Unlike any other food sector, aquaculture encompasses vast diversity in species, feeds, environments, intensities, and integration. This diversity is both a strength, supporting adaptive capacity and nutritional variety, and a challenge, as it renders environmental performance highly context dependent. Life Cycle Assessment (LCA) is widely used to quantify aquaculture’s impacts, yet many studies treat the sector as homogeneous or rely on averages that conceal differences. This risks misleading conclusions and poorly targeted policy. Our study addresses this gap by systematically assessing how aquaculture diversity shapes reported impacts and modelling practices in LCAs.
This study systematically reviewed 96 peer-reviewed aquaculture LCA studies, covering 93 models published over two decades. Using an eight-dimension diversity framework, systems were classified by feeding habits, taxa, production environment, integration level, intensity, infrastructure, water exchange, and exposure. Environmental impacts were converted to a consistent unit of one tonne of product and compared across diversity classes. We also examined how subsystem inclusion and modeller choices varied with system type.
Disaggregating results by diversity subclasses revealed extreme variation otherwise hidden by averages. Global warming potential ranged from 32 kg CO₂-eq in algae systems to nearly 14,000 kg CO₂-eq in aquaponics (Fig. 1), while marine eutrophication spanned from 0.05 kg N-eq in algae to 108 kg N-eq in recirculating aquaculture. Statistical tests confirmed significant differences across most diversity classes, particularly for global warming potential, acidification, and eutrophication. High-impact outliers such as RAS, IMTA, and aquaponics pushed sectoral means upward, despite these systems often being pointed as more sustainable, while extensive or nature-based systems consistently clustered at the low-impact end. Boundary analysis revealed that complex, high-tech systems often excluded infrastructure or downstream processes, whereas simpler systems were modelled more comprehensively, exaggerating their relative impacts. Collectively, these findings demonstrate that aquaculture sustainability cannot be generalised without accounting for system diversity.
These findings demonstrate that aquaculture diversity is a fundamental determinant of both environmental outcomes and how they are modelled. Robust, policy-relevant LCAs must therefore adopt diversity-aware frameworks, consistently include key subsystems, and transparently report modelling choices. Without this, averages and one-size-fits-all benchmarks risk misrepresenting sustainability patterns and undermining effective policy design.