Temporally-scheduled ROV-based monitoring to detect behavioural rhythms of deep-sea megafauna

Abstract Animal activity rhythms, especially expressed as periodic displacement in the case of motile megafauna, can affect the detection of species in a certain spot during a short sampling window, and thus affect our perception of local biodiversity. However, this temporal aspect of animal behavior is rarely included in the data collection strategies of ecological monitoring programs, potentially leading to biased scientific outcomes and/or management decisions. Here, we innovatively performed high-frequency Remotely Operated Vehicle video surveys to assess the effects of benthic species’ rhythmic displacement on the assessment of their densities and overall biodiversity indicators inside and outside (as control) a deep Mediterranean No-Take Zone, a Marine Protected Area where all fishery activity is banned year-round. These depths are hosting several targets of commercial interest for the local fisheries, including the Norway lobster Nephrops norvegicus . We used Bayesian Hierarchical Clustering on the waveform data of 22 identified taxa (18 species and 4 genera) and identified 6 groups of rhythmic behavior: nocturnal, crepuscular, diurnal (with N. norvegicus as a special, singular case), bimodal and arrhythmic. Species accumulation curves and Pielou´s J´ index showed that biodiversity indices of different orders (i.e., both richness and evenness) varied in relation to the time of sampling during the 24-h cycle (i.e., daytime, nighttime and crepuscular hours). Our work showcases the need for temporally structured data collection strategies in deep-sea ecological monitoring programs, and provides further evidence towards the integration of robotic technologies in Ecosystem-Based Management approaches to increase sampling capabilities while reducing operational costs.