Space-time dynamic evaluation method and system for habitat suitability of rice sheath blight disease

Abstract

The invention discloses a spatio-temporal dynamic evaluation method and system for habitat suitability of rice sheath blight disease. The spatial-temporal heterogeneity characterization of sensitive key habitat factors of the rice sheath blight disease is realized by integrating multi-source time sequence information such as remote sensing, weather and plant protection. And on the basis, a rice sheath blight habitat suitability evaluation model is constructed by taking a month as a time unit, so that space-time refined space-time dynamic evaluation of the rice sheath blight is realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of evaluation of the habitat suitability of plant diseases, and relates to a spatio-temporal dynamic evaluation method for the habitat suitability of diseases that integrates multi-source temporal information such as remote sensing, meteorology, and plant protection. Specifically, it relates to a spatio-temporal dynamic evaluation method and system for the habitat suitability of rice sheath blight disease. Background Art

[0002] Currently, the prevention and control of plant diseases and insect pests is still dominated by the trend prediction model that relies on meteorology and plant protection. There are problems such as over-expanding the prevention and control area and overusing pesticides, resulting in a continuous increase in the input of chemical pesticides. This not only increases the prevention and control cost but also easily causes agricultural non-point source pollution and threatens food safety. Evaluating the habitat suitability of plant diseases and insect pests in a region is the key to achieving accurate prediction of crop diseases and insect pests on a large scale, and can provide important background information for the prediction and prevention and control of crop diseases and insect pests on a large scale.

[0003] As a soil-borne disease, the rice sheath blight pathogen mainly relies on irrigation water flow to carry soil sclerotia for transmission. Therefore, the occurrence and prevalence of sheath blight are closely related to landscape patterns such as farmland ditch networks. At the same time, the planting pattern of mixed single and double cropping of rice further increases the spatio-temporal complexity of the prevalence and transmission of sheath blight. However, at present, the evaluation of the habitat suitability of crop diseases and insect pests in a region still mainly relies on meteorological and plant protection information, and it is difficult to effectively characterize the impact of complex habitat spatio-temporal heterogeneity, which greatly limits the effectiveness of the evaluation. In recent years, the deep integration of multi-disciplinary technical means such as space remote sensing and computer science has brought valuable opportunities for the innovative development of the technology for evaluating the habitat suitability of crop diseases and insect pests.

[0004] The present invention fully considers the prevalence and transmission characteristics of rice sheath blight, integrates multi-source temporal information such as remote sensing, meteorology, and plant protection, and extracts the habitat factors of rice sheath blight with a month as the time unit. The spatio-temporal causal inference theory GCCM method is introduced to characterize the spatio-temporal heterogeneity of the key habitat factors sensitive to rice sheath blight. On this basis, using the classical species distribution model MaxEnt, a spatio-temporal fine spatio-temporal dynamic evaluation model for the habitat suitability of rice sheath blight is constructed, realizing the evaluation of the habitat suitability of rice sheath blight at the monthly scale, overcoming the technical bottleneck of insufficient characterization of spatio-temporal heterogeneity in traditional evaluation methods for the habitat suitability of plant diseases and insect pests, improving the level of regional crop disease habitat suitability evaluation, and providing reliable background information for regional disease early warning modeling, etc. Summary of the Invention

[0005] The object of the present invention is to provide a spatio-temporal dynamic evaluation method for the habitat suitability of regional diseases that fully considers the epidemic transmission characteristics of rice sheath blight, aiming at the deficiency in the prior art that it is difficult to effectively reflect the complex influence of the spatio-temporal heterogeneity of the habitats of pests and diseases in the region. This includes the construction of a habitat suitability evaluation model that fully considers the influence of the spatio-temporal heterogeneity of the habitats of rice sheath blight in the mixed cropping mode of single-season and double-season rice. By comprehensively integrating multi-source temporal information such as remote sensing, meteorology, and plant protection, the method realizes the spatio-temporal heterogeneity characterization of the key habitat factors sensitive to rice sheath blight. On this basis, with a month as the time unit, a habitat suitability evaluation model for rice sheath blight is constructed, so as to realize the spatio-temporal refined spatio-temporal dynamic evaluation of rice sheath blight.

[0006] The technical solutions adopted by the present invention to solve the technical problems are specifically as follows:

[0007] In the first aspect, an embodiment of the present application provides a spatio-temporal dynamic evaluation method for the habitat suitability of rice sheath blight, which includes the following steps:

[0008] Extract the habitat factors of rice sheath blight at the monthly scale by integrating multi-source temporal information.

[0009] Obtain the remote sensing, meteorological, and plant protection temporal data of the area to be evaluated over the years, calculate the disease habitat factors at different spatio-temporal resolution scales based on this, and resample each habitat factor to form a temporal disease habitat factor with a unified spatial resolution.

[0010] Divide the temporal disease habitat factors on a monthly basis, calculate the mean value of the same disease habitat factor at multiple time phases within the same month, and perform masking processing in combination with the monthly rice planting spatial distribution data to obtain the disease habitat factors for each month of each year under the constraint of the rice planting spatial distribution.

[0011] For the disease habitat factors for each month of each year under the constraint of the rice planting spatial distribution obtained over multiple years, calculate the mean value according to the disease habitat factors of the same month in different years, and complete the extraction of the habitat factors of rice sheath blight at the monthly scale by integrating multi-source temporal information.

[0012] Optimize the habitat factors based on spatial causal inference to realize the spatio-temporal heterogeneity characterization of the key sensitive habitat factors of rice sheath blight.

[0013] Analyze the causal relationship between each habitat factor at the monthly scale and the disease plant rate of rice sheath blight in the current month through spatial causal inference, and use whether the causal effect index is greater than the set threshold to complete the preliminary screening of the key sensitive habitat factors affecting the epidemic transmission of rice sheath blight, and obtain the preliminary screening habitat factors.

[0014] Further utilize spatial causal inference to analyze the possible causal relationships between the initially screened habitat factors pairwise, draw a causal-directed relationship diagram among the sheath blight habitat factors, and combine the causal-directed relationship diagram to optimize the selection of the key sensitive habitat factors of rice sheath blight for each month.

[0015] Based on the species distribution model, construct and optimize a spatio-temporal dynamic evaluation model for the habitat suitability of rice sheath blight.

[0016] Respectively take the key sensitive habitat factors of rice sheath blight for each month selected as the input of the species distribution model MaxEnt, and construct a spatio-temporal dynamic evaluation model for the habitat suitability of rice sheath blight (GCCM-MaxEnt).

[0017] Optimize the parameters of the evaluation model by combining the corrected AIC information criterion and the evaluation index AUC, and form a group of monthly-scale habitat suitability evaluation models for rice sheath blight optimized by month.

[0018] Utilize the group of monthly-scale habitat suitability evaluation models for rice sheath blight to obtain the monthly habitat suitability evaluation results of rice sheath blight during the key period of rice in the target area.

[0019] Preferably, the monthly rice planting spatial distribution data is obtained based on the phenological period information of single-season and double-season rice: First, based on the transplanting date and harvesting date corresponding to the same spatial distribution of each rice cropping type, obtain the rice growth time range under the same spatial distribution of each rice cropping type; On this basis, use the corresponding time range of each month as the threshold to divide the rice growth time range, and obtain the monthly rice planting spatial distribution data.

[0020] Preferably, the diseased plant rate data of rice sheath blight for each month is calculated from the plant protection time series data: First, select representative fields with sheath blight occurrence as the survey points, obtain the corresponding plant protection time series data, and calculate the time series diseased plant rate data of sheath blight at each survey point through the ratio of the number of diseased plants to the total number of surveyed plants; Second, divide the time series data of the diseased plant rate of sheath blight at each survey point every year by month, and take the maximum diseased plant rate value within each month as the diseased plant rate of sheath blight at that survey point in that month; Finally, calculate the average value of the diseased plant rate of sheath blight in the same month in all years to obtain the diseased plant rate data of rice sheath blight for each month.

[0021] Preferably, the causal effect index reflects the relationship between the true observed value and the spatially causal inference predicted value through the covariance of the standardized true observed value and the spatially causal inference predicted value. The initial screening of habitat factors is achieved through a pre-set causal effect index threshold.

[0022] Preferably, the causal orientation relationship diagram among the sheath blight habitat factors is drawn as follows: First, the causal relationship between pairwise sheath blight habitat factors is quantified using spatial causal inference to obtain the causal effect index between pairwise habitat factors; based on the calculated causal effect index, the causal relationship is defined into a directed acyclic graph using a directed acyclic graph tool, and an interactive and adjustable directed acyclic graph is obtained using an interactive network visualization tool; finally, a causal orientation relationship diagram among the sheath blight habitat factors is drawn using a graphic grammar drawing tool.

[0023] Preferably, the optimization of the evaluation model parameters is achieved by combining the corrected AIC information criterion and the evaluation index AUC, specifically as follows: First, set the RM value range and step size, determine the RM gradient experimental group to quantitatively analyze the regulation effect of parameter adjustment on the model complexity; then select the RM with the smallest AICc value corresponding to the model complexity under different RMs within the RM value range; finally, compare the model AUC performance between the optimized RM and the default RM parameter (RM = 1) of the species distribution model, and select the RM value corresponding to the largest AUC value as the optimal RM of the model to complete the model optimization and establish a parameter optimization scheme suitable for the monthly-scale evaluation model.

[0024] In a second aspect, an embodiment of the present application provides a spatio-temporal dynamic evaluation system for the habitat suitability of rice sheath blight, which includes the following modules:

[0025] Habitat factor extraction module: Extract monthly-scale rice sheath blight habitat factors by integrating multi-source time-series information.

[0026] Habitat factor optimization module: Optimize key sensitive habitat factors based on spatial causal inference.

[0027] Evaluation model group acquisition module: Construct and optimize a spatio-temporal dynamic evaluation model for the habitat suitability of rice sheath blight based on the species distribution model.

[0028] Evaluation module: Input the monthly rice sheath blight-sensitive key habitat factors of the target area into the spatio-temporal dynamic evaluation model for the habitat suitability of rice sheath blight for each month to obtain the evaluation results of the habitat suitability of rice sheath blight in the target area for each month, that is, the spatial distribution range map of the potential distribution probability of rice sheath blight for each month.

[0029] In a third aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored, and when the computer-executable instructions are executed by a processor, they are used to implement the implementation manner of the first aspect as described above.

[0030] In a fourth aspect, an embodiment of the present application provides a computer program product, including a computer program, which when executed by a processor, implements the implementation manner of the first aspect as described above.

[0031] The present invention has the following beneficial effects:

[0032] 1. The method for characterizing the spatio-temporal heterogeneity of habitat factors in disease-sensitive areas coupling the GCCM causal inference theory of the present invention explores the influence mechanism of spatio-temporal heterogeneity of habitats on the epidemic spread of pests and diseases by integrating multi-source data such as meteorology, remote sensing, and plant protection. Compared with traditional methods, it can analyze the impact of habitat differences on crop diseases at a finer spatio-temporal scale, and at the same time can effectively improve the interpretability and reliability of the selected habitat factors.

[0033] 2. The present invention establishes a modeling strategy for spatio-temporal dynamic evaluation of habitat suitability for crop diseases by coupling the causal inference theory GCCM with the species distribution model MaxEnt. Through monthly-scale modeling, it can finely respond to the spatio-temporal process characteristics of the occurrence and epidemic of diseases within a year. Compared with existing evaluation methods, it has stronger mechanism and can effectively improve the effectiveness and reliability of habitat suitability evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is the technical roadmap of the embodiment of the present invention.

[0035] Figure 2 is the causal-directed relationship diagram among habitat factors of rice sheath blight in each month of the embodiment of the present invention.

[0036] Figure 3 is the accuracy comparison diagram of the habitat suitability evaluation model of rice sheath blight in each month of the embodiment of the present invention.

[0037] Figure 4a , Figure 4b , Figure 4c , Figure 4d , Figure 4e , Figure 4f and Figure 4g are the mapping of the evaluation results of habitat suitability of rice sheath blight in each month in 13 provinces of China in the embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0038] The present invention will be further explained below with reference to the accompanying drawings:

[0039] A spatio-temporal dynamic evaluation method for habitat suitability of rice sheath blight, which includes the following steps:

[0040] Step 1: Extract monthly-scale habitat factors of rice sheath blight by integrating multi-source time-series information.

[0041] 1-1. Obtain the remote sensing, meteorological, and plant protection time-series data of the area to be evaluated over the years, calculate and obtain the disease habitat factors at different spatio-temporal resolution scales, and resample each habitat factor to form a unified spatial resolution time-series disease habitat factor.

[0042] 1-2. Divide the temporal habitat factors on a monthly basis, calculate the mean value of the same habitat factor for multiple time phases within the same month, and perform masking processing in combination with the monthly rice planting spatial distribution data to obtain the disease habitat factors for each month of each year under the constraint of the rice planting spatial distribution.

[0043] In one embodiment, the monthly rice planting spatial distribution data is obtained based on the phenological period information of single-season and double-season rice: First, based on the transplanting period and harvesting period dates corresponding to the same spatial distribution of each rice cropping type, obtain the rice growth time range under the same spatial distribution of each rice cropping type; On this basis, use the corresponding time range of each month as a threshold to divide the rice growth time range to obtain the monthly rice planting spatial distribution data.

[0044] 1-3. Calculate the mean value of the disease habitat factors for each month of each year under the constraint of the rice planting spatial distribution for multiple years according to the disease habitat factors of the same month in different years, and complete the extraction of the habitat factors of rice sheath blight at the monthly scale by integrating multi-source temporal information.

[0045] Step two. Optimize the habitat factors based on the spatial causal inference GCCM to realize the spatio-temporal heterogeneity characterization of the key sensitive habitat factors of rice sheath blight.

[0046] 2-1. Analyze the causal relationship between each habitat factor at the monthly scale and the diseased plant rate of rice sheath blight in the current month through GCCM, and complete the preliminary screening of the key sensitive habitat factors affecting the epidemic spread of rice sheath blight based on whether the causal effect index is greater than the set threshold to obtain the preliminary screened habitat factors.

[0047] In one embodiment, the diseased plant rate data of rice sheath blight for each month is calculated from the plant protection temporal data: First, select representative fields with the occurrence of sheath blight as the survey points, and according to the national standard "GBT 15791-2011 Technical Specification for Forecasting Rice Sheath Blight", obtain the occurrence and epidemic situation of rice sheath blight at each survey point for multiple consecutive years, record the total number of surveyed plants and the number of diseased plants with rice sheath blight, and conduct 1 survey at each survey point at intervals during the critical period of rice every year to obtain the plant protection temporal data, and calculate the temporal diseased plant rate data of rice sheath blight at each survey point through the ratio of the number of diseased plants to the total number of surveyed plants; Second, divide the temporal diseased plant rate data of rice sheath blight at each survey point by month, and take the maximum diseased plant rate value within each month as the diseased plant rate of rice sheath blight at that survey point in that month; Finally, obtain the diseased plant rate data of rice sheath blight for each month by calculating the average value of the diseased plant rates of rice sheath blight in the same month in all years.

[0048] In one embodiment, the causal effect index is ρ, which reflects the relationship between the true observed value and the GCCM predicted value through the covariance between the standardized true observed value and the GCCM predicted value. The preliminary screening of habitat factors is achieved through a preset ρ value threshold.

[0049] 2-2. Further use GCCM to analyze the possible causal relationships between the preliminarily screened habitat factors pairwise, draw the causal directed relationship diagram among the sheath blight habitat factors, and combine the causal directed relationship diagram to realize the optimization of the key sensitive habitat factors of rice sheath blight for each month.

[0050] In one embodiment, the drawing direction of the causal directed relationship diagram among the sheath blight habitat factors is as follows: First, use GCCM to quantify the causal relationships between the sheath blight habitat factors pairwise to obtain the ρ values of the causal effect indexes between the habitat factors pairwise; based on the calculated ρ values, use the directed acyclic graph tool dagitty to define the causal relationships into the directed acyclic graph DAG, and use the interactive network visualization tool visNetwork to obtain an interactive and adjustable DAG; finally, use the graphic grammar drawing tools ggdag and ggplot2 to draw the causal directed relationship diagram among the sheath blight habitat factors.

[0051] Step three: Construct and optimize the spatio-temporal dynamic evaluation model of rice sheath blight habitat suitability based on the species distribution model.

[0052] 3-1. Respectively use the key sensitive habitat factors of rice sheath blight for each month optimized by GCCM as the input of the species distribution model MaxEnt to construct the spatio-temporal dynamic evaluation model of rice sheath blight habitat suitability (GCCM-MaxEnt).

[0053] 3-2. Achieve the optimization of the evaluation model parameters by combining the corrected AIC information criterion and the evaluation index AUC, and form a monthly-scale group of rice sheath blight habitat suitability evaluation models optimized by month.

[0054] In one embodiment, the optimization strategy of the regularization coefficient RM of the GCCM-MaxEnt model that combines the corrected AIC information criterion and the area under the curve AUC of the evaluation index is specifically as follows: First, set the RM value range and step size, determine the RM gradient experimental group to quantitatively analyze the regulatory effect of parameter adjustment on the model complexity; then select the RM with the smallest AICc value of the corresponding model complexity under different RMs within the RM value range; finally, compare the model AUC performance between the optimized RM and the MaxEnt model default RM parameter (RM = 1), and select the RM value corresponding to the largest AUC value as the optimal RM of the model to complete the model optimization and establish the parameter optimization scheme suitable for the monthly-scale evaluation model.

[0055] Step 4: Input the monthly key habitat factors of rice sheath blight in the target area into the spatio-temporal dynamic evaluation model of the habitat suitability of rice sheath blight for each month to obtain the evaluation results of the habitat suitability of rice sheath blight in the target area for each month, that is, the spatial distribution range map of the potential distribution probability of rice sheath blight in each month.

[0056] As Figure 1 shown, a spatio-temporal dynamic evaluation method for the habitat suitability of rice sheath blight that integrates multi-source temporal information is as follows:

[0057] Step 1: Integrate multi-source information to obtain spatio-temporally continuous habitat factors of rice sheath blight and realize the extraction of habitat factors of rice sheath blight at the monthly scale.

[0058] 1-1 Obtain multi-source temporal data, specifically including remote sensing, meteorological, and plant protection data. The time range is from 2008 to 2015. Due to the lack of plant protection survey data in 2012, all data in 2012 were removed in subsequent studies. The target area is located in the middle and lower reaches of the Yangtze River to the southwest rice-growing area, including 13 provinces of Jiangsu, Zhejiang, Anhui, Fujian, Jiangxi, Hubei, Hunan, Guangdong, Guangxi, Sichuan, Guizhou, Yunnan, and Chongqing.

[0059] 1-2 The habitat factors characterizing the growth status of rice include the normalized difference vegetation index (NDVI), enhanced vegetation index (EVI), leaf area index (LAI), fractional vegetation cover (FVC), fraction of photosynthetically active radiation absorbed by vegetation (FPAR), and net primary productivity (NPP), all of which are standardized MODIS vegetation products. The habitat factors characterizing environmental meteorological parameters include the maximum temperature (TMX), minimum temperature (TMN), average temperature (TMP), sunshine hours (NSUN), relative humidity (RH), precipitation (PRE), land surface temperature (LST), and elevation information (DEM). Among them, the meteorological data are from the National Tibetan Plateau Scientific Data Center, the National Earth System Science Data Center, and the China Surface Climate Data Daily Value Dataset (V3.0). The land surface temperature (LST) is from the standardized MODIS product, and the elevation information (DEM) is from the Data Center for Resources and Environmental Sciences, Chinese Academy of Sciences. The habitat factors characterizing the spatial landscape pattern state include patch density (PD), landscape shape index (LSI), edge density (ED), patch richness density (PRD), Shannon diversity index (SHDI), Simpson diversity index (SIDI), largest patch index (LPI), and mean fractal dimension (FRAC_MN), all of which are based on the 30m resolution land use product CLCD and are obtained by calculating through the landscape spatial pattern quantitative analysis software FRAGSTATS.

[0060] 1-3 For the habitat factors with different temporal and spatial resolutions obtained for each year, during the time period of April to October, which is the sensitive period for sheath blight occurrence, the average value of each habitat factor is calculated monthly for temporal monthly-scale standardization, and spatially, they are uniformly resampled to a resolution of 1 km for standardization, obtaining multi-dimensional habitat factors for each month of each year with spatio-temporal standardization. On this basis, further average the habitat factors of the same month from 2008 to 2015 (excluding 2012) to obtain the monthly-scale habitat factors of rice sheath blight in the target area.

[0061] 1-4 Based on the phenological time information of single and double cropping rice in the ChinaCropArea1km dataset, within the time range of April to October, using the time of each month as the phenological threshold range, obtain the spatial distribution data of rice planting in the target area each month, and perform masking on the obtained monthly-scale habitat factors of rice sheath blight, finally realizing the extraction of monthly-scale habitat factors of rice sheath blight under the constraint of the spatial distribution of rice planting.

[0062] The spatial distribution data of rice planting is obtained based on the phenological information of single and double cropping rice: First, based on the transplanting date and harvesting date corresponding to the same spatial distribution of each rice cropping type, obtain the rice growth time range under the same spatial distribution of each rice cropping type; on this basis, use the corresponding time range of each month as the threshold to divide the rice growth time range, and obtain the spatial distribution data of rice planting each month.

[0063] Step 2: Introduce the spatial causal inference GCCM to optimize the habitat factors, and realize the spatio-temporal heterogeneity characterization of the key sensitive habitat factors of rice sheath blight.

[0064] 2-1 The plant protection survey data of rice sheath blight are the diseased plant rate data every 5 - 7 days from 2008 to 2015 (excluding 2012). The calculation formula of the diseased plant rate DI is as follows:

[0065]

[0066] Among them, M is the total number of plants in the plant protection data, and M disease is the number of diseased plants.

[0067] For the diseased plant rate data of rice sheath blight for each month of each year (April to October), divide the plant protection survey data of the diseased plant rate of sheath blight every 5 - 7 days at the same plant protection site by month, and take the maximum value of the diseased plant rate within each month as the diseased plant rate of sheath blight at this site for this month; calculate the 7-year average diseased plant rate of the same month from 2008 to 2015 (excluding 2012), and use it as the diseased plant rate of each month of rice sheath blight finally used for causal inference analysis.

[0068] 2-2 Considering the characteristics of the plant protection survey data area / county, the monthly scale rice sheath blight habitat factors constrained by the spatial distribution of rice planting were obtained, and the districts and counties where the rice sheath blight survey and plant protection stations were located were taken as units. The mean values of the habitat factors in each district and county were counted to match the rice sheath blight plant protection survey while ensuring the representativeness of the habitat factors. Finally, a data set of key habitat factors for sheath blight that was suitable for GCCM method analysis and matched with the sheath blight plant protection survey data was formed.

[0069] 2-3 Use GCCM to quantify the causal relationship between the rice sheath blight disease rate and various habitat factors at the monthly scale. Based on the generalized embedding theorem, GCCM proposes to reconstruct the state space based on spatial lag, and realize causal inference based on the cross-mapping prediction of the state space. It can identify the causal direction and estimate the causal effect in weakly coupled relationships, and identify the dominant causal direction and estimate the causal effect in strongly coupled relationships. The generalized embedding theorem is a theory about the mapping between differentiable manifolds M and Euclidean space R, that is, M can be mapped to R through observation functions, and points in R can be used to construct M, so that a state space containing spatial cross-sectional data can be constructed. Specifically, GCCM defines the prediction based on the nearest mutual neighbors as a cross-mapping prediction, and the calculation formula is as follows:

[0070]

[0071] Among them, S is the spatial unit of the variable Y that needs to be predicted (rate of sheath blight diseased plants / habitat factor); is the shadow manifold based on the habitat factor x The predicted value of is the spatial unit used for prediction; For space unit Observation value; L is the embedding dimension; For space unit The corresponding weights are determined by an exponential decay function:

[0072]

[0073] in is the weight function between two states in the shadow manifold, and its definition formula is as follows:

[0074]

[0075] Where exp is the exponential function, dis(*,*) is the distance function between two states in the defined shadow manifold, and its definition formula is as follows

[0076]

[0077] Where |*| represents the absolute value, abs[*,*] represents the distance function between two vectors, The first element h in si (x) corresponds to the spatial focus unit, while the other elements in correspond to vectors with multiple spatial units respectively. For raster data, since the number and position of neighbors in a certain order are fixed, abs[*,*] is defined as follows:

[0078]

[0079] where is the spatial unit of the k-th order spatial lag of si in the direction d, and D is the number of spatial units at the k-th order.

[0080] Finally, the ability of GCCM cross-map prediction is measured by the Pearson correlation coefficient between the true observations and the corresponding predicted values, that is, the causal effect index ρ:

[0081]

[0082] where, represents the covariance between the observed values and the GCCM predicted values, and represent the variances of the observed values and the GCCM predicted values respectively.

[0083] In the embodiment, ρ > 0.2 is used as the threshold to preliminarily screen the sheath blight sensitive habitat factors, and the sensitive habitat factors for each month from April to October are as shown in Table 1 below: Table 1: Sensitive habitat factors for each month from April to October

[0084] Month Sheath blight sensitive habitat factors April LSI, PRD, TMX, PRE, RH, NDVI, LST May LSI, PRD, TMX, PRE, RH, LST, NSUN June LSI, TMP, PRE, RH, NSUN, LST, LAI, NDVI, FPAR, EVI, NPP July LSI, TMP, PRE, RH, NSUN, LST, LAI, NDVI, FPAR, EVI, NPP August LSI, TMP, PRE, RH, NSUN, LST, LAI, NDVI, FPAR, EVI, NPP September LSI, TMX, PRE, RH, NSUN, LST, NDVI, EVI, FPAR October LSI, TMX, PRE, RH, NSUN, LST, NDVI, EVI, FPAR

[0085] 2 - 4 To reduce the redundant information between factors, the causal relationships between the sheath blight sensitive habitat factors obtained by preliminary screening are further analyzed by GCCM, and the causal orientation relationship diagram between the key sheath blight habitat factors is drawn ( Figure 2 ): Based on the calculated causal effect index ρ values of pairwise quantification of sheath blight habitat factors using GCCM, the causal relationships are defined into a directed acyclic graph DAG using dagitty, and an interactive and adjustable DAG is obtained using visNetwork. Finally, the causal orientation relationship diagram between sheath blight habitat factors is drawn using ggdag and ggplot2. On this basis, combined with the causal orientation relationship diagram, for each causal relationship chain in the relationship diagram, the source habitat factors among all the habitat factors with an association relationship are selected to realize the optimization of the key sheath blight habitat factors for rice in each month. The finally obtained spatio-temporal heterogeneity key sensitive habitat factors for rice sheath blight are as shown in Table 2 below: Table 2: Spatio-temporal heterogeneity key sensitive habitat factors for rice sheath blight

[0086] Month Sheath blight sensitive key habitat factors April LSI, PRD, TMX, PRE, NDVI May LSI, PRD, PRE, TMX, NSUN June - August LSI, TMP, PRE, NSUN, LAI September - October LSI, TMX, PRE, NSUN, EVI

[0087] Step 3: Respectively take the key habitat factors of rice sheath blight in each month optimized by GCCM as the input of the species distribution model MaxEnt. Based on the species distribution model MaxEnt, construct the spatio-temporal dynamic evaluation model GCCM-MaxEnt for the habitat suitability of rice sheath blight, and optimize the model parameters by calibrating the AIC information criterion to form a group of monthly-scale evaluation models for the habitat suitability of rice sheath blight corresponding to April - October.

[0088] 3-1 The MaxEnt model is a probability model based on the principle of maximum entropy. It establishes a non-random relationship between the spatial distribution data of diseases and the habitat factor layer through ecological principles and outputs the potential distribution probability of diseases, thereby constructing a disease habitat suitability evaluation model with strong mechanism. MaxEnt maximizes the conditional probability distribution under the constraint of entropy value:

[0089]

[0090]

[0091] where characteristic function is the weight of the characteristic function.

[0092] 3-2 Use the ENMeval toolkit to optimize the regularization coefficient RM of the MaxEnt model in combination with the calibrated AIC information criterion: set an RM gradient experimental group with a range of 0.5 - 4.0 and a step size of 0.5, quantitatively analyze the regulatory effect of parameter adjustment on the model complexity; select the RM with the minimum AICc value of the corresponding model complexity under different RMs in the range of 0.5 - 4.0; compare the AUC performance of the optimized RM and the default RM parameter (RM = 1) of the MaxEnt model to reveal the marginal effect of model complexity on the prediction accuracy of species distribution, and establish a parameter optimization scheme suitable for the monthly-scale evaluation model.

[0093] AUC is the area under the curve, which characterizes the discrimination ability of the model for the presence / false absence samples of pests and diseases, and its value range is between 0 and 1. The evaluation criteria are: AUC < 0.7 indicates that the model accuracy is insufficient, 0.7 ≤ AUC < 0.9 is an effective prediction model, and AUC ≥ 0.9 is determined to be a high-precision model. The AUC calculation formula is as follows:

[0094]

[0095] where TPR: true positive rate; FPR: false positive rate.

[0096] After optimizing the RM parameters, the optimal RM parameters in the MaxEnt model of the monthly-scale rice sheath blight habitat suitability evaluation model from April to October are 1.5, 0.5, 3.0, 3.5, 0.5, 1.0, and 1.5 in sequence. After optimizing the RM parameters, the AUC values of the rice sheath blight habitat suitability evaluation model from April to October are 87.1%, 85.8%, 83.1%, 80.1%, 81.1%, 82.6%, and 86.2% respectively. Except for the unchanged AUC value in September, the AUC values of the evaluation model after optimizing the RM parameters are increased by 1.8%, 1.6%, 2.1%, 2.7%, 4.3%, and 1.9% compared with the default parameters in sequence ( Figure 3 ).

[0097] Step 4: Input the sensitive key habitat factors of rice sheath blight in each month of the target area into the monthly habitat suitability evaluation model GCCM-MaxEnt respectively to obtain the evaluation results of the habitat suitability of rice sheath blight in each month from April to October in the target area, that is, the spatial distribution range map of the potential distribution probability of rice sheath blight in each month.

[0098] By comparing the habitat suitability result maps of rice sheath blight in each month, it is found that the habitat suitability area of rice sheath blight in the whole year shows the characteristics of expanding from south to north over time c and then gradually retreating southward ( Figure 3 ).

[0099] The permutation importance is used to quantify the marginal contribution of each key habitat factor to the evaluation model. It is found that the dominant driving habitat factors of the rice sheath blight habitat suitability evaluation model in April are TMX and PRE in sequence, TMX in May, TMP in June - August, and TMX in September - October.

[0100] Furthermore, the ecological response curves are drawn using the built-in plotting module of the MaxEnt model software: with the standardized habitat factors as the abscissa and the occurrence risk probability of sheath blight as the ordinate, the "Cloglog" plotting module built in the MaxEnt software is used to draw the ecological response curves of the habitat suitability probability of sheath blight under different values of the dominant driving habitat factors in each month. Among them, the value range of the habitat suitability probability of sheath blight is 0 - 1. On this basis, the non - linear interaction mechanism between the dominant driving habitat factors in each month and the habitat suitability of sheath blight is analyzed in combination with the ecological response curves. It is found that the dominant driving habitat factor PRE in April reaches the highest point at 221.6 mm, the dominant driving habitat factor TMX in May reaches the highest point at 28.6 °C, the dominant driving habitat factor TMP in June - August reaches the highest point at 29.5 °C, 29.5 °C, and 28.4 °C respectively, and the dominant driving habitat factor TMX in September - October reaches the highest point at 32.0 °C and 30.6 °C respectively.

[0101] Compared with the traditional disease habitat suitability evaluation method that relies on meteorological plant protection information, expert experience, and statistical methods to select habitat factors, the spatio-temporal dynamic evaluation method for the habitat suitability of rice sheath blight proposed in this method fully considers the impact of the single- and double-cropping rice mixed cropping pattern on the epidemic spread of rice sheath blight, can analyze the impact of habitat differences on crop diseases at a finer spatio-temporal scale, the introduction of the spatial causal inference theory can effectively improve the interpretability and reliability of habitat factors, and at the same time, using the month as the time unit to establish a model for the evaluation of disease habitat suitability can finely respond to the spatio-temporal process characteristics of the occurrence and epidemic of diseases within a year, has stronger mechanism, improves the effectiveness and reliability of habitat suitability evaluation, provides new ideas and methods for the evaluation of the habitat suitability of rice sheath blight, provides important background information for the forecast and prevention and control of crop pests and diseases on a large scale, and has application potential.

Claims

1. A spatio-temporal dynamic evaluation method for the habitat suitability of rice sheath blight, characterized in that Including the following steps: Extract monthly-scale habitat factors of rice sheath blight by integrating multi-source time-series information; Optimize habitat factors based on spatial causal inference to realize the spatio-temporal heterogeneity characterization of key sensitive habitat factors of rice sheath blight; Construct and optimize a spatio-temporal dynamic evaluation model for the habitat suitability of rice sheath blight based on the species distribution model; Use the monthly-scale habitat suitability evaluation model group of rice sheath blight to obtain the monthly habitat suitability evaluation results of rice sheath blight during the critical period of rice in the target area.

2. The spatio-temporal dynamic evaluation method for the habitat suitability of rice sheath blight according to claim 1, characterized in that Extract monthly-scale habitat factors of rice sheath blight by integrating multi-source time-series information, specifically as follows: Obtain the remote sensing, meteorological, and plant protection time-series data of the area to be evaluated over the years, calculate the disease habitat factors at different spatio-temporal resolution scales from this, and resample each habitat factor to form a time-series disease habitat factor with a unified spatial resolution; Divide the time-series disease habitat factors on a monthly basis, calculate the mean of the same disease habitat factor at multiple time phases within the same month, and perform masking processing in combination with the monthly rice planting spatial distribution data to obtain the disease habitat factors for each month of each year under the constraint of the rice planting spatial distribution; Calculate the mean of the disease habitat factors for each month of each year under the constraint of the rice planting spatial distribution obtained over multiple years according to the disease habitat factors of the same month in different years, and complete the extraction of monthly-scale habitat factors of rice sheath blight by integrating multi-source time-series information.

3. A spatio-temporal dynamic evaluation method for the habitat suitability of rice sheath blight according to claim 1, characterized in that Optimize habitat factors based on spatial causal inference, specifically as follows: Analyze the causal relationship between each monthly-scale habitat factor and the diseased plant rate of rice sheath blight in the current month through spatial causal inference, and complete the preliminary screening of key sensitive habitat factors affecting the epidemic spread of rice sheath blight based on whether the causal effect index is greater than the set threshold to obtain the preliminarily screened habitat factors; Use spatial causal inference to analyze the possible causal relationships between the preliminarily screened habitat factors, draw a causal-directed relationship diagram among the habitat factors of sheath blight, and realize the optimization of key sensitive habitat factors of rice sheath blight for each month in combination with the causal-directed relationship diagram.

4. A method for spatio-temporal dynamic evaluation of the habitat suitability of rice sheath blight, according to claim 1, characterized in that Construction and optimization of a spatio-temporal dynamic evaluation model for the habitat suitability of rice sheath blight based on the species distribution model, specifically as follows: Respectively use the key sensitive habitat factors of rice sheath blight for each month obtained by optimization as the input of the species distribution model to construct a spatio-temporal dynamic evaluation model for the habitat suitability of rice sheath blight; Realize the optimization of the evaluation model parameters by combining the corrected AIC information criterion and the evaluation index AUC, and form a monthly-scale habitat suitability evaluation model group optimized by month.

5. The spatio-temporal dynamic evaluation method for the habitat suitability of rice sheath blight according to claim 2, characterized in that, The above-mentioned monthly rice planting spatial distribution data is obtained based on the phenological period information of single and double cropping rice: First, based on the transplanting period and harvesting period dates corresponding to the same spatial distribution of each rice cropping type, obtain the rice growth time range under the same spatial distribution of each rice cropping type; On this basis, use the corresponding time range of each month as a threshold to divide the rice growth time range to obtain the monthly rice planting spatial distribution data.

6. The spatio-temporal dynamic evaluation method for the habitat suitability of rice sheath blight according to claim 3, characterized in that, The diseased plant rate data of rice sheath blight for each month is calculated from the plant protection time series data. First, representative fields with sheath blight occurrence are selected as survey points to obtain the corresponding plant protection time series data, and the time series diseased plant rate data of sheath blight for each survey point is calculated by the ratio of the number of diseased plants to the total number of surveyed plants. Second, the plant protection time series data of the diseased plant rate of sheath blight at each survey point each year is divided by month, and the maximum diseased plant rate value within each month is taken as the diseased plant rate of sheath blight at that survey point for that month. Finally, the diseased plant rate data of rice sheath blight for each month is obtained by calculating the average value of the diseased plant rates of sheath blight in the same month over all years.

7. A method for spatio-temporal dynamic evaluation of the habitat suitability of rice sheath blight according to claim 3, characterized in that, The described causal effect index reflects the relationship between the true observed value and the spatially causal inference predicted value through the covariance of the standardized true observed value and the spatially causal inference predicted value; the initial screening of habitat factors is achieved through a preset causal effect index threshold.

8. The spatio-temporal dynamic evaluation method for the habitat suitability of rice sheath blight according to claim 3, characterized in that The drawing direction of the causal directed relationship diagram among the sheath blight habitat factors is as follows: First, the spatial causal inference is used to quantify the causal relationship between pairwise sheath blight habitat factors to obtain the causal effect index between pairwise habitat factors; based on the calculated causal effect index, the directed acyclic graph tool is used to define the causal relationship into the directed acyclic graph, and an interactive and adjustable directed acyclic graph is obtained using the interactive network visualization tool; finally, the graphical grammar drawing tool is used to draw the causal directed relationship diagram among the sheath blight habitat factors.

9. The spatio-temporal dynamic evaluation method for the habitat suitability of rice sheath blight according to claim 4, characterized in that, The optimization of the evaluation model parameters is achieved by combining the corrected AIC information criterion and the evaluation index AUC. Specifically: First, the RM value range and step size are set, and the RM gradient experimental group is determined to quantitatively analyze the regulatory effect of parameter adjustment on the model complexity; then, the RM with the smallest AICc value of the corresponding model complexity under different RMs within the RM value range is selected; finally, the model AUC performances between the optimized RM and the default RM parameter of the species distribution model are compared, and the RM value corresponding to the largest AUC value is selected as the optimal RM of the model to complete the model optimization, and a parameter optimization scheme suitable for the monthly scale evaluation model is established.

10. A spatio-temporal dynamic evaluation system for the habitat suitability of rice sheath blight, characterized in that, It includes the following modules: Habitat factor extraction module: comprehensively extract monthly scale rice sheath blight habitat factors from multi-source time series information; Habitat factor optimization module: optimize key sensitive habitat factors based on spatial causal inference; Evaluation model group acquisition module: construct and optimize the spatio-temporal dynamic evaluation model of rice sheath blight habitat suitability based on the species distribution model; Evaluation module: input the monthly sensitive key habitat factors of rice sheath blight in the target area into the spatio-temporal dynamic evaluation model of rice sheath blight habitat suitability for each month to obtain the evaluation results of the habitat suitability of rice sheath blight in the target area for each month, that is, the spatial distribution range map of the potential distribution probability of rice sheath blight in each month.

Images