Building the foundation for efficient monitoring of mesophotic fishes to support management of the uThukela Banks marine protected area

Abstract

Long-term monitoring (LTM) is essential to assess the ecological benefits of marine protected areas (MPAs), and baseline data are required as a foundation of an effective monitoring programme. However, to be successful, LTM programmes must be feasible, in terms of costs and resources, and rigorous, to ensure statistical trends are detectable. The uThukela Banks MPA is a relatively new (established in 2019) large conservation area on the North-East coast of KwaZulu-Natal, South Africa. uThukela Banks MPA protects important demersal fisheries species found on mesophotic reef systems in the mid- and outer-continental shelf (30-150 m depth). Mesophotic reefs are particularly important for larger sexually mature fish that undertake ontogenetic habitat shifts from shallow nursery grounds. Few studies, especially in the context of South Africa, have investigated fish population and community structure on mesophotic reefs or on method optimization for LTM of mesophotic reef ecosystems. As such, this study aimed to provide a description of the fish communities and propose a feasible and robust LTM programme for the mesophotic reef fishes in the uThukela Banks MPA. To do this, 11 mesophotic reefs were mapped with multi-beam echo-sounder and fish were surveyed with baited remote underwater stereo-video systems (stereo-BRUVs) following a stratified random sampling approach, informed by the reef maps. Fish assemblage structure on the mesophotic reefs was significantly influenced by depth, benthic habitat type and habitat complexity. Reef habitats hosted higher species richness and abundance than adjacent sand habitats. Slinger Chrysoblephus puniceus abundance was greatest on moderate-depth reefs with higher structural complexity. Multivariate analyses identified distinct fish assemblages among the management zones, with restricted and controlled-pelagic zones supporting different assemblages than adjacent fished sites. However, this was driven by habitat and depth gradients rather than management measures. Species accumulation curves, power analysis and cost-benefit analysis were used to identify the sampling parameters for a feasible and robust LTM programme. Species accumulation curves had very high variability with between 8 and 38 samples required to detect 90 % of the observed fish species. The power analysis showed that 36 samples per location are required to detect a 10 % annual increase in slinger abundance over five years. Sample size requirements to detect the same changes in the abundance of all fish, biomass of all fish and abundance of species caught by commercial fisheries (sexually mature fisheries targets) was 16, 28, and 32, respectively. The results from the cost-benefit analysis showed that the costs for collecting stereo-BRUVs data per reef was R42 450 including boat fuel, bait, consumables, skipper, accommodation etc. The cost per stereo-BRUVs sample was R1 180.56 and the cost of surveying one mesophotic reef in each of the four zones of the uThukela Banks MPA and an adjacent fished reef site would be R212 500. A cost of R467 500 was estimated to survey the 11 reef systems originally included in this study. These results highlighted the importance of cost-considerations in an LTM programme. The priorities of a LTM programme must be refined and predetermined, because financial constraints tend to restrict the scope of monitoring. The study’s outcome is an optimally designed LTM that can be used by Ezemvelo KZN Wildlife (the management authority for the uThukela Banks MPA). These comprehensive baseline data and their inclusion into a monitoring framework will help them to monitor changes in fish assemblages in the study area. The study also serves as a template that can be used to inform mesophotic fish population monitoring throughout the KwaZulu-Natal MPA network.