COĞRAFİ YAPAY ZEKA TEKNİKLERİ İLE AKDENİZ HAVZASINDA POTANSİYEL DENİZ ÇAYIRI ALANLARININ BELİRLENMESİ

In today's world, biodiversity faces many challenges due to global climate change and increasing environmental problems. Seagrass meadows, also referred to as the lungs of the sea, play a crucial role in ecosystems. Seagrasses found in shallow waters reached by sunlight are considered the lungs of the sea due to their capacity to produce oxygen. Additionally, seagrass meadows provide habitats for marine organisms and serve as a food source for grazers. This research aims to predict the habitat areas of seagrass meadows in the Mediterranean using geospatial artificial intelligence (GeoAI) techniques. GeoAI techniques were employed to determine potential seagrass areas. An artificial intelligence model was developed to create a habitat distribution model considering the conditions of regions where the species exists within the study area. Seven spatial variables influencing the habitat conditions of seagrasses were identified in this study. Data sets from the Copernicus Marine Monitoring Service for the year 2023 were utilized. Data on existing seagrass meadows from the same source were used to train the artificial intelligence model. Within the ArcGIS Pro 3.2 software, the Presence-only Prediction (MaxEnt) tool was employed to predict the probability of species presence according to the Maximum Entropy method. The cutoff value in the model was set to 0.5, resulting in a 67.7% accuracy in model prediction based on classification results. The resulting map shows areas with significantly high probability values, indicating widespread seagrass habitat presence. These areas include the southern coast of Turkey, the Gulf of Iskenderun, the eastern and western coasts of the Aegean Sea, the southern parts of the Eastern Mediterranean Sea, the coasts of Egypt in the Levantine Sea, the coasts of the Central and Northern Adriatic Sea, the southern part of Sicily, especially the coasts of Tunisia, and the eastern-southeastern coasts of Spain in the Western Mediterranean. The results demonstrate that geospatial artificial intelligence techniques can effectively be used to predict ecosystems and identify terrestrial or marine habitats. The findings of this study are believed to have the potential to guide efforts in determining the distribution and conservation of habitats.