Road counts expose widespread declines in South African raptors underestimated by atlas data

Checklist-based citizen science surveys are becoming increasingly popular, as they may provide a means to monitor species' distributions and population trends. However, such checklist-based data may be limited in their ability to accurately quantify trends in abundance at large spatial scales and across multiple species. In this study, we (i) explore the population trends of 18 raptors and another eight large terrestrial bird species in central South Africa, using 752 road count transects (totalling 391,789 km), conducted from 2009 to 2025. We then, (ii) compare our road count trends with trend reporting rate estimates derived from a checklist-based survey, the Southern African Bird Atlas Project (SABAP2), for the same species, within the same core area, and over the same time period. Road counts showed that 50% of the species (13 of 26) declined significantly, while only three species (12%) showed significant increases. The remaining ten species (38%) showed no significant trends. In contrast, SABAP2 data showed the opposite pattern with only 13% of the species (3 of 24) significantly declining and 11 species (46%) significantly increasing. The remaining ten species showed no significant trends. Comparing the direction of trends (regardless of significance) between the road counts and the SABAP2 trends, we found that only around half the trends agreed. Among the contrasting trends, almost all showed declining trends from road counts, but increasing trends from SABAP2. Our findings support the hypothesis that using changes in reporting rates from atlas data may not accurately capture abundance trends at large spatial scales and across multiple species. Beyond this comparison, the declining trends observed from road counts, raise concerns over the conservation status of multiple species in a region which is known to host important raptor species' populations.