Introduction This study investigates the size composition of Brachyplatystoma vaillantii specimens caught by the Amazon industrial fleet, identifying spatiotemporal patterns and assessing the vulnerability of juveniles to fishing. Methods The research was conducted at the Amazon River mouth, using data collected by onboard scientific observers between 2008 and 2011, including catch records and length sampling of goliath catfish. Statistical analysis employed spatiotemporal vector autoregressive mixed models, using the proportion of discarded fish as the response variable. Results The average discard rate was 22.7% in number and 9.4% in biomass. Statistical analysis revealed significant differences in discarded rates among fishing seasons (FS), with FS-C (2010–2011) showing the highest rates. Environmental and operational variables, such as river discharge, depth, trawl speed, and geographical location, significantly influenced discarded rates. River discharge had a positive linear effect on discards, whereas depth had a negative effect, indicating higher discarded rates in shallow waters. Spatial and temporal autocorrelation were significant, highlighting the importance of modeling these components to understand discard patterns. Discussion The analysis suggested that, despite spatial and temporal variability, fishing season and environmental conditions, such as depth and discharge, are key determinants of juvenile discard. The study underscores the need for more sophisticated models to capture the complex dynamics of discards and to support the identification of high-risk areas and seasons for juvenile catches, thereby supporting the design of more effective spatial and temporal fisheries regulations.
Trindade et al. (2026) studied this question.