acquired for various regions based on UAV remote sensing technology.

[0026] The soil moisture time series data, the wind speed time series data, the soil organic matter time series data, the surface vegetation coverage time series data, and the surface temperature time series data are aligned along the time axis to generate multidimensional ecological data time series sequences for each region.

[0027] Outlier detection is performed on the multidimensional ecological data time series. The outlier detection includes calculating the local outlier factor of each time series data, marking data points with local outlier factors greater than a preset outlier threshold as outliers and performing interpolation repair.

[0028] In one embodiment, constructing a dynamic change matrix of soil desertification in each region based on the multidimensional ecological data time series includes:

[0029] Principal component analysis was performed on the time series of the multidimensional ecological data to extract the dominant ecological factors of each region. The dominant ecological factors include the rate of change of soil moisture, the rate of change of surface vegetation coverage, and the rate of change of wind speed.

[0030] Based on the dominant ecological factors, ecological factor time-series vectors for each region are constructed, wherein the ecological factor time-series vectors are composed of the time-series values ​​of each dominant ecological factor arranged in chronological order.

[0031] The rate of change of the time series vector of the ecological factors in each region within adjacent time windows is calculated, and the rate of change is characterized by the ratio of the Euclidean distance between adjacent time windows to the length of the time window.

[0032] The change rates of each region are arranged in a matrix according to time order and spatial location to generate a dynamic change matrix of soil desertification for each region. The rows of the dynamic change matrix of soil desertification represent the time dimension, and the columns represent the spatial dimension.

[0033] In one embodiment, predicting the future desertification trend of each region based on the soil desertification dynamic change matrix includes:

[0034] The dynamic change matrix of soil desertification is decomposed over time to obtain the trend component, periodic component and noise component.

[0035] Based on the trend components, a desertification trend prediction model for each region is constructed. The desertification trend prediction model is trained using a long short-term memory network. The training input is the historical data of the trend components, and the training output is the future predicted value of the trend components.

[0036] Fourier transform is performed on the periodic components to extract the desertification periodic characteristics of each region, including the period length and period amplitude.

[0037] The future predicted values ​​of the desertification trend prediction model are superimposed with the desertification cycle characteristics to generate a desertification trend prediction sequence for each region.

[0038] Based on the desertification trend prediction sequence and the preset desertification degree threshold range, the regional desertification prediction result is generated, which includes the change in the desertification degree level of each region in the future time period.

[0039] In one embodiment, the strategy of dynamically adjusting the plant species distribution and spatial layout of the drought-resistant plant community spatial configuration model based on the regional desertification prediction results includes:

[0040] Based on the changes in the degree of desertification in each region in the regional desertification prediction results, the regions that need to be adjusted and the adjustment priorities are determined. The adjustment priorities are calculated by normalizing the product of the magnitude and rate of change of the degree of desertification.

[0041] For areas where desertification has worsened, increase the planting density of the outer windbreak layer and the root density of the middle water-locking layer, while adjusting the proportion of nitrogen-fixing plants in the inner soil improvement layer.

[0042] For areas where desertification has been improved, the planting density of the outer windbreak layer should be reduced, the diversity of plant species in the middle water-locking layer should be increased, and the symbiotic microbial community configuration of the inner soil improvement layer should be optimized.

[0043] Based on the adjusted plant species distribution and spatial layout strategy, the spatial stratification width and transition zone width of the outer windbreak layer, the middle water-locking layer and the inner soil improvement layer are recalculated to generate an updated spatial configuration model of the drought-resistant plant community.

[0044] In one embodiment, acquiring time-series data of surface vegetation coverage and surface temperature for each region based on UAV remote sensing technology includes:

[0045] Using drones equipped with multispectral and thermal imaging cameras, multispectral and thermal images of various regions were acquired.

[0046] The normalized vegetation index was calculated on the multispectral images to obtain time-series data of surface vegetation coverage in each region.

[0047] Temperature calibration is performed on the thermal imaging images to obtain time-series data of surface temperature for each region;

[0048] Spatial registration is performed between the time series data of surface vegetation coverage and the time series data of surface temperature. The spatial registration includes affine transformation based on terrain feature points and non-rigid registration based on vegetation edges.

[0049] In one embodiment, performing principal component analysis on the time series of the multidimensional ecological data to extract the dominant ecological factors for each region includes:

[0050] The multidimensional ecological data time series is standardized to generate a standardized time series matrix;

[0051] Calculate the covariance matrix of the standardized time series matrix, and obtain the eigenvalues ​​and eigenvectors of the covariance matrix;

[0052] Based on the size of the feature values, select the principal components corresponding to the top few feature vectors whose cumulative contribution rate is greater than a preset contribution rate threshold.

[0053] The principal components are linearly combined with the original ecological factors in the time series of the multidimensional ecological data to extract the dominant ecological factors. The number of dominant ecological factors is determined by the number of principal components.

[0054] Compared with existing technologies, the integrated construction method for drought-resistant plant communities in desert areas provided in this application can significantly improve the success rate and sustainability of desert ecological restoration, enabling the restored vegetation community to have stronger resilience and self-repair capabilities, and effectively cope with the impact of extreme climate events. By accurately predicting desertification trends and intervening in advance, the overall cost of desertification control is significantly reduced, and the input of human and material resources is decreased. The systematic spatial configuration design accelerates the self-succession process of the ecosystem, shortens the ecological restoration cycle, and enables the formation of stable vegetation cover in desert areas more quickly. Differentiated governance strategies optimize resource allocation efficiency and avoid resource waste. The synergistic configuration of plants and microorganisms promotes soil health restoration, forming a virtuous cycle ecosystem and reducing subsequent maintenance costs. The closed-loop optimization mechanism shifts ecological restoration from passive response to active management, improving overall governance efficiency. Attached Figure Description

[0055] Figure 1 This is a schematic diagram of an integrated construction method for drought-resistant plant communities in desert areas, as described in one embodiment of this application. Detailed Implementation

[0056] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0057] This embodiment provides an integrated method for constructing drought-resistant plant communities in desert areas. The entities executing this method include, but are not limited to, computer systems, mobile monitoring stations, edge computing devices, and plant ecological monitoring equipment. These entities can also be considered as general computing nodes in this application. The computer system includes, but is not limited to, at least one of servers, workstations, and embedded controllers.

[0058] In summary, the integrated construction method for drought-resistant plant communities in this desert region can include: acquiring soil desertification distribution data and environmental dynamic data of the target desert region; dividing the target desert region into regions based on the soil desertification distribution data; constructing a spatial configuration model of drought-resistant plant communities in each region based on the regional division results and environmental dynamic data; monitoring multidimensional ecological data of the spatial configuration model of drought-resistant plant communities in each region, obtaining multidimensional ecological data time series, and constructing a dynamic change matrix of soil desertification; predicting the future desertification trend of each region based on the dynamic change matrix of soil desertification, and generating regional desertification prediction results; and dynamically adjusting the plant species distribution and spatial layout strategy of the spatial configuration model of drought-resistant plant communities based on the regional desertification prediction results.

[0059] This method achieves three-dimensional construction of desert ecosystems through a hierarchical plant community spatial configuration model. It adopts a three-layer structure design of "outer windbreak layer, middle water-locking layer, and inner soil improvement layer" to improve the ecosystem's resilience and self-repair capacity. By monitoring and analyzing multi-dimensional ecological data, a dynamic change matrix of desertification is established to achieve accurate prediction of the changing trend of desert ecosystems. Based on the prediction results, the plant community configuration strategy is dynamically adjusted to form a closed-loop optimization mechanism, which significantly improves the success rate and sustainability of vegetation restoration in desert areas.

[0060] Specifically, the detailed implementation steps of the integrated construction method for drought-resistant plant communities in this desert region include:

[0061] First, soil desertification distribution data and environmental dynamic data for the target desert area were acquired using soil sampling and environmental monitoring equipment. Soil desertification distribution data included key indicators such as soil particle composition, moisture content, and organic matter content, and was collected using a grid-based sampling method covering the target area. Environmental dynamic data included meteorological factors such as wind speed distribution, precipitation patterns, and surface temperature changes, acquired through the fusion of automatic weather station and remote sensing data. All this data was transmitted to the data processing center using standardized data acquisition protocols to ensure accuracy and completeness.

[0062] Based on soil desertification distribution data, a hierarchical clustering algorithm was used to divide the target desert region into lightly desertified, moderately desertified, and severely desertified areas. The division process was based on the soil desertification index, taking into account the weighted influence of soil moisture content and organic matter content, generating continuous regional division results. This multi-dimensional index-based regional division method avoids the limitations of traditional single-index division and can more accurately reflect the complexity of the desert ecosystem.

[0063] Based on regional division results and environmental dynamic data, a spatial configuration model of drought-resistant plant communities was constructed for each region. The spatial configuration model includes an outer windbreak layer, a middle water-locking layer, and an inner soil amendment layer, forming a three-dimensional ecological protection structure. The outer windbreak layer uses trees and shrubs with high wind resistance, the middle water-locking layer uses herbaceous plants with deep root systems and high water-locking capacity, and the inner soil amendment layer uses legumes with nitrogen-fixing capabilities and their symbiotic microorganisms. This multi-layered spatial configuration fully considers the synergistic effects and niche complementarity among plants, constructing a plant community structure with high resilience.

[0064] Multidimensional ecological data monitoring was conducted on the spatial configuration models of drought-resistant plant communities in various regions to obtain multidimensional ecological data time series for each region. The monitoring system included a soil sensor network, meteorological monitoring stations, and a UAV remote sensing platform to continuously monitor key indicators such as soil moisture, wind speed, organic matter content, vegetation cover, and surface temperature. The acquired data underwent time axis alignment and outlier processing to form multidimensional ecological data time series, providing a high-quality data foundation for subsequent analysis.

[0065] Based on multidimensional ecological data time series, principal component analysis was used to extract dominant ecological factors and construct a dynamic change matrix of soil desertification in each region. This matrix unifies the temporal and spatial dimensions, and can intuitively reflect the dynamic changes in the desertification process. In the matrix construction process, the interactions and cumulative effects among ecological factors were considered, improving the ability to represent changes in complex ecosystems.

[0066] Based on the dynamic change matrix of soil desertification, a desertification trend prediction model was established using a long short-term memory network, and regional desertification prediction results were generated by combining periodic analysis. The prediction process not only considered historical trends but also incorporated the influence of seasonal variations and climate cycles, achieving multi-scale prediction of desertification trends. The obtained prediction results include changes in the degree of desertification in each region over future time periods, providing a scientific basis for the dynamic adjustment of plant communities.

[0067] Based on regional desertification predictions, the plant species distribution and spatial layout strategies of the drought-resistant plant community spatial configuration model are dynamically adjusted. For areas with worsening desertification, windbreak and water retention functions are enhanced; for areas with improved desertification, species diversity and symbiotic microbial community configuration are optimized. This prediction-based dynamic adjustment strategy forms a closed-loop optimization mechanism, significantly improving the adaptability and recovery efficiency of desert ecosystems.

[0068] Furthermore, the detailed implementation steps of this method for regionalizing the target desert area based on soil desertification distribution data may include:

[0069] First, soil particle size distribution, moisture content, and organic matter content data for the target desert area were acquired using soil sampling equipment. The sampling equipment included a portable soil particle size analyzer, a TDR soil moisture sensor, and a carbon and nitrogen analyzer. Sampling points were arranged in a grid pattern to cover the target area, with the grid density dynamically adjusted according to the terrain complexity to ensure representativeness. Soil particle size distribution data included the percentage of soil particles of different sizes, with particular attention paid to the proportion of fine particles smaller than 0.05 mm, as this indicator directly relates to the soil's resistance to wind erosion and its water retention capacity. Soil moisture content data was collected stratified to reflect the vertical distribution characteristics of soil moisture, while soil organic matter content data reflected soil fertility and microbial activity potential.

[0070] Based on the proportion of fine particles smaller than 0.05 mm in the soil particle size composition data, the soil desertification index was calculated for each sampling point. The calculation formula is:

[0071] Soil desertification index = 1 - (percentage of particles < 0.05 mm) / (reference standard value);

[0072] The reference standard value is the average proportion of fine particles in the same soil type in non-desertified areas. A higher index value indicates more severe soil desertification; a lower value indicates a more stable soil structure. During the calculation process, data from outlier sampling points are weighted to avoid the impact of local outliers on the overall assessment.

[0073] The soil desertification index was weighted to improve the comprehensiveness and accuracy of the assessment. The weighting process included negatively weighting soil moisture content data and soil organic matter content data to obtain a comprehensive desertification degree value for each sampling point. The weighted calculation formula is as follows:

[0074]

[0075] In this study, ZH represents the comprehensive desertification degree value, TZ represents the soil desertification index, SH represents the moisture content, SF represents the moisture reference value, YH represents the organic matter content, CH represents the organic matter reference value, and α and β are weighting coefficients, reflecting the influence of moisture and organic matter on the degree of desertification, respectively, determined through regression analysis of historical data. The moisture and organic matter reference values ​​are the standard values ​​for non-desertified soils in the region. This multi-factor weighted approach considers the inhibitory effects of soil moisture and organic matter on the desertification process, making the assessment results more consistent with ecological principles.

[0076] Based on preset desertification severity threshold ranges, the comprehensive desertification severity value is divided into mild, moderate, and severe desertification ranges, corresponding to mild, moderate, and severe desertification areas, respectively. The threshold ranges are set using a quantile method combined with expert knowledge to ensure that the classification results conform to both statistical distribution patterns and ecological significance. In special terrain areas (such as mountain-plain transition zones), the threshold ranges are adjusted for terrain features to improve the adaptability of the classification.

[0077] The boundaries of the divided regions are smoothed to obtain continuous regional division results. The smoothing process combines geospatial interpolation methods with morphological processing to eliminate the jagged effect of regional boundaries caused by limited sampling points, generating ecologically meaningful continuous regional interfaces. During the smoothing process, topographic factors are introduced as auxiliary variables to ensure that regional boundaries are consistent with topographic features, enhancing the practicality of the division results. The smoothed regional division results are stored in both raster and vector data formats for easy subsequent analysis and application.

[0078] Furthermore, the detailed implementation steps for constructing spatial configuration models of drought-resistant plant communities in each region in this method may include:

[0079] Based on soil desertification levels and wind speed distribution data for each region, the plant species and planting density for the outer windbreak layer were determined. Plant species included trees and shrubs with high wind resistance, such as tamarisk, oleaster, and saxaul, which possess well-developed root systems and leaf structures adapted to aeolian environments. The planting density was adaptively adjusted based on gradient changes in wind speed distribution data, calculated using the following formula:

[0080]

[0081] Where ZG represents planting density, JG represents base density, FG represents wind speed value, FH represents wind speed threshold, base density is the recommended planting density under standard conditions, γ is the wind speed response coefficient, and the wind speed threshold is 80% of the region's historical highest wind speed. This design allows for increased planting density in areas with high wind speeds to enhance wind protection; in areas with low wind speeds, density can be appropriately reduced to optimize resource utilization. In special areas with extremely high wind speeds, a "dense-sparse-dense" strip structure is used to form a wind buffer zone, further improving wind protection.

[0082] Based on soil moisture evaporation rates and surface runoff data for each region, the plant species and root distribution characteristics of the middle aquifer were determined. Plant species included herbaceous plants with deep root systems and high water retention capacity, such as *Stipa*, *Artemisia*, and *Agropyron*, which can reduce water transpiration loss through special physiological structures. Root distribution characteristics were characterized by the ratio between root depth and root density, reflecting the relationship between plant water acquisition strategies and soil water use efficiency. In areas with high evaporation rates, plant species with a large root-to-density ratio were selected to acquire deep water through vertical root development; in areas with large surface runoff, plant species with a small root-to-density ratio were selected to intercept surface runoff through horizontal root development. This plant configuration strategy based on hydrological characteristics significantly improved water use efficiency and is the core innovation of the middle aquifer design.

[0083] Based on soil organic matter content and microbial activity data for each region, the plant species and symbiotic microbial community configuration of the inner soil amendment layer were determined. Plant species included nitrogen-fixing legumes such as *Alternanthera philoxeroides*, *Sweet clover*, and *Ilex chinensis*, which can improve soil fertility through symbiotic mycorrhizal fungi and nitrogen-fixing bacteria. The symbiotic microbial community configuration was optimized using the soil microbial diversity index, calculated using the following formula:

[0084] Microbial community configuration index = proportion of nitrogen-fixing bacteria × (1 + δ × microbial diversity index);

[0085] δ represents the diversity response coefficient, determined experimentally. In areas with extremely low organic matter content, increasing the proportion of nitrogen-fixing bacteria accelerates organic matter accumulation; in areas with high microbial activity, emphasis is placed on optimizing the microbial community structure to improve ecosystem stability. The design concept of the inner soil amendment layer is to promote soil health through a mutually beneficial symbiotic system of "plants and microorganisms," providing a good soil environmental foundation for the entire plant community.

[0086] Based on topographic relief and hydrological connectivity data for each region, the spatial stratification widths and transition zone widths of the outer windbreak layer, middle water-locking layer, and inner soil amendment layer were determined. The transition zone width was calculated by normalizing the product of the standard deviation of the topographic relief data and the watershed convergence index of the hydrological connectivity data. The calculation formula is as follows:

[0087] Transition zone width = base width × (standard deviation of topographic relief × watershed convergence index) / normalization factor;

[0088] The base width is the standard transition zone width under flat terrain conditions, and a normalization factor ensures that the calculation results are within a reasonable range. In areas with drastic topographic changes, the transition zone width is increased to buffer environmental gradient changes; in areas with high hydrological connectivity, the transition zone width is also increased accordingly to promote effective water transport between different layers. This spatial layout optimization based on topographic and hydrological characteristics greatly improves the ecological adaptability and anti-disturbance ability of the plant community structure.

[0089] Furthermore, the detailed implementation steps of this method for monitoring multidimensional ecological data of the spatial configuration model of drought-resistant plant communities in each region and obtaining the time series sequence of multidimensional ecological data for each region may include:

[0090] Soil moisture sensors, wind speed sensors, and soil organic matter sensors are deployed in the outer windbreak layer, middle water-locking layer, and inner soil amendment layer of each region to construct a multi-layered, comprehensive ecological monitoring network. The soil moisture sensor employs TDR technology, enabling multi-depth monitoring and capturing the vertical distribution characteristics of soil moisture in the profile. The wind speed sensor utilizes ultrasonic principles, featuring low power consumption and high precision, monitoring near-surface wind field distribution. The soil organic matter sensor, based on near-infrared spectroscopy, enables in-situ rapid determination of soil organic matter content. These sensors are connected to edge computing units via a wireless sensor network to form a distributed monitoring system, acquiring time-series data on soil moisture, wind speed, and soil organic matter at each layer.

[0091] This study utilizes UAV remote sensing technology to acquire time-series data on land vegetation cover and land surface temperature for various regions. UAVs, equipped with multispectral and thermal imaging cameras, periodically conduct aerial surveys of target areas along preset routes to obtain high-resolution remote sensing data. Vegetation cover data is obtained by calculating the Normalized Difference Vegetation Index (NDVI), and land surface temperature data is obtained through thermal infrared radiation correction. During data acquisition, a sunlight angle compensation algorithm and an atmospheric influence correction model are introduced to improve the accuracy of the remote sensing data. To ensure the spatiotemporal continuity of the data, a data acquisition procedure with fixed flight routes and fixed time periods is established to minimize data fluctuations caused by non-ecological factors.

[0092] Time-series data on soil moisture, wind speed, soil organic matter, land cover, and land surface temperature were aligned along their time axes to generate multidimensional ecological data time-series sequences for each region. A dynamic time warping algorithm was employed for time-axis alignment to address inconsistencies in timestamps across different monitoring data sets, ensuring data comparability. During the alignment process, a time compensation mechanism was introduced to account for sensor data latency and differences in remote sensing data acquisition cycles, achieving accurate fusion of heterogeneous data sources. The aligned multidimensional ecological data time-series sequences are organized with a unified time scale, providing a structured data foundation for subsequent analysis.

[0093] Outlier detection is performed on multidimensional ecological time series data to improve data quality. Outlier detection includes calculating the local outlier factor for each time series data point, marking data points with local outlier factors exceeding a preset outlier threshold as outliers, and then performing interpolation for correction. The calculation of the local outlier factor is based on the statistical distance between a data point and its temporal neighbors, which can effectively identify abrupt changes and outliers in the time series data. In outlier correction, a constrained interpolation algorithm that maintains the ecological change trend is adopted to avoid introducing new data distortions during the correction process. The corrected multidimensional ecological time series data possesses high integrity and continuity, providing reliable data support for constructing a dynamic change matrix of soil desertification.

[0094] Furthermore, the detailed implementation steps of constructing the dynamic change matrix of soil desertification in each region based on the time series of multidimensional ecological data in this method may include:

[0095] Principal component analysis (PCA) was performed on the time series of multidimensional ecological data to extract the dominant ecological factors for each region. PCA can identify key variables from high-dimensional data, reduce data redundancy, and improve analytical efficiency. Dominant ecological factors include soil moisture change rate, surface vegetation cover change rate, and wind speed change rate, which are directly related to the dynamic changes in the desertification process. In the PCA process, the Kaiser criterion combined with scree plots was used to determine the optimal number of principal components, preserving key information while avoiding oversimplification. To improve the interpretability of the principal components, a Varimax orthogonal rotation was performed on the principal component loading matrix to clarify the correspondence between the principal components and the original ecological factors.

[0096] Based on dominant ecological factors, time-series vectors of ecological factors are constructed for each region. These vectors consist of the time-series values ​​of each dominant ecological factor arranged in chronological order. The time-series vectors employ multi-resolution representation, incorporating both original high-frequency sampled data and aggregated statistical values ​​at different time scales, enabling the simultaneous capture of short-term fluctuations and long-term trends. During vector construction, environmental background variables (such as seasonal cycles and climate anomalies) are introduced as auxiliary information to enhance the ecological interpretability of the time-series data. The constructed time-series vectors of ecological factors represent the dynamic characteristics of the desertification process, providing a foundation for subsequent calculations of the rate of change.

[0097] The rate of change of ecological factor time-series vectors in each region within adjacent time windows is calculated to quantify the dynamic characteristics of the desertification process. The rate of change is characterized by the ratio of the Euclidean distance between adjacent time windows to the length of the time window, and the calculation formula is as follows:

[0098]

[0099] Where VH represents the rate of change, V(t) and V(t+1) represent the time-series vectors of ecological factors in adjacent time windows, ||·|| represents the Euclidean norm, and Δt is the length of the time window. This calculation method can comprehensively consider the coordinated changes of multiple ecological factors, providing an overall measure of the dynamics of the desertification process. In the calculation of the rate of change, an adaptive time window technique is adopted, dynamically adjusting the window length according to the rate of data change. A shorter window is used for rapid changes, and a longer window is used for slow changes, improving the sensitivity and adaptability of the rate of change calculation.

[0100] The change rates of each region are arranged in chronological order and spatial location to form a matrix, generating a dynamic change matrix of soil desertification for each region. In this matrix, rows represent the time dimension, columns represent the spatial dimension, and matrix element values ​​represent the desertification change rate at a specific spatiotemporal point. This matrix representation unifies temporal evolution and spatial distribution into a single mathematical structure, facilitating global analysis and pattern recognition. During matrix construction, a geographically weighted normalization method is used to eliminate the influence of differences in region size and shape, ensuring the comparability of matrix elements across different regions. The completed dynamic change matrix of soil desertification provides a complete expression of the spatiotemporal characteristics of the desertification process, offering crucial input for predicting desertification trends.

[0101] Furthermore, the detailed implementation steps of this method for predicting the future desertification trend of each region based on the dynamic change matrix of soil desertification may include:

[0102] A time-series decomposition was performed on the dynamic change matrix of soil desertification to obtain trend, periodic, and noise components. The time-series decomposition employed a combination of wavelet transform and empirical mode decomposition, which effectively handles non-stationary time series and separates the variation characteristics at different time scales. The trend component reflects the long-term evolution direction of the desertification process, the periodic component reflects seasonal and interannual fluctuations, and the noise component includes random fluctuations and measurement errors. During the decomposition process, consistent decomposition parameters were used for matrices from different regions to ensure the comparability of component extraction. Simultaneously, an optimal decomposition level was set based on the characteristic time scales of the desert ecosystem to balance decomposition accuracy and computational complexity.

[0103] Based on trend components, a desertification trend prediction model is constructed for each region. This model is trained using a Long Short-Term Memory (LSTM) network. LSTM is a special type of recurrent neural network capable of capturing long-term dependencies, making it suitable for processing time series data such as desertification processes, which possess long-term memory characteristics. Model training employs a sliding window method to generate training samples, using historical data sequences as input and future data points as output. Backpropagation is used to optimize network parameters. During training, early stopping and regularization techniques are introduced to prevent overfitting; simultaneously, a learning rate decay strategy is employed to improve training stability and convergence speed. The trained model can predict future trend evolution based on historical trend components, providing the main framework for desertification trend prediction.

[0104] Fourier transforms were performed on the periodic components to extract the periodic characteristics of desertification in each region. These periodic characteristics include period length and period amplitude, reflecting the time scale and intensity of the periodic changes in the desertification process, respectively. Fourier transform converts time-domain signals into frequency-domain representations, visually displaying the periodic structure of the signal. In the periodic analysis, a power spectral density estimation method was used to identify significant periodic components; for non-integer periodic signals, the Lomb-Scargle periodogram method was employed to improve the accuracy of period detection. The extracted periodic characteristics include not only seasonal cycles but also long-period components related to climate cycles (such as the El Niño Southern Oscillation), comprehensively capturing the periodic variation patterns of the desertification process.

[0105] The future predictions from the desertification trend prediction model are overlaid with desertification cycle characteristics to generate desertification trend prediction sequences for each region. The overlay process considers the interaction between trend and cycle, and phase-locking technology is used to ensure time alignment between the cycle component and the trend component. The formula for generating the prediction sequence is:

[0106]

[0107] Where XU is the predicted sequence, YT is the trend prediction value, and Ai, Ti and These represent the amplitude, period length, and phase of the i-th periodic component, respectively. This trend-period superposition method retains the predictive power of long-term trends while incorporating the influence of periodic fluctuations, thus improving the comprehensiveness and accuracy of predictions.

[0108] Based on the desertification trend prediction sequence and the preset desertification degree threshold range, regional desertification prediction results are generated. These results include changes in the desertification degree level of each region over a future time period, as well as the time points and rates of change. During the result generation process, prediction uncertainty analysis is introduced, and prediction intervals are generated using Monte Carlo simulation to quantify the reliability of the prediction results. Simultaneously, sensitivity analysis is combined to identify key factors affecting the accuracy of the predictions, providing a scientific basis for interpreting the results. The final prediction results are presented in a spatiotemporal visualization, intuitively demonstrating the spatiotemporal evolution patterns of desertification trends in each region, providing decision support for the dynamic adjustment of plant community configurations.

[0109] Furthermore, the detailed implementation steps of this method for dynamically adjusting the plant species distribution and spatial layout strategy of the drought-resistant plant community spatial configuration model based on regional desertification prediction results may include:

[0110] Based on the changes in desertification severity levels in each region according to the regional desertification prediction results, the regions requiring adjustment and their adjustment priorities are determined. The adjustment priority is calculated by normalizing the product of the magnitude and rate of change in desertification severity level, using the following formula:

[0111] Adjustment priority = (Level change magnitude × Change rate) / Normalization factor

[0112] The classification method uses the magnitude of desertification level change as the maximum change in desertification level during the prediction period, the rate of change as the slope of the level change, and a normalization factor to ensure that the priority value is within the range of [0, 1]. This priority calculation method based on change characteristics can effectively identify areas requiring urgent intervention and optimize resource allocation efficiency. In the priority determination process, regional ecological sensitivity and protection importance are also considered, giving higher priority to ecologically sensitive areas or areas with important protective functions to ensure the stability of key nodes in the ecosystem.

[0113] For areas where desertification is worsening, an enhanced adjustment strategy is adopted, increasing the planting density of the outer windbreak layer and the root density of the middle water-locking layer, while adjusting the proportion of nitrogen-fixing plants in the inner soil amendment layer. Specifically, the increase in planting density in the outer windbreak layer is proportional to the predicted increase in wind speed, the increase in root density in the middle water-locking layer is proportional to the predicted decrease in water content, and the increase in the proportion of nitrogen-fixing plants in the inner soil amendment layer is proportional to the predicted decrease in organic matter. This targeted adjustment strategy can specifically enhance the plant community's ability to cope with deteriorating conditions and effectively curb the desertification process. In terms of plant species selection, the proportion of species with strong resilience is increased, such as the drought-resistant Tamarix chinensis and the wind-erosion-resistant Artemisia argyi, to improve the ecological resilience of the community.

[0114] For areas where desertification has improved, an optimization strategy is adopted. This involves reducing the planting density of the outer windbreak layer, increasing the diversity of plant species in the middle water-locking layer, and optimizing the symbiotic microbial community configuration in the inner soil amendment layer. The increase in plant diversity follows the principle of niche complementarity, selecting species with different niche characteristics, such as plants with different root depths and photosynthetic pathways, to maximize resource utilization efficiency. Optimization of the symbiotic microbial community focuses on functional diversity, increasing the proportion of functional bacteria such as phosphate-solubilizing bacteria and growth-promoting bacteria to promote soil nutrient cycling and plant growth. This optimization strategy, based on the improvement trend, aims to promote the transformation of the ecosystem from a "resistance-based" to a "sustainable" model, enhancing the system's self-sustaining capacity and ecological value.

[0115] Based on the adjusted plant species distribution and spatial layout strategy, the spatial stratification widths and transition zone widths of the outer windbreak layer, middle water-locking layer, and inner soil amendment layer were recalculated, generating an updated spatial configuration model of the drought-resistant plant community. The calculation of spatial stratification width comprehensively considers wind speed gradient, moisture gradient, and soil property gradient, establishing a mapping relationship between "environmental gradient and spatial structure" to match the spatial structure of the plant community with environmental characteristics. The optimization of the transition zone width is based on the intensity of species interactions; the transition zone width is increased in areas with strong mutualistic interactions to promote positive interactions, while the transition zone width is reduced in areas with strong competitive interactions to decrease negative interactions. The updated spatial configuration model was validated through computer simulation to assess its adaptability to predicted environmental conditions, ensuring the scientific validity and effectiveness of the adjustment scheme.

[0116] Furthermore, the detailed implementation steps of this method for acquiring time-series data of land surface vegetation coverage and land surface temperature in various regions based on UAV remote sensing technology may include:

[0117] The system utilizes drones equipped with multispectral and thermal imaging cameras to acquire multispectral and thermal images of various regions. The drones employ a hybrid fixed-wing and rotary-wing design, combining cruising efficiency with hovering capabilities to meet the long-distance monitoring needs of desert environments. The multispectral camera features five bands: red, green, blue, near-infrared, and red-edge, with a resolution better than 10 cm / pixel, capable of capturing subtle vegetation changes. The thermal imaging camera has a temperature resolution better than 0.1℃ and a spatial resolution better than 30 cm / pixel, enabling precise measurement of surface temperature distribution. Data acquisition follows a pre-set flight path, designed to consider regional area, terrain changes, and monitoring accuracy requirements, ensuring complete data coverage and appropriate overlap. Under special weather conditions (such as strong winds), the system automatically adjusts flight parameters or replans the flight path to guarantee data quality.

[0118] Normalized Difference Vegetation Index (NDVI) was calculated from multispectral images to obtain time-series data of land cover in each region. The NDVI is calculated based on the difference in reflectance between the near-infrared and red light bands, using the following formula:

[0119]

[0120] NIR and RED represent the reflectance in the near-infrared and red light bands, respectively. NDVI values ​​range from -1 to 1, with higher values ​​indicating better vegetation cover. To improve the accuracy of NDVI calculations, atmospheric scattering correction and soil background adjustment are introduced to reduce interference from non-vegetation factors. Under sparse desert vegetation conditions, the Enhanced Vegetation Index (EVI) is also used as an auxiliary indicator. EVI is less sensitive to soil background and atmospheric influences and can more accurately characterize low-coverage vegetation. Based on the NDVI and EVI calculation results, a regression model is used to convert them into absolute vegetation cover, forming standardized time-series vegetation cover data.

[0121] Temperature calibration is performed on thermal imaging images to obtain time-series surface temperature data for each region. The temperature calibration process includes three key steps: radiometric correction, atmospheric influence compensation, and surface emissivity correction. Radiometric correction converts the digital values ​​output by the thermal imager into radiance; atmospheric influence compensation eliminates the effects of atmospheric absorption and scattering on thermal radiation; and surface emissivity correction considers the emissivity differences among different surface types, improving the accuracy of temperature calculation. Surface emissivity correction is particularly important in desert environments because surfaces with different degrees of sandification exhibit different thermal radiation characteristics. This method innovatively introduces an emissivity estimation model based on spectral characteristics, automatically estimating surface emissivity from multispectral data, significantly improving the accuracy of temperature calibration. The calibrated temperature data undergoes spatial filtering and temporal smoothing to form high-quality time-series surface temperature data.

[0122] Spatial registration was performed on time-series data of land surface vegetation cover and land surface temperature to ensure spatial consistency. Spatial registration consisted of two stages: affine transformation based on topographic feature points and non-rigid registration based on vegetation edges. Affine transformation addressed overall geometric deformation, while non-rigid registration handled local deformation; the combination of both achieved high-precision registration. For feature point selection, time-stable topographic features (such as ridgelines and valley lines) were prioritized as control points to improve registration reliability. In non-rigid registration, a mutual information maximization method based on vegetation edges was employed to address the challenge of feature sparsity in desert environments. Registration accuracy was controlled using root mean square error (RMSE) to ensure the error was less than one pixel. The registered data possessed pixel-level spatial correspondences, supporting subsequent pixel-level analysis and spatiotemporal pattern recognition.

[0123] Furthermore, the detailed implementation steps of this method for performing principal component analysis on the time series of multidimensional ecological data to extract the dominant ecological factors of each region may include:

[0124] Multidimensional ecological time series data were standardized to generate a standardized time series matrix. Standardization eliminated dimensional differences between different ecological factors, making the data comparable. The Z-score method was used for standardization. This standardization resulted in a mean of 0 and a standard deviation of 1 for each ecological factor, eliminating the influence of different measurement scales. During the standardization process, considering the potential non-normal distribution of ecological data, Box-Cox transformation was used for preprocessing significantly skewed data to make it closer to a normal distribution, improving the effectiveness of principal component analysis. The standardized time series matrix preserved the original data's temporal structure, with each row representing a time point and each column representing an ecological factor.

[0125] The covariance matrix of the standardized time series matrix is ​​calculated to obtain its eigenvalues ​​and eigenvectors. The covariance matrix reflects the correlations between different ecological factors and is the core data structure of principal component analysis. The eigenvalues ​​and eigenvectors are calculated using the Singular Value Decomposition (SVD) algorithm. Compared to traditional eigenvalue decomposition methods, SVD has better numerical stability and is particularly suitable for processing high-dimensional ecological data. During the calculation, sparse matrix optimization techniques are employed to improve the efficiency of large-scale data processing; simultaneously, parallel computing accelerates eigenvalue decomposition to meet real-time analysis requirements. Eigenvalues ​​represent the variance explained by each principal component, while eigenvectors define the mapping relationship between the original ecological factors and the principal component space.

[0126] Based on the eigenvalues, the principal components are selected from the top few eigenvectors whose cumulative contribution rate exceeds a preset contribution rate threshold. The cumulative contribution rate, the proportion of the total variance explained by the selected principal components to the total variance of the original data, is a key indicator for determining the number of principal components. The preset contribution rate threshold is typically set between 85% and 95%, reasonably reducing dimensionality while retaining key information. In eigenvector selection, in addition to considering the cumulative contribution rate, the Kaiser criterion (eigenvalue greater than 1) and the scree plot method are also combined to comprehensively determine the optimal number of principal components. This multi-criteria selection method achieves a good balance between information preservation and dimensionality reduction, avoiding oversimplification or redundant retention caused by simply applying a single criterion.

[0127] The principal components are linearly combined with the original ecological factors in the time series of multidimensional ecological data to extract the dominant ecological factors. The dominant ecological factors are a dimensionality-reduced representation of the original multidimensional data, preserving key information while simplifying the data structure for easier subsequent analysis and interpretation. The weights of the linear combination are determined based on the element values ​​of the eigenvectors, calculated using the following formula:

[0128] Z_i = ∑(aij × Y_j)

[0129] Where aij is the j-th element of the i-th eigenvector, Z_i is the i-th dominant ecological factor, and Y_j is the j-th original ecological factor. To improve the interpretability of the dominant ecological factors, the principal component loading matrix is ​​subjected to Varimax orthogonal rotation, ensuring that each principal component is primarily correlated with a few original variables, reducing cross-loading, and clarifying the ecological significance of the principal components. The rotated dominant ecological factors typically correspond to indicators with clear ecological significance, such as soil moisture change rate, surface vegetation cover change rate, and wind speed change rate, providing a scientific framework for desertification dynamic analysis.

[0130] Through the above implementation methods, this application constructs a spatial configuration model of drought-resistant plant communities with a three-layer structure of "outer windbreak layer, middle water-locking layer, and inner soil improvement layer". Through multi-dimensional ecological data monitoring, desertification dynamic analysis and trend prediction, the dynamic optimization and adjustment of plant communities are realized, which significantly improves the restoration efficiency and sustainability of desert ecosystems.

[0131] It should be noted that the above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0132] The above are merely preferred embodiments of this application. The scope of protection of this application is not limited to the above embodiments. All technical solutions within the scope of this application's concept are within the scope of protection of this application. It should be noted that for users of ordinary technical skills, any improvements and modifications made without departing from the principles of this application should also be considered within the scope of protection of this application.

Claims

1. A method for the integrated construction of drought-resistant plant communities in desert areas, characterized in that, The method comprises the following steps: obtain the soil desertification degree distribution data and environmental dynamic data of the target desert area, and divide the target desert area according to the soil desertification degree distribution data to obtain a slight desertification region, a moderate desertification region and a severe desertification region; based on the regional division result and the environmental dynamic data, a spatial configuration model of a drought-resistant plant community in each region is constructed, and the spatial configuration model comprises an outer windproof layer, a middle water-locking layer and an inner soil improvement layer; multi-dimensional ecological data monitoring is performed on the spatial configuration model of the drought-resistant plant community in each region to obtain a multi-dimensional ecological data time sequence of each region, and a soil desertification dynamic change matrix of each region is constructed based on the multi-dimensional ecological data time sequence; based on the soil desertification dynamic change matrix, the future desertification trend of each region is predicted, and a regional desertification prediction result is generated; according to the regional desertification prediction result, the plant species distribution and spatial layout strategy of the spatial configuration model of the drought-resistant plant community are dynamically adjusted to realize integrated construction of the drought-resistant plant community in the desert area; wherein the construction of the spatial configuration model of the drought-resistant plant community in each region comprises: determining the plant species and planting density of the outer windproof layer according to the soil desertification degree and wind speed distribution data of each region, wherein the plant species comprises trees and shrubs with high wind resistance, and the planting density is adaptively adjusted based on the gradient change of the wind speed distribution data; determining the plant species and root distribution characteristics of the middle water-locking layer according to the soil water evaporation rate and surface runoff data of each region, wherein the plant species comprises herbaceous plants with deep root systems and high water-locking capacity, and the root distribution characteristics are characterized by the ratio between root depth and root density; determining the plant species and symbiotic flora configuration of the inner soil improvement layer according to the soil organic matter content and microbial activity data of each region, wherein the plant species comprises leguminous plants with nitrogen fixation ability, and the symbiotic flora configuration is optimized by a soil microbial diversity index; based on the terrain undulation data and hydrological connectivity data of each region, the spatial layering width and transition zone width of the outer windproof layer, the middle water-locking layer and the inner soil improvement layer are determined, and the transition zone width is normalized calculated by the product of the standard deviation of the terrain undulation data and the watershed convergence index of the hydrological connectivity data; the construction of the soil desertification dynamic change matrix of each region based on the multi-dimensional ecological data time sequence comprises: performing principal component analysis on the multi-dimensional ecological data time sequence to extract dominant ecological factors of each region, wherein the dominant ecological factors include soil moisture change rate, surface vegetation coverage change rate and wind speed change rate; constructing an ecological factor time sequence vector of each region based on the dominant ecological factors, wherein the ecological factor time sequence vector is composed of time sequence values of each dominant ecological factor arranged in time sequence; calculate the change rate of the ecological factor time sequence vector of each region in adjacent time windows, and the change rate is characterized by the ratio of the Euclidean distance of adjacent time windows to the length of the time window. The change rates of the regions are arranged in time sequence and space position in matrix form to generate a soil desertification dynamic change matrix of the regions, wherein a row of the soil desertification dynamic change matrix represents a time dimension and a column represents a space dimension.

2. The method according to claim 1, wherein the method is characterized by, The region division of the target desert area according to the soil desertification degree distribution data comprises: obtaining soil particle composition data, soil moisture content data and soil organic matter content data of the target desert area; calculating a soil desertification index of each sampling point according to a proportion of fine particles with a particle size less than 0.05 mm in the soil particle composition data; performing weighted processing on the soil desertification index, wherein the weighted processing comprises taking the soil moisture content data as a negative correlation weight and taking the soil organic matter content data as a negative correlation weight to obtain a comprehensive desertification degree value of each sampling point; dividing the comprehensive desertification degree value into a slight desertification interval, a moderate desertification interval and a severe desertification interval based on a preset desertification degree threshold interval, which correspond to the slight desertification region, the moderate desertification region and the severe desertification region respectively; and performing smoothing processing on the divided region boundary to obtain a continuous region division result.

3. The method according to claim 1, wherein the method is characterized by, The multi-dimensional ecological data monitoring of the drought-tolerant plant community space configuration model of each region comprises: deploying soil moisture sensors, wind speed sensors and soil organic matter sensors in the outer circle windproof layer, the middle circle water locking layer and the inner circle soil improvement layer of each region respectively to obtain soil moisture time series data, wind speed time series data and soil organic matter time series data of each layer; obtaining surface vegetation coverage rate time series data and surface temperature time series data of each region based on unmanned aerial vehicle remote sensing technology; aligning the soil moisture time series data, the wind speed time series data, the soil organic matter time series data, the surface vegetation coverage rate time series data and the surface temperature time series data on a time axis to generate a multi-dimensional ecological data time series sequence of each region; performing outlier detection on the multi-dimensional ecological data time series sequence, wherein the outlier detection comprises calculating a local outlier factor of each time series data, marking data points with a local outlier factor greater than a preset outlier threshold as outliers and performing interpolation repair.

4. The method according to claim 3, wherein the method is characterized by, The unmanned aerial vehicle remote sensing technology comprises: using an unmanned aerial vehicle to carry a multispectral camera and a thermal imaging camera to obtain multispectral images and thermal imaging images of each region; performing normalized vegetation index calculation on the multispectral images to obtain surface vegetation coverage rate time series data of each region; performing temperature calibration on the thermal imaging images to obtain surface temperature time series data of each region; performing spatial registration on the surface vegetation coverage rate time series data and the surface temperature time series data, wherein the spatial registration comprises affine transformation based on terrain feature points and non-rigid registration based on vegetation edges.

5. The method according to claim 1, wherein the method is characterized by, The prediction of future desertification trends of each region based on the soil desertification dynamic change matrix comprises: performing time series decomposition on the soil desertification dynamic change matrix to obtain a trend component, a periodic component and a noise component; construct a desertification trend prediction model for each region based on the trend component, the desertification trend prediction model being trained using a long short-term memory network, the training input being historical data of the trend component, and the training output being a future predicted value of the trend component; perform Fourier transform on the periodic component to extract desertification periodic features of each region, the desertification periodic features including a periodic length and a periodic amplitude; superimpose the future predicted value of the desertification trend prediction model and the desertification periodic features to generate a desertification trend prediction sequence for each region; generate a regional desertification prediction result according to the desertification trend prediction sequence and a preset desertification degree threshold interval, the regional desertification prediction result including a change in desertification degree level of each region in a future time period.

6. The method according to claim 1, wherein the method is characterized by, The plant species distribution and spatial layout strategy of the drought-tolerant plant community spatial configuration model is dynamically adjusted according to the regional desertification prediction result, including: determining the regions that need to be adjusted and the adjustment priority according to the change in desertification degree level of each region in the regional desertification prediction result, the adjustment priority being normalized calculated by the product of the amplitude and the speed of the change in desertification degree level; for the regions with deteriorating desertification degree, increasing the planting density of the outer windbreak layer and the root density of the middle water-locking layer, and adjusting the proportion of nitrogen-fixing plants in the inner soil improvement layer; for the regions with improved desertification degree, reducing the planting density of the outer windbreak layer, increasing the diversity of plant species in the middle water-locking layer, and optimizing the configuration of symbiotic bacteria in the inner soil improvement layer; based on the adjusted plant species distribution and spatial layout strategy, recalculating the spatial layering width of the outer windbreak layer, the middle water-locking layer and the inner soil improvement layer, and the width of the transition zone to generate an updated drought-tolerant plant community spatial configuration model.

7. The method according to claim 1, wherein the method is characterized by, The principal component analysis is performed on the multi-dimensional ecological data time series to extract dominant ecological factors of each region, including: standardizing the multi-dimensional ecological data time series to generate a standardized time series matrix; calculating the covariance matrix of the standardized time series matrix to obtain eigenvalues and eigenvectors of the covariance matrix; sorting the eigenvalues according to their sizes, and selecting the principal components corresponding to the first several eigenvectors whose cumulative contribution rate is greater than a preset contribution rate threshold; linearly combining the principal components and the original ecological factors in the multi-dimensional ecological data time series to extract the dominant ecological factors, the number of the dominant ecological factors being determined by the number of the principal components.

Images