The extent to which fluids may leak from sedimentary basins to the seabed is a critical issue for assessing the potential of a basin for carbon capture and storage. The Petrel Sub-basin, located beneath central and eastern Joseph Bonaparte Gulf in tropical northern Australia, was identified as potentially suitable for the geological storage of CO2 because of its geological characteristics and proximity to offshore gas and petroleum resources. In May 2012, a multidisciplinary marine survey (SOL5463) was undertaken to collect data in two targeted areas of the Petrel Sub-basin to facilitate an assessment of its CO2 storage potential. This paper focuses on Area 1 of that survey, a 471 km(2) area of sediment-starved shelf (water depths of 78 to 102 m), characterised by low-gradient plains, low-lying ridges, palaeo-channels and shallow pockmarks. Three pockmark types are recognised: small shallow unit pockmarks 10-20 m in diameter (generally <1 m, rarely to 2 m deep), composite pockmarks of 150-300 m diameter formed from the co-location of several crosscutting pockmarks forming a broad shallow depression (<1 m deep), and pockmark clusters comprised of shallow unit pockmarks co-located side by side (150-300 m width overall, <1 m deep). Pockmark distribution is non-random, focused within and adjacent to palaeo-channels, with pockmark clusters also located adjacent to ridges. Pockmark formation is constrained by AMS C-14 dating of in situ mangrove deposits and shells to have begun after 15.5 cal ka BP when a rapid marine transgression of Bonaparte Shelf associated with meltwater pulse 1A drowned coastal mangrove environments. Pockmark development is likely an ongoing process driven by fluid seepage at the seabed, and sourced from CO2 produced in the shallow sub-surface (<2 m) sediment. No evidence for direct connection to deeper features was observed. Crown Copyright (C) 2014 Published by Elsevier Ltd. All rights reserved.