1
High survival during hibernation affects onset and timing of reproduction
Claudia Bieber, Rimvydas Juškaitis, Christopher Turbill, Thomas Ruf
Electronic Supplemental Material (ESM)
Online Resource 2
Modelling Recapture Probability
Our recapture probability modelling revealed that interactions between the parameters age, season (ts), and sex fitted our data best. However, since including the factor birth seems also a possible explanation to describe our data (ΔQAICc < 2) we used the models from rank one to four (table A, bold printed) for further modeling of survival probability.
Table A. Models of local recapture probabilities in common dormice. Models were ranked by QAICc (quasi-likelihood corrected AIC for small sample size). Estimation of the survival parameters was fixed to the global model ((age*ts*sex*birth)) and is therefore not shown in the table. No. = model rank, Δ QAICc = difference between model QAICc and minimum QAICc, QAIC weight= relative strength of evidence for a model within the set of models computed, npar = number of parameters, Dev. = Deviance (total deviation between the computed model and a saturated model of the data). The first four models (bold printed) were chosen for further modelling procedures of .
No. / Survivalparameters / QAICc / ΔQAICc / QAIC
weight / npar / Dev.
1 / p(age*ts*sex) / 1948.95 / 0.00 / 0.28 / 36 / 967.80
2 / p(age*birth*ts*sex) / 1949.03 / 0.08 / 0.26 / 48 / 941.99
3 / p(age*birth*ts+sex) / 1949.95 / 1.00 / 0.17 / 37 / 966.66
4 / p(age*ts*sex+birth) / 1949.98 / 1.03 / 0.16 / 37 / 966.69
5 / p(age*birth*sex+ts) / 1951.54 / 2.58 / 0.08 / 34 / 974.64
6 / p(age*birth*ts) / 1953.53 / 4.58 / 0.03 / 36 / 972.38
7 / p(age*birth+ts+sex) / 1956.80 / 7.44 / 0.00 / 31 / 985.86
Parameter abbreviation: ts = time interval “season” (constant monthly estimates for early active season (May-August), late active season (August-October) and winter (hibernation, October-May).