Statistical Analysis --- Landscape heterogeneity and environmental dynamics improve predictions of establishment success of colonising small founding populations

Authors:

Pili, Arman N. (Corresponding author) armannorciopili@gmail.com, arman.pili@monash.edu School of Biological Sciences, Faculty of Science, Monash University, Clayton 3800, Australia Macroecology, Institute of Biochemistry and Biology, University of Potsdam, D-14469 Potsdam, Germany ORCID: 0000-0002-3952-9732

Schumaker, Nathan H. US Environmental Protection Agency, Pacific Ecological Systems Division, Corvallis, OR, United States ORCID: 0000-0002-4331-825X

Camacho-Cervantes, Morelia Instituto de Ciencias del Mar y Limnología, Universidad Nacional Autonoma de Mexico, Mexico City 04510, Mexico ORCID: 0000-0002-7123-7924

Tingley, Reid School of Biological Sciences, Faculty of Science, Monash University, Clayton 3800, Australia ORCID: 0000-0002-7630-7434

Chapple, David G. School of Biological Sciences, Faculty of Science, Monash University, Clayton 3800, Australia ORCID: 0000-0002-7720-6280

Content of the github repo

The R code for the Statistical analysis of the study:

Pili, A. N., Schumaker, N. H., Camacho-Cervantes, M., Tingley, R., and Chapple, D. G. (2024). Landscape heterogeneity and environmental dynamics improve predictions of establishment success of colonising small founding populations. Evolutionary Applications.

Using a spatially explicit, temporally dynamic, mechanistic, individual-based simulator of a model alien invader, the cane toad (Rhinella marina), we simulated colonisation scenarios to investigate how (1) the number of founding individuals, (2) the number of dispersal events, (3) landscape’s spatial composition and configuration of habitats (“spatially heterogenous landscapes”), and (4) timing of arrival with regards to dynamic environmental conditions (“dynamic environmental conditions”) influence the establishment success of small founding populations. We analysed the dynamic effects of these predictors on establishment success using running-window logistic regression models.