Assessing the ecological resilience of Ebola virus in Africa and potential influencing factors based on a synthesized model

Assessing the ecological resilience of Ebola virus in Africa and potential influencing factors based on a synthesized model | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Assessing the ecological resilience of Ebola virus in Africa and potential influencing factors based on a synthesized model Li Shen, Jiawei Song, Yibo Zhou, Xiaojie Yuan, Samuel Seery, Ting Fu, and 5 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3899519/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background The Ebola epidemic has persisted in Africa since it was firstly identified in 1976. However, few studies have focused on spatiotemporally assessing the ecological adaptability of this virus and the influence of multiple factors on outbreaks. This study quantitatively explores the ecological adaptability of Ebola virus and its response to different potential natural and anthropogenic factors from a spatiotemporal perspective. Methods Based on historical Ebola cases and relevant environmental factors collected from 2014 to 2022 in Africa, the spatiotemporal distribution of Ebola adaptability is characterized by integrating four distinct ecological models into one synthesized spatiotemporal framework. Maxent and Generalized Additive Models were applied to further reveal the potential responses of the Ebola virus niche to its ever changing environments. Results Ebola habitats appear to aggregate across the sub-Saharan region and in north Zambia and Angola, covering approximately 16% of the African continent. Countries presently unaffected by Ebola but at increased risk include Ethiopia, Tanzania, Côte d'Ivoire, Ghana, Cameroon, and Rwanda. In addition, among the thirteen key influencing factors, temperature seasonality and population density were identified as significantly influencing the ecological adaptability of Ebola. Specifically, those regions were prone to minimal temperature variations between seasons. Both the potential anthropogenic influence and vegetation coverage have a rise-to-decline impact on the outbreaks of Ebola virus across Africa. Conclusions Our findings suggest new ways to effectively respond to smaller potential Ebola outbreaks in Sub-Saharan Africa. We believe that this integrated modeling approach and response analysis provide a framework that can be extended to predict risk of similar epidemiological studies for other diseases across the world. Ebola ecological adaptability ecological niche models spatiotemporal distribution influencing factors Figures Figure 1 Figure 2 Figure 3 Figure 4 1. Background Ebola virus disease (EVD), also known as Ebola hemorrhagic fever, is an acute and severe infectious disease caused by the Ebola virus, a filamentous virus in the family Filoviridae [ 1 , 2 ]. As one of the most lethal viral diseases in human history, EVD typically causes organ damage, high fever, internal bleeding, and diarrhea with a high transmission rate and a case fatality ratio of between 50 and 90% [ 3 ]. The disease was first discovered in Sudan and the Democratic Republic of Congo in Africa in 1976 [ 4 ], and has since emerged in Central and West African nations [ 5 ], affecting Uganda, Democratic Republic of Congo (DRC), and Nigeria [ 6 ]. However over the past decade, three major pandemics have occurred: the 2014 West African Ebola outbreak, the 2018 outbreak in DRC, and the 2022 outbreak in Uganda, with the first one recorded as the largest outbreak to date [ 7 , 8 ]. As a result of these outbreaks, many scholars have investigated the influence of different factors on the dynamics of EVD outbreaks by constructing and optimizing traditional transmission models (e.g., Poisson transmission models, Susceptible-Infectious-Recovered models, and Susceptible-Exposed-Infectious-Recovered models), simulating and predicting Ebola outbreak trends [ 9 , 10 ]. These studies mainly consider the spread of EVD based on frequent interactions between the uninfected agents and infected animals or humans that have been exposed to the virus. They have also used general models to simulate the processes of transmission for predicting potential risk of EVD epidemics by optimizing epidemiological parameters. In recent years, epidemiological ecology has received increased attention to seek opportunities to examine relationships between certain factors and disease transmission [ 11 , 12 ], such as using time-series regression analysis to estimate both short-term and long-term associations between environmental exposures and health outcomes [ 13 , 14 ]. In addition, from a phenological perspective, the distribution and spread of filoviruses can be effectively predicted by exploring the environmental conditions required for the survival of pathogens [ 15 , 16 ]. This could help reduce uncertainty associated with traditional transmission models and can be applied to specific infectious diseases. However, most of these traditional mathematical models neglect to incorporate the spatial variation of Ebola pathogens for revealing its potential outbreak source and sink areas. In addition, these studies mainly regard EVD as an infectious disease to predict its scale change, but rarely consider specific transmission processes of the virus, resulting in sudden and rapid predictions with relatively low accuracy. In addition, these models have paid little attention to examining the impacts of both natural and anthropogenic factors on the host movement from a spatiotemporal perspective. Therefore, we conducted this study to quantitatively explore the spatiotemporal distribution of Ebola outbreaks in Africa as well as the impacts of environmental factors from an epidemiological ecology perspective. There were three specific objectives: 1) to predict the spatiotemporal distribution of Ebola by integrating various ecological niche models; 2) to investigating the pivotal factors that can affect the suitable niches of Ebola; and 3) to reveal the ecological adaptability of Ebola virus by characterizing its response to the influencing factors. 2. Materials and Methods 2.1 Study area and status Since the first recorded outbreaks in Sudan and the DRC in 1976, the transmission of the EVD has predominantly been observed within the African continent. This study focused on Western and Central African nations where the most recent outbreaks occurred. According to the World Health Organization (WHO), two major outbreaks have been recorded since 2014. The first occurred in West Africa with over 28,600 infections and 11,325 fatalities in 2014 [ 7 , 17 ]. The second was in 2018 in the DRC, and had a total of 3,470 cases with 2,287 fatalities. Since then, there have been sporadic, relatively small outbreaks across Africa, in the DRC from 2018 to 2022, Guinea in February 2021, and Uganda in September 2022 as shown in Fig. 1 . However, there are established epidemiological and genetic factors which link these outbreaks with the 2014 outbreak. Therefore, this study investigated the ecological adaptability of the EVD based on outbreaks since the West Africa nations outbreak in 2014. 2.2 Data Preparation 2.2.1 Surveyed EVD Cases Long-term EVD surveillance data was used in this study and included cases from the 2014–2016 outbreaks, the 2018 DRC outbreak, and the 2022 outbreak in Uganda. The first two groups of data were obtained from The Humanitarian Data Exchange (HDX) website ( https://data.humdata.org/ ) which serves as an open platform for facilitating easily accessible and useable data for this analysis. This data source includes data from over 250 countries and territories. Epidemic data were primarily based on information provided by the WHO, along with real-time reporting by national health authorities of different countries. Data from the 2022 Uganda outbreak were acquired from the Global Health website ( https://global.health/ ) which is purported to provide accurate real-time disease data in the early stages of an outbreak, tracking cases for the first 100 days. Geographical points of each case were collected from the aforementioned sources and then abstracted by locating them in the center of a specific region or in proximity to the medical facilities. This pre-process helped to avoid potential conflicts or errors without affecting the reliability and accuracy of the final results. 2.2.2 Environmental Data Previous research has demonstrated that the transmission of Ebola is influenced by multiple environmental factors such as climate conditions (e.g. precipitation and temperature) and vegetation abundance (e.g. deforestation and the consumption of forest prey) [ 18 ]. In this study, climate data were primarily obtained from WorldClim ( https://www.worldclim.org/ ), a database that offers high-resolution weather and climate data at a global scale. Surface coverage data were obtained from the Copernicus Climate Change Service ( https://climate.copernicus.eu/ ), which combines climate system observations with scientific research to produce authoritative and quality-assured information on European and global climate. In addition, vegetation information was acquired from the global map data archive established by the Global Map Transfer Program and long-term time series normalized difference vegetation index (NDVI) data were retrieved from NOAA's Advanced Very High Resolution Radiometer (AVHRR) sensor. 2.2.3 Human Impact Data Demographic data were extracted from Gridded Population of the World, version 4 data (GPW.v4) which is based on national censuses and population registers. We also utilized Human Influence Index (HII) grids data set which contains information such as population density raster, human land use raster, and constructed roads. In addition, night time light remote sensing data were also utilized due to its strong correlation with socioeconomic factors, which was obtained from the Operational LineScan System (OLS) of the U.S. Defense Meteorological Satellite Program (DMSP) and the VIIRS Plus DMSP light change dataset. All the aforementioned data were carefully examined and cleaned to meet the subsequent analysis. Finally, after integrating the epidemic case data, remote sensing data, surveyed attribute data, and the spatial vector boundaries, we used ArcGIS (version 10.8) software to establish a geospatial database for further analysis. 2.2.4 Variable Filter To prevent overfitting and enhance the interpretability of our models, we conducted variable filtering to select significant data. Initially, we implemented a JackKnife method based on the Maxent model and utilized the XGboost algorithm to assess the significance of each factor. In order to address the issue of strong interrelationships among predictor variables, which hampers the interpretation of variables [ 19 ]. we utilized Spearman’s correlation coefficient to remove variables exhibiting substantial correlation, thereby enhancing the precision of the predictor variables. This process retains representative and more practically significant factors, ultimately selecting 13 variables that are most likely to influence Ebola distribution, namely Isothermality (bio3), Temperature Seasonality (bio4), Min Temperature of Coldest Month (bio6), Temperature Annual Range (bio7), Mean Temperature of Coldest Quarter (bio11), Annual Precipitation (bio12), Precipitation of Driest Month (bio14), Gross Primary Productivity (GPP), Gridded Population of the World (GPW), Human Influence Index (HII), Land Cover Classification System (LCCS), Night Lights (light) and NDVI. 2.3 Methods Spatial-temporal distribution of EVD in Africa from 2014 to 2023 were visualized by ArcGIS (version 10.8) software. In order to accurately estimate the current spatiotemporal distribution of EVD as well as its potential risk, we established a synthesized predictive model by integrating four distinct ecological niche approaches which are Maxent, Bioclim, Domain and GARP models. The Maxent model provides probabilities ranging from 0 to 1, indicating the likelihood of Ebola presence. The Bioclim and Domain models yield larger values for regions more prone to Ebola outbreaks, and the GARP model distinguishes the study area into suitable and unsuitable regions. The description of those four models can be found in the supplementary materials (Appendix D). Our proposed synthesized model is given by $$\begin{array}{c}{\widehat{{y}_{j}}}^{i}={f}_{i}\left({{X}^{{\prime }}}_{j},{{\Theta }}_{i}\right)\#（1）\end{array}$$ where \(i\) represents the type of ecological niche model, \(j\) represents the prediction unit, \(f\) denotes the model mapping, and \(X\) is the feature vector used to describe multidimensional influencing factors. After training the model, input features \({{X}^{{\prime }}}_{j}\) are applied to each model for prediction, resulting in predicted values \({\widehat{{y}_{j}}}^{i}.\) By adjusting the model parameters \({{\Theta }}_{i}\) , we aimed to minimize the difference between the predicted values \({f}_{i}\left({X}_{j},{{\Theta }}_{i}\right)\) and the actual values \({y}_{j}\) , which is expressed as: $$\begin{array}{c}\underset{{{\Theta }}_{i}}{min} {\left({y}_{j}-{f}_{i}\left({X}_{j},{{\Theta }}_{i}\right)\right)}^{2}\#（2）\end{array}$$ Finally, the predicted values of the four models are weighted and summed to obtain the ultimate ecological niche prediction \({\widehat{\varvec{y}}}_{j}^{inter}\) . The weights \({W}_{i}\) are determined based on prediction accuracy and are calculated as follows: $$\begin{array}{c}\begin{array}{c}{\widehat{\varvec{y}}}_{j}^{inter}=\sum _{i=1}^{4} {W}_{i}\cdot {\widehat{\varvec{y}}}_{j}^{i}\end{array}\#\left(3\right)\end{array}$$ Based on the predictive accuracy of each model, the final outcome was retrieved by incorporating different weights according to the following standards: (1) the plotted Receiver Operating Characteristic curves (ROC); (2) the calculated Area Under the Curve (AUC) values, a widely used metric for evaluating the classification performance of heavily imbalanced data; (3) normalizing the different suitability representations of into five levels (1 for Not Suitable, 2 for Less Suitable, 3 for Moderate, 4 for More Suitable, and 5 for Suitable); (4) assigning weights to each model based on the corresponding AUC values, which were then standardized using a grid-based calculation method to obtain the final result. A schematic representation of the entire process was depicted in Figure S1 . The Maxent model and GAMs (generalized additive models) (see Appendix D) were performed to understand the influence of each environmental factor on the predicted risk. Throughout the modeling process, all other environmental variables were set the mean value but with one specific variable estimated. The resulting Maxent-based Environmental Marginal Response Curves demonstrated how the predicted probability varies with changes in the variable, thereby revealing the correspondence between ecological viral adaptability with certain variables. 3. Results 3.1 Ecological adaptability assessment of Ebola Each of the four models had distinctive characteristics for predicting Ebola occurrences across different regions. Since the Maxnet, Bioclim, and Domain models all derived specific values while the GARP model provides logistic variables, it is necessary to standardize these different results for effective comparison. Therefore, we reclassified values obtained from the first three model results shown in Fig. 2 a, b and c with darker colors representing higher adaptability of Ebola. The Maxent and Bioclim models demonstrated similar predictions for favorable habitats, primarily concentrated in West Africa, sub-Saharan nations, along the edge of the Congolese basin, around Lake Victoria, and on the Ethiopian plateau. These ecologically suitable zones were generally situated near the equator. In contrast, the Domain and GARP models utilized more stringent criteria to identify non-viable areas, and predicted a broader range of potential zones distributed across the African continent, spanning from sub-Saharan regions to the Kalahari Desert in the South African plateau. According to the accuracy assessment of ROC curves illustrated by Fig. 3 (e, f, g, and h), the AUC values derived from these four predict models were all relatively high with Maxent (0.99), Bioclim (0.92), Domain (0.93) and GARP (0.85), respectively. In order to synthesize the predictions of the four ecological niche models, we assigned 30%, 30%, 30%, and 10% weights to the four models mentione according to their respective accuracies, and obtained the integrated danger levels of Ebola outbreaks as well as the potential habitat of EVD in Africa as shown by Figure S2 (a), while higher levels indicate a greater risk of an Ebola outbreak. The areas at most risk predominantly lay in the western, central, and eastern regions of Africa, situated south of the Sahara Desert but close to the equator as illustrated by Figure S2 (b, c, d and e). Specific countries included Guinea, Liberia, Sierra Leone, Nigeria, DRC, Uganda, Ethiopia, among other. Using the Albers Equal Area Conic projection within corresponding coordinate system, the calculated total area occupied by these high-risk regions (areas above level moderately high in the map) was approximately 3, 509, 623.11 km2, accounting for nearly 11.61% of Africa’s entire land area. By integrating population density grids, we estimated the total population affected in each country located within the high-risk areas depicted in Figure S3. The red bars in Figure S3 (c) represented countries without reported Ebola outbreaks but with a substantial population at risk of infection. The blue bars signify countries with a history of Ebola outbreaks and documented cases. Based on the current predictive results, countries at higher risk yet without a history of Ebola include Ethiopia, Tanzania, Côte d'Ivoire, Ghana, Cameroon, Rwanda, Burundi, and others. Côte d'Ivoire has reported one single Ebola case in 2021 but this did not lead to a further outbreak. Countries that have experienced recent Ebola outbreaks (e.g., Guinea, the DRC, and Liberia), as well as those that had outbreaks in earlier periods (e.g., Gabon, South Africa, and the Sudan), all matched with the predicted Ebola-adapted zones, providing a good validation for the reliability 3.2 Significance analysis of impact factors Given the superior performance of the Maxent model in accuracy assessment owing to a high AUC value of 0.99, this study conducted an impact factor analysis to examine the crucial environmental variables for predicting Ebola-suitable habitats. Table 1 presented the contribution and permutation importance of 13 environmental variables to the ecological adaptability prediction of EVD, which was derived from the Maxent modeling. Averages of 100 experiments revealed that bio4, GPW, bio12, LCCS, and bio6 acted in rank as pivotal factors that can affect the suitable niches of Ebola. These key variables accounted for a cumulative percentage contribution of 97.8% and permutation importance of 98.5%. Table 1 Contribution and permutation importance of variables derived through Maxent modeling (average of 100 experiments). Variable Contribution (%) Permutation Importance bio4 44.89 48.19 GPW 33.43 32.69 bio12 9.76 9.09 LCCS 7.21 5.67 bio6 2.47 2.82 HII 1.55 0.59 bio3 0.52 0.44 GPP 0.14 0.43 NDVI 0 0.06 light 0 0 bio7 0 0 bio14 0 0 bio11 0 0 Jackknife testing (Fig. 3 ) showed that when considering individual environmental variables, bio3, bio4, bio11, bio12, GPP, and GPW had the most significant impact on regularized training gains, test gains, and AUC values. This suggests that these variables might have the most useful information and are crucial for the adaptability of the Ebola virus [ 20 ]. However, after excluding individual environmental variables, GPW, LCCS, and bio12 had the largest reductions in gains and AUC values, indicating that these variables hold more unique information and played an irreplaceable role compared to others. After combining these findings, it became evident that bio4 (Temperature Seasonality), bio12 (Annual Precipitation), and GPW (Population Density) exerted a significant impact on the suitable niches of Ebola. 3.3 Response of impact factors To generate response curves for different variables, we utilized individual variables to construct the Maxent model. Consequently, we obtained the model’s predicted probabilities as a function of increasing the values of each environmental factor (Figure S4). The x-axis represented values of the environmental variables, while the y-axis corresponded to the consequent Ebola suitability probability. The line graph demonstrated the dependency of predicted suitability on the selected variable, as well as the dependency arising from the correlation between the chosen variable and other variables. Since HII and LCCS were categorical variables, the results have been illustrated using a standard bar chart. Based on the GAM model, we investigated the relationship between each influencing factor and Ebola’s ecological adaptability. To achieve this, we generated 1000 random points in the study area and extracted the corresponding environmental data and Maxent model results for these points. The Maxent model results were used as the response variable, and we examined how the environmental variables influenced the model outcomes (Fig. 4 ). Each scatter plot in the figure represents a random point, with the x-axis representing the variable’s value and the y-axis showing its impact on Ebola’s ecological adaptability. Vertical lines on the x-axis depicted data distribution and density, while the light purple background signified the confidence interval. These lines can represent the fitting of the data points. 4. Discussion As an interdisciplinary piece of research, this study attempted to investigate suitable ecological niches for Ebola virus as well as its potential responses to different environmental factors. This study provides a practical example of how to apply geostatistical approaches to quantitatively explore the dynamic distribution, variation, and influential determinants of an epidemic, not only demonstrating an opportunity to address certain research gaps in the existing literature, but also presenting some potential to support public health decision-makers in formulating and implementing effective strategies. Regarding spatiotemporal predictions for Ebola habitats at a macroscopic scale, the forecasted outcomes of the integrated model mainly aggregate to the south of the Sahara Desert and north of Zambia and Angola, covering approximately 16% of the entire African continent. This finding relatively matches with the zoonotic niche map of Ebola virus revealed by Pigott et al.[ 21 ], but in comparison, our high-risk areas are more dispersed and better correspond to the previous outbreak patterns [ 22 ]. This may be attributed to the utilization of the latest infection case data, including the 2018 Congolese outbreak and the 2022 Uganda outbreak. Additionally, we incorporated a more comprehensive set of potential influencing factors, which improved identifying the most significant ecological determinants that possibly affect the adaptability of Ebola virus in Africa, thereby further increasing the accuracy of final prediction. Moreover, our results are supported by Redding et al [ 23 ]. who developed an Ebola system-dynamics model that yielded a similar prediction. However, our study integrated each unique feature of the four ecological niche models, achieving higher AUC values and able to provide response curves for various influencing factors, thus offering a better understanding of the association between Ebola outbreaks and multiple environmental factors. Nevertheless, our results provide only a weak explanation to distinguish indexed cases and secondary cases. Our research also compiled the demographic statistics within Ebola outbreak-prone regions across different African countries, estimating the total population at risk in these high-risk areas to be 20,504,885. Among this population, 11,464,743 individuals reside in regions with reported Ebola outbreaks, while 9,040,142 individuals live in areas with no reported outbreaks. Notably, Ethiopia, one of the countries with the highest at-risk population, although not previously documented with Ebola cases, has been identified as having the potential risk for an Ebola outbreak in our study, similar to the findings from Redding et al [ 23 ]. Both researches highlighted environmental factors in Ethiopia are conducive to a potential Ebola outbreak, warranting serious concern from local governments and global public health institutions. Other countries at high risk of Ebola without previous reports include Tanzania, Cote d'Ivoire, Ghana, Cameroon, and Rwanda, which are urged to promptly implement effective prevention measures against potential Ebola epidemics. Furthermore, this study scrutinized the responses of various influencing factors and explained partial triggers for Ebola outbreak from a spatiotemporal perspective. Utilizing the Jackknife test and Maxent model, we identified the key factors closely associated with Ebola's ecological adaptability, notably including bio4, GPW, bio12 and LCCS. In order to further investigate how these influencing factors affect Ebola ecological adaptability, we combined the outcomes yielded by Maxent and Generalized Additive Models to reveal the distinct patterns in Ebola distribution: (1) The ecological adaptability of Ebola virus decreases with a rise in temperature seasonality (bio4), indicating preference for regions with minimal temperature variation between seasons. This matches closely with the conclusions by Olivero et al. that the observation of Ebola outbreaks predominantly occurring in low-latitude areas near the equator [ 24 ]. Its adaptability also increases with higher winter average temperatures and annual precipitation (bio6, bio12), similar to the findings on environment temperature and precipitation trends [ 25 ]. (2) The ecological adaptability of Ebola virus escalates with greater population density (GPW), likely due to the rapid human-to-human virus transmission promoted by frequent interpersonal contacts [ 26 , 27 ]. Additionally, the adaptability increases initially and then decreases, becoming lower in more economically developed regions (HII), which can be explained by the fact that the areas less impacted by anthropogenic activities especially rapid urbanization are almost not susceptible to large-scale outbreaks of Ebola [ 28 ]. whereas the economically developed regions with better healthcare resources are able to effectively implement barrier care techniques to reduce the risk of infection [ 29 ]. (3) The ecological adaptability of Ebola virus follows a pattern of initial growth and then subsequent decline with vegetation increase, peaking within regions of moderate vegetation coverage. For equatorial area with more vegetation abundance, the reduction in vegetation possibly accelerates the risk of Ebola virus transmission. Other studies have also indicated that deforestation substantially elevates human interaction with infected wildlife, thereby favoring viral spread [ 30 ]. This perspective effectively corroborates our findings. (4) The ecological adaptability of Ebola virus responds positively to land cover and values of 10, 30, 60, 170, and 190 respectively represents drylands, croplands, broadleaf deciduous forests, forests, floodplain, and urban zones (LCCS), which is consistent with the findings of Redding et al. who claimed that the land-use/land-cover determines distributions across the different reservoir host species [ 23 ]. Nevertheless, we took a step further by incorporating a quantitative analysis, particularly focusing on the nuances within land cover types. Nonetheless, several limitations should be taken into consideration when interpreting our findings. Firstly, the available Ebola case data were restricted to the provincial level without accurate geographical location information (e.g., longitude and latitude) which cannot satisfy a finer spatial resolution of our analysis. Secondly, the evaluation did not take the hosts of Ebola virus into consideration due to the difficulty in obtaining such information, which may lead to some incomplete understanding of how those impact factors drive the risk of Ebola outbreaks. Thirdly, our study was focused on Africa where the major endemic areas distributed in, which is in necessity to be further extended to other regions such as the North and South America, Asia, Europe for a deepening exploration of the global outbreak and transmission of Ebola based on the One Health strategy. Therefore, further research is needed to promote a comprehensive forecast of Ebola across regions based on involving Ebola host data, higher-precision case data, and largescale geographical areas. 5. Conclusions This study integrated different ecological niche models to effectively predict the spatial distribution of suitable habitats for the Ebola virus. Our findings indicate that these suitable habitats are primarily concentrated in regions near the equator, south of the Sahara Desert. Additionally, our research delved into the response patterns of key influencing factors related to Ebola, elucidating specific relationships between Ebola outbreaks and variables such as climate, socio-economic conditions, vegetation, and land use types. These insights offer valuable directions for future research in the field of epidemic prevention and control. List Of Abbreviations AUC the Area Under the Curve DRC the Democratic Republic of Congo EVD Ebola virus disease GAM generalized additive model GPP Gross Primary Productivity GPW Gridded Population of the World data HDX the Humanitarian Data Exchange HII Human Influence Index NDVI Normalized difference vegetation index ROC Receiver Operating Characteristic curves WHO the World Health Organization Declarations Ethics approval and consent to participate Not applicable Consent for publication Not applicable Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Competing interests The authors declare that they have no competing interests. Funding This work was partly supported by grants from the National Natural Science Foundation of China [grant number 42201448 and 82273689], and Natural Science Foundation of Hubei Province [grant number 2022CFB610]. Authors' contributions All authors developed the study concept. LS, JS, and YZ designed and performed the analysis and wrote the original draft. XY, SS and TF performed the data collection and preparation. Reviewing and conceptualization was conducted by XL and YL. Supervision and review & editing was conducted by ZS, RL and KL. All co-authors provided input and critical review of the manuscript leading to the final version. All authors read and approved the final manuscript. Acknowledgements The project team would like to thank all the participants who consented to participate in the study. We thank our colleagues for their careful reading and editing of this manuscript. The funder had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication. References Feldmann H, Jones S, Klenk HD, Schnittler HJ: Ebola virus: from discovery to vaccine. Nature Reviews Immunology 2003, 3(8):677-685. Malvy D, McElroy AK, de Clerck H, Guenther S, van Griensven J: Ebola virus disease. Lancet 2019, 393(10174):936-948. Ansari AA: Clinical features and pathobiology of Ebolavirus infection. Journal of Autoimmunity 2014, 55:1-9. Colebunders R, Borchert N: Ebola haemorrhagic fever - a review. Journal of Infection 2000, 40(1):16-20. Pattyn S, van der Groen G, Jacob W, Piot P, Courteille G: Isolation of Marburg-like virus from a case of haemorrhagic fever in Zaire. Lancet (London, England) 1977, 1(8011):573-574. Jacob ST, Crozier I, Fischer WA, II, Hewlett A, Kraft CS, de La Vega M-A et al : Ebola virus disease. Nature Reviews Disease Primers 2020, 6(1). Kalenga OI, Moeti M, Sparrow A, Vinh-Kim N, Lucey D, Ghebreyesus TA: The Ongoing Ebola Epidemic in the Democratic Republic of Congo, 2018-2019. New England Journal of Medicine 2019, 381(4):373-383. Samarasekera U: Solidarity Against Ebola: an update. Lancet Microbe 2023, 4(3):E139-E139. Area I, Batarfi H, Losada J, Nieto JJ, Shammakh W, Torres A: On a fractional order Ebola epidemic model. Advances in Difference Equations 2015. Ristic B, Dawson P, Ieee: Real-time forecasting of an epidemic outbreak: Ebola 2014/2015 case study. In: 19th International Conference on Information Fusion (FUSION): 2016 Jul 05-08 2016; Heidelberg, GERMANY ; 2016: 1983-1990. Shahraiyni HT, Sodoudi S, Cubasch U: Weather conditions and their effect on the increase of the risk of type A acute aortic dissection onset in Berlin. International Journal of Biometeorology 2016, 60(8):1303-1305. Manlove K, Wilber M, White L, Bastille-Rousseau G, Yang A, Gilbertson MLJ et al : Defining an epidemiological landscape that connects movement ecology to pathogen transmission and pace-of-life. Ecology Letters 2022, 25(8):1760-1782. Bhaskaran K, Gasparrini A, Hajat S, Smeeth L, Armstrong B: Time series regression studies in environmental epidemiology. International Journal of Epidemiology 2013, 42(4):1187-1195. Kim H, Lee J-T, Fong KC, Bell ML: Alternative adjustment for seasonality and long-term time-trend in time-series analysis for long-term environmental exposures and disease counts. Bmc Medical Research Methodology 2021, 21(1). Peterson AT, Bauer JT, Mills JN: Ecologic and geographic distribution of filovirus disease. Emerging Infectious Diseases 2004, 10(1):40-47. Valero KCW, Isokpehi RD, Douglas NE, Sivasundaram S, Johnson B, Wootson K et al : Plant Phenology Supports the Multi-emergence Hypothesis for Ebola Spillover Events. Ecohealth 2018, 15(3):497-508. Coltart CEM, Lindsey B, Ghinai I, Johnson AM, Heymann DL: The Ebola outbreak, 2013-2016: old lessons for new epidemics. Philosophical Transactions of the Royal Society B-Biological Sciences 2017, 372(1721). Merler S, Ajelli M, Fumanelli L, Gomes MFC, Pastore y Piontti A, Rossi L et al : Spatiotemporal spread of the 2014 outbreak of Ebola virus disease in Liberia and the effectiveness of non-pharmaceutical interventions: a computational modelling analysis. Lancet Infectious Diseases 2015, 15(2):204-211. Kumar S, Stohlgren TJ, Chong GW: Spatial heterogeneity influences native and nonnative plant species richness. Ecology 2006, 87(12):3186-3199. Guo Y, Zhang S, Tang S, Pan J, Ren L, Tian X et al : Analysis of the prediction of the suitable distribution of Polygonatum kingianum under different climatic conditions based on the MaxEnt model. Frontiers in Earth Science 2023, 11. Pigott DM, Golding N, Mylne A, Huang Z, Henry AJ, Weiss DJ et al : Mapping the zoonotic niche of Ebola virus disease in Africa. Elife 2014, 3. Feldmann H, Sprecher A, Geisbert TW: Ebola. New England Journal of Medicine 2020, 382(19):1832-1842. Redding DW, Atkinson PM, Cunningham AA, Lo Iacono G, Moses LM, Wood JLN et al : Impacts of environmental and socio-economic factors on emergence and epidemic potential of Ebola in Africa. Nature Communications 2019, 10. Olivero J, Fa JE, Real R, Farfán MA, Márquez AL, Vargas JM et al : Mammalian biogeography and the Ebola virus in Africa. Mammal Review 2017, 47(1):24-37. Schmidt JP, Park AW, Kramer AM, Han BA, Alexander LW, Drake JM: Spatiotemporal Fluctuations and Triggers of Ebola Virus Spillover. Emerging Infectious Diseases 2017, 23(3):415-422. Matua GA, Van der Wal DM, Locsin RC: Ebola hemorrhagic fever outbreaks: strategies for effective epidemic management, containment and control. Brazilian Journal of Infectious Diseases 2015, 19(3):308-313. Longley JL, Danquah LO, Massa MS, Ross DA, Weiss HA: The impact of case and contact characteristics on contact tracing during the West Africa Ebola epidemic. Journal of Infection 2021, 83(4):501-504. Ali H, Dumbuya B, Hynie M, Idahosa P, Keil R, Perkins PJCC et al : The social and political dimensions of the Ebola response: Global inequality, climate change, and infectious disease. Climate Change and Health: Improving Resilience and Reducing Risks 2016:151-169. Grillet G, Marjanovic N, Diverrez JM, Tattevin P, Tadié JM, L'Her E: Intensive care medical procedures are more complicated, more stressful, and less comfortable with Ebola personal protective equipment: A simulation study. Journal of Infection 2015, 71(6):703-706. Olivero J, Fa JE, Real R, Marquez AL, Farfan MA, Vargas JM et al : Recent loss of closed forests is associated with Ebola virus disease outbreaks. Scientific Reports 2017, 7. Additional Declarations No competing interests reported. Supplementary Files Supplementalfiles.docx floatimage2.png Figure S1. Predicted ecologically suitable regions of Ebola virus based on Maxent, Bioclim, Domain, and GARP models with 13 independent, influential variables i.e., bio3, bio4, bio6, bio7, bio11, bio12, bio14, NDVI, GPP, GPW, HII, LCCS, light. floatimage5.png Figure S2. Map of potentially suitable habitats for Ebola virus in Africa and a closer look at four high-risk regions with (a) Integrated risk levels of Ebola based on results derived from four models, (b) Western Africa region, (c) Coastal regions of the Gulf of Guinea, (d) Central Africa region, and (e) Eastern African region. floatimage6.jpeg Figure S3. Populations in Ebola risk areas within each country with (a) Spatial distribution of countries within the Ebola risk zone with reported Ebola outbreak, (b) The spatial distribution of countries within the Ebola risk zone without any reported Ebola outbreaks, and (c) Histogram of populations at high risk of Ebola infection. floatimage8.png Figure S4. Response curves of environmental factors and the probability of suitability for Ebola based on Maxent model. Each curve (a-m) represents model-predicted probabilities associated with bio3, bio4, bio6, bio7, bio11, bio12, bio14, GPP, GPW, HII, LCCS, light, and NDVI. Curves display the mean response derived from 10 replicate Maxent runs (shown in red) with means and standard deviations (represented in blue, with two shades for categorical variables). Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-3899519","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":269789554,"identity":"82fa4eb1-ebfc-4367-8544-d9c4b07178fd","order_by":0,"name":"Li Shen","email":"","orcid":"","institution":"Wuhan University","correspondingAuthor":false,"prefix":"","firstName":"Li","middleName":"","lastName":"Shen","suffix":""},{"id":269789555,"identity":"88ae8590-1571-43d2-9e27-72f6c9647627","order_by":1,"name":"Jiawei Song","email":"","orcid":"","institution":"Wuhan University","correspondingAuthor":false,"prefix":"","firstName":"Jiawei","middleName":"","lastName":"Song","suffix":""},{"id":269789556,"identity":"1a301f94-b47c-4524-996d-f6713079002c","order_by":2,"name":"Yibo Zhou","email":"","orcid":"","institution":"Wuhan University","correspondingAuthor":false,"prefix":"","firstName":"Yibo","middleName":"","lastName":"Zhou","suffix":""},{"id":269789557,"identity":"0baa56f6-2655-4648-8658-22749e28b5e2","order_by":3,"name":"Xiaojie Yuan","email":"","orcid":"","institution":"Air Force Medical University","correspondingAuthor":false,"prefix":"","firstName":"Xiaojie","middleName":"","lastName":"Yuan","suffix":""},{"id":269789558,"identity":"29c8185f-4e64-4a8d-a10f-5a61bc6b755f","order_by":4,"name":"Samuel Seery","email":"","orcid":"","institution":"Lancaster University","correspondingAuthor":false,"prefix":"","firstName":"Samuel","middleName":"","lastName":"Seery","suffix":""},{"id":269789559,"identity":"f2136d33-9bf7-4e68-8bec-38a700835124","order_by":5,"name":"Ting Fu","email":"","orcid":"","institution":"Air Force Medical University","correspondingAuthor":false,"prefix":"","firstName":"Ting","middleName":"","lastName":"Fu","suffix":""},{"id":269789560,"identity":"bcd081f0-4f65-4bd6-b735-eac629dabab1","order_by":6,"name":"Xihao Liu","email":"","orcid":"","institution":"Air Force Medical University","correspondingAuthor":false,"prefix":"","firstName":"Xihao","middleName":"","lastName":"Liu","suffix":""},{"id":269789561,"identity":"8d8db492-abf4-4da2-81e9-7b69720bc2d3","order_by":7,"name":"Yihong Liu","email":"","orcid":"","institution":"Fourth Military Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yihong","middleName":"","lastName":"Liu","suffix":""},{"id":269789562,"identity":"86657101-b855-482d-b8a5-8e2bcdac7f37","order_by":8,"name":"Zhongjun Shao","email":"","orcid":"","institution":"Air Force Medical University","correspondingAuthor":false,"prefix":"","firstName":"Zhongjun","middleName":"","lastName":"Shao","suffix":""},{"id":269789563,"identity":"2e1e0945-6d4b-4752-82cf-cbb767c2ad8d","order_by":9,"name":"Rui Li","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABEElEQVRIiWNgGAWjYBACPmYgkfjHRo6BgbHxwAeYMA8eLWxgLQ1pxkAtDQdnEKUFRDA2HE5sANKH4SrxamHnfSbxcEda+tr2ww0HeBjqEufPSGB88LaNQd4cp8PYzSQSz9jkbjuT2HBAguFw4oYbCcyGc9sYDHc24NLCxiaRwJaWu+0AUIsBw4HEDUCuNG8bQ4LBAbxaDqebnX/YcCAB4jD23wS1JLYdTjC7AbTlAANzYsONBDZmAlqYLRLOpBluu/Gw4WCDwWHjDWceNkvOOSdhuAGHFn7+Y4w3f1TYyJudT3/4+E9Fnez89uSDH96U2cjjsgUNGDA4NgCjCciSIEo9GNgTr3QUjIJRMApGCgAAxGxc0DybN7YAAAAASUVORK5CYII=","orcid":"","institution":"Air Force Medical University","correspondingAuthor":true,"prefix":"","firstName":"Rui","middleName":"","lastName":"Li","suffix":""},{"id":269789564,"identity":"5870f886-8b5f-43f9-81a5-7783dfb511ac","order_by":10,"name":"Kun Liu","email":"","orcid":"","institution":"Air Force Medical University","correspondingAuthor":false,"prefix":"","firstName":"Kun","middleName":"","lastName":"Liu","suffix":""}],"badges":[],"createdAt":"2024-01-26 09:14:11","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3899519/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3899519/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":50458239,"identity":"6df4fa90-ea42-403b-b78d-44e8ebb8878c","added_by":"auto","created_at":"2024-01-31 20:08:50","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":20937112,"visible":true,"origin":"","legend":"\u003cp\u003eStudy area and spatial-temporal distribution of EVD in Africa since 2014. (a) Spatial distribution of EVD cases, and (b) Temporal variations of EVD cases and fatality rate in different countries.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-3899519/v1/d84bee394dfc5c2b58550a6f.png"},{"id":50458234,"identity":"8539b46b-7288-4748-bd8e-078524e68599","added_by":"auto","created_at":"2024-01-31 20:08:48","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":2927285,"visible":true,"origin":"","legend":"\u003cp\u003ePotential distribution of Ebola outbreak occurrences based on (a) Maxent, (b) Bioclim, (c) Domain, (d) GARP models plus ROC curves for (e) Maxent, (f) Boclim, (g) Domain, and (h) GARP.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-3899519/v1/c436c4cc1d55fe71525316d8.png"},{"id":50458835,"identity":"8101c288-0507-4b8c-b1e5-2d7439c988c2","added_by":"auto","created_at":"2024-01-31 20:16:48","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1150574,"visible":true,"origin":"","legend":"\u003cp\u003eJackknife test of the importance of environment variables in Maxent, including Jackknife of regularized training gain, test gain and AUC.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-3899519/v1/985d259564c0e9a218457a46.png"},{"id":50458236,"identity":"2dd1ec2b-5fe1-4ee5-836c-c587b0d40fa2","added_by":"auto","created_at":"2024-01-31 20:08:49","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":8218428,"visible":true,"origin":"","legend":"\u003cp\u003eGAM model-based impact factors and ecological adaptability response of Ebola (a-m), final fitness classes with bio3, bio4, bio6, bio7, bio11, bio12, bio14, GPP, GPW, HII, LCCS, light, and NDVI.\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-3899519/v1/c575116b8278e393212972f0.png"},{"id":50458231,"identity":"37cf2063-1489-4c29-80e8-268a5a34392f","added_by":"auto","created_at":"2024-01-31 20:08:48","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":2675011,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementalfiles.docx","url":"https://assets-eu.researchsquare.com/files/rs-3899519/v1/ede4401b09740519ff966ed9.docx"},{"id":50458237,"identity":"8563bdc6-f8da-4e5b-b250-190f98bd3cb9","added_by":"auto","created_at":"2024-01-31 20:08:49","extension":"png","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":16042041,"visible":true,"origin":"","legend":"\u003cp\u003eFigure S1. Predicted ecologically suitable regions of Ebola virus based on Maxent, Bioclim, Domain, and GARP models with 13 independent, influential variables i.e., bio3, bio4, bio6, bio7, bio11, bio12, bio14, NDVI, GPP, GPW, HII, LCCS, light.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-3899519/v1/24bad3ed3c60bd940d33ecc6.png"},{"id":50458240,"identity":"f9bbf4be-d647-433c-8b50-48b01d7972f4","added_by":"auto","created_at":"2024-01-31 20:08:50","extension":"png","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":26094577,"visible":true,"origin":"","legend":"\u003cp\u003eFigure S2. Map of potentially suitable habitats for Ebola virus in Africa and a closer look at four high-risk regions with (a) Integrated risk levels of Ebola based on results derived from four models, (b) Western Africa region, (c) Coastal regions of the Gulf of Guinea, (d) Central Africa region, and (e) Eastern African region.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-3899519/v1/55a4e9927c1f0a76d89571ca.png"},{"id":50458232,"identity":"021f1695-68ef-4a8b-8ef5-8518660bff48","added_by":"auto","created_at":"2024-01-31 20:08:48","extension":"jpeg","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":686386,"visible":true,"origin":"","legend":"\u003cp\u003eFigure S3. Populations in Ebola risk areas within each country with (a) Spatial distribution of countries within the Ebola risk zone with reported Ebola outbreak, (b) The spatial distribution of countries within the Ebola risk zone without any reported Ebola outbreaks, and (c) Histogram of populations at high risk of Ebola infection.\u003c/p\u003e","description":"","filename":"floatimage6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-3899519/v1/8738444c94401b0012698e20.jpeg"},{"id":50458235,"identity":"36541e68-3a81-4257-825f-98e50cfed2a3","added_by":"auto","created_at":"2024-01-31 20:08:49","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":9148663,"visible":true,"origin":"","legend":"\u003cp\u003eFigure S4. Response curves of environmental factors and the probability of suitability for Ebola based on Maxent model. Each curve (a-m) represents model-predicted probabilities associated with bio3, bio4, bio6, bio7, bio11, bio12, bio14, GPP, GPW, HII, LCCS, light, and NDVI. Curves display the mean response derived from 10 replicate Maxent runs (shown in red) with means and standard deviations (represented in blue, with two shades for categorical variables).\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-3899519/v1/c245fd42931c176eecb04c37.png"}],"financialInterests":"No competing interests reported.","formattedTitle":"Assessing the ecological resilience of Ebola virus in Africa and potential influencing factors based on a synthesized model","fulltext":[{"header":"1. Background","content":"\u003cp\u003eEbola virus disease (EVD), also known as Ebola hemorrhagic fever, is an acute and severe infectious disease caused by the Ebola virus, a filamentous virus in the family Filoviridae [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. As one of the most lethal viral diseases in human history, EVD typically causes organ damage, high fever, internal bleeding, and diarrhea with a high transmission rate and a case fatality ratio of between 50 and 90% [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. The disease was first discovered in Sudan and the Democratic Republic of Congo in Africa in 1976 [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e], and has since emerged in Central and West African nations [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e], affecting Uganda, Democratic Republic of Congo (DRC), and Nigeria [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. However over the past decade, three major pandemics have occurred: the 2014 West African Ebola outbreak, the 2018 outbreak in DRC, and the 2022 outbreak in Uganda, with the first one recorded as the largest outbreak to date [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eAs a result of these outbreaks, many scholars have investigated the influence of different factors on the dynamics of EVD outbreaks by constructing and optimizing traditional transmission models (e.g., Poisson transmission models, Susceptible-Infectious-Recovered models, and Susceptible-Exposed-Infectious-Recovered models), simulating and predicting Ebola outbreak trends [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. These studies mainly consider the spread of EVD based on frequent interactions between the uninfected agents and infected animals or humans that have been exposed to the virus. They have also used general models to simulate the processes of transmission for predicting potential risk of EVD epidemics by optimizing epidemiological parameters. In recent years, epidemiological ecology has received increased attention to seek opportunities to examine relationships between certain factors and disease transmission [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e], such as using time-series regression analysis to estimate both short-term and long-term associations between environmental exposures and health outcomes [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. In addition, from a phenological perspective, the distribution and spread of filoviruses can be effectively predicted by exploring the environmental conditions required for the survival of pathogens [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. This could help reduce uncertainty associated with traditional transmission models and can be applied to specific infectious diseases.\u003c/p\u003e \u003cp\u003eHowever, most of these traditional mathematical models neglect to incorporate the spatial variation of Ebola pathogens for revealing its potential outbreak source and sink areas. In addition, these studies mainly regard EVD as an infectious disease to predict its scale change, but rarely consider specific transmission processes of the virus, resulting in sudden and rapid predictions with relatively low accuracy. In addition, these models have paid little attention to examining the impacts of both natural and anthropogenic factors on the host movement from a spatiotemporal perspective. Therefore, we conducted this study to quantitatively explore the spatiotemporal distribution of Ebola outbreaks in Africa as well as the impacts of environmental factors from an epidemiological ecology perspective. There were three specific objectives: 1) to predict the spatiotemporal distribution of Ebola by integrating various ecological niche models; 2) to investigating the pivotal factors that can affect the suitable niches of Ebola; and 3) to reveal the ecological adaptability of Ebola virus by characterizing its response to the influencing factors.\u003c/p\u003e"},{"header":"2. Materials and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Study area and status\u003c/h2\u003e \u003cp\u003eSince the first recorded outbreaks in Sudan and the DRC in 1976, the transmission of the EVD has predominantly been observed within the African continent. This study focused on Western and Central African nations where the most recent outbreaks occurred. According to the World Health Organization (WHO), two major outbreaks have been recorded since 2014. The first occurred in West Africa with over 28,600 infections and 11,325 fatalities in 2014 [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. The second was in 2018 in the DRC, and had a total of 3,470 cases with 2,287 fatalities. Since then, there have been sporadic, relatively small outbreaks across Africa, in the DRC from 2018 to 2022, Guinea in February 2021, and Uganda in September 2022 as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e1\u003c/span\u003e. However, there are established epidemiological and genetic factors which link these outbreaks with the 2014 outbreak. Therefore, this study investigated the ecological adaptability of the EVD based on outbreaks since the West Africa nations outbreak in 2014.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Data Preparation\u003c/h2\u003e \u003cdiv id=\"Sec5\" class=\"Section3\"\u003e \u003ch2\u003e2.2.1 Surveyed EVD Cases\u003c/h2\u003e \u003cp\u003eLong-term EVD surveillance data was used in this study and included cases from the 2014\u0026ndash;2016 outbreaks, the 2018 DRC outbreak, and the 2022 outbreak in Uganda. The first two groups of data were obtained from The Humanitarian Data Exchange (HDX) website (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://data.humdata.org/\u003c/span\u003e\u003cspan address=\"https://data.humdata.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) which serves as an open platform for facilitating easily accessible and useable data for this analysis. This data source includes data from over 250 countries and territories.\u003c/p\u003e \u003cp\u003eEpidemic data were primarily based on information provided by the WHO, along with real-time reporting by national health authorities of different countries. Data from the 2022 Uganda outbreak were acquired from the Global Health website (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://global.health/\u003c/span\u003e\u003cspan address=\"https://global.health/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) which is purported to provide accurate real-time disease data in the early stages of an outbreak, tracking cases for the first 100 days.\u003c/p\u003e \u003cp\u003eGeographical points of each case were collected from the aforementioned sources and then abstracted by locating them in the center of a specific region or in proximity to the medical facilities. This pre-process helped to avoid potential conflicts or errors without affecting the reliability and accuracy of the final results.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section3\"\u003e \u003ch2\u003e2.2.2 Environmental Data\u003c/h2\u003e \u003cp\u003ePrevious research has demonstrated that the transmission of Ebola is influenced by multiple environmental factors such as climate conditions (e.g. precipitation and temperature) and vegetation abundance (e.g. deforestation and the consumption of forest prey) [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. In this study, climate data were primarily obtained from WorldClim (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.worldclim.org/\u003c/span\u003e\u003cspan address=\"https://www.worldclim.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), a database that offers high-resolution weather and climate data at a global scale.\u003c/p\u003e \u003cp\u003eSurface coverage data were obtained from the Copernicus Climate Change Service (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://climate.copernicus.eu/\u003c/span\u003e\u003cspan address=\"https://climate.copernicus.eu/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), which combines climate system observations with scientific research to produce authoritative and quality-assured information on European and global climate. In addition, vegetation information was acquired from the global map data archive established by the Global Map Transfer Program and long-term time series normalized difference vegetation index (NDVI) data were retrieved from NOAA's Advanced Very High Resolution Radiometer (AVHRR) sensor.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section3\"\u003e \u003ch2\u003e2.2.3 Human Impact Data\u003c/h2\u003e \u003cp\u003eDemographic data were extracted from Gridded Population of the World, version 4 data (GPW.v4) which is based on national censuses and population registers. We also utilized Human Influence Index (HII) grids data set which contains information such as population density raster, human land use raster, and constructed roads. In addition, night time light remote sensing data were also utilized due to its strong correlation with socioeconomic factors, which was obtained from the Operational LineScan System (OLS) of the U.S. Defense Meteorological Satellite Program (DMSP) and the VIIRS Plus DMSP light change dataset.\u003c/p\u003e \u003cp\u003eAll the aforementioned data were carefully examined and cleaned to meet the subsequent analysis. Finally, after integrating the epidemic case data, remote sensing data, surveyed attribute data, and the spatial vector boundaries, we used ArcGIS (version 10.8) software to establish a geospatial database for further analysis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section3\"\u003e \u003ch2\u003e2.2.4 Variable Filter\u003c/h2\u003e \u003cp\u003eTo prevent overfitting and enhance the interpretability of our models, we conducted variable filtering to select significant data. Initially, we implemented a JackKnife method based on the Maxent model and utilized the XGboost algorithm to assess the significance of each factor. In order to address the issue of strong interrelationships among predictor variables, which hampers the interpretation of variables [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. we utilized Spearman\u0026rsquo;s correlation coefficient to remove variables exhibiting substantial correlation, thereby enhancing the precision of the predictor variables. This process retains representative and more practically significant factors, ultimately selecting 13 variables that are most likely to influence Ebola distribution, namely Isothermality (bio3), Temperature Seasonality (bio4), Min Temperature of Coldest Month (bio6), Temperature Annual Range (bio7), Mean Temperature of Coldest Quarter (bio11), Annual Precipitation (bio12), Precipitation of Driest Month (bio14), Gross Primary Productivity (GPP), Gridded Population of the World (GPW), Human Influence Index (HII), Land Cover Classification System (LCCS), Night Lights (light) and NDVI.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Methods\u003c/h2\u003e \u003cp\u003eSpatial-temporal distribution of EVD in Africa from 2014 to 2023 were visualized by ArcGIS (version 10.8) software. In order to accurately estimate the current spatiotemporal distribution of EVD as well as its potential risk, we established a synthesized predictive model by integrating four distinct ecological niche approaches which are Maxent, Bioclim, Domain and GARP models. The Maxent model provides probabilities ranging from 0 to 1, indicating the likelihood of Ebola presence. The Bioclim and Domain models yield larger values for regions more prone to Ebola outbreaks, and the GARP model distinguishes the study area into suitable and unsuitable regions. The description of those four models can be found in the supplementary materials (Appendix D). Our proposed synthesized model is given by\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\begin{array}{c}{\\widehat{{y}_{j}}}^{i}={f}_{i}\\left({{X}^{{\\prime }}}_{j},{{\\Theta }}_{i}\\right)\\#（1）\\end{array}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ewhere \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(i\\)\u003c/span\u003e\u003c/span\u003e represents the type of ecological niche model, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(j\\)\u003c/span\u003e\u003c/span\u003e represents the prediction unit, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(f\\)\u003c/span\u003e\u003c/span\u003e denotes the model mapping, and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(X\\)\u003c/span\u003e\u003c/span\u003e is the feature vector used to describe multidimensional influencing factors. After training the model, input features \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({{X}^{{\\prime }}}_{j}\\)\u003c/span\u003e\u003c/span\u003e are applied to each model for prediction, resulting in predicted values \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({\\widehat{{y}_{j}}}^{i}.\\)\u003c/span\u003e\u003c/span\u003e By adjusting the model parameters \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({{\\Theta }}_{i}\\)\u003c/span\u003e\u003c/span\u003e, we aimed to minimize the difference between the predicted values \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({f}_{i}\\left({X}_{j},{{\\Theta }}_{i}\\right)\\)\u003c/span\u003e\u003c/span\u003eand the actual values\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({y}_{j}\\)\u003c/span\u003e\u003c/span\u003e, which is expressed as:\u003cdiv id=\"Equb\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equb\" name=\"EquationSource\"\u003e\n$$\\begin{array}{c}\\underset{{{\\Theta }}_{i}}{min} {\\left({y}_{j}-{f}_{i}\\left({X}_{j},{{\\Theta }}_{i}\\right)\\right)}^{2}\\#（2）\\end{array}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eFinally, the predicted values of the four models are weighted and summed to obtain the ultimate ecological niche prediction \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({\\widehat{\\varvec{y}}}_{j}^{inter}\\)\u003c/span\u003e\u003c/span\u003e. The weights \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({W}_{i}\\)\u003c/span\u003e\u003c/span\u003e are determined based on prediction accuracy and are calculated as follows:\u003cdiv id=\"Equc\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equc\" name=\"EquationSource\"\u003e\n$$\\begin{array}{c}\\begin{array}{c}{\\widehat{\\varvec{y}}}_{j}^{inter}=\\sum _{i=1}^{4} {W}_{i}\\cdot {\\widehat{\\varvec{y}}}_{j}^{i}\\end{array}\\#\\left(3\\right)\\end{array}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eBased on the predictive accuracy of each model, the final outcome was retrieved by incorporating different weights according to the following standards: (1) the plotted Receiver Operating Characteristic curves (ROC); (2) the calculated Area Under the Curve (AUC) values, a widely used metric for evaluating the classification performance of heavily imbalanced data; (3) normalizing the different suitability representations of into five levels (1 for Not Suitable, 2 for Less Suitable, 3 for Moderate, 4 for More Suitable, and 5 for Suitable); (4) assigning weights to each model based on the corresponding AUC values, which were then standardized using a grid-based calculation method to obtain the final result. A schematic representation of the entire process was depicted in Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003eThe Maxent model and GAMs (generalized additive models) (see Appendix D) were performed to understand the influence of each environmental factor on the predicted risk. Throughout the modeling process, all other environmental variables were set the mean value but with one specific variable estimated. The resulting Maxent-based Environmental Marginal Response Curves demonstrated how the predicted probability varies with changes in the variable, thereby revealing the correspondence between ecological viral adaptability with certain variables.\u003c/p\u003e \u003c/div\u003e"},{"header":"3. Results","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Ecological adaptability assessment of Ebola\u003c/h2\u003e \u003cp\u003eEach of the four models had distinctive characteristics for predicting Ebola occurrences across different regions. Since the Maxnet, Bioclim, and Domain models all derived specific values while the GARP model provides logistic variables, it is necessary to standardize these different results for effective comparison. Therefore, we reclassified values obtained from the first three model results shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e2\u003c/span\u003ea, b and c with darker colors representing higher adaptability of Ebola.\u003c/p\u003e \u003cp\u003eThe Maxent and Bioclim models demonstrated similar predictions for favorable habitats, primarily concentrated in West Africa, sub-Saharan nations, along the edge of the Congolese basin, around Lake Victoria, and on the Ethiopian plateau. These ecologically suitable zones were generally situated near the equator. In contrast, the Domain and GARP models utilized more stringent criteria to identify non-viable areas, and predicted a broader range of potential zones distributed across the African continent, spanning from sub-Saharan regions to the Kalahari Desert in the South African plateau.\u003c/p\u003e \u003cp\u003eAccording to the accuracy assessment of ROC curves illustrated by Fig.\u0026nbsp;\u003cspan refid=\"Fig14\" class=\"InternalRef\"\u003e3\u003c/span\u003e (e, f, g, and h), the AUC values derived from these four predict models were all relatively high with Maxent (0.99), Bioclim (0.92), Domain (0.93) and GARP (0.85), respectively. In order to synthesize the predictions of the four ecological niche models, we assigned 30%, 30%, 30%, and 10% weights to the four models mentione according to their respective accuracies, and obtained the integrated danger levels of Ebola outbreaks as well as the potential habitat of EVD in Africa as shown by Figure S2 (a), while higher levels indicate a greater risk of an Ebola outbreak.\u003c/p\u003e \u003cp\u003eThe areas at most risk predominantly lay in the western, central, and eastern regions of Africa, situated south of the Sahara Desert but close to the equator as illustrated by Figure S2 (b, c, d and e). Specific countries included Guinea, Liberia, Sierra Leone, Nigeria, DRC, Uganda, Ethiopia, among other. Using the Albers Equal Area Conic projection within corresponding coordinate system, the calculated total area occupied by these high-risk regions (areas above level moderately high in the map) was approximately 3, 509, 623.11 km2, accounting for nearly 11.61% of Africa\u0026rsquo;s entire land area.\u003c/p\u003e \u003cp\u003eBy integrating population density grids, we estimated the total population affected in each country located within the high-risk areas depicted in Figure S3. The red bars in Figure S3 (c) represented countries without reported Ebola outbreaks but with a substantial population at risk of infection. The blue bars signify countries with a history of Ebola outbreaks and documented cases. Based on the current predictive results, countries at higher risk yet without a history of Ebola include Ethiopia, Tanzania, C\u0026ocirc;te d'Ivoire, Ghana, Cameroon, Rwanda, Burundi, and others.\u003c/p\u003e \u003cp\u003eC\u0026ocirc;te d'Ivoire has reported one single Ebola case in 2021 but this did not lead to a further outbreak. Countries that have experienced recent Ebola outbreaks (e.g., Guinea, the DRC, and Liberia), as well as those that had outbreaks in earlier periods (e.g., Gabon, South Africa, and the Sudan), all matched with the predicted Ebola-adapted zones, providing a good validation for the reliability\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Significance analysis of impact factors\u003c/h2\u003e \u003cp\u003eGiven the superior performance of the Maxent model in accuracy assessment owing to a high AUC value of 0.99, this study conducted an impact factor analysis to examine the crucial environmental variables for predicting Ebola-suitable habitats. Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e presented the contribution and permutation importance of 13 environmental variables to the ecological adaptability prediction of EVD, which was derived from the Maxent modeling. Averages of 100 experiments revealed that bio4, GPW, bio12, LCCS, and bio6 acted in rank as pivotal factors that can affect the suitable niches of Ebola. These key variables accounted for a cumulative percentage contribution of 97.8% and permutation importance of 98.5%.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eContribution and permutation importance of variables derived through Maxent modeling (average of 100 experiments).\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eContribution (%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePermutation Importance\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ebio4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e44.89\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e48.19\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGPW\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e33.43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e32.69\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ebio12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e9.76\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e9.09\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLCCS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e7.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.67\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ebio6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2.82\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHII\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.55\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.59\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ebio3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.52\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.44\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGPP\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.43\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNDVI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.06\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003elight\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ebio7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ebio14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ebio11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eJackknife testing (Fig.\u0026nbsp;\u003cspan refid=\"Fig14\" class=\"InternalRef\"\u003e3\u003c/span\u003e) showed that when considering individual environmental variables, bio3, bio4, bio11, bio12, GPP, and GPW had the most significant impact on regularized training gains, test gains, and AUC values. This suggests that these variables might have the most useful information and are crucial for the adaptability of the Ebola virus [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. However, after excluding individual environmental variables, GPW, LCCS, and bio12 had the largest reductions in gains and AUC values, indicating that these variables hold more unique information and played an irreplaceable role compared to others.\u003c/p\u003e \u003cp\u003eAfter combining these findings, it became evident that bio4 (Temperature Seasonality), bio12 (Annual Precipitation), and GPW (Population Density) exerted a significant impact on the suitable niches of Ebola.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Response of impact factors\u003c/h2\u003e \u003cp\u003eTo generate response curves for different variables, we utilized individual variables to construct the Maxent model. Consequently, we obtained the model\u0026rsquo;s predicted probabilities as a function of increasing the values of each environmental factor (Figure S4). The x-axis represented values of the environmental variables, while the y-axis corresponded to the consequent Ebola suitability probability. The line graph demonstrated the dependency of predicted suitability on the selected variable, as well as the dependency arising from the correlation between the chosen variable and other variables. Since HII and LCCS were categorical variables, the results have been illustrated using a standard bar chart.\u003c/p\u003e \u003cp\u003eBased on the GAM model, we investigated the relationship between each influencing factor and Ebola\u0026rsquo;s ecological adaptability. To achieve this, we generated 1000 random points in the study area and extracted the corresponding environmental data and Maxent model results for these points. The Maxent model results were used as the response variable, and we examined how the environmental variables influenced the model outcomes (Fig.\u0026nbsp;\u003cspan refid=\"Fig15\" class=\"InternalRef\"\u003e4\u003c/span\u003e). Each scatter plot in the figure represents a random point, with the x-axis representing the variable\u0026rsquo;s value and the y-axis showing its impact on Ebola\u0026rsquo;s ecological adaptability. Vertical lines on the x-axis depicted data distribution and density, while the light purple background signified the confidence interval. These lines can represent the fitting of the data points.\u003c/p\u003e \u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eAs an interdisciplinary piece of research, this study attempted to investigate suitable ecological niches for Ebola virus as well as its potential responses to different environmental factors. This study provides a practical example of how to apply geostatistical approaches to quantitatively explore the dynamic distribution, variation, and influential determinants of an epidemic, not only demonstrating an opportunity to address certain research gaps in the existing literature, but also presenting some potential to support public health decision-makers in formulating and implementing effective strategies.\u003c/p\u003e \u003cp\u003eRegarding spatiotemporal predictions for Ebola habitats at a macroscopic scale, the forecasted outcomes of the integrated model mainly aggregate to the south of the Sahara Desert and north of Zambia and Angola, covering approximately 16% of the entire African continent. This finding relatively matches with the zoonotic niche map of Ebola virus revealed by Pigott et al.[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e], but in comparison, our high-risk areas are more dispersed and better correspond to the previous outbreak patterns [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. This may be attributed to the utilization of the latest infection case data, including the 2018 Congolese outbreak and the 2022 Uganda outbreak. Additionally, we incorporated a more comprehensive set of potential influencing factors, which improved identifying the most significant ecological determinants that possibly affect the adaptability of Ebola virus in Africa, thereby further increasing the accuracy of final prediction. Moreover, our results are supported by Redding et al [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. who developed an Ebola system-dynamics model that yielded a similar prediction. However, our study integrated each unique feature of the four ecological niche models, achieving higher AUC values and able to provide response curves for various influencing factors, thus offering a better understanding of the association between Ebola outbreaks and multiple environmental factors. Nevertheless, our results provide only a weak explanation to distinguish indexed cases and secondary cases.\u003c/p\u003e \u003cp\u003eOur research also compiled the demographic statistics within Ebola outbreak-prone regions across different African countries, estimating the total population at risk in these high-risk areas to be 20,504,885. Among this population, 11,464,743 individuals reside in regions with reported Ebola outbreaks, while 9,040,142 individuals live in areas with no reported outbreaks. Notably, Ethiopia, one of the countries with the highest at-risk population, although not previously documented with Ebola cases, has been identified as having the potential risk for an Ebola outbreak in our study, similar to the findings from Redding et al [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. Both researches highlighted environmental factors in Ethiopia are conducive to a potential Ebola outbreak, warranting serious concern from local governments and global public health institutions. Other countries at high risk of Ebola without previous reports include Tanzania, Cote d'Ivoire, Ghana, Cameroon, and Rwanda, which are urged to promptly implement effective prevention measures against potential Ebola epidemics.\u003c/p\u003e \u003cp\u003eFurthermore, this study scrutinized the responses of various influencing factors and explained partial triggers for Ebola outbreak from a spatiotemporal perspective. Utilizing the Jackknife test and Maxent model, we identified the key factors closely associated with Ebola's ecological adaptability, notably including bio4, GPW, bio12 and LCCS. In order to further investigate how these influencing factors affect Ebola ecological adaptability, we combined the outcomes yielded by Maxent and Generalized Additive Models to reveal the distinct patterns in Ebola distribution:\u003c/p\u003e \u003cp\u003e(1) The ecological adaptability of Ebola virus decreases with a rise in temperature seasonality (bio4), indicating preference for regions with minimal temperature variation between seasons. This matches closely with the conclusions by Olivero et al. that the observation of Ebola outbreaks predominantly occurring in low-latitude areas near the equator [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. Its adaptability also increases with higher winter average temperatures and annual precipitation (bio6, bio12), similar to the findings on environment temperature and precipitation trends [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e].\u003c/p\u003e \u003cp\u003e(2) The ecological adaptability of Ebola virus escalates with greater population density (GPW), likely due to the rapid human-to-human virus transmission promoted by frequent interpersonal contacts [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. Additionally, the adaptability increases initially and then decreases, becoming lower in more economically developed regions (HII), which can be explained by the fact that the areas less impacted by anthropogenic activities especially rapid urbanization are almost not susceptible to large-scale outbreaks of Ebola [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. whereas the economically developed regions with better healthcare resources are able to effectively implement barrier care techniques to reduce the risk of infection [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e].\u003c/p\u003e \u003cp\u003e(3) The ecological adaptability of Ebola virus follows a pattern of initial growth and then subsequent decline with vegetation increase, peaking within regions of moderate vegetation coverage. For equatorial area with more vegetation abundance, the reduction in vegetation possibly accelerates the risk of Ebola virus transmission. Other studies have also indicated that deforestation substantially elevates human interaction with infected wildlife, thereby favoring viral spread [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. This perspective effectively corroborates our findings.\u003c/p\u003e \u003cp\u003e(4) The ecological adaptability of Ebola virus responds positively to land cover and values of 10, 30, 60, 170, and 190 respectively represents drylands, croplands, broadleaf deciduous forests, forests, floodplain, and urban zones (LCCS), which is consistent with the findings of Redding et al. who claimed that the land-use/land-cover determines distributions across the different reservoir host species [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. Nevertheless, we took a step further by incorporating a quantitative analysis, particularly focusing on the nuances within land cover types.\u003c/p\u003e \u003cp\u003eNonetheless, several limitations should be taken into consideration when interpreting our findings. Firstly, the available Ebola case data were restricted to the provincial level without accurate geographical location information (e.g., longitude and latitude) which cannot satisfy a finer spatial resolution of our analysis. Secondly, the evaluation did not take the hosts of Ebola virus into consideration due to the difficulty in obtaining such information, which may lead to some incomplete understanding of how those impact factors drive the risk of Ebola outbreaks. Thirdly, our study was focused on Africa where the major endemic areas distributed in, which is in necessity to be further extended to other regions such as the North and South America, Asia, Europe for a deepening exploration of the global outbreak and transmission of Ebola based on the One Health strategy. Therefore, further research is needed to promote a comprehensive forecast of Ebola across regions based on involving Ebola host data, higher-precision case data, and largescale geographical areas.\u003c/p\u003e"},{"header":"5. Conclusions","content":"\u003cp\u003eThis study integrated different ecological niche models to effectively predict the spatial distribution of suitable habitats for the Ebola virus. Our findings indicate that these suitable habitats are primarily concentrated in regions near the equator, south of the Sahara Desert. Additionally, our research delved into the response patterns of key influencing factors related to Ebola, elucidating specific relationships between Ebola outbreaks and variables such as climate, socio-economic conditions, vegetation, and land use types. These insights offer valuable directions for future research in the field of epidemic prevention and control.\u003c/p\u003e"},{"header":"List Of Abbreviations","content":"\u003cp\u003eAUC the Area Under the Curve\u003c/p\u003e\u003cp\u003eDRC the Democratic Republic of Congo\u003c/p\u003e\u003cp\u003eEVD Ebola virus disease\u003c/p\u003e\u003cp\u003eGAM generalized additive model\u003c/p\u003e\u003cp\u003eGPP Gross Primary Productivity\u003c/p\u003e\u003cp\u003eGPW Gridded Population of the World data\u003c/p\u003e\u003cp\u003eHDX the Humanitarian Data Exchange\u003c/p\u003e\u003cp\u003eHII Human Influence Index\u003c/p\u003e\u003cp\u003eNDVI Normalized difference vegetation index\u003c/p\u003e\u003cp\u003eROC Receiver Operating Characteristic curves\u003c/p\u003e\u003cp\u003eWHO the World Health Organization\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was partly supported by grants from the National Natural Science Foundation of China [grant number 42201448 and 82273689], and Natural Science Foundation of Hubei Province [grant number 2022CFB610].\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026apos; contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors developed the study concept. LS, JS, and YZ designed and performed the analysis and wrote the original draft. XY, SS and TF performed the data collection and preparation. Reviewing and conceptualization was conducted by XL and YL. Supervision and review \u0026amp; editing was conducted by ZS, RL and KL. All co-authors provided input and critical review of the manuscript leading to the final version. All authors read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe project team would like to thank all the participants who consented to participate in the study. We thank our colleagues for their careful reading and editing of this manuscript. The funder had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eFeldmann H, Jones S, Klenk HD, Schnittler HJ: Ebola virus: from discovery to vaccine. \u003cem\u003eNature Reviews Immunology \u003c/em\u003e2003, 3(8):677-685.\u003c/li\u003e\n\u003cli\u003eMalvy D, McElroy AK, de Clerck H, Guenther S, van Griensven J: Ebola virus disease. \u003cem\u003eLancet \u003c/em\u003e2019, 393(10174):936-948.\u003c/li\u003e\n\u003cli\u003eAnsari AA: Clinical features and pathobiology of Ebolavirus infection. \u003cem\u003eJournal of Autoimmunity \u003c/em\u003e2014, 55:1-9.\u003c/li\u003e\n\u003cli\u003eColebunders R, Borchert N: Ebola haemorrhagic fever - a review. \u003cem\u003eJournal of Infection \u003c/em\u003e2000, 40(1):16-20.\u003c/li\u003e\n\u003cli\u003ePattyn S, van der Groen G, Jacob W, Piot P, Courteille G: Isolation of Marburg-like virus from a case of haemorrhagic fever in Zaire. \u003cem\u003eLancet (London, England) \u003c/em\u003e1977, 1(8011):573-574.\u003c/li\u003e\n\u003cli\u003eJacob ST, Crozier I, Fischer WA, II, Hewlett A, Kraft CS, de La Vega M-A\u003cem\u003e et al\u003c/em\u003e: Ebola virus disease. \u003cem\u003eNature Reviews Disease Primers \u003c/em\u003e2020, 6(1).\u003c/li\u003e\n\u003cli\u003eKalenga OI, Moeti M, Sparrow A, Vinh-Kim N, Lucey D, Ghebreyesus TA: The Ongoing Ebola Epidemic in the Democratic Republic of Congo, 2018-2019. \u003cem\u003eNew England Journal of Medicine \u003c/em\u003e2019, 381(4):373-383.\u003c/li\u003e\n\u003cli\u003eSamarasekera U: Solidarity Against Ebola: an update. \u003cem\u003eLancet Microbe \u003c/em\u003e2023, 4(3):E139-E139.\u003c/li\u003e\n\u003cli\u003eArea I, Batarfi H, Losada J, Nieto JJ, Shammakh W, Torres A: On a fractional order Ebola epidemic model. \u003cem\u003eAdvances in Difference Equations \u003c/em\u003e2015.\u003c/li\u003e\n\u003cli\u003eRistic B, Dawson P, Ieee: Real-time forecasting of an epidemic outbreak: Ebola 2014/2015 case study. In: \u003cem\u003e19th International Conference on Information Fusion (FUSION): 2016 \u003c/em\u003e\u003cem\u003eJul 05-08 2016; Heidelberg, GERMANY\u003c/em\u003e; 2016: 1983-1990.\u003c/li\u003e\n\u003cli\u003eShahraiyni HT, Sodoudi S, Cubasch U: Weather conditions and their effect on the increase of the risk of type A acute aortic dissection onset in Berlin. \u003cem\u003eInternational Journal of Biometeorology \u003c/em\u003e2016, 60(8):1303-1305.\u003c/li\u003e\n\u003cli\u003eManlove K, Wilber M, White L, Bastille-Rousseau G, Yang A, Gilbertson MLJ\u003cem\u003e et al\u003c/em\u003e: Defining an epidemiological landscape that connects movement ecology to pathogen transmission and pace-of-life. \u003cem\u003eEcology Letters \u003c/em\u003e2022, 25(8):1760-1782.\u003c/li\u003e\n\u003cli\u003eBhaskaran K, Gasparrini A, Hajat S, Smeeth L, Armstrong B: Time series regression studies in environmental epidemiology. \u003cem\u003eInternational Journal of Epidemiology \u003c/em\u003e2013, 42(4):1187-1195.\u003c/li\u003e\n\u003cli\u003eKim H, Lee J-T, Fong KC, Bell ML: Alternative adjustment for seasonality and long-term time-trend in time-series analysis for long-term environmental exposures and disease counts. \u003cem\u003eBmc Medical Research Methodology \u003c/em\u003e2021, 21(1).\u003c/li\u003e\n\u003cli\u003ePeterson AT, Bauer JT, Mills JN: Ecologic and geographic distribution of filovirus disease. \u003cem\u003eEmerging Infectious Diseases \u003c/em\u003e2004, 10(1):40-47.\u003c/li\u003e\n\u003cli\u003eValero KCW, Isokpehi RD, Douglas NE, Sivasundaram S, Johnson B, Wootson K\u003cem\u003e et al\u003c/em\u003e: Plant Phenology Supports the Multi-emergence Hypothesis for Ebola Spillover Events. \u003cem\u003eEcohealth \u003c/em\u003e2018, 15(3):497-508.\u003c/li\u003e\n\u003cli\u003eColtart CEM, Lindsey B, Ghinai I, Johnson AM, Heymann DL: The Ebola outbreak, 2013-2016: old lessons for new epidemics. \u003cem\u003ePhilosophical Transactions of the Royal Society B-Biological Sciences \u003c/em\u003e2017, 372(1721).\u003c/li\u003e\n\u003cli\u003eMerler S, Ajelli M, Fumanelli L, Gomes MFC, Pastore y Piontti A, Rossi L\u003cem\u003e et al\u003c/em\u003e: Spatiotemporal spread of the 2014 outbreak of Ebola virus disease in Liberia and the effectiveness of non-pharmaceutical interventions: a computational modelling analysis. \u003cem\u003eLancet Infectious Diseases \u003c/em\u003e2015, 15(2):204-211.\u003c/li\u003e\n\u003cli\u003eKumar S, Stohlgren TJ, Chong GW: Spatial heterogeneity influences native and nonnative plant species richness. \u003cem\u003eEcology \u003c/em\u003e2006, 87(12):3186-3199.\u003c/li\u003e\n\u003cli\u003eGuo Y, Zhang S, Tang S, Pan J, Ren L, Tian X\u003cem\u003e et al\u003c/em\u003e: Analysis of the prediction of the suitable distribution of Polygonatum kingianum under different climatic conditions based on the MaxEnt model. \u003cem\u003eFrontiers in Earth Science \u003c/em\u003e2023, 11.\u003c/li\u003e\n\u003cli\u003ePigott DM, Golding N, Mylne A, Huang Z, Henry AJ, Weiss DJ\u003cem\u003e et al\u003c/em\u003e: Mapping the zoonotic niche of Ebola virus disease in Africa. \u003cem\u003eElife \u003c/em\u003e2014, 3.\u003c/li\u003e\n\u003cli\u003eFeldmann H, Sprecher A, Geisbert TW: Ebola. \u003cem\u003eNew England Journal of Medicine \u003c/em\u003e2020, 382(19):1832-1842.\u003c/li\u003e\n\u003cli\u003eRedding DW, Atkinson PM, Cunningham AA, Lo Iacono G, Moses LM, Wood JLN\u003cem\u003e et al\u003c/em\u003e: Impacts of environmental and socio-economic factors on emergence and epidemic potential of Ebola in Africa. \u003cem\u003eNature Communications \u003c/em\u003e2019, 10.\u003c/li\u003e\n\u003cli\u003eOlivero J, Fa JE, Real R, Farf\u0026aacute;n MA, M\u0026aacute;rquez AL, Vargas JM\u003cem\u003e et al\u003c/em\u003e: Mammalian biogeography and the Ebola virus in Africa. \u003cem\u003eMammal Review \u003c/em\u003e2017, 47(1):24-37.\u003c/li\u003e\n\u003cli\u003eSchmidt JP, Park AW, Kramer AM, Han BA, Alexander LW, Drake JM: Spatiotemporal Fluctuations and Triggers of Ebola Virus Spillover. \u003cem\u003eEmerging Infectious Diseases \u003c/em\u003e2017, 23(3):415-422.\u003c/li\u003e\n\u003cli\u003eMatua GA, Van der Wal DM, Locsin RC: Ebola hemorrhagic fever outbreaks: strategies for effective epidemic management, containment and control. \u003cem\u003eBrazilian Journal of Infectious Diseases \u003c/em\u003e2015, 19(3):308-313.\u003c/li\u003e\n\u003cli\u003eLongley JL, Danquah LO, Massa MS, Ross DA, Weiss HA: The impact of case and contact characteristics on contact tracing during the West Africa Ebola epidemic. \u003cem\u003eJournal of Infection \u003c/em\u003e2021, 83(4):501-504.\u003c/li\u003e\n\u003cli\u003eAli H, Dumbuya B, Hynie M, Idahosa P, Keil R, Perkins PJCC\u003cem\u003e et al\u003c/em\u003e: The social and political dimensions of the Ebola response: Global inequality, climate change, and infectious disease. \u003cem\u003eClimate Change and Health: Improving Resilience and Reducing Risks \u003c/em\u003e2016:151-169.\u003c/li\u003e\n\u003cli\u003eGrillet G, Marjanovic N, Diverrez JM, Tattevin P, Tadi\u0026eacute; JM, L\u0026apos;Her E: Intensive care medical procedures are more complicated, more stressful, and less comfortable with Ebola personal protective equipment: A simulation study. \u003cem\u003eJournal of Infection \u003c/em\u003e2015, 71(6):703-706.\u003c/li\u003e\n\u003cli\u003eOlivero J, Fa JE, Real R, Marquez AL, Farfan MA, Vargas JM\u003cem\u003e et al\u003c/em\u003e: Recent loss of closed forests is associated with Ebola virus disease outbreaks. \u003cem\u003eScientific Reports \u003c/em\u003e2017, 7.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Ebola, ecological adaptability, ecological niche models, spatiotemporal distribution, influencing factors","lastPublishedDoi":"10.21203/rs.3.rs-3899519/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3899519/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eThe Ebola epidemic has persisted in Africa since it was firstly identified in 1976. However, few studies have focused on spatiotemporally assessing the ecological adaptability of this virus and the influence of multiple factors on outbreaks. This study quantitatively explores the ecological adaptability of Ebola virus and its response to different potential natural and anthropogenic factors from a spatiotemporal perspective.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eBased on historical Ebola cases and relevant environmental factors collected from 2014 to 2022 in Africa, the spatiotemporal distribution of Ebola adaptability is characterized by integrating four distinct ecological models into one synthesized spatiotemporal framework. Maxent and Generalized Additive Models were applied to further reveal the potential responses of the Ebola virus niche to its ever changing environments.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eEbola habitats appear to aggregate across the sub-Saharan region and in north Zambia and Angola, covering approximately 16% of the African continent. Countries presently unaffected by Ebola but at increased risk include Ethiopia, Tanzania, C\u0026ocirc;te d'Ivoire, Ghana, Cameroon, and Rwanda. In addition, among the thirteen key influencing factors, temperature seasonality and population density were identified as significantly influencing the ecological adaptability of Ebola. Specifically, those regions were prone to minimal temperature variations between seasons. Both the potential anthropogenic influence and vegetation coverage have a rise-to-decline impact on the outbreaks of Ebola virus across Africa.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eOur findings suggest new ways to effectively respond to smaller potential Ebola outbreaks in Sub-Saharan Africa. We believe that this integrated modeling approach and response analysis provide a framework that can be extended to predict risk of similar epidemiological studies for other diseases across the world.\u003c/p\u003e","manuscriptTitle":"Assessing the ecological resilience of Ebola virus in Africa and potential influencing factors based on a synthesized model","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-01-31 20:08:43","doi":"10.21203/rs.3.rs-3899519/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"89867fc7-acbc-4de5-921d-c6eece148eb6","owner":[],"postedDate":"January 31st, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-02-09T11:14:52+00:00","versionOfRecord":[],"versionCreatedAt":"2024-01-31 20:08:43","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-3899519","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3899519","identity":"rs-3899519","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}