PAN Song,PENG De-liang,,LI Ying-mei,CHEN Zhi-jie,ZHAI Ying-yan,LIU Chen,HONG Bo#

1 Shaanxi Key Laboratory of Plant Nematology,Bio-Agriculture Institute of Shaanxi,Shaanxi Academy of Sciences,Xi’an 710043,P.R.China

2 State Key Laboratory for Biology of Plant Diseases and Insect Pests,Institute of Plant Protection,Chinese Academy of Agricultural Sciences,Beijing 100193,P.R.China

Abstract In recent years, Meloidogyne enterolobii has emerged as a major parasitic nematode infesting many plants in tropical or subtropical areas. However,the regions of potential distribution and the main contributing environmental variables for this nematode are unclear. Under the current climate scenario,we predicted the potential geographic distributions of M. enterolobii worldwide and in China using a Maximum Entropy (MaxEnt) model with the occurrence data of this species. Furthermore,the potential distributions of M. enterolobii were projected under three future climate scenarios(BCC-CSM2-MR,CanESM5 and CNRM-CM6-1) for the periods 2050s and 2090s. Changes in the potential distribution were also predicted under different climate conditions. The results showed that highly suitable regions for M. enterolobii were concentrated in Africa,South America,Asia,and North America between latitudes 30°S to 30°N. Bio16 (precipitation of the wettest quarter),bio10 (mean temperature of the warmest quarter),and bio11 (mean temperature of the coldest quarter) were the variables contributing most in predicting potential distributions of M. enterolobii. In addition,the potential suitable areas for M. enterolobii will shift toward higher latitudes under future climate scenarios. This study provides a theoretical basis for controlling and managing this nematode.

Keywords: Meloidogyne enterolobii,species distribution model,MaxEnt,climate change,future climate scenarios,centroid change

1.lntroduction

Root-knot nematodes (Meloidogynespp.),known for their wide geographical distribution and wide host range,have become one of the most threatening pathogenic nematodes in the world. One species ofMeloidogynespp.,Meloidogyne enterolobii,has emerged as a major parasitic nematode infesting many plants including various ornamentals,fruits trees,vegetable crops,wild plants,and weeds (Limaet al.2005;Carneiroet al.2006;Britoet al.2010;Poornimaet al.2016;Schwarzet al.2020).Meloidogyne enterolobiiis also known as the most aggressiveMeloidogynespecies because of its high reproduction rate and ability to overcome the resistance of some important crops,such as potato with theMhgene,tomato with theMi-1gene,bell pepper with theNgene,sweet pepper with theTabascogene,and soybean with theMir1gene (Berthouet al.2003;Britoet al.2007;Cetintaset al.2008). In a microplot experiment,tomato yield losses of up to 65% caused byM.enterolobiialone have been observed (Cetintaset al.2007). Many farmers might not even realize their fields have been infested withM.enterolobiiuntil the end of the season when crops are harvested and obvious galled root systems can be observed (Schwarz 2019). In heavily infested areas,cultivation may become unviable,such as guava cultivation in Brazil and India (Carneiroet al.2007;Singh 2020).

Meloidogyne enterolobiiwas originally described on Pacara earpod trees (Enterolobium contortisiliquum) in Hainan Province of China in 1983 and has been reported in many countries (Yang and Eisenback 1983). Until recently,it was generally considered that the distribution ofM.enterolobiiwas restricted to regions with typical tropical or subtropical climatic conditions,such as Africa,South and Central America,Southeast Asia,and South Asia (EPPO 2022). In the United States,M.enterolobiiwas first reported in Puerto Rico and Florida. Since then,it has spread and has been reported in North and South Carolina (Britoet al.2004;Rutteret al.2019;Schwarzet al.2020). Apart from Hainan,M.enterolobiihas also been reported in other southern regions of China,including Guangdong,Guangxi,Fujian,Hunan,Yunnan,and Taiwan (Niuet al.2012;Wanget al.2015;Xiaoet al.2018;Lianget al.2020;Zhanget al.2020). In Brazil,M.enterolobiihas been detected in 18 states and is the most important pathogen for guava (Carneiroet al.2001;Souzaet al.2006;Limaet al.2007;de Almeidaet al.2008;Silvaet al.2008;Grothet al.2017;Silvaet al.2017;Luquiniet al.2019;Galbieriet al.2020). Apart from tropical and subtropical regions,M.enterolobiihas also been detected in commercial greenhouses in Switzerland,which is located in a temperate region (Kiewnicket al.2009).

Climate change can influence the distribution and abundance of invasive pathogens directly and indirectly.Many studies have analyzed the invasion and distribution of invasive insects influenced by climate change,but few studies have reported on plant nematodes in this field. Global warming provides a potential opportunity forM.enterolobiito spread from low-latitude areas to highlatitude areas. Since it is probable thatM.enterolobiican survive in warmer areas and greenhouses throughout regions where no disease occurs,the risk of its establishment is high. Therefore,modeling the impact of climate change on the distribution ofM.enterolobiican provide vital information for controlling and managing the spread of this nematode.

Some studies have applied species distribution models(SDMs) to predict the potential impact of invasive species to facilitate early warning and planning for future impacts(Ekesiet al.2016;Wei et al.2020). In this study,the current and future potential distribution ofM.enterolobiiwere estimated based on occurrence data using MaxEnt Software. In this study,we aimed to predict the trends of distribution and spread ofM.enterolobii,thereby providing a theoretical basis for the prevention and control of this nematode.

2.Materials and methods

2.1.Species occurrence data

Occurrence data ofM.enterolobiiwere derived from databases of the European and Mediterranean Plant Protection Organization (EPPO,https://gd.eppo.int/),Center for Agriculture and Bioscience International (CABI,https://www.cabi.org/) and Global Biodiversity Information Facility (GBIF,https://www.gbif.org/). In addition,we also performed a literature search was performed in the Web of Science database using “Meloidogyne enterolobii”as a keyword and the Chinese literature database of China National Knowledge Infrastructure (CNKI,https://www.cnki.net/). Geo-referenced coordinates for each distribution sites were either obtained from the literature or using Google Earth Pro v7.3.4 (https://earth.google.com/web/). To check and minimize spatial biases,ENMTools 1.4 (Warrenet al.2010) was used to trim duplicate occurrences such that only one distribution point was present in each grid cell with a spatial resolution of 2.5 arc-minutes (approximately 4.5 km). In total,199 points forM.enterolobiiwere used in this study (Fig.1;Appendix A).

2.2.Environmental variables

Fig. 1 Global occurrence data of Meloidogyne enterolobii (red dot) for the MaxEnt modeling.

Fig. 2 Receiver operating characteristic (ROC) curve generated by the MaxEnt model. The plot represents the sensitivity(true positive rate) and the specificity (false positive rate) of the model. The area under the ROC curve (AUC) represents the entire area underneath the ROC curve (red);the 95%confidence intervals are indicated in blue.

Fig. 3 Jackknife test of variable importance for the MaxEnt model of M.enterolobii distribution. Light sea green,without a variable;blue,with only a single variable;red,with all variables.

Fig. 4 Response curves of three environmental variables with the largest contribution. A,bio10 (mean temperature of warmest quarter). B,bio11 (mean temperature of coldest quarter). C,bio16 (precipitation of wettest quarter). The red curve represents the average of ten replicates;blue margins represent the standard deviation(±SD).

Fig. 6 Global potential distribution map for Meloidogyne enterolobii under the future climate scenario. A,ssp126 in the year 2050.B,ssp126 in the year 2090. C,ssp585 in the year 2050. D,ssp585 in the year 2090. Gray,unsuitable habitat;green,low habitat suitability;orange,medium habitat suitability;red,high habitat suitability.

Fig. 7 Potential distribution map for Meloidogyne enterolobii in China under the future climate scenario. A,ssp126 in the year 2050. B,ssp126 in the year 2090. C,ssp585 in the year 2050. D,ssp585 in the year 2090. White,unsuitable habitat;green,low habitat suitability;orange,medium habitat suitability;red,high habitat suitability.

Fig. 8 Changes in the centroid of the potential distribution of Meloidogyne enterolobii in China. The black star represents the current centroid;dots of different color represent future centroids based on two emission scenarios (ssp126 and ssp585) in 2050 and 2090. The red and blue broken lines indicate the shift of centroid under ssp126 and ssp585 scenarios,respectively.

Twenty-three environmental variables,including 19 bioclimatic variables,1 topographical variable,and 3 soil variables,were collected as the candidate variables for modeling. The topographical variable (altitude) and bioclimatic variables under current climate conditions(bio1-bio19,recorded from 1970 to 2000) were downloaded from the Worldclim Dataset v.2.1 (https://www.worldclim.org/data/worldclim21.html). Three soil variables (soil pH,soil organic carbon and soil texture)were extracted from the Harmonized World Soil Database v.1.2 (HWSD,http://www.fao.org/soils-portal/soil-survey/soil-maps-and-databases/harmonized-world-soildatabase-v12/en/). To avoid multicollinearity between these selected variables,we also calculated Pearson’s correlation coefficients using SDMtoolbox 2.4 (http://www.sdmtoolbox.org/) in ArcGIS 10.2 (ESRI,Redlands,CA,USA) to remove highly correlated variables (r≥|0.8|)(Appendix B;Brownet al.2017). Finally,12 variables(including bio2,bio3,bio10,bio11,bio16,bio17,bio18,bio19,altitude,topsoil pH,topsoil organic carbon,and topsoil texture) were used to predict the distribution ofM.enterolobii.

To predict the global potential distribution ofM.enterolobiiunder future climate conditions,8 selected bioclimatic variables available from three types of Global Climate Models (GCMs) including Beijing Climate Center Climate System Model version 2 (BCC-CSM2-MR),Canadian Earth System Model version 5 (CanESM5) and Centre National de Recherches Météorologiques Climate Model version 6 (CNRM-CM6-1) for 2050 (average for year 2041-2060) and 2090 (average for year 2081-2100),as well as 2 greenhouse gas emission scenarios including ssp126 (low greenhouse gas emission scenario)and ssp585 (high greenhouse gas emission scenario)according to the Shared Socioeconomic Pathways (SSP)under Phase 6 of the Coupled Model Intercomparison Project were obtained from the Worldclim Dataset. All of the current and future environmental variables were resampled to a 2.5-min spatial resolution using ArcGIS.To reduce the variations among different GCMs,a potential habitat suitability map forM.enterolobiiwas finally generated by averaging the projections of the three GCMs under all future climate scenarios. In addition,SDMtoolbox was used to analyze the centroid position changes of the total suitable habitat area ofM.enterolobiiin China under current and future climate scenarios.

2.3.Optimization of model parameters

A well-chosen set of model parameters can enhance the performance of the MaxEnt model. Recent studies have demonstrated that the default settings of MaxEnt may produce overfitted models and are not always appropriate for species distribution modeling (Merowet al.2013;Radosavijevic and Anderson 2014). Thus,to avoid overfitting and produce the optimal model forM.enterolobii,the kuenm package (Coboset al.2019) was used in R 4.1.0 (https://www.r-project.org/) to compare various combinations of the two most critical parameters,feature classes and regularization multiplier,to choose the best combination. MaxEnt contains 5 feature classes: linear(L),quadratic (Q),product (P),threshold (T),and hinge(H);thus,there were 31 combinations of feature classes in total. Moreover,8 values of the regularization multiplier varied from 0.5 to 4 with an interval of 0.5. Altogether,248 candidate models with all combinations of 8 regularization multiplier settings,31 feature class combinations,and a set of 12 environmental variables were evaluated.

Model parameters were optimized using the functionkuenm_cevalbased on statistical significance (partial receiver operating characteristic,ROC,with 500 iterations),predictive ability (omission rate,OR),and model complexity (AICc) (Petersonet al.2008;Zhuet al.2017). According to the “OR_AICc” criterion,the significant candidate models below the threshold of omission rate (for example,≤0.05 when possible) and with the lowest ΔAICc values (≤2) were considered the final models with the optimal model parameters (Coboset al.2019).

2.4.MaxEnt modeling

We used MaxEnt v.3.4.1 (https://biodiversityinformatics.amnh.org/open_source/maxent/) in this study to model the potential suitable habitats forM.enterolobii. In this study,75% of occurrences were randomly selected for model training,and the remaining occurrences (25%) were used for model testing. The analysis included 10 replicates and a maximum number of 5 000 iterations involving the abovementioned 199 distribution points forM.enterolobiiand 12 environmental variables. The percent contribution,permutation importance,as well as jackknife test were used to determine each environmental variable’s relative contribution to the models (Phillipset al.2017). When the current distribution model forM.enterolobiiwas built,we projected it to 2 future periods (2050 and 2090) in MaxEnt.

The logistic output was used to generate a map in raster format with species suitability values between 0 and 1. Fixed cumulative value 5 logistic threshold(FCV5) was used to define suitable and unsuitable habitats forM.enterolobii,and the areas with an output value below FCV5 were regarded as unsuitable habitats(Weiet al.2018). In ArcGIS,the Natural Breaks (Jenks)method was used to reclassify the suitable habitats forM.enterolobiiinto 3 levels: low habitat suitability,medium habitat suitability,and high habitat suitability. The areas of suitable habitats at different levels were calculated and distribution maps were generated using ArcGIS.Moreover,to render every model comparable in space and time series and simplify the analysis process,future climate scenario maps were also prepared using the same criteria as for the current climate scenario.

2.5.Model evaluation

The area under the ROC curve (AUC) was used to estimate the performance of the model (Swets 1988).The AUC value ranged from 0 to 1;AUC＜0.5 indicates a random prediction,0.5≤AUC＜0.7 indicates poor model performance,0.7≤AUC≤0.9 indicates moderate performance,and AUC＞0.9 indicates high performance(Araujoet al.2005;Petersonet al.2011).

3.Results

3.1.Model performance for M.enterolobii

Among the 248 candidate models,all models were statistically significant and only one model met OR and AICc criteria (delta AICc=0) (Appendices C and D). Thus,the model M_1.5_F_lt_Set1 (regularization multiplier=1.5,feature class combinations=L and T) was chosen in the final MaxEnt settings.

The MaxEnt model performance forM.enterolobiiwas better than random with an average test AUC value of 0.939±0.017 (Fig.2),indicating that the model had an excellent performance. The global habitat areas forM.enterolobiiwere divided into four levels: (1) ＜0.095,which indicated an unsuitable habitat;(2) 0.095-0.228,which indicated low habitat suitability;(3) 0.228-0.450,which indicated medium habitat suitability;and (4) ＞0.450,which indicated high habitat suitability.

3.2.Variable importance

The percent contribution and permutation importance of the 12 environmental variables are listed in Table 1. The results showed that bio16 (precipitation of the wettest quarter) was the variable contributing the most. Its contribution rate and permutation importance reached 39.6 and 39.1%,respectively,indicating that bio16 was the main rainfall factor affecting the distribution ofM.enterolobii.In addition,the percent contribution of bio11 (mean temperature of the coldest quarter) reached 17.9% and the permutation importance of bio10 (mean temperature of the warmest quarter) reached 19.7%,indicating that bio11 and bio10 were the main temperature factors.

Table 1 Percentage contribution of 12 environmental variables

The results of the jackknife test yielded the relative importance of each variable (Fig.3). Among these variables,bio16,bio10,and bio11 were the leading three variables in terms of regularized training gain. Thus,the main variables for predicting the potential distribution ofM.enterolobiiwere bio10,bio11,and bio16. The remaining environmental variables had relatively low influence.

According to response curves for the three environmental variables contributing most to the prediction ofM.enterolobiidistribution (Fig.4),the average ranges of each variable associated with high habitat suitability(probability＞0.450) were 25.4-29.0°C for bio10,12.3-22.8°C for bio11,and 360-1 170 mm for bio16.

3.3.Potential distribution of M.enterolobii under current climate scenario

The global potential distribution ofM.enterolobiibased on current environmental variables is shown in Fig.5-A. The results showed that the regions suitable forM.enterolobiiwere distributed on all continents except Antarctica,and most suitable areas were concentrated in Africa,South America,Asia,and North America between latitudes 30°S and 30°N. The global land area of potential suitable habitat forM.enterolobiiwas approximately 5 147.28×104km2,42.9% of which (2 209.69×104km2) comprised regions with low suitability,37.1% of which (1 909.02×104km2) had medium suitability,and 20.0% (1 028.57×104km2) were highly suitable regions (Table 2).

Table 2 Area with suitability for Meloidogyne enterolobii under different climate scenarios

In addition,the distribution pattern ofM.enterolobiiwas significantly different on different continents. Africa had the largest suitable area forM.enterolobii(Fig.5-B),and areas with medium and high habitat suitability were mainly distributed in western,central,and southern Africa.In South America,the areas suitable forM.enterolobiicovered all 12 countries. Highly suitable regions forM.enterolobiiwere distributed in Brazil,Argentina,Paraguay,Bolivia,Venezuela,and Guyana. In Asia,the areas suitable forM.enterolobiiwere primarily distributed in East Asia (China,Japan,and South Korea),South Asia(India,Nepal,Bangladesh,and Sri Lanka),and Southeast Asia (Burma,Thailand,Vietnam,Cambodia,and the Philippines). In North America,the areas of potential distribution mainly included the southeast region of the United States,Mexico,Honduras,Nicaragua,and Cuba. In Oceania,the areas suitable forM.enterolobiiwere mainly distributed in northern and eastern Australia. Europe had the fewest suitable areas,and the total estimated suitable habitat area forM.enterolobiiwas approximately 312.09×104km2(Fig.5-B). Except for a few small areas of medium habitat suitability located in France,Portugal,and Italy,the remaining regions were regarded as areas of low habitat suitability or unsuitable forM.enterolobii.

Under the current climate scenario,the area of potential suitable habitat forM.enterolobiiin China was approximately 300.14×104km2,accounting for 31.26%of the national area (Table 2). Highly suitable regions forM.enterolobiicovered 13.34% (128.11×104km2) of the total area of China and were mostly distributed in southern regions with latitudes lower than 35°N,including Hainan,Guangxi,Guangdong,and Fujian,most parts of Yunnan,Guizhou,Chongqing,Hunan,Jiangxi,Taiwan,eastern Sichuan,as well as small parts of Tibet and Zhejiang.Areas with medium habitat suitability accounted for 13.98%of the national area,covering 134.19×104km2. Areas with low suitability forM.enterolobiicovered 10.51% of the total area of China,totaling 37.84×104km2(Fig.5-C).

3.4.Potential distribution of M.enterolobii under future climate scenarios

The global potential distribution ofM.enterolobiibased on two emission scenarios (ssp126 and ssp585) in 2050 and 2090 is shown in Fig.6 and Table 2. Compared with the current climate scenario,the area of suitable habitat forM.enterolobiiin future climate scenarios increased to varying degrees. Under the ssp126 scenario,it is estimated that by 2050,the total area of suitable habitat will be 5 484.35×104km2,which is increased by 6.55%compared with the current area of suitable habitat. The area of highly suitable habitat will expand to 1 156.09×104km2,which is increased by 12.40% (Fig.6-A). In contrast to the current distribution,by 2090,the total suitable area and highly suitable area will increase by 7.34 and 12.54%,respectively,which exhibited a slightly increasing trend compared with that of 2050 (Fig.6-B). Under the ssp585 scenario,it is estimated that by 2050,the total suitable area and highly suitable area ofM.enterolobiiwill expand to 5 586.98×104and 1 212.99×104km2,which is increased by 8.54 and 17.93%,respectively (Fig.6-C). In contrast to the current distribution,by 2090,the total suitable area will increase by 18.57%,and the highly suitable area will increase by 4.11%. In that period,the total area of suitable habitat will be 6 103.06×104km2,which represents a significant increase compared with that of 2050. However,the area of highly suitable habitat in 2090 was smaller than that in 2050 for the ssp585 scenario(Fig.6-D). In addition,the highly suitable regions forM.enterolobiitended to spread from low latitudes to high latitudes. North America except for the southeast region of the United States,North and West Europe,Saharan Africa,Middle East,Central Asia,Russia,Mongolia,northern China,and southern Australia remained regions that had an unsuitable habitat forM.enterolobii(Fig.6).

The potential distribution ofM.enterolobiiin China under the ssp126 and ssp585 scenarios is shown in Fig.7 and Table 2. Compared with the current climate scenario,the total suitable area and the area of highly suitable habitat forM.enterolobiiincreased,but the area of moderately suitable habitat decreased under two future climate scenarios. Under the ssp126 scenario,in the 2050s,the total area of suitable habitat will be 329.92×104km2,which is increased by 9.92%. The area of highly suitable habitat will expand to 227.75×104km2,which is increased by 73.87%,and some provinces,such as Hubei,Anhui,the southern parts of Henan and Jiangsu,will become highly suitable regions forM.enterolobiicompared with the current area of suitable habitat (Fig.7-A). By 2090,the total suitable area and highly suitable area will increase by 11.93 and 76.88%,respectively (Fig.7-B). Under the ssp585 scenario,the area of suitable habitat forM.enterolobiifurther increased. In the 2050s,the total suitable area and highly suitable area ofM.enterolobiiwill expand to 344.03×104and 228.14×104km2,which is increased by 14.62 and 78.08%,respectively (Fig.7-C). By 2090,the total area of suitable habitat will be 399.43×104km2,which is increased by 33.08% compared with the current area of suitable habitat. In that period,the eastern and southern parts of Shandong will become highly suitable regions forM.enterolobii,whereas the area of highly suitable habitat in Anhui,Hubei,Hunan,and Jiangxi will decrease compared with that of 2050. In addition,Central Shaanxi,Southern Shanxi,Eastern Gansu,and Southwestern Hebei will become moderately suitable regions forM.enterolobii.However,Xinjiang,Qinghai,Inner Mongolia,Heilongjiang,and Jilin remained regions of unsuitable habitat forM.enterolobii(Fig.7-D).

3.5.Centroid changes in potential distribution

The centroids of the habitat ofM.enterolobiiin China under different climate scenarios are shown in Fig.8.For the range ofM.enterolobii,the current centroid was located in Guangxi (110.589°E,24.921°N). Under the ssp126 scenario,the centroid of the suitable area is expected to shift northeastward into Shaoyang City,Hunan Province (111.185°E,27.052°N) during the 2050s,and to 111.305°E,27.207°N during the 2090s. Under the ssp585 scenario,the centroid of the future suitable area will be located at a northeastern position (110.886°E,27.188°N) during the 2050s but shifts in the northwestern direction to Huaihua City (110.347°E,27.442°N) during the 2090s. In general,the centroid of the potential suitable area ofM.enterolobiiwill migrate to higher latitude areas under the future climate scenarios.

4.Discussion

Compared with other niche models,MaxEnt has several advantages,such as easy operation,low sample number requirement,low uncertainty,and high accuracy. Thus,MaxEnt has been widely used in ecology,conservation biology,and invasive biology,especially in the prediction of potential distribution of invasive alien species in recent years (Phillipset al.2006;Liet al.2020). An appropriate model complexity from MaxEnt is critical to obtain a better-performing model in species distribution prediction(Merowet al.2013). However,the default parameters of MaxEnt might lead to an increase in model complexity and yield over-fitted prediction models,which result in reduced cross-temporal transferability and predictive credibility of MaxEnt models for forecasting species distributions under future climate change (Warren and Seifert 2011;Moreno-Amatet al.2015). In this study,two main parameters,feature classes and regularization multiplier,known to influence the prediction results by MaxEnt,were optimized to constrain the complexity of the model.The results indicated that a regularization multiplier=1.5 and feature class combinations=LT were selected as the optimal parameter set,and a model with the lowest AICc was used for the potential distribution prediction ofM.enterolobiiunder different climate scenarios. The estimated test AUC value of 0.939 indicated an excellent predictive accuracy of the model.

Recently,bioclimatic variables and topographic variables have emerged as powerful tools for studying biological invasion and predicting potential global and regional distributions and habitat suitability (Jianget al.2018;Jinget al.2020;Weiet al.2020). However,different from insects,the life cycle of root-knot nematode is completed in the soil. Therefore,soil factors might significantly affect the distribution ofM.enterolobii. In this study,23 environmental variables (19 bioclimatic variables,1 topographical variable,and 3 soil variables) were obtained and their correlation analyses were performed using Pearson’s correlation.Finally,12 variables were selected for the distribution prediction ofM.enterolobii. Based on the jackknife test,the permutation importance,and contribution rate of these variables calculated using the MaxEnt model,the predicted results indicated that the mean temperature of the warmest quarter (bio10),mean temperature of the coldest quarter(bio11),and precipitation of the wettest quarter (bio16)were the main environmental variables affecting the distribution and regional changes in suitable habitat forM.enterolobii. Therefore,its distribution is determined by both temperature and precipitation factors. By contrast,other environmental factors,such as altitude,topsoil pH,topsoil organic carbon,and topsoil texture,had less impact on the distribution ofM.enterolobii. According to previous studies,temperature and humidity factors had significant effects on the survival of nematodes,with different temperatures and humidity required for the development and reproduction of different root-knot nematode species(Charwatet al.2002;Dávila-Negrón and Dickson 2013;Wuet al.2018). Compared with other species,M.enterolobiiwere more suited for reproducing and surviving in a high-temperature environment,and exposure to lower temperature decreased root penetration and slowed the developmental cycle ofM.enterolobii(Vellosoet al.2022). Rainfall also had a significant effect on nematode abundance and diversity. A significant trend for increasedMeloidogynespp.diversity and greater prevalence with increasing rainfall has been observed (Jordaanet al.1989;Fleminget al.2016). This is consistent with the characteristics of high temperature and precipitation in highly suitable habitats forM.enterolobiifound in this study(Fig.4). When temperature and precipitation increased,the temperature and humidity of the soil also increased accordingly,which was more conducive to the survival ofM.enterolobii.Conversely,cold and arid environments wereunfavorable for the survival ofM.enterolobii.

In this study,the global potential distribution ofM.enterolobiiwas first simulated using MaxEnt. The potential global distribution ofM.enterolobiiunder the current climate scenario was concentrated in Africa,South America,Asia,and North America between latitudes from 30°S to 30°N,which is consistent with the occurrence data ofM.enterolobii. The suitable areas forM.enterolobiiwere mostly tropical and subtropical regions with large quantities of precipitation and high temperatures,which not only are more conducive to the reproduction and survival ofM.enterolobiibut alsoallow for greater plant diversity,which is more beneficial for the distribution and spread ofM.enterolobii. On the contrary,although North Africa is a tropical region,it was not suitable for the distribution ofM.enterolobiiowing to little rainfall and an extremely dry climate. It should be noted that in European countries such as France,Belgium,Switzerland,and Portugal,M.enterolobiiis mainly distributed in greenhouse environments,which differs from local field environments(Kiewnicket al.2008). To a certain degree,this might affect the accuracy of model prediction. The area of potential suitable habitat forM.enterolobiiin China was located south of 35°N. The reason for this phenomenon might be that the field environment is too cold for the survival ofM.enterolobiiin the winter. In some high northern latitudes where the occurrence ofM.enterolobiihas been reported,such as France and Shaanxi (China),the nematodes could not be detected one or two years after first being detected,which indicated that these areas might not be suitable for the distribution ofM.enterolobii(Santoset al.2019).

The potential distribution ofM.enterolobiiunder future climate scenarios indicated that from now to 2090s,the suitable area ofM.enterolobiishowed a trend of gradual expansion and movement toward higher latitudes,which might be related to the trend of global warming. These results are consistent with some other invasive pests,such asDysmicoccus neobrevipes(Weiet al.2020) andCulex pipiens pallens(Liuet al.2020). Based on previous studies,under the medium greenhouse gas emission scenario (RCP4.5),the average temperature will rise by 3.0-5.2°C,and annual precipitation will increase by 4.2-6.1% in northern China by the end of the 21st century(Moet al.2018). These results suggested that compared with the current period,a higher temperature and humidity environment in northern China in the future will provide an ideal habitat forM.enterolobii. Therefore,the potential suitable area ofM.enterolobiiwill shift toward the northern part of China under future climate scenarios.

In this study,SDMToolbox Software was used to analyze changes in the centroid of the habitat ofM.enterolobiiin China under different climate scenarios. Currently,the centroid is located in Guilin City,in Guangxi Zhuang Autonomous Region (110.589°E,24.921°N). In the 2050s and 2090s,the centroid of the total suitable area ofM.enterolobiishowed a trend of migration to Hunan Province under two different greenhouse gas emission scenarios (ssp126 and ssp585),which validated the results that climate warming would lead to the spread ofM.enterolobiifrom low-latitude regions to high-latitude regions. To better predict changes in the distribution ofM.enterolobiiin the future,it is necessary to analyze the ecological and physiological mechanisms of its adaptation to environmental factors,such as temperature and humidity.

Considering the wide global distribution of the host andthe high level of agricultural trade between countries,there is a high risk ofM.enterolobiitransmission to non-suitable areas through susceptible plant materials(e.g.,parts of the roots or tuber stems),although the global distribution area of these nematodes is mainly concentrated in tropical and subtropical regions at present. In 2010,M.enterolobiiwasadded to the European and Mediterranean Plant Protection Organization A2 Alert list and recommended for regulation as quarantine pests in some states of the United States(USDA and PCIT 2014;EPPO 2022). A highly specific and sensitive diagnostic method to detectM.enterolobiiis necessary for monitoring the transmission and preventing the spread of this highly destructive nematode. To control the spread of this nematode species,it is essential to cultivate non-infected seedlings and the infected plant material,infested soil,and contaminated farm-equipment from infested fields withM.enterolobiishould not be moved to non-infested regions (Schwarzet al.2020).Meanwhile,several management strategies,such as soil solarization,crop rotation with non-host crops,biological control,new natural resistance sources,and fumigant and non-fumigant nematicide have been employed to control the spread ofM.enterolobiiworldwide (Zasadaet al.2010;Carrillo-Fasioet al.2020).

As an endoparasite,M.enterolobiifeeds and matures to the adult stage of its life cycle fully inside host plant tissue (Elling 2013;Sureshet al.2019). Therefore,biological factors cannot be ignored in predicting the distribution ofM.enterolobii.In this study,the model used considered only the impact caused by abiotic factors but not that caused by biological factors,which may lead to predicting wider suitable areas ofM.enterolobii.It will be necessary to consider the impact of biological factors,such as the distribution ofM.enterolobiihost and different crop planting modes in our subsequent research.

5.Conclusion

In the present study,maps of the potential distribution ofM.enterolobiiin China and worldwide based on current and future climate scenarios are provided for the first time. The range of highly suitable habitats of this nematode increased and shifted toward higher latitudes under future climate scenarios compared with the current climate scenario. Therefore,the northern part of China is considered to have a risk of invasion in the future. We sought to predict the trends of distribution and spread ofM.enterolobii,which provides a theoretical basis for the prevention and control of this species.

Acknowledgements

This work was supported by the Key R&D Project of Shaanxi Province,China (2020ZDLNY07-06) and the Science and Technology Program of Shaanxi Academy of Sciences (2022k-11). We thank LetPub (www.letpub.com) for its linguistic assistance during the preparation of this manuscript.

Declaration of competing interest

The authors declare that they have no conflict of interest.

Appendicesassociated with this paper are available on https://doi.org/10.1016/j.jia.2023.06.022