打印

作者

来源

LANDSCAPE AND URBAN PLANNING,Vol.171,P.98-106

语言

英文

关键字

Carpobrotus sp.; Generalized additive model; Invasion species distribution model; Light detection and ranging; COME land cover map; SPECIES DISTRIBUTION MODELS; BIOLOGICAL INVASIONS; PROPAGULE PRESSURE; ALIEN PLANTS; HABITAT SUITABILITY; CARPOBROTUS-EDULI

作者单位

[Bazzichetto, Manuele; Carranza, Maria Laura] Univ Molise, EnviX Lab, Dipartimento Biosci & Terr, I-86090 Pesche, IS, Italy. [Malavasi, Marco; Bartak, Vojtech; Moudry, Vitezslav] Czech Univ Life Sci, Dept Appl Geoinformat & Spatial Planning, Fac Environm Sci, Kamycka 129, Prague 16521 6, Czech Republic. [Acosta, Alicia Teresa Rosario] Univ Roma Tre, Dipartimento Sci, Vle Marconi 446, I-00146 Rome, Italy. Bazzichetto, M (reprint author), Univ Molise, EnviX Lab, Dipartimento Biosci & Terr, I-86090 Pesche, IS, Italy. E-Mail: manuele.bazzichetto@gmail.com; malavasi@fzp.czu.cz; aliciateresarosario.acosta@uniroma3.it; moudry@fzp.czu.cz; carranza@unimol.it

摘要

Although many hypotheses explaining invasion success have been formulated, invasion drivers are usually tested in isolation. This work aims to analyze the combined influence of propagule pressure (P), abiotic (A) and biotic (B) factors (PAB) on determining the invasion process of an exotic plant taxon (Carpobrotus sp.) in Mediterranean coastal landscapes. Specifically, we used a binomial Generalized Additive Model for exploring the relation between the occurrence of the invasive species and a set of PAB proxy variables derived from high-resolution remote sensed imagery (LiDAR - Light Detection - and Ranging and orthophotos). We evaluated the predictive power of the model by computing the mean of the AUC scores obtained through a 5-fold cross-validation and visual inspection of the Hosmer-Lemeshow plot. The integrated PAB approach efficiently captured the different roles played by the drivers of invasion in affecting the presence of the species. Invasion does not proceed homogeneously across the coastal landscape, but is promoted wherever the combined action of the PAB factors is favorable for establishment of the invader. Moreover, the use of remotely sensed data allowed us to model the invader-landscape relationship on a large geographic extent and to highlight the coastal sectors that are most likely to be invaded in the future.