Table 2.

Summary of the (generalized) linear mixed models (#1a–#5) used in this manuscript, with reference to the variables described in Table 1

Model	Response	Fixed	Random	Model type	Subset	n
Reuse level and spatiotemporal variation
#1a	Reuse	Area + Female.age	Territory + Year.f + Year.Area	Binomial	All	1334
#1b	Reuse	Area + Male.age	Territory + Year.f + Year.Area	Binomial	All	469
#1c	Reuse	Replacement	Territory + Year.f	Binomial	DK	1129
#2	Laying.day	Reuse + Replacement + Year.c	Territory + Year.f	Gaussian	DK	902
Consequences by nest reuse
#3a	Nest.predated	0 + Area + Area:Reuse + Year.c	Territory + Year.f + Year.Area	Binomial	All	1257
#3b	Nest.predated	Reuse + Detect.day	Territory + Year.f	Binomial	OS	29
#4	Clutch.size.f	Reuse + Replacement + Laying.day	Territory + Year.f	Ordinal	DK	736
#5	N.young/Clutch.size	Reuse + Replacement + Laying.day	Territory + Year.f	Binomial	DK	736

The model structures are described in terms of the response variables used (Response), explanatory variables with fixed effect (Fixed) and random effects on the intercept (Random), type of model (Model type; Binomial—logit link, binomial error; Gaussian—identity link, Gaussian error; Ordinal—logit link, multinomial error), the subset of data used (Subset; All—all data, DK—Denmark, OS—Oslo) and the total sample size in the analysis (n). The models are divided into those describing patterns of reuse and timing of breeding (models #1a–#2), and those investigating the consequences to breeding performance by reuse (models #3a–#5). We centralized all numeric and binary explanatory variables (in column fixed) prior to the analyses