Table 3.

AIC table with estimates of R ² for all linear mixed-effect models with <4 delta AIC from top model

Model description		Fixed effects						n	df	∆AICc	AICcwt	mR 2	cR 2
		Area	Elev	LakeD	Littoral	IceD	Snow
Landscape 1				6.49	−0.35			43	5	0.00	0.36	0.51	0.64
Landscape 1				6.51				43	4	0.94	0.22	0.45	0.64
Landscape 2			0.09	29.63				27	5	0.00	0.28	0.28	0.28
Landscape 2			0.04					27	4	2.24	0.09	0.08	0.29
Landscape 2					−0.23			27	4	2.85	0.07	0.04	0.27
Landscape 2				10.13				27	4	3.08	0.06	0.04	0.26
Dynamic 1				5.30	−0.56	−1.10		21	6	0.00	0.64	0.64	0.89
Dynamic 1				5.33		−1.05		21	5	1.17	0.36	0.49	0.89
Dynamic 2						−1.11		13	4	0.00	0.34	0.25	0.74
Dynamic 2							−2.33	13	4	2.44	0.10	0.18	0.63
Dynamic 2			0.07					13	4	2.72	0.09	0.21	0.21
Dynamic 2		−7.97E-06						13	4	3.81	0.05	0.14	0.20
Dynamic 2					−0.38			13	4	3.97	0.05	0.12	0.41

Models: Landscape 1 landscape model with all sites, Landscape 2 landscape model with only shallow (lake depth <4 m) sites, Dynamic 1 dynamic model with all sites that have ice on dates and snow depth, Dynamic 2 dynamic model with shallow sites (lake depth <4 m) that have ice on dates and snow depth. Fixed effects used in the model: Depth maximum lake depth, Area surface area (m²) of the lake using radar imagery, Littoral an estimate of percent lake littoral area (area <1.6 m deep) using radar imagery, LDP lake drainage position rank (values; 0 no connection–5 highly connected). Dynamic variables, IceD the number of days that the lake was covered in ice, Snow snow depth on top of the lake ice. Additional info; DF the degrees of freedom in model, AIC Akaike information criterion, ∆AIC change in AIC values, AICwt the relative likelihood of a model, mR ² variation explained by the fixed effects model, cR ² variation explained by the random effects model