Article révisé par les pairs
Résumé : Large ecosystem processes often take place beyond the observation time of a researcher. Yet, through retrospective research scientists can approach and understand ecosystem changes. This contributes to the fundamental understanding of both human-induced and natural dynamics in ecosystems world-wide. This also holds for fast changing coastal areas with mangrove ecosystems, which are important for biodiversity, for coastal protection, and for the daily livelihood of millions of people in tropical coastal developing countries. In addition, retrospective research generates a basis for predictions that can be used early on to protect an ecosystem. In attempting to protect ecosystems from adverse human-induced change and destruction, and to manage them for sustainability, scientists are only beginning to investigate and understand natural ecosystem dynamics. It is important and advisable to gather, combine and analyse all possible data that allow a researcher to look back in time. This paper reviews the available retrospective methods, and highlights the transdisciplinary way (i.e. combination between basic and applied sciences on one hand, and social and human sciences on the other) in which retrospective research on a scale between months and centuries can be carried out, but it also includes methods on larger scales that may be marginally relevant. The paper particularly emphasizes the lack of transdisciplinary (not interdisciplinary) integration between sciences in retrospective research on mangrove forests in the past. © 2008 Elsevier B.V. All rights reserved.