Distribuição do golfinho-nariz-de-garrafa (Tursiops truncatus Montagu, 1821) e do golfinho-de-dentes-rugosos (Steno bredanensis Lesson, 1828) no estado do Espírito Santo, Brasil

Abstract

Understanding the environmental determinants of top predator distribution is essential for marine conservation. This study investigated the distribution patterns of Tursiops truncatus and Steno bredanensis in Espírito Santo, Brazil. Passive Acoustic Monitoring (PAM) was conducted between 2019 and 2022 aboard a 40-foot sailing vessel following zig-zag transects. A towed hydrophone recorded acoustic signals, which were analyzed in real-time using PAMGuard. Acoustic classification via Random Forest achieved an accuracy exceeding 89%. Species Distribution Models (SDMs) incorporated temperature, salinity, pH, depth, and distance from the coast. The models exhibited high predictive performance (AUC > 0.90), revealing non-random distribution patterns. For T. truncatus, depth and distance from the coast emerged as the primary predictors, corroborating its affinity for coastal environments. For S. bredanensis, a significant influence of physicochemical variables, such as pH and salinity, was observed, alongside a shift in predictor importance throughout the sampling period. The results indicate that depth acts as the main mechanism for niche segregation during sympatry events, suggesting vertical habitat partitioning to reduce interspecific competition. The temporal dynamics in variable hierarchy reinforce the ecological plasticity of these species in response to oceanographic fluctuations. In conclusion, the integration of acoustic methods and predictive modeling is effective for characterizing ecological niches and providing baseline data for cetacean management strategies in tropical ecosystems.