Ecological Archives E093-171-A7

Alexandre Robert, Vítor H. Paiva, Mark Bolton, Frédéric Jiguet, and Joël Bried. 2012. The interaction between reproductive cost and individual quality is mediated by oceanic conditions in a long-lived bird. Ecology 93:1944–1952. http://dx.doi.org/10.1890/11-1840.1

Appendix G. Details of comparisons among final capture–recapture models (four-state models) in the Monteiro's Storm-Petrel (Oceanodroma monteiroi) population monitored at Praia Islet (Azores, Portugal) over the period 1993–2010.

Environmental and population covariates considered in best three-state models (see Appendix D) were used in models with four states (successful breeders, unsuccessful breeders, nonbreeders, and absent). The analysis brought strong evidence that time and covariate effects on breeders survival (Table G1) and transition (Table G2) were mostly due to temporal variance in unsuccessful breeder survival and unsuccessful breeder→nonbreeder transition.

Table G1. Model selection for survival parameters of the 'successful breeder' and 'unsuccessful breeder' states, with respect to year to year variations, breeding success in the previous year (BSlag) and sea surface temperatures within a 50 km radius from the colony (SST50, 2nd qtr.). Model selection was performed according to Akaike's criterion corrected for small sample size (AICc). ΔAICc is the difference in AICc between each model and the best model. The AICc weight (wi.) is the amount of evidence in favor of a model. Dev and n.p. are respectively the deviance and the number of estimated parameters for each model. In all presented models, survival rates of absent and nonbreeder are assumed constant, recapture rates vary additively between breeding states and years, and transition rates vary according to age and breeding state (see Table D1, Mod. A1 of Appendix D for detailed structure). * : parameters assumed equal among states.

Model
name
Successful
breeder
Unsuccessful
breeder
AICc ΔAICc wi. n.p. Dev.
Mod. 1 constant BSlag + SST50
(2nd qrt.)
4470.00 0.00 0.639 38 2631.29
Mod. 2 constant BSlag × SST50
(2nd qrt.)
4472.11 2.11 0.223 39 2631.29
Mod. 3 constant BSlag 4473.70 3.70 0.101 37 2637.10
Mod. 4 constant SST50
(2nd qrt.)
4475.75 5.75 0.036 37 2639.15
Mod. 5 BSlag time 4484.91 14.91 0 53 2614.26
Mod. 6 constant time 4485.54 15.54 0 51 2619.19
Mod. 7 constant constant 4485.68 15.68 0 35 2653.29
Mod. 8 time BSlag 4489.45 19.45 0 53 2618.79
Mod. 9 SST50
(2nd qrt.)
constant 4489.66 19.66 0 37 2653.06
Mod. 10 constant * constant * 4494.77 24.78 0 34 2664.48
Mod. 11 SST50
(2nd qrt.)
time 4495.27 25.27 0 53 2624.62
Mod. 12 BSlag constant 4495.32 25.32 0 37 2658.72
Mod. 13 time SST50
(2nd qrt.)
4503.91 33.91 0 53 2633.26
Mod. 14 time constant 4505.49 35.49 0 51 2639.13
Mod. 15 time time 4516.70 46.7 0 67 2615.61


Table G2. Model selection for transition parameters among breeding states (BS: successful breeder; BU: unsuccessful breeder; NB: nonbreeder), with respect to year to year variations, breeding success in the previous year (BSlag) and sea surface temperatures within a 1500 km radius from the colony (SST1500, 1st quarter). Model selection was performed according to Akaike's criterion corrected for small sample size (AICc). ΔAICc is the difference in AICc between each model and the best model. The AICc weight (wi.) is the amount of evidence in favor of a model. Dev and n.p. are respectively the deviance and the number of estimated parameters for each model. In all models survival rates are constrained as in Mod. 1 of Table G1. Year and covariate effects on transition parameters are only applied on transition rates of the older age class. In all models, long term transitions of individuals marked as chicks and adults are assumed equal. * : transitions assumed equal among states.

Model name Transition rates AICc ΔAICc wi. n.p. Dev.
ψUB→NB ψSB→NB ψSB→UB ψUB→SB ψNB→UB ψNB→SB
Mod. 1 BSlag × SST1500
(1st qrt.)
constant constant constant constant constant 4442.36 0.00 0.189 42 2595.20
Mod. 2 Bslag + SST1500
(1st qrt.)
constant constant constant constant constant 4443.25 0.89 0.122 41 2598.21
Mod. 3 SST1500
(1st qrt.)
constant constant constant constant constant 4447.84 5.48 0.012 40 2604.92
Mod. 4 BSlag constant constant constant constant constant 4462.23 19.87 0 40 2619.31
Mod. 5 time constant constant constant constant constant 4464.82 22.46 0 54 2592.02
Mod. 6 constant constant constant constant constant constant 4470.00 27.64 0 38 2631.30
Mod. 7 constant constant constant time constant constant 4478.58 36.22 0 54 2605.77
Mod. 8 constant constant constant constant time constant 4482.94 40.58 0 54 2610.13
Mod. 9   constant *   constant * constant constant constant constant 4487.00 44.64 0 37 2648.20
Mod. 10 constant constant time constant constant constant 4492.40 50.04 0 54 2619.59
Mod. 11 constant constant constant constant constant time 4497.04 54.68 0 54 2624.24
Mod. 12 constant time constant constant constant constant 4503.14 60.78 0 54 2630.34
Mod. 13 time time time time time time 4585.57 143.21 0 135 2527.71

[Back to E093-171]