Phenology impact on mangrove area estimation pre- and post a cyclone in Fiji using Sentinel-1 imagery

Mangroves offer many socioeconomic and ecological benefits which can be lost due to damage caused by tropical cyclones. Thus, it is important to monitor the damage and recovery more accurately after these events to protect their benefits. In this study, mapping of mangrove damage and recovery was done using Sentinel-1 radar data after Tropical Cyclone Yasa in Suva-Navua, Fiji. Mature mangroves first needed to be discriminated from herbaceous, sapling and damaged mangroves, and then the distribution of mangroves corresponding to phenology seasons was calculated over three years. The ability to discriminate mature mangroves was found to vary with phenological season and before/after the cyclone event. Before the cyclone, poor classification of mature mangroves was observed in the Start of Season and the Peak Growth seasons due to full canopy cover causing radar signal attenuation. After the cyclone event, when there was damage to the mature mangroves, the reduced canopy cover improved the ability to discriminate the mature mangroves as the radar attenuation was reduced. A 7% loss of mature mangroves was observed from the End of Season to Start of Season in the post-cyclone period and a subsequent 6% gain was then observed from Start of Season to Peak Growth season in Suva-Navua. Time series Sentinel-1 data can be used to discriminate mature mangroves from other vegetation types in tropical regions where optical data acquisition is significantly impacted by cloud cover; however, it is important to account for phenological seasons to improve estimates of mature mangrove damage and recovery due to canopy-signal attenuation relationship. The next step in this research is to explore if there is optimal time series of radar data for monitoring mature mangroves in a post-cyclone period.