Working with imperfect trawl survey data: the role of sampling effort in biodiversity estimates at different spatial scales

Abstract

Scientific trawl surveys are conducted worldwide for fisheries management and ecosystem assessments. However, maintaining constant sampling protocols is rarely achievable due to logistical and financial limitations. Here, we used a data set from the San Matías Gulf (Patagonia, Argentina) to evaluate how variation in sampling effort can bias estimates of species richness and occupancy. We also explored analytical approaches that improve the reliability of biodiversity indicators when standardized data sets are not available. Species incidence data from 7 surveys (2006-2024) were analysed using generalized additive models (GAMs) and species-accumulation and rarefaction/extrapolation curves to assess the effects that varying sampling effort had on species richness at local (haul) and regional (entire Gulf) scales. We also evaluated whether rare species were disproportionally affected by sampling effort reduction and whether there have been significant changes in species occupancy within the gulf. Our results reveal that local species richness estimates are largely explained by depth, longitude and sampling effort. At the regional scale, all surveys showed similar richness values at comparable effort. These findings suggest that the lower richness observed in the 2 most recent surveys is likely an artifact of reduced sampling effort, leading to the under-detection of rare species. However, asymptotic richness shows a constant decrease from 2018 onwards, which indicates that loss of biodiversity cannot be ruled out. Additionally, we detected changes in the occupancy of some species. Our results highlight the importance of explicitly modelling effort-related variables and provide a methodological framework applicable to biodiversity surveys facing sampling effort constraints.