Table 1.

Final model selection, where “age” refers to the age class (juvenile vs. adult and yearling), “ay” and “j” are used for models with partial interactions, “wint” refers to winter temperatures, “marm” refers to March maximum temperature, and “snow” refers to snow depth. Number of parameters (np), adjusted AIC (QAICc), change in QAICc relative to the top model (ΔQAICc), weight of each model (Wt), deviance (QDeviance), and relative reduction in deviance (R) are also presented. We present the parameterization associated with the survival probability ϕ while the best parameterization for p is maintained

Model	Np	QAICc	ΔQAICc	Wt	QDeviance	R
~age * wint	5	1782.88	0.00	0.29	788.49	0.277
~ay +j + j:wint	4	1783.42	0.53	0.22	791.03	0.172
~j + ay + ay:wint	4	1785.09	2.20	0.10	792.70	0.102
~j + ay + ay:marm	4	1785.32	2.44	0.09	792.93	0.092
~age	3	1785.54	2.66	0.08	795.16	0.000
~j + ay + ay:snow	4	1785.60	2.72	0.07	793.21	0.081
~ay + j + j:snow	4	1786.79	3.90	0.04	794.40	0.032
~age * snow	5	1786.86	6.29	0.04	792.46	0.112
~age * marm	5	1787.07	7.48	0.04	792.67	0.103
~ay + j + j:marm	4	1787.31	7.55	0.03	794.92	0.010
~1	2	1799.84	7.76	0.00	811.46	NA