Abstract
Blue whales in the Indian Ocean have been severely depleted by previous extensive commercial whaling. A good understanding of their spatio-temporal distribution is crucial for conservation. The songs of three blue whale acoustic populations - Antarctic blue whales (Balaenoptera musculus intermedia, ANT BW) and pygmy blue whales (B. musculus brevicauda) from the Southeast (SEIO PBW) and Southwest Indian Ocean (SWIO PBW) - were analyzed using 13 years of passive acoustic recordings from 10 sites in the southwest Indian Ocean. Although blue whale vocalizations comprise both songs and non-song calls (e.g., D-calls), the present study concentrates on the examination of songs. Generalized additive models (GAMs) were used to relate acoustic presence, measured by the number of positive minutes per day (averaged weekly), to environmental drivers such as sea surface temperature (SST), chlorophyll-a concentrations, and sea ice extent. These models allowed predictions of blue whale acoustic presence across the region. Empirical orthogonal functions (EOFs) were applied for dimensionality reduction to identify key habitats, including the Kerguelen Plateau and Madagascar Basin, which may serve as important feeding and resting zones based on acoustic presence and environmental data. Antarctic blue whales were predominantly detected in austral winter and spring, associated with lower SST and higher chlorophyll-a. In contrast, SEIO and SWIO pygmy blue whales were more frequent in summer and autumn, with some overlap suggesting ecological interactions. These findings lay the groundwork for targeted conservation efforts to protect critical blue whale habitats in a rapidly changing ocean.