1254394725

137

Table 5.3 Model selection testing for the eflfect of mean size-independent body mass (M,), haematocrit (Het), basal metabolic ratę (BMR) and summit metabolic ratę (Msum) on the return ratę of both cohorts during winter. (W_Aicc: weight of the model, Par: number of parameters, Dev.: Deviance)

	Model	AlCc	AA! Cc	W/uCc	Ukellhood	Par	Dev. 1
1	CD(BMR-t-Msum) p(.)	63.6	0.0	0.49		4	55.1
2	0(Hct+BMR+Msum) p(.)	65.9	2.3	0.16	0.32	5	55.1
3	<D(M,+8MR+M$um) p(.)	65.9	2.3	0.16	0.32	5	55.1
4	<D(M,+Hct+BMR+Msum) p(.)	65.9	2.3	0.16	0.32	5	55.1
5	OlMj+M^+Hct+HcP+BMR+BMR^Msum+Msum^2) p(.)	68.9	5.2	0.04	0.07	7	53.2
6	<D(Hct+Msum) p(.)	76.2	12.6	0.00	0.00	4	67.6
7	(D(Msum) p(.)	76.5	12.9	0.00	0.00	3	70.2
8	<D(M,+Hct+Msum) p(.)	78.5	14.9	0.00	0.00	5	67.6
9	<D(M,+Msum) p(.)	78.7	15.1	0.00	0.00	4
10	<D(Msum+Msum^2) p(.)	78.7	15.1	0.00	0.00	4	70.2
11	<D(M.+M.^J) p(.)	80.7	17.1	0.00	0.00	4	72.1
12	(D(Hct+Hct^2) p(.)	87.5	23.9	0.00	0.00	4	79.0
13	O(-) P(-)	87.5	23.9	0.00	0.00	2	83.4
14	O(Hct) p(.)	88.0	24.3	0.00	0.00	3	81.6
15	<D(BMR) p(.)	89.4	25.8	0.00	0.00	3	83.1
16	<D(M,) p(.)	89.7	26.1	0.00	0.00	3	83.4
17	<D(MS+Hct) p(.)	90.0	26.3	0.00	0.00	4	81.4
18	<D(Hct+BMR) p(.)	90.1	26.5	0.00	0.00	4	81.6
19	<D(BMR+BMR^2) p(.)	91.4	27.8	0.00	0.00	4	82.8
20	<D(M,+8MR) p(.)	91.6		0.00	0.00	4	83.1
21	0(Ms+Hct+BMR) p(.)	92.2	28.6	0.00	0.00	5	81.3

mean Ć = 1.0

5.5.2 Among year survival

Among years, the return ratę did not depend on group (X² = 0.9, p = 0.3) but was influenced by the period (X² = 12.8, p<0.05). Based on the QAICc, the best model explaining our data was the model <D₍.)p_(f) (table 5.4). Hence, the apparent survival probability was high and constant among periods and ages (O = 0.96 ± 0.02) but the encounter probability varied over time. Specifically, encounter probability decreased between the end of a winter and the beginning of the next one (figurę 5.3).