Combined Effects of Global Climate Suitability and Regional Environmental Variables on the Distribution of an Invasive Marsh Species Spartina alterniflora

Invasion by Spartina alterniflora seriously threatens native ecosystem along Chinese coast. Determining the main influential factors and their relationships with the distribution of S. alterniflora is thus crucial for invasion control. However, the distribution is influenced by environmental variables at different scales and the relative importance of cross-scale variables is unclear. Based on the MaxEnt modelling technique, a combined regional environmental niche (CREN) model was built by integrating the global climate niche (GCN) model into the regional environment niche (REN) model to study the combined effects of global climate suitability and regional environmental variables on species distribution. The CREN model performed much better than the GCN model with AUC, TSS, specificity and sensitivity values increasing by 0.12, 0.04, 0.05 and 0.45, whereas it performed as well as the REN model, but reduced the overfitting. When considering the combined effects, the predicted suitable area decreased from 66.90% at the global scale to 18.53% at the regional scale. Global sensitivity analysis showed there were strong interactions among different variables, especially for elevation and global climate suitability, the most influential variables. Interactions reduced the importance of soil salinity, but enhanced that of soil percentage sand. The presence probability increased with increasing of global climate suitability and soil salinity, while decreased with increasing of elevation, soil organic carbon and percentage sand. The presence probability was the highest in moderately well drained and lowest in poorly drained soil. Ignoring the combined effects of cross-scale variables will prevent comprehensive elucidation of their relationship with species distribution, which should be considered to take effective measures against biological invasion.