Multi-source data fusion-based intelligent assessment system and method for comprehensive risk of Cytospora chrysosperma

Abstract

The invention discloses a system and a method for intelligently evaluating the comprehensive risk of Cytosporoma based on multi-source data fusion, and relates to the technical field of forestry disease risk prediction.The system comprises a valsa ceratosperma data acquisition module, a valsa ceratosperma data integration module and a risk evaluation and early warning module, and the valsa ceratosperma data acquisition module, the valsa ceratosperma data integration module and the risk evaluation and early warning module are connected with one another. In combination with database retrieval, data entry, data downloading and meteorological data assimilation, obtaining multi-modal data of the forest tree valsa ceratosperma; the valsa ceratosperma data integration module integrates the suitability index and the forest valsa ceratosperma propagation frequency through a random forest model, and the ecological niche model construction technology, the HYSPLIT model application technology and the modern information technology in the method are closely combined to obtain the suitability index and the forest valsa ceratosperma propagation frequency information; a random forest model is used for integrating multi-source data, and real-time and comprehensive monitoring of the risk level of the forest tree valsa ceratosperma is achieved.

Description

Technical Field

[0001] The invention relates to the technical field of forestry disease risk prediction, and in particular to a Cytospora chrysosperma comprehensive risk intelligent assessment system and method based on multi-source data fusion. Background Art

[0002] Tree rot is a typical forestry disaster caused by Ascosporum fungi, which mainly harms the main branches, side branches, bifurcations and other parts of the trees. Studies have shown that tree rot is mostly caused by drought, frost damage, mechanical damage and other reasons that lead to the peak of the incidence of Ascosporum pathogenic fungi. It mainly infects the subbark tissue. Its hyphae grow in or between the cells of the phloem to the xylem. In the early stage, it causes phloem discoloration. In the later stage, buried fruiting bodies are formed and protrude under the dead branches or damaged bark. It is highly contagious and spreads passively by wind. The transmission pathway involves three processes: release, transmission and deposition. In my country, due to the vast geographical area, the climate conditions and forest species vary greatly from place to place, and the damage caused by rot diseases in different places is also inconsistent, but the basic infection and disease patterns are similar. Therefore, before the symptoms appear in the trees, the niche model is derived. Based on the monitoring and analysis of pathogens, hosts, diseases and environment, the possibility and scale of the disease are calculated and forecasted in time to minimize or even avoid the disease and reduce losses. Considering the size of Cytospora spores, known transmission characteristics, and similarities with other fungal spores that have been proven to be able to spread over long distances, the use of the HYSPLIT model to study the potential transmission path and range of Cytospora spores has important scientific value;

[0003] Although the existing C. chrysosperma comprehensive risk intelligent assessment system and method based on multi-source data fusion has made great progress, there are still some problems that need to be optimized. The occurrence of forest rot disease is affected by environmental changes and transmission conditions. The existing C. chrysosperma comprehensive risk intelligent assessment system and method based on multi-source data fusion cannot predict the long-distance transmission range, diffusion trend, and risk spatiotemporal dynamics of forest rot pathogen spores in forest rot disease with air flow, scientifically analyze the possibility of epidemic disasters of rot disease in different regions, and capture the spatiotemporal dynamic changes of the risk of invasion of forest rot pathogens. The risk assessment of forest rot disease is not comprehensive and cannot provide data support for the targeted formulation of forest rot disease prevention and control measures. Summary of the invention

[0004] The present invention aims to provide a C. chrysosperma comprehensive risk intelligent assessment system and method based on multi-source data fusion to solve the problems raised in the above background technology.

[0005] To solve the above technical problems, the technical solutions adopted by the present invention are: in the first aspect, a C. chrysosperma comprehensive risk intelligent assessment system based on multi-source data fusion includes a decay pathogen data acquisition module, a decay pathogen data integration module and a risk assessment and early warning module;

[0006] The decay pathogen data collection module combines database retrieval, data entry, data download and meteorological data assimilation to obtain multimodal data of forest decay pathogens, providing data support for subsequent data integration;

[0007] The decay pathogen data integration module is divided into an ecological niche model construction unit, a HYSPLIT model construction unit, a meteorological data calculation unit, a weighted overlay unit and a factor integration unit, and integrates the suitability index and the transmission frequency of forest decay pathogens through a random forest model;

[0008] The risk assessment and early warning module combines the ecological niche model and the HYSPLIT model to analyze the multimodal data of forest wood decay pathogens, evaluate the risk level of forest wood decay pathogens under the conditions of multiple influencing factors, and perform early warning operations.

[0009] A further improvement of the technical solution of the present invention is that in the decay pathogen data acquisition module, the process of acquiring multimodal data of forest tree decay pathogens by combining database retrieval, data entry, data download and meteorological data assimilation includes:

[0010] The multimodal data of tree rot pathogens include distribution data, environmental data, and high-altitude meteorological data of tree rot pathogens, wherein the distribution data of tree rot pathogens is in .CSV format; the environmental data of tree rot pathogens are bioclimatic variables, topographic variables, geomorphic variables, and soil variables in .asc format; the high-altitude meteorological data of tree rot pathogens are air pressure, temperature, meridional wind, and latitudinal wind on an isobaric surface;

[0011] The distribution data of forest decay pathogens were obtained by searching the Global Biodiversity Information Facility, China National Knowledge Infrastructure (CNKI) and ScienceNet, and the entered distribution data were converted into a .CSV file. The .CSV file contains three columns: the first column is the species name, the second column is the longitude, and the third column is the latitude.

[0012] Search the CHELSA database and the ENVIREM database, and obtain 38 bioclimatic variables through data download; search the EarthEnv database, and obtain 8 topographic variables and geomorphic variables through data download; search the DAAC database, and obtain 1 soil variable through data download, convert the downloaded data into .asc format, and obtain the environmental data of forest decay pathogens;

[0013] The global data assimilation system data is used to obtain the high-altitude meteorological data of forest rot pathogens through the meteorological data assimilation method. The high-altitude meteorological data include wind speed, air pressure on isobaric surfaces, temperature, meridional wind and latitudinal wind.

[0014] The further improvement of the technical solution of the present invention is that: the decay pathogen data integration module is divided into an ecological niche model construction unit, a HYSPLIT model construction unit, a meteorological data calculation unit, a weighted superposition unit and a factor integration unit, and the process of integrating the suitability index and the transmission frequency of forest decay pathogens through the random forest model includes:

[0015] The ecological niche model construction unit is used to preprocess the collected data, optimize the ecological niche model parameters through the Kuenm package, and construct the ecological niche model; the HYSPLIT model construction unit is used to determine the release source point of the spores of the tree rot pathogen and construct the HYSPLIT model; the meteorological data calculation unit is used to calculate the propagation trajectory and propagation frequency of the spores of the tree rot pathogen; the weighted superposition unit is used to obtain the initial risk map of the tree rot pathogen through the weighted superposition method; the factor integration unit is used to integrate the suitability index and the propagation frequency of the tree rot pathogen through the random forest model.

[0016] A further improvement of the technical solution of the present invention is that: the ecological niche model construction unit preprocesses the collected data, optimizes the ecological niche model parameters through the Kuenm package, and the process of constructing the ecological niche model includes:

[0017] A1. Remove redundancy from the distribution data of tree rot pathogens to obtain normal distribution data of tree rot pathogens; use ENMTools software to analyze the correlation between bioclimatic variables, topographic variables, geomorphic variables and soil variables in the environmental data of tree rot pathogens; use Pearson correlation analysis and Jackknife method to screen variables affecting the suitable habitat of tree rot pathogens; evaluate the Pearson correlation coefficient R between the normal distribution data of tree rot pathogens and the environmental variables in the environmental data of tree rot pathogens; use R ≥ 0.8 as the threshold for determining the correlation between the normal distribution data of tree rot pathogens and the environmental variables in the environmental data of tree rot pathogens; combine MaxEnt software modeling to obtain the percentage contribution rate of bioclimatic variables, topographic variables, geomorphic variables and soil variables to the ecological niche model; retain the environmental variables in the environmental data of tree rot pathogens with the top 10 contribution rates;

[0018] A2. The parameters of the niche model include feature class and regularization multiplier. The feature class is represented by FC, the regularization multiplier is represented by RM, and the features of the model are divided into linear features, quadratic features, fragmented features and multiplicative features. The linear features are represented by L, the quadratic features are represented by Q, the fragmented features are represented by H, and the multiplicative features are represented by P. The four features L, Q, H, and P are combined into 15 FC combinations. The FC combinations are L, Q, H, P, LQ, LP, LH, QP, QH, PH, LQH, LPH, QPH and LQPH. RM is set to 0.1-4 with an interval of 0.1.

[0019] A3. Select the OR_AICC model with a missing rate of less than 5%, a DeltaAICc value of no more than 2, and the minimum DeltaAICc value as the basis of the optimal model, obtain the MaxEnt pre-modeling, output the results through the MaxEnt pre-modeling, and obtain the dominant environmental factors affecting the ecological and geographical distribution of forest wood decay pathogens based on the regularized training gain contribution rate of the Jackknife method and the single factor influence curve;

[0020] A4. The dominant environmental factors that affect the ecological and geographical distribution of forest decay pathogens were used as a data set, of which 75% was used as a training set and 25% was used as a test set. The MaxEnt model was trained using the training set through MaxEnt3.4.4. The performance of the MaxEnt model was evaluated through the test set combined with the area under the curve of the receiver operating characteristic method, and the MaxEnt model parameters were adjusted to obtain the ecological niche model.

[0021] A5. The dominant environmental factors that affect the eco-geographical distribution of tree rot pathogens are taken as input, and the niche model outputs a raster layer of the distribution of suitable habitats for tree rot pathogens. The spatial resolution is 30 arc seconds. The relative probability of the existence of tree rot pathogens is represented by a value of 0-1 in each grid cell. The relative probability of the existence of tree rot pathogens is also called the suitability index. The larger the value of the suitability index, the more suitable the environmental variables are for the survival of tree rot pathogens, and the higher the invasion risk.

[0022] A further improvement of the technical solution of the present invention is that the process of the HYSPLIT model construction unit determining the release source point of the spores of the forest rot pathogen and constructing the HYSPLIT model includes:

[0023] The normal distribution data of the woody decay pathogen after removing redundancy is determined as the release source of the spores of the woody decay pathogen;

[0024] The HYSPLIT model was established using high-altitude meteorological data and spore release sources of wood rot pathogens to simulate the propagation trajectory of wood rot spores along airflow in time series.

[0025] The HYSPLIT model uses the Lagrangian method to calculate the propagation trajectory of tree rot spores along the air flow as follows:

[0026] U 1 =U' (t+Δt) =U (t) +V (U , t) Δt

[0027]

[0028] Where V is the wind speed; Δt is the integration time step; U (t) is the initial position of the spores of forest rot; U 1 It is the first estimated location of spores of forest rot; U 2 It is the second estimated location of the spores of forest rot; U 3 is the final location of the spores of forest rot; V (U , t) is the wind speed at the initial location of the spores of forest rot; is the wind speed at the first estimated location of the wood rot spores.

[0029] A further improvement of the technical solution of the present invention is that the process of calculating the propagation trajectory and propagation frequency of the spores of the tree rot pathogen by the meteorological data calculation unit includes:

[0030] According to the peak period of forest rot pathogen transmission in spring and autumn, the HYSPLIT model is set from March 1 to October 31 each year, and March 1 to May 1 is divided into the spring category; June 1 to August 1 is divided into the summer category; September 1 to October 31 is divided into the autumn category; March 1 to October 31 is divided into the whole year category. Each source point starts to spread from 10m, 50m, and 100m above the ground at 00:00, 06:00, 12:00, and 18:00 every day. Each forward propagation simulation lasts 48 hours, and the simulation lasts for 5 years in total. The propagation trajectory of forest rot spores along the airflow in the time series is calculated.

[0031] The forest-covered area where the high-altitude meteorological data of tree rot pathogens were collected was divided into grid cells with a resolution of 10 km. ArcGIS 10.4 was used to analyze the propagation trajectory of tree rot spores along the air flow in the time series output by the HYSPLIT model, and the trajectory frequency of each grid cell was calculated as follows:

[0032]

[0033] Among them, TF ij is the transmission frequency of spores of forest rot pathogens; T ijis the number of propagation trajectories of tree rot spores along with the airflow in the time series that pass through the forest coverage area where the high-altitude meteorological data of tree rot pathogens are collected; N is the total number of propagation trajectories of tree rot spores along with the airflow in the forest area where the high-altitude meteorological data of tree rot pathogens are collected; i, j represent the row and column numbers of the grid unit, respectively.

[0034] A further improvement of the technical solution of the present invention is that the weighted superposition unit obtains the initial risk map of forest wood decay pathogens through the weighted superposition method, including:

[0035] S1. Use the niche model to obtain the grid layer of the distribution of the suitable habitat of the wood rot pathogen and the suitability index in each grid unit; use the HYSPLIT model to obtain the propagation trajectory of the wood rot spores along the air flow in the time series, and combine the calculation to obtain the spring propagation frequency map, summer propagation frequency map, autumn propagation frequency map, annual propagation frequency map and the propagation frequency of the wood rot pathogen in each grid unit;

[0036] S2. Resample and standardize the propagation frequency map so that the propagation frequency layer and the distribution grid layer of the suitable habitat of the wood decay pathogen maintain the same spatial resolution and projection;

[0037] S3. Clarify the importance of suitability index and propagation frequency in each grid unit. Based on the occurrence characteristics of tree rot diseases, compare the suitability index, the propagation frequency of tree rot pathogens in spring, summer, autumn and the propagation frequency of tree rot pathogens throughout the year in pairs, create a 5×5 judgment matrix, use the 1-9 assignment method to assign weights to the suitability index, the propagation frequency of tree rot pathogens in spring, summer, autumn and the propagation frequency of tree rot pathogens throughout the year, and record them in the judgment matrix to meet x cn × nc =1, where x cn and x nc is an element in the judgment matrix;

[0038] S4. Use the eigenvalue decomposition method to calculate the eigenvalues ​​and eigenvectors of the judgment matrix:

[0039] |A-λI|=0

[0040] Among them, A is the judgment matrix, λ is the eigenvalue, I is the unit matrix, solve multiple λ values, and take the largest λ max As the maximum eigenvalue, the corresponding maximum eigenvector is A max =(x 1 +x 2 +x 3 +x4 +x 5 ) T , and x 1 , x 2 , x 3 , x 4 , x 5 They are the suitability index corresponding to the largest eigenvector, the frequency of propagation of forest rot pathogens in spring, the frequency of propagation of forest rot pathogens in summer, the frequency of propagation of forest rot pathogens in autumn, and the frequency of propagation of forest rot pathogens throughout the year;

[0041] S5, against A max =(x 1 +x 2 +x 3 +x 4 +x 5 ) T Normalization processing was performed to calculate the weights corresponding to the suitability index, the frequency of propagation of forest rot pathogens in spring, the frequency of propagation of forest rot pathogens in summer, the frequency of propagation of forest rot pathogens in autumn, and the frequency of propagation of forest rot pathogens throughout the year:

[0042]

[0043] Among them, w c is the weight, c = 1, 2, 3, 4, 5, corresponding to x 1 , x 2 , x 3 , x 4 , x 5 , that is, w 1 , w 2 , w 3 , w 4 , w 5 They are the weights of suitability index, the frequency of spread of forest rot pathogens in spring, the frequency of spread of forest rot pathogens in summer, the frequency of spread of forest rot pathogens in autumn, and the frequency of spread of forest rot pathogens throughout the year;

[0044] S6. Pass The formula calculates the consistency index, where CI is the consistency index, c = 1, 2, 3, 4, 5, λ max is the maximum eigenvalue of the judgment matrix. The actual random consistency index value is obtained by querying the random consistency index table, which is represented by RI. The consistency ratio is calculated, and when the consistency ratio is less than 0.1, the matrix is ​​judged to be reasonable;

[0045] S7. Use weighted overlay to calculate the initial risk index corresponding to each grid cell:

[0046] F c =xc1 × 1 +x c2 × 2 +x c3 × 3 +x c4 × 4 +x c5 × 5

[0047] Among them, F c is the initial risk index corresponding to the cth row of the judgment matrix; w 1 , w 2 , w 3 , w 4 , w 5 They are the suitability index, the frequency of propagation of forest rot pathogens in spring, summer, autumn and the weight of the frequency of propagation of forest rot pathogens throughout the year; x c1 , x c2 , x c3 , x c4 , x c5 They are the elements corresponding to each row of the judgment matrix, c = 1, 2, 3, 4, 5;

[0048] S8. Use GIS software to assign the calculated initial risk index to the corresponding grid to obtain the initial risk map of forest wood decay pathogens.

[0049] A further improvement of the technical solution of the present invention is that the process of integrating the suitability index and the transmission frequency of forest rot pathogens by the factor integration unit through the random forest model includes:

[0050] A sampling layer was created in ArcGIS, sampling points were randomly generated, and a table was created. Each row in the table represented a grid unit, and each column of data included the occurrence probability of the suitability index, the propagation frequency of the forest rot pathogen in spring, the propagation frequency of the forest rot pathogen in summer, the propagation frequency of the forest rot pathogen in autumn, the propagation frequency of the forest rot pathogen throughout the year, and the initial risk index. Each column of data was used as input, and the comprehensive risk index was used as the target variable.

[0051] Using the random forest algorithm, the occurrence probability of the suitability index, the propagation frequency of the forest rot pathogen in spring, the propagation frequency of the forest rot pathogen in summer, the propagation frequency of the forest rot pathogen in autumn, the propagation frequency of the forest rot pathogen throughout the year, and the initial risk index were used as data sets, 80% of the data sets were divided into training sets, and 20% of the data sets were divided into test sets;

[0052] The random forest model was trained using the data set to learn the relationship between the occurrence probability of the suitability index, the propagation frequency of the forest rot pathogens in spring, summer, autumn, and the propagation frequency of the forest rot pathogens throughout the year and the initial risk index, as well as their influence on the comprehensive risk index, to obtain the random forest model; the test set was combined with the root mean square error, square root error, and determination coefficient to evaluate the performance of the trained random forest model, adjust the model parameters, and obtain the comprehensive risk assessment model;

[0053] Use the comprehensive risk assessment model to predict each grid unit, obtain the comprehensive risk index of each grid unit, generate a comprehensive risk map, and divide the risk level according to the comprehensive risk index. When the comprehensive risk index is between 0 and 0.3, it is classified as a low risk level; when the comprehensive risk index is between 0.3 and 0.6, it is classified as a medium risk level; when the comprehensive risk index is between 0.6 and 1, it is classified as a high risk level.

[0054] A further improvement of the technical solution of the present invention is that the risk assessment and early warning module, in combination with the niche model and the HYSPLIT model, analyzes the multimodal data of forest rot pathogens, assesses the risk level of forest rot pathogens under the conditions of multiple influencing factors, and performs early warning operations, including:

[0055] Combining the niche model and the HYSPLIT model, the multimodal data of tree rot pathogens were analyzed to obtain the comprehensive risk index of each grid unit. The tree rot pathogens with a comprehensive risk index between 0 and 0.3 were at a low risk level; the tree rot pathogens with a comprehensive risk index between 0.3 and 0.6 were at a medium risk level; and the tree rot pathogens with a comprehensive risk index between 0.6 and 1 were at a high risk level.

[0056] When the risk level of wood decay pathogens is low, the monitoring frequency will be increased and a low-risk warning signal will be issued; when the risk level of wood decay pathogens is medium, the monitoring frequency will be increased for areas corresponding to grid cells with high comprehensive risk indexes, and a warning signal will be issued that the areas corresponding to grid cells with high comprehensive risk indexes are medium-risk; when the risk level of wood decay pathogens is high, all areas will be intensively monitored and a high-risk warning signal will be issued.

[0057] In the second aspect, a C. chrysosperma comprehensive risk intelligent assessment method based on multi-source data fusion is used to implement the C. chrysosperma comprehensive risk intelligent assessment system based on multi-source data fusion, which is composed of the following steps:

[0058] Step 1: combining database retrieval, data entry, data download and meteorological data assimilation to collect multimodal data of forest rot pathogens, wherein the multimodal data of forest rot pathogens includes distribution data, environmental data and high-altitude meteorological data of forest rot pathogens;

[0059] Step 2: Remove redundancy from the distribution data of the tree rot pathogens, perform dimensionality reduction on the environmental data of the tree rot pathogens, optimize the niche model parameters through the Kuenm package, and construct the niche model;

[0060] Step 3: Determine the release source of the spores of the tree rot pathogen, build the HYSPLIT model, and calculate the propagation trajectory and frequency of the spores of the tree rot pathogen;

[0061] Step 4: Obtain the initial risk index corresponding to each grid unit through eigenvalue decomposition method, normalization processing, consistency index calculation and weighted overlay calculation. Combined with GIS software, assign the calculated initial risk index to the corresponding grid to obtain the initial risk map of forest wood decay pathogens.

[0062] Step 5: Integrate the suitability index and the frequency of tree decay pathogen transmission through the random forest model;

[0063] Step 6. Combine the niche model and the HYSPLIT model to analyze the multimodal data of wood decay pathogens, evaluate the risk level of wood decay pathogens under multiple influencing factors, and when the risk level of wood decay pathogens is low, strengthen the monitoring frequency and issue a low-risk warning signal; when the risk level of wood decay pathogens is medium, increase the monitoring frequency for areas corresponding to grid cells with high comprehensive risk indexes, and issue a medium-risk warning signal for areas corresponding to grid cells with high comprehensive risk indexes; when the risk level of wood decay pathogens is high, conduct intensive monitoring of all areas and issue a high-risk warning signal.

[0064] Beneficial effects of the present invention: Compared with the traditional C. chrysosperma comprehensive risk intelligent assessment system and method, the C. chrysosperma comprehensive risk intelligent assessment system and method based on multi-source data fusion are closely integrated with the niche model construction technology and HYSPLIT model application technology in the method of the present invention with modern information technology. Through the effective collection and redundancy removal of forest rot pathogen distribution data and the dimensionality reduction processing of environmental data, the key data of forest rot pathogens are accurately captured, the suitability index and the frequency information of forest rot pathogen transmission are obtained, and the random forest model is used to integrate multi-source data to achieve real-time and comprehensive monitoring of the risk level of forest rot pathogens. The complex relationship and change process of multiple factors such as environmental suitability and transmission dynamics are comprehensively considered, and the existing multi-source data are solved. The integrated C. chrysosperma comprehensive risk intelligent assessment system and method cannot predict the long-distance transmission range, diffusion trend, and risk spatiotemporal dynamics of tree rot pathogen spores in tree rot disease with air flow, and the risk assessment of tree rot disease is not comprehensive, and cannot provide data support for the targeted formulation of tree rot disease prevention and control measures. It is ensured that the method in the present invention can refine the dynamic monitoring standards for the C. chrysosperma comprehensive risk intelligent assessment system and method based on multi-source data fusion within a more precise range, so that the monitored data becomes a more accurate indicator under the same conditions. The development and application of this method have significantly enhanced the degree of intelligence in the process of forestry disease risk assessment, provided a strong guarantee for the scientific formulation of tree rot disease prevention and control strategies, and effectively improved the protection effectiveness and management level of forestry ecology. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0066] Figure 1 It is a block diagram of the C. chrysosperma comprehensive risk intelligent assessment system based on multi-source data fusion of the present invention;

[0067] Figure 2 The figure is a flow chart of the C. chrysosperma comprehensive risk intelligent assessment method based on multi-source data fusion of the present invention. DETAILED DESCRIPTION

[0068] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0069] Embodiment 1, as Figure 1 As shown, the present invention provides a C. chrysosperma comprehensive risk intelligent assessment system based on multi-source data fusion, including a decay pathogen data acquisition module, a decay pathogen data integration module and a risk assessment and early warning module;

[0070] The decay pathogen data collection module combines database retrieval, data entry, data download and meteorological data assimilation to obtain multimodal data of forest decay pathogens and provide data support for subsequent data integration;

[0071] The module for integrating data of tree rot pathogens is divided into the ecological niche model building unit, the HYSPLIT model building unit, the meteorological data calculation unit, the weighted overlay unit and the factor integration unit. It integrates the suitability index and the frequency of tree rot pathogen transmission through the random forest model.

[0072] The risk assessment and early warning module combines the ecological niche model and the HYSPLIT model to analyze the multimodal data of forest wood decay pathogens, evaluate the risk level of forest wood decay pathogens under multiple influencing factors, and perform early warning operations.

[0073] In the decay pathogen data collection module, the process of obtaining multimodal data of forest decay pathogens by combining database retrieval, data entry, data download and meteorological data assimilation includes:

[0074] The multimodal data of tree rot pathogens include distribution data, environmental data, and high-altitude meteorological data of tree rot pathogens, wherein the distribution data of tree rot pathogens is in .CSV format; the environmental data of tree rot pathogens are bioclimatic variables, topographic variables, geomorphic variables, and soil variables in .asc format; the high-altitude meteorological data of tree rot pathogens are air pressure, temperature, meridional wind, and latitudinal wind on an isobaric surface;

[0075] The distribution data of forest decay pathogens were obtained by searching the Global Biodiversity Information Facility, China National Knowledge Infrastructure (CNKI) and ScienceNet, and the entered distribution data were converted into a .CSV file. The .CSV file contains three columns: the first column is the species name, the second column is the longitude, and the third column is the latitude.

[0076] Search the CHELSA database and the ENVIREM database, and obtain 38 bioclimatic variables through data download; search the EarthEnv database, and obtain 8 topographic variables and geomorphic variables through data download; search the DAAC database, and obtain 1 soil variable through data download, convert the downloaded data into .asc format, and obtain the environmental data of forest decay pathogens;

[0077] The global data assimilation system data is used to obtain the high-altitude meteorological data of forest rot pathogens through the meteorological data assimilation method. The high-altitude meteorological data include wind speed, air pressure on isobaric surfaces, temperature, meridional wind and latitudinal wind.

[0078] The decay pathogen data integration module is divided into a niche model construction unit, a HYSPLIT model construction unit, a meteorological data calculation unit, a weighted overlay unit, and a factor integration unit. The process of integrating the suitability index and the transmission frequency of tree decay pathogens through the random forest model includes:

[0079] The niche model construction unit is used to preprocess the collected data, optimize the niche model parameters through the Kuenm package, and construct the niche model; the HYSPLIT model construction unit is used to determine the release source point of the spores of the tree rot pathogen and construct the HYSPLIT model; the meteorological data calculation unit is used to calculate the propagation trajectory and frequency of the spores of the tree rot pathogen; the weighted superposition unit is used to obtain the initial risk map of the tree rot pathogen through the weighted superposition method; the factor integration unit is used to integrate the suitability index and the propagation frequency of the tree rot pathogen through the random forest model.

[0080] The niche model building unit preprocesses the collected data and optimizes the niche model parameters through the Kuenm package. The process of building the niche model includes:

[0081] A1. Remove redundancy from the distribution data of tree rot pathogens to obtain normal distribution data of tree rot pathogens; use ENMTools software to analyze the correlation between bioclimatic variables, topographic variables, geomorphic variables and soil variables in the environmental data of tree rot pathogens; use Pearson correlation analysis and Jackknife method to screen variables affecting the suitable habitat of tree rot pathogens; evaluate the Pearson correlation coefficient R between the normal distribution data of tree rot pathogens and the environmental variables in the environmental data of tree rot pathogens; use R ≥ 0.8 as the threshold for determining the correlation between the normal distribution data of tree rot pathogens and the environmental variables in the environmental data of tree rot pathogens; combine MaxEnt software modeling to obtain the percentage contribution rate of bioclimatic variables, topographic variables, geomorphic variables and soil variables to the ecological niche model; retain the environmental variables in the environmental data of tree rot pathogens with the top 10 contribution rates;

[0082] A2. The parameters of the niche model include feature class and regularization multiplier. The feature class is represented by FC, the regularization multiplier is represented by RM, and the features of the model are divided into linear features, quadratic features, fragmented features and multiplicative features. The linear features are represented by L, the quadratic features are represented by Q, the fragmented features are represented by H, and the multiplicative features are represented by P. The four features L, Q, H, and P are combined into 15 FC combinations. The FC combinations are L, Q, H, P, LQ, LP, LH, QP, QH, PH, LQH, LPH, QPH and LQPH. RM is set to 0.1-4 with an interval of 0.1.

[0083] A3. Select the OR_AICC model with a missing rate of less than 5%, a DeltaAICc value of no more than 2, and the minimum DeltaAICc value as the basis of the optimal model, obtain the MaxEnt pre-modeling, output the results through the MaxEnt pre-modeling, and obtain the dominant environmental factors affecting the ecological and geographical distribution of forest wood decay pathogens based on the regularized training gain contribution rate of the Jackknife method and the single factor influence curve;

[0084] A4. The dominant environmental factors that affect the ecological and geographical distribution of forest decay pathogens were used as a data set, of which 75% was used as a training set and 25% was used as a test set. The MaxEnt model was trained using the training set through MaxEnt3.4.4. The performance of the MaxEnt model was evaluated through the test set combined with the area under the curve of the receiver operating characteristic method, and the MaxEnt model parameters were adjusted to obtain the ecological niche model.

[0085] A5. The dominant environmental factors that affect the eco-geographical distribution of tree rot pathogens are taken as input, and the niche model outputs a raster layer of the distribution of suitable habitats for tree rot pathogens. The spatial resolution is 30 arc seconds. The relative probability of the existence of tree rot pathogens is represented by a value of 0-1 in each grid cell. The relative probability of the existence of tree rot pathogens is also called the suitability index. The larger the value of the suitability index, the more suitable the environmental variables are for the survival of tree rot pathogens, and the higher the invasion risk.

[0086] The process of constructing the HYSPLIT model unit, determining the release source of the spores of the tree rot pathogen, and constructing the HYSPLIT model includes:

[0087] The normal distribution data of the woody decay pathogen after removing redundancy is determined as the release source of the spores of the woody decay pathogen;

[0088] The HYSPLIT model was established using high-altitude meteorological data and spore release sources of wood rot pathogens to simulate the propagation trajectory of wood rot spores along airflow in time series.

[0089] The HYSPLIT model uses the Lagrangian method to calculate the propagation trajectory of tree rot spores along the air flow as follows:

[0090] U 1 =U' (t+Δt) =U (t) +V (U , t) Δt

[0091]

[0092] Where V is the wind speed; Δt is the integration time step; U (t) is the initial position of the spores of forest rot; U 1 It is the first estimated location of spores of forest rot; U 2 It is the second estimated location of the spores of forest rot; U 3 is the final location of the spores of forest rot; V (U , t) is the wind speed at the initial location of the spores of forest rot; is the wind speed at the first estimated location of the wood rot spores.

[0093] The meteorological data calculation unit calculates the propagation trajectory and propagation frequency of the spores of the tree rot pathogen, including:

[0094] According to the peak period of forest rot pathogen transmission in spring and autumn, the HYSPLIT model is set from March 1 to October 31 each year, and March 1 to May 1 is divided into the spring category; June 1 to August 1 is divided into the summer category; September 1 to October 31 is divided into the autumn category; March 1 to October 31 is divided into the whole year category. Each source point starts to spread from 10m, 50m, and 100m above the ground at 00:00, 06:00, 12:00, and 18:00 every day. Each forward propagation simulation lasts 48 hours, and the simulation lasts for 5 years in total. The propagation trajectory of forest rot spores along the airflow in the time series is calculated.

[0095] The forest-covered area where the high-altitude meteorological data of tree rot pathogens were collected was divided into grid cells with a resolution of 10 km. ArcGIS 10.4 was used to analyze the propagation trajectory of tree rot spores along the air flow in the time series output by the HYSPLIT model, and the trajectory frequency of each grid cell was calculated as follows:

[0096]

[0097] Among them, TF ij is the transmission frequency of spores of forest rot pathogens; T ijis the number of propagation trajectories of tree rot spores along with the airflow in the time series that pass through the forest coverage area where the high-altitude meteorological data of tree rot pathogens are collected; N is the total number of propagation trajectories of tree rot spores along with the airflow in the forest area where the high-altitude meteorological data of tree rot pathogens are collected; i, j represent the row and column numbers of the grid unit, respectively.

[0098] The weighted overlay unit, through the weighted overlay method, obtains the initial risk map of forest decay pathogens, including:

[0099] S1. Use the niche model to obtain the grid layer of the distribution of the suitable habitat of the wood rot pathogen and the suitability index in each grid unit; use the HYSPLIT model to obtain the propagation trajectory of the wood rot spores along the air flow in the time series, and combine the calculation to obtain the spring propagation frequency map, summer propagation frequency map, autumn propagation frequency map, annual propagation frequency map and the propagation frequency of the wood rot pathogen in each grid unit;

[0100] S2. Resample and standardize the propagation frequency map so that the propagation frequency layer and the distribution grid layer of the suitable habitat of the wood decay pathogen maintain the same spatial resolution and projection;

[0101] S3. Clarify the importance of suitability index and propagation frequency in each grid unit. Based on the occurrence characteristics of tree rot diseases, compare the suitability index, the propagation frequency of tree rot pathogens in spring, summer, autumn and the propagation frequency of tree rot pathogens throughout the year in pairs, create a 5×5 judgment matrix, use the 1-9 assignment method to assign weights to the suitability index, the propagation frequency of tree rot pathogens in spring, summer, autumn and the propagation frequency of tree rot pathogens throughout the year, and record them in the judgment matrix to meet x cn × nc =1, where x cn and x nc is an element in the judgment matrix;

[0102] S4. Use the eigenvalue decomposition method to calculate the eigenvalues ​​and eigenvectors of the judgment matrix:

[0103] |A-λI|=0

[0104] Among them, A is the judgment matrix, λ is the eigenvalue, I is the unit matrix, solve multiple λ values, and take the largest λ max As the maximum eigenvalue, the corresponding maximum eigenvector is A max =(x 1 +x 2 +x 3 +x 4 +x 5) T , and x 1 , x 2 , x 3 , x 4 , x 5 They are the suitability index corresponding to the largest eigenvector, the frequency of propagation of forest rot pathogens in spring, the frequency of propagation of forest rot pathogens in summer, the frequency of propagation of forest rot pathogens in autumn, and the frequency of propagation of forest rot pathogens throughout the year;

[0105] S5, against A max =(x 1 +x 2 +x 3 +x 4 +x 5 ) T Normalization processing was performed to calculate the weights corresponding to the suitability index, the frequency of propagation of forest rot pathogens in spring, the frequency of propagation of forest rot pathogens in summer, the frequency of propagation of forest rot pathogens in autumn, and the frequency of propagation of forest rot pathogens throughout the year:

[0106]

[0107] Among them, w c is the weight, c = 1, 2, 3, 4, 5, corresponding to x 1 , x 2 , x 3 , x 4 , x 5 , that is, w 1 , w 2 , w 3 , w 4 , w 5 They are the weights of suitability index, the frequency of spread of forest rot pathogens in spring, the frequency of spread of forest rot pathogens in summer, the frequency of spread of forest rot pathogens in autumn, and the frequency of spread of forest rot pathogens throughout the year;

[0108] S6. Pass The formula calculates the consistency index, where CI is the consistency index, c = 1, 2, 3, 4, 5, λ max is the maximum eigenvalue of the judgment matrix. The actual random consistency index value is obtained by querying the random consistency index table, which is represented by RI. The consistency ratio is calculated, and when the consistency ratio is less than 0.1, the matrix is ​​judged to be reasonable;

[0109] S7. Use weighted overlay to calculate the initial risk index corresponding to each grid cell:

[0110] F c =x c1 × 1+x c2 × 2 +x c3 × 3 +x c4 × 4 +x c5 × 5

[0111] Among them, F c is the initial risk index corresponding to the cth row of the judgment matrix; w 1 , w 2 , w 3 , w 4 , w 5 They are the suitability index, the frequency of propagation of forest rot pathogens in spring, summer, autumn and the weight of the frequency of propagation of forest rot pathogens throughout the year; x c1 , x c2 , x c3 , x c4 , x c5 They are the elements corresponding to each row of the judgment matrix, c = 1, 2, 3, 4, 5;

[0112] S8. Use GIS software to assign the calculated initial risk index to the corresponding grid to obtain the initial risk map of forest wood decay pathogens.

[0113] The process of integrating the suitability index and the frequency of tree decay pathogen transmission through the random forest model includes:

[0114] A sampling layer was created in ArcGIS, sampling points were randomly generated, and a table was created. Each row in the table represented a grid unit, and each column of data included the occurrence probability of the suitability index, the propagation frequency of the forest rot pathogen in spring, the propagation frequency of the forest rot pathogen in summer, the propagation frequency of the forest rot pathogen in autumn, the propagation frequency of the forest rot pathogen throughout the year, and the initial risk index. Each column of data was used as input, and the comprehensive risk index was used as the target variable.

[0115] Using the random forest algorithm, the occurrence probability of the suitability index, the propagation frequency of the forest rot pathogen in spring, the propagation frequency of the forest rot pathogen in summer, the propagation frequency of the forest rot pathogen in autumn, the propagation frequency of the forest rot pathogen throughout the year, and the initial risk index were used as data sets, 80% of the data sets were divided into training sets, and 20% of the data sets were divided into test sets;

[0116] The random forest model was trained using the data set to learn the relationship between the occurrence probability of the suitability index, the propagation frequency of the forest rot pathogens in spring, summer, autumn, and the propagation frequency of the forest rot pathogens throughout the year and the initial risk index, as well as their influence on the comprehensive risk index, to obtain the random forest model; the test set was combined with the root mean square error, square root error, and determination coefficient to evaluate the performance of the trained random forest model, adjust the model parameters, and obtain the comprehensive risk assessment model;

[0117] Use the comprehensive risk assessment model to predict each grid unit, obtain the comprehensive risk index of each grid unit, generate a comprehensive risk map, and divide the risk level according to the comprehensive risk index. When the comprehensive risk index is between 0 and 0.3, it is classified as a low risk level; when the comprehensive risk index is between 0.3 and 0.6, it is classified as a medium risk level; when the comprehensive risk index is between 0.6 and 1, it is classified as a high risk level.

[0118] The risk assessment and early warning module combines the niche model and the HYSPLIT model to analyze the multimodal data of forest rot pathogens, evaluate the risk level of forest rot pathogens under multiple influencing factors, and conduct early warning operations, including:

[0119] Combining the niche model and the HYSPLIT model, the multimodal data of tree rot pathogens were analyzed to obtain the comprehensive risk index of each grid unit. The tree rot pathogens with a comprehensive risk index between 0 and 0.3 were at a low risk level; the tree rot pathogens with a comprehensive risk index between 0.3 and 0.6 were at a medium risk level; and the tree rot pathogens with a comprehensive risk index between 0.6 and 1 were at a high risk level.

[0120] When the risk level of wood decay pathogens is low, the monitoring frequency will be increased and a low-risk warning signal will be issued; when the risk level of wood decay pathogens is medium, the monitoring frequency will be increased for areas corresponding to grid cells with high comprehensive risk indexes, and a warning signal will be issued that the areas corresponding to grid cells with high comprehensive risk indexes are medium-risk; when the risk level of wood decay pathogens is high, all areas will be intensively monitored and a high-risk warning signal will be issued.

[0121] Embodiment 2, as Figure 2 As shown, on the basis of Example 1, the present invention provides a technical solution: a C. chrysosperma comprehensive risk intelligent assessment method based on multi-source data fusion, which is used to implement the above-mentioned C. chrysosperma comprehensive risk intelligent assessment system based on multi-source data fusion, and is composed of the following steps:

[0122] Step 1: combining database retrieval, data entry, data download and meteorological data assimilation to collect multimodal data of forest rot pathogens, wherein the multimodal data of forest rot pathogens includes distribution data, environmental data and high-altitude meteorological data of forest rot pathogens;

[0123] Step 2: Remove redundancy from the distribution data of the tree rot pathogens, perform dimensionality reduction on the environmental data of the tree rot pathogens, optimize the niche model parameters through the Kuenm package, and construct the niche model;

[0124] Step 3: Determine the release source of the spores of the tree rot pathogen, build the HYSPLIT model, and calculate the propagation trajectory and frequency of the spores of the tree rot pathogen;

[0125] Step 4: Obtain the initial risk index corresponding to each grid unit through eigenvalue decomposition method, normalization processing, consistency index calculation and weighted overlay calculation. Combined with GIS software, assign the calculated initial risk index to the corresponding grid to obtain the initial risk map of forest wood decay pathogens.

[0126] Step 5: Integrate the suitability index and the frequency of tree decay pathogen transmission through the random forest model;

[0127] Step 6. Combine the niche model and the HYSPLIT model to analyze the multimodal data of wood decay pathogens, evaluate the risk level of wood decay pathogens under multiple influencing factors, and when the risk level of wood decay pathogens is low, strengthen the monitoring frequency and issue a low-risk warning signal; when the risk level of wood decay pathogens is medium, increase the monitoring frequency for areas corresponding to grid cells with high comprehensive risk indexes, and issue a medium-risk warning signal for areas corresponding to grid cells with high comprehensive risk indexes; when the risk level of wood decay pathogens is high, conduct intensive monitoring of all areas and issue a high-risk warning signal.

[0128] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A comprehensive risk intelligent assessment system for Cytospora chrysosperma based on multi-source data fusion, including a decay pathogen data collection module, a decay pathogen data integration module and a risk assessment and early warning module, characterized in that: The decay pathogen data collection module combines database retrieval, data entry, data download and meteorological data assimilation to obtain multimodal data of forest decay pathogens; The decay pathogen data integration module is divided into an ecological niche model construction unit, a HYSPLIT model construction unit, a meteorological data calculation unit, a weighted overlay unit and a factor integration unit, and integrates the suitability index and the transmission frequency of forest decay pathogens through a random forest model; The risk assessment and early warning module combines the ecological niche model and the HYSPLIT model to analyze the multimodal data of forest wood decay pathogens, evaluate the risk level of forest wood decay pathogens under the conditions of multiple influencing factors, and perform early warning operations.

2. The C. chrysosperma comprehensive risk intelligent assessment system based on multi-source data fusion according to claim 1, characterized in that: In the decay pathogen data collection module, the process of obtaining multimodal data of forest decay pathogens by combining database retrieval, data entry, data download and meteorological data assimilation includes: The multimodal data of tree rot pathogens include distribution data, environmental data, and high-altitude meteorological data of tree rot pathogens, wherein the distribution data of tree rot pathogens is in .CSV format; the environmental data of tree rot pathogens are bioclimatic variables, topographic variables, geomorphic variables, and soil variables in .asc format; the high-altitude meteorological data of tree rot pathogens are air pressure, temperature, meridional wind, and latitudinal wind on an isobaric surface; The distribution data of forest decay pathogens were obtained by searching the Global Biodiversity Information Facility, China National Knowledge Infrastructure (CNKI) and ScienceNet, and the entered distribution data were converted into a .CSV file. The .CSV file contains three columns: the first column is the species name, the second column is the longitude, and the third column is the latitude. Search the CHELSA database and the ENVIREM database, and obtain 38 bioclimatic variables through data download; search the EarthEnv database, and obtain 8 topographic variables and geomorphic variables through data download; search the DAAC database, and obtain 1 soil variable through data download, convert the downloaded data into .asc format, and obtain the environmental data of forest decay pathogens; The global data assimilation system data is used to obtain the high-altitude meteorological data of forest rot pathogens through the meteorological data assimilation method. The high-altitude meteorological data include wind speed, air pressure on isobaric surfaces, temperature, meridional wind and latitudinal wind.

3. The C. chrysosperma comprehensive risk intelligent assessment system based on multi-source data fusion according to claim 2 is characterized by: The rot pathogen data integration module is divided into an ecological niche model construction unit, a HYSPLIT model construction unit, a meteorological data calculation unit, a weighted overlay unit and a factor integration unit. The process of integrating the suitability index and the transmission frequency of forest rot pathogens through the random forest model includes: The ecological niche model construction unit is used to preprocess the collected data, optimize the ecological niche model parameters through the Kuenm package, and construct the ecological niche model; the HYSPLIT model construction unit is used to determine the release source point of the spores of the tree rot pathogen and construct the HYSPLIT model; the meteorological data calculation unit is used to calculate the propagation trajectory and propagation frequency of the spores of the tree rot pathogen; the weighted superposition unit is used to obtain the initial risk map of the tree rot pathogen through the weighted superposition method; the factor integration unit is used to integrate the suitability index and the propagation frequency of the tree rot pathogen through the random forest model.

4. The C. chrysosperma comprehensive risk intelligent assessment system based on multi-source data fusion according to claim 3 is characterized by: The ecological niche model construction unit preprocesses the collected data and optimizes the ecological niche model parameters through the Kuenm package. The process of constructing the ecological niche model includes: A1. Remove redundancy from the distribution data of tree rot pathogens to obtain normal distribution data of tree rot pathogens; use ENMTools software to analyze the correlation between bioclimatic variables, topographic variables, geomorphic variables and soil variables in the environmental data of tree rot pathogens; use Pearson correlation analysis and Jackknife method to screen variables affecting the suitable habitat of tree rot pathogens; evaluate the Pearson correlation coefficient R between the normal distribution data of tree rot pathogens and the environmental variables in the environmental data of tree rot pathogens; use |R|≥0.8 as the threshold for determining the correlation between the normal distribution data of tree rot pathogens and the environmental variables in the environmental data of tree rot pathogens; combine MaxEnt software modeling to obtain the percentage contribution rate of bioclimatic variables, topographic variables, geomorphic variables and soil variables to the niche model; retain the environmental variables in the environmental data of tree rot pathogens with the top 10 contribution rates; A2. The parameters of the niche model include feature class and regularization multiplier. The feature class is represented by FC, the regularization multiplier is represented by RM, and the features of the model are divided into linear features, quadratic features, fragmented features and multiplicative features. The linear features are represented by L, the quadratic features are represented by Q, the fragmented features are represented by H, and the multiplicative features are represented by P. The four features L, Q, H, and P are combined into 15 FC combinations. The FC combinations are L, Q, H, P, LQ, LP, LH, QP, QH, PH, LQH, LPH, QPH and LQPH. RM is set to 0.1-4 with an interval of 0.

1. A3. Select the OR_AICC model with a missing rate of less than 5%, a Delta AICc value of no more than 2, and the minimum Delta AICc value as the basis of the optimal model, obtain the MaxEnt pre-modeling, output the results through the MaxEnt pre-modeling, and obtain the dominant environmental factors affecting the ecological and geographical distribution of forest wood decay pathogens based on the regularized training gain contribution rate of the Jackknife method and the single factor influence curve; A4. The dominant environmental factors that affect the ecological and geographical distribution of forest decay pathogens were used as a data set, of which 75% was used as a training set and 25% was used as a test set. The MaxEnt model was trained using the training set through MaxEnt3.4.

4. The performance of the MaxEnt model was evaluated through the test set combined with the area under the curve of the receiver operating characteristic method, and the MaxEnt model parameters were adjusted to obtain the ecological niche model. A5. The dominant environmental factors that affect the eco-geographical distribution of tree rot pathogens are taken as input, and the niche model outputs a raster layer of the distribution of suitable habitats for tree rot pathogens. The spatial resolution is 30 arc seconds. The relative probability of the existence of tree rot pathogens is represented by a value of 0-1 in each grid cell. The relative probability of the existence of tree rot pathogens is also called the suitability index. The larger the value of the suitability index, the more suitable the environmental variables are for the survival of tree rot pathogens, and the higher the invasion risk.

5. The C. chrysosperma comprehensive risk intelligent assessment system based on multi-source data fusion according to claim 4 is characterized by: The HYSPLIT model construction unit determines the release source of the spores of the tree rot pathogen and constructs the HYSPLIT model, and the process includes: The normal distribution data of the woody decay pathogen after removing redundancy is determined as the release source of the spores of the woody decay pathogen; The HYSPLIT model was established using high-altitude meteorological data and spore release sources of wood rot pathogens to simulate the propagation trajectory of wood rot spores along airflow in time series. The HYSPLIT model uses the Lagrangian method to calculate the propagation trajectory of tree rot spores along the air flow as follows: U1=U' (t+Δt) =U (t) +V (U , t) Δt Where V is the wind speed; Δt is the integration time step; U (t) is the initial position of the spores of the forest rot disease; U1 is the first estimated position of the spores of the forest rot disease; U2 is the second estimated position of the spores of the forest rot disease; U3 is the final position of the spores of the forest rot disease; V (U , t) is the wind speed at the initial location of the spores of forest rot; is the wind speed at the first estimated location of the wood rot spores.

6. The C. chrysosperma comprehensive risk intelligent assessment system based on multi-source data fusion according to claim 5, characterized in that: The process of the meteorological data calculation unit calculating the propagation trajectory and propagation frequency of the spores of the tree rot pathogen includes: According to the peak period of forest rot pathogen transmission in spring and autumn, the HYSPLIT model is set from March 1 to October 31 each year, and March 1 to May 1 is divided into the spring category; June 1 to August 1 is divided into the summer category; September 1 to October 31 is divided into the autumn category; March 1 to October 31 is divided into the whole year category. Each source point starts to spread from 10m, 50m, and 100m above the ground at 00:00, 06:00, 12:00, and 18:00 every day. Each forward propagation simulation lasts 48 hours, and the simulation lasts for 5 years in total. The propagation trajectory of forest rot spores along the airflow in the time series is calculated. The forest-covered area where the high-altitude meteorological data of tree rot pathogens were collected was divided into grid cells with a resolution of 10 km. ArcGIS 10.4 was used to analyze the propagation trajectory of tree rot spores along the air flow in the time series output by the HYSPLIT model, and the trajectory frequency of each grid cell was calculated as follows: Among them, TF ij is the transmission frequency of spores of forest rot pathogens; T ij is the number of propagation trajectories of tree rot spores along with the airflow in the time series that pass through the forest coverage area where the high-altitude meteorological data of tree rot pathogens are collected; N is the total number of propagation trajectories of tree rot spores along with the airflow in the forest area where the high-altitude meteorological data of tree rot pathogens are collected; i, j represent the row and column numbers of the grid unit, respectively.

7. The C. chrysosperma comprehensive risk intelligent assessment system based on multi-source data fusion according to claim 6, characterized in that: The weighted overlay unit obtains the initial risk map of forest wood decay pathogens by weighted overlay method, including: S1. Use the niche model to obtain the grid layer of the distribution of the suitable habitat of the wood rot pathogen and the suitability index in each grid unit; use the HYSPLIT model to obtain the propagation trajectory of the wood rot spores along the air flow in the time series, and combine the calculation to obtain the spring propagation frequency map, summer propagation frequency map, autumn propagation frequency map, annual propagation frequency map and the propagation frequency of the wood rot pathogen in each grid unit; S2. Resample and standardize the propagation frequency map so that the propagation frequency layer and the distribution grid layer of the suitable habitat of the wood decay pathogen maintain the same spatial resolution and projection; S3. Clarify the importance of suitability index and propagation frequency in each grid unit. Based on the occurrence characteristics of tree rot diseases, compare the suitability index, the propagation frequency of tree rot pathogens in spring, summer, autumn and the propagation frequency of tree rot pathogens throughout the year in pairs, create a 5×5 judgment matrix, use the 1-9 assignment method to assign weights to the suitability index, the propagation frequency of tree rot pathogens in spring, summer, autumn and the propagation frequency of tree rot pathogens throughout the year, and record them in the judgment matrix to meet x cn × nc =1, where x cn and x nc is an element in the judgment matrix; S4. Use the eigenvalue decomposition method to calculate the eigenvalues ​​and eigenvectors of the judgment matrix: |A-λI|=0 Among them, A is the judgment matrix, λ is the eigenvalue, I is the unit matrix, solve multiple λ values, and take the largest λ max As the maximum eigenvalue, the corresponding maximum eigenvector is A max =(x1+x2+x3+x4+x5) T , and x1, x2, x3, x4, and x5 are the suitability index corresponding to the maximum eigenvector, the propagation frequency of forest rot pathogens in spring, the propagation frequency of forest rot pathogens in summer, the propagation frequency of forest rot pathogens in autumn, and the propagation frequency of forest rot pathogens throughout the year; S5, against A max =(x1+x2+x3+x4+x5) T Normalization processing was performed to calculate the weights corresponding to the suitability index, the frequency of propagation of forest rot pathogens in spring, the frequency of propagation of forest rot pathogens in summer, the frequency of propagation of forest rot pathogens in autumn, and the frequency of propagation of forest rot pathogens throughout the year: Among them, w c is the weight, c = 1, 2, 3, 4, 5, corresponding to x1, x2, x3, x4, x5, that is, w1, w2, w3, w4, w5 are the weights of the suitability index, the frequency of propagation of forest rot pathogens in spring, the frequency of propagation of forest rot pathogens in summer, the frequency of propagation of forest rot pathogens in autumn, and the frequency of propagation of forest rot pathogens throughout the year; S6. Pass The formula calculates the consistency index, where CI is the consistency index, c = 1, 2, 3, 4, 5, λ max is the maximum eigenvalue of the judgment matrix. The actual random consistency index value is obtained by querying the random consistency index table, which is represented by RI. The consistency ratio is calculated, and when the consistency ratio is less than 0.1, the matrix is ​​judged to be reasonable; S7. Use weighted overlay to calculate the initial risk index corresponding to each grid cell: F c =x c1 ×w1+x c2 ×w2+x c3 ×w3+x c4 ×w4+x c5 ×w5 Among them, F c is the initial risk index corresponding to the cth row of the judgment matrix; w1, w2, w3, w4, w5 are the weights of the suitability index, the frequency of propagation of forest rot pathogens in spring, the frequency of propagation of forest rot pathogens in summer, the frequency of propagation of forest rot pathogens in autumn, and the frequency of propagation of forest rot pathogens throughout the year, respectively; x c1 , x c2 , x c3 , x c4 , x c5 They are the elements corresponding to each row of the judgment matrix, c = 1, 2, 3, 4, 5; S8. Use GIS software to assign the calculated initial risk index to the corresponding grid to obtain the initial risk map of forest wood decay pathogens.

8. The C. chrysosperma comprehensive risk intelligent assessment system based on multi-source data fusion according to claim 7, characterized in that: The process of integrating the suitability index and the frequency of tree decay pathogen transmission by the factor integration unit through the random forest model includes: A sampling layer was created in ArcGIS, sampling points were randomly generated, and a table was created. Each row in the table represented a grid unit, and each column of data included the occurrence probability of the suitability index, the propagation frequency of the forest rot pathogen in spring, the propagation frequency of the forest rot pathogen in summer, the propagation frequency of the forest rot pathogen in autumn, the propagation frequency of the forest rot pathogen throughout the year, and the initial risk index. Each column of data was used as input, and the comprehensive risk index was used as the target variable. Using the random forest algorithm, the occurrence probability of the suitability index, the propagation frequency of the forest rot pathogen in spring, the propagation frequency of the forest rot pathogen in summer, the propagation frequency of the forest rot pathogen in autumn, the propagation frequency of the forest rot pathogen throughout the year, and the initial risk index were used as data sets, 80% of the data sets were divided into training sets, and 20% of the data sets were divided into test sets; The random forest model was trained using the data set to learn the relationship between the occurrence probability of the suitability index, the propagation frequency of the forest rot pathogens in spring, summer, autumn, and the propagation frequency of the forest rot pathogens throughout the year and the initial risk index, as well as their influence on the comprehensive risk index, to obtain the random forest model; the test set was combined with the root mean square error, square root error, and determination coefficient to evaluate the performance of the trained random forest model, adjust the model parameters, and obtain the comprehensive risk assessment model; Use the comprehensive risk assessment model to predict each grid unit, obtain the comprehensive risk index of each grid unit, generate a comprehensive risk map, and divide the risk level according to the comprehensive risk index. When the comprehensive risk index is between 0 and 0.3, it is classified as a low risk level; when the comprehensive risk index is between 0.3 and 0.6, it is classified as a medium risk level; when the comprehensive risk index is between 0.6 and 1, it is classified as a high risk level.

9. The C. chrysosperma comprehensive risk intelligent assessment system based on multi-source data fusion according to claim 8, characterized in that: The risk assessment and early warning module combines the niche model and the HYSPLIT model to analyze the multimodal data of forest rot pathogens, evaluate the risk level of forest rot pathogens under multiple influencing factors, and perform early warning operations, including: Combining the niche model and the HYSPLIT model, the multimodal data of tree rot pathogens were analyzed to obtain the comprehensive risk index of each grid unit. The tree rot pathogens with a comprehensive risk index between 0 and 0.3 were at a low risk level; the tree rot pathogens with a comprehensive risk index between 0.3 and 0.6 were at a medium risk level; and the tree rot pathogens with a comprehensive risk index between 0.6 and 1 were at a high risk level. When the risk level of wood decay pathogens is low, the monitoring frequency will be increased and a low-risk warning signal will be issued; when the risk level of wood decay pathogens is medium, the monitoring frequency will be increased for areas corresponding to grid cells with high comprehensive risk indexes, and a warning signal will be issued that the areas corresponding to grid cells with high comprehensive risk indexes are medium-risk; when the risk level of wood decay pathogens is high, all areas will be intensively monitored and a high-risk warning signal will be issued.

10. A C. chrysosperma comprehensive risk intelligent assessment method based on multi-source data fusion, which is implemented based on the C. chrysosperma comprehensive risk intelligent assessment system based on multi-source data fusion as described in any one of claims 1 to 9, characterized in that: It consists of the following steps: Step 1: combining database retrieval, data entry, data download and meteorological data assimilation to collect multimodal data of forest rot pathogens, wherein the multimodal data of forest rot pathogens includes distribution data, environmental data and high-altitude meteorological data of forest rot pathogens; Step 2: Remove redundancy from the distribution data of the tree rot pathogens, perform dimensionality reduction on the environmental data of the tree rot pathogens, optimize the niche model parameters through the Kuenm package, and construct the niche model; Step 3: Determine the release source of the spores of the tree rot pathogen, build the HYSPLIT model, and calculate the propagation trajectory and frequency of the spores of the tree rot pathogen; Step 4: Obtain the initial risk index corresponding to each grid unit through eigenvalue decomposition method, normalization processing, consistency index calculation and weighted overlay calculation. Combined with GIS software, assign the calculated initial risk index to the corresponding grid to obtain the initial risk map of forest wood decay pathogens. Step 5: Integrate the suitability index and the frequency of tree decay pathogen transmission through the random forest model; Step 6. Combine the niche model and the HYSPLIT model to analyze the multimodal data of wood decay pathogens, evaluate the risk level of wood decay pathogens under multiple influencing factors, and when the risk level of wood decay pathogens is low, strengthen the monitoring frequency and issue a low-risk warning signal; when the risk level of wood decay pathogens is medium, increase the monitoring frequency for areas corresponding to grid cells with high comprehensive risk indexes, and issue a medium-risk warning signal for areas corresponding to grid cells with high comprehensive risk indexes; when the risk level of wood decay pathogens is high, conduct intensive monitoring of all areas and issue a high-risk warning signal.

Images