ECOLOGY AND GEOGRAPHY OF AVIAN VIRUSES USING NICHE MODELS AND WILD BIRD SURVEILLANCE

The emergence of highly pathogenic avian influenza strain H5N1 (hereafter "H5N1"), and other bird-associated viruses, have raised serious concerns about impacts on human, livestock, and wildlife populations. Ecological niche modelling (ENM) techniques were used to test the hypothesis that spatial distributions of H5N1 cases are related to coarse-scale environmental features in West Africa and in the Middle East and north-eastern Africa. Areas of drought-sensitive vegetation phenology were identified as key to H5N1 transmission, notwithstanding a small minority of models which indicated more variable transmission environments. ENMs were further used to estimate the environmental distribution of H5N1 relative to host group (poultry, wild birds, etc) in Europe. Results revealed no distinct ecological niche requirements among H5N1 host groups, suggesting that transmission cycles are broadly interwoven. Finally, avian virus surveillance was carried out in Ghana and Peru to assess patterns of host association and to test the assumption that avian influenza (AI) prevalence is low or nil in land birds. 600 Peruvian land birds of 177 species were tested for AI using rRT-PCR, revealing an infection prevalence of 1.3%. 564 Ghanaian land birds of 146 species were tested for AI (and Alphaviruses, and Flaviviruses) using PCR techniques. Samples were negative for Alphaviruses and AI, but amplified one sequence of a Yaoundé-like Flavivirus. Results of AI surveillance highlight the spatial variation of AI prevalences. Nonetheless, the prevalence in Peru demonstrates that surveillance programs for monitoring spread and identification of AI viruses should not focus solely on water birds