Species distribution models can be highly sensitive to algorithm configuration

•Configuration settings in SDM algorithms can strongly affect modelled distribution.•SDM algorithms vary widely in their sensitivity to altered configuration settings.•Algorithm sensitivity to configuration settings shows some dependence on the species being modelled•CTA, SRE, ANN and FDA showed strong sensitivity to configuration settings, while MaxEnt and GLM showed less sensitivity.•It is important to carefully configure SDMs and document these settings for reliable provenance and scientific robustness.