With advances in global biodiversity data sharing, particularly following the Census of Marine Life, understanding of marine biodiversity has improved but remains incomplete. The Ocean Biodiversity Information System and Global Biodiversity Information Facility host over 150 million marine occurrence records, enabling reassessment of global biodiversity and data gaps. Here, we compile a quality-controlled dataset of ca. 48 million records covering 184,141 marine animal species, representing ~87% of accepted World Register of Marine Species and 91% of Ocean Biodiversity Information System taxa. Generalised Linear and Additive Models assess how geoecological drivers and human impact influence species richness while accounting for sampling effort and spatial autocorrelation across depth and taxa. Approximately 50% of the global ocean remains insufficiently sampled, with more than 160 million km² below 200 m lacking data. Sampling is biased toward developed regions, especially the North Atlantic, with major gaps in equatorial and Global South regions. Central tropical areas ( - 5° to 5°) contribute only <2.5% of global records, helping explain non-significant bimodal latitudinal patterns. Shallow-water richness is mainly associated with temperature, while deep-sea patterns relate to human impact (sampling intensity) and nitrate-driven remineralisation. These results highlight major global data gaps and the need for depth-explicit, bias-aware biodiversity assessment and monitoring to support conservation and the UN Ocean Decade.
🌊
