Skip directly to site content Skip directly to page options Skip directly to A-Z link Skip directly to A-Z link Skip directly to A-Z link
Volume 10, Number 1—January 2004
Research

Ecologic and Geographic Distribution of Filovirus Disease

Author affiliations: *University of Kansas, Lawrence, Kansas, USA; †Centers for Disease Control and Prevention, Atlanta, Georgia, USA

Cite This Article

Abstract

We used ecologic niche modeling of outbreaks and sporadic cases of filovirus-associated hemorrhagic fever (HF) to provide a large-scale perspective on the geographic and ecologic distributions of Ebola and Marburg viruses. We predicted that filovirus would occur across the Afrotropics: Ebola HF in the humid rain forests of central and western Africa, and Marburg HF in the drier and more open areas of central and eastern Africa. Most of the predicted geographic extent of Ebola HF has been observed; Marburg HF has the potential to occur farther south and east. Ecologic conditions appropriate for Ebola HF are also present in Southeast Asia and the Philippines, where Ebola Reston is hypothesized to be distributed. This first large-scale ecologic analysis provides a framework for a more informed search for taxa that could constitute the natural reservoir for this virus family.

The natural maintenance cycles of filoviruses (Order Mononegavirales, family Filoviridae) are unknown (1).1Although dynamics of filoviruses as causes of epidemic diseases among humans, great apes, and other primates have been described in detail (213), the natural reservoir, mode of transmission to hominids and pongids (humans, gorillas, and chimpanzees), and temporal dynamics remain unclear. Diverse taxa have been suggested as potential reservoirs, including bats, rodents, arthropods, and plants (1418).

Two observations provide clues about the nature of the host-virus relationship. First, filovirus transmission to humans is not common, and most occurrences can be traced to a single index case (2,6,19) (exceptions occur—e.g., the Durba Marburg outbreak appears to have involved multiple independent infections of humans from a reservoir population presumably associated with a mine). We assume that introductions to nonhuman hominid populations also generally begin with single index cases, but this hypothesis is more difficult to investigate. This rarity argues against a common arthropod vector for transmission: if anthrophilic arthropod vectors were to carry filoviruses, multiple index cases would be more common, as many hominids in an area would have the opportunity for infection. In addition, filoviruses generally do not replicate in arthropods or arthropod cell lines, leading several authors to speculate on more incidental modes of transfer (e.g., direct contact) (20).

Second, filoviruses show clear geographically related phylogeographic structure. Viruses and subtypes from particular geographic areas cluster together phylogenetically, even when occurrences from different years are studied. This phylogeographic structure suggests a stable host-parasite relationship, in which viruses are maintained in permanent local-regional pools. This host would not experience high death rates, as primates do (7,9); evolution of avirulence in long-term host-parasite relationships is expected on theoretical grounds (21).

Searches for the natural reservoir of filoviruses have taken several paths. Epidemiologic studies designed to trace lineages of transmission in outbreaks have identified index cases but have not succeeded in specifying the mode of “jump” to hominids (2,6,19). Testing large numbers of organisms from the vicinity of outbreaks has failed to identify even a single nonhominid infection (1416). Finally, laboratory tests of reservoir competence of species have documented the following: 1) no, or very limited, infection of plants or arthropods; 2) a single marginally successful infection of snakes but with very low levels of virus circulation; 3) successful infection of bats and possibly rodents; and 4) frequent successful, but fatal, infection in nonhuman primates (e.g., Chlorocebus, Macaca) (1,17,18). While these investigations have shed some light, they have not provided convincing evidence for a particular reservoir.

An unexplored approach to identifying the natural reservoir of filoviruses is large-scale ecologic and geographic comparisons to detect patterns of co-occurrence and codistribution of viruses with potential hosts. This approach has been applied successfully to identifying reservoir rodent species for Chagas disease (22). Our general approach is as follows: 1) to understand the large-scale ecology and geography of disease occurrences by using ecologic niche modeling (23), and 2) to compare these characteristics with ecologic and geographic patterns of potential reservoirs. Here, we address the first step and document broad-scale ecologic and geographic patterns in filovirus-associated HF occurrences.

Methods

Distributional data for filovirus-associated HF occurrences in hominids were accumulated from the literature (Table). Occurrences of unknown origin were excluded from analysis, but when reasonable guesses could be made as to point or general area of origin (e.g., 1995 outbreak of hemorrhagic fever due to Ebola Ivory Coast as originating at Plibo, Liberia), they were included. All occurrences were georeferenced (available from: URL: http://www.calle.com/world) to the nearest 0.001°. Although assigned geographic coordinates may not fix the exposure point precisely, they represent our best guess as to its position and are likely to be representative of the coarse-scale ecologic conditions. (The relatively crude spatial resolution at which analyses were conducted makes some error in georeferencing irrelevant.)

Ecologic niches and potential geographic distributions were modeled by using the Genetic Algorithm for Rule-set Prediction (GARP) (3032) (available from: URL: http://www.lifemapper.org/desktopgarp/). In general, GARP focuses on modeling ecologic niches (the conjunction of ecologic conditions wherein a species can maintain populations without immigration) (33). Specifically, GARP relates ecologic characteristics of occurrence points to those of points sampled randomly from the rest of the study region, developing a series of decision rules that best summarize factors associated with presence (23).

Occurrence points are divided evenly into training (for model building) and test (for model evaluation) datasets. GARP works in an iterative process of rule selection, evaluation, testing, and incorporation or rejection: a method is chosen from a set of possibilities (e.g., logistic regression, bioclimatic rules) and applied to the training data to develop or evolve a rule. Predictive accuracy is evaluated on the basis of the test data. Rules may evolve in ways that mimic DNA evolution (e.g., point mutations, deletions). Change in predictive accuracy between iterations is used to evaluate whether particular rules should be incorporated into the model; the algorithm runs 1,000 iterations or until convergence. Model quality was evaluated through independent test dataset reserved prior to modeling; a chi-square test was used to compare observed success in predicting the distribution of test points with that expected under a random model (proportional area predicted present provides an estimate of occurrence points correctly predicted, were the prediction random with respect to the distribution of the test points).

To characterize environments, we used 11 GIS coverages summarizing elevation, slope, aspect, flow direction, flow accumulation, and tendency to pool water (from the USGS Hydro-1K dataset [available from: URL: http://edcdaac.usgs.gov/gtopo30/hydro/), and climate characteristics, including daily temperature range; mean annual precipitation; maximum, minimum, and mean annual temperatures; solar radiation; frost days; wet days; and vapor pressure (1960–1990; Intergovernmental Panel on Climate Change [available from: URL: http://www.ipcc.ch/]). These coverages are worldwide and provide a consistent view of ecologic variation across regions studied. GARP’s predictive ability has been tested under diverse circumstances (22,23,3447).

To optimize model performance, we developed 100 replicate models of ecologic niches based on independent random subsamples from available occurrences. We chose a “best subset” of these models on the basis of optimal error distributions for individual replicate models (34): median area predicted across all replicate modes was calculated, and the 20 models with predicted areas closest to the median were chosen for further consideration. These geographic predictions were combined to provide a summary of potential geographic distributions. Projection of the Africa-based rule-sets onto maps of Asia and the Pacific provided hypotheses of potential distributional areas in other regions (46).

To permit visualization of the ecologic dimensions of models, we combined best-subsets predictions with maps of the ecologic parameters used to build them in a GIS environment (COMBINE in ArcView 3.2). The resulting dataset represents unique combinations of environments and predictions; its attributes table provides the model prediction for all environmental combinations to permit visualization of ecologic variation. We also compared ecologic conditions inside and outside of the modeled Ebola HF distribution within 11 regularly spaced circular windows (radius 50 km); comparisons were summarized through Mann-Whitney U-statistics, permitting a nonparametric visualization of the strength of association of each ecologic dimension (temperature, precipitation, elevation) with the range limit.

Results

Figure 1

Thumbnail of Summary of known and predicted geography of filoviruses in Africa. (A) Known occurrence points of filovirus hemorrhagic fevers (HFs) identified by virus species. (B) Geographic projection of ecologic niche model based on all known filovirus disease occurrences in Africa. (C) Geographic projection of ecologic niche model based on all known Ebola HF occurrences (i.e., eliminating Marburg HF occurrences). (D) Geographic projection of ecologic niche model based on all known occurrences

Figure 1. Summary of known and predicted geography of filoviruses in Africa. (A) Known occurrence points of filovirus hemorrhagic fevers (HFs) identified by virus species. (B) Geographic projection of ecologic niche model based...

The geographic distribution of filovirus disease spreads generally across the humid Afrotropics (Figure 1A). Outlier occurrences lie at the eastern extreme of the distribution, consisting of occurrences associated with Ebola Sudan and Marburg viruses. Preliminary analyses of these geographic distributions, based on random subsets of the few data points available, indicated high statistical significance to model predictions: predictions of the geographic distribution of filovirus HFs correctly included random independent subsets much better than random model expectations (all p < 10-7). Although subsequent modeling was done without subsetting to maximize occurrence data, these preliminary results nonetheless indicated excellent predictivity of our distributional hypoptheses.

Modeling the distribution of Filoviridae in general (all points in Figure 1A) produced a broad potential distribution across the Afrotropics, including areas from which filovirus HF occurrences have not been reported (Tanzania, Mozambique, Madagascar; Figure 1B). Predicted distributions of the two major Filoviridae clades—Ebola and Marburg viruses—showed different geographic patterns. When just the three African Ebola virus species were analyzed together, areas of overprediction in eastern Africa disappeared, and predicted distributional areas included only areas surrounding known occurrence points, except for a few small disjunct areas in West Africa (Figure 1C). The predicted distribution did not include all of the Afrotropics—coastal central Africa and most of West Africa appeared not to be included, although these models are based on very small samples of occurrences.

When we analyzed the relatively few Marburg HF occurrences for which distributional data exist (n = 4 occurrences), a complementary distributional area was predicted (Figure 1D). Marburg HF was predicted to be absent in the humid Afrotropics, rather appearing focused in drier areas in eastern and south-central Africa. In contrast to Ebola HF, Marburg virus appears to have the potential to occur in areas from which filovirus disease has not yet been described.

Figure 2

Thumbnail of Geographic projection of ecologic niche models in which two Ebola virus species were modeled and used to predict the distributional area of the third. (A) Ebola Zaire and Ebola Sudan predicting (Ebola Ivory Coast omitted; note that distributional area is predicted in Ivory Coast). (B) Ebola Sudan and Ebola Ivory Coast predicting (Ebola Zaire omitted). (C) Ebola Zaire and Ebola Ivory Coast predicting (Ebola Sudan omitted). Darker shades of red represent increasing confidence in predi

Figure 2. Geographic projection of ecologic niche models in which two Ebola virus species were modeled and used to predict the distributional area of the third. (A) Ebola Zaire and Ebola Sudan predicting...

Sequential omission of Ebola virus species from analyses provided a view of ecologic similarity of species (45): if omission of a particular species causes little overall change, then its ecologic characteristics are not distinct from those of the remaining species. Omission of Ebola Ivory Coast had little effect on the prediction (Figure 2A; note predicted area in Ivory Coast); similarly, predictions omitting Ebola Zaire included at least part of the distribution of Ebola Zaire (southern portion omitted; Figure 2B). Eliminating Ebola Sudan, however, yielded a prediction completely excluding the distribution of Ebola Sudan (Figure 2C), which suggests that Ebola Sudan occurs under a distinct ecologic regime.

Figure 3

Thumbnail of Ecologic distribution of predicted potential distributional areas for Ebola hemorrhagic fever (HF) occurrences, visualized in a few dimensions of climate. (A) Large-scale view (all of Africa), in which the basic concentration of Ebola HF occurrences in hot, wet climates is shown. (B) Distribution by land use/land-cover type, summarized as the proportion of overall area of land-cover types that is predicted to be present at the highest confidence level. (C,D) Regional scale (distribu

Figure 3. Ecologic distribution of predicted potential distributional areas for Ebola hemorrhagic fever (HF) occurrences, visualized in a few dimensions of climate. (A) Large-scale view (all of Africa), in which the basic concentration...

Inspection of niche models of Ebola HF occurrences (Marburg HF excluded) in ecologic space (Figure 3) provided insight into their ecologic distribution. Predicted Ebola HF occurrences were concentrated in regions presenting high precipitation combined with moderate-to-high temperatures (Figure 3A), coinciding with the ecologic distribution of evergreen broadleaf forest, although in specific cases that forest may be highly disturbed. In fact, >50% of African evergreen broadleaf forest is predicted to be within the niche of Ebola HF; no other land-cover type exceeded 5% within the Ebola HF niche (Figure 3B). In other dimensions, Ebola HF occurrences were distributed centrically in African environments and did not include extremes (Figure 3C–D).

Figure 4

Thumbnail of Preliminary exploration of patterns of ecologic variation along the modeled distributional limits (highest confidence level) for Ebola viruses in central Africa. The histograms represent relative values of Mann-Whitney U-tests for inside versus outside the prediction area for temperature (red bars), precipitation (blue bars), and elevation (green bars).

Figure 4. Preliminary exploration of patterns of ecologic variation along the modeled distributional limits (highest confidence level) for Ebola viruses in central Africa. The histograms represent relative values of Mann-Whitney U-tests for inside...

Distributional limits are complex results of multiple causal agents. A species is seldom limited on all sides by a single factor; rather, distributional limits are the combined result of many such factors. Inspection of the ecologic dimensions coincident with modeled geographic limits of Ebola HF occurrences (Figure 4) showed some of this complexity. At points around the distributional limit of Ebola HF distributional areas in central Africa, precipitation dominates the range limit at point 11, but temperature and elevation dominate at points 2, 3, and 6. Moreover, gradients are steeper in some areas than others (e.g., point 6 vs. 3). This preliminary analysis thus illustrates the complex relationships between ecologic dimensions and distributional limits.

Figure 5

Thumbnail of Projection of filovirus ecologic niche models onto southeastern Asia and the Philippines to assess the degree to which possible Philippine distributional areas are predictable on the basis of the ecologic characteristics of African filovirus hemorrhagic fever (HF) occurrences. (A) Projection of model for Marburg HF occurrences (Figure 1D) to southeastern Asia. (B) Projection of model for all filovirus disease occurrences (Figure 1B) to southeastern Asia (the projection of models for

Figure 5. Projection of filovirus ecologic niche models onto southeastern Asia and the Philippines to assess the degree to which possible Philippine distributional areas are predictable on the basis of the ecologic characteristics...

Given the mysterious origin of Ebola Reston virus (Ebola HF among macaques in a breeding facility on Luzon, Philippines) (9,13), a key question regarding Ebola HF distribution and ecology is whether similar ecologic conditions exist in Southeast Asia (e.g., Philippines). Projecting ecologic niche models for Marburg HF in Africa onto Asian environments identified few “appropriate” areas: only a few scattered areas in Papua New Guinea and Indonesia (Figure 5A). Projection of Ebola HF models, however, identified broader potential distributional areas in Southeast Asia (Figure 5B), including the lowlands of Mindanao (Figure 5, inset), a finding that suggests that similar ecologic conditions exist in the Philippines.

Discussion

Ecology and Geography of Filovirus Occurrences

The ecologic niche characteristics reconstructed for filovirus species disease outbreaks coincided closely with phylogenetic patterns in the group (1,48). That is, disease sites for Ebola Ivory Coast and Ebola Zaire coincided ecologically, and these viruses are phylogenetically sister taxa. Ebola Sudan is genetically and ecologically most distinct among Ebola virus species, and (with Ebola Reston) forms the sister clade to Ebola Ivory Coast + Ebola Zaire. Correspondence between phylogenetic and ecologic patterns suggests that ecologic distributions of Ebola Sudan and Ebola Reston may prove similar; hence, the ecologic characteristics of Ebola Sudan may provide clues about the origin of Ebola Reston.

Marburg HF occurrence sites are quite distinct, with minimal overlap with Ebola HF ecologic distributions, coinciding with Marburg virus’ distant position in the phylogeny of the Filoviridae. This pattern suggests that Marburg virus and the Ebola viruses may have host species with markedly different ecologic requirements.

Ebola Reston

The geographic origin of Ebola Reston virus has been subject of controversy (9,49). Although the Ebola virus-infected monkeys initially documented in Reston, Virginia, originated in the Philippines, whether Ebola Reston occurs naturally in the Philippines has been debated. Nevertheless, the virus is distinct, and its geographic distribution is otherwise unknown. Given the phylogeny-ecology correspondence documented above, the ecology of Ebola Sudan may prove key in predicting the distribution of Ebola Reston, but the scanty occurrence data make species-specific models difficult. Our results are relevant in that ecologic conditions under which Ebola HF occurs in Africa are also found in the Philippines.

In previous analyses of animals, the conservative nature of ecologic niches has been documented to lead to prediction into regions inhabited by congener species (45). To the extent that host-parasite codistribution and cospeciation may be involved in the virus-reservoir relationships of filoviruses, prediction of potential distributional areas in the Philippines may reflect conservative niche evolution in the host taxon. Of course, because of historical effects (e.g., limited dispersal) on species’ distributions, potential distributional areas are often predicted in areas not inhabited (44), so this evidence is not definitive.

Limitations of the Approach

Limitations of our approach should be recognized. First, small sample sizes become critical. Although predictive models can be developed with relatively small samples of occurrence points (39), sample sizes for filovirus HF disease outbreaks are so minimal that single data points can change overall results. Examples of this sensitivity include the Zimbabwe Marburg HF disease outbreak and the Booue, Gabon, Ebola Zaire HF outbreak; inclusion of these points causes geographic predictions to be expanded considerably.

Other limitations center on the ecologic dimensions in which the niche is modeled. If additional dimensions exist that limit species’ distributions (and they certainly do) , GARP predictions will be overly large. Jackknife manipulations (systematic omission of ecologic dimensions to assess sensitivity to coverage density) can, to some degree, help in assessing sensitivity to coverage completeness (42), but dimensions more important than the set actually used may exist. Particularly relevant is climate variability—extreme events such as droughts and heavy rainfall may prove particularly relevant to filovirus transmission but are not included herein; such more complex models are under development (A.T. Peterson et al., unpub. data). Spurious associations between occurrence points and ecologic dimensions, though usually detected through independent test datasets, can limit distributional predictions overmuch.

Natural Reservoirs for Filoviruses

Detailed understanding of the geography and ecology of filovirus HF outbreaks represents an underexplored avenue of investigation regarding natural transmission cycles of filoviruses. We assembled available information regarding filovirus HF outbreaks and used various analytical tools to arrive at a detailed understanding of geography and ecology of filovirus disease occurrences. Consequently, we can now assemble criteria by which potential reservoir taxa might be judged. If one assumes a fair degree of host specificity in this host-parasite system, patterns of codistribution and cophylogeny can be expected. Hence, criteria include the following: 1) African Ebola virus reservoirs would be distributed principally in evergreen broadleaf forest; 2) the main focus of the geographic distribution of the reservoir(s) would be in the Congo Basin; 3) a disjunct (allopatric) distributional area would be present in West Africa; 4) a related taxon in eastern Africa would range in more arid habitats; 5) the reservoir would belong to a clade more broadly distributed across Africa and Southeast Asia.

Assessment of potential reservoir taxa by using these criteria has begun (A.T. Peterson et al., unpub. data), with the idea of eventually testing hypotheses of host associations through ecologic niche comparison methods (22). The goal, to be explored in future publications, is to develop reduced lists of taxa of highest priority for virus survey.

Dr. Peterson is associate professor in the Department of Ecology and Evolutionary Biology and Curator in the Natural History Museum and Biodiversity Research Center, University of Kansas. His research interests focus on the geography of species’ distributions

Top

Acknowledgments

We thank David Vieglais, Ricardo Scachetti-Pereira, and David Stockwell for technical innovations; Rigoberto Esono Anvene for assistance with georeferencing localities; Emily Jentes for literature research; Tom Ksiazek, Pierre Rollin, Dan Bausch, and Darin Carroll for valuable comments on versions of the manuscript; and Ellen Salehi for the spark of interest in infectious diseases.

Much of the writing of this contribution was carried out at the Centro de Referência em Informação Ambiental, funded by the Fundação de Amparo à Pesquisa do Estado de São Paulo, Brazil, and the U.S. National Science Foundation.

Top

References

  1. Murphy  FA, Peters  CJ. Ebola virus: where does it come from and where is it going? In: Krause RM, editor. Emerging infections: biomedical research reports. San Diego: Academic Press; 1998. p. 375–410.
  2. Ebola haemorrhagic fever in Zaire, 1976. Bull World Health Organ. 1978;56:27193.PubMedGoogle Scholar
  3. Ebola haemorrhagic fever in Sudan, 1976. Bull World Health Organ. 1978;56:24770.PubMedGoogle Scholar
  4. Baron  RC, McCormick  JB, Zubeir  OA. Ebola virus disease in southern Sudan: hospital dissemination and intrafamilial spread. Bull World Health Organ. 1983;61:9971003.PubMedGoogle Scholar
  5. Bertherat  EJ, Renaut  AA, Nabias  RJ, Dubreuil  G, Georges-Courbot  M-C. Leptospirosis and Ebola virus infection in five gold-panning villages in northeastern Gabon. Am J Trop Med Hyg. 1999;60:6105.PubMedGoogle Scholar
  6. Conrad  JL, Isaacson  M, Smith  EB, Wulff  H, Crees  M, Geldenhuys  P, Epidemiologic investigation of Marburg virus disease, southern Africa, 1975. Am J Trop Med Hyg. 1978;27:12105.PubMedGoogle Scholar
  7. Formentry  P, Boesch  C, Wyers  M, Steiner  C, Donati  F, Dind  F, Ebola virus outbreak among wild chimpanzees living in a rain forest of Cote d'Ivoire. J Infect Dis. 1999;179:S1206. DOIPubMedGoogle Scholar
  8. Georges  AJ, Leroy  E, Renaut  AA, Benissan  CT, Nabias  RJ, Trinh Ngoc  M, Ebola hemorrhagic fever outbreaks in Gabon, 1994–1997: epidemiologic and health control issues. J Infect Dis. 1999;179:S6575. DOIPubMedGoogle Scholar
  9. Hayes  CG, Burans  JP, Ksiazek  TG, del Rosario  RA, Miranda  ME, Manaloto  CR, Outbreak of fatal illness among captive macaques in the Philippines caused by an Ebola-related filovirus. Am J Trop Med Hyg. 1992;46:66471.PubMedGoogle Scholar
  10. Heymann  DL, Weisfeld  JS, Webb  PA, Johnson  KM, Cairns  T, Berquist  H. Ebola hemorrhagic fever: Tandala, Zaire, 1977–1978. J Infect Dis. 1980;142:3726.PubMedGoogle Scholar
  11. Khan  AS, Kweteminga  T, Heymann  DL, Le Guenno  B, Nabeth  P, Kerstiens  B, The reemergence of Ebola hemorrhagic fever, Democratic Republic of the Congo, 1995. J Infect Dis. 1999;179:S7686. DOIPubMedGoogle Scholar
  12. Leroy  E, Souquiere  S, Rouquet  P, Drevet  D. Re-emergence of Ebola haemorrhagic fever in Gabon. Lancet. 2002;359:712. DOIGoogle Scholar
  13. Miranda  ME, Ksiazek  TG, Retuya  TJ, Khan  AS, Sanchez  A, Fulhorst  CF, Epidemiology of Ebola (subtype Reston) virus in the Philippines, 1996. J Infect Dis. 1999;179:S1159. DOIPubMedGoogle Scholar
  14. Arata  AA, Johnson  B. Approaches toward studies on potential reservoirs of viral haemorrhagic fever in southern Sudan (1977). In: Pattyn SRS, editor. Ebola virus haemorrhagic fever. New York: Elsevier; 1978. p. 191–200.
  15. Germain  M. Collection of mammals and arthropods during the epidemic of haemorrhagic fever in Zaire. In: Pattyn SRS, editor. Ebola virus haemorrhagic fever. New York: Elsevier; 1978. p. 185–9.
  16. Leirs  H, Mills  JN, Krebs  JW, Childs  JE, Akaibe  D, Woollen  N, Search for the Ebola virus reservoir in Kikwit, Democratic Republic of the Congo: Reflections on a vertebrate collection. J Infect Dis. 1999;179:S15563. DOIPubMedGoogle Scholar
  17. Swanepoel  R, Leman  PA, Burt  FJ, Zachariades  NA, Braack  LEO, Ksiazek  TG, Experimental inoculation of plants and animals with Ebola virus. Emerg Infect Dis. 1996;2:3215.PubMedGoogle Scholar
  18. Turell  MJ, Bressler  DS, Rossi  CA. Lack of virus replication in arthropods after intrathoracic inoculation of Ebola Reston virus. Am J Trop Med Hyg. 1996;55:8990.PubMedGoogle Scholar
  19. Smith  DH, Johnson  BK, Isaacson  M, Swanepoel  R, Johnson  KM, Killey  M, Marburg-virus disease in Kenya. Lancet. 1982;1:81620. DOIPubMedGoogle Scholar
  20. Murphy  FA, Kiley  MP, Fisher-Hoch  SP. Filoviridae: Marburg and Ebola viruses. In: Fields BN, Knipe DM, editors. Virology. New York: Raven Press, Ltd.; 1990. p. 933–42.
  21. Clayton  DH, Moore  J. Host-parasite co-evolution. General principles and avian models. Oxford: Oxford University Press; 1997.
  22. Peterson  AT, Sanchez-Cordero  V, Beard  CB, Ramsey  JM. Ecologic niche modeling and potential reservoirs for Chagas disease, Mexico. Emerg Infect Dis. 2002;8:6627.PubMedGoogle Scholar
  23. Peterson  AT, Stockwell  DRB, Kluza  DA. Distributional prediction based on ecological niche modeling of primary occurrence data. In: Scott JM, editor. Predicting species occurrences: issues of scale and accuracy. Washington: Island Press; 2002. p. 617–23.
  24. Le Guenno  B, Formentry  P, Wyers  M, Guonon  P, Walker  F, Boesch  C. Isolation and partial characterisation of a new strain of Ebola virus. Lancet. 1995;345:12714. DOIPubMedGoogle Scholar
  25. World Health Organization. Suspect cases of Ebola in Liberia. Press Release 11 December 1995. Available from: URL: http://www.who.int/archives/inf-pr-1995/r95-90.html
  26. Centers for Disease Control and Prevention. Outbreak of Ebola hemorrhagic fever—Uganda, August 2000–January 2001. MMWR Morb Mortal Wkly Rep. 2001;50:737.PubMedGoogle Scholar
  27. Bertherat  EJ, Talarmin  A, Zeller  H. [Democratic Republic of the Congo: between civil war and the Marburg virus]. Med Trop. 1999;59:2014.
  28. Zeller  H. [Lessons from the Marburg virus epidemic in Durba, Democratic Republic of the Congo (1998–2000)]. Med Trop. 2000;60:236.
  29. World Health Organization. Marburg disease in Democratic Republic of Congo—update 2. Communicable Disease Surveillance and Response Disease Outbreak News. 6 May 1999.
  30. Stockwell  DRB, Peters  DP. The GARP modelling system: problems and solutions to automated spatial prediction. International Journal of Geographic Information Systems. 1999;13:14358. DOIGoogle Scholar
  31. Stockwell  DRB, Noble  IR. Induction of sets of rules from animal distribution data: A robust and informative method of analysis. Math Comput Simul. 1992;33:38590. DOIGoogle Scholar
  32. Stockwell  DRB. Genetic algorithms II. In: Fielding AH, editor. Machine learning methods for ecological applications. Boston: Kluwer Academic Publishers; 1999. p. 123–44.
  33. MacArthur  R. Geographical ecology. Princeton (NJ): Princeton University Press; 1972.
  34. Anderson  RP, Lew  D, Peterson  AT. Using intermodel variation in error components to select best subsets of ecological niche models. Ecol Modell. 2003;162:21132. DOIGoogle Scholar
  35. Anderson  RP, Laverde  M, Peterson  AT. Using niche-based GIS modeling to test geographic predictions of competitive exclusion and competitive release in South American pocket mice. Oikos. 2002;93:316. DOIGoogle Scholar
  36. Anderson  RP, Laverde  M, Peterson  AT. Geographical distributions of spiny pocket mice in South America: insights from predictive models. Glob Ecol Biogeogr. 2002;11:13141. DOIGoogle Scholar
  37. Feria  TP, Peterson  AT. Using point occurrence data and inferential algorithms to predict local communities of birds. Divers Distrib. 2002;8:4956. DOIGoogle Scholar
  38. Stockwell  DRB, Peterson  AT. Controlling bias in biodiversity data. In: Scott JM, editor. Predicting species occurrences: issues of scale and accuracy. Washington: Island Press; 2002. p. 537–46.
  39. Stockwell  DRB, Peterson  AT. Effects of sample size on accuracy of species distribution models. Ecol Modell. 2002;148:113. DOIGoogle Scholar
  40. Peterson  AT. Predicting species’ geographic distributions based on ecological niche modeling. Condor. 2001;103:599605. DOIGoogle Scholar
  41. Peterson  AT, Ball  LG, Cohoon  KC. Predicting distributions of tropical birds. Ibis. 2002;144:e2732. DOIGoogle Scholar
  42. Peterson  AT, Cohoon  KC. Sensitivity of distributional prediction algorithms to geographic data completeness. Ecol Modell. 1999;117:15964. DOIGoogle Scholar
  43. Peterson  AT, Papes  M, Kluza  DA. Predicting the potential invasive distributions of four alien plant species in North America. Weed Sci. In press.
  44. Peterson  AT, Sanchez-Cordero  V, Soberon  J, Bartley  J, Buddemeier  RH, Navarro-Siguenza  AG. Effects of global climate change on geographic distributions of Mexican Cracidae. Ecol Modell. 2001;144:2130. DOIGoogle Scholar
  45. Peterson  AT, Soberon  J, Sanchez-Cordero  V. Conservatism of ecological niches in evolutionary time. Science. 1999;285:12657. DOIPubMedGoogle Scholar
  46. Peterson  AT, Vieglais  DA. Predicting species invasions using ecological niche modeling. Bioscience. 2001;51:36371. DOIGoogle Scholar
  47. Peterson  AT, Scachetti-Pereira  R, Hargrove  WW. Potential distribution of Asian longhorned beetles (Anoplophora glabripennis) in North America. Am Midl Nat. 2004. In press.
  48. Georges-Courbot  M-C, Sanchez  A, Lu  C-Y, Baize  S, Leroy  E, Lansout-Soukate  J, Isolation and phylogenetic characterization of Ebola viruses causing different outbreaks in Gabon. Emerg Infect Dis. 1997;3:5962.PubMedGoogle Scholar
  49. Peters  CJ, Johnson  ED, Jahrling  PB, Ksiazek  TG, Rollin  PE, White  J, In: Morse SS, editor. Emerging viruses. Oxford: Oxford University Press; 1993. p. 159–75.

Top

Figures
Table

Top

Cite This Article

DOI: 10.3201/eid1001.030125

1Although filovirus taxonomy undergoes frequent revision, we follow nomenclature most recently established by the International Committee of Taxonomy of Viruses in 2002. Filoviruses consist of two genera. The genus Marburgvirus contains one species, Lake Victoria marburgvirus, with several recognized strains, and Ebolavirus contains four species: Ivory Coast ebolavirus, Reston ebolavirus, Sudan ebolavirus, and Zaire ebolavirus. For simplicity, we refer to the viruses using the unitalicized vernacular (e.g., Ebola Zaire). We use “Ebola viruses” to refer in general to members of Ebolavirus and “Marburg viruses” to refer in general to members of Marburgvirus. The diseases caused by filoviruses are termed Ebola hemorrhagic fever (HF; diseases caused by Ebola viruses) and Marburg HF (diseases caused by Marburg viruses).

Table of Contents – Volume 10, Number 1—January 2004

EID Search Options
presentation_01 Advanced Article Search – Search articles by author and/or keyword.
presentation_01 Articles by Country Search – Search articles by the topic country.
presentation_01 Article Type Search – Search articles by article type and issue.

Top

Comments

Please use the form below to submit correspondence to the authors or contact them at the following address:

A. Townsend Peterson, Natural History Museum and Biodiversity Research Center, University of Kansas, Lawrence, Kansas 66045, USA; fax: 785-864-5335

Send To

10000 character(s) remaining.

Top

Page created: December 21, 2010
Page updated: December 21, 2010
Page reviewed: December 21, 2010
The conclusions, findings, and opinions expressed by authors contributing to this journal do not necessarily reflect the official position of the U.S. Department of Health and Human Services, the Public Health Service, the Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.
file_external