Integrating Empirical Dynamic Modeling (EDM) into the Fisheries Integrated Modeling System (FIMS)

This project proposes the integration of Empirical Dynamic Modeling (EDM) into the Fisheries Integrated Modeling System (FIMS), an open-source framework used for fisheries stock assessment. While FIMS currently supports parametric models such as catch-at-age and surplus production models, ecological systems often exhibit nonlinear dynamics that are difficult to capture with predefined equations. EDM offers a data-driven alternative that reconstructs system dynamics directly from time-series observations using delay embedding techniques. The project will implement core EDM components within FIMS, including delay-embedding generation and prediction algorithms such as Simplex projection, S-map, and Gaussian Process EDM. These methods will be integrated with the existing FIMS statistical inference framework to enable empirical forecasting and hybrid modeling approaches. The final outcome will be a modular EDM module, tested using GoogleTest and testthat, along with documentation and example workflows to demonstrate its application in fisheries modeling.