Predicting ecosystem changes by a new model of ecosystem evolution.

In recent years, computer simulation has been increasingly used to predict changes in actual ecosystems. In these studies, snapshots of ecosystems at certain points in time were instantly constructed without considering their evolutionary histories. However, it may not be possible to correctly predict future events unless their evolutionary processes are considered. In this study, we developed a new ecosystem model for reproducing the evolutionary process on an oceanic island, targeting Nakoudojima Island of the Ogasawara Islands. This model successfully reproduced the primitive ecosystem (the entire island covered with forest) prior to the invasion of alien species. Also, by adding multiple alien species to this ecosystem, we were able to reproduce temporal changes in the ecosystem of Nakoudojima Island after invasion of alien species. Then, we performed simulations in which feral goats were eradicated, as had actually been done on the island; these suggested that after the eradication of feral goats, forests were unlikely to be restored. In the ecosystems in which forests were not restored, arboreous plants with a high growth rate colonized during the early stage of evolution. As arboreous plants with a high growth rate consume a large amount of nutrient in soil, creating an oligotrophic state. As a result, plants cannot grow, and animal species that rely on plants cannot maintain their biomass. Consequently, many animals and plants become extinct as they cannot endure disturbances by alien species, and the ecosystem loses its resilience. Therefore, even if feral goats are eradicated, forests are not restored. Thus, the founder effect from the distant past influences future ecosystem changes. Our findings show that it is useful to consider the evolutionary process of an ecosystem in predicting its future events.