We developed an epidemiological model of avian malaria (Plasmodium relictum) across an altitudinal gradient on the island of Hawaii that includes the dynamics of the host, vector, and parasite. This introduced mosquito-borne disease is hypothesized to have contributed to extinctions and major shifts in the altitudinal distribution of highly susceptible native forest birds. Our goal was to better understand how biotic and abiotic factors influence the intensity of malaria transmission and impact on susceptible populations of native Hawaiian forest birds. Our model illustrates key patterns in the malaria-forest bird system: high malaria transmission in low-elevation forests with minor seasonal or annual variation in infection;episodic transmission in mid-elevation forests with site-to-site, seasonal, and annual variation depending on mosquito dynamics;and disease refugia in high-elevation forests with only slight risk of infection during summer. These infection patterns are driven by temperature and rainfall effects on parasite incubation period and mosquito dynamics across an elevational gradient and the availability of larval habitat, especially in mid-elevation forests. The results from our model suggest that disease is likely a key factor in causing population decline or restricting the distribution of many susceptible Hawaiian species and preventing the recovery of other vulnerable species. The model also provides a framework for the evaluation of factors influencing disease transmission and alternative disease control programs, and to evaluate the impact of climate change on disease cycles and bird populations. ??2011 by the Ecological Society of America.