Journal of Biogeography
SUPPORTING INFORMATION
Combining trade data and niche modelling improves predictions of the origin and
distribution of non-native European populations of a globally invasive species
Laura Cardador, Martina Carrete, Belinda Gallardo, José L. Tella
APPENDIX S1
Table S1. Results of exploratory analyses to select minimum-maximum distances of
occurrence points at which pseudo-absences should be selected to increase model
fit. TSS, sensitivity (percentage of correctly classified presences) and specificity
(percentage of correctly classified pseudo-absences) according to Maxent models
are shown. Note that exploratory models were based on total occurrences of ringnecked parakeets in their native (Asia and Africa) and invasive (Europe) ranges.
Model performance was assessed by dividing the species occurrence data into
random training (70%) and test (30%) datasets. Maxent models were implemented
in library “biomod2” in R software. Default cutoff values of “biomod2” were used
to compute sensitivity and specificity scores. Distances selected are shown in bold.
Distance range (km)
TSS
Sensitivity
Specificity
0-250
0.487
79.23
69.57
0-500
0.553
83.30
71.83
10-500
0.548
78.00
76.90
100-250
0.652
85.54
79.57
100-500
0.736
87.58
85.87
100-750
0.652
85.54
79.57
Table S2. Characteristics of occurrence localities of the ring-necked parakeet in Africa,
Asia and Europe. Means and ranges (in parenthesis) of annual mean temperature (TMEAN),
temperature seasonality (TRANGE, standard deviation *100), maximum temperature of the
warmest month (TMAX), minimum temperature of the coldest month (TMIN), annual
precipitation (PAN), precipitation of the driest month (PMIN), precipitation seasonality
(PRANGE, coefficient of variation), altitude (ALT) and population density (POP) are
provided. Percentage of occurrences in different land-use types is also provided. Sample
size: 140 for Africa, 1 070 for Asia and 535 for Europe.
Variable
Africa
Asia
Europe
TMEAN (°C)
26.9 (14.3 - 29.7)
25.2 (-2.4 - 29.1)
11.8 (5-20.0)
TMAX (°C)
36.8 (7.2 - 42.4)
36.3 (11.9 - 43.0)
24.5 (17.3 - 38.7)
TMIN (°C)
16.3 (5.7 - 22.7)
13.3 (-17.7 - 23.4)
1.5 (-9.0 - 13.8)
TRANGE (°C)
191 (57 - 363)
359 (51 - 764)
530 (229 - 913)
PMEAN (mm)
859 (72 - 3,768)
1,388 (125 - 4,588)
682 (109 - 1,478)
3 (0 - 66)
8 (0 - 162)
33 (0 - 97)
PRANGE (mm)
122 (34 - 184)
108 (29 -163)
28 (8 - 100)
ALT (m)
223 (3 - 2,623)
382 (0 - 5,033)
157 (0 - 2,267)
864 (0 - 30,379)
918 (0 - 36,732)
758 (0 - 20,846)
PMIN (mm)
POP (number of people)
Land-use type (%)
Woodlands
34.9
31.4
14.4
Open lands
25.6
3.1
3.2
Crop lands
21.7
55.9
42.3
Urban
10.9
8.6
38.9
Others
7.0
1.1
1.2
Figure S1. Distribution of the ring-necked parakeet in the study area. Occurrences
according
to
(a)
information
used
in
this
study
(http://ebird.org/content/ebird/) and GBIF (http://data.gbif.org)
based
on
eBird
datasets, (b) CABI
(http://www.cabi.org/), (c) DAISIE (http://www.europe-aliens.org/) and (d) countries
where the species has been recorded according to ISSG (http://www.issg.org/database).
In grey, the native and invasive areas considered in this study; in red, occurrences or
countries classified as invasive; in blue, occurrences or countries classified as native. In
yellow data not classified as native/invasive or not used in this study.
Figure S2. Binary predictions of suitable habitat in Europe for the ring-necked parakeet
according to models calibrated in African and Asian ranges. Results of models based on
different techniques are shown: generalized additive models (GAM), generalized linear
models (GLM) and random forest (RF).
Figure S3. Deviance explained by pure and joint effects of environmental suitability,
propagule pressure and geographic origin, obtained from generalized linear models
(GLM) for occurrence rate of ring-necked parakeets in 25 European countries. Results
of deviance partitioning for GLMs without (a) and with (b) year of first importation as
control variable are shown.