Comparing Vessel Monitoring System and Logbook Data for Skipjack Tuna Habitat Modeling in Eastern Indian Ocean

Abstract

Skipjack tuna (Katsuwonus pelamis) is important to Indonesian fisheries, which lead to the need of accurate habitat prediction for sustainable management. This study assesses the spatiotemporal coverage and predictive utility of logbook and Vessel Monitoring System (VMS) data for skipjack habitat modeling using MaxEnt, with sea surface temperature (SST), chlorophyll-a, sea surface height (SSH), and salinity as predictors. Findings indicate VMS offers broader positional coverage but suffers from behavioral ambiguity, whereas logbook data, though spatially limited, provides higher accuracy due to direct catch reporting. Model evaluations showed comparable performance: the VMS-derived model achieved an AUC of 0.760 and an F1-score of 0.658, while the logbook-derived model yielded an AUC of 0.742 and an F1-score of 0.624. However, distribution analysis revealed the logbook-derived model performed better, with 87.5% of fishing events occurring in higher Habitat Suitability Index (HSI) areas compared to 73.1% for the VMS-derived model. These results suggest VMS data presents a viable alternative and comparative data source to logbook records for habitat modeling, offering opportunities to enhance fisheries management.