Ticks, belonging to the Ixodoidea family, stand out as critical vectors for diseases affecting both humans and animals, surpassing other invertebrate carriers in the diversity of pathogens they transmitted. Tick-borne rickettsioses represent a globally distributed zoonotic threat caused by the obligate intracellular bacteria, rickettsia (1). Japanese spotted fever (JSF) emerges as a severe acute tick-borne disease, attributed to an alpha-proteobacterial agent (Rickettsia japonica) belonging to the spotted fever group within the genus Rickettsia (2). The presence of R. japonica has been documented in various regions, including Japan, South Korea, and Thailand, with reports dating back to 1984 (3). Subsequent findings have identified the bacterium in both ixodes ticks and the blood of individuals in central, southeastern, and northeastern China (4–6). A large number of studies have shown that Haemaphysalis longicornis is the main vector for R. japonica (5, 7, 8). The clinical manifestation of JSF is characterized by a sudden onset of symptoms, including headache, fever, shaking chills, skin eruptions, tick bite eschars, and overall malaise (9). Research indicates that following infection with JSF, if not promptly treated, there is a possibility of developing eruptive purpura, disseminated intravascular coagulation complications, and multiple organ failure (10, 11). Hence, it is crucial to conduct public health surveillance to detect cases and outbreaks promptly, enabling the implementation of preventive measures. This underscores the significance of identifying potential high-risk areas (12).

Global climate change is reshaping the distribution and density of tick populations by altering annual mean temperature, annual precipitation, and temperature seasonality, leading to extreme dry months (13). Climate significantly influences species distribution by modifying life cycles and environmental variables such as temperature, precipitation, and wind speed. These changes can directly impact ecology, resulting in population increases or decreases, and potentially leading to species evolution or extinction (14). Consequently, climate change poses an urgent and potentially irreversible threat to human societies, fauna, and vectors. According to the IPCC report, human activities have caused global warming of approximately 1°C above preindustrial levels, with a probable range of 0.8°C to 1.2°C. It is expected that global warming will reach 1.5°C between 2030 and 2052 if it continues at the current rate (15). Numerous studies and models (16–18) predict a significant expansion in tick habitats due to climate change. The northern global climate is anticipated to become warmer and wetter, facilitating the survival and adaptation of ticks like Haemaphysalis longicornis (H. longicornis) in many regions (19). This expansion of ticks has been observed on a broad scale across regions like Japan and the wider Asian continent (20). Therefore, the growing need to predict the impact of global changes has driven ecology towards becoming a more predictive science (21).

Even though the precise worldwide distribution of JSF remains unclear, ecological niche models have established associations between species occurrence and diverse environmental and climatic factors. These models enable the anticipation of distribution maps (22). GARP, Bioclim, Domain, and Maxent are extensively utilized for predicting species distribution, with the Maximum Entropy model (Maxent) proving particularly effective on a global scale (23, 24). Maxent asserts that, given precisely stated prior data, the probability distribution that most accurately reflects the current knowledge state of the system is the one with maximum entropy. Lorenz et al. employed Maxent modeling to identify regions with the highest prevalence of Mayaro virus occurrence in South America. Their study, based on confirmed cases, serological detection spanning the last two decades, and socio-environmental variables, aimed to map out areas of heightened risk for Mayaro virus presence in the region (12). Cheng et al. utilized epidemiological models and Maxent to assess the potential risk map of the spread of Usutu virus in Europe, providing data for the epidemiology and risk assessment of disease transmission (25).

There are few studies on risk prediction of R. japonica and JSF, and the classification of risk areas is crucial for formulating public policy related to control and prevention. Given this context, we used overlapping datasets on the global distribution of R. japonica potential vectors ticks and R. japonica to identify potential areas where JSF occurs on a global scale and assess the influential bioclimatic variables for JSF using Maxent. Furthermore, it also aims to identify prospective areas for conservation planning under future climate change scenarios.

In this study, three common metrics, AUC, TSS, and Kappa, were selected as the basis for the assessment of model accuracy. We obtained an AUC value of 0.990, a TSS value of 0.857, and a Kappa value of 0.763 for the model of R. japonica. The mean omission of test data of R. japonica was in general agreement with the predicted omission (Figure 1A), indicating that the models were well constructed. The AUC value (Figure 1B) was close to 1; The TSS value was >0.85, and the Kappa value was >0.70. According to the model accuracy evaluation reference standards (Table 2), the results of the three evaluation indicators suggest good to excellent levels. These results indicate that the obtained Maxent model predictions were highly accurate. Maxent model revealed that the most influential climatic variables affecting the suitability of R. japonica, were average precipitation in June (prec6), precipitation of warmest quarter (bio18),temperature seasonality (bio4), minimum temperature in August (tmin8), elevation (ele), average precipitation in September (prec9), precipitation of driest month (bio14), isothermality (bio3), minimum temperature in January (tmin1), average precipitation in August (prec8), average precipitation in November (prec11). Pearson correlation test was further conducted on the selected climatic variables (Figure 2). Variables exhibiting high correlation but low contribution were subsequently removed, resulting in the retention of eight climatic variables in the final analysis (Table 1).

Given the substantial contribution rates of prec6 and bio4, we conducted a univariate analysis to elucidate their individual influences. As depicted in Figure 3A, the probability of R. japonica presence exhibited an initial increase followed by a subsequent decrease with the rise in Average Precipitation in June. Notably, within the 160–600 mm interval of Average Precipitation in June, the probability exceeded 0.7, indicating highly suitable range. The probability reached its peak value (approximately 0.9) when the Average Precipitation in June was around 200 mm. Furthermore, the Temperature Seasonality variation coefficient within highly suitable habitats ranged from 600 to 900, with a peak around 800, as illustrated in Figure 3B.

The distribution of R. japonica occurrence points was predominantly in East Asia, Southeast Asia, and several European countries (Figure 4). The predicted potentially suitable area for R. japonica under historical climate conditions spanned 17.67 million square kilometers. Within these suitable regions, the proportions of highly suitable, moderately suitable, and less suitable areas were 7.05, 14.64, and 78.31%, respectively. Highly and moderately suitable areas were primarily concentrated in Eastern Asia (including the southeast of China, Korea, and Japan), and Southern Asia (encompassing the east of India, Thailand, Cambodia, Myanmar, Laos, Bangladesh, Vietnam, and the Philippines). Marginally suitable areas were observed not only in the vicinity of moderately and highly suitable regions but also in Europe (including Norway, United Kingdom, Netherlands, Belgium, France, Germany, Switzerland, Austria, Latvia, Estonia, Lithuania, Sweden, Poland, Czech Republic, Italy, Slovakia, Hungary, Croatia, Bosnia and Herzegovina, Ukraine, Romania, Russia, Kazakhstan), in the middle and south of Australia and New Zealand.

Among 16 future climate scenarios, the potential range of middle to high suitable areas for R. japonica expanded, encompassing regions such as China, Bhutan, Bangladesh, India, Vietnam, Laos, Burma, Cambodia, Thailand and other countries in Southeast Asia (Figure 5). Under the SSP5-8.5 scenario model, the projected maximum extent of the suitable area for Asia and its adjacent regions is anticipated to reach 1525.51 × 10⁴ km², with an area change rate of 118.65% from 2021 to 2040. The maximum area of high suitability for R. japonica across different time periods is observed in the SSP3-7.0 scenario for the years 2041–2060, where the high suitability area reached its peak at 193.42 × 10⁴ km², compared to the current suitable area of 95.13 × 10⁴ km² (Table 3).

In the historical climate scenario, China’s suitable areas are primarily concentrated in the southeast coastal regions, as well as the cities along the Qinling Mountains and the Huai River. Moderately and less suitable areas are predominantly found in Liaoning, Hebei, Tianjin, Beijing, Shandong, Shaanxi, and Yunnan. Highly suitable areas are distributed across Henan, Jiangsu, Sichuan, Chongqing, Hubei, Anhui, Hunan, Jiangxi, Zhejiang, Guangdong, Guangxi, Fujian, Hainan, and Taiwan (Figure 6). Under the SSP5-8.5 scenario models, the suitable area of R. japonica in China during 2021–2040 is the largest, which is 3.95 × 10⁶ km², and the area change rate is 50.42%. The areas potentially suitable for R. japonica under each climate scenario demonstrate an overall increasing trend in the future (Table 4), with suitability areas gradually expanding northward.

We employed the Maxent model with optimized parameters to forecast the future suitable areas of R. japonica Asian and adjacent regions under four SSPs scenarios, serving as an early warning. Our findings indicate that mean precipitation in June, temperature seasonality, minimum temperature in August, and elevation are the primary factors influencing the potential geographical distribution of R. japonica, with mean precipitation in June exhibiting the greatest contribution rates and permutation importance. The projected suitable areas of R. japonica in different periods under each climate scenario, both Asian and within China, surpass those under the current climate scenario. In Asian and adjacent regions, Japan emerges as the most favorable region for R. japonica. In Southeast Asia, suitable zones transition from moderate-low to moderate-high suitability. Moreover, potential suitable zones in Europe show signs of expansion. Under future climate scenarios, the highly suitable areas for R. japonica in central and southern China are projected to increase, with the suitable areas gradually expanding northward.

Currently, there were few literatures exploring the relationship between climate and JSF. The majority of research were concentrated on clinical symptoms of cases and serological testing (9, 40–43). According to the latest monitoring data from Japan, the incidence rate of JSF had been consistently rising until 2016 (44). It was observed that JSF had become more prevalent among the older adult population by 2020. More importantly, the incidence of JSF is increasing in areas not previously considered high-incidence areas (45). We suspect that this phenomenon is related to changes in temperature and increased precipitation. It was predicted that Tick-Borne Diseases (TBD) would become prevalent in new regions due to global warming, prompting ticks that prefer warmer environments to become more active (17, 46). As observed in mainland China, climate factors such as temperature, sunshine duration, precipitation, and seasonal indicators are correlated with the incidence rate of TBD (19). Our research has also found the average precipitation in June, temperature seasonality, and the minimum temperature in August contributed strongly to the distribution of R. japonica. In regions with high precipitation and relatively high temperatures, the suitable habitat area for R. japonica is larger. These findings were consistent with a study conducted in Japan, which revealed that the incidence rate of JSF is related to year-round seasonal temperature fluctuations (47). The possible reason for this is the continuous increase in the population size of the vector ticks, allowing them to attack hosts even during winter. The prevalence of R. japonica is mainly from April to December, and the vector ticks activity shift with the change of climate, which can attack the hosts at the end of calendar winter and in early spring (48).

In southern China, the distribution areas of R. japonica share the same Köppen climate classification, known as Cfa, characterized by a temperate climate with hot summers and an absence of dry seasons. This climate type entails a minimum monthly mean temperature ranging from 0 to 18°C and a maximum monthly mean temperature of 22°C or higher throughout the year, coupled with high humidity (31). Over the past 40 years, the Cfa climate classification has notably expanded in China, mirroring the findings of our study, where highly suitable areas for R. japonica have demonstrated a tendency to expand in southern China. Moreover, within suitable areas of China, R. japonica has expanded northward. This expansion can primarily be attributed to the projected warmer and wetter climate in the North, which enhances the climate suitability for tick survival in many regions (49, 50).

The expansion of the suitable area implies an increasing public health risk for the population residing within these regions. We unexpectedly found that the incidence of JSF is increasing in eastern Japan (Fukushima, Ibaraki, Tochigi, Gunma, and Shizuoka Prefectures), where the disease has previously been regarded as non-prevalent regions (51, 52). Meanwhile, reports of JSF cases in China have also gradually increased over the past 5 years (53–56). Therefore, it is imperative to enhance health education regarding tick prevention and control measures among the population residing in the potentially suitable areas, disseminating knowledge about the course and etiology of JSF among healthcare professionals as well as in society (57). Simultaneously, promoting and guiding individuals to take protective measures to prevent tick contact and bites. This includes using repellents, wearing untreated or permethrin-treated protective clothing, and conducting tick checks after entering indoor spaces, followed by showering to assist in detecting ticks on the skin (58). In addition, regular monitoring of tick density and activity, assessing the potential health risks posed by ticks to humans and animals, effective prevention through environmental modification, biological control, and physical barriers, developing an early warning system for emerging tick-borne diseases, and conducting targeted vaccination campaigns in high-incidence areas (59).

Our research presents distinct advantages. Firstly, there is currently no existing study on the risk prediction of R. japonica and JSF. This study introduces a systematic approach for Asian and adjacent regions predicting the distribution of R. japonica. The Maxent model, based on the principle of maximum entropy, offers several advantages: it exhibits higher accuracy in prediction outcomes and is capable of achieving superior modeling results with smaller sample sizes, which is importantly notable (60, 61). Additionally, Maxent facilitates the identification and interpretation of non-linear responses. Contributions and Jackknife analyses promptly highlight crucial variables, while response curves illustrate the variation in probabilities with dependent variables, providing comprehensive information (62). Maxent further excels over other models such as GLMs, GAMs, BIOCLIM, and GARP in predictive accuracy due to its ability to model complex environmental variables (63). Furthermore, predictions were enhanced by overlaying coordinates of H. longicornis and R. japonica, significantly improving the accuracy of risk distribution forecasts. This approach identifies numerous prospective regions susceptible to R. japonica invasion in China. While not yet intercepted or detected in China, these regions could potentially be colonized under specific conditions. Lastly, the expansion of suitable areas for R. japonica towards northern China is attributed to global warming and the significant increase in the Cfa climate, highlighting a notable trend in our findings.

Several limitations were identified in our study that highlight areas for improvement in future research. Firstly, our reliance on historical climate data spanning from 1970 to 2000 restricted access to current climate information, potentially impacting the accuracy of current suitability estimates. Moreover, the exclusive use of the Maxent model focused solely on species’ climatic adaptability and specific climatic variables, neglecting the inclusion of other critical factors. To enhance precision in estimating distribution within suitable habitats, future studies should incorporate a broader range of variables including soil moisture levels, vegetation cover types, population densities, tick vector densities, and land use patterns. Finally, and of critical importance, due to the limitations of current data, the geographical scope for future predictions is limited to analysis of Asia and adjacent regions. Although this study has enhanced model performance by refining feature selection and adjusting regularization parameters to optimize model fitting, it is imperative to exercise caution when extrapolating the model. Future research should focus on collecting more data from various regions to further validate and refine the model.

Detecting potential regions suitable for species invasion holds significant importance in preventing alien incursions. Utilizing the Maxent model, we evaluated the potential invasion ranges of R. japonica under both current and future climate conditions. Our findings reveal that the potential invasion range of R. japonica predominantly spans subtropical and warmer temperate regions, encompassing areas such as Japan, South Korea, southeast China, various southeast Asian countries, and European nations. Temperature seasonality and precipitation emerged as pivotal climatic factors influencing the potential distribution of R. japonica. Looking ahead, the invasion range of R. japonica is poised to expand globally. Within China, the southeastern region witnesses an increase in high-suitability areas, with a discernible trend of expansion towards the north. These insights furnish valuable reference points for devising appropriate management strategies aimed at thwarting the establishment and further dissemination of R. japonica throughout the world.