Aim: Species introductions have reshaped island faunas for the last 200 years, often threatening native biodiversity. Approximately equal numbers of native and alien species currently co-occur in the New Zealand avifauna, but they show distinct habitat use. Antagonistic interactions, habitat affinities and legacies of introduction history may concur to explain their segregation along habitat gradients. To investigate these processes, we explored how habitat, ecological traits and introduction history relate with the current composition of bird assemblages.

Location: New Zealand

Taxon: Birds

Methods: We analysed 917 bird point counts spread along habitat and elevation gradients in the Canterbury region, South Island, and related 10 ecological traits to landscape composition using a three-table ordination method known as ‘RLQ analysis’, accounting for spatial autocorrelation and phylogeny. We tested whether alien species’ positions in the RLQ were related to proxies of introduction history.

Results: Eighteen endemic, 11 native and 19 alien species were distributed along a gradient from forest to open-habitat assemblages, in relation to foraging mode, nesting site and body size. A second gradient segregated species between native and exotic forests according to territoriality, sedentarity and diet. Traits accounted for the separation of native and alien bird species in forests, but not in open habitats. Phylogenetic signals emerged from the separation of native and alien species by forest type, and spatial structures suggested a landscape-level, rather than regional or local determinism. These correlations were independent of introduction history, although open-habitat assemblages tended to host alien species introduced later in time. 

Main conclusions: Habitat type and resource availability explain the spatial partitioning of New Zealand bird assemblages between native and alien species more consistently than competitive exclusion. We conclude that trait-mediated ecological differences among species have likely played a predominant role in species’ segregation among landscapes, while maintaining endemic bird assemblages in native forests. 

All the necessary information with respect to the data is provided in the article. It is unwise to use these data without reading the article before, and we strongly encourage any user to get in touch with us. Raw bird and habitat data are provided for reproducibility, but any potential user should keep in mind that they are the property of a private institution, Scion Research, which has its own data policy and contracts. To comply with these privacy rules, species names have been blinded. Hence, the data provided here are sufficient to reproduce entirely the analyses performed in the article but not to perform fully novel research on this data set. 

Corresponding author: please contact Jean-Yves Barnagaud: jean-yves.barnagaud@ephe.psl.eu

#### Data for : Trait-habitat associations explain novel bird assemblages mixing native and alien species across New-Zealand landscapes ####

These datafiles and R scripts relate to : Barnagaud, JY, Brockerhoff,E., Mossion, R., Dufour, P., Pavoine, S., Deconchat, M. & Barbaro, L. 2021. Trait-habitat associations explain novel bird assemblages mixing native and alien species across New-Zealand landscapes. Diversity & Distributions. 

All the necessary information with respect to the data is provided in the article, but species names are blinded. It is unwise to use these data without reading the article before, and we strongly encourage any user to get in touch with us ahead. The original bird count and habitat data are supplied for reproducibility but their public use and diffusion is subject to the contractual rules and authorization of Scion Research, Christchurch, NZ. Please contact Eckehard Brockerhoff (eckehard.brockerhoff@gmail.com) or Jean-Yves Barnagaud (jean-yves.barnagaud@ephe.psl.eu) if you are willing to re-use these data. Only the raw data that are necessary to replicate the analyses of the above-mentioned article are provided here. They are not sufficient to engage into a fully new study. 

#### Authors ####

Jean-Yves Barnagaud (1)* – jean-yves.barnagaud@ephe.psl.eu (corresponding author)
ORCID: https://orcid.org/0000-0002-8438-7809
Eckehard G. Brockerhoff (2,3)*
Raphael Mossion (1)
Paul Dufour (4)
Sandrine Pavoine (5)
Marc Deconchat (6)
Luc Barbaro (5,6)
* equal contributions
 
#### Corresponding author ####

Please contact Jean-Yves Barnagaud : jean-yves.barnagaud@ephe.psl.eu
 
#### Usage notes ####

Species, variables and traits acronyms are defined in the article. The data files have been blinded to comply with Scion Research's data privacy rules.

#### List of files ####

## Folder “data” contains all the necessary data to replicate the analyses (except species' names which are blinded). 

traits_bird_NZ_bin_blinded.txt and Q_blinded.txt : these two matrices contain bird trait data with different formatting. The two files originate from Q.txt but are provided twice to facilitate the run of R codes. Authors : Jean-Yves Barnagaud & Raphaël Mossion.

species-status_blinded.txt : species status and biome of origin (compilation from literature sources). Author : Jean-Yves Barnagaud

R.txt : Habitat on point counts. The first column contains point count indices for correspondence with L.txt and geographic_coordinates.txt. Author : Eckehard Brockerhoff. Note : these data are owned by a private company, Scion Research, which has its own data policy. They should not be reused for new studies without their permission, to be applied for through E. Brockerhoff.

NZBirds_Prum_blinded.tre : Phylogenetic tree for the species considered in the study. Author : Paul Dufour (paul.dufour80@gmail.com)

L_blinded.txt : Bird counts per point. Author : Eckehard Brockerhoff. Note : these data are owned by a private company, Scion Research, which has its own data policy.  Species' names are blinded. They should not be reused for new studies without their permission, to be applied for through E. Brockerhoff.

Introduction_effort_blinded.txt : data on the number of introduction events and introduction effort. Author: Raphaël Mossion. 

hab.txt: forest category at each point count. Author : Raphaël Mossion. These categories are derived from R.txt

geographic_coordinates.txt: geographic coordinates of point counts (EPSG27200-NZGD49). Author : Eckehard Brockerhoff. Note : these data are owned by a private company, Scion Research, which has its own data policy. They should not be reused for new studies without their permission, to be applied for through E. Brockerhoff

## Folder “scripts” contains all the necessary data to replicate the analyses. 
Barnagaud_et_al_div_dis_2021_blinded.R : R script to replicate all the analyses described in the article. Authors : Raphaël Mossion and Jean-Yves Barnagaud
functions_Barnagaud_DD_2021.R : additional functions adapted from Pavoine et al, 2011. Authors : Raphaël Mossion and Sandrine Pavoine.