Bycatch information complements the understanding of spatial distribution for commercially-important fish species

The distribution of most commercially-important marine species was estimated with target catch data, while bycatch information has often been neglected. In this study, we aim to investigate the effect of including bycatch data on species distribution modeling (SDM) and to determine if the improvement was because of sampling -related data characteristics or other gear-specific effects. We used logbook data from stick-held dip net and pole-and-line fisheries, which were collected in the East China Sea from September to November, 2009-2019. The stick-held dip net caught hairtail (Trichiurus spp.) as bycatch, while the pole-and-line fishery targeted hairtail. We conducted presence-only SDM using maximum entropy approach based on the bycatch, target, and integrated data, then compared the predictability among models. We found that models with integrated data had the best predictability for species distribution. However, in certain months, the bycatch or target models per-formed as well as the integrated models. The better performance of bycatch models may be because of the larger sample size, wider sampling range, and lower sampling bias; however, after ruling out these effects, the target model had a better distribution predictability, indicating other factors related to fishing method may influence model performance as well. Our finding supports the need to incorporate bycatch information for SDM when species spatial information is only recorded in fishery-dependent data. Data availability: Data will be shared on request to the corresponding author.