Matching expert range maps with species distribution model predictions.

Area of Science:

• Ecology and Conservation Biology
• Biodiversity Informatics
• Geographic Information Systems (GIS)

Background:

• Expert-derived species range maps are vital for conservation planning.
• Existing expert maps often lack integration with habitat data, limiting analytical use.
• Internal range heterogeneity, like habitat associations, is typically ignored in map analyses.

Purpose of the Study:

• To develop a quantitative metric (expert score) for assessing expert map accuracy against species distribution models (SDMs).
• To evaluate the performance of expert range maps from two major North American sources (Glassberg and Scott).
• To identify factors influencing the predictive performance of expert range maps.

Main Methods:

• A quantitative metric (expert score) was devised to compare expert maps with SDM-derived habitat probability surfaces.
• Expert maps for 330 butterfly species from Glassberg (1999, 2001) and Scott (1986) were analyzed.
• Species-wise expert scores were computed, and map performance was correlated with map topology and species traits.

Main Results:

• Glassberg maps achieved significantly higher expert scores (0.61) than Scott maps (0.41), attributed to stricter range inclusion criteria.
• Expert map inaccuracies were primarily due to including unsuitable sites (commission errors), not excluding suitable ones (omission errors).
• Map topology, specifically map detail, was a stronger predictor of expert map performance than species characteristics like mobility.

Conclusions:

• Detailed expert range maps with finer topology show greater agreement with SDMs, minimizing both commission and omission errors.
• The developed expert score provides a valuable tool for evaluating and improving the quality of expert-based range maps.
• Conservation planning can benefit from using more detailed expert maps that better reflect habitat suitability and species distribution.