A cost-effective method to map mangrove forest extent, composition, and condition in small islands based on Sentinel-2 data: Implications for management

• Published in: Ecological indicators 2024-02, Vol.159, p.111696, Article 111696
• Main Authors: Casal, Gema, Trégarot, Ewan, Cornet, Cindy C., McCarthy, Tim, van der Geest, Matthijs
• Format: Article
• Language: English
• Subjects: Coastal monitoring; Ecological condition; Leaf Area Index; Net Primary Productivity; SL2P
• ISSN: 1470-160X; 1872-7034
• DOI: 10.1016/j.ecolind.2024.111696

•Sentinel-2 effectively mapped extent, composition, and health of mangrove forests in Lac Bay, Bonaire.•Maximum Likelihood outperformed NDVI, showing robust results for mapping mangrove forest extent.•NDVI-based algorithms are effective for deriving accurate LAI using Sentinel-2 data.•SL2P underestimated LAI values for the mangrove canopy in Lac Bay. Despite their ecological, economic, and social importance, mangrove ecosystems suffer from high levels of degradation caused by a combination of anthropogenic stressors and the effects of climate change. Their degradation inevitably reduces the provision of ecosystem services and ultimately impacts human well-being, especially in coastal communities of Small Islands Developing States (SIDS). To timely identify and manage stressors causing local mangrove degradation, in situ monitoring is required. However, the financial means and human capacity to do so are often limited in SIDS, hampering adequate management of their mangrove forests. In search of a cost-effective alternative, we evaluated the use of Sentinel-2 satellites to monitor mangrove extent and species distribution in Lac Bay, a bay located on the small tropical island of Bonaire (Caribbean Netherlands). We also evaluated the mangrove’s ecological condition through two biophysical variables 1) Effective Leaf Area Index (LAIe) and 2) Net Primary Productivity (NPP). Our results showed that Sentinel-2 data are a valuable tool for mapping the extent of mangrove forests in Bonaire and species composition (mean overall accuracy > 95 %). Using five Sentinel-2 images from 2021 and 2022, the extent of mangrove forests in Lac Bay was estimated to be on average 222.3 ha, of which 136.0 ha were classified as Rhizophora mangle (red mangrove) and 77.1 ha as Avicennia germinans (black mangrove). Mean values for predicted LAIe ranged from 3.37 to 3.85 for Lac Bay, with significantly higher values in the wet season (3.82 ± 0.57) compared to the dry season (3.40 ± 0.56). The generic Simplified Level-2 Prototype Processor (SL2P) underestimated the LAIe values in Lac Bay, with moderate differences between SL2P values and in situ data (BDE = 0.41, RMSE = 1.09). Mean NPP values were estimated to be 8.82 ± 1.46 (g Cm−2 d -1). LAIe and NPP maps showed a zonal distribution, with highest values in the mid-West and East on the seaward side, and lowest values in the northern landward part of Lac Bay. The method developed in this study provides a cost-effective way to monitor the