Marine protected areas (MPAs) such as the Tasman Fracture Australian Marine Park (AMP), are widely used to assist conservation and resource management. Effective monitoring and management of these areas is underpinned by understanding the ecosystems they aim to protect. Knowledge of deep, mesophotic ecosystems has been limited to date. However, the increasing development of video-based surveying provides a quantitative sampling tool for previously inaccessible areas, with multibeam sonar mapping followed by baited remote underwater videos (BRUVs) becoming widely adopted as a typical survey method.
This study uses BRUV imagery from a series of 2021 deployments in the Tasman Fracture AMP to build upon a 2015 study of the region to quantify changes in abundance, length, and biomass within demersal fish assemblages in the AMP. It aims to assess protection effects on demersal reef fish, further develop understanding of spatio-temporal variation in the fish communities, and identify drivers of this variation to aid monitoring of the reef system.
Spatially balanced sampling of 100 sites was undertaken using stereo-BRUVs in the Tasman Fracture AMP Marine National Park Zone (MNPZ) and adjacent fished areas. Impacts of protection alongside environmental drivers (rugosity and depth) on the distribution and composition of assemblages were assessed using a range of recognised indicator metrics at both community and species-specific levels. These metrics revealed no significant influence of the MNPZ, with implications that low fishing pressure in the region may prevent observation of significant park effect; however, indications of recovery in fished areas correlated with regionally decreased fishing effort over this period. Total demersal fish abundance was significantly related to depth and habitat complexity, with responses varying at a species-specific level. This study highlights the importance of understanding both environmental and external drivers within mesophotic ecosystems and how this knowledge may be used to better inform sampling designs.