Résumé
Mangrove restoration has been attempted in the 7 km² Perancak Estuary, Bali since the 90's following shrimp farming loss of profitability. We characterized the surface types of estuarine zone and monitored mangrove changes since 2001 using a series of 20 very high resolution satellite images acquired over the estuary. Mangrove development was further assessed by forest inventory conducted in 2014 and diameter increment measurement using dendrometer bands installed on 83 mangrove trees in ponds and intact forests since 2011. The Perancak Estuary was cove red by over 1350 shrimp ponds accounting for 63% of the estuary area in 2014. Rhizophora planted mangroves were mostly found in the ponds whereas natural mangroves mainly established outside ponds with a clear dominance of Avicennia and Sonneratia genera. It is worth saying that, in ponds hydrologically reconnected to tidal flows due to damage dykes, natural recruitment of the two previous genera was observed. The 2001-2014 spatio-temporal analysis over the entire estuary revealed that natural mangroves have expanded from 20 to 65 ha whereas planted mangroves extent varied from 20 to 50 ha. Stretching newly mangroves in the ponds and intact areas resulted in 'greening' event in the estuary. However different characteristics of mangrove development between Rhizophora monocultures and natural mangroves were found. Sonneratia and Avicennia became dominant in the ponds as they demonstrated rapid increase in size compared to Rhizophora trees with annual mean individual tree diameter increment reaching 1.1, 0.9 cm and 0.5 cm, respectively. Even if variation in mangrove growth is observed between in the pond and intact areas, our findings proved that Avicennia and Sonneratia perform better than replanted mangroves of Rhizophora, suggesting that the adequacy of monospecies Rhizophora plantations for sustainable restoration of estuarine functions should be further questioned. (Texte intégral)