Delimiting the coastal geographic background to predict potential distribution of Spartina alterniflora

Abstract

Ecological niche modeling is an important tool in studying biological invasion; however, the geographic background on niche model transferability received scant attention. The salt marsh grass Spartina alterniflora, natively distributed along the eastern coasts of the Americas, is considered a global invasive species. In this study, we first compared the climate space among geographically separated populations. The classic niche model approaches involving the calibration of native range climate envelop of S. alterniflora and transferring worldwide were then used to predict potential invasion. Niche models based on two geographic backgrounds were compared, namely a large squared area delimited by a bounding box containing all known occurrences, which is usually used in former studies and a small coastal area defined as the geographic space available to the species. Both area-based models showed good performance in native range predictions, however, when models were transferred, niche model calibrated on the coastal area showed higher predictability in capturing the introduced occurrences. Given the potential substantial effect of geographic background on niche model transferability, caution is warranted when interpreting low-niche model transferability with niche differentiation and when predicting other (coastal) species’ invasion.