The Study
This study presents a systematic review of the development and use of quantitative models applied to Chilean salmon aquaculture, based on the analysis of nearly 100 scientific studies selected from an initial pool of 1,454 identified documents, covering the period from 1990 to 2025. The objective was to characterize the types of numerical models used, their fields of application, accessibility for decision making , state and forcing variables, and their potential to support health, environmental, genetic, and production management in Chile’s salmon sector.
Main Findings
The greatest development of predictive models is concentrated in disease epidemiology (50%), followed by environmental impact assessment (13.6%), resistance genetics (10.6%), and bioeconomic analyses (4.5%). However, the review detected limited external validation in real-world decision-making contexts and a scarce incorporation of social or territorial variables. Most models are not, or do not report being integrated into management platforms and lack operational versions for use in the industry.
Projections
The study identifies opportunities to advance toward open, interoperable, and collaborative models through the development of digital public goods, integration with real-time monitoring systems, the use of hybrid modelling approaches (e.g., combining physics-based and machine learning methods), and their incorporation into integrated models of socioecological and production systems. It also proposes assessing the use of these models as professional training tools and strengthening priority lines such as climate change adaptation, animal welfare, and operational efficiency.
Impact or Value for the Industry
The models reviewed have the potential to become strategic assets for a more efficient, resilient, and sustainable salmon aquaculture industry. Their adoption could enable the anticipation of health risks, optimization of production decisions, reduction of environmental impacts, and improvement of traceability and crisis response. Achieving this requires stronger coordination between science and industry, software packaging, and specialized technical training to ensure effective use in real farming environments.