Predicting species distributions from small numbers of occurrence records: A test case using cryptic geckos in Madagascar

摘要: Aim: Techniques that predict species potential distributions by combining observed occurrence records with environmental variables show much for application across a range of biogeographical analyses. Some the most promising applications relate to which are scarce, due cryptic habits, locally restricted or low sampling effort. However, minimum sample sizes required yield useful predictions remain difficult determine. Here we developed and tested novel jackknife validation approach assess ability when fewer than 25 available. Location: Madagascar. Methods: Models were evaluated 13 secretive leaf-tailed geckos (Uroplatus spp.) endemic Madagascar, available from 4 23 localities (at 1 km2 grid resolution). Predictions based on 20 data layers generated using two modelling approaches: method principle maximum entropy (Maxent) genetic algorithm (GARP). Results: We found high success rates statistical significance in tests as five Maxent model was applied. Results GARP at very (less c. 10) less good. When experimentally reduced those records, variability among different combinations demonstrated models greatly influenced exactly observations included. Main conclusions: emphasize this small should be interpreted identifying regions have similar conditions where is known occur, not predicting actual limits species. The proposed here enables assessment predictive built sizes, although use test larger may lead overoptimistic estimates power. Our analyses demonstrate geographical numbers great value, example targeting field surveys accelerate discovery unknown populations © 2007 Authors.