Search for a command to run...
Abstract Aim Long‐term monitoring of biodiversity is necessary to identify population declines and to develop conservation management. Because long‐term monitoring is labour‐intensive, resources to implement robust monitoring programmes are lacking in many countries. The increasing availability of citizen science data in online public databases can potentially fill gaps in structured monitoring programmes, but only if trends estimated from unstructured citizen science data match those estimated from structured monitoring programmes. We therefore aimed to assess the correlation between trends estimated from structured and unstructured data. Location Denmark. Methods We compared population trends for 103 bird species estimated over 28 years from a structured monitoring programme and from unstructured citizen science data to assess whether trends estimated from the two data sources were correlated. Results Trends estimated from the two data sources were generally positively correlated, but less than half the population declines identified from the structured monitoring data were recovered from the unstructured citizen science data. The mismatch persisted when we reduced the structured monitoring data from count data to occurrence data to mimic the information content of unstructured citizen science data and when we filtered the unstructured data to reduce the number of incomplete lists reported. Mismatching trends were especially prevalent for the most common species. Worryingly, more than half the species showing significant declines in the structured monitoring showed significant positive trends in the citizen science data. Main conclusions We caution that unstructured citizen science databases cannot replace structured monitoring data because the former are less sensitive to population changes. Thus, unstructured data may not fulfil one of the most critical functions of structured monitoring programmes, namely to act as an early warning system that detects population declines.