Habitat suitability models of Indian mackerel (Rastrelliger kanagurta) in Indonesia Fishery Management Area 713 using remote sensing data and machine learning technique

Indian mackerel (Rastrelliger kanagurta) is an important pelagic species within Indonesia fishery management area 713. Increasing fishing pressure is considered to be the primary cause of the species’ population decline. To support the sustainable utilization of Indian mackerel, this study integrates monthly presence data from January 2018 to December 2022 with remotely sensed environmental variables to predict the species' habitat distribution using random forest algorithm. The environmental variables utilized include chlorophyll-a (Chl), sea surface height (SSH), sea surface salinity (SSS), and sea surface temperature (SST). The results indicate that the random forest algorithm performed satisfactorily, as evidenced by the overall accuracy of 0.944±0.026, true skill statistic (TSS) of 0.889±0.051, and area under the curve (AUC) of the receiver operating characteristic of 0.982±0.019. Chl, SSH, and SSS emerge as the most influential predictors affecting the spatial distribution of the species. The highest potential habitats (HSI > 0.8) were observed along the coast of South Kalimantan to Kangean Island, with smaller patches identified in the waters of West Sulawesi and South Sulawesi. This study accentuates the effectiveness of remote sensing data and machine learning techniques in determining the habitat of pelagic fish. Moreover, the findings of the study can aid decision-making related to sustainable fishery management in the region. © 2025, BIOFLUX SRL. All rights reserved.