Introduction

Climate change is significantly exacerbating the frequency and intensity of pest outbreaks in crops, posing a serious threat to global agricultural ecosystems and food security1,2,3. Rising temperatures, altered precipitation patterns, and an increase in extreme weather events not only affect crop growth but also create favorable conditions for the spread and establishment of plant pathogens and pests, leading to an expansion of their suitable habitats and the invasion of new regions4,5,6. For example, notable insect pests such as Ostrinia furnacalis, Helicoverpa armigera, and the vector of Bursaphelenchus xylophilus, Monochamus alternatus, have shown significant range expansion driven by climate change, posing severe threats to agricultural and forestry ecosystems7,8,9. Similarly, pathogens like Puccinia spp., Phytophthora infestans, and Erysiphe necator are also spreading rapidly in warmer and more humid environments10,11,12. These climate-driven pest and pathogen outbreaks have caused severe ecological imbalances across multiple regions, disrupting biodiversity and leading to substantial economic losses in agriculture, which in turn affects farmers’ livelihoods and regional economic stability13,14,15. Therefore, understanding and predicting the potential distribution and spread of pests and diseases under climate change is utmost importance, as it is crucial for developing effective disease management strategies and addressing the agricultural risks posed by a changing climate16,17,18.

As the ancestor of many cultivated fruit trees such as apples, walnuts, apricots, plums and so on around the globe, Tianshan Wild Fruit Forest is a rare natural genetic treasure trove in the world and an important germplasm resource bank for genetic diversity and gene evolution of temperate fruit trees19.The wild fruit forests in the Tianshan Mountains stretch over 500 km across Kazakhstan, Kyrgyzstan and Xinjiang Uygur Autonomous Region (Xinjiang) in China20with a total area of about 2.40 × 1012 m2. Among them, the wild fruit forests in the Tianshan Mountains within China’s territory account for nearly 40% of the total area of wild fruit forests in the world. It is mainly distributed in the low hills of the valleys of Ili and the Tacheng Basin in Xinjiang, and is a class II key wild plant under the protection of China and Xinjiang. However, the area of wild fruit forests has been shrinking and the habitat of wild apricot has deteriorated due to personnel activities, overgrazing, and the occurrence and spread of pests and diseases21. The severe incidence of pests and diseases is a major constraint on the healthy growth of wild apricot, with shot-hole disease caused by W. carpophilus showing more severe manifestations22.The average incidence of shot-hole disease on leaves and fruits of wild apricot was 89.33% and 73.17%, respectively, and on fruits of wild cherry plum was 57.45%, and it was found that ecological health is seriously threatened due to shot-hole disease caused by W. carpophilus by the field surveys conducted in the years 2021−202423,24.W. carpophilus, the causal agent of shot-hole disease, was first described in France and has since been reported across Europe, Africa, Asia, North and South America, and Australia, infecting a broad range of Prunus hosts25.

Wilsonomyces carpophilus is capable of causing shot-hole disease in drupe fruit trees globally26which severely affects healthy plant growth and fruit quality(Fig. 1d & h). The disease first appeared in France in 1927, and subsequently appeared worldwide27,28. W. carpophilus mainly affects apricot, peach and nectarine and to a lesser extent cherries, and is capable of infesting not only the leaves and fruits of the host, but also the shoots, dormant buds and calyxes29,30. Symptoms on leaves consisted of 1–2 mm circular spots, purplish in colorwith a yellow bordering halo, which eventually enlarged and became necrotic, causing abscission in the center of the lesion, giving the leaf the typical shot-hole appearance (Fig. 1a ~ c). Symptoms on fruits appeared as form scabs, and the fruits were necrotic (Fig. 1e ~ g). W. carpophilus has become a major threat to the health of wild fruit forests in the Tianshan Mountains of Xinjiang and stone fruit trees, with its severity showing an escalating trend.

Fig. 1
figure 1

Natural symptoms on Prunus. armeniaca leave and fruit after infection by Wilsonomyces carpophilus in the field. Note (a ~ c) The symptoms of wild apricot leaves were infected on 1 d, 4 d, and 8 d; (e ~ g)The symptoms of wild apricot fruit were infected on 1 d, 6 d, and 12 d; (d) natural disease symptoms on Prunus armeniaca leaves in the field; (h) natural disease symptoms on Prunus armeniaca fruits in the field.

Full size image

Species Distribution Models (SDMs) have become essential tools for studying the distribution, spread dynamics, and potential expansion of pathogens. SDMs effectively predict the suitable habitats of pathogens, under current and future environmental conditions, providing a scientific basis for disease early warning and control measures by integrating environmental variables with historical data. The study of plant pathogen distributions using SDMs has increased in recent years. Kriticos successfully predicted the potential distribution of Puccinia psidii in Australia using an SDM, revealing the profound impact of climate change on the pathogen’s distribution patterns31. Similar approaches have been applied in the distribution prediction of Erysiphe necator, offering valuable support for disease management in global grape-growing regions32. SDMs not only help uncover changes in the suitable habitats of pathogens, but also enable the analysis of the effects of climate change, land use, and other factors on pathogen spread and invasion potential, providing predictive insights for the development of effective control strategies33. Particularly in the context of global climate change, SDMs offer forward-looking decision support for agricultural and forestry disease management, helping to identify potential high-risk areas and optimize resource allocation to minimize the economic losses caused by disease34. Species distribution modeling (SDM) encompasses a suite of correlative approaches that statistically relate species occurrence records to spatially explicit environmental predictors35. Regression-based algorithms, including generalized linear models (GLMs) and classification tree analysis (CTA), provide transparent, hypothesis-driven frameworks but may lack flexibility in capturing complex, nonlinear responses and interactions36,37. Envelope and presence-only methods—such as Bioclim and Surface Range Envelope (SRE)—offer efficient screening tools with minimal data requirements but can oversimplify ecological niches and are sensitive to sampling bias and outlier records38,39. The advent of machine-learning frameworks—exemplified by Maximum Entropy (MaxEnt), Random Forest (RF), and Gradient Boosting Machines (GBM)—has markedly improved the capacity to model complex ecological responses and interactions, often achieving superior discrimination ability at the expense of reduced mechanistic interpretability40,41,42. Recognizing the variability inherent in individual model projections, ensemble forecasting platforms (e.g., BIOMOD) systematically combine multiple algorithms and parameterizations to quantify inter-model uncertainty and produce consensus predictions, thereby enhancing the reliability of distribution forecasts for conservation planning43,44. By combining outputs from six complementary algorithms into a consensus suitability map, our ensemble modeling approach pinpoints high-risk hotspots for W. carpophilus, thereby enabling prioritized surveillance and targeted control measures in the most vulnerable regions.

This study aims to confront the critical challenge of predicting the spread and potential distribution of W. carpophilus in the context of climate change and environmental variability. First to forecast the current and future suitable habitats of W. carpophilus using SDMs, which will provide insights into potential areas at risk of pathogen spread under varying climatic conditions; Next to utilize the mess mod multivariate environmental similarity and dissimilarity analysis to examine which environmental factors are likely to drive changes in the pathogen’s spread. Ultimate to explore the relationship between environmental factors and the pathogen’s distribution using environmental factor curves, which will assist in identifying thresholds and optimal conditions for pathogen establishment, as well as the potential for its invasion into new regions. Through the above research this study will provide valuable predictive insights into the future dynamics of W. carpophilus, helping to inform disease management strategies in affected agricultural systems.

Materials and methods

Species geographical distribution data

A total of 90 sample plots were established in the natural distribution area of wild apricot in Ili, Xinjiang during 2021–2024 based on tree age, slope, elevation, stand structure, temperature and humidity. The survey of the overwintering sites of W. carpophilus was carried out in each winter (November to February of the following year). The regularity of shot-hole disease occurrence was investigated from March to October every year. Selected plants within each plot were collected weekly from leaves, bark, twigs, winter buds, and undertree weeds. The disease cycle of shot-hole disease was studied by the isolation rate of W. carpophilus isolated from the samples. The data was integrated to comprehensively evaluate the relationship between disease occurrence and key environmental drivers. We compiled occurrence records of W. carpophilus from multiple sources, including the Global Biodiversity Information Facility (GBIF, www.gbif.org), relevant literature, and reports published by governmental agencies. This resulted in a total of 641 occurrence points. To ensure compatibility with environmental variables, we refined the occurrence data using ENMTools45which involved filtering the data to retain only one occurrence point per 5 × 5 km grid cell. This approach effectively minimized spatial autocorrelation, a potential issue arising from clustered data points that, if left unaddressed, could lead to biased model outcomes46,47. Although we did not use an explicit bias-background file, all occurrence records were spatially thinned to a 5 km resolution using the ENMTools algorithm, a procedure shown to effectively reduce spatial clustering and mitigate sampling bias in presence-only SDMs46,47. Following this filtering process, we retained a total of 302 valid distribution points of W. carpophilus (Fig. 2). This refined dataset formed the foundation of our predictive models, ensuring a robust analysis of species distribution in relation to environmental factors.

Fig. 2
figure 2

Study area geographic locations of Wilsonomyces carpophilus population distribution.Note: overlapping purple circles indicate multiple records at the same or adjacent locations; point size was chosen to balance legibility at the global scale without distorting spatial accuracy.

Full size image

Environmental data sources and screening

The environmental data used in this study include 19 climate variables. Climate data were obtained from the WorldClim database (https://worldclim.org/), with a resolution of 2.5 arcminutes48. For future climate projections, we selected data from the sixth phase of the Coupled Model Intercomparison Project (CMIP6), specifically from the BCC-CSM2-MR model under two Representative Concentration Pathways (RCPs): SSP126 and SSP585. Geospatial data were sourced from the National Geomatics Center of China (http://www.ngcc.cn/ngcc/)49.

To mitigate the potential for multicollinearity among environmental variables and avoid model overfitting, we employed the “usdm” package in R (version 4.2.3) to filter variables. We retained only those factors with a correlation coefficient lower than 0.8 and a Variance Inflation Factor (VIF) below 10. Additionally, we ensured that ecologically significant variables were preserved for the analysis50. Ultimately, six climate variables were selected for the model (Fig. 3).

Distribution data for wild apricot host trees were collected from literature, field surveys, and the Global Biodiversity Information Facility (GBIF, http://www.gbif.org). The collected distribution points, stored in a.csv file, were imported into ArcGIS 10.4. Use the Join tool to smooth distribution points into continuous TIFF raster files using kernel density estimation. Subsequently, the rasterized host variable was converted to an ASCII file to ensure data consistency. This process included aligning coordinate systems, pixel sizes, rows, columns, and spatial extents with those of the climate variables. Incorporating the host factor in this manner ensures that the model is more accurate and reliable by capturing the ecological relationships relevant to W. carpophilus distribution.

Fig. 3
figure 3

Correlation analysis of environmental factors.

Full size image

Host factor quantification

The distribution of the host plant is a key ecological driver of plant disease occurrence, and for W. carpophilus, the range of its principal host, Prunus armeniaca, directly determines habitat extent and ecological suitability35. To integrate this biotic factor, we acquired host-distribution records and processed them into an independent environmental predictor for the SDM51.

  1. 1.

    Data sources.

    Host records for Prunus armeniaca were compiled from the Global Biodiversity Information Facility (GBIF)52World Flora Online53government monitoring reports, and Xinjiang regional vegetation surveys54.

  2. 2.

    Spatial resolution matching.

    To align with the climate layers, raw occurrence points were aggregated into a 0.3 × 0.3 km grid. Any cell containing one or more points was treated as a single presence record, thereby reducing spatial autocorrelation from point clustering55.

  3. 3.

    Habitat Suitability Index (HSI) calculation.

    Using R, we rasterized the aggregated host grid and performed a weighted overlay of host presence with its most influential climate and soil predictors to derive a Habitat Suitability Index (HSI)56. Each factor was assigned a weight reflecting its ecological importance; the composite score per grid cell yielded an HSI raster ranging from 0 (unsuitable) to 1 (highly suitable)57.

  4. 4.

    Data standardization.

    The HSI values were linearly normalized to the [0–1] range to facilitate their integration alongside other environmental variables in the SDM, ensuring consistency and scientific rigor40.

These additions clarify the processing steps and confirm that six predictors—Bio2, Bio3, Bio8, Bio9, Bio13, Bio15, and host HSI—were used in the final model.

Species distribution model construction

To construct the optimal ensemble model (EM), the study used the biomod2 (version 4.2.4) package in RStudio and employed 12 algorithms: Artificial Neural Network (ANN), Classification Tree Analysis (CTA), Flexible Discriminant Analysis (FDA), Generalized Additive Model (GAM), Gradient Boosting Machine (GBM), Generalized Linear Model (GLM), Multivariate Adaptive Regression Splines (MARS), Maximum Entropy Model (MAXENT), Maximum Entropy Network Model (MAXNET), Random Forest (RF), Surface Range Envelope (SRE), and Extreme Gradient Boosting (XGBOOST). Each model was optimized using the bm tuning function58. To model the potential geographic distributions of W. carpophilus, 75% of the distribution records were used for training, and the remaining 25% were used for testing59. To ensure result stability, ten randomizations were conducted, generating 1,000 pseudo-absence points, and this process was repeated three times, yielding 360 single models. Pseudo-absences (1 000 points per replicate, 3 replicates) were generated using the Surface Range Envelope (SRE) method in the biomod2 R package. To limit bias from clustered presences, absences were drawn only within the accessible area (M), defined as a 500 km buffer around the minimum convex polygon of occurrence records. We applied a 10 km exclusion buffer around all presences and stratified sampling by environmental predictor quantiles to ensure coverage of the full climatic space accessible to the species. Ten models with a True Skill Statistic (TSS) value greater than 0.9 were selected based on accuracy: CTA, FDA, GAM, GBM, GLM, MARS, MAXENT, MAXNET, RF, and XGBOOST. These 10 models were used to construct six ensemble models: Mean Ensemble (EMmean), Coefficient of Variation Ensemble (EMcv), Confidence Interval Ensemble (EMci), Median Ensemble (EMmedian), Weighted Arithmetic Ensemble (EMca), and Weighted Mean Ensemble (EMwmean). The ensemble model with the highest TSS value was selected to predict the potential geographic distribution of W. carpophilus. Model accuracy was assessed using TSS, AUC60. Generally, higher values for these metrics indicate greater accuracy in the species distribution mode61,62,63.

Multivariate environmental similarity analysis

To identify global climate suitability zones for W. carpophilus, this study employed Multivariate Environmental Similarity Surfaces (MESS) analysis. MESS evaluates the degree to which a region’s environmental conditions resemble those within the model’s training area, making it a valuable tool for assessing both suitability and uncertainty in novel environments. This approach calculates the differences in multidimensional environmental variables (e.g., temperature, precipitation) relative to baseline climate conditions. The global climate dataset used was WorldClim 2.1, which includes key variables such as annual temperature, precipitation, and seasonal variation64. These data formed an environmental matrix that reflects the climatic characteristics of both native and potentially invaded regions of W. carpophilus.

Initially, the species distribution model (SDM) was used to predict the species’ suitability by comparing climate conditions in known distribution areas with global climate data. The MESS analysis then assessed the similarity of each climate variable in the multidimensional environmental space, highlighting regions where the environmental conditions fell outside the model’s training set. This indicated areas of uncertainty or risk in climate suitability. By calculating the MESS values for each grid cell within the current suitable habitat of W. carpophilus, regions where the climate conditions align closely with the species’ known distribution were identified. MESS values close to 100 indicate high similarity, while negative values reveal substantial differences65.

Statement on guidelines

All experimental studies and materials involved in this research fully comply with the IUCN Policy Statement on Research Involving Species at Risk of Extinction and the Convention on the Trade in Endangered Species of Wild Fauna and Flora (CITES), and are in full compliance with relevant institutional, national, and international guidelines and regulations.

Results

Model accuracy evaluation

This study evaluated the accuracy of various models, including ANN, CTA, FDA, GAM, GBM, GLM, MARS, MAXENT, MAXNET, RF, XGBOOST, and SRE, along with an ensemble model (EM) for W. carpophilus. The TSS values for CTA, FDA, GAM, GBM, GLM, MARS, MAXENT, MAXNET, RF, and XGBOOST models were all greater than 0.8, while the AUC values exceeded 0.9 (Fig. 4). Consequently, these ten high-accuracy models were selected to construct the EM using six ensemble approaches: EMmean, EMcv, EMci, EMmedian, EMca, and EMwmean.

The developed EM demonstrated the highest accuracy among all models evaluated in this study. Therefore, the EM derived from individual high-accuracy models significantly enhanced predictive precision while reducing uncertainty during the modeling process. These findings suggest that the predicted potential geographic distribution of W. carpophilus, as inferred from the EM, is highly reliable.

Fig. 4
figure 4

Evaluation indexes of the individual predictive models. True Skill Statistic (TSS) values for each algorithm. Area Under the Curve (AUC) values for each algorithm.

Full size image

Current potential geographic distribution of Wilsonomyces Carpophilus worldwide

Based on the current potential geographic spread of W. carpophilus, the species distribution model (SDM) results indicate distinct gradients of habitat suitability globally. The current distribution is categorized into four classes: unsuitable habitat, low suitability habitat, moderate suitability habitat, and high suitability habitat (Fig. 5).

The unsuitable habitat represents the largest area, covering approximately 1.30 × 108km², characterized by regions where the climatic and environmental conditions are not conducive to the establishment or spread of W. carpophilus. These areas are primarily located in regions with climates that are either too extreme or do not provide the necessary environmental requirements for the pathogen’s survival and proliferation.

Low suitability habitats account for an estimated 9.73 × 106 km². These regions are characterized by marginal conditions that may support the pathogen but are not optimal for widespread establishment. The low suitability areas are distributed predominantly across parts of North America, Eastern Europe, Central Asia, and Southern Australia. These areas reflect environments where some climate variables align with the known conditions suitable for W. carpophilus, but other factors may limit its establishment or survival.

Moderate suitability habitats are estimated at 5.71 × 106 km². These areas are considered suitable for W. carpophilus, with favorable climatic conditions that can facilitate the species’ establishment and survival, albeit with some limitations. Moderate suitability habitats are particularly prevalent across parts of Europe, North America, and temperate regions of East Asia. The distribution pattern suggests that these regions may have a combination of moderate temperature and precipitation conditions that are favorable for the development of W. carpophilus but may still experience some constraints on optimal growth.

High suitability habitats, which represent areas most favorable for the establishment and spread of W. carpophilus, cover approximately 3.33 × 106 km². These regions are concentrated mainly across central Europe, extending into parts of Eastern Europe and Western Asia, as well as localized areas in North America. High suitability habitats provide the most conducive environmental conditions for the proliferation of W. carpophilus, with optimal temperature, precipitation, and seasonal variations that align well with the species’ ecological requirements. This suggests that these regions are at the highest risk of invasion and potential establishment of the pathogen, necessitating targeted monitoring and management strategies.

Overall, the results indicate that W. carpophilus has a significant potential range under current climatic conditions, with particularly favorable conditions across temperate regions of Europe, parts of Asia, and North America. The identification of these suitability zones provides crucial insights for prioritizing areas for pathogen monitoring and developing appropriate management interventions.

Current potential geographic distribution of Wilsonomyces Carpophilus in China

Based on the results of the biomod2 model for the prediction of the potential habitable zones of Wilsonomyces carpophilus, its distribution in China showed significant spatial heterogeneity and was classified into four grades, namely, non-habitable zone, low-habitable zone, medium-habitable zone and high-habitable zone. The areas of each class of suitable areas were 7.88 × 106 km² for non-suitable areas, 1.25 × 106 km² for low suitable areas, 2.87 × 105 km² for medium suitable areas, and 1.67 × 105 km² for high suitable areas, which accounted for 84.70%, 13.40%, 3.10%, and 1.80% of the national land area, respectively (Fig. 5).

The largest area of the non-adapted zone covers most of the country, mainly in the north-central part of the three northeastern provinces (around the Daxinganling, Xiaoxinganling, and Zhang Guangcailing), as well as in the high-altitude and desert regions of the northwestern region (the Taklamakan Desert, the Qinghai-Tibetan Plateau, the Qilian Mountains, and the Tian Shan Mountain Range). In addition, the North China Plain, the middle and lower reaches of the Yangtze River Plain, and the southern tropical region (most of Hainan Island) also belong to the unsuitable areas. The climatic conditions in these regions are not suitable for the expansion of W. carpophilus.

The area of low habitat is 1.25 × 106 km², mainly distributed in the transition area with lower climate suitability. Typical distribution areas include the edge of the the valleys of Ili, the Turpan Basin and the northern foothills of the Tianshan Mountains in Xinjiang, the Jiayuguan, Jiuquan and Zhangye areas of the Hexi Corridor in Gansu, parts of the Yunnan-Guizhou Plateau (Dali and Kunming in Yunnan, and Guiyang in Guizhou), and parts of the Loess Plateau (the central part of Shanxi and the eastern part of the Guanzhong Plain in Shaanxi). The low-fitness areas usually show a banded or blocky distribution with weak fitness.

The mesophilic zone covers an area of 2.87 × 105 km², and is mainly concentrated in mountainous and hilly areas with more favorable climatic conditions. The edge of the core area of the valleys of Ili in Xinjiang, the low-altitude basins on the north and south sides of the Tianshan Mountains, and the area to the east of the Qilian Mountains in Gansu are the main distribution ranges of the mesophilic zone. In addition, the central part of the Yunnan-Guizhou Plateau (Yunnan Dianzhong Plateau) and the edge of the Sichuan Basin (south of the Chengdu Plain) are also moderately hospitable.

The area of highly suitable zone is the smallest, only 1.67 × 105 km², but the distribution is concentrated, mainly in the climatic conditions that are most suitable for the growth of W. carpophilus. In Xinjiang, the core area of the valleys of Ili and the northern foot of Tianshan Mountain are the distribution hotspots of highly suitable areas. In addition, the west-central part of the Hexi Corridor in Gansu (near Zhangye, Jiuquan and Dunhuang), part of the Sichuan Basin (around Panzhihua), and part of the Yunnan-Guizhou Plateau (Dianzhong Plateau) also belong to highly suitable areas. These areas are characterized by moderate temperatures, adequate and stable precipitation, and provide ideal environments for the survival of W. carpophilus. W. carpophilus in China is concentrated in the Tianshan Mountains and the Hexi Corridor in the northwest, the Yunnan-Guizhou Plateau and the southern edge of the Loess Plateau. The suitable areas are mainly in the low suitable areas, while the medium and high suitable areas are small in size but concentrated in distribution. The non-viable areas occupied most of the country, indicating that the distribution of W. carpophilus is significantly limited by climatic conditions.

Current potential geographic distribution of Wilsonomyces Carpophilus in Xinjiang

The results of the analysis of the potential fitness zones of Wilsonomyces carpophilus in Xinjiang based on the biomod2 model showed that the fitness zones of this pathogen have significant heterogeneity in spatial distribution, the area of each grade of suitable zone is 1.17 × 106 km² in non-suitable zone, 2.37 × 106 km² in low suitable zone, 1.83 × 105 km² in medium suitable zone and 1.48 × 105 km² in high suitable zone, accounting for 31.50%, 63.70%, 4.90% and 4.00% of the total area of Xinjiang, respectively (Fig. 5).(https://desktop.arcgis.com).

The non-viable areas are mainly located in the Taklamakan Desert and the Kunlun and Altun mountain ranges in the southwestern part of the southern border, as well as in the northern Altai Mountains in the northern border, and other high-altitude or extremely arid areas. These areas are characterized by harsh climatic conditions, low temperatures and scanty precipitation, which are unsuitable for the growth and spread of W. carpophilus.

The low-fitness zone is the most widely distributed area, covering the oasis agricultural areas of Xinjiang and some of the mountainous fringe areas. It mainly includes Urumqi, Changji periphery and Shihezi in the northern foothills of the Tianshan Mountains, Korla and Yuli areas in the southern foothills of the Tianshan Mountains, the periphery of the valleys of Ili, and low altitude areas such as Burqin, Fuhai and Jimune counties in the southern foothills of the Altai Mountains. These areas have suitable hydrothermal conditions, but local limiting climatic factors reduce their habitability.

The medium-fitness zone is smaller in size and is concentrated in areas with better climatic conditions, including Changji and Fukang in the middle and low elevation areas of the Tianshan Mountains, Yining City, Huocheng County, and Nilek County in the core area of the valleys of Ili, and part of Hami City in the extended area of the Hexi Corridor. These areas are transitional zones from low to high habitability zones, with climatic conditions in between.

Highly suitable areas are the smallest in size and highly concentrated in distribution, mainly located in the core area of the valleys of Ili in the counties of Tekes, Gongliu and Zhaosu, as well as in the oasis areas at the northern foothills of the Tianshan Mountains. These areas have warm and humid climate, sufficient precipitation, and frequent agricultural activities, which provide the most suitable environment for W. carpophilus to survive.

Overall, the potential fitness zones of W. carpophilus in Xinjiang were mainly distributed in the Tianshan Mountains and its surrounding oases, the valleys of Ili and the southern foothills of the Altai Mountains, with the low fitness zones covering the widest areas, while the high fitness zones were concentrated in the areas with optimal climatic conditions, suggesting that the hydrothermal conditions and the topographic features were the key influencing factors for the distribution of the pathogen’s fitness zones.

Fig. 5
figure 5

Current potential geographic distribution of Wilsonomyces carpophilus.(ArcGIS Version 10.4).

Full size image

Future potential global geographic distribution of Wilsonomyces Carpophilus

Based on the projections under different Shared Socioeconomic Pathways (SSPs), the future potential global geographic distribution of W. carpophilus reveals substantial shifts across different suitability levels under various climate scenarios (Fig. 6). The results are presented for two future periods (2050s and 2090 s) under two contrasting climate scenarios, SSP126 and SSP585.

In the 2050 s under SSP126, the unsuitable habitat constitutes approximately 1.26 × 108 km², which is slightly reduced compared to the current distribution, suggesting minor shifts in habitat suitability. The low suitability habitat is projected to expand to 1.14 × 107 km², indicating an increase in marginally favorable conditions for the pathogen’s establishment. Moderate suitability habitats are projected to cover 6.88 × 106 km², and high suitability areas are expected to occupy 4.49 × 106 km², with notable expansion in regions such as Europe, North America, and parts of Asia, indicating more favorable climatic conditions for W. carpophilus compared to the present.

Under SSP585 in the 2050 s, the unsuitable habitat area slightly increases to 1.27 × 108 km². The low suitability habitat is projected to decrease to 1.09 × 107 km² compared to SSP126, reflecting slight variability in areas with marginal suitability due to more extreme climatic conditions. Moderate suitability areas are expected to shrink to 6.33 × 106 km², whereas the high suitability habitat expands slightly to 4.66 × 106 km². These trends suggest a shift in habitat quality with increased concentrations of suitable areas under more severe climate conditions, particularly in central Europe and North America.

In the 2090 s, under SSP126, the unsuitable habitat decreases slightly to 1.25 × 108 km², indicating some areas transitioning to more favorable conditions for pathogen establishment. The low suitability habitat expands further to 1.23 × 107 km², suggesting a gradual broadening of marginal habitats. Moderate suitability areas increase to 6.99 × 106 km², and high suitability regions also expand to 5.13 × 106 km². These findings indicate a trend towards increased suitability under the less severe climate scenarios, particularly in regions like Europe, which remain consistently suitable for the pathogen.

In contrast, the 2090 s under SSP585 reveal notable changes, with the unsuitable habitat slightly decreasing to 1.24 × 108 km², similar to SSP126. The low suitability areas are projected to occupy 1.23 × 107 km², indicating stability in areas with marginal conditions. Moderate suitability habitats expand further to 7.25 × 106 km², and high suitability regions increase significantly to 5.76 × 106 km². This trend highlights a notable expansion of highly suitable areas under extreme climate conditions, suggesting a considerable increase in favorable environments for W. carpophilus, particularly across Europe, parts of Asia, and North America, making these regions potential hotspots for the pathogen’s future distribution.

Overall, the future projections indicate a general trend of increasing habitat suitability for W. carpophilus, particularly under more extreme climate scenarios (SSP585). High suitability habitats are projected to expand significantly by the 2090 s, suggesting that climate change may create increasingly favorable conditions for the spread and establishment of this pathogen. These insights underscore the need for targeted monitoring and proactive management strategies, especially in regions projected to transition to higher suitability classes.

Fig. 6
figure 6

Geographic distribution of future scenarios for Wilsonomyces carpophilus.(ArcGIS Version 10.4).

Full size image

MESS and mod variable analysis

The Multivariate Environmental Similarity Surface (MESS) analysis for W. carpophilus under future climate scenarios indicates notable shifts in environmental similarity across the pathogen’s current and projected distribution range. The mean multivariate environmental similarity values for the 302 known occurrence sites under SSP126 and SSP585 scenarios were calculated as 26.83 and 26.89, respectively, during the 2050 s, and 27.22 and 23.51, respectively, during the 2090 s (Fig. 7). These values reflect variations in the degree to which future environmental conditions resemble the baseline conditions used to train the model.

The analysis highlights significant climatic anomalies, particularly in the northern parts of the current habitat range, where environmental conditions are projected to diverge considerably from the baseline. The climate scenarios ranked by decreasing degrees of climatic anomalies are as follows: SSP585 during the 2090 s, SSP585 during the 2050 s, SSP126 during the 2090 s, and SSP126 during the 2050s. This order indicates that the severity of environmental changes, which affect the suitability of habitats for W. carpophilus, is most pronounced under the more extreme emissions scenario (SSP585), especially towards the end of the century.

The dominant environmental factor influencing habitat suitability was identified as Precipitation Seasonality (bio-15), which exhibited drastic changes beginning with SSP126 in the 2050 s and continued to dominate as the primary driver of habitat shifts. This suggests that fluctuations in the distribution of precipitation throughout the year are likely to have a substantial impact on the future distribution of W. carpophilus. Alongside Precipitation Seasonality, other climatic variables such as Mean Diurnal Range (bio-2), Isothermality (bio-3), and Precipitation of the Wettest Month (bio-13) also contributed to the overall suitability. However, their influence on the pathogen’s potential range gradually decreased over time. As a result, the combination of increased variability in precipitation and a reduction in the stabilizing influence of other climatic factors is expected to negatively affect the fitness and potential range expansion of W. carpophilus.

The spatial analysis presented in Fig. 8 shows that the most dissimilar variables under future scenarios are concentrated in northern regions of the current habitat, primarily in parts of North America, Europe, and Asia. This highlights regions where environmental conditions will diverge most significantly from those suitable for the current distribution, introducing greater uncertainty in model predictions for these areas. Notably, under SSP585 in the 2090 s, the extent of regions characterized by high levels of environmental dissimilarity is substantially increased, suggesting that climate extremes will play a critical role in altering habitat suitability for the pathogen.

Overall, the results show underscore the impact of projected climate change on the distribution of W. carpophilus, with Precipitation Seasonality emerging as the key factor driving habitat shifts. The increasing divergence of environmental conditions from the present indicates the potential challenges for the pathogen’s survival in certain regions, while also highlighting the areas that may become more susceptible to its spread. These insights are crucial for developing adaptive management strategies aimed at mitigating the risks posed by W. carpophilus under future climate scenarios.

Fig. 7
figure 7

Multivariate environmental similarity surface (MESS) analysis of Wilsonomyces carpophilus under future climate scenarios.(ArcGIS Version 10.4).

Full size image
Fig. 8
figure 8

Most dissimilar variable analysis of Wilsonomyces carpophilus under future climate scenarios.(ArcGIS Version 10.4).

Full size image

Environmental factor response curves for Wilsonomyces Carpophilus

The environmental factor response curves for W. carpophilus illustrate the relationship between key climatic variables and the predicted probability of the species’ occurrence. These response curves provide insights into the optimal environmental conditions for the pathogen’s survival and proliferation (Fig. 9). For each climatic variable, the range corresponding to a predicted occurrence probability greater than 0.5 is considered as the optimal habitat for W. carpophilus.

The analysis of Mean Diurnal Range (bio-2) suggests that W. carpophilus is most likely to occur in areas where the mean diurnal temperature range is between 9.97 °C and 13.65 °C. This indicates that moderate daily temperature fluctuations are favorable for the establishment of the pathogen. The response curve shows a sharp decline beyond these values, suggesting that significant deviations in temperature variability negatively affect habitat suitability.

For Isothermality (bio-3), which measures the ratio of diurnal to annual temperature range, the optimal range for W. carpophilus lies between 25.83% and 40.58%. This indicates that environments with moderate stability between daily and annual temperature variability are conducive to the pathogen. Beyond this range, particularly at higher values of isothermality, the probability of occurrence decreases significantly, suggesting that highly stable temperature environments may not favor the pathogen’s establishment.

The Mean Temperature of the Wettest Quarter (bio-8) is another key factor influencing W. carpophilus distribution. The optimal range for bio-8 is between 8.54 °C and 22.17 °C, indicating that the pathogen prefers moderate temperatures during the wettest part of the year. Lower or higher temperatures during this critical growth phase are likely to limit its ability to thrive.

For Mean Temperature of the Driest Quarter (bio-9), the optimal range falls between 0.051 °C and 18.34 °C. This suggests that the pathogen can tolerate cool conditions during the driest quarter but prefers areas where temperatures are not excessively high. A gradual decline in suitability is observed with increasing temperatures beyond this range, highlighting the importance of cool, stable conditions during drier periods.

Precipitation of the Wettest Month (bio-13) shows an optimal range from 29.87 mm to 141.90 mm, indicating that W. carpophilus favors environments where precipitation during the wettest month is moderate. Extremely high levels of precipitation beyond this range are associated with a rapid decline in habitat suitability, possibly due to adverse effects on the pathogen’s ability to establish under excessively wet conditions.

Precipitation Seasonality (bio-15), which measures the variation in monthly precipitation over the year, shows an optimal range of 3.72–51.45%. This suggests that W. carpophilus prefers areas with low to moderate variation in precipitation. Higher variability in precipitation throughout the year appears to be detrimental to its habitat suitability, indicating that stable moisture availability is crucial for the pathogen’s growth.

In summary, the environmental response curves provide a detailed understanding of the optimal climatic conditions for W. carpophilus. Moderate temperature ranges, stable precipitation patterns, and specific thresholds of seasonal variability are all crucial for the pathogen’s establishment and spread. Understanding these preferences can inform management strategies by identifying areas at greatest risk of invasion under both current and future climate scenarios, helping to mitigate the impact of this pathogen on susceptible agricultural and forestry systems.

Fig. 9
figure 9

Environmental factor curves.

Full size image

Discussion

Climatic drivers of Wilsonomyces Carpophilus distribution

The distribution of W. carpophilus is strongly influenced by several key climatic drivers, including Precipitation Seasonality (bio-15), Mean Diurnal Range (bio-2), and Precipitation of the Wettest Month (bio-13). The environmental factor response curves indicated that these variables play significant roles in determining the suitability of habitats for the pathogen. Precipitation Seasonality (bio-15) emerged as the most critical factor, suggesting that W. carpophilus is highly responsive to seasonal variations in precipitation. Moderate variability in precipitation was found to be optimal for the pathogen’s establishment, as excessive fluctuations could hinder its ability to maintain a stable population (Fig. 9). The dominance of Precipitation Seasonality as a driver of suitability is consistent with findings from other studies examining plant pathogens that are sensitive to seasonal moisture availability, such as Phytophthora infestans, which also prefers environments with moderate seasonal variability to support its life cycle1,2. Inusa Jacob Ajene, predicted that the future distribution of the fitness zones of Candidatus Liberibacter asiaticus and Candidatus Liberibacter africanus in Africa is also related to changes in climatic environments66and the study by Muhammad Riaz Ejaz, The results showed that future potential distribution areas of Fusarium could be affected under different climatic environments in the future67.

The interplay between precipitation variability and temperature fluctuations is crucial in defining the habitat suitability for W. carpophilus. Moderate precipitation seasonality provides a stable moisture regime that supports the pathogen’s life cycle, while extreme variability poses a significant barrier to its establishment. Similarly, the preference for moderate diurnal temperature range highlights the pathogen’s need for stable but not extreme thermal conditions. This sensitivity to both moisture and temperature aligns with previous research on climate-sensitive pathogens such as Erysiphe necator, where similar climatic requirements were noted for successful establishment and spread6. Such dependencies on stable yet moderate environmental conditions suggest that W. carpophilus is highly susceptible to climate-induced habitat changes, which may either expand or restrict its potential distribution, depending on future climatic shifts. Empirical studies show that moderate precipitation seasonality ensures prolonged leaf-wetness periods (> 24 h) required for conidial germination while minimizing wash‐off of inoculum, and that mean diurnal temperature ranges of ~ 10–15℃ optimize germ‐tube growth and sporulation—whereas extreme thermal amplitudes or dry spells disrupt pathogen fitness and reduce infection efficiency68.

In conclusion, Precipitation Seasonality, Mean Diurnal Range, and Precipitation of the Wettest Month are key climatic drivers that collectively determine the distribution of W. carpophilus. The preference for moderate levels of these factors underscores the pathogen’s reliance on stable environmental conditions for successful establishment. These findings are consistent with those reported for other fungal pathogens, highlighting the broader relevance of precipitation and temperature stability in influencing pathogen dynamics and emphasizing the need for targeted management strategies in regions where climatic conditions are expected to become more favorable for W. carpophilus under future climate scenarios.

Future climate change impacts and potential range expansion

The results of this study demonstrate significant shifts in the potential distribution of W. carpophilus under future climate scenarios (SSP126 and SSP585) for the 2050 s and 2090s. Under both scenarios, the projected expansion of suitable habitats highlights the potential impact of climate change on the pathogen’s distribution. The extent of suitable habitat is projected to increase more significantly under SSP585, the scenario representing higher greenhouse gas emissions and more extreme climate change conditions, compared to SSP126, which assumes a lower emissions pathway (Fig. 6).

In the 2050 s, under SSP126, W. carpophilus is expected to see a modest expansion of highly suitable habitats, primarily across Europe and North America, with an estimated 4.49 × 106 km² identified as highly suitable. Under SSP585, this area increases to 4.66 × 106 km², indicating a more pronounced expansion in favorable conditions. By the 2090 s, these trends become more evident. Under SSP126, the high suitability area expands to 5.13 × 106 km², while under SSP585, it reaches 5.76 × 106 km², suggesting that higher greenhouse gas emissions will lead to more favorable conditions for the pathogen’s establishment across a wider area (Fig. 6). These findings are consistent with other studies that have shown the potential for increased disease risk in a warming climate, particularly for plant pathogens that are sensitive to temperature and precipitation regimes69,70.

The increased extent of suitable habitats under SSP585 highlights a trend towards greater expansion under more extreme climate conditions. This suggests that regions currently unsuitable for W. carpophilus could become highly favorable by the end of the century, particularly in parts of Europe, North America, and Asia. For example, the northern regions of Europe, which are currently marginal for the pathogen, are projected to become increasingly suitable under SSP585. Similar trends have been observed for other fungal pathogens, such as Phytophthora ramorum, where climate change scenarios have predicted an expansion into higher latitudes and elevations due to increasing temperatures and changing precipitation patterns71,72. These findings underscore the importance of climate change in shaping the future distribution of pathogens and highlight the potential for the emergence of new disease hotspots in previously unaffected areas.

The implications of these projected shifts are significant for the management of W. carpophilus, particularly in at-risk regions such as Europe, North America, and parts of Asia. As the pathogen’s suitable range expands, new habitats conducive to its establishment could lead to increased risks for agricultural and forestry systems in these regions. The expansion of high-suitability areas under SSP585 suggests that warmer temperatures and altered precipitation regimes could create conditions that favor pathogen establishment and spread. This is consistent with previous research on other pathogens, such as Erysiphe necator and Puccinia striiformis, which have been shown to benefit from warmer temperatures and increased humidity, leading to an expansion of their geographic range73,74.

Effective management strategies must consider the potential for range expansion due to climate change, especially in regions predicted to transition from unsuitable to suitable conditions. Proactive measures, such as enhanced monitoring of W. carpophilus in regions identified as at-risk, could help mitigate the potential impacts of its spread. Additionally, incorporating predictive models like the SDM and MESS analysis used in this study could provide valuable tools for policymakers and land managers to prioritize areas for disease prevention efforts, thereby reducing the potential economic and ecological impacts of W. carpophilus establishment in newly suitable areas.

In summary, climate change is projected to significantly alter the distribution of W. carpophilus, with more extreme scenarios (SSP585) leading to larger expansions of suitable habitats. The creation of new favorable environments, particularly in Europe, North America, and parts of Asia, highlights the need for targeted management strategies to mitigate the risks associated with the spread of this pathogen. These findings align with broader trends observed for other plant pathogens, emphasizing the importance of integrating climate change projections into disease management frameworks to safeguard vulnerable agricultural and forestry ecosystems.

Implications for disease management and control strategies

In this study, demonstrate that regions projected to have high suitability under future climate scenarios, such as Europe, North America, and parts of Asia, are at elevated risk of pathogen invasion and establishment. The expansion of suitable habitats under more extreme climate conditions (SSP585) highlights the urgent need to address the potential impacts on agricultural and forestry systems in these areas71. Identifying these high-risk areas enables stakeholders to prioritize resource allocation for disease management, ensuring that vulnerable regions are adequately monitored and prepared for potential outbreaks.

Proactive monitoring and management are essential to mitigating the risks posed by W. carpophilus. For regions identified as highly suitable under SSP585, early detection of the pathogen will be critical to prevent widespread establishment. Implementing robust monitoring networks, particularly in new areas predicted to become suitable, can help detect early signs of invasion. Monitoring programs should focus on both wild and cultivated host plants, as these provide reservoirs for the pathogen. Studies on other fungal pathogens, such as Phytophthora ramorum, have shown that early detection and rapid response are key factors in successfully managing invasive pathogens and limiting their spread73,75. Therefore, establishing a comprehensive monitoring framework that includes field surveys, sentinel sites, and regular sampling can help detect the pathogen before it becomes established.

Management strategies should also include preventive measures aimed at reducing the risk of pathogen introduction and establishment. For example, quarantine measures and restrictions on the movement of potentially infected plant material could help prevent the spread of W. carpophilus to new areas. In high-risk regions, promoting agricultural practices that reduce host susceptibility, such as optimizing irrigation schedules to avoid excessive moisture, can also contribute to reducing disease incidence. Similar approaches have been recommended for managing grapevine powdery mildew caused by Erysiphe necator, where minimizing conditions conducive to infection has been shown to effectively reduce disease outbreaks76. Targeted fungicide applications, when necessary, should be used in conjunction with cultural practices to manage disease pressure effectively, particularly in regions where the pathogen is already present.

The integration of predictive models such as the species distribution model (SDM) and Multivariate Environmental Similarity Surface (MESS) analysis into policy and management frameworks is crucial for mitigating potential economic losses. Predictive models allow for the identification of regions most vulnerable to pathogen spread, enabling targeted disease prevention efforts. The application of SDMs has proven beneficial in identifying high-risk areas for other plant pathogens, such as Puccinia striiformis (the causal agent of wheat stripe rust), allowing for the implementation of timely management strategies to minimize crop loss77,78. Incorporating these models into national and regional policy frameworks will allow for proactive planning, particularly in regions where agricultural and forestry sectors are economically important but vulnerable to climate-induced changes in pathogen dynamics.

In summary, the identification of high-risk areas under future climate scenarios underscores the importance of proactive monitoring, preventive measures, and the integration of predictive models into disease management frameworks. By implementing targeted strategies based on predictive modeling, stakeholders can mitigate the risks associated with the spread of W. carpophilus and protect economically significant agricultural and forestry sectors from the potential impacts of climate change.

Limitations of the study and directions for future research

This study has several limitations that must be acknowledged to provide a context for the interpretation of the results. First, the reliance on specific climate data sources, such as WorldClim, introduces potential biases, as these datasets may not fully capture microclimatic variations that are crucial for local-scale predictions of W. carpophilus distribution79. Moreover, the modeling process involved assumptions about static land-use patterns, which may not accurately reflect future landscape changes, such as agricultural expansion or urbanization, that could significantly impact the pathogen’s spread. This limitation is especially pertinent given that land-use changes have been demonstrated to influence the spread of other plant pathogens, such as Phytophthora cinnamomi80. Furthermore, while spatial filtering was used to minimize the impact of spatial autocorrelation in the occurrence data, potential biases due to clustering may still exist, which could influence the accuracy of the predictions81.We acknowledge that, despite thinning, residual spatial sampling bias may remain and could lead to overestimated TSS and AUC values in intensively surveyed regions. In future work, incorporation of a target-group background or explicit bias file would further improve model calibration and transferability. We acknowledge that densely clustered presence records in Central Asia may still bias discrimination metrics. By constraining pseudo-absences to the accessible area and stratifying by environmental conditions, we mitigate exaggeration of TSS and AUC. In addition, we now report omission rates at the selected threshold to explicitly assess model omission errors. We acknowledge that the use of random data splits rather than spatial cross-validation may leave residual spatial autocorrelation between training and test sets, potentially inflating discrimination metrics. Future studies should implement spatially explicit validation frameworks (e.g., checkerboard partitioning or k-fold cross-validation by latitude) to robustly evaluate model transferability and avoid overoptimistic performance estimates.

Future research could address these limitations by incorporating additional factors that may refine the predictions. One potential improvement involves including land-use data, as well as anthropogenic influences, to capture a more comprehensive picture of W. carpophilus habitat suitability. Studies on similar pathogens have highlighted the importance of anthropogenic factors in determining pathogen distribution82,83. Additionally, further studies should explore the ecological interactions between W. carpophilus and other species, such as mutualists or competitors, to better understand the drivers of establishment beyond climatic factors. Ecological interactions have been shown to significantly influence the dynamics of plant pathogens, as evidenced by research on Erysiphe necator, where interactions with host phenology and natural enemies played a crucial role in determining its success84.

Moreover, the application of more advanced modeling approaches, such as Joint Species Distribution Models (JSDMs), could offer valuable insights into the community-level dynamics involving W. carpophilus. JSDMs have been increasingly used to study the co-occurrence patterns of species under changing climatic conditions, providing a more nuanced understanding of how species interactions and shared environmental drivers impact their distributions85. These advancements could enhance the predictive power of models and improve our ability to manage emerging threats posed by pathogens like W. carpophilus.

Conclusion

In the present study, significant insights into the potential spread of W. carpophilus under various climate change scenarios, highlighting the critical role of key climatic drivers in determining the pathogen’s current and future distribution. By utilizing species distribution models (SDMs) and Multivariate Environmental Similarity Surface (MESS) analyses, the research demonstrated how climatic shifts, particularly under more extreme future conditions, are likely to expand the suitable habitat for W. carpophilus. These findings underscore the pathogen’s vulnerability to changing climatic factors and emphasize the expanding risk it poses to new regions, particularly in Europe, North America, and parts of Asia.To effectively mitigate the risks associated with W. carpophilus, there is a pressing need for continuous monitoring and adaptive strategies, particularly in areas predicted to become more conducive to the pathogen’s establishment. Proactive surveillance, early detection, and rapid response will be crucial in minimizing the impact of W. carpophilus outbreaks, particularly in vulnerable agricultural and forestry regions. As climate change accelerates, adaptive disease management strategies must be prioritized, and predictive models should be further refined and integrated into policy frameworks to ensure a resilient and sustainable approach to managing emerging plant pathogens.