Remote sensing mangrove forest phenology

Mangrove forests are evergreen intertidal salt-tolerant communities that offer numerous ecosystems goods and services, and play an important role in the carbon cycle on the coastal area. Despite the recognition of their benefits, these tropical and subtropical forests are under anthropogenic pressure such as urban expansion, agriculture, aquaculture, degradation and deforestation. Therefore, there is a need to have up-to-date, spatially continuous information to assess mangrove status. This can be achieved through the accurate estimation of biophysical variables and through the understanding of spatio-temporal dynamics. This thesis focuses on the dynamics of mangrove forests from regional to global scale using in situ and multi-temporal remote sensing data to retrieve biophysical information, investigates mangrove forest phenology and its environmental drivers. The particular objectives are to: (i) assess the potential leaf hyperspectral data and Landast-8 using OLS and Machine Learning algorithms to estimate mangrove chlorophyll concentration, (ii) characterise the mangrove forest phenology at regional scale and to validate with ground biophysical data, (iii) for the first time characterise the mangrove forest phenology at global scale and provide insights on the global environmental drivers. The results from these studies could provide reliable information for rapid assessment of the status of mangroves forests. They also serve as baseline to compare future changes derived from anthropogenic degradation and the changing climate. Future work should focus on the use of new optical satellites with improved capabilities to retrieve biophysical variables. Given the importance of mangroves as carbon sinks future work should also focus on the link between remotely estimated biophysical properties and carbon exchange at ground level