Recovering the drivers of sampling bias in Bignonieae (Bignoniaceae) and identifying priority areas for new survey efforts

Identifying knowledge gaps and the potential biases and limitations of biological databases is essential for biogeographical research, to efficiently plan biodiversity surveys, and to accurately design conservation efforts. Here we describe the taxonomic, temporal, and spatial coverage of a comprehensive database mainly composed of data collected between 1963 and 2003 of the largest clade of Neotropical lianas, the tribe Bignonieae (Bignoniaceae). We assess the level of database completeness and propose new survey areas to fill knowledge gaps and optimize sampling coverage. The Bignonieae database includes 28,763 records representing 98% of the known species. The database covers 72% of the Neotropical region and includes data collected mainly during the last 40 years of the 20th century. Members of the tribe are conspicuous components of lowland forests, with most species showing narrow range sizes. The Amazon rainforest is the most under-sampled region and the area with the lowest sampling rate. On the other hand, the best sampled areas are scattered across Central America, the Peruvian and Bolivian Amazon, and selected Brazilian cities. Sampling rate across the geographical extent of Bignonieae was best predicted by the distance from cities. Collection effort is needed across the Neotropics so that a higher number of localities can be sampled, especially in the Amazon, where Bignonieae is centered. New surveys are urgently needed to maximize new species discoveries and to effectively design conservation plans that maximize biodiversity-rich regions facing increased threat.