Abstract
The situation of world’s mangroves remains a matter of concern, despite the public awareness about their role in the sustainability of coastal biodiversity. Extensive and intensive aquaculture practies are mainly called into question. The need for a better integrated coastal zone management arises since a couple of decades. It must be recognized that progress is insufficient especially in Indonesia.The present PhD work has been carried out in the frame of the INDESO project. The objective was to describe mangrove changes in terms of area expansion, forest structure and species composition before studying, at fine scales, the diversity of mangrove stands, their regeneration capacity and the evolution trends of the whole ecosystem in a given region.In order to achieve this aim, i used 28 very high spatial resolution (VHSR) optical satellite images aquired between 2001 and 2015. I also participated in forest inventories and species identification from the seedling to the adult stages. I visually analysed the set of images for allowing the discrimination of planted mangroves from natural ones. I examined the influence of acquisition geometry parameters on the variability of mangrove spectral signatures. Thanks to supervised maximum likelihood classifications, I succeeded in evaluating the expansion area of mangrove with an accurracy of few square meters and I proceeded in assessing the robustness of fine scale mapping of mangrove zonations.Results obtained in the analysis of the Perancak estuary highlighted that mangrove area is expanding since 2001. The plantation practices were however presented as disrespectful of any sustainable management plan. Indeed, our simple field data indicated strong differences between mangrove plantations and natural forests. Plantations were shown to be cultures of only Rhizophora species with high planting density whereas natural forest stands are composed of Sonneratia and Avicennia mangrove trees. In addition, we highlighted that the regeneration capacity in planted areas remains very lower than the ones clearly observed in natural areas. Moreover, the only regenetive process in plantations seems to be with native species and not with the planted ones. From the analysis of the VHSR images, we identified aquaculture ponds that were planted and their planting year. Consequently, we suceeded in monitoring the mangrove expansion inside every planted pond. Important variability of expansion rate was pointed out, this suggesting contrasting environments in each pond and the inadequacy of cultures of Rhizophora everywhere. The observation of natural colonization in many areas of the estuary, pond floors included, was an interesting point to mention.Besides, results obtained in the study of the protected mangrove region of Nusa Lembongan assessed the potential of multispectral VHSR satellite images for mapping mangrove zonations. Four mangrove zonations mainly dominated by different species exhibit signatures that can be differentiated over years. Different image acquisition parameters may, however, generate confusion in classification. This is particularyly true when forest stands with open canopies are observed with sun frontward or with a near vertical viewing angle because of sun light interaction with the clear waters. We probably obtained the first detailed maps of mangrove zonations in the region. They showed a complex spatial organization of mangrove zonations.The whole of my work advocated for further remote sensing works aiming at the physical interpretation of spectral and textural signatures of mangroves. It provided recommendations such as a moratorium onf Rhizophora plantation and simple methods that will be, we hope, starting points for closing the knowledge gap that hampers monitoring of future coastal zone management plans.