Bias in neotropical and reef biodiversity monitoring programs may prevent detecting changes in species diversity through time

Abstract

Biodiversity monitoring networks are essential to quantify changes in biodiversity over time. Despite the importance of monitoring, studies using these data fail to find the widely accepted negative human impact on biodiversity, raising concerns about potential sampling issues in monitoring programs. To quantify the magnitude of these biases and their implications for future studies, we investigated the spatial, temporal, and taxonomic distribution of biodiversity monitoring networks in the Neotropics. Our dataset spans 59 years (1961–2020) and includes data from 30 countries and terrestrial and marine ecosystems. To understand how the duration of the time series could impact future studies, we simulated communities under different levels of colonization and extinction. We then evaluated whether colonization and extinction could be reliably estimated from a short time series. Most networks cover specific geographical regions (overrepresentation in southeastern Brazil), are short (<10 years), and focus on plants and mammals on land and fishes on marine reefs. In addition, most studies (81 %) sample areas with no significant habitat changes over time, which is significantly different from a random sampling expectation. We found long-term data (>50 years) to be important to achieve greater accuracy in temporal analyses, but short time series would still produce unbiased measures of colonization and extinction. Because monitoring sites are primarily in areas with low human impact and are restricted taxonomically, future studies using these data are unlikely to detect most species extinctions. Biodiversity monitoring programs should aim for a more equitable geographic distribution of projects and diversify taxonomic and geographical representation.