Forest recolonization monitoring based on HR and VHR imagery: The case of the Maido Forest fire exploiting Pléiades and SPOT KALIDEOS database

Abstract

This paper explores the potential of the Pléiades images for the monitoring of the recolonization of Maido Forest in a post-fire framework. This site, located in the heart of the National Park of Reunion and classified as a UNESCO World Heritage Site, has been affected by several wildfires late October 2011, ravaging an area of approximately 2700 ha covered mostly by endemic species. To understand the state and the phenological behavior of pre and post fire vegetation, this study uses the SPOT KALIDEOS archive (since 2005) combined with field information. Then, a series of post-fire Pleiades image acquired from June 2012 to September 2013 plus one pre-event Worldview-2 image were used to assess the contribution of very high spatial resolution to monitor forest recolonization at a plot scale. A total of 66 SPOT and 10 very high resolution (Pléiades and WorldView-2) images have been integrated and exploited. Spectral vegetation indexes like NDVI or TSAVI have been analyzed on multi-temporal series and have highlighted the forest vegetation cycle as well as outside perturbation phases than closer to the event and after. The NRI index allowed characterization and monitoring of post-fire vegetation recolonization. Finally, the texture provided by the very high resolution images helped to confirm vegetation recovery and to qualify this recolonization phenomenon consisting of a vegetation apparently different from the original.