An automatic generation method of an invasive species prevention and control level map

Abstract

The application discloses an automatic generation method of an invasive species prevention and control grade map, comprising the following steps: acquiring suitable habitat data of a target species in a target area and spatial data of at least two human interference factors; respectively performing spatial analysis on the suitable habitat data and the data of each human interference factor, dividing a risk grade, and obtaining a plurality of single-factor risk grade maps; weighting and superimposing the plurality of single-factor risk grade maps, calculating a comprehensive risk distribution according to a comprehensive evaluation rule, and generating a prevention and control grade map through reclassification; the method can be encapsulated as a geographic processing model to realize one-key automatic generation. Through systematic integration of natural suitable habitats and human interference factor systems, the application solves the problems of the existing prevention and control division methods, such as disconnection with actual conditions, difficulty in quickly responding to multi-scenario analysis, strong subjectivity and low repeatability, and realizes high-precision fitting of the division results with actual conditions, as well as an objective, stable and repeatable division process.

Description

Technical Field

[0001] This invention belongs to the field of invasive species control levels, and specifically relates to an automated method for generating invasive species control level maps. Background Technology

[0002] Currently, generating suitable habitat maps for invasive species using species distribution models (such as Biomod2) is quite common. However, suitable habitats only represent the survival potential of a species, and the results cannot directly guide control actions targeting "spreading hotspots" and "core harm areas." Managers need to know "where to control" and "how strong the control measures should be." Existing technologies typically involve simple reclassification of suitable habitats or rely on qualitative superposition of a few factors based on human experience. There is a lack of a systematic and quantitative method to integrate multiple key spread and harm drivers (such as roads and farmland) with suitable habitats to directly output control maps with clear hierarchical significance.

[0003] One existing technology is a "simple classification method based on a single suitable habitat zone". This method directly reclassifies the predicted suitable habitat zone results (such as the suitability index) into high, medium, and low suitability zones using methods such as equal intervals or quantiles, and then treats these zones as control zones in a general sense. This method does not consider the actual dispersal pathways of species and the risk of localization hazards.

[0004] The drawback of existing technology one is that it only divides regions based on the survival potential of species under natural conditions, completely ignoring the decisive influence of key anthropogenic disturbance factors such as roads, logistics, farmland, human activities, and different land types on the spread speed and severity of invasive species. Its output of "suitable habitats" fails to answer the two core questions of prevention and control decisions: "Through what pathways and at what speed will the species spread?" and "Where is the greatest economic loss likely to occur?" This leads to a serious disconnect between the allocation of prevention and control resources and the actual spread risk and economic vulnerability.

[0005] The second existing technology, the "manual GIS overlay analysis method," requires technicians to manually overlay layers such as suitable areas, road buffer zones, and farmland distribution in GIS software. The prevention and control level is determined through visual interpretation or simple calculation, and this operation needs to be repeated when changing to a different area.

[0006] The drawback of the existing technology is that it heavily relies on the operator's personal experience, and its overlay rules and level thresholds are subjectively set, lacking unified and quantifiable standards. When dealing with different species or regions, the entire process must be restarted, making it impossible to form a standardized analysis process that can be replicated and compared. In addition, manual operation is difficult to efficiently handle simulation analysis under multiple scenarios (such as different climate scenarios and different planned roads), and cannot meet the needs of dynamic decision-making. Summary of the Invention

[0007] To address the aforementioned technical problems, this invention proposes an automated method for generating invasive species control level maps. This method solves the problems of control zone division being out of sync with actual conditions, difficulty in quickly responding to multi-scenario analysis, high subjectivity, different operators obtaining different division results, and low repeatability.

[0008] The technical solution of the present invention is as follows:

[0009] An automated method for generating invasive species control level maps includes the following steps:

[0010] Step S1: Prepare and acquire data on the suitable habitat of the target species in the target area, as well as spatial data on at least two anthropogenic disturbance factors;

[0011] Step S2: Perform independent spatial analysis on the suitable habitat data of the target species in the target area and the spatial data of at least two anthropogenic disturbance factors. Specifically, the suitable habitat data is classified into risk levels based on its survival potential, and the anthropogenic disturbance factors are classified into risk levels based on their intensity of promoting species dispersal or their sensitivity to potential harm, resulting in multiple single-factor risk level maps.

[0012] Step S3: Overlay the single-factor risk level maps and use a weighted overlay model to perform a comprehensive calculation to obtain the comprehensive risk distribution;

[0013] Step S4: Reclassify the raster layer of the calculated comprehensive risk distribution to generate the final prevention and control level map;

[0014] Step S5: To improve the ease of use of the method, steps S2 to S4 are constructed into an executable geoprocessing model. The geoprocessing model can receive suitable area data and human interference factor data as input, and automatically execute the entire spatial analysis and calculation process, and finally output a prevention and control level map.

[0015] Preferably, step S2 includes the following steps:

[0016] Step S21: Divide the suitability probability grid according to the threshold;

[0017] Step S22: Create multi-ring buffers for the road network;

[0018] Step S23: Extract land types related to agricultural production or intensive human activities from land use data.

[0019] Preferably, step S3 includes the following steps:

[0020] Step S31: Assign a weight Wi to the suitable area factor and each type of human interference factor, ensuring that the sum of the weights Wi is 1;

[0021] Step S32: Based on the biological diffusion characteristics, historical invasion pathways and potential economic and ecological harm of the target invasive species, the weight Wi of each factor is determined by expert scoring; among them, the total weight of human interference factors is set to be no less than 50% to highlight the dominant role of human activities in invasion risk.

[0022] Step S33: Based on the weight W i Calculate the overall risk value for each grid cell.

[0023] Preferably, step S5 includes the following steps:

[0024] Step S51: The user drags and drops the "Suitable Area Raster", "Road Vector", and "Land Use Vector" data to the model input port. The model automatically performs buffer analysis, reclassification, raster calculation, and final map output.

[0025] Step S52: Save the model generated in step S51 as a tool file for distribution to different users for repeated use, ensuring the standardization of the analysis process.

[0026] Preferably, the multiple single-factor risk level maps in step S2 are based on the suitable habitat data of the target species in the target area and the spatial data of at least two anthropogenic disturbance factors, as well as the ease of dispersal of the target species by each factor or the level of harm that it may cause in a specific area.

[0027] Preferably, the artificial interference factors in step S2 include roads, farmland, land use type, houses, and ports.

[0028] Compared with existing technologies, the advantages of the automated generation method for invasive species control level maps in this invention are as follows:

[0029] 1. Precision: By integrating suitable habitats with various actual diffusion and hazard factors, the delineated control zones are more precise and better reflect the actual invasion and disaster patterns of species.

[0030] 2. Direct Decision Support: The output "Prevention and Control Level Map" directly indicates the prevention and control intensity required for different regions, achieving a "seamless conversion" from scientific research data to management actions and forming a closed-loop decision-making process.

[0031] 3. Highly targeted prevention and control: The final map not only reflects the places where species can "survive", but also highlights the areas that are "easy to spread" and "cause great losses", achieving a qualitative change from "suitable habitat" to "prevention and control area".

[0032] 4. High degree of standardization and automation: The entire process, especially the factor ranking and fusion in the second stage, is encapsulated as an automated tool, ensuring the repeatability and efficiency of the results. Attached Figure Description

[0033] To more clearly illustrate the purpose, design concept, and innovation of the automated generation method for invasive species control level maps proposed in this invention, the invention will be described in detail below with reference to the accompanying drawings and tables.

[0034] Figure 1 This is a flowchart of the present invention.

[0035] Figure 2 This is the diagram for generating and outputting the prevention and control level map of the present invention. Detailed Implementation

[0036] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0037] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0038] An automated method for generating invasive species control level maps includes the following steps:

[0039] Step S1: Prepare and acquire data on the suitable habitat of the target species in the target area, as well as spatial data on at least two anthropogenic disturbance factors;

[0040] Step S2: Perform independent spatial analysis on the suitable habitat data of the target species in the target area and the spatial data of at least two anthropogenic disturbance factors. Specifically, the suitable habitat data is classified into risk levels based on its survival potential, and the anthropogenic disturbance factors are classified into risk levels based on their intensity of promoting species dispersal or their sensitivity to potential harm, resulting in multiple single-factor risk level maps.

[0041] Step S3: Overlay the single-factor risk level maps and use a weighted overlay model to perform a comprehensive calculation to obtain the comprehensive risk distribution;

[0042] Step S4: Reclassify the raster layer of the calculated comprehensive risk distribution to generate the final prevention and control level map;

[0043] Step S5: To improve the ease of use of the method, steps S2 to S4 are constructed into an executable geographic processing model. The geographic model can receive the suitable area data and human interference factor data as input, and automatically execute the entire spatial analysis and calculation process, and finally output the prevention and control level map.

[0044] Step S2 of this implementation plan includes the following steps:

[0045] Step S21: Divide the suitability probability grid according to the threshold;

[0046] Step S22: Create multi-ring buffers for the road network;

[0047] Step S23: Extract land types related to agricultural production or intensive human activities from land use data.

[0048] Step S3 of this implementation plan includes the following steps:

[0049] Step S31: Assign a weight Wi to the suitable area factor and each type of human interference factor, ensuring that the sum of the weights Wi is 1;

[0050] Step S32: Based on the biological diffusion characteristics, historical invasion pathways and potential economic and ecological harm of the target invasive species, the weight Wi of each factor is determined by expert scoring; among them, the total weight of human interference factors is set to be no less than 50% to highlight the dominant role of human activities in invasion risk.

[0051] Step S33: Based on the weight W i Calculate the overall risk value for each grid cell.

[0052] Step S5 of this implementation plan includes the following steps:

[0053] Step S51: The user drags and drops the "Suitable Area Raster", "Road Vector", and "Land Use Vector" data to the model input port. The model automatically performs buffer analysis, reclassification, raster calculation, and final map output.

[0054] Step S52: Save the model generated in step S51 as a tool file for distribution to different users for repeated use, ensuring the standardization of the analysis process.

[0055] The multiple single-factor risk level maps in step S2 of this implementation plan are based on the data of the target species’ suitable habitat in the target area and the spatial data of at least two anthropogenic disturbance factors, as well as the ease of dispersal of the target species by each factor or the level of harm that it may cause in a specific area.

[0056] Step S2 of this implementation plan involves selecting at least two artificial interference factors from a sub-database consisting of roads, farmland, ports, land use types, logistics hubs, and settlements, based on the transmission vectors and targets of the target species.

[0057] The core of this implementation plan lies in its adoption of a two-stage, factor-based decision-making framework: In the first stage, a "suitable habitat map" representing the intrinsic survival potential of species is generated by integrating only natural climatic environmental factors using species distribution models (such as Biomod2 and MaxEnt). In the second stage, based on this "suitable habitat map," various anthropogenic disturbance factors, such as roads, farmland, and ports, are specifically integrated to conduct risk assessment and classification. This design is based on the fact that natural factors are relatively stable, while anthropogenic factors change rapidly. Processing them separately allows for the reuse of stable suitable habitat results while flexibly adjusting to respond to rapidly changing human activity scenarios, making prevention and control decisions more timely and targeted.

[0058] When this implementation plan is implemented,

[0059] This invention aims to address how to systematically process different types of influencing factors in stages. First, it integrates natural factors in a species distribution model to assess "survival potential." Then, it incorporates human-mediated diffusion factors in a prevention and control decision-making model to assess "actual risk." Finally, through a clearly defined process, it generates a prevention and control level map that accurately reflects the overall invasion risk. Furthermore, natural influencing factors are relatively stable, and the predicted and assessed suitable habitat maps for invasive species can be reused repeatedly. Human-mediated diffusion factors, on the other hand, change rapidly and are more unstable, allowing for timely adjustments to obtain a more real-time and effective prevention and control decision-making map. Separating the two types of factors also reduces operational steps.

[0060] This implementation scheme further describes the construction of a two-stage, factor-classification decision-making framework. Its core lies in explicitly classifying influencing factors into two categories, using them separately in two stages, and finally merging them. The technical process is as follows: Figure 1 As shown,

[0061] 1. Survival potential modeling based on natural factors

[0062] Using species distribution models (such as biomod2), natural factors selected by random forests are input for training and prediction to obtain a map of suitable areas representing survival potential under natural conditions. At this stage, only natural factors are considered.

[0063] 2. Actual risk assessment of integrating humanistic diffusion factors

[0064] This stage uses the suitable habitat distribution map obtained in the first stage and various human diffusion factors as input to conduct a risk assessment and comprehensively generate a prevention and control level map. Spatial analysis is performed on each input human impact factor, and risk levels are independently assigned.

[0065] Classification of suitable habitat areas by prevention and control: Using species distribution models (such as biomod2) to predict the distribution map of suitable habitat areas under future climate scenarios, suitable habitat areas are divided into three levels, such as low, medium and high.

[0066] Human factor classification: Each human diffusion factor (such as road buffer zone, farmland distribution area, etc.) is analyzed independently and classified into corresponding risk levels according to its intensity of promoting diffusion or causing harm.

[0067] 1) Road Buffer Factor: This factor comprehensively considers both road class and distance attenuation. Different base diffusion risk weights are assigned to roads of different classes (e.g., national highways, provincial highways, rural roads) (higher weights indicate stronger propagation). Based on this, distance attenuation buffer zones (e.g., 0-1km, 1-5km, >5km) are generated for each road; the closer the distance, the higher the risk attenuation coefficient. Finally, the comprehensive road diffusion risk level for each spatial location is calculated by multiplying (or weighting) the weights and attenuation coefficients.

[0068] 2) Farmland distribution factor: Based on land use type, farmland is assigned a high-risk level, while other land types are assigned a medium or low-risk level. The classification can also be based on the types of crops grown on the farmland, as the impact of the same invasive species varies among different crops, thus allowing for further classification.

[0069] 3) Other diffusion influencing factors: such as rivers, ports, and settlements, are spatially and hierarchically classified according to their contribution to species diffusion.

[0070] 3. Multi-factor hierarchical spatial overlay and comprehensive calculation

[0071] Finally, a spatial overlay analysis process is used to integrate all the single-factor layers that have been classified into different levels, generating a comprehensive invasive species control decision map. To ensure the versatility and convenience of this method, we have solidified this process into a model tool. This tool receives "target area suitable habitat map" and "series of influencing factor data" as input on one end, and automatically outputs a complete "control level map" on the other end, forming an automated method for "one-click generation" of precise control plans for specific regions and species.

[0072] Specific implementation methods,

[0073] Step 1: Data Preparation and Input Input data includes two parts:

[0074] 1) Data on the suitable habitat of the target species in the target area. This data is usually pre-generated by species distribution models such as MaxEnt and Biomod2, and is a raster layer, where each cell value is the survival suitability probability (0-1) of the species.

[0075] 2) Spatial data with at least two artificial interference factors.

[0076] For example: Road data: vector line layer, containing road classification attributes (e.g., expressway, national highway, provincial highway, county road). Land use data: vector polygon or raster layer, containing land type attributes (e.g., farmland, orchard, residential land, forest land, etc.). Crop distribution or economic value data can be further integrated. Other factors: point or polygon vector data such as ports, airports, logistics parks, scenic areas, etc. Step Two: Single-Factor Risk Hierarchy

[0077] This step processes each input layer independently, transforming it into a comparable risk level map (e.g., mapping the value range to 1, 2, 3, representing low, medium, and high risk, respectively). Habitat grading: The habitability probability grid is divided according to a threshold. For example, a probability <0.3 indicates low potential (1), 0.3-0.6 indicates medium potential (2), and >0.6 indicates high potential (3). Road factor grading (see...) Figure 2 ( ) Create multi-ring buffer zones for the road network. For example, create buffer zones of 0-1 km (high-risk zone, level 3) and 1-3 km (medium-risk zone, level 2) for national highways; and create buffer zones of 0-0.5 km (high-risk zone, level 3) and 0.5-2 km (medium-risk zone, level 2) for provincial highways. The area outside the buffer zones is a low-risk zone (level 1). Different buffer distances can be used for different road levels to reflect the differences in the dispersal capacity of invasive species.

[0078] Farmland factor grading: Extract “cultivated land” and “orchard” from land use data. Further subdivision is possible, such as assigning the highest risk level to high-economic-value crop areas such as “vegetable bases” and “orchards” (3), assigning a medium risk level to ordinary field crop areas (2), and assigning a low risk level to other non-agricultural land (1). At the same time, invasive species can assign a high risk level to easily invasive crops and a low risk level to species that are not easily invasive.

[0079] Step 3: Multi-factor spatial superposition and comprehensive calculation

[0080] Overlay all single-factor risk level raster layers generated in Step 2 (spatial reference and resolution must be unified). Use a weighted overlay model for comprehensive calculation. Assign a weight Wi to each factor, with the sum of weights being 1. Weights can be determined through expert scoring (e.g., suitable area potential weight 0.4, road diffusion weight 0.35, farmland hazard weight 0.25). For each spatial raster unit, perform the following calculation: Comprehensive Risk Value = W Suitable * R Suitable + W Road * R Road + W Farmland * R Farmland + … where R Suitable, R Road, and R Farmland represent the quantified risk level value of the unit in the corresponding factor layer. Step 4: Generation and Output of Prevention and Control Level Map Figure 2As shown, the calculated "Comprehensive Risk Value" raster layer is reclassified to generate the final prevention and control level map. For example: Comprehensive Risk Value ≥ 2.5: Classified as "Level 1 Prevention and Control Zone" (red), requiring emergency monitoring and eradication measures.

[0081] 1.5 ≤ Overall Risk Value < 2.5: Classified as "Level II Prevention and Control Zone" (Orange), requiring strengthened patrols and prevention and control. Overall Risk Value < 1.5: Classified as "Level III Prevention and Control Zone / General Monitoring Zone" (Yellow), requiring routine monitoring.

[0082] The classification results are then used to enhance the map (adding legends, scale bars, compasses, etc.) and output as images or interactive electronic maps.

[0083] Step 5: Automated tool packaging

[0084] To improve the ease of use of the method, steps two through four above can be constructed into a geoprocessing model in ArcGIS Pro's Model Builder. Users only need to drag and drop data such as "suitable area raster," "road vector," and "land use vector" into the model input port, and the model can automatically perform a series of operations such as buffer analysis, reclassification, raster calculation, and final map output, achieving "one-click generation." This model can be saved as a tool file for easy distribution to different users for repeated use, ensuring the standardization of the analysis process.

[0085] Case 1: Control of Canada Ipomoea – Shanghai as an Example

[0086] 1. Target species

[0087] The Canadian single-stem is a perennial herbaceous plant of the Asteraceae family native to North America. It is a noxious weed that spreads rapidly through its rhizomes and seeds. Due to its strong reproductive capacity and invasiveness, it has become a key invasive alien species under management in China.

[0088] 2. Data Preparation

[0089] 1) Suitable habitat data

[0090] Using the MaxEnt model, 300 occurrence sites of this species in the Yangtze River Delta region (sourced from the China Digital Herbarium) and 1km resolution climate data (annual average temperature, precipitation, etc.) of Shanghai were input to generate a local adaptability probability raster map.

[0091] 2) Humanistic diffusion factors

[0092] Road data: Obtain the vector map layer of roads in Shanghai, including "expressways", "provincial roads", "county roads", "township roads", etc.

[0093] Land use data: Obtain the current land use map of Shanghai. Based on the growth habits of the Canadian variegated plant, we need to focus on relatively open land types such as "vacant land", "parks and green spaces", and "industrial land".

[0094] Human interference points: Obtain point data for "nursery" and "logistics distribution point" and generate point vector map.

[0095] 3. Single-factor risk grading

[0096] 1) Hierarchical classification of suitable habitats: Using reclassification, areas with a probability value < 0.25 are assigned to level 1 (low potential); areas with a probability value < 0.25 are assigned to level 2 (medium potential); and areas with a probability value > 0.6 are assigned to level 3 (high potential).

[0097] 2) Road factor grading: Using buffer zone analysis tools, grading rules are formulated based on the correlation between road grade and diffusion capacity: For expressways and main roads, a 100-meter core buffer zone is generated on both sides and designated as a high-risk corridor (level 3); for secondary roads, a 60-meter buffer zone is generated and designated as a medium-risk corridor (level 2); all areas outside the buffer zones are considered low-risk (level 1).

[0098] Subsequently, through fusion and raster-to-area conversion, the aforementioned vector buffer is transformed into spatially continuous and standardized raster data.

[0099] 3) Land use / disturbance factor grading: The land use / disturbance factor grading process focuses on identifying high-risk habitat types that are subject to strong human disturbance and are prone to the colonization of alien species.

[0100] Using selection tools, land use data was used to extract two types of land: "vacant land" and "construction sites," which are characterized by strong human disturbance, low vegetation cover, and susceptibility to invasive alien species. These were directly designated as high-risk colonization areas (Level 3). At the same time, sensitive areas such as "parks and green spaces" and "nurseries," which may introduce and spread invasive species through human activities, were designated as medium-risk colonization areas (Level 2) by establishing a 100-meter buffer zone around them. All other land types were classified as low-risk areas (Level 1).

[0101] Subsequently, spatially continuous raster data is generated through fusion and surface-to-raster conversion.

[0102] 4. Risk Integration and Mapping

[0103] 1) Weighting: Given that the spread of Canadian goldenrod in Shanghai primarily relies on road traffic and human-caused land disturbance (such as construction sites and vacant land) for propagation and establishment, its harm to traditional agriculture is relatively low. Therefore, the following weights were determined through expert consultation: road dispersal factor (W2=0.4), land use disturbance factor (W3=0.35), and suitability potential factor (W1=0.25). This weighting allocation reflects a targeted response to the species' "human-driven dispersal" characteristic.

[0104] 2) Comprehensive Calculation: Calculate the comprehensive risk value using the raster calculator. Comprehensive Risk Value = 0.3 * Suitable Area Map + 0.4 * Road Buffer Zone Map + 0.3 * Land Use Disturbance Map, ultimately outputting a comprehensive risk map.

[0105] 5. Generation of Prevention and Control Level Map

[0106] Based on a comprehensive risk grid obtained through multi-factor weighted overlay, a reclassification tool is used to classify key decisions according to preset scientific thresholds: areas with a comprehensive risk value ≥ 2.4 are designated as Level I prevention and control zones (red), representing urgent actions requiring immediate cleanup; areas with risk values ​​between 1.8 and 2.4 are designated as Level II prevention and control zones (orange), requiring intensive monitoring; and areas with risk values ​​< 1.8 are designated as Level III monitoring zones (yellow), requiring routine patrols. Ultimately, an automated decision-making map, "Canada Goldenrod Prevention and Control Level Map," is generated that can directly guide on-site work.

[0107] Case Study 2: Assessing the Risk of Red Imported Fire Ant Spread – Shenzhen as an Example

[0108] 1. Target species

[0109] Red imported fire ants, a type of hymenopteran insect native to South America, are recognized by the International Union for Conservation of Nature (IUCN) as one of the 100 most destructive invasive species globally. Their spread occurs primarily through two routes: natural dispersal via the nuptial flights of reproductive ants; and long-distance dispersal of ant nests through human activities (such as the transport of seedlings, turf, and goods with soil). Red imported fire ants attack humans and livestock, damage power facilities, and disrupt agricultural and forestry production, demonstrating their extreme invasiveness, harmfulness, and quarantine importance.

[0110] 2. Data Preparation

[0111] 1) Suitable habitat data

[0112] Based on global bioclimatic data and historical occurrence data of red imported fire ants in South China (Guangdong, Guangxi, and Fujian), the Biomod2 ensemble model platform was used to select random forest and generalized additive model for ensemble prediction, generating a 1-kilometer resolution raster map of the suitability probability of red imported fire ants in the entire Shenzhen area. Each pixel value (0-1) represents the potential colonization probability under the natural climatic conditions of that location.

[0113] 2) Humanistic diffusion factors

[0114] Logistics network and key node data:

[0115] Key ports: Obtain precise boundary vector data for Yantian Port Area and Shenzhen Bao'an International Airport.

[0116] Logistics channels: Obtain GPS trajectory data of container trucks on key routes such as "Yantian Port-Pinghu Logistics Base" and "Yantian Port-Qianhai Free Trade Zone" over the past year. Generate freight heat maps through kernel density analysis to identify high-frequency, stable inland diffusion corridors.

[0117] Secondary nodes: Obtain vector data of locations in large logistics parks, seedling trading markets, and construction sites throughout the city.

[0118] Sensitive habitat data:

[0119] Obtain boundary vector data for all city parks, community green spaces, and golf courses; obtain distribution data for major agricultural areas and nursery bases.

[0120] 3. Single-factor risk grading

[0121] 1) Hierarchical classification of suitable habitats: Using a reclassification tool, the suitability probability grid is divided into three potential levels: a probability value < 0.3 is a low potential area (level 1); a probability value ≤ 0.3 is a medium potential area (level 2); and a probability value ≥ 0.7 is a high potential area (level 3).

[0122] 2) Hierarchical classification of logistics diffusion factors:

[0123] Extremely high risk source delineation: Buffer analysis was performed on the vector data of airport and port boundaries, and the area extending 3 kilometers outward from the boundaries of Yantian Port Area and Bao'an International Airport was designated as an extremely high risk source area (Level 4).

[0124] High-risk corridor designation: Based on the freight GPS heat map, the top 20% of main roads with the highest kernel density values ​​are buffered 1 km to both sides and designated as high-risk diffusion corridors (level 3).

[0125] Medium-risk channels are designated as medium-risk diffusion channels (Level 2), with a 500-meter buffer zone on each side of other major roads connecting core ports and logistics nodes.

[0126] All areas outside the buffer zone are designated as low-risk zones (Level 1).

[0127] Finally, a unified raster map of logistics diffusion risk is generated by fusion and surface-to-raster conversion tools.

[0128] 3) Hierarchy of sensitive habitat factors:

[0129] Large green spaces with frequent human interference and high ecological value, such as "city parks" and "golf courses", are directly designated as high-hazard-value areas (level 3).

[0130] "Community green space", "agricultural area" and "nursery" are designated as medium hazard value areas (level 2).

[0131] The remaining construction land, water areas, etc., are designated as low hazard value zones (Level 1), and a sensitive habitat risk raster map is generated.

[0132] 4. Risk Integration and Mapping

[0133] 1) Weighting: In port cities like Shenzhen, the invasion risk of red imported fire ants exhibits a dual characteristic: the risk of long-distance introduction mediated by international logistics and the risk of secondary diffusion driven by domestic logistics. Given that the climate conditions in this region fully meet the survival threshold of red imported fire ants, experts used analytic hierarchy process (AHP) to set weights: logistics diffusion factor (W1=0.5), survival potential factor (W2=0.3), and sensitive habitat factor (W3=0.2). This weighting significantly strengthens the control weight of the dominant risk factor, 'human logistics activities'.

[0134] 2) Comprehensive Calculation: Using a raster calculator, a weighted overlay analysis was performed, i.e., the comprehensive risk value = 0.5 * logistics diffusion risk raster map + 0.3 * suitable habitat risk map + 0.2 * sensitive habitat risk raster map. The comprehensive risk raster map of red imported fire ant invasion in Shenzhen was obtained.

[0135] 5. Generation and application verification of prevention and control level maps

[0136] 1) Generation of Prevention and Control Level Map: The comprehensive risk grid is reclassified. Based on the frequency distribution of risk values ​​and expert experience, thresholds are set: areas with a risk value ≥ 2.7 are designated as Level I emergency prevention and control zones (red); risk values ​​between 2.0 and 2.7 are designated as Level II key monitoring zones (orange); and risk values ​​< 2.0 are designated as Level III routine monitoring zones (yellow). The final automated output is the "Shenzhen Red Imported Fire Ant Invasion Risk Prevention and Control Level Map".

Claims

1. A method for automatically generating an invasive species control level map, characterized in that, Includes the following steps: Step S1: Prepare and acquire data on the suitable habitat of the target species in the target area, as well as spatial data on at least two anthropogenic disturbance factors; Step S2: Perform independent spatial analysis on the suitable habitat data of the target species in the target area and the spatial data of at least two anthropogenic disturbance factors, and divide each data into multiple risk levels to obtain multiple single-factor risk level maps; Step S3: Overlay the single-factor risk level maps and use a weighted overlay model to perform a comprehensive calculation to obtain the comprehensive risk distribution; Step S4: Reclassify the raster layer of the calculated comprehensive risk distribution to generate the final prevention and control level map; Step S5: To improve the ease of use of the method, steps S2 to S4 are constructed into an executable geoprocessing model. The geoprocessing model can receive suitable area data and human interference factor data as input, and automatically execute the entire spatial analysis and calculation process, and finally output a prevention and control level map.

2. The method for automatically generating an invasive species control level map according to claim 1, characterized in that, Step S2 includes the following steps: Step S21: Divide the suitability probability grid according to the threshold; Step S22: Create multi-ring buffers for the road network; Step S23: Extract land types related to agricultural production or intensive human activities from land use data.

3. The method for automatically generating an invasive species control level map according to claim 1, characterized in that, Step S3 includes the following steps: Step S31: Assign a weight Wi to the suitable area factor and each type of human interference factor, ensuring that the sum of the weights Wi is 1; Step S32: Based on the biological diffusion characteristics, historical invasion pathways and potential economic and ecological harm of the target invasive species, the weight Wi of each factor is determined by expert scoring; among them, the total weight of human interference factors is set to be no less than 50% to highlight the dominant role of human activities in invasion risk. Step S33: Based on the weight W i Calculate the overall risk value for each grid cell.

4. The method for automatically generating an invasive species control level map according to claim 1, characterized in that, Step S5 includes the following steps: Step S51: The user drags and drops the "Suitable Area Raster", "Road Vector", and "Land Use Vector" data to the model input port. The model automatically performs buffer analysis, reclassification, raster calculation, and final map output. Step S52: Save the model generated in step S51 as a tool file for distribution to different users for repeated use, ensuring the standardization of the analysis process.

5. The method for automatically generating an invasive species control level map according to claim 1, characterized in that, The multiple single-factor risk level maps in step S2 are based on the target species’ suitable habitat data in the target area and spatial data of at least two anthropogenic disturbance factors, as well as the ease of spread of the target species by each factor or the level of harm that it may cause in a specific area.

6. The method for automatically generating an invasive species control level map according to claim 1, characterized in that, The artificial interference factors in step S2 are selected from a database of land factors, including roads, farmland, ports, land use types, logistics hubs, and settlements, based on the transmission vectors and harmful objects of the target species.

Images