Abstract
Mangroves are ecosystems that link freshwater, land, and oceans. They are also home to vast biodiversity of species and provide vital resources to many coastal communities worldwide. However, in various parts of the world, these natural ecosystems have been subjected to anthropogenic actions that compromise their subsistence. Therefore, the aim of this study was to analyse seasonal changes in vegetation cover in the La Caimanera marsh (Colombia) using remote sensing and satellite sensor imagery. Methodologically, machine learning techniques were used, including semi-supervised and supervised learning, using Landsat, Copernicus, and Planet images. The influence of mangrove vegetation cover on pollutants was explored using remote sensing data and in-situ measurements. The results obtained indicated that in 2015 and 2024, the ecosystem showed stability, with slight reductions in mangrove cover and water bodies. In contrast, a slight increase in urbanized areas was also evident. The NDWI, MNDWI, OC3, and QAA spectral indices were used to monitor water dynamics and water quality. The results reflected stability in water conditions between 2010 and 2020, with a slight reduction in 2023. The increase in chlorophyll-a and reduction in turbidity in 2023 showed alterations in water quality. In addition, the predominance of mature mangrove trees, which comprised 83% of the vegetation, is noteworthy, reflecting a healthy and stable ecosystem. The greater density and homogeneity of the mangrove canopy observed in this study suggests a positive ecosystem response and greater resilience to past environmental changes. It is concluded that the changes experienced by the La Caimanera marsh reflect a balance between conservation and development, highlighting the effectiveness of territorial management policies implemented to maintain the ecosystem's resilience in the face of human pressures.