Multi-scale use of cleaning stations by manta rays in Mozambique: Insights from remote underwater video surveillance

Cleaning stations play a critical ecological role within reef ecosystems by facilitating mutualistic interactions between manta rays (Mobula alfredi and M. birostris) and cleaner fish and have been suggested to contribute to the health and site fidelity of these threatened elasmobranchs. However, fine-scale behavioural data on manta ray cleaning ecology and their responses to human disturbance remain limited. This study used a long-life remote underwater video (LL-RUV) system to conduct non-invasive, long-term monitoring of manta ray behaviour at two cleaning stations near Praia do Tofo, Mozambique, between July 2021 and October 2024 (610 h of video across eight deployments, mean duration 64 h). The research aimed to: (1) characterise temporal patterns in manta ray visitation, (2) investigate cleaner-client interactions with an emphasis on body part specialisation and (3) assess fine-scale temporal overlap between reef and oceanic manta rays, and between manta rays and divers. A total of 174 manta ray sightings were recorded: 82 reef, 65 oceanic, and 27 unidentified manta rays, which were excluded from species-level comparisons. Both species showed clear temporal trends in visitation, with peaks from morning to midday and reduced activity in the afternoon, supported by a significant non-linear effect of time of day. Cleaning interactions involved three cleaner fish species, with preliminary observations suggesting spatial differences in how bluestreak cleaner wrasse (Labroides dimidiatus) and white-spotted butterflyfish (Chaetodon kleinii) engaged with different manta ray body regions. Interval-based analyses indicated that, despite overlapping diel visitation patterns, simultaneous use of cleaning stations by reef and oceanic manta rays was extremely rare. Manta ray presence was also uncommon within minutes of diver activity, with models indicating reduced probability of occurrence in temporal proximity to divers. These results highlight the value of LL-RUV for detecting fine-scale behavioural patterns relevant to the management of critical cleaning sites.