Decision analysis to guide recovery of the po'ouli, a critically endangered Hawaiian honeycreeper

We used a probabilistic decision tree to help select among alternative recovery strategies for the Po'ouli, a critically endangered forest bird endemic to the Hawaiian island of Maui. The Po'ouli is one of the rarest birds in the world, with only three individuals known, and no breeding pairs. The most urgent conservation need for the species is to create a breeding pair and obtain eggs for captive propagation. Seven recovery strategies had been proposed, and there was disagreement among conservation workers about which strategy to pursue. In addition to lengthy discussions, a decision tree was introduced to provide an objective way of quantifying the chance of success under each alternative. All available information was used to attempt evaluation of each node in the decision tree. The overall chance of obtaining eggs was calculated by multiplying estimates at each node under each alternative. All options produced low estimates of potential success, but two options, removal of all birds to captivity and placement of the birds in a field aviary in an accessible location, produced similar estimates of success. The US Fish and Wildlife Service and the Hawaii Division of Forestry and Wildlife eventually agreed that removal to captivity was the preferred alternative because it could be implemented quickly, which was important considering the advanced age of the birds, and would make it easier to provide veterinary care and ensure their safety from predators, severe weather, and vandals. Decision trees and other structured models should not be relied on exclusively, but they can provide an objective method of helping to make difficult conservation decisions and provide a record of complex thought processes used in reaching a determination. The case of the Po'ouli may serve as a template for navigation towards a decision to recover other species of extreme rarity.