An ANN-based non-destructive model for age and maturity estimation of the data-deficient Goldblotch Grouper ( <i>Epinephelus costae</i> ): a practical tool for data-poor fisheries in conflict-affected Mediterranean coasts

The Goldblotch Grouper ( Epinephelus costae ) is listed as “Data Deficient” by the IUCN, yet it faces intense fishing pressure across the Eastern Mediterranean. In Syria, more than a decade of conflict has dismantled fisheries monitoring infrastructure, leaving managers without essential biological parameters for evidence-based conservation. Traditional methods for age and maturity assessment are destructive, require specialized laboratories, and are infeasible in post-conflict settings. To address this gap, this study presents the first non-destructive artificial neural network (ANN) model to simultaneously estimate age and maturity status using total length as the sole input. The model was trained exclusively on 150 specimens collected from six Syrian coastal sites, Ras al-Bassit to Tartous, during a 12-month period from March 2024 to March 2025. Age was determined by scale reading, and maturity status was assessed by macroscopic examination of gonads after dissection, revealing a contemporary length-at-first-maturity (L m = 30 cm) in Syrian waters. The feedforward multilayer perceptron (MLP) architecture (1–10–2) achieved high predictive accuracy, with a Pearson’s correlation coefficient of R = 0.9995 for age estimation (MSE = 0.0091 on the test set), and 100% maturity classification accuracy on the current independent test subset. However, real-world performance may vary in other populations. Our dataset includes individuals up to 74.3 cm and 12 years old, which are substantially larger and older than those reported in recent studies from neighboring regions. This underscores the value of targeted sampling methods, such as speargun, in accessing cryptic, large-bodied spawners. This ANN-based tool provides the first rapid, non-destructive, and field-applicable method for estimating population structure of E. costae under data-poor, conflict-affected conditions. By requiring only a length measurement, it offers a potential pathway for real-time decision support in data-poor contexts, supporting science-informed management where traditional monitoring systems have collapsed.