Cloud Computing and Machine Learning for Analyzing Spatiotemporal Dynamics of Mangrove Ecosystems in the Grand Saloum (Senegal and Gambia)

Abstract

The Grand Saloum is characterized by a vast coastal plain cut by a dense hydrographic network and populated by mangrove plant formations. It is an ecosystem of capital importance in view of its ecological, socio-economic and environmental role. However, the Saloum delta remains a complex and very sensitive environment, particularly in the context of climate change. It therefore deserves special attention for better conservation. The objective of this study is to analyze the spatiotemporal dynamics of its mangrove ecosystems in relation to the variability of rainfall. The methodology is based on the exploitation of Landsat satellite images time series using Machine Learning technic from the Google Earth Engine platform to make the diachronic maps of mangrove ecosystems and analyze its relationship with rainfall. The results showed an expansion of mangrove areas in the Gambian part where the surface increased from 9 381 ha in 1988 to 11611 ha in 2020 which represents an overall growth of 23,8%. In the Senegalese part, mangrove surface increased from 52 616 ha to 62 300 between 1988 and 2020 which is +18% growth. The detection of changes showed an important development of mangrove along the Saloum during the first decade and a strong growth in the Gambian part from the 2000s. The vegetation index showed a regeneration of the mangrove between 2000 and 2020. The temporal dynamics of the mangrove is strongly correlated with the rainfall variability.