Spatial and Temporal Distribution of Pine Wood Nematodes in Yiling District Based on GIS

Dan ZHOU

Abstract [Objectives] This study was conducted to accurately understand the spatial distribution of the outbreak of pine wood nematode disease， explore its spatial laws， and provide a more scientific and efficient analysis and decision-making basis for the prevention and management of pine wood nematode disease. [Methods] Taking the GIS management of Bursaphelenchus xylophilus tree felling in Yiling District as an example， the spatial positions of epidemic wood in the two years from 2019 to 2020 were collected. The spatial data management function of ArcGIS software was used to collect point and surface layers. Spatial information maps were drawn by spatial visualization， trend surface， spatial autocorrelation， nearest neighbor distance index and spatial-temporal clustering analysis of kernel density， so as to intuitively understand the evolution of spatial pattern trend of pine wood nematode disease， and to explain the variation raw and diffusion characteristics of pine wood nematode disease. [Results] ①Under the control measures， the affected pine forests and epidemic sites （subcompartments） decreased in the following year， the spread of the epidemic was still spreading， which was spreading from the central to the western and northern regions. The distribution of the plane epidemic sites was sparse from high concentration. The new epidemic sites occurred in a jump， and the number of affected pine trees was high. The outbreak was concentrated in a certain area， and the risk was high. The natural transmission and diffusion would coexist for a long time. ②The epidemic trend of pine nematode in the two years was in the east and west， with a high value in the east， and the trend line was almost straight line. The outbreak of pine nematode showed cone-shaped central point aggregation and spread. ③The spatial and temporal clustering characteristics of the epidemic show spatial clustering in the overall pattern， high clustering of pine victims caused by the spread of the epidemic source and clustering distribution pattern of epidemic points. ④In the two years， the damaged pine trees were searched within the range of 5 km by using the distance threshold. The kernel density maps showed that the damaged pine trees in 2019 were more serious than those in 2020. The result of density map was consistent with the nearest neighbor distance index， and the damaged pine forests were mainly concentrated. ⑤The epidemic situation of B. xylophilus disease in this region has a certain number of epidemic sites and a wide distribution area， the epidemic situation is difficult to eradicate in a short time. On the comprehensive measures， we should increase the clearing of epidemic trees， block the spread of the epidemic situation in the new areas， reduce the epidemic sites， regulate the density of pine trees by forest management， reduce the density of Monochamus alternatus and establish a broad-leaved forest ecosystem. [Conclusions] The study provides assistive technology support for the prevention and control of pine wood nematode disease.

Key words GIS; Bursaphelenchus xylophius; Time and space; Spatial characteristics; Yiling District

Pine wood nematode disease， also known as pine wood nematode wilt disease and pine wilt disease[1]， is a major international quarantine disease and one of the most serious natural disasters. The disease spreads rapidly in tropical and subtropical regions and causes serious damage[2]， and is also one of the more harmful alien invasive species[3]. Once pine trees are infected by pine wood nematodes， they basically cannot survive. The infected pine forests died in patches， posing serious damage and threat to the pine forest ecosystem. The research on pine wood nematodes involves many fields from classification to control， pathogens， vectors， genes， behaviors and symbiotic bacteria， and there is not much breakthrough in general[4].

The pathogen is Bursaphelenchus xylophilus of Bursaphelenchus in Aphelenchoididae[1，5]. For pine wood nematode invasion and expansion in China， it is likely to infect most pine forest ecosystems in almost all climate zones[2]. Its transmission ways are mainly natural transmission and human-made transmission. Natural transmission is mainly completed by the vector insect longhorn beetles. There are 13 species of insects that can serve as vector insects[1]， and 6 species of insects that can carry pine wood nematodes have been found in China[6]. Besides Monochamus alternatus， other insects that can carry pine wood nematodes and thereby spread pine wood nematode disease have been reported at home and abroad[7]. The spread of this disease has not been effectively curbed.

The spread and epidemic of pine wood nematode disease has certain geographical spatial location distribution and occurrence time laws， and shows correlation. In this study， the geographic spatial location data information of pine wood nematode disease was analyzed to accurately understand the spatial distribution of the epidemic situation and dig its spatial law. We quantified the geospatial process of pine wood nematode disease and constructed the corresponding time-space model to assist decision-making for scientific monitoring， prediction， investigation and evaluation， and performed space-time simulation of the density， distribution and changes of the damaged pine forests through geographic information system analysis and research， so as to assist the prevention and control of pine wood nematode disease and provide auxiliary technology support.

Materials and Methods

General situation of the study area

Yiling District is located in the middle and upper reaches of the Yangtze River in Hubei Province， northwest of Hubei， and the geographical position is in 110°51′5.8″-111°39′3.0″ E， 30°32′3.3″-31°28′3.0″ N. The world-renowned Three Gorges Dam is located in the Yiling District. It is located in the Daba Mountain extension， and the main landform is dominated by middle and low mountains. The terrain gradually rises from the southeast to the northwest. The topography is undulating， complex， and the height varies greatly. The altitude is 51-2 005 m， with an average of 650 m. This area， from south to north， has the climate characteristics of central subtropical， northern subtropical and warm temperate zone. The climate is suitable， while the natural conditions are superior， and the sunshine is sufficient. In general， this area has an annual average temperature of 16.6 ℃， an extreme maximum temperature of 41.4 ℃， a minimum temperature of -12 ℃， and annual average sunshine hours of 1 669.2 h， and the accumulated temperature of ≥10 ℃ is 540 8 ℃; the frost-free period is 278 d; the annual average sunshine hours is 1 669.2 h; and the annual average relative humidity is 78.5%. The average annual precipitation in the whole region is 997-1 370 mm. The rainfall season and rainfall are extremely uneven. From May to September， the rainfall is concentrated， with an amount of 800 mm， and the natural climate is warm and humid， suitable for the growth and reproduction of invasive alien pine wood nematodes.

The forest area in this region is 24.838×104 hm2. The existing forest vegetation is mainly natural secondary vegetation， with a wide variety of plants. The forest vegetation type belongs to subtropical deciduous broad-leaved forests. It forest vegetation is dominated by various types of plant structures such as mixed forests of deciduous broad-leaved and evergreen coniferous forests， and there are both mid-subtropical evergreen broad-leaved forests and northern subtropical evergreen broad-leaved forests. The forest coverage rate is 74.87%. The main type of regional forest vegetation is deciduous broad-leaved mixed forest， and most of the forest communities are mainly vegetation composed of Fagaceae species. Masson pine （Pinus massoniana） of pinaceae accounts for 27.90%， and the masson pine forest is in pure and mixed distribution. The broad-leaved mixed forests account for 45.33%， and other forests account for 26.77%. Masson pine occupies a certain ecological space and distribution range in the forest vegetation types in this region.

Pine wood nematode disease was first detected in the pine forest of Meiziya Village， Xiaoxitajie Street in 2006[8]， which became an epidemic area in 2009[9]. The region is relatively densely populated， with developed transportation， frequent economic activities， and more frequent transfer of pine wood between different regions， which is conducive to the spread of pine wood nematode epidemics. Pine wood nematode disease invades masson pine forests with strong pathogenicity， fast death rate， large number， rapid spread and difficult management， and is thus caught off guard. It has caused serious economic and ecological losses to the area. The disease has caused heavy losses to the national economy， destroyed the natural landscape and ecological environment， and posed a serious threat to pine forest resources.

Data source

The data was obtained from logging management subcompartment database （*.SHP format） of wood died for pine wood nematode disease in the two years from 2019 to 2020 in Yiling District. The database was established based on the spatial location of each diseased dead tree with Erqing as the base map， for field investigation and diseased wood （dead wood） logging management， monitoring， inspection and acceptance. The SpatialJoin tool of ArcGIS software platform was used to integrate the diseased dead wood location and plant number into point and surface layer database.

Research method

In order to facilitate data analysis and the mutual conversion between point and surface layers of the logging management subcompartment database of wood died for pine wood nematode disease， the point mode analysis was adopted. The point pattern analysis can be used to describe any type of event data[10]， and each subcompartment （point） data is regarded as an epidemic site （or source of infection）. Because the death of pine trees infected with pine wood nematodes has a spatial location and time of occurrence， and the death time period of pine trees is mainly concentrated in autumn， it is difficult to obtain the epidemic data of each epidemic site （subcompartment） on the ground every month （week）， and the time analysis adopted the annual comparison method. Data descriptive analysis and trend analysis were carried out. The spatio-temporal cluster analysis was performed by the spatial autocorrelation， nearest neighbor distance index， and nuclear density analysis[10-14]. The spatial information map drawing and calculation were performed with ArcGIS 10.8 software.

Descriptive analysis

It was to describe the basic situation of the outbreak of pine wood nematode disease， and compare the evolution of time and space trends. It used data to quantify the spatial transmission and distribution characteristics of the pine wood nematode disease epidemic， and the epidemic map was drawn （including： the spatial distribution of pine forest damage degree， the distribution of epidemic sites， and the three-dimensional trend of epidemic sites）.

Trend surface analysis

Linear model and mathematical surface were used to fit sample data， so as to establish a binary polynomial regression model. The trend analysis tool projected the numerical values of the affected pine trees of forests at the epidemic sites onto orthogonal planes （an east-west plane and a north-south plane）， and fitted the curve on the projection plane to form a trend graph， showing the overall distribution of the epidemic in space， thereby reflecting the regional system variation and establishing a trend surface model.

Spatiotemporal cluster analysis

Spatial autocorrelation （Global Moran's I）

Exploratory spatial data analysis （ESDA） is a method to test whether there is a statistically significant spatial distribution and to further understand the spatial process that generates this distribution[16]. Spatial autocorrelation is an important method of ESDA. Spatial autocorrelation is divided into global and local autocorrelation， and can reveal the regional structure of spatial variables. It is to test whether the attribute value of a variable at a specific spatial position is significantly related to the attribute value of its neighboring spatial position[13-16]. If the attribute values of adjacent objects are similar， they have a positive correlation. If the attribute values of adjacent objects are different， they have a negative correlation. If the attribute values of adjacent objects are random， the adjacent objects have no correlation.

Empirical analysis of the spatial autocorrelation of the spatial spread of pine wood nematode epidemic sites can describe the overall spatial distribution of the epidemic sites and check whether the epidemic sites are related to the attribute values of their neighboring spatial points. The positive correlation result indicates that the attribute value change has the same changing trend as its neighboring spatial units， which represents the existence of aggregation of spatial phenomena， while negative correlation is a random phenomenon. When an area has a high density of damaged pine forests and is positively correlated with the surrounding area， the area can be considered a hot spot.

The analysis uses global autocorrelation. The spatial autocorrelation index Moran's I reflects the spatial correlation of the study area as a whole， and its value range is [-1， 1]. When I>0， it indicates a positive correlation， and a value closer to 1 means greater aggregation; when I<0， it indicates a negative correlation; and when I=0， it indicates a random phenomenon[10-11]. Local spatial autocorrelation is to check whether there is a small and possibly overlooked cluster based on global analysis when no cluster appears. It can reflect the degree of correlation between sub-regions. The calculation formula of Moran's I for global spatial autocorrelation is：

wherein is zi the deviation of the attribute of the epidemic site i from its average value （xi-X）; wi， j is the spatial weight between the epidemic sites i and j; n is the total number of epidemic sites; and S0 is aggregation of all spatial weights. E[I] and V[I] respectively represent the expected value and variance of Moran's I. When p is very small， the observed spatial feature attributes are unlikely to be generated by random processes （small probability events）， so the null hypothesis can be rejected. Both the z score and p value are related to the standard normal distribution.

Nearest neighbor index （NNI）

The nearest neighbor index is to measure the distance between the centroid of each element and the centroid of its nearest neighbor. Calculating the average of the nearest neighbor distances and describing the spatial distribution pattern with the distance between the nearest neighbors is a spatial measurement method for judging the spatial distribution characteristics and spatial pattern characteristics of point elements. With the epidemic site as the element point， the distance between any epidemic site and the nearest epidemic site in the range area can be calculated as dmin， and the actual shortest distance dmin can be obtained by taking the average value. Compared with the theoretical nearest distance E（dmin） under the random distribution mode， the nearest neighbor index R can be obtained. When R=1， it means that the event distribution pattern is random distribution; when R<1， it means that the event distribution pattern is aggregate distribution; and when R>1， the performance pattern tends to spread.

Calculation formula：

Wherein di is the distance between the epidemic site i and the nearest epidemic site; A is the area of the region （using the default area， determined by the smallest enclosing rectangle）; n is the number of epidemic sites; and dmin is the nearest neighbor distance， which is an average distance.

Kernel density analysis （KDE）

The kernel density analysis tool （Kernel Density） is used to calculate the density of the affected pine trees of an epidemic site in its surrounding neighborhood. With the sample disease point as the center， by searching the circle radius， the closer the distance to the center of the circle， the higher the density value of the grid units. The kernel density can intuitively express the distribution density of the research object， and the value of the kernel density represents the degree of aggregation of research objects in the spatial distribution[16]. This method can calculate the density of damaged pine trees within the radius of any epidemic area， and it can reflect the relative concentration of the spatial distribution of the damaged pine forests， by which we can judge the overall spatial distribution characteristics of damaged pine forests.

Calculation formula：

It is to predict the numerical density of damaged pine plants in the forest within the range of the epidemic site location （x， y） ， wherein i=1， …， n is the input epidemic site， which is located with the radius distance of position （x， y）， and includes the epidemic sites in the sum; popi is the field value parameter of the population of the point i， which is the value of the victimized deaths in all subcompartments; and disti is the distance between the epidemic site and the position （x， y）.

Agricultural Biotechnology2020

Results and Analysis

Basic information

Since the outbreak of pine wood nematode disease in this region was identified and confirmed by the State Forestry Administration in 2009， the epidemic has spread rapidly. In just over 10 years， a total of 208.97×104 pine trees have died， covering an area of 7.49×104 hm2. The natural vector is M. alternatus adults， which occur once a year， has a long emergence cycle， hides in the pine bark or trunk， and is difficult to control. It mainly damages masson pine and Pinus armandi， and has caused economic and ecological losses of 210 million yuan in the region. According to the statistical analysis in Table 1， in 2020， the scope of occurrence spread to 108 villages in 11 townships， through natural transmission and human-made transmission. The area of the affected pine forests was 1.126×104 hm2， and the number of outbreak sites （subcompartments） was 7 360. 36.20×104 dead trees were cut， and the damage degree of small class was 1-5 000. The disposal method was on-site burning. The forest disaster rate was less than 3‰， and the damage was mild. There were many epidemic areas and sites （subcompartments） in the region， and there was a high risk of natural spread and diffusion. Compared with the same period in 2019， the degree of harm in 2020 was less than that of the previous year under the control measures. The value of pine trees was still relatively large， and there was a trend of aggravation or spreading. At present， capital investment has been continuously increased， control measures have been taken， and the degree of damage has been contained.

Visual analysis

The ArcGIS software was used to process and output for visualization， analyze the epidemic layer， and divide the affected pine forests into 10 levels from low to high by the natural discontinuity point classification method. The color from lighter to darker and the size of the point were used to visualize the map， and the distribution maps of the affected pine forest degree and the distribution maps of the epidemic sites are as shown in Fig. 1 to Fig. 4. Areas where pine wood nematode disease caused a high degree of damage to pine forests were Fenxiang town， Huanghua Town， Yaqueling， Longquan Town， Dengcun Township. The horizontal distribution of the affected pine forests was characterized by clustered and concentrated continuous outbreaks， many epidemic sites， large number of damaged plants， high damage degree， existence of certain areas with high risk of spreading in the distribution plane， and coexistence of natural transmission and diffusion and outbreak risk for a long term.

When comparing the epidemic situation in transmission time， there were 5 542 epidemic sites （subcompartments）， with a range of damage degree from 1 to 4 917 plants in 2019; and there were 7 360 epidemic sites （subcompartments） with a range of damage degree from 1 to 2 209 plants in 2020. Under the control measures， the value of damaged pine forests declined， but the decline was small. When comparing the epidemic situation in transmission space， with the passage of time， the spread of the 2020 epidemic spread to the periphery， from the central to the western and northern regions. Under the control measures， the degree of damage of the epidemic sites tended to be sparsely distributed from highly concentrated distribution， and the distribution was relatively scattered. From the overall time and space characteristics of the two years of epidemics， there were many epidemic sites （subcompartments）， which were distributed in a certain area. It is difficult to prevent and control all epidemic sites， and it is difficult to eradicate the epidemic in the short term.

Trend surface analysis

With the numerical value of damaged pine trees at the location coordinate of the epidemic site （x， y） as independent variables， the trend surface analysis of the epidemic sites was performed by the space exploration data analysis. For the three-dimensional data points （X， Y， Z）， X and Y represents the latitude and longitude of the geometric center of each region， of which the X axis is the direction east， and the Y axis is the direction north， and Z represents the independent variable， i.e.， the numerical value of damaged pine trees （i）. As shown in Fig. 5 and Fig. 6， the heights of the rods at the epidemic spot represent the numerical values of the damaged plants. i was projected on planes XY and YZ and thus formed scatter plots， respectively， and binomial fitting was performed through the scatter points to obtain a certain curve trend of the numerical value of damaged pine trees in the latitude and longitude direction. The results showed that the epidemic trend of pine nematode disease in the two years from 2019 to 2020 was in the east and west， with a high value in the east. In 2020， it spread to the middle. From the analysis of the projected trend line value， the trend line was proximately a straight line， and the overall trend was flat and low. The epidemic sites presented cone-shaped central point aggregation and spread.

Analysis of the spatial clustering characteristics of epidemic sites

According to the distribution of epidemic sites and the density distribution characteristics of damaged pine trees of forests， by setting different distance thresholds several times， it was found that the effect was best when the distance radius （R） reached 5 km. The smoother the density grids generated and the higher the generalization degree， the higher the relative concentration degree of the density of pine trees. The numbers of damaged pine trees in the forests within a radius of 5 km at each epidemic spot were collected， and the regional density distribution map of the number of damaged pine plants was obtained. Different levels of color grids represented the change of the number of damaged pine trees， regional classification， diffusion trend， and spatial distribution aggregation change characteristics. The results are shown in Fig. 6 and Fig. 7. The kernel density map showed that the degree of aggregation in 2019 was based on 290 plants and 8 natural discontinuity point classification， and the degree of aggregation ranged from 0 and 2 321 plants; and the degree of aggregation in 2020 was based on 152 plants and 8 natural discontinuity point classification， and the degree of aggregation was between 0 and 1 361 trees. The epidemic areas in the two years were roughly the same， and the situation and the degree of damage in 2019 were more serious; the scope of the epidemic expanded in 2020， and the severer epidemic areas shifted. The densely damaged pine forests were distributed in Letianxi Town， Fenxiang Town， and Huanghua Town. The density maps are similar to a heat map， which is highly concentrated from the center point and spreads around. The obtained results were consistent with those of the nearest neighbor index， that is to say， the pine forest damage was characterized by aggregate distribution. The density distribution map could be used as the basis for zoning the damage degree of pine wood nematode invasion.

Global autocorrelation analysis

The calculation results are shown in Table 2. Samples from all epidemic sites in the two annual periods were calculated separately， and the distance thresholds involved were respectively 1 and 2 km. The pine forest values of the epidemic sites were searched within the radius of 1 and 2 km， so as to check whether the attribute value of an epidemic site in the space position was correlated to the attribute value of the epidemic site in the adjacent space position. The results showed that the Moran index （Moran's I） under the selected distance thresholds in the two years was greater than 0， and the P value （probability） was 0， so the probability of randomly generating this clustering pattern was less than 1%. There was a positive correlation between the epidemic site in the region， and the overall pattern showed spatial aggregation， that is， the infected plants in the pine forests were highly clustered due to the spread of the epidemic source.

Nearest neighbor index analysis

Spatial autocorrelation analysis can only analyze the spatial aggregation of data， but cannot reveal the spatial distribution pattern. Spatial distribution is divided into three types： aggregate， uniform and random. The Nearest neighbor index of the point-like elements in uniform distribution is the largest， followed by random distribution， and the smallest in aggregate distribution. The nearest neighbor distance is a geographic indicator that indicates the proximity of point-like objects in geographic space. The nearest-neighbor point index can well reflect the spatial distribution characteristics of point-like elements. The Average Nearest Neighbor analysis tool was used to calculate the nearest neighbor index based on the average distance of each disease point and the expected nearest disease point. The expected average distance is the average distance between fields in the assumed random distribution pattern. The area involved was the default area of the system determined by the smallest enclosing rectangle of the epidemic site. The calculation results are shown in Table 2. The sample epidemic sites of the two years were calculated separately. The average observed values of the epidemic sites were 173.34 and 169.82 m， respectively， which were smaller than the expected averages of 433.4 and 348.9 m， respectively， and the nearest neighbor ratio index values were 0.399 9 and 0.486 7 （between 0 and 1）， respectively， smaller than 1. The corresponding Z test values were -85.46 and -84.24， respectively， and the P value was 0， that is， the probability of random distribution was 0， indicating that the distribution pattern was present in aggregate distribution.

Results and Discussion

Conclusions

The two-year data of the pine wood nematode epidemic in the region was analyzed by spatial visualization. Through time and space comparison， the dead wood of pine trees in 2019 reached the historical peak. Under the control measures， the damaged pine trees in 2020 showed a decline， but decline in epidemic sites （subcompartments） was small， and the number of damaged pine trees was still large. The horizontal distribution of the affected pine forests was characterized by clustering and concentrated continuous outbreaks， large distribution area with many epidemic sites， existence of certain regions in the distribution plane with high risk of spreading， and coexistence of natural spread and outbreaking for a long time. When comparing the epidemic situation in transmission space， with the passage of time， the spreading range of the epidemic spread to the periphery; the degree of damage at the epidemic sites had a tendency from a high aggregate concentration to a sparse distribution trend; and the new epidemic sites occurred in a leaping manner， and the distribution was relatively scattered. It is difficult to prevent and control all epidemic sites， and it is difficult to eradicate the epidemic in the short term.

The three-dimensional trend surface image visually showed the changing trend of the pine wood nematode epidemic on a three-dimensional scale. The results showed that the epidemic trend of pine nematode disease in the two years from 2019 to 2020 was in the east and west， with a high value in the east. In 2020， it spread to the middle. From the analysis of the projected trend line value， the trend line was proximately a straight line， and the overall trend was flat and low. The epidemic sites presented cone-shaped central point aggregation and spread.

The results of space-time clustering： ① The distance threshold radii for the epidemic site samples in the two years were， respectively， 1 and 2 km， with which the Moran's I value was greater than 0， which indicated a positive correlation. The overall pattern showed spatial clustering， and the damaged pine trees were highly clustered due to the spread of the epidemic source. ② The average observed values of the nearest neighbor index were respectively， 173.34 and 169.82， which were smaller than the expected averages of 433.4 and 348.9 m， respectively. The nearest neighbor ratio index was 0.399 9 and 0.486 7， respectively. The corresponding test scores were -85.46 and -84.24， and the P value was 0. The distribution mode of the epidemic sites was aggregate distribution.

The spatial aggregation characteristic kernel density maps of the epidemic sites showed that the numbers of damaged pine trees in forests in the two years within the threshold radius （R） of 5 km were subjected to 8 natural discontinuity point classification， and the aggregation degree was in the range of 0-2 321 and 0-1 361 plants， respectively. The degree of damage in 2019 was relatively serious， and there was an expansion trend in 2020. The epidemic areas in the two years were roughly the same， and the areas suffered from more serious damage shifted in 2020. The density maps were similar to a heat map， which is highly concentrated from the center point and spreads around. The obtained results were consistent with those of the nearest neighbor index， that is to say， the pine forest damage was characterized by aggregate distribution. The density distribution map could be used as the basis for zoning the damage degree of pine wood nematode invasion.

Discussion

There are many factors for the spread and epidemic of pine wood nematode disease. This study has certain limitations， because the spreading factors of the epidemic are closely related to the density of vector insects， altitude， climate， distribution of pine forests， density of pine forests， environmental factors， interference from human activities， etc.， and the data of epidemic sites over the years is incomplete. It is necessary to superimpose research. Spatial management and spatial analysis of the pine wood nematode disease epidemic is an effective scientific management method.

Based on the analysis of the occurrence time of the epidemic in two years， under the prevention and control measures， the number of damaged pine trees showed a downward trend， but there was a trend of spreading in the distribution area. The overall development of the epidemic was relatively rapid， and the spread of the epidemic in some areas was accelerating. The numbers of damaged pine trees in local areas were increasing， and a trend of concentrated outbreak was observed， accompanied with a leaping distribution. This region has a wide distribution of epidemics and scattered epidemic sites. Once a certain number of epidemic sites and certain distribution area are reached， it will be difficult to eradicate in a short period of time.

From the spatial analysis of the occurrence of epidemics in two years， most areas were concentrated， and some were scattered. New epidemic sites were distributed in a leaping manner. The epidemics were widely distributed and the epidemic sites were scattered. The spatial distribution and spatial pattern showed a high degree of clustering. In terms of the spatial development trend of the epidemic， the overall epidemic was developing rapidly， and the spread of the epidemic in some areas were accelerating. The number of damaged pine trees in new areas had increased year by year， and the area of damage had continued to expand. There was a trend of concentrated outbreaks in some areas. We believe that the pine wood nematode epidemic sites will develop from discrete distribution to clustered distribution at the beginning， and it is a long-term process to achieve aggregation when the pine forest will be extremely damaged.

With regard to the prevention and control of pine wood nematode disease， there is currently no better prevention and control method in China， and it is unable to make a definite diagnosis about whether a ground living pine forest is infected. Pine wood nematode does not actively infect pine trees by itself， and needs M. alternatus to spread and invade healthy pine trees. We can transfer the direction to M. alternatus， which is the vector for the prevention and control of pine wood nematode disease， under comprehensive control measures. We can prevent the spread of the epidemic by reducing the density of the vector insect population， blocking the epidemic sites spatially， reducing the epidemic sites year by year， increasing the strength of cleaning diseased wood， and early detecting and removing of infected trees. From the perspectives of forest management， we should adjust the density of pine trees and cultivate and establish a broad-leaved forest ecosystem.

References

[1] National Forestry and Grassland Administration Forest and Grassland Pest Control Station， Nanjing Forestry University. Occurrence and control techniques of pine wood nematode in China[M]. Beijing： China Forestry Publishing House， 2019. （in Chinese）

[2] LI YX， ZHANG XY. Research progress of pathogenetic mechanism of pine wilt Disease[J]. Journal of Environmental Entomology， 2018， 40（2）： 231-241. （in Chinese）

[3] REN BZ， TANG LH. Progress in quarantine pests of Bursaphelenchus xylophilus[J]. Journal of Jilin Agricultural University， 2020， 42（1）： 8-13. （in Chinese）

[4] PAN CS. Research progress on pine wood nematode disease[J]. Journal of Xiamen University： Natural Science， 2011， 50（2）： 476-483. （in Chinese）

[5] XU XW， WANG YX， LIU X， et al. Research progress of pathogenetic mechanism of pine wilt disease[J]. Hubei Agricultural Sciences， 2018， 57（20）： 5-8. （in Chinese）

[6] ZHANG Y， RAO LJ， HE LX， et al. The vector insect species of pine wilt disease and advances in research on integrated management[J]. Biological Disaster Science， 2019， 42（3）： 171-178. （in Chinese）

[7] CAO DD， WEI JR， WANG WC， et al.： Nanjing Oxytelus batiuculus （Coleoptera： Staphylinidae） under the bark of wood died for Bursaphelenchus xylophilus[J]. Scientia Silvae Sinicae， 2019， 55（1）： 160-165. （in Chinese）

[8] WAN YM， CUI GJ. Integrated control measures of pine wood nematode disease in Yiling District， Yichang City[J]. Mod Agric Sci Technol， 2018（18）： 116-117. （in Chinese）

[9] State Forestry and Grassland Administration Government Network. Announcement of the State Forestry and Grassland Administration （No. 4 of 2019） （2019 pine wood nematode epidemic area） [EB/OL]http：∥www.forestry.gov.cn/portal/main/govfile/13/govfile_1635.html.2019-07-07. （in Chinese）

[10] ZHOU CH， PEI T， et al. Spatial analysis principles of geographic information system[M]. Beijing： Science Press， 2011. （in Chinese）

[11] TANG， GA， et al. ArcGIS geographic information system spatial analysis experiment tutorial[M]. Beijing： Science Press， 2008. （in Chinese）

[12] XIAO GX. Actual combat of spatial statistics[M]. Beijing： Science Press， 2018. （in Chinese）

[13] Maribeth Price. ArcGIS geographic information system tutorial[M]. LI YL， translator. Beijing： Publishing House of Electronics Industry， 2012. （in Chinese）

[14] ESRI： GIS mapping software and spatial analysis[EB/OL]https：∥www.esri.com/zh-cn/home.2020-07-26

[15] YING LG， NING YM. Spatial data： Its nature， effects and analysis[J]. Advances in Earth Science， 2005， 20（1）： 49-55. （in Chinese）

[16] Quantitative method and application based on GIS[M]. JIANG SG， TENG JH， translator. Beijing： Commercial Press， 2009. （in Chinese）