Constraints on interpretation of ecological niche models by limited environmental ranges on calibration areas

Potential for spread of the white-nose fungus (Pseudogymnoascus destructans) in the Americas: use of Maxent and NicheA to assure strict model transference

“…As could be appreciated in NicheA visualisations, different continental study areas differed markedly in size, shape and position in environmental space, such that we considered continents individually and together in generating projections. Consequently, predictions of potential spread based solely on the plausible native range in Europe do not permit rigorous delineation of the niche or the invasive potential of the species; this situation limits the completeness and rigor of model projections to North America (Owens et al, 2013) and led us to calibrate models based on both northern continents for transfer to South America as a best means of mapping invasive potential of P. destructans there with least potential for extrapolative situations.…”

“…Considering that the current distribution of P. destructans in North America and Europe is entirely north of latitude 23.5°N (the Tropic of Cancer). In order to avoid projection onto entirely novel conditions (Elith et al, 2010;Owens et al, 2013), we explored the invasive potential of the agent in continental areas south of latitude 23.5°S (the Tropic of Capricorn) including portions of Argentina, Brazil, Chile, Paraguay, and Uruguay. We developed and explored six maps representing the following niche model projections: North America, Europe, North America and Europe pooled, North America projected to South America, Europe projected to South America and North America and Europe pooled projected to South America.…”

“…We used these minimum-volume ellipsoids under the assumption that fundamental ecological niches should be be convex taking the general form of ellipsoids (Soberón and Nakamura, 2009). Projecting Maxent models using NicheA indentifies environments falling inside ellipsoids representative of species' niches under some circumstances providing biologically-realistic transfers, even when environments differ from those in the calibration area (Owens et al, 2013;Qiao et al, 2013). Models were transferred to geographical space to explore potential suitable areas for P. destructans in South America.…”

“…Similarity to environments between calibration areas and novel areas for projection (e.g. South America) was assessed using mobility-oriented parity (MOP) according to Owens et al (2013), which refines the earlier multivariate environmental similarity surface (MESS) metrics implemented within Maxent for identifying areas of simple transfer (i.e. interpretation is acceptable) versus areas where extrapolation is occurring and interpretation should involve considerable caution (Elith et al, 2010).…”

Combining environmental suitability and population abundances to evaluate the invasive potential of the tunicate Ciona intestinalis along the temperate South American coast

“…Non-analog environments in an invaded range represent those habitats outside the range of values considered to quantify the native range and so, correspond to environmental conditions that have not been experienced by the species before the invasion (Fitzpatrick & Hargrove, 2009). Therefore, conclusions about these areas must be taken with caution (Owens et al, 2013). Second, results of ecological niche models (ENMs) are usually expressed as quantitative suitability indexes or as probability of presence, which are not necessarily linked to population abundances, a key parameter for pest managers or conservation biologists (VanDerWal et al, 2009).…”