Do revisitation studies overestimate local extinction?

Abstract

Revisitation studies use historical information from literature or museum collections to assess rates of local extinction. However, revisitation studies suffer from the bias that they can only detect decline or stasis relative to the number of historical sites, as newly colonized sites are not detected. This drawback can be avoided with complete resurveys of study areas. We used 100-year-old historical information on 99 mountain plants from a 174 km 2 area in Switzerland and performed a revisitation study and a complete resurvey. The resurvey was used to determine the magnitude of the bias of revisitation. In the revisitation study, we found an average loss of historical sites of plant species of 51.1% (SE = 3.4%). When sites newly observed in the resurvey were also considered and assumed to represent new colonizations, the average loss in sites declined to 26.8% (SE = 5.7%). However, if newly observed sites were treated as historically overlooked sites the loss of sites was only 45.7% (SE = 3.4%). Our results thus show that revisitation studies can overestimate local extinction, but that the corresponding bias depends on whether newly observed sites are considered as historically overlooked sites or as new colonizations.