A habitat suitability assessment method for Yemeni desert locust based on MaxEnt and space-time cube

Abstract

The present invention relates to a method for assessing the habitat suitability of the Yemeni desert locust based on MaxEnt and a space-time cube. The method comprises the following steps: acquiring multi-source data; preprocessing the multi-source data; screening habitat factors; obtaining a MaxEnt model with optimal parameters; training the MaxEnt model with optimal parameters to obtain inter-annual habitat suitability results for the Yemeni desert locust; constructing a space-time cube; performing a space-time integrated analysis; and grading the suitability of the Yemeni desert locust based on the results of the space-time integrated analysis. The method comprehensively evaluates ecological niche conditions, with the MaxEnt model facilitating comprehensive consideration of the niche conditions of the Yemeni desert locust; predicts potential suitable areas, taking into account spatiotemporal changes, with the space-time cube method allowing for tracking changes in desert locust habitat suitability in time and space; and provides early warning. The niche model and the space-time cube method can provide early warning of the potential emergence of new habitats for the desert locust, allowing preventive measures to be taken to mitigate its impact on agriculture and ecosystems.

Description

Technical Field

[0001] The present invention relates to the technical field of niche modeling and multi-temporal analysis, in particular to a method for assessing the habitat suitability of Yemeni desert locusts based on MaxEnt and space-time cubes. Background Art

[0002] The Desert Locust (Cyrtacanthacridinae), a genus of the family Catantopidae, superfamily Acrididae, is widely distributed across Africa, West Asia, and parts of South Asia. It is the world's most destructive migratory pest. A mature Desert Locust can consume its own weight in crops in a single day. A swarm can reach a maximum size of approximately 800 square kilometers, containing nearly 40 billion locusts, whose daily food intake could feed 400,000 people annually. Desert Locust outbreaks can easily cause a global food crisis, seriously threatening ecological security, regional stability, and human well-being. Before a locust outbreak occurs, it is crucial to conduct large-scale analysis of locust populations and their habitats, to guide timely and effective population control measures, keep population density within safe thresholds, and prevent the formation of swarms.

[0003] Currently, analysis of locust populations and their habitats is primarily based on technical means such as ground surveys based on expert experience and biological models that incorporate expert knowledge. This typically requires a high level of professional expertise, is costly, and suffers from poor representativeness and timeliness. With the advancement of remote sensing technology, it is increasingly being incorporated into locust management activities. Currently, the combination of remote sensing and ecological niche models is widely used to analyze locust populations and their habitats, but long-term, spatially and temporally integrated analysis is lacking.

[0004] Therefore, how to use niche models and spatiotemporal integration analysis technology to realize the habitat suitability assessment of the Yemen Desert Locust is a technical problem that needs to be urgently solved in the control of the Yemen Desert Locust. Summary of the Invention

[0005] To address the problems of crop losses and environmental damage caused by desert locusts, the present invention aims to provide a Yemen Desert Locust habitat suitability assessment method based on MaxEnt and space-time cubes, which can capture the temporal and spatial variation trends of desert locust habitat suitability and effectively assess the habitat suitability of desert locusts in Yemen.

[0006] To achieve the above object, the present invention adopts the following technical solution: a method for assessing the habitat suitability of Yemeni desert locust based on MaxEnt and space-time cube, the method comprising the following steps in order:

[0007] (1) Acquisition of multi-source data: Acquisition of meteorological data and satellite data, including normalized difference vegetation index, surface temperature, precipitation, annual average effective accumulated temperature, digital elevation data, soil moisture, and land cover classification data; acquisition of desert locust ground survey data and soil sand content;

[0008] (2) Preprocessing of multi-source data: resampling and temporal aggregation of multi-source data to ensure that data from different sources have consistent spatial range, spatial resolution, and temporal resolution; using land cover classification to mask permanent water bodies, forests, and buildings; and cleaning desert locust ground survey data;

[0009] (3) Screening of habitat factors: Calculate the Pearson coefficient between multi-source data and retain the multi-source data with a Pearson coefficient less than 0.8 as habitat factors;

[0010] (4) Constructing the MaxEnt model and parameter tuning: Obtaining the MaxEnt model with the optimal parameters, training the MaxEnt model with the optimal parameters, and obtaining the interannual habitat suitability results for the Yemeni desert locust;

[0011] (5) Constructing a space-time cube: adding time attributes to the inter-annual habitat suitability results of the Yemeni desert locust, constructing a multidimensional grid, and creating a space-time cube based on the inter-annual habitat suitability results of the Yemeni desert locust;

[0012] (6) Conducting space-time integration analysis: Conducting space-time integration analysis on the space-time cube to obtain space-time integration analysis results;

[0013] (7) Assessment of habitat suitability of desert locusts in Yemen: Based on the results of spatiotemporal integration analysis, the suitability of desert locusts in Yemen is classified into different levels.

[0014] The step (1) specifically includes the following steps:

[0015] (1a) Obtain the normalized vegetation index, which is as follows:

[0016] NDVI = (NIR - Red) / (NIR + Red)

[0017] Among them, NIR represents the reflectivity of the near-infrared band, and Red represents the reflectivity of the red band;

[0018] (1b) Obtain the surface temperature using the following formula:

[0019] LST=(day+night)×0.5×0.02-273.15

[0020] Wherein, day and night are the observed values during the day and night, respectively, and the daily LST is the average of the day and night;

[0021] (1c) Obtain precipitation: directly obtain it through the precipitation band in the CHIRPS dataset;

[0022] (1d) Obtain the annual average effective accumulated temperature:

[0023] ET yearly =DD year / total number of days in a year

[0024] Among them, DD year The effective accumulated temperature of desert locusts is obtained through the Day Degree model (degree-day model) in one year, and is obtained through the following formula:

[0025]

[0026] Among them, n is the total number of days in a year, T i (t) is the surface temperature at hour t on day i of the year. LT = 15.5°C is the theoretical starting temperature for desert locust egg development. UT = 41°C is the maximum threshold temperature for egg hatching. When the temperature is lower than LT or higher than UT, locust eggs stop developing. Represents the DD value of the i-th day in a year. The DD value refers to the effective accumulated temperature of the Day Degree model;

[0027] (1e) Obtain digital elevation data: Digital elevation data are directly obtained from the 30m spatial resolution NASA DEM dataset;

[0028] (1f) Obtain soil moisture: Obtain soil moisture at 7-28 cm from ERA5-Land monthly data;

[0029] (1g) Obtaining land cover classification data: Land cover classification data are directly obtained from the MODIS annual land classification dataset MCD12Q1;

[0030] (1h) Obtaining ground survey point data: Obtaining desert locust ground survey data;

[0031] (1i) Obtain soil sediment content: Obtain soil sediment content from the SoilGrids250m 2.0 dataset.

[0032] The step (2) specifically includes the following steps:

[0033] (2a) Unify the temporal resolution of multi-source data to interannual: perform annual maximum aggregation on the NDVI, annual mean aggregation on soil moisture, and annual sum aggregation on precipitation;

[0034] (2b) Unify the spatial resolution of multi-source data: resample the multi-source data to 5 km;

[0035] (2c) Masking permanent water bodies, forests, and buildings: Use land cover classification data as mask input to perform masking operations on multi-source data to exclude permanent water bodies, forests, and building locations;

[0036] (2d) Unified spatial scope: The spatial scope of multi-source data was unified and limited to Yemen;

[0037] (2e) Cleaning redundant ground survey data: The ground survey data were extracted by year, and then the data of different years were cleaned using one of the multi-source data grids as a background reference to ensure that there was at most one ground survey point in each grid. The cleaned ground survey data were saved as a CSV file for future use.

[0038] The step (4) specifically includes the following steps:

[0039] (4a) Construction of MaxEnt model: Constructing the MaxEnt niche model of the Yemeni desert locust;

[0040] (4b) Optimizing MaxEnt parameters: Automatically optimizing the characteristic classes and regularization multipliers of the MaxEnt ecological niche model for the Yemeni desert locust, selecting the model with the best parameters, i.e., the MaxEnt model with the best parameters, and recording the parameters of the MaxEnt model with the best parameters;

[0041] (4c) Model training: Input the selected habitat factors and cleaned ground survey data into the MaxEnt model with optimal parameters to complete the model training; repeat the training 25 times to obtain 25 MaxEnt models with optimal parameters;

[0042] (4d) Obtaining the inter-annual habitat suitability results for the Yemeni desert locust: Based on the 25 optimal MaxEnt models, the area under the receiver operating characteristic curve (AUC) was used as the threshold. The models with an area under the receiver operating characteristic curve (AUC) greater than 0.7 among the 25 optimal MaxEnt models were fused to obtain the final MaxEnt model. The habitat factors corresponding to the final MaxEnt model year were input into the final MaxEnt model to obtain the inter-annual habitat suitability results for the Yemeni desert locust.

[0043] The step (5) specifically includes the following steps:

[0044] (5a) Create a blank mosaic dataset;

[0045] (5b) Add the interannual habitat suitability results of the Yemeni desert locust to the blank mosaic dataset, add time attributes to it, construct multidimensional information, and thus obtain multidimensional raster data;

[0046] (5c) Create a space-time cube using multidimensional raster data.

[0047] The step (6) specifically includes the following steps:

[0048] (6a) Emerging hotspot analysis: Emerging hotspot analysis was performed using a space-time cube to obtain the changing trends of habitat suitability in Yemen;

[0049] (6b) Time series cluster analysis: Time series cluster analysis was performed using a space-time cube to obtain a time-space integration analysis result, which refers to the spatial and temporal clustering results of the Yemen Desert Locust suitability values.

[0050] It can be seen from the above technical solution that the beneficial effects of the present invention are: first, comprehensive evaluation of niche conditions. The MaxEnt model helps to comprehensively consider the niche conditions of the Yemeni desert locust, including factors such as temperature, soil moisture, and vegetation; second, prediction of potential suitable areas. The MaxEnt model can predict potential suitable areas that may exist in Yemen based on known niche condition data; third, considering spatiotemporal changes. The space-time cube method allows tracking changes in the suitability of desert locust habitats in time and space; fourth, early warning. Through the niche model and the space-time cube method, early warning of possible new habitats of desert locusts can be provided, so that preventive measures can be taken to reduce their impact on agriculture and ecosystems. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 is a flow chart of the method of the present invention;

[0052] Figure 2 This is a diagram showing the suitability of the Yemen Desert Locust obtained using the MaxEnt model of the present invention;

[0053] Figure 3 This is a graph showing the changing trends of habitat suitability in Yemen;

[0054] Figure 4 It is the cluster mean curve in the time series cluster analysis of the present invention;

[0055] Figure 5 This is a diagram for grading the suitability of the Yemen Desert Locust in the present invention;

[0056] Figure 6 The figure shows the distribution map of the known core breeding area of desert locust and the ground survey points described in the present invention. DETAILED DESCRIPTION

[0057] like Figure 1 As shown in FIG, a method for assessing habitat suitability of the Yemeni Desert Locust based on MaxEnt and space-time cubes comprises the following steps in sequence:

[0058] (1) Acquisition of multi-source data: Using the Google Earth Engine platform (Gee) to acquire meteorological data and satellite data, including normalized difference vegetation index, surface temperature, precipitation, annual average effective accumulated temperature, digital elevation data, soil moisture, and land cover classification data; obtaining desert locust ground survey data from the Food and Agriculture Organization of the United Nations, and obtaining soil sand content from the International Soil Reference and Information Center;

[0059] (2) Preprocessing of multi-source data: resampling and temporal aggregation of multi-source data to ensure that data from different sources have consistent spatial range, spatial resolution, and temporal resolution; using land cover classification to mask permanent water bodies, forests, and buildings; and cleaning desert locust ground survey data;

[0060] (3) Screening of habitat factors: Calculate the Pearson coefficient between multi-source data and retain the multi-source data with a Pearson coefficient less than 0.8 as habitat factors;

[0061] (4) Constructing the MaxEnt model and parameter tuning: Obtaining the MaxEnt model with the optimal parameters, training the MaxEnt model with the optimal parameters, and obtaining the interannual habitat suitability results for the Yemeni desert locust;

[0062] (5) Constructing a space-time cube: adding time attributes to the inter-annual habitat suitability results of the Yemeni desert locust, constructing a multidimensional grid, and creating a space-time cube based on the inter-annual habitat suitability results of the Yemeni desert locust;

[0063] (6) Conducting space-time integration analysis: Conducting space-time integration analysis on the space-time cube to obtain space-time integration analysis results;

[0064] (7) Assessment of desert locust habitat suitability in Yemen: Based on the results of spatiotemporal integration analysis, the suitability of desert locusts in Yemen is classified into different levels, such as Figure 5 As shown, Figure 5 The distribution of habitat suitability of desert locusts in Yemen after the suitability level classification is shown in Figure 2. This distribution is consistent with the known distribution of desert locust breeding areas in Yemen, such as Figure 6 shown.

[0065] The step (1) specifically includes the following steps:

[0066] (1a) Obtain the normalized vegetation index. The normalized vegetation index is calculated using the red and near-infrared bands of the MODIS surface reflectance product MOD09A1. The formula is as follows:

[0067] NDVI = (NIR - Red) / (NIR + Red)

[0068] Among them, NIR represents the reflectance of the near-infrared band, and Red represents the reflectance of the red band, which correspond to band 1 (sur_refl_b01) and band 2 (sur_refl_b02) of MOD09A1 respectively.

[0069] (1b) Obtain the surface temperature using the following formula:

[0070] LST=(day+night)×0.5×0.02-273.15

[0071] Where day and night are the observed values during the day and night, respectively, and the daily LST is the average of the day and night values; 0.02 is the scaling factor for the MOD11A1 band, and 273.15 is the offset to convert Kelvin to Celsius.

[0072] (1c) Obtain precipitation: directly obtain it through the precipitation band in the CHIRPS dataset;

[0073] (1d) Obtain the annual average effective accumulated temperature. The annual average effective accumulated temperature is obtained based on the hourly surface temperature of the ERA5-Land reanalysis data using the following formula:

[0074] ET yearly =DD year / total number of days in a year

[0075] Among them, DD year The effective accumulated temperature of desert locusts is obtained through the Day Degree model (degree-day model) in one year, and is obtained through the following formula:

[0076]

[0077] Among them, n is the total number of days in a year, T i (t) is the surface temperature at hour t on day i of the year. LT = 15.5°C is the theoretical starting temperature for desert locust egg development. UT = 41°C is the maximum threshold temperature for egg hatching. When the temperature is lower than LT or higher than UT, locust eggs stop developing. Represents the DD value of the i-th day in a year. The DD value refers to the effective accumulated temperature of the Day Degree model;

[0078] (1e) Obtain digital elevation data: Digital elevation data are directly obtained from the 30m spatial resolution NASA DEM dataset;

[0079] (1f) Obtain soil moisture: Obtain soil moisture at 7-28 cm from ERA5-Land monthly data;

[0080] (1g) Obtaining land cover classification data: Land cover classification data are directly obtained from the MODIS annual land classification dataset MCD12Q1;

[0081] (1 hour) Obtaining ground survey point data: Obtain desert locust ground survey data from the Food and Agriculture Organization of the United Nations' freely available dataset at the Locust Center;

[0082] (1i) Obtaining soil sediment content: The soil sediment content was obtained from the SoilGrids250m2.0 dataset provided by the International Soil Reference and Information Center.

[0083] The step (2) specifically includes the following steps:

[0084] (2a) Unify the temporal resolution of multi-source data to interannual: perform annual maximum aggregation on the NDVI, annual mean aggregation on soil moisture, and annual sum aggregation on precipitation;

[0085] (2b) Unify the spatial resolution of multi-source data: resample the multi-source data to 5 km;

[0086] (2c) Masking permanent water bodies, forests, and buildings: Use land cover classification data as mask input to perform masking operations on multi-source data to exclude permanent water bodies, forests, and building locations;

[0087] (2d) Unified spatial range: The spatial range of multi-source data was unified using the SDM toolbox and limited to Yemen;

[0088] (2e) Cleaning redundant ground survey data: The ground survey data were extracted by year, and then one of the multi-source data grids was used as a background reference. ENMtools was used to clean the data of different years to ensure that there was at most one ground survey point in each grid. The cleaned ground survey data were saved as a CSV file for future use.

[0089] The step (4) specifically includes the following steps:

[0090] (4a) Construction of MaxEnt model: Constructing the MaxEnt niche model of the Yemeni desert locust;

[0091] (4b) Optimizing MaxEnt parameters: Automatically optimizing the feature class and regularization multiplier of the MaxEnt niche model for the Yemeni desert locust, selecting the model with the optimal parameters, i.e., the MaxEnt model with the optimal parameters, and recording the parameters of the MaxEnt model with the optimal parameters; that is, automatically optimizing the feature class (FC) and regularization multiplier (RM) of the MaxEnt model, and selecting the MaxEnt model with the optimal parameters from 50 different models obtained by combining FC (L, LQ, H, LQH, LQHPT) and RM (0.5:0.5:5);

[0092] (4c) Model training: Input the selected habitat factors and cleaned ground survey data into the MaxEnt model with optimal parameters to complete the model training; repeat the training 25 times to obtain 25 MaxEnt models with optimal parameters;

[0093] (4d) Obtaining the inter-annual habitat suitability results of the Yemeni Desert Locust: Based on the 25 optimal MaxEnt models, the area under the receiver operating characteristic curve (AUC) was used as the threshold, and the models with an area under the receiver operating characteristic curve (AUC) greater than 0.7 among the 25 optimal MaxEnt models were fused to obtain the final MaxEnt model. The habitat factors corresponding to the final MaxEnt model year were input into the final MaxEnt model to obtain the inter-annual habitat suitability results of the Yemeni Desert Locust. Figure 2 As shown, suitability ranges from 0 to 1, with higher suitability values indicating greater suitability for desert locust survival and reproduction. The results indicate that suitability for desert locusts in Yemen varies from year to year, but the western Red Sea plains of Yemen have consistently been an ideal habitat for desert locusts.

[0094] The step (5) specifically includes the following steps:

[0095] (5a) Create a blank mosaic dataset;

[0096] (5b) Add the interannual habitat suitability results of the Yemeni desert locust to the blank mosaic dataset, add time attributes to it, construct multidimensional information, and thus obtain multidimensional raster data;

[0097] (5c) Create a space-time cube using multidimensional raster data.

[0098] The step (6) specifically includes the following steps:

[0099] (6a) Emerging hotspot analysis: Emerging hotspot analysis was performed using the space-time cube to obtain the changing trends of habitat suitability in Yemen, such as Figure 3As shown;

[0100] (6b) Time series cluster analysis: Time series cluster analysis was performed using a space-time cube to obtain a time-space integration analysis result, which refers to the spatial and temporal clustering results of the Yemen Desert Locust suitability values. Figure 4 The mean value of habitat suitability for each cluster at each time step is shown, i.e. the mean value is used to summarize the categories of habitat suitability.

[0101] MaxEnt niche modeling was performed in years with sufficient ground survey data. Table 1 shows the optimal model parameters for each year in Yemen and the final MaxEnt training results. Kappa coefficients ranged from 0.492 to 0.893, TSS ranged from 0.508 to 0.919, and AUC ranged from 0.832 to 0.992, indicating that the constructed MaxEnt model is capable of effectively assessing habitat suitability for desert locusts in Yemen.

[0102] Table 1 Optimal model parameters and training results

[0103]

[0104] The present invention is applicable not only to Yemen, but also to habitat suitability assessment of desert locusts in any place.

[0105] This paper constructs a habitat suitability assessment framework for the Yemeni desert locust using an ecological model and a space-time cube method. The MaxEnt model is constructed using data such as the Normalized Difference Vegetation Index, surface temperature, precipitation, average annual effective accumulated temperature, digital elevation data, soil moisture, land cover classification data, desert locust ground survey data, and soil sediment content. The MaxEnt model results are then analyzed using the space-time cube method for a spatiotemporal integrated analysis, resulting in a grading system that fully accounts for temporal and spatial variations in habitat suitability for the Yemeni desert locust. Results demonstrate that the constructed MaxEnt model performs well and can provide relatively accurate habitat suitability distribution results. The spatiotemporal integrated analysis using the space-time cube method is consistent with actual results and can accurately assess the habitat suitability of the Yemeni desert locust.

Claims

1. A method for assessing desert locust habitat suitability in Yemen based on MaxEnt and space-time cube, characterized by: The method comprises the following steps in sequence: (1) Acquisition of multi-source data: Acquisition of meteorological data and satellite data, including normalized difference vegetation index, surface temperature, precipitation, annual average effective accumulated temperature, digital elevation data, soil moisture, and land cover classification data; acquisition of desert locust ground survey data and soil sand content; (2) Preprocessing of multi-source data: resampling and temporal aggregation of multi-source data to ensure that data from different sources have consistent spatial range, spatial resolution, and temporal resolution; Use land cover classification for permanent water bodies, forests, and built-up masks; Cleaning of desert locust ground survey data; (3) Screening of habitat factors: Calculate the Pearson coefficient between multi-source data and retain the multi-source data with a Pearson coefficient less than 0.8 as habitat factors; (4) Constructing the MaxEnt model and parameter tuning: Obtaining the MaxEnt model with the optimal parameters, training the MaxEnt model with the optimal parameters, and obtaining the interannual habitat suitability results for the Yemeni desert locust; (5) Constructing a space-time cube: adding time attributes to the inter-annual habitat suitability results of the Yemeni desert locust, constructing a multidimensional grid, and creating a space-time cube based on the inter-annual habitat suitability results of the Yemeni desert locust; (6) Conducting space-time integration analysis: Conducting space-time integration analysis on the space-time cube to obtain space-time integration analysis results; (7) Assessment of habitat suitability of desert locusts in Yemen: Based on the results of spatiotemporal integration analysis, the suitability of desert locusts in Yemen is classified into different levels.

2. the Yemen Desert Locust habitat suitability assessment method based on MaxEnt and space-time cube according to claim 1, is characterized in that: The step (1) specifically includes the following steps: (1a) Obtain the normalized vegetation index, which is as follows: NDVI = (NIR - Red) / (NIR + Red) Among them, NIR represents the reflectivity of the near-infrared band, and Red represents the reflectivity of the red band; (1b) Obtain the surface temperature using the following formula: LST=(day+night)×0.5×0.02-273.15 Wherein, day and night are the observed values during the day and night, respectively, and the daily LST is the average of the day and night; (1c) Obtain precipitation: directly obtain it through the precipitation band in the CHIRPS dataset; (1d) Obtain the annual average effective accumulated temperature: ET yearly =DD year / total number of days in a year Among them, DD year The effective accumulated temperature of desert locusts is obtained through the Day Degree model (degree-day model) in one year, and is obtained through the following formula: Among them, n is the total number of days in a year, T i (t) is the surface temperature at hour t on day i of the year. LT = 15.5°C is the theoretical starting temperature for desert locust egg development. UT = 41°C is the maximum threshold temperature for egg hatching. When the temperature is lower than LT or higher than UT, locust eggs stop developing. Represents the DD value of the i-th day in a year. The DD value refers to the effective accumulated temperature of the Day Degree model; (1e) Obtain digital elevation data: Digital elevation data are directly obtained from the 30m spatial resolution NASA DEM dataset; (1f) Obtain soil moisture: Obtain soil moisture at 7-28 cm from ERA5-Land monthly data; (1g) Obtaining land cover classification data: Land cover classification data are directly obtained from the MODIS annual land classification dataset MCD12Q1; (1h) Obtaining ground survey point data: Obtaining desert locust ground survey data; (1i) Obtain soil sediment content: Obtain soil sediment content from the SoilGrids250m 2.0 dataset.

3. the Yemen Desert Locust habitat suitability assessment method based on MaxEnt and space-time cube according to claim 1, is characterized in that: The step (2) specifically includes the following steps: (2a) Unify the temporal resolution of multi-source data to interannual: perform annual maximum aggregation on the NDVI, annual mean aggregation on soil moisture, and annual sum aggregation on precipitation; (2b) Unify the spatial resolution of multi-source data: resample the multi-source data to 5 km; (2c) Masking permanent water bodies, forests, and buildings: Use land cover classification data as mask input to perform masking operations on multi-source data to exclude permanent water bodies, forests, and building locations; (2d) Unified spatial scope: The spatial scope of multi-source data was unified and limited to Yemen; (2e) Cleaning redundant ground survey data: The ground survey data were extracted by year, and then the data of different years were cleaned using one of the multi-source data grids as a background reference to ensure that there was at most one ground survey point in each grid. The cleaned ground survey data were saved as a CSV file for future use.

4. the Yemen Desert Locust habitat suitability assessment method based on MaxEnt and space-time cube according to claim 1, is characterized in that: The step (4) specifically includes the following steps: (4a) Construction of MaxEnt model: Constructing the MaxEnt niche model of the Yemeni desert locust; (4b) Optimizing MaxEnt parameters: Automatically optimizing the characteristic classes and regularization multipliers of the MaxEnt ecological niche model for the Yemeni desert locust, selecting the model with the best parameters, i.e., the MaxEnt model with the best parameters, and recording the parameters of the MaxEnt model with the best parameters; (4c) Model training: Input the selected habitat factors and cleaned ground survey data into the MaxEnt model with optimal parameters to complete the model training; repeat the training 25 times to obtain 25 MaxEnt models with optimal parameters; (4d) Obtaining the inter-annual habitat suitability results for the Yemeni desert locust: Based on the 25 optimal MaxEnt models, the area under the receiver operating characteristic curve (AUC) was used as the threshold. The models with an area under the receiver operating characteristic curve (AUC) greater than 0.7 among the 25 optimal MaxEnt models were fused to obtain the final MaxEnt model. The habitat factors corresponding to the final MaxEnt model year were input into the final MaxEnt model to obtain the inter-annual habitat suitability results for the Yemeni desert locust.

5. the Yemen Desert Locust habitat suitability assessment method based on MaxEnt and space-time cube according to claim 1, is characterized in that: The step (5) specifically includes the following steps: (5a) Create a blank mosaic dataset; (5b) Add the interannual habitat suitability results of the Yemeni desert locust to the blank mosaic dataset, add time attributes to it, construct multidimensional information, and thus obtain multidimensional raster data; (5c) Create a space-time cube using multidimensional raster data.

6. the Yemen Desert Locust habitat suitability assessment method based on MaxEnt and space-time cube according to claim 1, is characterized in that: The step (6) specifically includes the following steps: (6a) Emerging hotspot analysis: Emerging hotspot analysis was performed using a space-time cube to obtain the changing trends of habitat suitability in Yemen; (6b) Time series cluster analysis: Time series cluster analysis was performed using a space-time cube to obtain a time-space integration analysis result, which refers to the spatial and temporal clustering results of the Yemen Desert Locust suitability values.

Images