Impacts of aspergillosis on sea fan coral demography: modeling a moving target

Little is known about how epizootics in natural populations affect vital rates and population structure, or about the process of recovery after an outbreak subsides. We investigated the effects of aspergillosis, an infectious disease caused by the fungal pathogen Aspergillus sydowii, on the demography of a gorgonian coral, Gorgonia ventalina. Caribbean sea fans were affected by a seven-year epizootic, marked by an initial period in 1994 of high infection prevalence, high mortality rates, and almost complete reproductive failure of infected fans. Post epizootic, in 2005, host populations were relatively healthy, with low disease prevalence. Using longitudinal data from populations on coral reefs in the Florida Keys (USA) and the Yucatan Peninsula (Mexico), we documented changes in the epidemiology of sea fan aspergillosis over the course of the epizootic. We developed an "integral projection model" that scales disease impacts from individual to population levels using direct estimates of vital rates. Within-colony lesion growth rate and host mortality were higher during the peak of the epizootic. Effects on individuals and populations changed substantially post-epizootic; recruitment increased, mortality of infected adults decreased, and the size dependence of infection was reduced. Elasticity analysis indicated that population growth is more sensitive to changes in the growth and survival of established colonies than to recruitment, due to slow colony growth and the longevity and fecundity of large adults. Disease prevalence in our monitored populations decreased from similar to 50% in 1997 to <10% by 2003 and <1% in 2007 and was accompanied by very high mortality during the early stages of the epizootic. The population model suggested that host evolution (due to selection for higher disease resistance through differential mortality) could proceed quickly enough to explain the observed changes in prevalence and in the size independence of infection risk. Our model indicates that the time required for population recovery following an outbreak is largely determined by the percentage of healthy tissue lost from the population. However, recovery following an especially severe outbreak (i.e., 80% or more tissue loss) is much faster if the affected population receives an external supply of recruits from unaffected areas.