| Abstrakt: |
Habitat fragmentation has the potential to impact animal diversity by diminishing ecological connectivity, particularly in estuarine and mangrove environments, where numerous aquatic organisms undergo their early developmental stages. In this study, an investigation was conducted to explore this scenario by employing deep learning models and collecting two species of Sillago larvae and juveniles from the Ka Long estuary in northern Vietnam from September 2022 to August 2023. The utilization of combined model, consisting of particle swarm optimization and gradient boosting regression, allowed for the forecasting of the spatial distribution of Sillago fish in the context of habitat fragmentation. The findings indicate that areas adjacent to mangroves exhibit a higher abundance of Sillago spp. larvae and juveniles. Through the implementation of a deep convolutional neural network approach, it is evident that there have been significant changes in land use/land cover between 2013 and 2023, characterized by a decrease in mangrove forests and waterbodies due to urbanization. In addition, landscape indices reveal a decline in habitat connectivity and an increase in the number of edges, indirectly affecting the fish larvae and juveniles inhabiting mangroves. By employing genetic algorithm analysis, this study forecasts the distribution of Sillago larval and juvenile populations under the influence of landscape fragmentation in 2033 and 2043, predicting a dramatic decline in Sillago density. However, it is projected that the fish will still choose to reside in mangrove forests, underscoring the importance of conserving this environment for the preservation of this commercially valuable fish both at the study site and in neighboring mangrove forests in northern Vietnam. [ABSTRACT FROM AUTHOR] |
Full text from SpringerLink