Range maps and species richness patterns: errors of commission and estimates of uncertainty

Authors

  • Frank A. La Sorte,

  • Bradford A. Hawkins


F. A. La Sorte (flasorte@ucsd.edu), Dept of Fisheries and Wildlife Sciences, Univ. of Missouri-Columbia, Columbia, MO 65211, USA (present address: Div. of Biological Sciences, Univ. of California, San Diego, La Jolla, CA 92093, USA). – B. A. Hawkins, Dept of Ecology and Evolutionary Biology, Univ. of California, Irvine, CA 92697, USA.

Abstract

Range maps are often combined into “range overlap maps” to estimate spatial variation in species richness. Range maps are, in most cases, designed to represent a species’ maximum geographical extent and not patterns of occupancy within the range. As a consequence, range maps overestimate occupancy by presenting false occupancy (errors of commission) within the interior of the range. To assess the implications of errors of commission when developing and applying range overlap maps, we used neutral landscapes to simulate range maps and patterns of occupancy within ranges. We explored several scenarios based on combinations of six parameters defining biogeographical and cartographic factors typically encountered by investigators. Our results suggest that, in general, uncertainty is lowest when map resolutions are moderately fine, the majority of species have geographically restricted ranges, species occur throughout their range, patterns of occupancy within the range are not correlated among species, and geographically local and widespread species tend to occupy different regions. Several of these outcomes are associated with broad geographical extents, the scale at which range overlap maps are typically applied. Thus, under most circumstances, reasonably accurate and precise representation of species richness patterns can be achieved. However, these representations can be improved by enhancing occupancy data for widespread species – a primary source of uncertainty – and selecting a map resolution that captures relevant biological and environmental heterogeneity. Hence, by determining where a study is situated within the scenarios examined in our simulations, uncertainty and its sources and implications can be ascertained. With this knowledge, research goals, methods, and data sources can be reassessed and refined and, in the end, conclusions and recommendations can be qualified. Alternatively, unique regional, taxonomic, or cartographic factors could be included in future simulations to provide direct assessments of uncertainty.

Ancillary