How Much of the Forest Do Citizen Scientists Actually Record? A Decade of iNaturalist Sampling in the Hyrcanian Forests

Citizen-science records increasingly inform biodiversity research and conservation planning, yet how much of a large forest they actually cover is rarely measured. We analyzed 1,700 research-grade iNaturalist records from the Hyrcanian Forest, northern Iran (2015–2025), on a 10 × 10 km grid, comparing observed coverage with null models. Annual records rose from 48 to 386, but occupied only 43.7% of forest cells, 41% fewer than an even spread would give, and 56% of cells had no record. Fifteen observers (top 10%) provided 69% of records. In a presence–background model with clus-ter-robust errors, only elevation was independently associated with record presence (odds ratio 0.92 per 100 m); distances to roads and settlements were not. Because steep terrain limits where people can walk and record, similar eleva-tion-related gaps may occur in other mountain forests, where record counts could overstate coverage. Models using such records should adjust for elevation bias.