Spatial error analysis of species richness for a gap analysis map

Abstract

Variation in the distribution of species richness as a result of introduced errors of omission and commission in the Gap Analysis database for Oregon was evaluated using Monte Carlo simulations. Random errors, assumed to be independent of a species' distribution, and boundary errors, assumed to be dependent on the species' distribution, were simulated using ten rodent species. Error rates of omission and commission equal to 5 and 20 percent were used in the simulations. Indications are that predictions of species richness within a Gap Analysis database can be very sensitive to both types of errors with sensitivity to random error being much greater. Implications are that the inclusion of error modeling in applied GIS databases is critical to spatially explicit conservation recommendations.