Table 2. Model selection results for the subsets and full dataset examining the influence of anthropogenic noise, artificial night light and weather components on the onset of blackbird dawn song.

Subset	Candidate modelsa	df	K	AIC	ΔAIC	ωi
All data
	Noi, Ld, Day, Temp, Cloud, Habitat, Noi*Ld	2613	11	29917.7	0.00	0.43
	Noi, Ld, Day, Temp, Cloud, Habitat, Noi*Ld, Cloud*Ld	2612	12	29918.6	1.03	0.26
	Noi, Ld, Day, Temp, Cloud, Habitat, Moon, Noi*Ld,	2612	12	29919.3	1.61	0.19
	Noi, Ld, Day, Temp, Cloud, Habitat, Moon, Noi*Ld, Cloud*Ld	2611	13	29920.3	2.62	0.12
Centre
	Noi, Day, Temp, Cloud, Moon	891	9	9751.7	0.00	0.40
	Noi, Ld, Day, Temp, Cloud, Moon	890	10	9751.7	0.04	0.39
	Noi, Ld, Day, Temp, Cloud, Moon, Cloud*Ld	889	11	9753.6	1.91	0.15
	Noi, Ld, Day, Temp, Cloud, Moon, Noi*Ld, Cloud*Ld	888	12	9755.6	3.90	0.06
Park
	Noi, Ld, Day, Temp, Cloud, Noi*Ld, Cloud*Ld	1714	11	19798.8	0.00	0.47
	Noi, Ld, Day, Temp, Cloud, Moon, Noi*Ld, Cloud*Ld	1713	12	19799.1	0.34	0.40
	Noi, Ld, Day, Temp, Cloud, Noi*Ld	1715	10	19801.3	2.49	0.13

All models contain a correction factor for spatial autocorrelation and a random intercept to account for repeated measures of individual birds.

^aNoi – PCA factor noise, Ld – lamp density, Day – Julian date, Temp – temperature, Cloud – cloud cover, Moon – lunar phase; Habitat – habitat factor.

Model selection was based on Akaike’s information criteria (AIC). The parameter df indicates the residual degrees of freedom, K represents the number of parameters in the candidate model and ΔAIC the difference of AIC values to the top-ranking model. Only candidate models with ΔAIC <4 are presented, because they are considered to receive support, and included to calculate Akaike weights (ω_i).