Soaring over open waters: horizontal winds provide lift to soaring migrants in weak thermal conditions

Table 2 Fixed effects on sea crossing probability, soaring probability and soaring patters

Model	Predictor	Estimate	Std. Error	z value	p value	P.A. %	R²_c	R²_m
(a) Crossing initiation	(Intercept)	− 1.97	0.46	4.26	0.00	75	0.52	0.31
	Wind support	0.99	0.27	4.38	0.00
	Side wind	0.30	0.19	1.58	0.11
	Cloud cover	− 0.07	0.20	0.38	0.70
	Precipitation	− 0.47	0.30	1.57	0.12
(b) Soaring over sea	(Intercept)	0.01	0.23	0.05	0.96	83	0.68	0.49
	Wind support	1.51	0.35	4.36	0.00
	Side wind	1.65	0.48	3.44	0.00
	ΔT	1.97	0.45	4.42	0.00
	Wind support: ΔT	1.65	0.27	6.06	0.00
(c) Soaring over land	(Intercept)	0.15	0.08	1.87	0.06	81	0.18	0.10
	poly(Wind support,2)1	0.01	0.09	0.07	0.94
	poly(Wind support,2)2	− 0.27	0.10	− 2.68	0.01
	Side wind	− 0.15	0.09	− 1.71	0.09
	ΔT	0.35	0.09	3.85	0.00
	Wind support: ΔT	− 0.15	0.13	− 1.19	0.23
(d) Soaring pattern over sea	(Intercept)	− 2.79	0.77	− 3.60	0.00	80	0.72	0.54
	Wind support	3.42	1.47	2.31	0.02
	Side wind	3.18	2.94	1.30	0.19
	ΔT	3.82	2.61	1.21	0.22
(e) Soaring pattern over land	(Intercept)	0.61	0.11	5.37	0.00	68	0.26	0.12
	poly(Wind support,2)1	− 0.60	0.13	− 4.61	0.00
	poly(Wind support,2)2	− 0.80	0.14	− 5.63	0.00
	Side wind	− 0.24	0.12	− 2.03	0.04
	ΔT	0.14	0.12	1.19	0.28

Fixed effects on sea crossing probability, soaring probability and soaring pattern as estimated by averaging our best GLMMs (ΔAIC or ΔAICc < 2) with individual ID included as a random factor (except d and e)
Significant results are highlighted in bold
GLMM generalized linear mixed model, ID identification number, ΔT difference between surface and air temperatures. P.A. predictive accuracy, R²_c conditional variance explained, R²_m marginal variance explained

ISSN: 2051-3933